diff --git "a/trainer_state.json" "b/trainer_state.json" new file mode 100644--- /dev/null +++ "b/trainer_state.json" @@ -0,0 +1,38443 @@ +{ + "best_global_step": null, + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 14.951603498542275, + "eval_steps": 500, + "global_step": 1600, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3699.0, + "completions/mean_length": 610.0078125, + "completions/mean_terminated_length": 534.484619140625, + "completions/min_length": 22.0, + "completions/min_terminated_length": 22.0, + "epoch": 0.009329446064139942, + "grad_norm": 0.13819736242294312, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 555247.0, + "reward": 0.5345982313156128, + "reward_std": 0.26035600900650024, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 1 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3926.0, + "completions/mean_length": 596.0346069335938, + "completions/mean_terminated_length": 540.4796142578125, + "completions/min_length": 55.0, + "completions/min_terminated_length": 55.0, + "epoch": 0.018658892128279883, + "grad_norm": 0.1229914054274559, + "learning_rate": 1e-06, + "loss": -0.0022, + "num_tokens": 1119206.0, + "reward": 0.4776785969734192, + "reward_std": 0.21755866706371307, + "rewards/verify_math_reward/mean": 0.4776785671710968, + "rewards/verify_math_reward/std": 0.49978047609329224, + "step": 2 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3738.0, + "completions/mean_length": 596.0223388671875, + "completions/mean_terminated_length": 532.3863525390625, + "completions/min_length": 18.0, + "completions/min_terminated_length": 18.0, + "epoch": 0.027988338192419825, + "grad_norm": 0.13747826218605042, + "learning_rate": 1e-06, + "loss": -0.0121, + "num_tokens": 1686250.0, + "reward": 0.4988839626312256, + "reward_std": 0.23206350207328796, + "rewards/verify_math_reward/mean": 0.4988839328289032, + "rewards/verify_math_reward/std": 0.5002779960632324, + "step": 3 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3539.0, + "completions/mean_length": 564.5960083007812, + "completions/mean_terminated_length": 520.7028198242188, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "epoch": 0.037317784256559766, + "grad_norm": 0.16430562734603882, + "learning_rate": 1e-06, + "loss": 0.0092, + "num_tokens": 2226504.0, + "reward": 0.5457589626312256, + "reward_std": 0.24920079112052917, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 4 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 565.3895263671875, + "completions/mean_terminated_length": 545.576904296875, + "completions/min_length": 19.0, + "completions/min_terminated_length": 19.0, + "epoch": 0.04664723032069971, + "grad_norm": 0.16295573115348816, + "learning_rate": 1e-06, + "loss": 0.0166, + "num_tokens": 2796181.0, + "reward": 0.5055803656578064, + "reward_std": 0.28827616572380066, + "rewards/verify_math_reward/mean": 0.5055803656578064, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 5 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 603.372802734375, + "completions/mean_terminated_length": 551.952392578125, + "completions/min_length": 89.0, + "completions/min_terminated_length": 89.0, + "epoch": 0.05597667638483965, + "grad_norm": 0.14144107699394226, + "learning_rate": 1e-06, + "loss": 0.0155, + "num_tokens": 3373979.0, + "reward": 0.5111607313156128, + "reward_std": 0.25476565957069397, + "rewards/verify_math_reward/mean": 0.5111607313156128, + "rewards/verify_math_reward/std": 0.5001546144485474, + "step": 6 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2267.0, + "completions/mean_length": 557.7288208007812, + "completions/mean_terminated_length": 521.8274536132812, + "completions/min_length": 4.0, + "completions/min_terminated_length": 4.0, + "epoch": 0.0653061224489796, + "grad_norm": 0.15719428658485413, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 3935296.0, + "reward": 0.5323660969734192, + "reward_std": 0.2519795596599579, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 7 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3053.0, + "completions/mean_length": 619.966552734375, + "completions/mean_terminated_length": 568.7904663085938, + "completions/min_length": 57.0, + "completions/min_terminated_length": 57.0, + "epoch": 0.07463556851311953, + "grad_norm": 0.13995982706546783, + "learning_rate": 1e-06, + "loss": -0.012, + "num_tokens": 4511930.0, + "reward": 0.5078125, + "reward_std": 0.24746696650981903, + "rewards/verify_math_reward/mean": 0.5078125, + "rewards/verify_math_reward/std": 0.5002182126045227, + "step": 8 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3136.0, + "completions/mean_length": 615.185302734375, + "completions/mean_terminated_length": 567.9343872070312, + "completions/min_length": 69.0, + "completions/min_terminated_length": 69.0, + "epoch": 0.08396501457725948, + "grad_norm": 0.13150471448898315, + "learning_rate": 1e-06, + "loss": 0.0117, + "num_tokens": 5102960.0, + "reward": 0.5368303656578064, + "reward_std": 0.22883078455924988, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 9 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3174.0, + "completions/mean_length": 596.9308471679688, + "completions/mean_terminated_length": 565.4076538085938, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 0.09329446064139942, + "grad_norm": 0.12639360129833221, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 5700282.0, + "reward": 0.5189732313156128, + "reward_std": 0.23198682069778442, + "rewards/verify_math_reward/mean": 0.5189732313156128, + "rewards/verify_math_reward/std": 0.49991893768310547, + "step": 10 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3772.0, + "completions/mean_length": 646.614990234375, + "completions/mean_terminated_length": 579.9032592773438, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "epoch": 0.10262390670553936, + "grad_norm": 0.12482870370149612, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 6294361.0, + "reward": 0.4933035969734192, + "reward_std": 0.19501692056655884, + "rewards/verify_math_reward/mean": 0.4933035671710968, + "rewards/verify_math_reward/std": 0.5002344250679016, + "step": 11 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2617.0, + "completions/mean_length": 601.154052734375, + "completions/mean_terminated_length": 553.7127075195312, + "completions/min_length": 67.0, + "completions/min_terminated_length": 67.0, + "epoch": 0.1119533527696793, + "grad_norm": 0.12638172507286072, + "learning_rate": 1e-06, + "loss": -0.0039, + "num_tokens": 6886323.0, + "reward": 0.5412946939468384, + "reward_std": 0.20342837274074554, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 12 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2701.0, + "completions/mean_length": 635.5167846679688, + "completions/mean_terminated_length": 588.5418701171875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.12128279883381925, + "grad_norm": 0.11367672681808472, + "learning_rate": 1e-06, + "loss": 0.0062, + "num_tokens": 7494586.0, + "reward": 0.5546875, + "reward_std": 0.21763604879379272, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 13 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2779.0, + "completions/mean_length": 568.1730346679688, + "completions/mean_terminated_length": 528.3555297851562, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.1306122448979592, + "grad_norm": 0.13681843876838684, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 8058821.0, + "reward": 0.5479910969734192, + "reward_std": 0.23022836446762085, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 14 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2455.0, + "completions/mean_length": 616.7120971679688, + "completions/mean_terminated_length": 573.4666748046875, + "completions/min_length": 31.0, + "completions/min_terminated_length": 31.0, + "epoch": 0.13994169096209913, + "grad_norm": 0.12186138331890106, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 8659803.0, + "reward": 0.559151828289032, + "reward_std": 0.18036124110221863, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 15 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3734.0, + "completions/mean_length": 637.3917846679688, + "completions/mean_terminated_length": 562.4617919921875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 0.14927113702623906, + "grad_norm": 0.12683585286140442, + "learning_rate": 1e-06, + "loss": 0.0083, + "num_tokens": 9244546.0, + "reward": 0.5390625, + "reward_std": 0.20489898324012756, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 16 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2093.0, + "completions/mean_length": 592.6116333007812, + "completions/mean_terminated_length": 557.064208984375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 0.158600583090379, + "grad_norm": 0.1239178404211998, + "learning_rate": 1e-06, + "loss": 0.0051, + "num_tokens": 9824286.0, + "reward": 0.5625, + "reward_std": 0.22368666529655457, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 17 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3810.0, + "completions/mean_length": 679.9553833007812, + "completions/mean_terminated_length": 581.9058227539062, + "completions/min_length": 25.0, + "completions/min_terminated_length": 25.0, + "epoch": 0.16793002915451896, + "grad_norm": 0.12310385704040527, + "learning_rate": 1e-06, + "loss": 0.017, + "num_tokens": 10412358.0, + "reward": 0.5301339626312256, + "reward_std": 0.2187958061695099, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936985969543457, + "step": 18 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2945.0, + "completions/mean_length": 600.3370971679688, + "completions/mean_terminated_length": 528.6719970703125, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 0.1772594752186589, + "grad_norm": 0.14129288494586945, + "learning_rate": 1e-06, + "loss": -0.0168, + "num_tokens": 10970988.0, + "reward": 0.6082589626312256, + "reward_std": 0.25784680247306824, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 19 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3255.0, + "completions/mean_length": 574.2232666015625, + "completions/mean_terminated_length": 530.44970703125, + "completions/min_length": 5.0, + "completions/min_terminated_length": 5.0, + "epoch": 0.18658892128279883, + "grad_norm": 0.12139517813920975, + "learning_rate": 1e-06, + "loss": 0.0175, + "num_tokens": 11531604.0, + "reward": 0.6261160969734192, + "reward_std": 0.20846940577030182, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 20 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 586.8192138671875, + "completions/mean_terminated_length": 539.1832885742188, + "completions/min_length": 54.0, + "completions/min_terminated_length": 54.0, + "epoch": 0.19591836734693877, + "grad_norm": 0.12492037564516068, + "learning_rate": 1e-06, + "loss": 0.0156, + "num_tokens": 12090434.0, + "reward": 0.582589328289032, + "reward_std": 0.19298996031284332, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 21 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2767.0, + "completions/mean_length": 645.552490234375, + "completions/mean_terminated_length": 602.66552734375, + "completions/min_length": 75.0, + "completions/min_terminated_length": 75.0, + "epoch": 0.20524781341107873, + "grad_norm": 0.11325538158416748, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 12719225.0, + "reward": 0.5078125, + "reward_std": 0.21368838846683502, + "rewards/verify_math_reward/mean": 0.5078125, + "rewards/verify_math_reward/std": 0.5002182126045227, + "step": 22 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2971.0, + "completions/mean_length": 629.3482666015625, + "completions/mean_terminated_length": 558.2779541015625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.21457725947521866, + "grad_norm": 0.11964946985244751, + "learning_rate": 1e-06, + "loss": -0.0061, + "num_tokens": 13293729.0, + "reward": 0.652901828289032, + "reward_std": 0.18246740102767944, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 23 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3421.0, + "completions/mean_length": 614.9185791015625, + "completions/mean_terminated_length": 579.5974731445312, + "completions/min_length": 5.0, + "completions/min_terminated_length": 5.0, + "epoch": 0.2239067055393586, + "grad_norm": 0.1198398619890213, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 13892128.0, + "reward": 0.53125, + "reward_std": 0.20331665873527527, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 24 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3672.0, + "completions/mean_length": 670.3013916015625, + "completions/mean_terminated_length": 608.015869140625, + "completions/min_length": 75.0, + "completions/min_terminated_length": 75.0, + "epoch": 0.23323615160349853, + "grad_norm": 0.11495624482631683, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 14512414.0, + "reward": 0.543526828289032, + "reward_std": 0.179909810423851, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838003516197205, + "step": 25 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4095.0, + "completions/mean_length": 632.630615234375, + "completions/mean_terminated_length": 589.5830688476562, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 0.2425655976676385, + "grad_norm": 0.12420367449522018, + "learning_rate": 1e-06, + "loss": 0.0135, + "num_tokens": 15121667.0, + "reward": 0.4977678656578064, + "reward_std": 0.20959801971912384, + "rewards/verify_math_reward/mean": 0.4977678656578064, + "rewards/verify_math_reward/std": 0.5002742409706116, + "step": 26 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 618.3236694335938, + "completions/mean_terminated_length": 555.0931396484375, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 0.2518950437317784, + "grad_norm": 0.1307782381772995, + "learning_rate": 1e-06, + "loss": 0.0028, + "num_tokens": 15698269.0, + "reward": 0.578125, + "reward_std": 0.22623679041862488, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 27 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 576.2064819335938, + "completions/mean_terminated_length": 548.4915771484375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 0.2612244897959184, + "grad_norm": 0.1312188059091568, + "learning_rate": 1e-06, + "loss": 0.0111, + "num_tokens": 16275870.0, + "reward": 0.535714328289032, + "reward_std": 0.2146647721529007, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 28 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 609.4330444335938, + "completions/mean_terminated_length": 566.09716796875, + "completions/min_length": 96.0, + "completions/min_terminated_length": 96.0, + "epoch": 0.2705539358600583, + "grad_norm": 0.12355566024780273, + "learning_rate": 1e-06, + "loss": 0.0106, + "num_tokens": 16865850.0, + "reward": 0.4988839626312256, + "reward_std": 0.20162531733512878, + "rewards/verify_math_reward/mean": 0.4988839328289032, + "rewards/verify_math_reward/std": 0.5002779960632324, + "step": 29 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2581.0, + "completions/mean_length": 551.9319458007812, + "completions/mean_terminated_length": 532.0438232421875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 0.27988338192419826, + "grad_norm": 0.12701798975467682, + "learning_rate": 1e-06, + "loss": 0.016, + "num_tokens": 17435229.0, + "reward": 0.566964328289032, + "reward_std": 0.20947928726673126, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 30 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2770.0, + "completions/mean_length": 672.4888916015625, + "completions/mean_terminated_length": 606.277587890625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 0.2892128279883382, + "grad_norm": 0.1089508980512619, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 18053027.0, + "reward": 0.5412946939468384, + "reward_std": 0.1957337111234665, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 31 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2854.0, + "completions/mean_length": 585.7991333007812, + "completions/mean_terminated_length": 554.1756591796875, + "completions/min_length": 5.0, + "completions/min_terminated_length": 5.0, + "epoch": 0.29854227405247813, + "grad_norm": 0.1300877034664154, + "learning_rate": 1e-06, + "loss": 0.014, + "num_tokens": 18635975.0, + "reward": 0.566964328289032, + "reward_std": 0.21771204471588135, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 32 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2357.0, + "completions/mean_length": 540.849365234375, + "completions/mean_terminated_length": 520.8989868164062, + "completions/min_length": 26.0, + "completions/min_terminated_length": 26.0, + "epoch": 0.30787172011661806, + "grad_norm": 0.12996546924114227, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 19186632.0, + "reward": 0.5479910969734192, + "reward_std": 0.1891927272081375, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 33 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2823.0, + "completions/mean_length": 601.3839721679688, + "completions/mean_terminated_length": 565.925537109375, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 0.317201166180758, + "grad_norm": 0.12991100549697876, + "learning_rate": 1e-06, + "loss": 0.0075, + "num_tokens": 19772000.0, + "reward": 0.5401785969734192, + "reward_std": 0.2281140238046646, + "rewards/verify_math_reward/mean": 0.5401785969734192, + "rewards/verify_math_reward/std": 0.49866142868995667, + "step": 34 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3997.0, + "completions/mean_length": 627.1451416015625, + "completions/mean_terminated_length": 576.07470703125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 0.32653061224489793, + "grad_norm": 0.11559727042913437, + "learning_rate": 1e-06, + "loss": 0.0166, + "num_tokens": 20364330.0, + "reward": 0.5915178656578064, + "reward_std": 0.20151470601558685, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 35 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2236.0, + "completions/mean_length": 648.239990234375, + "completions/mean_terminated_length": 585.5534057617188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 0.3358600583090379, + "grad_norm": 0.12591630220413208, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 20972441.0, + "reward": 0.535714328289032, + "reward_std": 0.21714679896831512, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 36 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 620.8527221679688, + "completions/mean_terminated_length": 565.6915893554688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 0.34518950437317786, + "grad_norm": 0.12799470126628876, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 21548109.0, + "reward": 0.546875, + "reward_std": 0.21128980815410614, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 37 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3290.0, + "completions/mean_length": 648.4710083007812, + "completions/mean_terminated_length": 565.7302856445312, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 0.3545189504373178, + "grad_norm": 0.11907628178596497, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 22132459.0, + "reward": 0.5301339626312256, + "reward_std": 0.19674541056156158, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936985969543457, + "step": 38 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2611.0, + "completions/mean_length": 571.8895263671875, + "completions/mean_terminated_length": 544.140625, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 0.3638483965014577, + "grad_norm": 0.13216789066791534, + "learning_rate": 1e-06, + "loss": 0.019, + "num_tokens": 22712064.0, + "reward": 0.535714328289032, + "reward_std": 0.22364458441734314, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 39 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3386.0, + "completions/mean_length": 591.5390625, + "completions/mean_terminated_length": 559.9673461914062, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 0.37317784256559766, + "grad_norm": 0.11718650162220001, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 23311667.0, + "reward": 0.5100446939468384, + "reward_std": 0.19294719398021698, + "rewards/verify_math_reward/mean": 0.5100446343421936, + "rewards/verify_math_reward/std": 0.5001782774925232, + "step": 40 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4017.0, + "completions/mean_length": 679.5089721679688, + "completions/mean_terminated_length": 621.3394165039062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 0.3825072886297376, + "grad_norm": 0.13004381954669952, + "learning_rate": 1e-06, + "loss": -0.0117, + "num_tokens": 23960851.0, + "reward": 0.4899553656578064, + "reward_std": 0.25212928652763367, + "rewards/verify_math_reward/mean": 0.4899553656578064, + "rewards/verify_math_reward/std": 0.5001782774925232, + "step": 41 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3935.0, + "completions/mean_length": 595.8984375, + "completions/mean_terminated_length": 564.3660278320312, + "completions/min_length": 80.0, + "completions/min_terminated_length": 80.0, + "epoch": 0.39183673469387753, + "grad_norm": 0.11121316254138947, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 24553208.0, + "reward": 0.5267857313156128, + "reward_std": 0.18967169523239136, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608627796173, + "step": 42 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3857.0, + "completions/mean_length": 604.8850708007812, + "completions/mean_terminated_length": 561.49267578125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 0.40116618075801747, + "grad_norm": 0.13258139789104462, + "learning_rate": 1e-06, + "loss": 0.0059, + "num_tokens": 25136265.0, + "reward": 0.5814732313156128, + "reward_std": 0.21362309157848358, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 43 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 547.6830444335938, + "completions/mean_terminated_length": 523.7617797851562, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 0.41049562682215746, + "grad_norm": 0.12222025543451309, + "learning_rate": 1e-06, + "loss": 0.0156, + "num_tokens": 25688605.0, + "reward": 0.609375, + "reward_std": 0.17941944301128387, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 44 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4000.0, + "completions/mean_length": 621.8515625, + "completions/mean_terminated_length": 570.7032470703125, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 0.4198250728862974, + "grad_norm": 0.11610330641269684, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 26285928.0, + "reward": 0.6015625, + "reward_std": 0.20095019042491913, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 45 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2399.0, + "completions/mean_length": 691.325927734375, + "completions/mean_terminated_length": 613.5935668945312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 0.4291545189504373, + "grad_norm": 0.12367964535951614, + "learning_rate": 1e-06, + "loss": 0.0085, + "num_tokens": 26911884.0, + "reward": 0.5457589626312256, + "reward_std": 0.2368360310792923, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 46 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 660.140625, + "completions/mean_terminated_length": 601.641357421875, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 0.43848396501457726, + "grad_norm": 0.11936866492033005, + "learning_rate": 1e-06, + "loss": 0.0141, + "num_tokens": 27526826.0, + "reward": 0.535714328289032, + "reward_std": 0.2177773267030716, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 47 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 593.5625, + "completions/mean_terminated_length": 554.0316162109375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 0.4478134110787172, + "grad_norm": 0.13356272876262665, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 28112346.0, + "reward": 0.5212053656578064, + "reward_std": 0.2054987996816635, + "rewards/verify_math_reward/mean": 0.5212053656578064, + "rewards/verify_math_reward/std": 0.49982914328575134, + "step": 48 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3727.0, + "completions/mean_length": 602.919677734375, + "completions/mean_terminated_length": 575.4151000976562, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 0.45714285714285713, + "grad_norm": 0.11741620302200317, + "learning_rate": 1e-06, + "loss": 0.0277, + "num_tokens": 28706906.0, + "reward": 0.504464328289032, + "reward_std": 0.18400903046131134, + "rewards/verify_math_reward/mean": 0.5044642686843872, + "rewards/verify_math_reward/std": 0.5002593398094177, + "step": 49 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3559.0, + "completions/mean_length": 675.6328125, + "completions/mean_terminated_length": 565.2984008789062, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.46647230320699706, + "grad_norm": 0.13388733565807343, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 29300121.0, + "reward": 0.5245535969734192, + "reward_std": 0.22691050171852112, + "rewards/verify_math_reward/mean": 0.5245535969734192, + "rewards/verify_math_reward/std": 0.4996756613254547, + "step": 50 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3378.0, + "completions/mean_length": 658.2489013671875, + "completions/mean_terminated_length": 579.7614135742188, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 0.47580174927113705, + "grad_norm": 0.13145190477371216, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 29889200.0, + "reward": 0.5915178656578064, + "reward_std": 0.24213102459907532, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 51 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3970.0, + "completions/mean_length": 594.7667846679688, + "completions/mean_terminated_length": 539.1915893554688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.485131195335277, + "grad_norm": 0.1295243501663208, + "learning_rate": 1e-06, + "loss": 0.0264, + "num_tokens": 30444455.0, + "reward": 0.629464328289032, + "reward_std": 0.19279566407203674, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 52 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 595.5457763671875, + "completions/mean_terminated_length": 548.0283203125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 0.4944606413994169, + "grad_norm": 0.11764854192733765, + "learning_rate": 1e-06, + "loss": 0.004, + "num_tokens": 31014664.0, + "reward": 0.5546875, + "reward_std": 0.19200357794761658, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 53 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3689.0, + "completions/mean_length": 545.8627319335938, + "completions/mean_terminated_length": 509.84100341796875, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 0.5037900874635568, + "grad_norm": 0.12535780668258667, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 31556733.0, + "reward": 0.5881696939468384, + "reward_std": 0.19681818783283234, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924396276473999, + "step": 54 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 559.9308471679688, + "completions/mean_terminated_length": 520.0203247070312, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 0.5131195335276968, + "grad_norm": 0.1356523185968399, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 32111863.0, + "reward": 0.6049107313156128, + "reward_std": 0.19238516688346863, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 55 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3399.0, + "completions/mean_length": 614.3392944335938, + "completions/mean_terminated_length": 563.0803833007812, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 0.5224489795918368, + "grad_norm": 0.12192821502685547, + "learning_rate": 1e-06, + "loss": 0.0152, + "num_tokens": 32700471.0, + "reward": 0.5145089626312256, + "reward_std": 0.20383884012699127, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 56 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2324.0, + "completions/mean_length": 566.9553833007812, + "completions/mean_terminated_length": 535.1621704101562, + "completions/min_length": 70.0, + "completions/min_terminated_length": 70.0, + "epoch": 0.5317784256559767, + "grad_norm": 0.13773545622825623, + "learning_rate": 1e-06, + "loss": 0.0225, + "num_tokens": 33262447.0, + "reward": 0.5859375, + "reward_std": 0.19261200726032257, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 57 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2196.0, + "completions/mean_length": 596.4085083007812, + "completions/mean_terminated_length": 552.9107666015625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 0.5411078717201167, + "grad_norm": 0.12552210688591003, + "learning_rate": 1e-06, + "loss": 0.0098, + "num_tokens": 33841637.0, + "reward": 0.543526828289032, + "reward_std": 0.23059673607349396, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 58 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3837.0, + "completions/mean_length": 595.4308471679688, + "completions/mean_terminated_length": 543.8935546875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 0.5504373177842565, + "grad_norm": 0.1274687796831131, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 34416039.0, + "reward": 0.5837053656578064, + "reward_std": 0.18513628840446472, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321892857551575, + "step": 59 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3239.0, + "completions/mean_length": 590.310302734375, + "completions/mean_terminated_length": 538.6976318359375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 0.5597667638483965, + "grad_norm": 0.12446974217891693, + "learning_rate": 1e-06, + "loss": 0.0148, + "num_tokens": 34976717.0, + "reward": 0.5892857313156128, + "reward_std": 0.1707427203655243, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 60 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2976.0, + "completions/mean_length": 683.3627319335938, + "completions/mean_terminated_length": 597.4610595703125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 0.5690962099125364, + "grad_norm": 0.11948134750127792, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 35582922.0, + "reward": 0.5535714626312256, + "reward_std": 0.21211464703083038, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973994791507721, + "step": 61 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3982.0, + "completions/mean_length": 641.0089721679688, + "completions/mean_terminated_length": 609.8828735351562, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 0.5784256559766764, + "grad_norm": 0.1262824833393097, + "learning_rate": 1e-06, + "loss": 0.0077, + "num_tokens": 36208466.0, + "reward": 0.5133928656578064, + "reward_std": 0.24821694195270538, + "rewards/verify_math_reward/mean": 0.5133928656578064, + "rewards/verify_math_reward/std": 0.500099778175354, + "step": 62 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 609.0892944335938, + "completions/mean_terminated_length": 549.7208251953125, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 0.5877551020408164, + "grad_norm": 0.12981727719306946, + "learning_rate": 1e-06, + "loss": -0.0043, + "num_tokens": 36771426.0, + "reward": 0.5792410969734192, + "reward_std": 0.20711390674114227, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 63 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3503.0, + "completions/mean_length": 627.7199096679688, + "completions/mean_terminated_length": 576.657958984375, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.5970845481049563, + "grad_norm": 0.13456951081752777, + "learning_rate": 1e-06, + "loss": 0.0172, + "num_tokens": 37370383.0, + "reward": 0.6049107313156128, + "reward_std": 0.20549951493740082, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 64 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2873.0, + "completions/mean_length": 632.9308471679688, + "completions/mean_terminated_length": 569.9658813476562, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.6064139941690962, + "grad_norm": 0.11932039260864258, + "learning_rate": 1e-06, + "loss": 0.0103, + "num_tokens": 37969633.0, + "reward": 0.546875, + "reward_std": 0.1770561784505844, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 65 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2319.0, + "completions/mean_length": 580.9185791015625, + "completions/mean_terminated_length": 537.228271484375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 0.6157434402332361, + "grad_norm": 0.11847219616174698, + "learning_rate": 1e-06, + "loss": 0.0122, + "num_tokens": 38543616.0, + "reward": 0.5892857313156128, + "reward_std": 0.18257686495780945, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 66 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2846.0, + "completions/mean_length": 675.6272583007812, + "completions/mean_terminated_length": 629.1968383789062, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.6250728862973761, + "grad_norm": 0.10224071890115738, + "learning_rate": 1e-06, + "loss": -0.0045, + "num_tokens": 39195498.0, + "reward": 0.5290178656578064, + "reward_std": 0.19561424851417542, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943605065345764, + "step": 67 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 618.34375, + "completions/mean_terminated_length": 575.11865234375, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 0.634402332361516, + "grad_norm": 0.12181031703948975, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 39788102.0, + "reward": 0.5491071939468384, + "reward_std": 0.1716417372226715, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 68 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 678.0245971679688, + "completions/mean_terminated_length": 611.9203491210938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 0.643731778425656, + "grad_norm": 0.12990140914916992, + "learning_rate": 1e-06, + "loss": 0.0202, + "num_tokens": 40416348.0, + "reward": 0.5647321939468384, + "reward_std": 0.2387114316225052, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606892466545105, + "step": 69 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2959.0, + "completions/mean_length": 644.1663208007812, + "completions/mean_terminated_length": 589.3753051757812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 0.6530612244897959, + "grad_norm": 0.13729752600193024, + "learning_rate": 1e-06, + "loss": 0.0037, + "num_tokens": 41026217.0, + "reward": 0.559151828289032, + "reward_std": 0.24611146748065948, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 70 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1841.0, + "completions/mean_length": 605.0379638671875, + "completions/mean_terminated_length": 565.6365966796875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 0.6623906705539359, + "grad_norm": 0.12223481386899948, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 41626955.0, + "reward": 0.5334821939468384, + "reward_std": 0.19384802877902985, + "rewards/verify_math_reward/mean": 0.5334821343421936, + "rewards/verify_math_reward/std": 0.49915632605552673, + "step": 71 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3670.0, + "completions/mean_length": 667.7288208007812, + "completions/mean_terminated_length": 581.43359375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 0.6717201166180758, + "grad_norm": 0.13852868974208832, + "learning_rate": 1e-06, + "loss": -0.0109, + "num_tokens": 42232568.0, + "reward": 0.5089285969734192, + "reward_std": 0.2327456921339035, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 72 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3816.0, + "completions/mean_length": 641.5502319335938, + "completions/mean_terminated_length": 590.6919555664062, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 0.6810495626822157, + "grad_norm": 0.12090474367141724, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 42850925.0, + "reward": 0.5691964626312256, + "reward_std": 0.19088521599769592, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 73 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2885.0, + "completions/mean_length": 686.5792846679688, + "completions/mean_terminated_length": 604.7531127929688, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 0.6903790087463557, + "grad_norm": 0.123042993247509, + "learning_rate": 1e-06, + "loss": 0.0128, + "num_tokens": 43469788.0, + "reward": 0.543526828289032, + "reward_std": 0.23014530539512634, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 74 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3779.0, + "completions/mean_length": 678.4085083007812, + "completions/mean_terminated_length": 592.3821411132812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.6997084548104956, + "grad_norm": 0.11929791420698166, + "learning_rate": 1e-06, + "loss": -0.0076, + "num_tokens": 44072594.0, + "reward": 0.5178571939468384, + "reward_std": 0.2101197987794876, + "rewards/verify_math_reward/mean": 0.5178571343421936, + "rewards/verify_math_reward/std": 0.4999600946903229, + "step": 75 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3359.0, + "completions/mean_length": 691.9777221679688, + "completions/mean_terminated_length": 634.0204467773438, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 0.7090379008746356, + "grad_norm": 0.12030773609876633, + "learning_rate": 1e-06, + "loss": 0.0072, + "num_tokens": 44724406.0, + "reward": 0.5145089626312256, + "reward_std": 0.23878629505634308, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 76 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2971.0, + "completions/mean_length": 657.1730346679688, + "completions/mean_terminated_length": 574.64111328125, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 0.7183673469387755, + "grad_norm": 0.140348881483078, + "learning_rate": 1e-06, + "loss": 0.0083, + "num_tokens": 45318841.0, + "reward": 0.535714328289032, + "reward_std": 0.2462616264820099, + "rewards/verify_math_reward/mean": 0.5357142686843872, + "rewards/verify_math_reward/std": 0.4990014135837555, + "step": 77 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3232.0, + "completions/mean_length": 612.5592041015625, + "completions/mean_terminated_length": 577.2141723632812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 0.7276967930029155, + "grad_norm": 0.12273713946342468, + "learning_rate": 1e-06, + "loss": 0.016, + "num_tokens": 45916742.0, + "reward": 0.5892857313156128, + "reward_std": 0.21327468752861023, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 78 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2177.0, + "completions/mean_length": 578.146240234375, + "completions/mean_terminated_length": 546.453857421875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 0.7370262390670554, + "grad_norm": 0.15242497622966766, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 46491081.0, + "reward": 0.5993303656578064, + "reward_std": 0.24901039898395538, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 79 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3639.0, + "completions/mean_length": 584.9721069335938, + "completions/mean_terminated_length": 533.2808227539062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 0.7463556851311953, + "grad_norm": 0.1354115754365921, + "learning_rate": 1e-06, + "loss": 0.0159, + "num_tokens": 47052016.0, + "reward": 0.609375, + "reward_std": 0.234275221824646, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 80 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2721.0, + "completions/mean_length": 672.7667846679688, + "completions/mean_terminated_length": 602.5866088867188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 0.7556851311953353, + "grad_norm": 0.12694555521011353, + "learning_rate": 1e-06, + "loss": 0.0109, + "num_tokens": 47676903.0, + "reward": 0.5658482313156128, + "reward_std": 0.2230786681175232, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 81 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3950.0, + "completions/mean_length": 652.4676513671875, + "completions/mean_terminated_length": 581.871337890625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 0.7650145772594752, + "grad_norm": 0.11660673469305038, + "learning_rate": 1e-06, + "loss": 0.0015, + "num_tokens": 48281954.0, + "reward": 0.5770089626312256, + "reward_std": 0.17405030131340027, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 82 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4055.0, + "completions/mean_length": 615.458740234375, + "completions/mean_terminated_length": 580.1431274414062, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 0.7743440233236152, + "grad_norm": 0.11257554590702057, + "learning_rate": 1e-06, + "loss": -0.001, + "num_tokens": 48881757.0, + "reward": 0.6261160969734192, + "reward_std": 0.18513768911361694, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 83 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4054.0, + "completions/mean_length": 665.0234375, + "completions/mean_terminated_length": 602.6420288085938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 0.7836734693877551, + "grad_norm": 0.12210234254598618, + "learning_rate": 1e-06, + "loss": 0.0052, + "num_tokens": 49507698.0, + "reward": 0.4955357313156128, + "reward_std": 0.2192864567041397, + "rewards/verify_math_reward/mean": 0.4955357015132904, + "rewards/verify_math_reward/std": 0.500259280204773, + "step": 84 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3932.0, + "completions/mean_length": 649.6473388671875, + "completions/mean_terminated_length": 578.9931640625, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 0.793002915451895, + "grad_norm": 0.11843501776456833, + "learning_rate": 1e-06, + "loss": 0.0062, + "num_tokens": 50105782.0, + "reward": 0.5714285969734192, + "reward_std": 0.19990073144435883, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 85 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3070.0, + "completions/mean_length": 610.638427734375, + "completions/mean_terminated_length": 591.0797119140625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 0.8023323615160349, + "grad_norm": 0.12302396446466446, + "learning_rate": 1e-06, + "loss": -0.0033, + "num_tokens": 50719298.0, + "reward": 0.5915178656578064, + "reward_std": 0.2105405479669571, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 86 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3432.0, + "completions/mean_length": 557.1529541015625, + "completions/mean_terminated_length": 537.2940673828125, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 0.8116618075801749, + "grad_norm": 0.12430504709482193, + "learning_rate": 1e-06, + "loss": 0.0167, + "num_tokens": 51278691.0, + "reward": 0.6015625, + "reward_std": 0.18809686601161957, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 87 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 581.357177734375, + "completions/mean_terminated_length": 541.6884765625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 0.8209912536443149, + "grad_norm": 0.12935684621334076, + "learning_rate": 1e-06, + "loss": 0.0221, + "num_tokens": 51838699.0, + "reward": 0.6305803656578064, + "reward_std": 0.19738341867923737, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 88 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3783.0, + "completions/mean_length": 656.5022583007812, + "completions/mean_terminated_length": 597.9409790039062, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 0.8303206997084548, + "grad_norm": 0.12048202753067017, + "learning_rate": 1e-06, + "loss": 0.0119, + "num_tokens": 52467333.0, + "reward": 0.5457589626312256, + "reward_std": 0.19516101479530334, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 89 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3795.0, + "completions/mean_length": 644.53125, + "completions/mean_terminated_length": 581.7772827148438, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 0.8396501457725948, + "grad_norm": 0.11943095922470093, + "learning_rate": 1e-06, + "loss": 0.0053, + "num_tokens": 53066849.0, + "reward": 0.5703125, + "reward_std": 0.18745999038219452, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 90 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3328.0, + "completions/mean_length": 613.0814819335938, + "completions/mean_terminated_length": 569.7909545898438, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 0.8489795918367347, + "grad_norm": 0.1321592479944229, + "learning_rate": 1e-06, + "loss": 0.0226, + "num_tokens": 53659074.0, + "reward": 0.6015625, + "reward_std": 0.21458598971366882, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 91 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2999.0, + "completions/mean_length": 666.7835083007812, + "completions/mean_terminated_length": 600.4618530273438, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 0.8583090379008746, + "grad_norm": 0.11667869985103607, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 54273960.0, + "reward": 0.6305803656578064, + "reward_std": 0.18051601946353912, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 92 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2585.0, + "completions/mean_length": 600.3292846679688, + "completions/mean_terminated_length": 572.8043212890625, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 0.8676384839650145, + "grad_norm": 0.1307196170091629, + "learning_rate": 1e-06, + "loss": 0.0152, + "num_tokens": 54880719.0, + "reward": 0.5758928656578064, + "reward_std": 0.24070632457733154, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 93 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 640.8828125, + "completions/mean_terminated_length": 593.9807739257812, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 0.8769679300291545, + "grad_norm": 0.11762183904647827, + "learning_rate": 1e-06, + "loss": 0.0063, + "num_tokens": 55482550.0, + "reward": 0.590401828289032, + "reward_std": 0.1821310967206955, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 94 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3517.0, + "completions/mean_length": 659.5502319335938, + "completions/mean_terminated_length": 616.8372802734375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 0.8862973760932945, + "grad_norm": 0.11842679977416992, + "learning_rate": 1e-06, + "loss": 0.0066, + "num_tokens": 56115947.0, + "reward": 0.546875, + "reward_std": 0.19952097535133362, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 95 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.005580357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2713.0, + "completions/mean_length": 617.068115234375, + "completions/mean_terminated_length": 597.5454711914062, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 0.8956268221574344, + "grad_norm": 0.13340096175670624, + "learning_rate": 1e-06, + "loss": 0.0091, + "num_tokens": 56731816.0, + "reward": 0.5770089626312256, + "reward_std": 0.25370344519615173, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 96 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 675.0625, + "completions/mean_terminated_length": 624.6976318359375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 0.9049562682215744, + "grad_norm": 0.12861862778663635, + "learning_rate": 1e-06, + "loss": 0.0123, + "num_tokens": 57374720.0, + "reward": 0.5636160969734192, + "reward_std": 0.2177562117576599, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 97 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3863.0, + "completions/mean_length": 735.8783569335938, + "completions/mean_terminated_length": 651.298583984375, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 0.9142857142857143, + "grad_norm": 0.12009000033140182, + "learning_rate": 1e-06, + "loss": 0.0195, + "num_tokens": 58029883.0, + "reward": 0.5055803656578064, + "reward_std": 0.20448850095272064, + "rewards/verify_math_reward/mean": 0.5055803656578064, + "rewards/verify_math_reward/std": 0.5002480745315552, + "step": 98 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 678.6785888671875, + "completions/mean_terminated_length": 608.61962890625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 0.9236151603498542, + "grad_norm": 0.11858928203582764, + "learning_rate": 1e-06, + "loss": 0.004, + "num_tokens": 58663467.0, + "reward": 0.4799107313156128, + "reward_std": 0.21324008703231812, + "rewards/verify_math_reward/mean": 0.4799107015132904, + "rewards/verify_math_reward/std": 0.4998753070831299, + "step": 99 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3930.0, + "completions/mean_length": 676.4263916015625, + "completions/mean_terminated_length": 606.3212280273438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 0.9329446064139941, + "grad_norm": 0.110529325902462, + "learning_rate": 1e-06, + "loss": 0.0175, + "num_tokens": 59286081.0, + "reward": 0.5770089626312256, + "reward_std": 0.17281359434127808, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3954.0, + "completions/mean_length": 642.6652221679688, + "completions/mean_terminated_length": 591.8233032226562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 0.9422740524781341, + "grad_norm": 0.11486780643463135, + "learning_rate": 1e-06, + "loss": -0.002, + "num_tokens": 59894181.0, + "reward": 0.5691964626312256, + "reward_std": 0.1902037262916565, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2266.0, + "completions/mean_length": 625.890625, + "completions/mean_terminated_length": 594.62841796875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 0.9516034985422741, + "grad_norm": 0.12553581595420837, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 60522123.0, + "reward": 0.5535714626312256, + "reward_std": 0.20320791006088257, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 770.4531860351562, + "completions/mean_terminated_length": 667.1277465820312, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 0.960932944606414, + "grad_norm": 0.10908481478691101, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 61189321.0, + "reward": 0.5267857313156128, + "reward_std": 0.17671987414360046, + "rewards/verify_math_reward/mean": 0.5267857313156128, + "rewards/verify_math_reward/std": 0.4995608329772949, + "step": 103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4075.0, + "completions/mean_length": 626.3214721679688, + "completions/mean_terminated_length": 583.1954956054688, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 0.970262390670554, + "grad_norm": 0.13236378133296967, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 61798393.0, + "reward": 0.5770089626312256, + "reward_std": 0.2327795773744583, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4045.0, + "completions/mean_length": 624.4453125, + "completions/mean_terminated_length": 561.3261108398438, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 0.9795918367346939, + "grad_norm": 0.13066381216049194, + "learning_rate": 1e-06, + "loss": -0.014, + "num_tokens": 62380256.0, + "reward": 0.5323660969734192, + "reward_std": 0.2111382633447647, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3616.0, + "completions/mean_length": 658.6652221679688, + "completions/mean_terminated_length": 592.1865844726562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 0.9889212827988338, + "grad_norm": 0.12282067537307739, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 62982980.0, + "reward": 0.5647321939468384, + "reward_std": 0.19321031868457794, + "rewards/verify_math_reward/mean": 0.5647321343421936, + "rewards/verify_math_reward/std": 0.49606892466545105, + "step": 106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022727272727272707, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2180.0, + "completions/mean_length": 607.0625, + "completions/mean_terminated_length": 525.9244384765625, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 0.9982507288629737, + "grad_norm": 0.12465576082468033, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 63573256.0, + "reward": 0.5457589626312256, + "reward_std": 0.20880205929279327, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981797933578491, + "step": 107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2778.0, + "completions/mean_length": 623.8671875, + "completions/mean_terminated_length": 580.7107543945312, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.00932944606414, + "grad_norm": 0.12487806379795074, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 64177609.0, + "reward": 0.578125, + "reward_std": 0.2260114848613739, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 641.9888916015625, + "completions/mean_terminated_length": 583.1804809570312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.01865889212828, + "grad_norm": 0.13888868689537048, + "learning_rate": 1e-06, + "loss": 0.0296, + "num_tokens": 64787839.0, + "reward": 0.5691964626312256, + "reward_std": 0.23921138048171997, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2985.0, + "completions/mean_length": 679.6942138671875, + "completions/mean_terminated_length": 613.6222534179688, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.0279883381924197, + "grad_norm": 0.11260522902011871, + "learning_rate": 1e-06, + "loss": 0.0024, + "num_tokens": 65420213.0, + "reward": 0.5290178656578064, + "reward_std": 0.18171106278896332, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943605065345764, + "step": 110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3918.0, + "completions/mean_length": 690.7299194335938, + "completions/mean_terminated_length": 592.9896240234375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.0373177842565597, + "grad_norm": 0.1303315907716751, + "learning_rate": 1e-06, + "loss": 0.0116, + "num_tokens": 66041083.0, + "reward": 0.5089285969734192, + "reward_std": 0.1854732781648636, + "rewards/verify_math_reward/mean": 0.5089285969734192, + "rewards/verify_math_reward/std": 0.5001994967460632, + "step": 111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3618.0, + "completions/mean_length": 623.7154541015625, + "completions/mean_terminated_length": 556.5608520507812, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 1.0466472303206997, + "grad_norm": 0.12855571508407593, + "learning_rate": 1e-06, + "loss": 0.0102, + "num_tokens": 66624876.0, + "reward": 0.5892857313156128, + "reward_std": 0.20940370857715607, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 629.0960083007812, + "completions/mean_terminated_length": 586.0045166015625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 1.0559766763848397, + "grad_norm": 0.12134411185979843, + "learning_rate": 1e-06, + "loss": -0.0033, + "num_tokens": 67234122.0, + "reward": 0.5580357313156128, + "reward_std": 0.19907476007938385, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689781665802, + "step": 113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2329.0, + "completions/mean_length": 668.825927734375, + "completions/mean_terminated_length": 602.5437622070312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 1.0653061224489795, + "grad_norm": 0.1247580423951149, + "learning_rate": 1e-06, + "loss": 0.0029, + "num_tokens": 67875206.0, + "reward": 0.5234375, + "reward_std": 0.2039564698934555, + "rewards/verify_math_reward/mean": 0.5234375, + "rewards/verify_math_reward/std": 0.49972933530807495, + "step": 114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2129.0, + "completions/mean_length": 619.6027221679688, + "completions/mean_terminated_length": 576.3932495117188, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 1.0746355685131195, + "grad_norm": 0.12528985738754272, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 68473226.0, + "reward": 0.5837053656578064, + "reward_std": 0.1946403682231903, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 662.2377319335938, + "completions/mean_terminated_length": 611.6840209960938, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.0839650145772595, + "grad_norm": 0.12204413115978241, + "learning_rate": 1e-06, + "loss": 0.0029, + "num_tokens": 69101183.0, + "reward": 0.6316964626312256, + "reward_std": 0.20223332941532135, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3165.0, + "completions/mean_length": 637.161865234375, + "completions/mean_terminated_length": 578.2713012695312, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 1.0932944606413995, + "grad_norm": 0.12889699637889862, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 69695760.0, + "reward": 0.5636160969734192, + "reward_std": 0.223838210105896, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2714.0, + "completions/mean_length": 623.7142944335938, + "completions/mean_terminated_length": 592.4324340820312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.1026239067055394, + "grad_norm": 0.11693552881479263, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 70296192.0, + "reward": 0.6383928656578064, + "reward_std": 0.17382459342479706, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 623.4921875, + "completions/mean_terminated_length": 556.3333129882812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 1.1119533527696792, + "grad_norm": 0.12983927130699158, + "learning_rate": 1e-06, + "loss": 0.0103, + "num_tokens": 70858833.0, + "reward": 0.629464328289032, + "reward_std": 0.20347003638744354, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3785.0, + "completions/mean_length": 661.78125, + "completions/mean_terminated_length": 611.2208251953125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 1.1212827988338192, + "grad_norm": 0.1284332573413849, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 71495181.0, + "reward": 0.5993303656578064, + "reward_std": 0.19738341867923737, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3249.0, + "completions/mean_length": 651.6239013671875, + "completions/mean_terminated_length": 592.9796142578125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.1306122448979592, + "grad_norm": 0.1337193250656128, + "learning_rate": 1e-06, + "loss": -0.0037, + "num_tokens": 72114796.0, + "reward": 0.546875, + "reward_std": 0.2297341525554657, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3067.0, + "completions/mean_length": 636.8326416015625, + "completions/mean_terminated_length": 589.8756103515625, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 1.1399416909620992, + "grad_norm": 0.13222360610961914, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 72714814.0, + "reward": 0.6194196939468384, + "reward_std": 0.1995958387851715, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 674.1596069335938, + "completions/mean_terminated_length": 623.7814331054688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.149271137026239, + "grad_norm": 0.11800728738307953, + "learning_rate": 1e-06, + "loss": 0.0194, + "num_tokens": 73355701.0, + "reward": 0.6015625, + "reward_std": 0.2128635197877884, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2949.0, + "completions/mean_length": 626.6674194335938, + "completions/mean_terminated_length": 591.465576171875, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.158600583090379, + "grad_norm": 0.11077390611171722, + "learning_rate": 1e-06, + "loss": 0.0118, + "num_tokens": 73969843.0, + "reward": 0.6015625, + "reward_std": 0.1770094931125641, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 691.4453735351562, + "completions/mean_terminated_length": 629.5443115234375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.167930029154519, + "grad_norm": 0.109529510140419, + "learning_rate": 1e-06, + "loss": 0.0084, + "num_tokens": 74614186.0, + "reward": 0.5111607313156128, + "reward_std": 0.15698787569999695, + "rewards/verify_math_reward/mean": 0.5111607313156128, + "rewards/verify_math_reward/std": 0.5001546144485474, + "step": 125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2955.0, + "completions/mean_length": 622.9866333007812, + "completions/mean_terminated_length": 575.8416748046875, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 1.177259475218659, + "grad_norm": 0.12651211023330688, + "learning_rate": 1e-06, + "loss": 0.0043, + "num_tokens": 75211710.0, + "reward": 0.6183035969734192, + "reward_std": 0.19509612023830414, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2372.0, + "completions/mean_length": 627.5714721679688, + "completions/mean_terminated_length": 584.4610595703125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 1.186588921282799, + "grad_norm": 0.11638539284467697, + "learning_rate": 1e-06, + "loss": 0.0118, + "num_tokens": 75811558.0, + "reward": 0.637276828289032, + "reward_std": 0.17742638289928436, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3709.0, + "completions/mean_length": 642.8895263671875, + "completions/mean_terminated_length": 599.969482421875, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.1959183673469387, + "grad_norm": 0.12096329778432846, + "learning_rate": 1e-06, + "loss": 0.0053, + "num_tokens": 76434451.0, + "reward": 0.5814732313156128, + "reward_std": 0.2248164415359497, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3591.0, + "completions/mean_length": 649.7767944335938, + "completions/mean_terminated_length": 606.9423828125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.2052478134110787, + "grad_norm": 0.12809109687805176, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 77064683.0, + "reward": 0.5881696939468384, + "reward_std": 0.22352485358715057, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 726.8984985351562, + "completions/mean_terminated_length": 638.1363525390625, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 1.2145772594752187, + "grad_norm": 0.13690784573554993, + "learning_rate": 1e-06, + "loss": -0.0077, + "num_tokens": 77714920.0, + "reward": 0.5133928656578064, + "reward_std": 0.24900856614112854, + "rewards/verify_math_reward/mean": 0.5133928656578064, + "rewards/verify_math_reward/std": 0.500099778175354, + "step": 130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.006696428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4015.0, + "completions/mean_length": 579.279052734375, + "completions/mean_terminated_length": 555.57080078125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.2239067055393587, + "grad_norm": 0.1269935518503189, + "learning_rate": 1e-06, + "loss": 0.0119, + "num_tokens": 78294530.0, + "reward": 0.6339285969734192, + "reward_std": 0.18927009403705597, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 636.5647583007812, + "completions/mean_terminated_length": 557.5821533203125, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 1.2332361516034984, + "grad_norm": 0.13009558618068695, + "learning_rate": 1e-06, + "loss": 0.0015, + "num_tokens": 78868468.0, + "reward": 0.6082589626312256, + "reward_std": 0.2007683664560318, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2822.0, + "completions/mean_length": 642.8616333007812, + "completions/mean_terminated_length": 572.068359375, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 1.2425655976676384, + "grad_norm": 0.12400022894144058, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 79461616.0, + "reward": 0.5613839626312256, + "reward_std": 0.1953885406255722, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 611.7890625, + "completions/mean_terminated_length": 584.3543701171875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.2518950437317784, + "grad_norm": 0.12356150895357132, + "learning_rate": 1e-06, + "loss": 0.0115, + "num_tokens": 80062371.0, + "reward": 0.606026828289032, + "reward_std": 0.21334956586360931, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3541.0, + "completions/mean_length": 601.8002319335938, + "completions/mean_terminated_length": 550.3567504882812, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.2612244897959184, + "grad_norm": 0.1391543745994568, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 80632352.0, + "reward": 0.6584821939468384, + "reward_std": 0.21429498493671417, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2227.0, + "completions/mean_length": 620.6317138671875, + "completions/mean_terminated_length": 553.41748046875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.2705539358600584, + "grad_norm": 0.11735261231660843, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 81210406.0, + "reward": 0.5948660969734192, + "reward_std": 0.1582338809967041, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3883.0, + "completions/mean_length": 658.0592041015625, + "completions/mean_terminated_length": 603.4886474609375, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 1.2798833819241984, + "grad_norm": 0.12435299158096313, + "learning_rate": 1e-06, + "loss": 0.0216, + "num_tokens": 81828907.0, + "reward": 0.5714285969734192, + "reward_std": 0.1993023008108139, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2638.0, + "completions/mean_length": 671.578125, + "completions/mean_terminated_length": 597.3887939453125, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 1.2892128279883381, + "grad_norm": 0.1195169985294342, + "learning_rate": 1e-06, + "loss": 0.0075, + "num_tokens": 82445129.0, + "reward": 0.546875, + "reward_std": 0.19628825783729553, + "rewards/verify_math_reward/mean": 0.546875, + "rewards/verify_math_reward/std": 0.4980759024620056, + "step": 138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2622.0, + "completions/mean_length": 663.15625, + "completions/mean_terminated_length": 600.7409057617188, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 1.2985422740524781, + "grad_norm": 0.13412754237651825, + "learning_rate": 1e-06, + "loss": -0.0019, + "num_tokens": 83069565.0, + "reward": 0.5100446939468384, + "reward_std": 0.2442018687725067, + "rewards/verify_math_reward/mean": 0.5100446343421936, + "rewards/verify_math_reward/std": 0.5001782774925232, + "step": 139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2407.0, + "completions/mean_length": 684.0881958007812, + "completions/mean_terminated_length": 614.14013671875, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 1.3078717201166181, + "grad_norm": 0.11663468927145004, + "learning_rate": 1e-06, + "loss": 0.0019, + "num_tokens": 83701836.0, + "reward": 0.504464328289032, + "reward_std": 0.2003546804189682, + "rewards/verify_math_reward/mean": 0.5044642686843872, + "rewards/verify_math_reward/std": 0.5002593398094177, + "step": 140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2325.0, + "completions/mean_length": 698.7109985351562, + "completions/mean_terminated_length": 617.1759643554688, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 1.3172011661807579, + "grad_norm": 0.1169997975230217, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 84335793.0, + "reward": 0.512276828289032, + "reward_std": 0.20813307166099548, + "rewards/verify_math_reward/mean": 0.5122767686843872, + "rewards/verify_math_reward/std": 0.500128448009491, + "step": 141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2892.0, + "completions/mean_length": 649.28125, + "completions/mean_terminated_length": 618.229736328125, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 1.3265306122448979, + "grad_norm": 0.1337287276983261, + "learning_rate": 1e-06, + "loss": 0.0091, + "num_tokens": 84978189.0, + "reward": 0.559151828289032, + "reward_std": 0.22692006826400757, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3251.0, + "completions/mean_length": 717.7288208007812, + "completions/mean_terminated_length": 600.6986083984375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.3358600583090379, + "grad_norm": 0.1338474601507187, + "learning_rate": 1e-06, + "loss": 0.0091, + "num_tokens": 85589962.0, + "reward": 0.5290178656578064, + "reward_std": 0.22192324697971344, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943602085113525, + "step": 143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 691.9464721679688, + "completions/mean_terminated_length": 641.830078125, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 1.3451895043731779, + "grad_norm": 0.11507385224103928, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 86252914.0, + "reward": 0.5479910969734192, + "reward_std": 0.22202880680561066, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3833.0, + "completions/mean_length": 692.3717041015625, + "completions/mean_terminated_length": 578.5247802734375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 1.3545189504373178, + "grad_norm": 0.12863509356975555, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 86841151.0, + "reward": 0.6004464626312256, + "reward_std": 0.19887904822826385, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3616.0, + "completions/mean_length": 679.7701416015625, + "completions/mean_terminated_length": 613.6996459960938, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.3638483965014578, + "grad_norm": 0.14015160501003265, + "learning_rate": 1e-06, + "loss": -0.0018, + "num_tokens": 87483977.0, + "reward": 0.5524553656578064, + "reward_std": 0.24220699071884155, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3925.0, + "completions/mean_length": 707.6641235351562, + "completions/mean_terminated_length": 622.3740844726562, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.3731778425655976, + "grad_norm": 0.13800568878650665, + "learning_rate": 1e-06, + "loss": 0.0206, + "num_tokens": 88115540.0, + "reward": 0.5491071939468384, + "reward_std": 0.221356600522995, + "rewards/verify_math_reward/mean": 0.5491071343421936, + "rewards/verify_math_reward/std": 0.49786055088043213, + "step": 147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3267.0, + "completions/mean_length": 718.4453735351562, + "completions/mean_terminated_length": 625.4850463867188, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 1.3825072886297376, + "grad_norm": 0.12576371431350708, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 88744411.0, + "reward": 0.5848214626312256, + "reward_std": 0.2235715538263321, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2045.0, + "completions/mean_length": 630.9017944335938, + "completions/mean_terminated_length": 555.8312377929688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.3918367346938776, + "grad_norm": 0.1185988187789917, + "learning_rate": 1e-06, + "loss": 0.0008, + "num_tokens": 89316499.0, + "reward": 0.5915178656578064, + "reward_std": 0.1904633492231369, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3367.0, + "completions/mean_length": 646.8951416015625, + "completions/mean_terminated_length": 588.1702880859375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.4011661807580174, + "grad_norm": 0.12802352011203766, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 89930933.0, + "reward": 0.6205357313156128, + "reward_std": 0.1799500733613968, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4094.0, + "completions/mean_length": 699.1160888671875, + "completions/mean_terminated_length": 589.5391845703125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 1.4104956268221573, + "grad_norm": 0.12424515187740326, + "learning_rate": 1e-06, + "loss": -0.006, + "num_tokens": 90539853.0, + "reward": 0.5479910969734192, + "reward_std": 0.19189481437206268, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2411.0, + "completions/mean_length": 640.5078125, + "completions/mean_terminated_length": 573.6780395507812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.4198250728862973, + "grad_norm": 0.1243554949760437, + "learning_rate": 1e-06, + "loss": 0.0077, + "num_tokens": 91128596.0, + "reward": 0.6640625, + "reward_std": 0.19189481437206268, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3125.0, + "completions/mean_length": 656.1819458007812, + "completions/mean_terminated_length": 581.6590576171875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.4291545189504373, + "grad_norm": 0.12588343024253845, + "learning_rate": 1e-06, + "loss": 0.0167, + "num_tokens": 91727935.0, + "reward": 0.574776828289032, + "reward_std": 0.1839316487312317, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1980.0, + "completions/mean_length": 643.2511596679688, + "completions/mean_terminated_length": 572.4658813476562, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 1.4384839650145773, + "grad_norm": 0.11955566704273224, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 92313088.0, + "reward": 0.6339285969734192, + "reward_std": 0.1943764090538025, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.012276785714285698, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 607.9888916015625, + "completions/mean_terminated_length": 564.6350708007812, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.4478134110787173, + "grad_norm": 0.1298545002937317, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 92897118.0, + "reward": 0.6272321939468384, + "reward_std": 0.19167795777320862, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2315.0, + "completions/mean_length": 614.3850708007812, + "completions/mean_terminated_length": 551.0829467773438, + "completions/min_length": 197.0, + "completions/min_terminated_length": 197.0, + "epoch": 1.457142857142857, + "grad_norm": 0.1303454339504242, + "learning_rate": 1e-06, + "loss": -0.0003, + "num_tokens": 93471375.0, + "reward": 0.6004464626312256, + "reward_std": 0.19166657328605652, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3856.0, + "completions/mean_length": 695.2645263671875, + "completions/mean_terminated_length": 641.2846069335938, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 1.466472303206997, + "grad_norm": 0.10413635522127151, + "learning_rate": 1e-06, + "loss": -0.0042, + "num_tokens": 94128492.0, + "reward": 0.5703125, + "reward_std": 0.16837765276432037, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 695.3236694335938, + "completions/mean_terminated_length": 617.6826171875, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 1.475801749271137, + "grad_norm": 0.1186751201748848, + "learning_rate": 1e-06, + "loss": 0.0049, + "num_tokens": 94760438.0, + "reward": 0.5848214626312256, + "reward_std": 0.1894555538892746, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3324.0, + "completions/mean_length": 686.5569458007812, + "completions/mean_terminated_length": 628.5073852539062, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 1.485131195335277, + "grad_norm": 0.11440914124250412, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 95392697.0, + "reward": 0.598214328289032, + "reward_std": 0.18295525014400482, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 683.2332763671875, + "completions/mean_terminated_length": 601.3268432617188, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 1.4944606413994168, + "grad_norm": 0.12605339288711548, + "learning_rate": 1e-06, + "loss": 0.0103, + "num_tokens": 96008962.0, + "reward": 0.598214328289032, + "reward_std": 0.20647267997264862, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 160 + }, + { + "clip_ratio/high_max": 0.012116559606511146, + "clip_ratio/high_mean": 0.005343925469787791, + "clip_ratio/low_mean": 0.0034358479824732058, + "clip_ratio/low_min": 0.00012867694385931827, + "clip_ratio/region_mean": 0.008779773372225463, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3620.0, + "completions/mean_length": 608.171875, + "completions/mean_terminated_length": 532.60888671875, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 1.5037900874635568, + "grad_norm": 0.1880822628736496, + "learning_rate": 1e-06, + "loss": 0.0103, + "num_tokens": 96565900.0, + "reward": 0.625, + "reward_std": 0.18840177357196808, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2976.0, + "completions/mean_length": 580.9777221679688, + "completions/mean_terminated_length": 541.3047485351562, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.5131195335276968, + "grad_norm": 0.1276920586824417, + "learning_rate": 1e-06, + "loss": 0.0096, + "num_tokens": 97139888.0, + "reward": 0.6149553656578064, + "reward_std": 0.1839316487312317, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 689.239990234375, + "completions/mean_terminated_length": 591.4569091796875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 1.5224489795918368, + "grad_norm": 0.13044473528862, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 97746455.0, + "reward": 0.5424107313156128, + "reward_std": 0.1896805614233017, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763264656067, + "step": 163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3935.0, + "completions/mean_length": 636.3292846679688, + "completions/mean_terminated_length": 557.34130859375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 1.5317784256559768, + "grad_norm": 0.12751109898090363, + "learning_rate": 1e-06, + "loss": 0.0281, + "num_tokens": 98321438.0, + "reward": 0.582589328289032, + "reward_std": 0.170254185795784, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3678.0, + "completions/mean_length": 649.8527221679688, + "completions/mean_terminated_length": 591.17822265625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.5411078717201168, + "grad_norm": 0.12062021344900131, + "learning_rate": 1e-06, + "loss": 0.0172, + "num_tokens": 98932130.0, + "reward": 0.578125, + "reward_std": 0.19284433126449585, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 643.3839721679688, + "completions/mean_terminated_length": 560.5211181640625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 1.5504373177842565, + "grad_norm": 0.1180112361907959, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 99516730.0, + "reward": 0.5970982313156128, + "reward_std": 0.1597748100757599, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2018.0, + "completions/mean_length": 598.8471069335938, + "completions/mean_terminated_length": 543.3367309570312, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.5597667638483965, + "grad_norm": 0.137054905295372, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 100080961.0, + "reward": 0.609375, + "reward_std": 0.1894962340593338, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3689.0, + "completions/mean_length": 723.7611694335938, + "completions/mean_terminated_length": 626.968994140625, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 1.5690962099125363, + "grad_norm": 0.13055726885795593, + "learning_rate": 1e-06, + "loss": -0.0037, + "num_tokens": 100711075.0, + "reward": 0.5479910969734192, + "reward_std": 0.21819737553596497, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3872.0, + "completions/mean_length": 715.3582763671875, + "completions/mean_terminated_length": 630.261962890625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 1.5784256559766763, + "grad_norm": 0.12505494058132172, + "learning_rate": 1e-06, + "loss": 0.0166, + "num_tokens": 101345892.0, + "reward": 0.5379464626312256, + "reward_std": 0.22395020723342896, + "rewards/verify_math_reward/mean": 0.5379464030265808, + "rewards/verify_math_reward/std": 0.4988364577293396, + "step": 169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3662.0, + "completions/mean_length": 666.8939819335938, + "completions/mean_terminated_length": 580.5777587890625, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.5877551020408163, + "grad_norm": 0.12575773894786835, + "learning_rate": 1e-06, + "loss": 0.0042, + "num_tokens": 101931973.0, + "reward": 0.6261160969734192, + "reward_std": 0.19392317533493042, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3605.0, + "completions/mean_length": 695.0513916015625, + "completions/mean_terminated_length": 593.413818359375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 1.5970845481049563, + "grad_norm": 0.12819413840770721, + "learning_rate": 1e-06, + "loss": -0.0096, + "num_tokens": 102538011.0, + "reward": 0.6506696939468384, + "reward_std": 0.19279521703720093, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3791.0, + "completions/mean_length": 659.7756958007812, + "completions/mean_terminated_length": 593.3185424804688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.6064139941690962, + "grad_norm": 0.12030527740716934, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 103157218.0, + "reward": 0.5680803656578064, + "reward_std": 0.17145811021327972, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3619.0, + "completions/mean_length": 623.0480346679688, + "completions/mean_terminated_length": 559.9033813476562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.6157434402332362, + "grad_norm": 0.12983718514442444, + "learning_rate": 1e-06, + "loss": 0.0155, + "num_tokens": 103748469.0, + "reward": 0.5915178656578064, + "reward_std": 0.20054012537002563, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3047.0, + "completions/mean_length": 739.8683471679688, + "completions/mean_terminated_length": 671.0637817382812, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 1.6250728862973762, + "grad_norm": 0.11184944212436676, + "learning_rate": 1e-06, + "loss": 0.0045, + "num_tokens": 104433335.0, + "reward": 0.5535714626312256, + "reward_std": 0.2003892958164215, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973995089530945, + "step": 174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3619.0, + "completions/mean_length": 634.9832763671875, + "completions/mean_terminated_length": 576.0556640625, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.634402332361516, + "grad_norm": 0.12884528934955597, + "learning_rate": 1e-06, + "loss": 0.0071, + "num_tokens": 105024464.0, + "reward": 0.578125, + "reward_std": 0.17171843349933624, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 711.5078735351562, + "completions/mean_terminated_length": 622.3402099609375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 1.643731778425656, + "grad_norm": 0.13214844465255737, + "learning_rate": 1e-06, + "loss": 0.0198, + "num_tokens": 105658135.0, + "reward": 0.5770089626312256, + "reward_std": 0.25254225730895996, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3531.0, + "completions/mean_length": 671.8660888671875, + "completions/mean_terminated_length": 613.56640625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.6530612244897958, + "grad_norm": 0.1255129724740982, + "learning_rate": 1e-06, + "loss": 0.0087, + "num_tokens": 106288727.0, + "reward": 0.5703125, + "reward_std": 0.22367849946022034, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.011160714285714302, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 623.1484375, + "completions/mean_terminated_length": 583.9514770507812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 1.6623906705539357, + "grad_norm": 0.1195969507098198, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 106905692.0, + "reward": 0.5535714626312256, + "reward_std": 0.18524546921253204, + "rewards/verify_math_reward/mean": 0.5535714030265808, + "rewards/verify_math_reward/std": 0.4973994791507721, + "step": 178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2288.0, + "completions/mean_length": 664.732177734375, + "completions/mean_terminated_length": 598.370849609375, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 1.6717201166180757, + "grad_norm": 0.12886367738246918, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 107529100.0, + "reward": 0.5625, + "reward_std": 0.21312496066093445, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3655.0, + "completions/mean_length": 658.9241333007812, + "completions/mean_terminated_length": 604.3673706054688, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 1.6810495626822157, + "grad_norm": 0.11670944094657898, + "learning_rate": 1e-06, + "loss": -0.0054, + "num_tokens": 108158928.0, + "reward": 0.5714285969734192, + "reward_std": 0.1662733107805252, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4014.0, + "completions/mean_length": 775.3951416015625, + "completions/mean_terminated_length": 624.2824096679688, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 1.6903790087463557, + "grad_norm": 0.13003426790237427, + "learning_rate": 1e-06, + "loss": 0.0046, + "num_tokens": 108783642.0, + "reward": 0.5959821939468384, + "reward_std": 0.2073410153388977, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3712.0, + "completions/mean_length": 706.9063110351562, + "completions/mean_terminated_length": 625.5679931640625, + "completions/min_length": 33.0, + "completions/min_terminated_length": 33.0, + "epoch": 1.6997084548104957, + "grad_norm": 0.11653265357017517, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 109416078.0, + "reward": 0.5412946939468384, + "reward_std": 0.17314989864826202, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 713.6194458007812, + "completions/mean_terminated_length": 648.20361328125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 1.7090379008746357, + "grad_norm": 0.13665492832660675, + "learning_rate": 1e-06, + "loss": 0.0222, + "num_tokens": 110079089.0, + "reward": 0.5558035969734192, + "reward_std": 0.2716216742992401, + "rewards/verify_math_reward/mean": 0.5558035969734192, + "rewards/verify_math_reward/std": 0.49715369939804077, + "step": 183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3900.0, + "completions/mean_length": 681.8170166015625, + "completions/mean_terminated_length": 583.8208618164062, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 1.7183673469387755, + "grad_norm": 0.14328347146511078, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 110679221.0, + "reward": 0.5703125, + "reward_std": 0.22638945281505585, + "rewards/verify_math_reward/mean": 0.5703125, + "rewards/verify_math_reward/std": 0.49530795216560364, + "step": 184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3380.0, + "completions/mean_length": 669.5234375, + "completions/mean_terminated_length": 583.2734375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 1.7276967930029155, + "grad_norm": 0.12823760509490967, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 111274914.0, + "reward": 0.6104910969734192, + "reward_std": 0.20786207914352417, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2426.0, + "completions/mean_length": 625.755615234375, + "completions/mean_terminated_length": 570.67236328125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 1.7370262390670554, + "grad_norm": 0.13960646092891693, + "learning_rate": 1e-06, + "loss": 0.0165, + "num_tokens": 111867335.0, + "reward": 0.6272321939468384, + "reward_std": 0.23401561379432678, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3851.0, + "completions/mean_length": 624.4420166015625, + "completions/mean_terminated_length": 557.3014526367188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 1.7463556851311952, + "grad_norm": 0.1303594410419464, + "learning_rate": 1e-06, + "loss": -0.0027, + "num_tokens": 112447251.0, + "reward": 0.606026828289032, + "reward_std": 0.21165889501571655, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3712.0, + "completions/mean_length": 698.1785888671875, + "completions/mean_terminated_length": 628.5194091796875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 1.7556851311953352, + "grad_norm": 0.12363781780004501, + "learning_rate": 1e-06, + "loss": 0.0036, + "num_tokens": 113094907.0, + "reward": 0.5613839626312256, + "reward_std": 0.21706831455230713, + "rewards/verify_math_reward/mean": 0.5613839030265808, + "rewards/verify_math_reward/std": 0.496494859457016, + "step": 188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3446.0, + "completions/mean_length": 684.7288208007812, + "completions/mean_terminated_length": 610.8244018554688, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.7650145772594752, + "grad_norm": 0.11238560825586319, + "learning_rate": 1e-06, + "loss": -0.0006, + "num_tokens": 113724768.0, + "reward": 0.613839328289032, + "reward_std": 0.1637873649597168, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2210.0, + "completions/mean_length": 645.2020263671875, + "completions/mean_terminated_length": 590.4274291992188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 1.7743440233236152, + "grad_norm": 0.1156289279460907, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 114330741.0, + "reward": 0.6517857313156128, + "reward_std": 0.16420035064220428, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 724.5111694335938, + "completions/mean_terminated_length": 635.6861572265625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 1.7836734693877552, + "grad_norm": 0.12296874076128006, + "learning_rate": 1e-06, + "loss": 0.0031, + "num_tokens": 114981311.0, + "reward": 0.5100446939468384, + "reward_std": 0.20978277921676636, + "rewards/verify_math_reward/mean": 0.5100446343421936, + "rewards/verify_math_reward/std": 0.5001782774925232, + "step": 191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3943.0, + "completions/mean_length": 677.099365234375, + "completions/mean_terminated_length": 603.0296020507812, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 1.7930029154518952, + "grad_norm": 0.1256035715341568, + "learning_rate": 1e-06, + "loss": 0.0142, + "num_tokens": 115599896.0, + "reward": 0.5859375, + "reward_std": 0.20674553513526917, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.010044642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3521.0, + "completions/mean_length": 634.6239013671875, + "completions/mean_terminated_length": 599.5028076171875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.802332361516035, + "grad_norm": 0.11491705477237701, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 116218519.0, + "reward": 0.6071428656578064, + "reward_std": 0.1791905015707016, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 618.935302734375, + "completions/mean_terminated_length": 559.7344360351562, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 1.811661807580175, + "grad_norm": 0.12553679943084717, + "learning_rate": 1e-06, + "loss": 0.0112, + "num_tokens": 116792309.0, + "reward": 0.609375, + "reward_std": 0.19118830561637878, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3036.0, + "completions/mean_length": 611.9832763671875, + "completions/mean_terminated_length": 556.681396484375, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 1.820991253644315, + "grad_norm": 0.13754640519618988, + "learning_rate": 1e-06, + "loss": 0.0237, + "num_tokens": 117363374.0, + "reward": 0.645089328289032, + "reward_std": 0.20016175508499146, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2981.0, + "completions/mean_length": 666.614990234375, + "completions/mean_terminated_length": 592.318115234375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.8303206997084547, + "grad_norm": 0.12373079359531403, + "learning_rate": 1e-06, + "loss": -0.0088, + "num_tokens": 117984685.0, + "reward": 0.559151828289032, + "reward_std": 0.18295595049858093, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3246.0, + "completions/mean_length": 678.3705444335938, + "completions/mean_terminated_length": 600.3424682617188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 1.8396501457725947, + "grad_norm": 0.13277728855609894, + "learning_rate": 1e-06, + "loss": 0.0041, + "num_tokens": 118598137.0, + "reward": 0.590401828289032, + "reward_std": 0.20824436843395233, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4003.0, + "completions/mean_length": 693.3080444335938, + "completions/mean_terminated_length": 603.6609497070312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 1.8489795918367347, + "grad_norm": 0.1427263468503952, + "learning_rate": 1e-06, + "loss": 0.0423, + "num_tokens": 119213093.0, + "reward": 0.6026785969734192, + "reward_std": 0.21582846343517303, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3991.0, + "completions/mean_length": 703.4397583007812, + "completions/mean_terminated_length": 618.04345703125, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 1.8583090379008746, + "grad_norm": 0.11273408681154251, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 119840343.0, + "reward": 0.6171875, + "reward_std": 0.18904300034046173, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 650.982177734375, + "completions/mean_terminated_length": 580.3554077148438, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 1.8676384839650146, + "grad_norm": 0.14073939621448517, + "learning_rate": 1e-06, + "loss": 0.0205, + "num_tokens": 120447431.0, + "reward": 0.6160714626312256, + "reward_std": 0.2494611144065857, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3474.0, + "completions/mean_length": 652.9955444335938, + "completions/mean_terminated_length": 598.3446655273438, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 1.8769679300291546, + "grad_norm": 0.12974786758422852, + "learning_rate": 1e-06, + "loss": 0.003, + "num_tokens": 121051923.0, + "reward": 0.6049107313156128, + "reward_std": 0.1859629601240158, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2348.0, + "completions/mean_length": 697.5636596679688, + "completions/mean_terminated_length": 627.891845703125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.8862973760932946, + "grad_norm": 0.11633289605379105, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 121690708.0, + "reward": 0.5658482313156128, + "reward_std": 0.18167714774608612, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3117.0, + "completions/mean_length": 688.2611694335938, + "completions/mean_terminated_length": 602.4827880859375, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 1.8956268221574344, + "grad_norm": 0.12267957627773285, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 122300734.0, + "reward": 0.598214328289032, + "reward_std": 0.1968199908733368, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3783.0, + "completions/mean_length": 704.739990234375, + "completions/mean_terminated_length": 615.39404296875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 1.9049562682215744, + "grad_norm": 0.125501811504364, + "learning_rate": 1e-06, + "loss": 0.0118, + "num_tokens": 122929269.0, + "reward": 0.5636160969734192, + "reward_std": 0.20666630566120148, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3833.0, + "completions/mean_length": 751.8560791015625, + "completions/mean_terminated_length": 647.9528198242188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.9142857142857141, + "grad_norm": 0.11129347234964371, + "learning_rate": 1e-06, + "loss": -0.0013, + "num_tokens": 123578268.0, + "reward": 0.543526828289032, + "reward_std": 0.1782180517911911, + "rewards/verify_math_reward/mean": 0.5435267686843872, + "rewards/verify_math_reward/std": 0.49838000535964966, + "step": 205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2494.0, + "completions/mean_length": 714.2098388671875, + "completions/mean_terminated_length": 621.1329956054688, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 1.9236151603498541, + "grad_norm": 0.1164608746767044, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 124219112.0, + "reward": 0.5223214626312256, + "reward_std": 0.1895272135734558, + "rewards/verify_math_reward/mean": 0.5223214030265808, + "rewards/verify_math_reward/std": 0.49978047609329224, + "step": 206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 717.8125610351562, + "completions/mean_terminated_length": 616.8551635742188, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.9329446064139941, + "grad_norm": 0.12072525918483734, + "learning_rate": 1e-06, + "loss": 0.0048, + "num_tokens": 124846040.0, + "reward": 0.5725446939468384, + "reward_std": 0.1901698261499405, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 713.8314819335938, + "completions/mean_terminated_length": 636.6129760742188, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 1.9422740524781341, + "grad_norm": 0.11601443588733673, + "learning_rate": 1e-06, + "loss": -0.0056, + "num_tokens": 125489233.0, + "reward": 0.5758928656578064, + "reward_std": 0.1690174639225006, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4011.0, + "completions/mean_length": 675.7879638671875, + "completions/mean_terminated_length": 613.6022338867188, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 1.951603498542274, + "grad_norm": 0.1186499372124672, + "learning_rate": 1e-06, + "loss": -0.0096, + "num_tokens": 126129115.0, + "reward": 0.5691964626312256, + "reward_std": 0.20666815340518951, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2952.0, + "completions/mean_length": 760.0156860351562, + "completions/mean_terminated_length": 672.1260375976562, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 1.960932944606414, + "grad_norm": 0.12476658821105957, + "learning_rate": 1e-06, + "loss": 0.0036, + "num_tokens": 126803345.0, + "reward": 0.5457589626312256, + "reward_std": 0.16570988297462463, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 678.8482666015625, + "completions/mean_terminated_length": 612.7599487304688, + "completions/min_length": 86.0, + "completions/min_terminated_length": 86.0, + "epoch": 1.970262390670554, + "grad_norm": 0.1370057910680771, + "learning_rate": 1e-06, + "loss": 0.012, + "num_tokens": 127434905.0, + "reward": 0.6127232313156128, + "reward_std": 0.2277858853340149, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 637.9375, + "completions/mean_terminated_length": 571.0579833984375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 1.9795918367346939, + "grad_norm": 0.13216781616210938, + "learning_rate": 1e-06, + "loss": 0.0032, + "num_tokens": 128024761.0, + "reward": 0.5814732313156128, + "reward_std": 0.19948704540729523, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2139.0, + "completions/mean_length": 652.1998291015625, + "completions/mean_terminated_length": 589.585205078125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 1.9889212827988338, + "grad_norm": 0.1154385507106781, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 128625788.0, + "reward": 0.5792410969734192, + "reward_std": 0.16691452264785767, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3451.0, + "completions/mean_length": 671.6217041015625, + "completions/mean_terminated_length": 573.3329467773438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 1.9982507288629736, + "grad_norm": 0.1378655582666397, + "learning_rate": 1e-06, + "loss": 0.0189, + "num_tokens": 129221665.0, + "reward": 0.5736607313156128, + "reward_std": 0.2290155440568924, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3731.0, + "completions/mean_length": 652.9185791015625, + "completions/mean_terminated_length": 582.3314819335938, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.00932944606414, + "grad_norm": 0.1272018551826477, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 129823376.0, + "reward": 0.59375, + "reward_std": 0.20000699162483215, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3848.0, + "completions/mean_length": 635.0647583007812, + "completions/mean_terminated_length": 584.1109619140625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.01865889212828, + "grad_norm": 0.1249706968665123, + "learning_rate": 1e-06, + "loss": 0.017, + "num_tokens": 130435594.0, + "reward": 0.625, + "reward_std": 0.17860347032546997, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3391.0, + "completions/mean_length": 724.5301513671875, + "completions/mean_terminated_length": 611.7589111328125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.02798833819242, + "grad_norm": 0.1199277713894844, + "learning_rate": 1e-06, + "loss": -0.002, + "num_tokens": 131058989.0, + "reward": 0.5424107313156128, + "reward_std": 0.18288108706474304, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763264656067, + "step": 217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3985.0, + "completions/mean_length": 696.896240234375, + "completions/mean_terminated_length": 607.3436889648438, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 2.03731778425656, + "grad_norm": 0.11401030421257019, + "learning_rate": 1e-06, + "loss": -0.0049, + "num_tokens": 131693576.0, + "reward": 0.5323660969734192, + "reward_std": 0.1718367636203766, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3053.0, + "completions/mean_length": 626.357177734375, + "completions/mean_terminated_length": 559.253662109375, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 2.0466472303206995, + "grad_norm": 0.1312752068042755, + "learning_rate": 1e-06, + "loss": -0.0045, + "num_tokens": 132279736.0, + "reward": 0.6160714626312256, + "reward_std": 0.19155782461166382, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 670.8761596679688, + "completions/mean_terminated_length": 600.6572265625, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 2.0559766763848395, + "grad_norm": 0.11966001242399216, + "learning_rate": 1e-06, + "loss": -0.0077, + "num_tokens": 132897745.0, + "reward": 0.5803571939468384, + "reward_std": 0.18553857505321503, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761127948761, + "step": 220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3201.0, + "completions/mean_length": 688.9230346679688, + "completions/mean_terminated_length": 591.130859375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.0653061224489795, + "grad_norm": 0.12390231341123581, + "learning_rate": 1e-06, + "loss": -0.0003, + "num_tokens": 133524068.0, + "reward": 0.5323660969734192, + "reward_std": 0.16792230308055878, + "rewards/verify_math_reward/mean": 0.5323660969734192, + "rewards/verify_math_reward/std": 0.4992299973964691, + "step": 221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3010.0, + "completions/mean_length": 646.1049194335938, + "completions/mean_terminated_length": 583.3795166015625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 2.0746355685131195, + "grad_norm": 0.12680338323116302, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 134125354.0, + "reward": 0.5926339626312256, + "reward_std": 0.1713072657585144, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3255.0, + "completions/mean_length": 663.560302734375, + "completions/mean_terminated_length": 613.0260620117188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 2.0839650145772595, + "grad_norm": 0.1163000613451004, + "learning_rate": 1e-06, + "loss": 0.0133, + "num_tokens": 134754496.0, + "reward": 0.660714328289032, + "reward_std": 0.18077746033668518, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3616.0, + "completions/mean_length": 661.8013916015625, + "completions/mean_terminated_length": 595.3833618164062, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 2.0932944606413995, + "grad_norm": 0.14293843507766724, + "learning_rate": 1e-06, + "loss": 0.0296, + "num_tokens": 135362958.0, + "reward": 0.5926339626312256, + "reward_std": 0.23375527560710907, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161848425865173, + "step": 224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2762.0, + "completions/mean_length": 677.6373291015625, + "completions/mean_terminated_length": 599.5924682617188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.1026239067055394, + "grad_norm": 0.10091352462768555, + "learning_rate": 1e-06, + "loss": -0.0018, + "num_tokens": 135962553.0, + "reward": 0.6808035969734192, + "reward_std": 0.1557076871395111, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3848.0, + "completions/mean_length": 645.3270263671875, + "completions/mean_terminated_length": 558.4679565429688, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 2.1119533527696794, + "grad_norm": 0.12225000560283661, + "learning_rate": 1e-06, + "loss": -0.003, + "num_tokens": 136524278.0, + "reward": 0.6417410969734192, + "reward_std": 0.179006889462471, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2817.0, + "completions/mean_length": 666.8605346679688, + "completions/mean_terminated_length": 600.5403442382812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 2.1212827988338194, + "grad_norm": 0.12009953707456589, + "learning_rate": 1e-06, + "loss": 0.0097, + "num_tokens": 137148793.0, + "reward": 0.609375, + "reward_std": 0.17261996865272522, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3462.0, + "completions/mean_length": 714.5167846679688, + "completions/mean_terminated_length": 613.4609375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.130612244897959, + "grad_norm": 0.11147555708885193, + "learning_rate": 1e-06, + "loss": 0.0052, + "num_tokens": 137779704.0, + "reward": 0.5424107313156128, + "reward_std": 0.16746655106544495, + "rewards/verify_math_reward/mean": 0.5424107313156128, + "rewards/verify_math_reward/std": 0.4984763264656067, + "step": 228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3432.0, + "completions/mean_length": 682.4832763671875, + "completions/mean_terminated_length": 604.549072265625, + "completions/min_length": 103.0, + "completions/min_terminated_length": 103.0, + "epoch": 2.139941690962099, + "grad_norm": 0.12987785041332245, + "learning_rate": 1e-06, + "loss": 0.0262, + "num_tokens": 138387857.0, + "reward": 0.6383928656578064, + "reward_std": 0.20398928225040436, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 718.1920166015625, + "completions/mean_terminated_length": 629.2005004882812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 2.149271137026239, + "grad_norm": 0.11622848361730576, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 139027237.0, + "reward": 0.6194196939468384, + "reward_std": 0.20944397151470184, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2505.0, + "completions/mean_length": 655.216552734375, + "completions/mean_terminated_length": 588.6712036132812, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 2.158600583090379, + "grad_norm": 0.1269744485616684, + "learning_rate": 1e-06, + "loss": 0.0022, + "num_tokens": 139634191.0, + "reward": 0.6316964626312256, + "reward_std": 0.16871324181556702, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3614.0, + "completions/mean_length": 768.4676513671875, + "completions/mean_terminated_length": 621.0944213867188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 2.167930029154519, + "grad_norm": 0.09723895788192749, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 140257434.0, + "reward": 0.5368303656578064, + "reward_std": 0.129900723695755, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3163.0, + "completions/mean_length": 641.6674194335938, + "completions/mean_terminated_length": 590.8108520507812, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 2.177259475218659, + "grad_norm": 0.10979227721691132, + "learning_rate": 1e-06, + "loss": 0.0086, + "num_tokens": 140867600.0, + "reward": 0.6417410969734192, + "reward_std": 0.16134923696517944, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3440.0, + "completions/mean_length": 654.4163208007812, + "completions/mean_terminated_length": 595.819580078125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 2.186588921282799, + "grad_norm": 0.11229733377695084, + "learning_rate": 1e-06, + "loss": 0.0115, + "num_tokens": 141475117.0, + "reward": 0.6339285969734192, + "reward_std": 0.15687911212444305, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 706.1272583007812, + "completions/mean_terminated_length": 608.8289184570312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.195918367346939, + "grad_norm": 0.1288706362247467, + "learning_rate": 1e-06, + "loss": 0.0014, + "num_tokens": 142097327.0, + "reward": 0.5970982313156128, + "reward_std": 0.21617180109024048, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3080.0, + "completions/mean_length": 698.5714721679688, + "completions/mean_terminated_length": 613.0526123046875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.205247813411079, + "grad_norm": 0.13237717747688293, + "learning_rate": 1e-06, + "loss": 0.0072, + "num_tokens": 142726223.0, + "reward": 0.606026828289032, + "reward_std": 0.21680979430675507, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3362.0, + "completions/mean_length": 739.3035888671875, + "completions/mean_terminated_length": 674.384521484375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 2.2145772594752184, + "grad_norm": 0.1296413689851761, + "learning_rate": 1e-06, + "loss": 0.0004, + "num_tokens": 143412151.0, + "reward": 0.5457589626312256, + "reward_std": 0.24176080524921417, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.008928571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 591.75, + "completions/mean_terminated_length": 560.18017578125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.2239067055393584, + "grad_norm": 0.12846998870372772, + "learning_rate": 1e-06, + "loss": 0.011, + "num_tokens": 143994743.0, + "reward": 0.6595982313156128, + "reward_std": 0.1666869819164276, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2262.0, + "completions/mean_length": 610.8504638671875, + "completions/mean_terminated_length": 559.5401611328125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 2.2332361516034984, + "grad_norm": 0.1109478622674942, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 144574313.0, + "reward": 0.6238839626312256, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3155.0, + "completions/mean_length": 672.4453125, + "completions/mean_terminated_length": 594.2819213867188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.2425655976676384, + "grad_norm": 0.1168147623538971, + "learning_rate": 1e-06, + "loss": 0.0144, + "num_tokens": 145185776.0, + "reward": 0.574776828289032, + "reward_std": 0.1638326644897461, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 240 + }, + { + "clip_ratio/high_max": 0.018368792487308383, + "clip_ratio/high_mean": 0.00804357198649086, + "clip_ratio/low_mean": 0.004262413567630574, + "clip_ratio/low_min": 0.0005176992708584294, + "clip_ratio/region_mean": 0.012305985379498452, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3030.0, + "completions/mean_length": 671.1484375, + "completions/mean_terminated_length": 600.9351196289062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 2.2518950437317784, + "grad_norm": 0.2730906009674072, + "learning_rate": 1e-06, + "loss": 0.01, + "num_tokens": 145801517.0, + "reward": 0.6283482313156128, + "reward_std": 0.20873039960861206, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3254.0, + "completions/mean_length": 662.3828125, + "completions/mean_terminated_length": 595.97607421875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.2612244897959184, + "grad_norm": 0.13381898403167725, + "learning_rate": 1e-06, + "loss": 0.0074, + "num_tokens": 146415372.0, + "reward": 0.6004464626312256, + "reward_std": 0.19077390432357788, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3416.0, + "completions/mean_length": 734.7645263671875, + "completions/mean_terminated_length": 646.2096557617188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 2.2705539358600584, + "grad_norm": 0.1320689618587494, + "learning_rate": 1e-06, + "loss": 0.01, + "num_tokens": 147068497.0, + "reward": 0.5334821939468384, + "reward_std": 0.23856060206890106, + "rewards/verify_math_reward/mean": 0.5334821343421936, + "rewards/verify_math_reward/std": 0.49915629625320435, + "step": 243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3375.0, + "completions/mean_length": 616.2109375, + "completions/mean_terminated_length": 568.9739990234375, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 2.2798833819241984, + "grad_norm": 0.1339106410741806, + "learning_rate": 1e-06, + "loss": 0.0134, + "num_tokens": 147666798.0, + "reward": 0.6037946939468384, + "reward_std": 0.20749256014823914, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2897.0, + "completions/mean_length": 712.638427734375, + "completions/mean_terminated_length": 651.1226806640625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 2.2892128279883384, + "grad_norm": 0.11931008845567703, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 148324666.0, + "reward": 0.5881696939468384, + "reward_std": 0.1803637593984604, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924396276473999, + "step": 245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 632.78125, + "completions/mean_terminated_length": 581.7938842773438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.298542274052478, + "grad_norm": 0.13235315680503845, + "learning_rate": 1e-06, + "loss": 0.0061, + "num_tokens": 148929230.0, + "reward": 0.609375, + "reward_std": 0.19107066094875336, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3745.0, + "completions/mean_length": 621.4955444335938, + "completions/mean_terminated_length": 550.2642822265625, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.307871720116618, + "grad_norm": 0.13729095458984375, + "learning_rate": 1e-06, + "loss": -0.0038, + "num_tokens": 149498898.0, + "reward": 0.590401828289032, + "reward_std": 0.20203717052936554, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 698.8158569335938, + "completions/mean_terminated_length": 609.3139038085938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.317201166180758, + "grad_norm": 0.12783533334732056, + "learning_rate": 1e-06, + "loss": -0.0147, + "num_tokens": 150114221.0, + "reward": 0.5736607313156128, + "reward_std": 0.19528047740459442, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3251.0, + "completions/mean_length": 712.4564819335938, + "completions/mean_terminated_length": 603.3099365234375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.326530612244898, + "grad_norm": 0.13753677904605865, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 150721550.0, + "reward": 0.6316964626312256, + "reward_std": 0.21623891592025757, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3964.0, + "completions/mean_length": 667.1752319335938, + "completions/mean_terminated_length": 616.6942138671875, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 2.335860058309038, + "grad_norm": 0.13102750480175018, + "learning_rate": 1e-06, + "loss": 0.0101, + "num_tokens": 151358915.0, + "reward": 0.5870535969734192, + "reward_std": 0.20993433892726898, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263837933540344, + "step": 250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2862.0, + "completions/mean_length": 672.671875, + "completions/mean_terminated_length": 586.5010986328125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.345189504373178, + "grad_norm": 0.13345371186733246, + "learning_rate": 1e-06, + "loss": -0.0074, + "num_tokens": 151948245.0, + "reward": 0.6049107313156128, + "reward_std": 0.204860121011734, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3697.0, + "completions/mean_length": 702.0313110351562, + "completions/mean_terminated_length": 592.5484008789062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 2.354518950437318, + "grad_norm": 0.11863140761852264, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 152551913.0, + "reward": 0.5948660969734192, + "reward_std": 0.1940726339817047, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3275.0, + "completions/mean_length": 645.2020263671875, + "completions/mean_terminated_length": 586.4483642578125, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 2.363848396501458, + "grad_norm": 0.1312142312526703, + "learning_rate": 1e-06, + "loss": 0.009, + "num_tokens": 153164438.0, + "reward": 0.574776828289032, + "reward_std": 0.2070397585630417, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 685.5513916015625, + "completions/mean_terminated_length": 599.7047729492188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 2.373177842565598, + "grad_norm": 0.12271033227443695, + "learning_rate": 1e-06, + "loss": -0.0156, + "num_tokens": 153790932.0, + "reward": 0.578125, + "reward_std": 0.19411788880825043, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2977.0, + "completions/mean_length": 746.0714721679688, + "completions/mean_terminated_length": 630.0230712890625, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 2.3825072886297374, + "grad_norm": 0.11483033001422882, + "learning_rate": 1e-06, + "loss": -0.001, + "num_tokens": 154438316.0, + "reward": 0.5546875, + "reward_std": 0.18975545465946198, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 637.5111694335938, + "completions/mean_terminated_length": 582.614501953125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.3918367346938774, + "grad_norm": 0.11873350292444229, + "learning_rate": 1e-06, + "loss": 0.0108, + "num_tokens": 155043382.0, + "reward": 0.5479910969734192, + "reward_std": 0.16010111570358276, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3515.0, + "completions/mean_length": 642.818115234375, + "completions/mean_terminated_length": 588.0056762695312, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.4011661807580174, + "grad_norm": 0.1250615119934082, + "learning_rate": 1e-06, + "loss": 0.0035, + "num_tokens": 155648139.0, + "reward": 0.5859375, + "reward_std": 0.19685529172420502, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2073.0, + "completions/mean_length": 590.3928833007812, + "completions/mean_terminated_length": 534.748291015625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 2.4104956268221573, + "grad_norm": 0.12371426075696945, + "learning_rate": 1e-06, + "loss": -0.001, + "num_tokens": 156205979.0, + "reward": 0.6462053656578064, + "reward_std": 0.17224974930286407, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3599.0, + "completions/mean_length": 720.5100708007812, + "completions/mean_terminated_length": 595.4918823242188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 2.4198250728862973, + "grad_norm": 0.12302586436271667, + "learning_rate": 1e-06, + "loss": -0.0049, + "num_tokens": 156813876.0, + "reward": 0.6261160969734192, + "reward_std": 0.1852782517671585, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3948.0, + "completions/mean_length": 708.2031860351562, + "completions/mean_terminated_length": 634.8072509765625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 2.4291545189504373, + "grad_norm": 0.12513022124767303, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 157459050.0, + "reward": 0.5915178656578064, + "reward_std": 0.21432706713676453, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 725.5513916015625, + "completions/mean_terminated_length": 612.8142700195312, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 2.4384839650145773, + "grad_norm": 0.14809367060661316, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 158075256.0, + "reward": 0.5691964626312256, + "reward_std": 0.24089428782463074, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652786254883, + "step": 261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4066.0, + "completions/mean_length": 703.4732666015625, + "completions/mean_terminated_length": 589.9976806640625, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 2.4478134110787173, + "grad_norm": 0.1256195306777954, + "learning_rate": 1e-06, + "loss": -0.0071, + "num_tokens": 158681432.0, + "reward": 0.5479910969734192, + "reward_std": 0.18686267733573914, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2689.0, + "completions/mean_length": 672.0546875, + "completions/mean_terminated_length": 597.8756713867188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 2.4571428571428573, + "grad_norm": 0.13301654160022736, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 159288785.0, + "reward": 0.5368303656578064, + "reward_std": 0.2112138420343399, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2814.0, + "completions/mean_length": 743.8326416015625, + "completions/mean_terminated_length": 587.189208984375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 2.466472303206997, + "grad_norm": 0.13447239995002747, + "learning_rate": 1e-06, + "loss": -0.0087, + "num_tokens": 159893955.0, + "reward": 0.5479910969734192, + "reward_std": 0.2173650562763214, + "rewards/verify_math_reward/mean": 0.5479910969734192, + "rewards/verify_math_reward/std": 0.49796950817108154, + "step": 264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3562.0, + "completions/mean_length": 734.3594360351562, + "completions/mean_terminated_length": 593.6395263671875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 2.4758017492711373, + "grad_norm": 0.13209767639636993, + "learning_rate": 1e-06, + "loss": 0.0026, + "num_tokens": 160485693.0, + "reward": 0.6350446939468384, + "reward_std": 0.1871197670698166, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.4816865026950836, + "step": 265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3720.0, + "completions/mean_length": 639.0803833007812, + "completions/mean_terminated_length": 556.1142578125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.485131195335277, + "grad_norm": 0.12567594647407532, + "learning_rate": 1e-06, + "loss": -0.0075, + "num_tokens": 161051981.0, + "reward": 0.660714328289032, + "reward_std": 0.15989314019680023, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3693.0, + "completions/mean_length": 664.140625, + "completions/mean_terminated_length": 565.63720703125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.494460641399417, + "grad_norm": 0.1250281184911728, + "learning_rate": 1e-06, + "loss": -0.0122, + "num_tokens": 161630403.0, + "reward": 0.5948660969734192, + "reward_std": 0.19102534651756287, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4018.0, + "completions/mean_length": 601.2332763671875, + "completions/mean_terminated_length": 525.5199584960938, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.503790087463557, + "grad_norm": 0.13596193492412567, + "learning_rate": 1e-06, + "loss": 0.013, + "num_tokens": 162181124.0, + "reward": 0.6361607313156128, + "reward_std": 0.19456438720226288, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.014508928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3085.0, + "completions/mean_length": 592.1953125, + "completions/mean_terminated_length": 540.6104125976562, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.513119533527697, + "grad_norm": 0.12503011524677277, + "learning_rate": 1e-06, + "loss": 0.0043, + "num_tokens": 162752875.0, + "reward": 0.640625, + "reward_std": 0.17171843349933624, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 674.1317138671875, + "completions/mean_terminated_length": 575.9150390625, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 2.522448979591837, + "grad_norm": 0.1236991137266159, + "learning_rate": 1e-06, + "loss": 0.0009, + "num_tokens": 163345905.0, + "reward": 0.5546875, + "reward_std": 0.17851974070072174, + "rewards/verify_math_reward/mean": 0.5546875, + "rewards/verify_math_reward/std": 0.4972778558731079, + "step": 270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 622.8705444335938, + "completions/mean_terminated_length": 559.7227172851562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 2.5317784256559768, + "grad_norm": 0.124686598777771, + "learning_rate": 1e-06, + "loss": 0.017, + "num_tokens": 163925213.0, + "reward": 0.6049107313156128, + "reward_std": 0.1755893975496292, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3927.0, + "completions/mean_length": 663.7388916015625, + "completions/mean_terminated_length": 577.3432006835938, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 2.5411078717201168, + "grad_norm": 0.12142544984817505, + "learning_rate": 1e-06, + "loss": 0.0011, + "num_tokens": 164519675.0, + "reward": 0.5915178656578064, + "reward_std": 0.18468162417411804, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.022321428571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2865.0, + "completions/mean_length": 634.958740234375, + "completions/mean_terminated_length": 555.939453125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 2.5504373177842563, + "grad_norm": 0.1158989742398262, + "learning_rate": 1e-06, + "loss": 0.0051, + "num_tokens": 165100822.0, + "reward": 0.6238839626312256, + "reward_std": 0.14083585143089294, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 616.0770263671875, + "completions/mean_terminated_length": 544.734619140625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.5597667638483967, + "grad_norm": 0.11704935133457184, + "learning_rate": 1e-06, + "loss": -0.0003, + "num_tokens": 165664107.0, + "reward": 0.6149553656578064, + "reward_std": 0.15495911240577698, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3581.0, + "completions/mean_length": 715.4029541015625, + "completions/mean_terminated_length": 622.3589477539062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 2.5690962099125363, + "grad_norm": 0.13065844774246216, + "learning_rate": 1e-06, + "loss": -0.0013, + "num_tokens": 166290828.0, + "reward": 0.5714285969734192, + "reward_std": 0.20854565501213074, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 738.8047485351562, + "completions/mean_terminated_length": 642.4442749023438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 2.5784256559766763, + "grad_norm": 0.11663296073675156, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 166934365.0, + "reward": 0.5725446939468384, + "reward_std": 0.18829230964183807, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3474.0, + "completions/mean_length": 653.8404541015625, + "completions/mean_terminated_length": 579.2667846679688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.5877551020408163, + "grad_norm": 0.14339974522590637, + "learning_rate": 1e-06, + "loss": -0.0011, + "num_tokens": 167521038.0, + "reward": 0.637276828289032, + "reward_std": 0.21463198959827423, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3880.0, + "completions/mean_length": 715.4174194335938, + "completions/mean_terminated_length": 594.2635498046875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.5970845481049563, + "grad_norm": 0.1284037083387375, + "learning_rate": 1e-06, + "loss": -0.0101, + "num_tokens": 168124844.0, + "reward": 0.6618303656578064, + "reward_std": 0.18186122179031372, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 695.4699096679688, + "completions/mean_terminated_length": 629.7030639648438, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 2.6064139941690962, + "grad_norm": 0.11293873935937881, + "learning_rate": 1e-06, + "loss": -0.0033, + "num_tokens": 168776033.0, + "reward": 0.5848214626312256, + "reward_std": 0.16645735502243042, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3984.0, + "completions/mean_length": 678.2489013671875, + "completions/mean_terminated_length": 559.8510131835938, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 2.6157434402332362, + "grad_norm": 0.13129226863384247, + "learning_rate": 1e-06, + "loss": 0.0016, + "num_tokens": 169359400.0, + "reward": 0.6305803656578064, + "reward_std": 0.15570586919784546, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2915.0, + "completions/mean_length": 728.3560791015625, + "completions/mean_terminated_length": 643.5869140625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 2.6250728862973762, + "grad_norm": 0.11450564861297607, + "learning_rate": 1e-06, + "loss": -0.002, + "num_tokens": 170017567.0, + "reward": 0.5725446939468384, + "reward_std": 0.18738332390785217, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3966.0, + "completions/mean_length": 688.8426513671875, + "completions/mean_terminated_length": 574.8777465820312, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 2.6344023323615158, + "grad_norm": 0.12026088684797287, + "learning_rate": 1e-06, + "loss": 0.0058, + "num_tokens": 170599610.0, + "reward": 0.5837053656578064, + "reward_std": 0.17577557265758514, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3024.0, + "completions/mean_length": 742.7489013671875, + "completions/mean_terminated_length": 614.52490234375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 2.643731778425656, + "grad_norm": 0.12936659157276154, + "learning_rate": 1e-06, + "loss": -0.0053, + "num_tokens": 171220313.0, + "reward": 0.609375, + "reward_std": 0.212449848651886, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 692.2277221679688, + "completions/mean_terminated_length": 622.4464721679688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 2.6530612244897958, + "grad_norm": 0.13705040514469147, + "learning_rate": 1e-06, + "loss": 0.0116, + "num_tokens": 171856861.0, + "reward": 0.6015625, + "reward_std": 0.23453807830810547, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3671.0, + "completions/mean_length": 691.8114013671875, + "completions/mean_terminated_length": 594.1021728515625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.6623906705539357, + "grad_norm": 0.14273203909397125, + "learning_rate": 1e-06, + "loss": -0.0028, + "num_tokens": 172473908.0, + "reward": 0.5736607313156128, + "reward_std": 0.19261160492897034, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4031.0, + "completions/mean_length": 712.622802734375, + "completions/mean_terminated_length": 611.5103149414062, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.6717201166180757, + "grad_norm": 0.11666254699230194, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 173103362.0, + "reward": 0.5803571939468384, + "reward_std": 0.1758490353822708, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2968.0, + "completions/mean_length": 731.0870971679688, + "completions/mean_terminated_length": 582.0582885742188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 2.6810495626822157, + "grad_norm": 0.11857963353395462, + "learning_rate": 1e-06, + "loss": -0.0181, + "num_tokens": 173699544.0, + "reward": 0.551339328289032, + "reward_std": 0.16458265483379364, + "rewards/verify_math_reward/mean": 0.5513392686843872, + "rewards/verify_math_reward/std": 0.4976350665092468, + "step": 287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4067.0, + "completions/mean_length": 760.5123291015625, + "completions/mean_terminated_length": 628.9500732421875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 2.6903790087463557, + "grad_norm": 0.12353627383708954, + "learning_rate": 1e-06, + "loss": -0.0097, + "num_tokens": 174331403.0, + "reward": 0.5892857313156128, + "reward_std": 0.18765361607074738, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3551.0, + "completions/mean_length": 738.036865234375, + "completions/mean_terminated_length": 609.6326904296875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 2.6997084548104957, + "grad_norm": 0.13085076212882996, + "learning_rate": 1e-06, + "loss": -0.0098, + "num_tokens": 174942580.0, + "reward": 0.559151828289032, + "reward_std": 0.18344561755657196, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2508.0, + "completions/mean_length": 766.6295166015625, + "completions/mean_terminated_length": 647.3109741210938, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 2.7090379008746357, + "grad_norm": 0.1308446228504181, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 175595744.0, + "reward": 0.5736607313156128, + "reward_std": 0.24472320079803467, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2632.0, + "completions/mean_length": 692.6875610351562, + "completions/mean_terminated_length": 586.9459228515625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 2.7183673469387752, + "grad_norm": 0.13336847722530365, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 176197424.0, + "reward": 0.6004464626312256, + "reward_std": 0.2052348405122757, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2644.0, + "completions/mean_length": 672.328125, + "completions/mean_terminated_length": 590.1599731445312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 2.7276967930029157, + "grad_norm": 0.12759360671043396, + "learning_rate": 1e-06, + "loss": -0.0024, + "num_tokens": 176799726.0, + "reward": 0.6238839626312256, + "reward_std": 0.20237983763217926, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4081.0, + "completions/mean_length": 648.7221069335938, + "completions/mean_terminated_length": 590.0283813476562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 2.7370262390670552, + "grad_norm": 0.13208197057247162, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 177408629.0, + "reward": 0.637276828289032, + "reward_std": 0.19456368684768677, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3258.0, + "completions/mean_length": 622.5803833007812, + "completions/mean_terminated_length": 559.42724609375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 2.746355685131195, + "grad_norm": 0.13532754778862, + "learning_rate": 1e-06, + "loss": -0.0015, + "num_tokens": 177990973.0, + "reward": 0.6495535969734192, + "reward_std": 0.20233957469463348, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2364.0, + "completions/mean_length": 733.8928833007812, + "completions/mean_terminated_length": 613.401123046875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.755685131195335, + "grad_norm": 0.13921375572681427, + "learning_rate": 1e-06, + "loss": -0.0132, + "num_tokens": 178617381.0, + "reward": 0.5770089626312256, + "reward_std": 0.22303910553455353, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 758.0045166015625, + "completions/mean_terminated_length": 618.2744140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.765014577259475, + "grad_norm": 0.12173353880643845, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 179243265.0, + "reward": 0.5948660969734192, + "reward_std": 0.1637125015258789, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0267857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3279.0, + "completions/mean_length": 692.1317138671875, + "completions/mean_terminated_length": 598.4472045898438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 2.774344023323615, + "grad_norm": 0.11701571941375732, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 179850327.0, + "reward": 0.6752232313156128, + "reward_std": 0.15488353371620178, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2600.0, + "completions/mean_length": 727.0614013671875, + "completions/mean_terminated_length": 622.3878173828125, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 2.783673469387755, + "grad_norm": 0.13242870569229126, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 180486798.0, + "reward": 0.5301339626312256, + "reward_std": 0.20163418352603912, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936988949775696, + "step": 298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 686.1339721679688, + "completions/mean_terminated_length": 588.2617797851562, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 2.793002915451895, + "grad_norm": 0.11804566532373428, + "learning_rate": 1e-06, + "loss": 0.0051, + "num_tokens": 181088902.0, + "reward": 0.613839328289032, + "reward_std": 0.18452827632427216, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2634.0, + "completions/mean_length": 642.2199096679688, + "completions/mean_terminated_length": 595.3359985351562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 2.8023323615160347, + "grad_norm": 0.12089167535305023, + "learning_rate": 1e-06, + "loss": 0.0076, + "num_tokens": 181702043.0, + "reward": 0.65625, + "reward_std": 0.16660960018634796, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.016741071428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2934.0, + "completions/mean_length": 608.560302734375, + "completions/mean_terminated_length": 549.1828002929688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 2.811661807580175, + "grad_norm": 0.13452772796154022, + "learning_rate": 1e-06, + "loss": 0.0073, + "num_tokens": 182266537.0, + "reward": 0.6350446939468384, + "reward_std": 0.18930783867835999, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0189732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3091.0, + "completions/mean_length": 622.9342041015625, + "completions/mean_terminated_length": 555.7644653320312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 2.8209912536443147, + "grad_norm": 0.13278339803218842, + "learning_rate": 1e-06, + "loss": 0.0112, + "num_tokens": 182835126.0, + "reward": 0.6729910969734192, + "reward_std": 0.16596952080726624, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.017857142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2644.0, + "completions/mean_length": 656.6038208007812, + "completions/mean_terminated_length": 594.0692749023438, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 2.8303206997084547, + "grad_norm": 0.1224234402179718, + "learning_rate": 1e-06, + "loss": -0.0035, + "num_tokens": 183459755.0, + "reward": 0.574776828289032, + "reward_std": 0.17394152283668518, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3826.0, + "completions/mean_length": 671.6038208007812, + "completions/mean_terminated_length": 597.4150390625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.8396501457725947, + "grad_norm": 0.12831582129001617, + "learning_rate": 1e-06, + "loss": -0.0007, + "num_tokens": 184071240.0, + "reward": 0.6071428656578064, + "reward_std": 0.18155770003795624, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2412.0, + "completions/mean_length": 695.2857666015625, + "completions/mean_terminated_length": 573.410400390625, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 2.8489795918367347, + "grad_norm": 0.13423438370227814, + "learning_rate": 1e-06, + "loss": 0.0017, + "num_tokens": 184655200.0, + "reward": 0.6350446939468384, + "reward_std": 0.19391925632953644, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3999.0, + "completions/mean_length": 719.5324096679688, + "completions/mean_terminated_length": 622.6188354492188, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.8583090379008746, + "grad_norm": 0.13095982372760773, + "learning_rate": 1e-06, + "loss": 0.0261, + "num_tokens": 185284581.0, + "reward": 0.6350446939468384, + "reward_std": 0.18915177881717682, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.013392857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3411.0, + "completions/mean_length": 632.029052734375, + "completions/mean_terminated_length": 585.0068359375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 2.8676384839650146, + "grad_norm": 0.1279597133398056, + "learning_rate": 1e-06, + "loss": -0.0068, + "num_tokens": 185899263.0, + "reward": 0.6439732313156128, + "reward_std": 0.20079933106899261, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3013.0, + "completions/mean_length": 674.880615234375, + "completions/mean_terminated_length": 600.7628173828125, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 2.8769679300291546, + "grad_norm": 0.11159589141607285, + "learning_rate": 1e-06, + "loss": 0.0003, + "num_tokens": 186502884.0, + "reward": 0.6071428656578064, + "reward_std": 0.16799716651439667, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865824937820435, + "step": 308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3311.0, + "completions/mean_length": 750.2176513671875, + "completions/mean_terminated_length": 630.3109741210938, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 2.8862973760932946, + "grad_norm": 0.12934349477291107, + "learning_rate": 1e-06, + "loss": -0.0089, + "num_tokens": 187135599.0, + "reward": 0.574776828289032, + "reward_std": 0.17912591993808746, + "rewards/verify_math_reward/mean": 0.5747767686843872, + "rewards/verify_math_reward/std": 0.49465295672416687, + "step": 309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2810.0, + "completions/mean_length": 706.4866333007812, + "completions/mean_terminated_length": 609.1986083984375, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 2.8956268221574346, + "grad_norm": 0.12779387831687927, + "learning_rate": 1e-06, + "loss": -0.0157, + "num_tokens": 187749667.0, + "reward": 0.613839328289032, + "reward_std": 0.2099011391401291, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3368.0, + "completions/mean_length": 730.6652221679688, + "completions/mean_terminated_length": 614.0831298828125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 2.904956268221574, + "grad_norm": 0.133760005235672, + "learning_rate": 1e-06, + "loss": -0.0189, + "num_tokens": 188372759.0, + "reward": 0.606026828289032, + "reward_std": 0.21034803986549377, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3346.0, + "completions/mean_length": 779.9877319335938, + "completions/mean_terminated_length": 669.0714721679688, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 2.914285714285714, + "grad_norm": 0.12504105269908905, + "learning_rate": 1e-06, + "loss": 0.0069, + "num_tokens": 189038772.0, + "reward": 0.5368303656578064, + "reward_std": 0.18644827604293823, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 787.1998291015625, + "completions/mean_terminated_length": 636.624267578125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 2.923615160349854, + "grad_norm": 0.1323768049478531, + "learning_rate": 1e-06, + "loss": 0.0101, + "num_tokens": 189683575.0, + "reward": 0.5345982313156128, + "reward_std": 0.19178743660449982, + "rewards/verify_math_reward/mean": 0.5345982313156128, + "rewards/verify_math_reward/std": 0.4990801215171814, + "step": 313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 753.294677734375, + "completions/mean_terminated_length": 625.4739379882812, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 2.932944606413994, + "grad_norm": 0.13254469633102417, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 190313623.0, + "reward": 0.6127232313156128, + "reward_std": 0.18667609989643097, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3437.0, + "completions/mean_length": 732.1339721679688, + "completions/mean_terminated_length": 615.602783203125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 2.942274052478134, + "grad_norm": 0.11032266914844513, + "learning_rate": 1e-06, + "loss": 0.0002, + "num_tokens": 190932255.0, + "reward": 0.6171875, + "reward_std": 0.14984887838363647, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3254.0, + "completions/mean_length": 714.7489013671875, + "completions/mean_terminated_length": 617.6980590820312, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 2.951603498542274, + "grad_norm": 0.12519097328186035, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 191570182.0, + "reward": 0.5848214626312256, + "reward_std": 0.18576611578464508, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 769.6652221679688, + "completions/mean_terminated_length": 674.1905517578125, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 2.960932944606414, + "grad_norm": 0.12233088165521622, + "learning_rate": 1e-06, + "loss": 0.0038, + "num_tokens": 192244866.0, + "reward": 0.590401828289032, + "reward_std": 0.17025348544120789, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3460.0, + "completions/mean_length": 694.8158569335938, + "completions/mean_terminated_length": 593.1712646484375, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 2.970262390670554, + "grad_norm": 0.14599263668060303, + "learning_rate": 1e-06, + "loss": -0.0225, + "num_tokens": 192853869.0, + "reward": 0.613839328289032, + "reward_std": 0.22140008211135864, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4094.0, + "completions/mean_length": 669.122802734375, + "completions/mean_terminated_length": 582.8626708984375, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 2.979591836734694, + "grad_norm": 0.13434584438800812, + "learning_rate": 1e-06, + "loss": -0.0098, + "num_tokens": 193451187.0, + "reward": 0.5814732313156128, + "reward_std": 0.1926429718732834, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2348.0, + "completions/mean_length": 667.9967041015625, + "completions/mean_terminated_length": 585.7245483398438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 2.9889212827988336, + "grad_norm": 0.12504513561725616, + "learning_rate": 1e-06, + "loss": -0.0107, + "num_tokens": 194045888.0, + "reward": 0.6082589626312256, + "reward_std": 0.15341928601264954, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.021205357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3788.0, + "completions/mean_length": 646.3817138671875, + "completions/mean_terminated_length": 571.646484375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 2.9982507288629736, + "grad_norm": 0.13025930523872375, + "learning_rate": 1e-06, + "loss": 0.0046, + "num_tokens": 194643726.0, + "reward": 0.6082589626312256, + "reward_std": 0.19343574345111847, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3061.0, + "completions/mean_length": 701.8683471679688, + "completions/mean_terminated_length": 584.2886962890625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.00932944606414, + "grad_norm": 0.13409017026424408, + "learning_rate": 1e-06, + "loss": -0.0136, + "num_tokens": 195240144.0, + "reward": 0.590401828289032, + "reward_std": 0.23525162041187286, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3124.0, + "completions/mean_length": 705.5402221679688, + "completions/mean_terminated_length": 575.8933715820312, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.01865889212828, + "grad_norm": 0.1364138275384903, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 195833588.0, + "reward": 0.625, + "reward_std": 0.19452865421772003, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3178.0, + "completions/mean_length": 727.4085083007812, + "completions/mean_terminated_length": 602.6458129882812, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.02798833819242, + "grad_norm": 0.13122336566448212, + "learning_rate": 1e-06, + "loss": -0.0229, + "num_tokens": 196447274.0, + "reward": 0.5736607313156128, + "reward_std": 0.15405938029289246, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4068.0, + "completions/mean_length": 711.8225708007812, + "completions/mean_terminated_length": 622.6632690429688, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.03731778425656, + "grad_norm": 0.1151110902428627, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 197095235.0, + "reward": 0.5457589626312256, + "reward_std": 0.1594706028699875, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 668.1350708007812, + "completions/mean_terminated_length": 565.693115234375, + "completions/min_length": 94.0, + "completions/min_terminated_length": 94.0, + "epoch": 3.0466472303206995, + "grad_norm": 0.1365208923816681, + "learning_rate": 1e-06, + "loss": 0.0099, + "num_tokens": 197681964.0, + "reward": 0.6395089626312256, + "reward_std": 0.19643980264663696, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3048.0, + "completions/mean_length": 681.6317138671875, + "completions/mean_terminated_length": 595.6864624023438, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 3.0559766763848395, + "grad_norm": 0.1353214979171753, + "learning_rate": 1e-06, + "loss": 0.0077, + "num_tokens": 198293226.0, + "reward": 0.6004464626312256, + "reward_std": 0.17266343533992767, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2767.0, + "completions/mean_length": 765.779052734375, + "completions/mean_terminated_length": 601.9976196289062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.0653061224489795, + "grad_norm": 0.12885475158691406, + "learning_rate": 1e-06, + "loss": -0.013, + "num_tokens": 198918780.0, + "reward": 0.5412946939468384, + "reward_std": 0.17731650173664093, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3713.0, + "completions/mean_length": 693.1160888671875, + "completions/mean_terminated_length": 587.3878173828125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.0746355685131195, + "grad_norm": 0.12753824889659882, + "learning_rate": 1e-06, + "loss": -0.0068, + "num_tokens": 199517132.0, + "reward": 0.6149553656578064, + "reward_std": 0.14617222547531128, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3190.0, + "completions/mean_length": 690.5245971679688, + "completions/mean_terminated_length": 600.8041381835938, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 3.0839650145772595, + "grad_norm": 0.1325608193874359, + "learning_rate": 1e-06, + "loss": -0.0185, + "num_tokens": 200129474.0, + "reward": 0.6651785969734192, + "reward_std": 0.17841166257858276, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3476.0, + "completions/mean_length": 683.2388916015625, + "completions/mean_terminated_length": 581.248291015625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 3.0932944606413995, + "grad_norm": 0.13901954889297485, + "learning_rate": 1e-06, + "loss": -0.0004, + "num_tokens": 200720280.0, + "reward": 0.606026828289032, + "reward_std": 0.209706112742424, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2190.0, + "completions/mean_length": 686.6596069335938, + "completions/mean_terminated_length": 588.802490234375, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 3.1026239067055394, + "grad_norm": 0.12468723952770233, + "learning_rate": 1e-06, + "loss": -0.0038, + "num_tokens": 201307479.0, + "reward": 0.6986607313156128, + "reward_std": 0.16086140275001526, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3565.0, + "completions/mean_length": 670.6283569335938, + "completions/mean_terminated_length": 551.9664916992188, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 3.1119533527696794, + "grad_norm": 0.13708895444869995, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 201859106.0, + "reward": 0.65625, + "reward_std": 0.1816764622926712, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3241.0, + "completions/mean_length": 691.099365234375, + "completions/mean_terminated_length": 601.39404296875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.1212827988338194, + "grad_norm": 0.1290242075920105, + "learning_rate": 1e-06, + "loss": 0.0053, + "num_tokens": 202480763.0, + "reward": 0.6238839626312256, + "reward_std": 0.16416895389556885, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.48468026518821716, + "step": 334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3713.0, + "completions/mean_length": 751.4922485351562, + "completions/mean_terminated_length": 599.291748046875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 3.130612244897959, + "grad_norm": 0.13000929355621338, + "learning_rate": 1e-06, + "loss": -0.0057, + "num_tokens": 203091556.0, + "reward": 0.5602678656578064, + "reward_std": 0.17821593582630157, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2734.0, + "completions/mean_length": 712.9642944335938, + "completions/mean_terminated_length": 591.7225341796875, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 3.139941690962099, + "grad_norm": 0.13226984441280365, + "learning_rate": 1e-06, + "loss": -0.0157, + "num_tokens": 203681964.0, + "reward": 0.65625, + "reward_std": 0.16491642594337463, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3666.0, + "completions/mean_length": 761.5011596679688, + "completions/mean_terminated_length": 633.9942016601562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.149271137026239, + "grad_norm": 0.12499138712882996, + "learning_rate": 1e-06, + "loss": -0.0079, + "num_tokens": 204319189.0, + "reward": 0.645089328289032, + "reward_std": 0.19951984286308289, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2691.0, + "completions/mean_length": 689.1685791015625, + "completions/mean_terminated_length": 591.3834838867188, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 3.158600583090379, + "grad_norm": 0.1238812655210495, + "learning_rate": 1e-06, + "loss": 0.0042, + "num_tokens": 204923796.0, + "reward": 0.6551339626312256, + "reward_std": 0.16837625205516815, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2879.0, + "completions/mean_length": 843.3516235351562, + "completions/mean_terminated_length": 626.5083618164062, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.167930029154519, + "grad_norm": 0.11512185633182526, + "learning_rate": 1e-06, + "loss": -0.0156, + "num_tokens": 205540407.0, + "reward": 0.5290178656578064, + "reward_std": 0.1513572782278061, + "rewards/verify_math_reward/mean": 0.5290178656578064, + "rewards/verify_math_reward/std": 0.49943605065345764, + "step": 339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3706.0, + "completions/mean_length": 689.1574096679688, + "completions/mean_terminated_length": 587.3436889648438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 3.177259475218659, + "grad_norm": 0.1282913088798523, + "learning_rate": 1e-06, + "loss": 0.0044, + "num_tokens": 206139876.0, + "reward": 0.6395089626312256, + "reward_std": 0.1639835089445114, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0234375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 652.5402221679688, + "completions/mean_terminated_length": 569.8971557617188, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.186588921282799, + "grad_norm": 0.1217779591679573, + "learning_rate": 1e-06, + "loss": -0.0044, + "num_tokens": 206721136.0, + "reward": 0.65625, + "reward_std": 0.14496827125549316, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2097.0, + "completions/mean_length": 749.0402221679688, + "completions/mean_terminated_length": 600.8065185546875, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 3.195918367346939, + "grad_norm": 0.13950563967227936, + "learning_rate": 1e-06, + "loss": -0.0148, + "num_tokens": 207328548.0, + "reward": 0.6082589626312256, + "reward_std": 0.23296253383159637, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3329.0, + "completions/mean_length": 709.9475708007812, + "completions/mean_terminated_length": 600.7200317382812, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 3.205247813411079, + "grad_norm": 0.14613395929336548, + "learning_rate": 1e-06, + "loss": -0.0065, + "num_tokens": 207943061.0, + "reward": 0.5803571939468384, + "reward_std": 0.2227773815393448, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4092.0, + "completions/mean_length": 790.3058471679688, + "completions/mean_terminated_length": 663.9003295898438, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 3.2145772594752184, + "grad_norm": 0.13230329751968384, + "learning_rate": 1e-06, + "loss": -0.0084, + "num_tokens": 208609151.0, + "reward": 0.5680803656578064, + "reward_std": 0.21868480741977692, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0200892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 642.0513916015625, + "completions/mean_terminated_length": 571.241455078125, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 3.2239067055393584, + "grad_norm": 0.12874354422092438, + "learning_rate": 1e-06, + "loss": -0.0089, + "num_tokens": 209195853.0, + "reward": 0.6674107313156128, + "reward_std": 0.16470026969909668, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2255.0, + "completions/mean_length": 684.8281860351562, + "completions/mean_terminated_length": 558.4884033203125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.2332361516034984, + "grad_norm": 0.12298430502414703, + "learning_rate": 1e-06, + "loss": -0.0103, + "num_tokens": 209763883.0, + "reward": 0.6127232313156128, + "reward_std": 0.1428307294845581, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3553.0, + "completions/mean_length": 703.7879638671875, + "completions/mean_terminated_length": 586.2748413085938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 3.2425655976676384, + "grad_norm": 0.12920339405536652, + "learning_rate": 1e-06, + "loss": -0.0002, + "num_tokens": 210362469.0, + "reward": 0.5993303656578064, + "reward_std": 0.1734083741903305, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.025669642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2806.0, + "completions/mean_length": 677.5480346679688, + "completions/mean_terminated_length": 587.4857177734375, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 3.2518950437317784, + "grad_norm": 0.13406428694725037, + "learning_rate": 1e-06, + "loss": 0.0111, + "num_tokens": 210956608.0, + "reward": 0.6428571939468384, + "reward_std": 0.19538643956184387, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3504.0, + "completions/mean_length": 684.5123291015625, + "completions/mean_terminated_length": 558.160888671875, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 3.2612244897959184, + "grad_norm": 0.14350523054599762, + "learning_rate": 1e-06, + "loss": -0.009, + "num_tokens": 211522875.0, + "reward": 0.6830357313156128, + "reward_std": 0.18952831625938416, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 706.3605346679688, + "completions/mean_terminated_length": 564.4686279296875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.2705539358600584, + "grad_norm": 0.13202689588069916, + "learning_rate": 1e-06, + "loss": -0.0036, + "num_tokens": 212099918.0, + "reward": 0.6339285969734192, + "reward_std": 0.1633320301771164, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3190.0, + "completions/mean_length": 771.1730346679688, + "completions/mean_terminated_length": 595.3595581054688, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 3.2798833819241984, + "grad_norm": 0.11288853734731674, + "learning_rate": 1e-06, + "loss": -0.0131, + "num_tokens": 212692793.0, + "reward": 0.6194196939468384, + "reward_std": 0.13834992051124573, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 715.8214721679688, + "completions/mean_terminated_length": 602.7589111328125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 3.2892128279883384, + "grad_norm": 0.12326151132583618, + "learning_rate": 1e-06, + "loss": -0.0243, + "num_tokens": 213307697.0, + "reward": 0.5915178656578064, + "reward_std": 0.16897033154964447, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2603.0, + "completions/mean_length": 739.9408569335938, + "completions/mean_terminated_length": 607.5672607421875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 3.298542274052478, + "grad_norm": 0.1377718150615692, + "learning_rate": 1e-06, + "loss": 0.0052, + "num_tokens": 213927204.0, + "reward": 0.5368303656578064, + "reward_std": 0.20831994712352753, + "rewards/verify_math_reward/mean": 0.5368303656578064, + "rewards/verify_math_reward/std": 0.49892017245292664, + "step": 353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3698.0, + "completions/mean_length": 812.130615234375, + "completions/mean_terminated_length": 650.6287841796875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 3.307871720116618, + "grad_norm": 0.13006362318992615, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 214575897.0, + "reward": 0.5412946939468384, + "reward_std": 0.1814585030078888, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2778.0, + "completions/mean_length": 803.1964721679688, + "completions/mean_terminated_length": 624.9976196289062, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 3.317201166180758, + "grad_norm": 0.1344502717256546, + "learning_rate": 1e-06, + "loss": -0.0101, + "num_tokens": 215201073.0, + "reward": 0.5758928656578064, + "reward_std": 0.20665815472602844, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448275566101074, + "step": 355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3572.0, + "completions/mean_length": 754.8392944335938, + "completions/mean_terminated_length": 619.0197143554688, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 3.326530612244898, + "grad_norm": 0.13514229655265808, + "learning_rate": 1e-06, + "loss": -0.0009, + "num_tokens": 215827457.0, + "reward": 0.5792410969734192, + "reward_std": 0.18874765932559967, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3045.0, + "completions/mean_length": 778.068115234375, + "completions/mean_terminated_length": 614.8910522460938, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 3.335860058309038, + "grad_norm": 0.1412975788116455, + "learning_rate": 1e-06, + "loss": -0.0097, + "num_tokens": 216444142.0, + "reward": 0.5580357313156128, + "reward_std": 0.20741769671440125, + "rewards/verify_math_reward/mean": 0.5580357313156128, + "rewards/verify_math_reward/std": 0.49689778685569763, + "step": 357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 776.2600708007812, + "completions/mean_terminated_length": 669.171630859375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.345189504373178, + "grad_norm": 0.1185586154460907, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 217121199.0, + "reward": 0.5848214626312256, + "reward_std": 0.17618785798549652, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3623.0, + "completions/mean_length": 744.7813110351562, + "completions/mean_terminated_length": 588.1822509765625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.354518950437318, + "grad_norm": 0.13598987460136414, + "learning_rate": 1e-06, + "loss": 0.0042, + "num_tokens": 217711339.0, + "reward": 0.6439732313156128, + "reward_std": 0.17852112650871277, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909069061279297, + "step": 359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4065.0, + "completions/mean_length": 770.849365234375, + "completions/mean_terminated_length": 619.52978515625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 3.363848396501458, + "grad_norm": 0.12691107392311096, + "learning_rate": 1e-06, + "loss": -0.0116, + "num_tokens": 218345660.0, + "reward": 0.6037946939468384, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 360 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 804.091552734375, + "completions/mean_terminated_length": 617.757080078125, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 3.373177842565598, + "grad_norm": 0.126580610871315, + "learning_rate": 1e-06, + "loss": 0.0001, + "num_tokens": 218957126.0, + "reward": 0.5691964626312256, + "reward_std": 0.16819261014461517, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652786254883, + "step": 361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2645.0, + "completions/mean_length": 818.2188110351562, + "completions/mean_terminated_length": 640.8328857421875, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 3.3825072886297374, + "grad_norm": 0.12290062010288239, + "learning_rate": 1e-06, + "loss": 0.0, + "num_tokens": 219585282.0, + "reward": 0.6127232313156128, + "reward_std": 0.18193678557872772, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3306.0, + "completions/mean_length": 702.0022583007812, + "completions/mean_terminated_length": 580.3676147460938, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 3.3918367346938774, + "grad_norm": 0.14238247275352478, + "learning_rate": 1e-06, + "loss": -0.0047, + "num_tokens": 220171924.0, + "reward": 0.613839328289032, + "reward_std": 0.18550649285316467, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3450.0, + "completions/mean_length": 769.0904541015625, + "completions/mean_terminated_length": 605.4718627929688, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.4011661807580174, + "grad_norm": 0.1256083846092224, + "learning_rate": 1e-06, + "loss": -0.0184, + "num_tokens": 220785253.0, + "reward": 0.6462053656578064, + "reward_std": 0.1413230001926422, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4046.0, + "completions/mean_length": 799.5335083007812, + "completions/mean_terminated_length": 604.7068481445312, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 3.4104956268221573, + "grad_norm": 0.13669243454933167, + "learning_rate": 1e-06, + "loss": -0.0133, + "num_tokens": 221394035.0, + "reward": 0.5859375, + "reward_std": 0.1702534705400467, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.033482142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3142.0, + "completions/mean_length": 690.3739013671875, + "completions/mean_terminated_length": 572.3960571289062, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 3.4198250728862973, + "grad_norm": 0.14225222170352936, + "learning_rate": 1e-06, + "loss": -0.0027, + "num_tokens": 221974210.0, + "reward": 0.684151828289032, + "reward_std": 0.18603530526161194, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2425.0, + "completions/mean_length": 751.1506958007812, + "completions/mean_terminated_length": 594.8493041992188, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 3.4291545189504373, + "grad_norm": 0.1299821138381958, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 222572625.0, + "reward": 0.6004464626312256, + "reward_std": 0.1774769276380539, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3020.0, + "completions/mean_length": 691.0636596679688, + "completions/mean_terminated_length": 585.2716064453125, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 3.4384839650145773, + "grad_norm": 0.12235794216394424, + "learning_rate": 1e-06, + "loss": -0.013, + "num_tokens": 223163754.0, + "reward": 0.6618303656578064, + "reward_std": 0.16258454322814941, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.024553571428571397, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3107.0, + "completions/mean_length": 652.9609375, + "completions/mean_terminated_length": 566.2940063476562, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 3.4478134110787173, + "grad_norm": 0.13919448852539062, + "learning_rate": 1e-06, + "loss": -0.0046, + "num_tokens": 223743023.0, + "reward": 0.6796875596046448, + "reward_std": 0.17333675920963287, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3094.0, + "completions/mean_length": 680.2589721679688, + "completions/mean_terminated_length": 582.2181396484375, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 3.4571428571428573, + "grad_norm": 0.12423066794872284, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 224339439.0, + "reward": 0.6227678656578064, + "reward_std": 0.16769154369831085, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3473.0, + "completions/mean_length": 794.4006958007812, + "completions/mean_terminated_length": 652.1897583007812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.466472303206997, + "grad_norm": 0.1391342729330063, + "learning_rate": 1e-06, + "loss": -0.0117, + "num_tokens": 224991174.0, + "reward": 0.5970982313156128, + "reward_std": 0.18621936440467834, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4096.0, + "completions/mean_length": 746.4263916015625, + "completions/mean_terminated_length": 622.3680419921875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 3.4758017492711373, + "grad_norm": 0.12099606543779373, + "learning_rate": 1e-06, + "loss": -0.0065, + "num_tokens": 225619756.0, + "reward": 0.5948660969734192, + "reward_std": 0.16548095643520355, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3272.0, + "completions/mean_length": 808.802490234375, + "completions/mean_terminated_length": 630.9070434570312, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.485131195335277, + "grad_norm": 0.12439883500337601, + "learning_rate": 1e-06, + "loss": -0.0174, + "num_tokens": 226234571.0, + "reward": 0.6272321939468384, + "reward_std": 0.17299722135066986, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0379464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 755.1183471679688, + "completions/mean_terminated_length": 623.3433837890625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 3.494460641399417, + "grad_norm": 0.11875832825899124, + "learning_rate": 1e-06, + "loss": 0.008, + "num_tokens": 226861997.0, + "reward": 0.6551339626312256, + "reward_std": 0.16018669307231903, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3489.0, + "completions/mean_length": 756.9141235351562, + "completions/mean_terminated_length": 617.1383666992188, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 3.503790087463557, + "grad_norm": 0.12686309218406677, + "learning_rate": 1e-06, + "loss": -0.0246, + "num_tokens": 227475128.0, + "reward": 0.645089328289032, + "reward_std": 0.17269553244113922, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 811.6886596679688, + "completions/mean_terminated_length": 642.0762939453125, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 3.513119533527697, + "grad_norm": 0.13370086252689362, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 228117617.0, + "reward": 0.5714285969734192, + "reward_std": 0.19313082098960876, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2335.0, + "completions/mean_length": 756.302490234375, + "completions/mean_terminated_length": 608.3904418945312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.522448979591837, + "grad_norm": 0.1330876648426056, + "learning_rate": 1e-06, + "loss": -0.0167, + "num_tokens": 228741232.0, + "reward": 0.5725446939468384, + "reward_std": 0.18257758021354675, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3415.0, + "completions/mean_length": 703.6652221679688, + "completions/mean_terminated_length": 594.2350463867188, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.5317784256559768, + "grad_norm": 0.11800918728113174, + "learning_rate": 1e-06, + "loss": 0.0047, + "num_tokens": 229350516.0, + "reward": 0.621651828289032, + "reward_std": 0.14992444217205048, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3395.0, + "completions/mean_length": 806.7199096679688, + "completions/mean_terminated_length": 644.9519653320312, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 3.5411078717201168, + "grad_norm": 0.14276005327701569, + "learning_rate": 1e-06, + "loss": -0.0268, + "num_tokens": 229993193.0, + "reward": 0.5848214626312256, + "reward_std": 0.21158403158187866, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 651.7142944335938, + "completions/mean_terminated_length": 544.6996459960938, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.5504373177842563, + "grad_norm": 0.13489356637001038, + "learning_rate": 1e-06, + "loss": -0.0008, + "num_tokens": 230560097.0, + "reward": 0.6071428656578064, + "reward_std": 0.15367820858955383, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 705.4955444335938, + "completions/mean_terminated_length": 596.1244506835938, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.5597667638483967, + "grad_norm": 0.14186830818653107, + "learning_rate": 1e-06, + "loss": 0.0002, + "num_tokens": 231160437.0, + "reward": 0.6328125, + "reward_std": 0.21117034554481506, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 811.3705444335938, + "completions/mean_terminated_length": 645.7913208007812, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 3.5690962099125363, + "grad_norm": 0.13240933418273926, + "learning_rate": 1e-06, + "loss": -0.0057, + "num_tokens": 231795641.0, + "reward": 0.5691964626312256, + "reward_std": 0.19171005487442017, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3423.0, + "completions/mean_length": 774.6897583007812, + "completions/mean_terminated_length": 611.3465576171875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 3.5784256559766763, + "grad_norm": 0.1344354897737503, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 232405531.0, + "reward": 0.613839328289032, + "reward_std": 0.20959803462028503, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 736.0100708007812, + "completions/mean_terminated_length": 635.5965576171875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.5877551020408163, + "grad_norm": 0.13360871374607086, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 233047668.0, + "reward": 0.6305803656578064, + "reward_std": 0.1902337223291397, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.48291724920272827, + "step": 384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3651.0, + "completions/mean_length": 783.1439819335938, + "completions/mean_terminated_length": 620.2166137695312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.5970845481049563, + "grad_norm": 0.13681259751319885, + "learning_rate": 1e-06, + "loss": -0.0095, + "num_tokens": 233668421.0, + "reward": 0.5848214626312256, + "reward_std": 0.1968969702720642, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.029017857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3249.0, + "completions/mean_length": 737.3170166015625, + "completions/mean_terminated_length": 636.9425048828125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 3.6064139941690962, + "grad_norm": 0.13389994204044342, + "learning_rate": 1e-06, + "loss": 0.0023, + "num_tokens": 234314313.0, + "reward": 0.6473214626312256, + "reward_std": 0.21556490659713745, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3528.0, + "completions/mean_length": 745.9732666015625, + "completions/mean_terminated_length": 581.2177734375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 3.6157434402332362, + "grad_norm": 0.12165497988462448, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 234891121.0, + "reward": 0.6417410969734192, + "reward_std": 0.17903897166252136, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 734.6920166015625, + "completions/mean_terminated_length": 589.9091796875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 3.6250728862973762, + "grad_norm": 0.14059767127037048, + "learning_rate": 1e-06, + "loss": 0.0012, + "num_tokens": 235488165.0, + "reward": 0.6149553656578064, + "reward_std": 0.17746874690055847, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3469.0, + "completions/mean_length": 745.9553833007812, + "completions/mean_terminated_length": 601.6577758789062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 3.6344023323615158, + "grad_norm": 0.1238742545247078, + "learning_rate": 1e-06, + "loss": -0.0115, + "num_tokens": 236113357.0, + "reward": 0.5948660969734192, + "reward_std": 0.17400823533535004, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3224.0, + "completions/mean_length": 767.8281860351562, + "completions/mean_terminated_length": 656.5051879882812, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 3.643731778425656, + "grad_norm": 0.13081805408000946, + "learning_rate": 1e-06, + "loss": -0.0322, + "num_tokens": 236772851.0, + "reward": 0.6171875, + "reward_std": 0.20974071323871613, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1949.0, + "completions/mean_length": 763.8392944335938, + "completions/mean_terminated_length": 612.20068359375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 3.6530612244897958, + "grad_norm": 0.14588944613933563, + "learning_rate": 1e-06, + "loss": 0.0054, + "num_tokens": 237388123.0, + "reward": 0.5680803656578064, + "reward_std": 0.16671909391880035, + "rewards/verify_math_reward/mean": 0.5680803656578064, + "rewards/verify_math_reward/std": 0.4956200420856476, + "step": 391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3377.0, + "completions/mean_length": 757.2879638671875, + "completions/mean_terminated_length": 584.8662109375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 3.6623906705539357, + "grad_norm": 0.13380998373031616, + "learning_rate": 1e-06, + "loss": -0.0175, + "num_tokens": 237969885.0, + "reward": 0.6506696939468384, + "reward_std": 0.16660889983177185, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4033.0, + "completions/mean_length": 705.9141235351562, + "completions/mean_terminated_length": 580.3553466796875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 3.6717201166180757, + "grad_norm": 0.1253570020198822, + "learning_rate": 1e-06, + "loss": -0.003, + "num_tokens": 238553496.0, + "reward": 0.6149553656578064, + "reward_std": 0.18460313975811005, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4051.0, + "completions/mean_length": 720.4531860351562, + "completions/mean_terminated_length": 583.2357788085938, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 3.6810495626822157, + "grad_norm": 0.1264265775680542, + "learning_rate": 1e-06, + "loss": -0.0041, + "num_tokens": 239142566.0, + "reward": 0.625, + "reward_std": 0.18730594217777252, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3665.0, + "completions/mean_length": 754.6842041015625, + "completions/mean_terminated_length": 598.5478515625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 3.6903790087463557, + "grad_norm": 0.1310785710811615, + "learning_rate": 1e-06, + "loss": -0.0154, + "num_tokens": 239737771.0, + "reward": 0.6540178656578064, + "reward_std": 0.17325684428215027, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3529.0, + "completions/mean_length": 876.2455444335938, + "completions/mean_terminated_length": 665.678955078125, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 3.6997084548104957, + "grad_norm": 0.12681274116039276, + "learning_rate": 1e-06, + "loss": -0.0111, + "num_tokens": 240402111.0, + "reward": 0.5301339626312256, + "reward_std": 0.18547147512435913, + "rewards/verify_math_reward/mean": 0.5301339030265808, + "rewards/verify_math_reward/std": 0.49936985969543457, + "step": 396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 667.7824096679688, + "completions/mean_terminated_length": 561.2669677734375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 3.7090379008746357, + "grad_norm": 0.12940621376037598, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 240985628.0, + "reward": 0.668526828289032, + "reward_std": 0.1590908169746399, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3782.0, + "completions/mean_length": 799.2589721679688, + "completions/mean_terminated_length": 637.1240844726562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 3.7183673469387752, + "grad_norm": 0.12849397957324982, + "learning_rate": 1e-06, + "loss": -0.0226, + "num_tokens": 241622580.0, + "reward": 0.606026828289032, + "reward_std": 0.17326006293296814, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3870.0, + "completions/mean_length": 792.8047485351562, + "completions/mean_terminated_length": 618.1351318359375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 3.7276967930029157, + "grad_norm": 0.1368076503276825, + "learning_rate": 1e-06, + "loss": 0.0065, + "num_tokens": 242234397.0, + "reward": 0.5502232313156128, + "reward_std": 0.18419378995895386, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2744.0, + "completions/mean_length": 750.0513916015625, + "completions/mean_terminated_length": 614.0371704101562, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 3.7370262390670552, + "grad_norm": 0.1369449347257614, + "learning_rate": 1e-06, + "loss": -0.0035, + "num_tokens": 242854931.0, + "reward": 0.5691964626312256, + "reward_std": 0.21545502543449402, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3763.0, + "completions/mean_length": 680.1796875, + "completions/mean_terminated_length": 557.7630004882812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.746355685131195, + "grad_norm": 0.12297762185335159, + "learning_rate": 1e-06, + "loss": -0.0285, + "num_tokens": 243425140.0, + "reward": 0.6741071939468384, + "reward_std": 0.1634521782398224, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3484.0, + "completions/mean_length": 852.9342041015625, + "completions/mean_terminated_length": 644.9465942382812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 3.755685131195335, + "grad_norm": 0.1394238919019699, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 244059201.0, + "reward": 0.6283482313156128, + "reward_std": 0.19027970731258392, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3385.0, + "completions/mean_length": 783.7756958007812, + "completions/mean_terminated_length": 628.9988403320312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 3.765014577259475, + "grad_norm": 0.12618130445480347, + "learning_rate": 1e-06, + "loss": -0.0225, + "num_tokens": 244694312.0, + "reward": 0.5948660969734192, + "reward_std": 0.18663515150547028, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3943.0, + "completions/mean_length": 803.2935791015625, + "completions/mean_terminated_length": 661.4656372070312, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 3.774344023323615, + "grad_norm": 0.13492321968078613, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 245347295.0, + "reward": 0.6194196939468384, + "reward_std": 0.20177754759788513, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3137.0, + "completions/mean_length": 703.8538208007812, + "completions/mean_terminated_length": 578.21875, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 3.783673469387755, + "grad_norm": 0.11796524375677109, + "learning_rate": 1e-06, + "loss": -0.0118, + "num_tokens": 245937228.0, + "reward": 0.65625, + "reward_std": 0.1425590217113495, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0279017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3520.0, + "completions/mean_length": 716.3426513671875, + "completions/mean_terminated_length": 619.3375244140625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 3.793002915451895, + "grad_norm": 0.1338043510913849, + "learning_rate": 1e-06, + "loss": -0.0004, + "num_tokens": 246560319.0, + "reward": 0.629464328289032, + "reward_std": 0.22202950716018677, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 858.6027221679688, + "completions/mean_terminated_length": 667.2671508789062, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.8023323615160347, + "grad_norm": 0.12642014026641846, + "learning_rate": 1e-06, + "loss": -0.0075, + "num_tokens": 247217323.0, + "reward": 0.5892857313156128, + "reward_std": 0.1669113039970398, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2653.0, + "completions/mean_length": 747.6105346679688, + "completions/mean_terminated_length": 574.68896484375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 3.811661807580175, + "grad_norm": 0.13587544858455658, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 247792414.0, + "reward": 0.621651828289032, + "reward_std": 0.17370514571666718, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2308.0, + "completions/mean_length": 773.2332763671875, + "completions/mean_terminated_length": 605.7315063476562, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 3.8209912536443147, + "grad_norm": 0.13434529304504395, + "learning_rate": 1e-06, + "loss": -0.0026, + "num_tokens": 248393487.0, + "reward": 0.613839328289032, + "reward_std": 0.16247648000717163, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3665.0, + "completions/mean_length": 774.4330444335938, + "completions/mean_terminated_length": 623.2765502929688, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 3.8303206997084547, + "grad_norm": 0.13005517423152924, + "learning_rate": 1e-06, + "loss": -0.0271, + "num_tokens": 249010507.0, + "reward": 0.6506696939468384, + "reward_std": 0.19110456109046936, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 936.55810546875, + "completions/mean_terminated_length": 713.8494262695312, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 3.8396501457725947, + "grad_norm": 0.1262396275997162, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 249708807.0, + "reward": 0.4843750298023224, + "reward_std": 0.19310013949871063, + "rewards/verify_math_reward/mean": 0.484375, + "rewards/verify_math_reward/std": 0.5000349283218384, + "step": 411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2598.0, + "completions/mean_length": 779.1082763671875, + "completions/mean_terminated_length": 611.9026489257812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.8489795918367347, + "grad_norm": 0.15091830492019653, + "learning_rate": 1e-06, + "loss": -0.0191, + "num_tokens": 250326296.0, + "reward": 0.5502232313156128, + "reward_std": 0.19767513871192932, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3112.0, + "completions/mean_length": 798.5658569335938, + "completions/mean_terminated_length": 578.7369384765625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 3.8583090379008746, + "grad_norm": 0.12895776331424713, + "learning_rate": 1e-06, + "loss": -0.0198, + "num_tokens": 250893003.0, + "reward": 0.6026785969734192, + "reward_std": 0.15860295295715332, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3987.0, + "completions/mean_length": 835.818115234375, + "completions/mean_terminated_length": 634.9537963867188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 3.8676384839650146, + "grad_norm": 0.13272640109062195, + "learning_rate": 1e-06, + "loss": -0.0148, + "num_tokens": 251520296.0, + "reward": 0.6517857313156128, + "reward_std": 0.17224019765853882, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3594.0, + "completions/mean_length": 748.3192138671875, + "completions/mean_terminated_length": 604.1234130859375, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 3.8769679300291546, + "grad_norm": 0.1382935494184494, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 252119854.0, + "reward": 0.6116071939468384, + "reward_std": 0.21237428486347198, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 822.4955444335938, + "completions/mean_terminated_length": 665.5204467773438, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 3.8862973760932946, + "grad_norm": 0.12975755333900452, + "learning_rate": 1e-06, + "loss": -0.013, + "num_tokens": 252768954.0, + "reward": 0.6015625, + "reward_std": 0.21440306305885315, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3475.0, + "completions/mean_length": 815.5156860351562, + "completions/mean_terminated_length": 613.4005126953125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 3.8956268221574346, + "grad_norm": 0.12999078631401062, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 253378456.0, + "reward": 0.5602678656578064, + "reward_std": 0.19915145635604858, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.036830357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3918.0, + "completions/mean_length": 759.974365234375, + "completions/mean_terminated_length": 632.4089965820312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 3.904956268221574, + "grad_norm": 0.13193072378635406, + "learning_rate": 1e-06, + "loss": -0.0032, + "num_tokens": 254021177.0, + "reward": 0.6328125, + "reward_std": 0.18794602155685425, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2361.0, + "completions/mean_length": 775.6428833007812, + "completions/mean_terminated_length": 620.4859619140625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 3.914285714285714, + "grad_norm": 0.14001573622226715, + "learning_rate": 1e-06, + "loss": 0.0, + "num_tokens": 254638129.0, + "reward": 0.5725446939468384, + "reward_std": 0.17810533940792084, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3936.0, + "completions/mean_length": 891.5614013671875, + "completions/mean_terminated_length": 673.858154296875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 3.923615160349854, + "grad_norm": 0.13152745366096497, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 255283120.0, + "reward": 0.5970982313156128, + "reward_std": 0.19107136130332947, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.49075525999069214, + "step": 420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.030133928571428603, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4077.0, + "completions/mean_length": 737.513427734375, + "completions/mean_terminated_length": 633.16455078125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.932944606413994, + "grad_norm": 0.13388368487358093, + "learning_rate": 1e-06, + "loss": 0.0014, + "num_tokens": 255926788.0, + "reward": 0.629464328289032, + "reward_std": 0.16841013729572296, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3894.0, + "completions/mean_length": 841.8594360351562, + "completions/mean_terminated_length": 677.8170776367188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 3.942274052478134, + "grad_norm": 0.1300029158592224, + "learning_rate": 1e-06, + "loss": -0.0057, + "num_tokens": 256595958.0, + "reward": 0.5859375, + "reward_std": 0.19936829805374146, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3005.0, + "completions/mean_length": 695.739990234375, + "completions/mean_terminated_length": 557.5180053710938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.951603498542274, + "grad_norm": 0.13235510885715485, + "learning_rate": 1e-06, + "loss": -0.0301, + "num_tokens": 257161205.0, + "reward": 0.7020089626312256, + "reward_std": 0.13790595531463623, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.03125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3845.0, + "completions/mean_length": 704.7745971679688, + "completions/mean_terminated_length": 595.3801879882812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 3.960932944606414, + "grad_norm": 0.12774226069450378, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 257760619.0, + "reward": 0.6774553656578064, + "reward_std": 0.1627379208803177, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0390625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2694.0, + "completions/mean_length": 747.3069458007812, + "completions/mean_terminated_length": 611.18115234375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 3.970262390670554, + "grad_norm": 0.1442898064851761, + "learning_rate": 1e-06, + "loss": -0.0285, + "num_tokens": 258376766.0, + "reward": 0.629464328289032, + "reward_std": 0.2099011391401291, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3606.0, + "completions/mean_length": 730.7623291015625, + "completions/mean_terminated_length": 581.7191162109375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 3.979591836734694, + "grad_norm": 0.14264552295207977, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 258960417.0, + "reward": 0.6238839626312256, + "reward_std": 0.19366355240345, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0323660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2921.0, + "completions/mean_length": 681.7154541015625, + "completions/mean_terminated_length": 567.5120849609375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 3.9889212827988336, + "grad_norm": 0.11849533021450043, + "learning_rate": 1e-06, + "loss": 0.0056, + "num_tokens": 259545202.0, + "reward": 0.6707589626312256, + "reward_std": 0.13557225465774536, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.051136363636363646, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3115.0, + "completions/mean_length": 769.1818237304688, + "completions/mean_terminated_length": 589.8922119140625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 3.9982507288629736, + "grad_norm": 0.13029074668884277, + "learning_rate": 1e-06, + "loss": -0.0045, + "num_tokens": 260138544.0, + "reward": 0.6662946939468384, + "reward_std": 0.16086068749427795, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 720.5658569335938, + "completions/mean_terminated_length": 599.5964965820312, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 4.0093294460641395, + "grad_norm": 0.1270969659090042, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 260745683.0, + "reward": 0.645089328289032, + "reward_std": 0.17750760912895203, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3923.0, + "completions/mean_length": 827.1607666015625, + "completions/mean_terminated_length": 621.6465454101562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.01865889212828, + "grad_norm": 0.13756363093852997, + "learning_rate": 1e-06, + "loss": -0.005, + "num_tokens": 261359411.0, + "reward": 0.6104910969734192, + "reward_std": 0.17863735556602478, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3725.0, + "completions/mean_length": 785.8627319335938, + "completions/mean_terminated_length": 598.4964599609375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 4.0279883381924195, + "grad_norm": 0.1424919068813324, + "learning_rate": 1e-06, + "loss": -0.0211, + "num_tokens": 261960632.0, + "reward": 0.6104910969734192, + "reward_std": 0.17765280604362488, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 835.1217041015625, + "completions/mean_terminated_length": 678.7520141601562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.03731778425656, + "grad_norm": 0.1345667988061905, + "learning_rate": 1e-06, + "loss": -0.018, + "num_tokens": 262632277.0, + "reward": 0.5859375, + "reward_std": 0.1917860209941864, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2559.0, + "completions/mean_length": 703.8538208007812, + "completions/mean_terminated_length": 553.618896484375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.0466472303206995, + "grad_norm": 0.14514771103858948, + "learning_rate": 1e-06, + "loss": -0.0048, + "num_tokens": 263194626.0, + "reward": 0.691964328289032, + "reward_std": 0.17179511487483978, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3971.0, + "completions/mean_length": 841.4777221679688, + "completions/mean_terminated_length": 653.1995239257812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.05597667638484, + "grad_norm": 0.13529257476329803, + "learning_rate": 1e-06, + "loss": -0.0145, + "num_tokens": 263842422.0, + "reward": 0.6026785969734192, + "reward_std": 0.19986502826213837, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3431.0, + "completions/mean_length": 794.2824096679688, + "completions/mean_terminated_length": 615.6011352539062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.0653061224489795, + "grad_norm": 0.11853773891925812, + "learning_rate": 1e-06, + "loss": -0.0196, + "num_tokens": 264453595.0, + "reward": 0.6428571939468384, + "reward_std": 0.14943519234657288, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3527.0, + "completions/mean_length": 833.7299194335938, + "completions/mean_terminated_length": 657.1835327148438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.07463556851312, + "grad_norm": 0.13476435840129852, + "learning_rate": 1e-06, + "loss": -0.0191, + "num_tokens": 265103849.0, + "reward": 0.6116071939468384, + "reward_std": 0.17479917407035828, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.4876568913459778, + "step": 436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3311.0, + "completions/mean_length": 750.9063110351562, + "completions/mean_terminated_length": 594.5934448242188, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 4.0839650145772595, + "grad_norm": 0.13191767036914825, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 265713941.0, + "reward": 0.6863839626312256, + "reward_std": 0.15413424372673035, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 837.0089721679688, + "completions/mean_terminated_length": 615.5995483398438, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 4.093294460641399, + "grad_norm": 0.1257736086845398, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 266316133.0, + "reward": 0.6595982313156128, + "reward_std": 0.16078399121761322, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3974.0, + "completions/mean_length": 859.2053833007812, + "completions/mean_terminated_length": 680.0188598632812, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 4.1026239067055394, + "grad_norm": 0.14162789285182953, + "learning_rate": 1e-06, + "loss": -0.0199, + "num_tokens": 266991877.0, + "reward": 0.559151828289032, + "reward_std": 0.21601955592632294, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3119.0, + "completions/mean_length": 830.7969360351562, + "completions/mean_terminated_length": 658.1362915039062, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.111953352769679, + "grad_norm": 0.14433181285858154, + "learning_rate": 1e-06, + "loss": -0.0023, + "num_tokens": 267643799.0, + "reward": 0.6462053656578064, + "reward_std": 0.17239172756671906, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2661.0, + "completions/mean_length": 781.982177734375, + "completions/mean_terminated_length": 594.396240234375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 4.121282798833819, + "grad_norm": 0.1444917917251587, + "learning_rate": 1e-06, + "loss": -0.0268, + "num_tokens": 268230599.0, + "reward": 0.6808035969734192, + "reward_std": 0.20426028966903687, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3348.0, + "completions/mean_length": 834.7745971679688, + "completions/mean_terminated_length": 617.3595581054688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.130612244897959, + "grad_norm": 0.11794041097164154, + "learning_rate": 1e-06, + "loss": -0.0193, + "num_tokens": 268837589.0, + "reward": 0.6428571939468384, + "reward_std": 0.15631386637687683, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2801.0, + "completions/mean_length": 918.2332763671875, + "completions/mean_terminated_length": 698.2924194335938, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 4.139941690962099, + "grad_norm": 0.11573206633329391, + "learning_rate": 1e-06, + "loss": -0.0252, + "num_tokens": 269517694.0, + "reward": 0.578125, + "reward_std": 0.16386516392230988, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3152.0, + "completions/mean_length": 788.6596069335938, + "completions/mean_terminated_length": 580.7247924804688, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 4.149271137026239, + "grad_norm": 0.14195089042186737, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 270092837.0, + "reward": 0.645089328289032, + "reward_std": 0.19632920622825623, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4025.0, + "completions/mean_length": 797.8147583007812, + "completions/mean_terminated_length": 619.32470703125, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 4.158600583090379, + "grad_norm": 0.12240738421678543, + "learning_rate": 1e-06, + "loss": -0.0142, + "num_tokens": 270708999.0, + "reward": 0.5970982313156128, + "reward_std": 0.15372125804424286, + "rewards/verify_math_reward/mean": 0.5970982313156128, + "rewards/verify_math_reward/std": 0.4907552897930145, + "step": 445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3886.0, + "completions/mean_length": 907.42529296875, + "completions/mean_terminated_length": 682.6630859375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 4.167930029154519, + "grad_norm": 0.114719919860363, + "learning_rate": 1e-06, + "loss": -0.0142, + "num_tokens": 271370764.0, + "reward": 0.6004464626312256, + "reward_std": 0.146052747964859, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 835.1529541015625, + "completions/mean_terminated_length": 678.7847900390625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 4.1772594752186585, + "grad_norm": 0.13486167788505554, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 272042389.0, + "reward": 0.5959821939468384, + "reward_std": 0.19599655270576477, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3218.0, + "completions/mean_length": 750.0859985351562, + "completions/mean_terminated_length": 597.8214721679688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 4.186588921282799, + "grad_norm": 0.15013878047466278, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 272661986.0, + "reward": 0.5758928656578064, + "reward_std": 0.21155081689357758, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3594.0, + "completions/mean_length": 727.7891235351562, + "completions/mean_terminated_length": 574.5099487304688, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 4.1959183673469385, + "grad_norm": 0.15057620406150818, + "learning_rate": 1e-06, + "loss": -0.0063, + "num_tokens": 273241637.0, + "reward": 0.6350446939468384, + "reward_std": 0.19325028359889984, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3797.0, + "completions/mean_length": 758.8928833007812, + "completions/mean_terminated_length": 590.668212890625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 4.205247813411079, + "grad_norm": 0.13374288380146027, + "learning_rate": 1e-06, + "loss": -0.0005, + "num_tokens": 273834909.0, + "reward": 0.6462053656578064, + "reward_std": 0.15225812792778015, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3210.0, + "completions/mean_length": 862.8326416015625, + "completions/mean_terminated_length": 639.0573120117188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 4.214577259475218, + "grad_norm": 0.13761775195598602, + "learning_rate": 1e-06, + "loss": -0.0124, + "num_tokens": 274453847.0, + "reward": 0.5948660969734192, + "reward_std": 0.20511652529239655, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3378.0, + "completions/mean_length": 808.6172485351562, + "completions/mean_terminated_length": 597.7874145507812, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.223906705539359, + "grad_norm": 0.12827613949775696, + "learning_rate": 1e-06, + "loss": -0.0105, + "num_tokens": 275038920.0, + "reward": 0.590401828289032, + "reward_std": 0.1660016030073166, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3856.0, + "completions/mean_length": 804.8739013671875, + "completions/mean_terminated_length": 585.4655151367188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.233236151603498, + "grad_norm": 0.1444123089313507, + "learning_rate": 1e-06, + "loss": -0.0141, + "num_tokens": 275621903.0, + "reward": 0.6473214626312256, + "reward_std": 0.16529481112957, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 795.1842041015625, + "completions/mean_terminated_length": 595.9633178710938, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 4.242565597667639, + "grad_norm": 0.12860849499702454, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 276200636.0, + "reward": 0.6953125596046448, + "reward_std": 0.17250937223434448, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3479.0, + "completions/mean_length": 761.5904541015625, + "completions/mean_terminated_length": 601.6947021484375, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 4.251895043731778, + "grad_norm": 0.14034169912338257, + "learning_rate": 1e-06, + "loss": -0.012, + "num_tokens": 276803565.0, + "reward": 0.6495535969734192, + "reward_std": 0.17337700724601746, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2779.0, + "completions/mean_length": 755.388427734375, + "completions/mean_terminated_length": 615.548828125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 4.261224489795918, + "grad_norm": 0.15026754140853882, + "learning_rate": 1e-06, + "loss": -0.0136, + "num_tokens": 277412905.0, + "reward": 0.6930803656578064, + "reward_std": 0.19764302670955658, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2898.0, + "completions/mean_length": 794.3080444335938, + "completions/mean_terminated_length": 599.172607421875, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 4.270553935860058, + "grad_norm": 0.13028138875961304, + "learning_rate": 1e-06, + "loss": -0.0331, + "num_tokens": 278005461.0, + "reward": 0.6015625, + "reward_std": 0.18494194746017456, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2918.0, + "completions/mean_length": 747.7332763671875, + "completions/mean_terminated_length": 545.6484985351562, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 4.279883381924198, + "grad_norm": 0.13466346263885498, + "learning_rate": 1e-06, + "loss": -0.0082, + "num_tokens": 278543598.0, + "reward": 0.6584821939468384, + "reward_std": 0.11626280844211578, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3224.0, + "completions/mean_length": 805.2266235351562, + "completions/mean_terminated_length": 614.8511962890625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 4.289212827988338, + "grad_norm": 0.13225360214710236, + "learning_rate": 1e-06, + "loss": -0.0237, + "num_tokens": 279147633.0, + "reward": 0.6339285969734192, + "reward_std": 0.14989416301250458, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3653.0, + "completions/mean_length": 818.0413208007812, + "completions/mean_terminated_length": 644.7062377929688, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.298542274052478, + "grad_norm": 0.14628779888153076, + "learning_rate": 1e-06, + "loss": -0.0135, + "num_tokens": 279789270.0, + "reward": 0.5770089626312256, + "reward_std": 0.222105473279953, + "rewards/verify_math_reward/mean": 0.5770089030265808, + "rewards/verify_math_reward/std": 0.4943099319934845, + "step": 460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2195.0, + "completions/mean_length": 782.8392944335938, + "completions/mean_terminated_length": 623.9625854492188, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 4.307871720116618, + "grad_norm": 0.13780836760997772, + "learning_rate": 1e-06, + "loss": -0.011, + "num_tokens": 280418446.0, + "reward": 0.6104910969734192, + "reward_std": 0.17923328280448914, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3502.0, + "completions/mean_length": 794.208740234375, + "completions/mean_terminated_length": 594.9290161132812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 4.317201166180758, + "grad_norm": 0.12764322757720947, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 281012161.0, + "reward": 0.6227678656578064, + "reward_std": 0.1743106245994568, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 883.8917846679688, + "completions/mean_terminated_length": 640.9591674804688, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.326530612244898, + "grad_norm": 0.12587527930736542, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 281636736.0, + "reward": 0.6227678656578064, + "reward_std": 0.143612802028656, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2662.0, + "completions/mean_length": 760.5391235351562, + "completions/mean_terminated_length": 608.7503051757812, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 4.335860058309038, + "grad_norm": 0.12094590067863464, + "learning_rate": 1e-06, + "loss": -0.0201, + "num_tokens": 282241499.0, + "reward": 0.7064732313156128, + "reward_std": 0.1525944322347641, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3679.0, + "completions/mean_length": 830.0033569335938, + "completions/mean_terminated_length": 595.6016235351562, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 4.345189504373177, + "grad_norm": 0.133238285779953, + "learning_rate": 1e-06, + "loss": 0.001, + "num_tokens": 282831262.0, + "reward": 0.5870535969734192, + "reward_std": 0.15781019628047943, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263834953308105, + "step": 465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 813.7801513671875, + "completions/mean_terminated_length": 586.6098022460938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 4.354518950437318, + "grad_norm": 0.14964893460273743, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 283408169.0, + "reward": 0.6272321939468384, + "reward_std": 0.18622256815433502, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2530.0, + "completions/mean_length": 789.7734985351562, + "completions/mean_terminated_length": 602.6285400390625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 4.363848396501457, + "grad_norm": 0.12583027780056, + "learning_rate": 1e-06, + "loss": -0.0049, + "num_tokens": 284004334.0, + "reward": 0.6640625, + "reward_std": 0.16191306710243225, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3776.0, + "completions/mean_length": 883.3750610351562, + "completions/mean_terminated_length": 673.274658203125, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 4.373177842565598, + "grad_norm": 0.1296965479850769, + "learning_rate": 1e-06, + "loss": -0.0072, + "num_tokens": 284653014.0, + "reward": 0.606026828289032, + "reward_std": 0.17614509165287018, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3794.0, + "completions/mean_length": 722.724365234375, + "completions/mean_terminated_length": 597.7882080078125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 4.382507288629737, + "grad_norm": 0.13411414623260498, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 285254535.0, + "reward": 0.6796875596046448, + "reward_std": 0.149361714720726, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4065.0, + "completions/mean_length": 846.5670166015625, + "completions/mean_terminated_length": 650.4473266601562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 4.391836734693878, + "grad_norm": 0.1258615404367447, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 285897859.0, + "reward": 0.6026785969734192, + "reward_std": 0.14293991029262543, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3597.0, + "completions/mean_length": 835.8761596679688, + "completions/mean_terminated_length": 630.9098510742188, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 4.401166180758017, + "grad_norm": 0.1337389349937439, + "learning_rate": 1e-06, + "loss": -0.0016, + "num_tokens": 286518228.0, + "reward": 0.6339285969734192, + "reward_std": 0.16627302765846252, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3720.0, + "completions/mean_length": 726.411865234375, + "completions/mean_terminated_length": 585.3593139648438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.410495626822158, + "grad_norm": 0.15227915346622467, + "learning_rate": 1e-06, + "loss": -0.0092, + "num_tokens": 287106541.0, + "reward": 0.6674107313156128, + "reward_std": 0.2021552175283432, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2800.0, + "completions/mean_length": 782.6652221679688, + "completions/mean_terminated_length": 599.241455078125, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 4.419825072886297, + "grad_norm": 0.12876202166080475, + "learning_rate": 1e-06, + "loss": -0.0333, + "num_tokens": 287706833.0, + "reward": 0.6774553656578064, + "reward_std": 0.1628890484571457, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3949.0, + "completions/mean_length": 801.9386596679688, + "completions/mean_terminated_length": 635.8839111328125, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.429154518950437, + "grad_norm": 0.13935180008411407, + "learning_rate": 1e-06, + "loss": 0.0093, + "num_tokens": 288340874.0, + "reward": 0.6004464626312256, + "reward_std": 0.18036304414272308, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2607.0, + "completions/mean_length": 813.2455444335938, + "completions/mean_terminated_length": 619.2293090820312, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 4.438483965014577, + "grad_norm": 0.12817791104316711, + "learning_rate": 1e-06, + "loss": -0.0189, + "num_tokens": 288960022.0, + "reward": 0.637276828289032, + "reward_std": 0.16503477096557617, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3492.0, + "completions/mean_length": 847.286865234375, + "completions/mean_terminated_length": 597.3858642578125, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 4.447813411078717, + "grad_norm": 0.13169068098068237, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 289545887.0, + "reward": 0.6015625, + "reward_std": 0.1699160635471344, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 817.9252319335938, + "completions/mean_terminated_length": 591.0418090820312, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 4.457142857142857, + "grad_norm": 0.14308656752109528, + "learning_rate": 1e-06, + "loss": -0.0197, + "num_tokens": 290125988.0, + "reward": 0.640625, + "reward_std": 0.17359565198421478, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3143.0, + "completions/mean_length": 797.6864013671875, + "completions/mean_terminated_length": 598.6165771484375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 4.466472303206997, + "grad_norm": 0.13955950736999512, + "learning_rate": 1e-06, + "loss": 0.0128, + "num_tokens": 290726051.0, + "reward": 0.6339285969734192, + "reward_std": 0.16589100658893585, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3376.0, + "completions/mean_length": 837.4620971679688, + "completions/mean_terminated_length": 616.0834350585938, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 4.475801749271137, + "grad_norm": 0.1139809638261795, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 291331689.0, + "reward": 0.6004464626312256, + "reward_std": 0.13034509122371674, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2763.0, + "completions/mean_length": 832.505615234375, + "completions/mean_terminated_length": 589.8956909179688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.485131195335277, + "grad_norm": 0.14736969769001007, + "learning_rate": 1e-06, + "loss": -0.027, + "num_tokens": 291907006.0, + "reward": 0.6808035969734192, + "reward_std": 0.1905835121870041, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3747.0, + "completions/mean_length": 782.0558471679688, + "completions/mean_terminated_length": 623.1415405273438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.494460641399417, + "grad_norm": 0.11887058615684509, + "learning_rate": 1e-06, + "loss": -0.008, + "num_tokens": 292526000.0, + "reward": 0.645089328289032, + "reward_std": 0.13534656167030334, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4049.0, + "completions/mean_length": 1020.98779296875, + "completions/mean_terminated_length": 681.8599853515625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.503790087463557, + "grad_norm": 0.13074375689029694, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 293175237.0, + "reward": 0.5848214626312256, + "reward_std": 0.19557330012321472, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4087.0, + "completions/mean_length": 815.9989013671875, + "completions/mean_terminated_length": 609.782958984375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 4.513119533527696, + "grad_norm": 0.1316675990819931, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 293774124.0, + "reward": 0.652901828289032, + "reward_std": 0.16153216361999512, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2819.0, + "completions/mean_length": 824.9386596679688, + "completions/mean_terminated_length": 606.8678588867188, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 4.522448979591837, + "grad_norm": 0.13047119975090027, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 294377669.0, + "reward": 0.6037946939468384, + "reward_std": 0.15826597809791565, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3534.0, + "completions/mean_length": 906.6183471679688, + "completions/mean_terminated_length": 665.404541015625, + "completions/min_length": 194.0, + "completions/min_terminated_length": 194.0, + "epoch": 4.531778425655976, + "grad_norm": 0.13221052289009094, + "learning_rate": 1e-06, + "loss": -0.0114, + "num_tokens": 295042951.0, + "reward": 0.5848214626312256, + "reward_std": 0.1698397994041443, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3916.0, + "completions/mean_length": 765.7891235351562, + "completions/mean_terminated_length": 593.8063354492188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.541107871720117, + "grad_norm": 0.12726761400699615, + "learning_rate": 1e-06, + "loss": -0.0067, + "num_tokens": 295643962.0, + "reward": 0.6149553656578064, + "reward_std": 0.15443845093250275, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3385.0, + "completions/mean_length": 843.3672485351562, + "completions/mean_terminated_length": 609.9246215820312, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 4.550437317784256, + "grad_norm": 0.14112919569015503, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 296235995.0, + "reward": 0.6428571939468384, + "reward_std": 0.19561424851417542, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.47942501306533813, + "step": 487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3770.0, + "completions/mean_length": 1007.0357666015625, + "completions/mean_terminated_length": 708.3475952148438, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 4.559766763848397, + "grad_norm": 0.11537722498178482, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 296899523.0, + "reward": 0.5636160969734192, + "reward_std": 0.1654042750597, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4050.0, + "completions/mean_length": 832.739990234375, + "completions/mean_terminated_length": 639.8759155273438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.569096209912536, + "grad_norm": 0.14216546714305878, + "learning_rate": 1e-06, + "loss": -0.0264, + "num_tokens": 297527938.0, + "reward": 0.6026785969734192, + "reward_std": 0.199411079287529, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2496.0, + "completions/mean_length": 832.958740234375, + "completions/mean_terminated_length": 636.0177612304688, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 4.578425655976677, + "grad_norm": 0.13702218234539032, + "learning_rate": 1e-06, + "loss": -0.0179, + "num_tokens": 298162645.0, + "reward": 0.6540178656578064, + "reward_std": 0.170254185795784, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 810.966552734375, + "completions/mean_terminated_length": 604.4342041015625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.587755102040816, + "grad_norm": 0.11966714262962341, + "learning_rate": 1e-06, + "loss": 0.0133, + "num_tokens": 298759119.0, + "reward": 0.6205357313156128, + "reward_std": 0.1426691859960556, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2667.0, + "completions/mean_length": 831.6842041015625, + "completions/mean_terminated_length": 642.8394165039062, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 4.597084548104956, + "grad_norm": 0.14490832388401031, + "learning_rate": 1e-06, + "loss": -0.0241, + "num_tokens": 299396196.0, + "reward": 0.5803571939468384, + "reward_std": 0.1980990767478943, + "rewards/verify_math_reward/mean": 0.5803571343421936, + "rewards/verify_math_reward/std": 0.4937761425971985, + "step": 492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3047.0, + "completions/mean_length": 859.5636596679688, + "completions/mean_terminated_length": 581.033935546875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 4.606413994169096, + "grad_norm": 0.13457715511322021, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 299962573.0, + "reward": 0.6729910969734192, + "reward_std": 0.16739028692245483, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2886.0, + "completions/mean_length": 804.2969360351562, + "completions/mean_terminated_length": 617.9740600585938, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 4.615743440233236, + "grad_norm": 0.13072776794433594, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 300570999.0, + "reward": 0.6104910969734192, + "reward_std": 0.1743113249540329, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4014.0, + "completions/mean_length": 754.1395263671875, + "completions/mean_terminated_length": 552.44140625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 4.625072886297376, + "grad_norm": 0.14342716336250305, + "learning_rate": 1e-06, + "loss": -0.0266, + "num_tokens": 301118732.0, + "reward": 0.6941964626312256, + "reward_std": 0.1770215779542923, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.46100425720214844, + "step": 495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0401785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3442.0, + "completions/mean_length": 761.3973388671875, + "completions/mean_terminated_length": 621.809326171875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 4.634402332361516, + "grad_norm": 0.13373427093029022, + "learning_rate": 1e-06, + "loss": -0.0312, + "num_tokens": 301738152.0, + "reward": 0.715401828289032, + "reward_std": 0.16304102540016174, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3391.0, + "completions/mean_length": 761.9520263671875, + "completions/mean_terminated_length": 593.881591796875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 4.643731778425656, + "grad_norm": 0.1349562555551529, + "learning_rate": 1e-06, + "loss": -0.0154, + "num_tokens": 302325245.0, + "reward": 0.6897321939468384, + "reward_std": 0.17389734089374542, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0502232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2602.0, + "completions/mean_length": 738.4051513671875, + "completions/mean_terminated_length": 560.8590087890625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 4.653061224489796, + "grad_norm": 0.13668808341026306, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 302897792.0, + "reward": 0.7042410969734192, + "reward_std": 0.1671295464038849, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 841.0078735351562, + "completions/mean_terminated_length": 644.5526733398438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.662390670553936, + "grad_norm": 0.12017830461263657, + "learning_rate": 1e-06, + "loss": -0.022, + "num_tokens": 303532007.0, + "reward": 0.640625, + "reward_std": 0.15759406983852386, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 935.2701416015625, + "completions/mean_terminated_length": 692.1370239257812, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 4.671720116618076, + "grad_norm": 0.153817281126976, + "learning_rate": 1e-06, + "loss": 0.0058, + "num_tokens": 304196825.0, + "reward": 0.5390625, + "reward_std": 0.18521295487880707, + "rewards/verify_math_reward/mean": 0.5390625, + "rewards/verify_math_reward/std": 0.4987502098083496, + "step": 500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 868.7109985351562, + "completions/mean_terminated_length": 637.0872802734375, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 4.681049562682215, + "grad_norm": 0.13420456647872925, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 304824046.0, + "reward": 0.5848214626312256, + "reward_std": 0.18667538464069366, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 806.4777221679688, + "completions/mean_terminated_length": 644.6978759765625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 4.690379008746356, + "grad_norm": 0.13107848167419434, + "learning_rate": 1e-06, + "loss": -0.0302, + "num_tokens": 305459466.0, + "reward": 0.6205357313156128, + "reward_std": 0.1738552749156952, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3316.0, + "completions/mean_length": 835.8170166015625, + "completions/mean_terminated_length": 647.2113037109375, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 4.699708454810495, + "grad_norm": 0.1311369091272354, + "learning_rate": 1e-06, + "loss": -0.021, + "num_tokens": 306091942.0, + "reward": 0.59375, + "reward_std": 0.18419085443019867, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3426.0, + "completions/mean_length": 766.0324096679688, + "completions/mean_terminated_length": 598.1676025390625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 4.709037900874636, + "grad_norm": 0.14386148750782013, + "learning_rate": 1e-06, + "loss": -0.0225, + "num_tokens": 306689627.0, + "reward": 0.6495535969734192, + "reward_std": 0.20883847773075104, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0424107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3040.0, + "completions/mean_length": 762.669677734375, + "completions/mean_terminated_length": 615.0396118164062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.718367346938775, + "grad_norm": 0.12091245502233505, + "learning_rate": 1e-06, + "loss": -0.0139, + "num_tokens": 307307627.0, + "reward": 0.6383928656578064, + "reward_std": 0.13850508630275726, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3922.0, + "completions/mean_length": 923.802490234375, + "completions/mean_terminated_length": 654.97216796875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 4.727696793002916, + "grad_norm": 0.1387169510126114, + "learning_rate": 1e-06, + "loss": -0.0079, + "num_tokens": 307951002.0, + "reward": 0.5569196939468384, + "reward_std": 0.17942126095294952, + "rewards/verify_math_reward/mean": 0.5569196343421936, + "rewards/verify_math_reward/std": 0.4970270097255707, + "step": 506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3613.0, + "completions/mean_length": 874.2254638671875, + "completions/mean_terminated_length": 651.2387084960938, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 4.737026239067055, + "grad_norm": 0.15239207446575165, + "learning_rate": 1e-06, + "loss": -0.006, + "num_tokens": 308587700.0, + "reward": 0.6104910969734192, + "reward_std": 0.21019718050956726, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3988.0, + "completions/mean_length": 857.1563110351562, + "completions/mean_terminated_length": 612.20166015625, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 4.746355685131196, + "grad_norm": 0.13962629437446594, + "learning_rate": 1e-06, + "loss": -0.0166, + "num_tokens": 309198496.0, + "reward": 0.606026828289032, + "reward_std": 0.16314977407455444, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 766.5089721679688, + "completions/mean_terminated_length": 594.5634155273438, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 4.755685131195335, + "grad_norm": 0.14048907160758972, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 309790984.0, + "reward": 0.668526828289032, + "reward_std": 0.17577417194843292, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3132.0, + "completions/mean_length": 861.4408569335938, + "completions/mean_terminated_length": 608.4368286132812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.765014577259475, + "grad_norm": 0.12899498641490936, + "learning_rate": 1e-06, + "loss": -0.0317, + "num_tokens": 310385259.0, + "reward": 0.6261160969734192, + "reward_std": 0.15957936644554138, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2696.0, + "completions/mean_length": 814.7131958007812, + "completions/mean_terminated_length": 604.2743530273438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 4.774344023323615, + "grad_norm": 0.14420810341835022, + "learning_rate": 1e-06, + "loss": -0.0288, + "num_tokens": 310982346.0, + "reward": 0.5814732313156128, + "reward_std": 0.17284567654132843, + "rewards/verify_math_reward/mean": 0.5814732313156128, + "rewards/verify_math_reward/std": 0.4935929775238037, + "step": 511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3594.0, + "completions/mean_length": 872.1473388671875, + "completions/mean_terminated_length": 611.5947265625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 4.783673469387755, + "grad_norm": 0.13967280089855194, + "learning_rate": 1e-06, + "loss": -0.0135, + "num_tokens": 311581630.0, + "reward": 0.5959821939468384, + "reward_std": 0.1749500334262848, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3479.0, + "completions/mean_length": 788.052490234375, + "completions/mean_terminated_length": 592.5473022460938, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 4.793002915451895, + "grad_norm": 0.14434507489204407, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 312170757.0, + "reward": 0.590401828289032, + "reward_std": 0.18524505198001862, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3974.0, + "completions/mean_length": 883.7199096679688, + "completions/mean_terminated_length": 653.1734008789062, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.802332361516035, + "grad_norm": 0.13082991540431976, + "learning_rate": 1e-06, + "loss": -0.0166, + "num_tokens": 312800770.0, + "reward": 0.6428571939468384, + "reward_std": 0.1623242348432541, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3669.0, + "completions/mean_length": 826.0547485351562, + "completions/mean_terminated_length": 599.73388671875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 4.811661807580175, + "grad_norm": 0.13562647998332977, + "learning_rate": 1e-06, + "loss": -0.0203, + "num_tokens": 313383995.0, + "reward": 0.6116071939468384, + "reward_std": 0.16863657534122467, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3187.0, + "completions/mean_length": 826.4620971679688, + "completions/mean_terminated_length": 612.6397094726562, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 4.820991253644315, + "grad_norm": 0.1445397287607193, + "learning_rate": 1e-06, + "loss": -0.0095, + "num_tokens": 313982913.0, + "reward": 0.6183035969734192, + "reward_std": 0.17659832537174225, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3135.0, + "completions/mean_length": 899.85498046875, + "completions/mean_terminated_length": 649.8555908203125, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 4.830320699708455, + "grad_norm": 0.1289319097995758, + "learning_rate": 1e-06, + "loss": -0.0262, + "num_tokens": 314612807.0, + "reward": 0.6183035969734192, + "reward_std": 0.1680738478899002, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2997.0, + "completions/mean_length": 816.982177734375, + "completions/mean_terminated_length": 581.6459350585938, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 4.839650145772595, + "grad_norm": 0.14406071603298187, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 315190087.0, + "reward": 0.6674107313156128, + "reward_std": 0.16811592876911163, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2498.0, + "completions/mean_length": 822.6897583007812, + "completions/mean_terminated_length": 579.3501586914062, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 4.848979591836734, + "grad_norm": 0.14825446903705597, + "learning_rate": 1e-06, + "loss": -0.0115, + "num_tokens": 315772529.0, + "reward": 0.625, + "reward_std": 0.17430922389030457, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2776.0, + "completions/mean_length": 826.4910888671875, + "completions/mean_terminated_length": 562.2484741210938, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 4.858309037900875, + "grad_norm": 0.15536543726921082, + "learning_rate": 1e-06, + "loss": -0.0224, + "num_tokens": 316321769.0, + "reward": 0.609375, + "reward_std": 0.16991646587848663, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3257.0, + "completions/mean_length": 825.3236694335938, + "completions/mean_terminated_length": 586.3880615234375, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 4.867638483965014, + "grad_norm": 0.1325286477804184, + "learning_rate": 1e-06, + "loss": -0.037, + "num_tokens": 316894963.0, + "reward": 0.6171875, + "reward_std": 0.15154093503952026, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2749.0, + "completions/mean_length": 779.1752319335938, + "completions/mean_terminated_length": 570.6441650390625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 4.876967930029155, + "grad_norm": 0.13267797231674194, + "learning_rate": 1e-06, + "loss": -0.0318, + "num_tokens": 317462824.0, + "reward": 0.6383928656578064, + "reward_std": 0.15131382644176483, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2893.0, + "completions/mean_length": 755.8035888671875, + "completions/mean_terminated_length": 562.5690307617188, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 4.886297376093294, + "grad_norm": 0.1674153357744217, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 318026456.0, + "reward": 0.5926339626312256, + "reward_std": 0.2234174758195877, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161848425865173, + "step": 523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2776.0, + "completions/mean_length": 845.4576416015625, + "completions/mean_terminated_length": 649.27099609375, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 4.895626822157435, + "grad_norm": 0.1450287252664566, + "learning_rate": 1e-06, + "loss": -0.0221, + "num_tokens": 318660426.0, + "reward": 0.5524553656578064, + "reward_std": 0.2036636620759964, + "rewards/verify_math_reward/mean": 0.5524553656578064, + "rewards/verify_math_reward/std": 0.49751853942871094, + "step": 524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 716.7433471679688, + "completions/mean_terminated_length": 550.5503540039062, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 4.904956268221574, + "grad_norm": 0.13918952643871307, + "learning_rate": 1e-06, + "loss": -0.0158, + "num_tokens": 319229780.0, + "reward": 0.6975446939468384, + "reward_std": 0.1494358777999878, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2711.0, + "completions/mean_length": 761.9051513671875, + "completions/mean_terminated_length": 593.8323364257812, + "completions/min_length": 97.0, + "completions/min_terminated_length": 97.0, + "epoch": 4.914285714285715, + "grad_norm": 0.11945682764053345, + "learning_rate": 1e-06, + "loss": -0.0234, + "num_tokens": 319826663.0, + "reward": 0.660714328289032, + "reward_std": 0.1371113657951355, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4059.0, + "completions/mean_length": 929.9364013671875, + "completions/mean_terminated_length": 610.9962768554688, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 4.923615160349854, + "grad_norm": 0.13910621404647827, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 320417670.0, + "reward": 0.613839328289032, + "reward_std": 0.15139049291610718, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3243.0, + "completions/mean_length": 881.7131958007812, + "completions/mean_terminated_length": 626.1192626953125, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 4.932944606413994, + "grad_norm": 0.15693318843841553, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 321026821.0, + "reward": 0.5758928656578064, + "reward_std": 0.1835525780916214, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 813.091552734375, + "completions/mean_terminated_length": 590.0571899414062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.942274052478134, + "grad_norm": 0.15590135753154755, + "learning_rate": 1e-06, + "loss": -0.0176, + "num_tokens": 321616591.0, + "reward": 0.6071428656578064, + "reward_std": 0.1770561784505844, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0345982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3626.0, + "completions/mean_length": 712.0033569335938, + "completions/mean_terminated_length": 590.7271728515625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 4.9516034985422746, + "grad_norm": 0.12643130123615265, + "learning_rate": 1e-06, + "loss": -0.004, + "num_tokens": 322222642.0, + "reward": 0.6774553656578064, + "reward_std": 0.16145406663417816, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 869.5000610351562, + "completions/mean_terminated_length": 625.4789428710938, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 4.960932944606414, + "grad_norm": 0.15435025095939636, + "learning_rate": 1e-06, + "loss": -0.0196, + "num_tokens": 322830978.0, + "reward": 0.5993303656578064, + "reward_std": 0.208243265748024, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2920.0, + "completions/mean_length": 735.5067138671875, + "completions/mean_terminated_length": 549.4723510742188, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 4.970262390670554, + "grad_norm": 0.12676119804382324, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 323386632.0, + "reward": 0.6573660969734192, + "reward_std": 0.13899043202400208, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3907.0, + "completions/mean_length": 848.2913208007812, + "completions/mean_terminated_length": 602.666259765625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 4.979591836734694, + "grad_norm": 0.15604153275489807, + "learning_rate": 1e-06, + "loss": -0.0308, + "num_tokens": 323978437.0, + "reward": 0.6037946939468384, + "reward_std": 0.19644230604171753, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0357142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3702.0, + "completions/mean_length": 723.1295166015625, + "completions/mean_terminated_length": 598.2083129882812, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 4.988921282798834, + "grad_norm": 0.13526904582977295, + "learning_rate": 1e-06, + "loss": -0.0162, + "num_tokens": 324591993.0, + "reward": 0.6941964626312256, + "reward_std": 0.1867866814136505, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.05965909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2796.0, + "completions/mean_length": 800.039794921875, + "completions/mean_terminated_length": 590.9305419921875, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 4.998250728862974, + "grad_norm": 0.1538151651620865, + "learning_rate": 1e-06, + "loss": -0.0349, + "num_tokens": 325172889.0, + "reward": 0.6473214626312256, + "reward_std": 0.18963709473609924, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 1009.6607666015625, + "completions/mean_terminated_length": 656.4974975585938, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.0093294460641395, + "grad_norm": 0.1324121356010437, + "learning_rate": 1e-06, + "loss": -0.0309, + "num_tokens": 325791001.0, + "reward": 0.53125, + "reward_std": 0.1785971224308014, + "rewards/verify_math_reward/mean": 0.53125, + "rewards/verify_math_reward/std": 0.4993011951446533, + "step": 536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 851.0580444335938, + "completions/mean_terminated_length": 634.7285766601562, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 5.01865889212828, + "grad_norm": 0.13117796182632446, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 326406669.0, + "reward": 0.6261160969734192, + "reward_std": 0.1670207679271698, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3977.0, + "completions/mean_length": 798.5725708007812, + "completions/mean_terminated_length": 591.260986328125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 5.0279883381924195, + "grad_norm": 0.13558591902256012, + "learning_rate": 1e-06, + "loss": -0.0143, + "num_tokens": 326994766.0, + "reward": 0.5915178656578064, + "reward_std": 0.16100971400737762, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2884.0, + "completions/mean_length": 788.7310791015625, + "completions/mean_terminated_length": 584.9656372070312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.03731778425656, + "grad_norm": 0.13228186964988708, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 327579069.0, + "reward": 0.637276828289032, + "reward_std": 0.163679301738739, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2887.0, + "completions/mean_length": 768.8114013671875, + "completions/mean_terminated_length": 563.8187255859375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 5.0466472303206995, + "grad_norm": 0.1459277719259262, + "learning_rate": 1e-06, + "loss": -0.0301, + "num_tokens": 328135196.0, + "reward": 0.6595982313156128, + "reward_std": 0.15529540181159973, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3063.0, + "completions/mean_length": 856.8928833007812, + "completions/mean_terminated_length": 607.7307739257812, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.05597667638484, + "grad_norm": 0.14010262489318848, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 328732980.0, + "reward": 0.6071428656578064, + "reward_std": 0.16401489078998566, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2267.0, + "completions/mean_length": 843.3605346679688, + "completions/mean_terminated_length": 588.9422607421875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 5.0653061224489795, + "grad_norm": 0.15258248150348663, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 329315143.0, + "reward": 0.6160714626312256, + "reward_std": 0.18228264153003693, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.4866124987602234, + "step": 542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 816.5770263671875, + "completions/mean_terminated_length": 542.9613037109375, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.07463556851312, + "grad_norm": 0.1298285722732544, + "learning_rate": 1e-06, + "loss": -0.0447, + "num_tokens": 329854980.0, + "reward": 0.6506696939468384, + "reward_std": 0.14575035870075226, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3784.0, + "completions/mean_length": 807.8850708007812, + "completions/mean_terminated_length": 650.2093505859375, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 5.0839650145772595, + "grad_norm": 0.12721940875053406, + "learning_rate": 1e-06, + "loss": 0.0068, + "num_tokens": 330506645.0, + "reward": 0.5636160969734192, + "reward_std": 0.1452706754207611, + "rewards/verify_math_reward/mean": 0.5636160969734192, + "rewards/verify_math_reward/std": 0.49621346592903137, + "step": 544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3290.0, + "completions/mean_length": 772.6808471679688, + "completions/mean_terminated_length": 572.1017456054688, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 5.093294460641399, + "grad_norm": 0.1556389182806015, + "learning_rate": 1e-06, + "loss": -0.0282, + "num_tokens": 331076847.0, + "reward": 0.6696428656578064, + "reward_std": 0.183292955160141, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3547.0, + "completions/mean_length": 802.2991333007812, + "completions/mean_terminated_length": 574.3341674804688, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 5.1026239067055394, + "grad_norm": 0.12720397114753723, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 331645355.0, + "reward": 0.65625, + "reward_std": 0.1528993397951126, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 735.5279541015625, + "completions/mean_terminated_length": 541.1204223632812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 5.111953352769679, + "grad_norm": 0.1253661811351776, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 332199012.0, + "reward": 0.6808035969734192, + "reward_std": 0.12967249751091003, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3018.0, + "completions/mean_length": 900.9609985351562, + "completions/mean_terminated_length": 617.5614013671875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 5.121282798833819, + "grad_norm": 0.13752582669258118, + "learning_rate": 1e-06, + "loss": -0.0262, + "num_tokens": 332808097.0, + "reward": 0.6037946939468384, + "reward_std": 0.17476637661457062, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3720.0, + "completions/mean_length": 755.9699096679688, + "completions/mean_terminated_length": 545.9798583984375, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.130612244897959, + "grad_norm": 0.1480942666530609, + "learning_rate": 1e-06, + "loss": -0.0194, + "num_tokens": 333349902.0, + "reward": 0.6941964626312256, + "reward_std": 0.1615314781665802, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.46100425720214844, + "step": 549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3536.0, + "completions/mean_length": 832.1819458007812, + "completions/mean_terminated_length": 610.444580078125, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 5.139941690962099, + "grad_norm": 0.1550217717885971, + "learning_rate": 1e-06, + "loss": -0.0169, + "num_tokens": 333952401.0, + "reward": 0.5959821939468384, + "reward_std": 0.19629710912704468, + "rewards/verify_math_reward/mean": 0.5959821343421936, + "rewards/verify_math_reward/std": 0.490975022315979, + "step": 550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3957.0, + "completions/mean_length": 746.810302734375, + "completions/mean_terminated_length": 561.4016723632812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.149271137026239, + "grad_norm": 0.14474813640117645, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 334516335.0, + "reward": 0.6863839626312256, + "reward_std": 0.1497069001197815, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 1035.47216796875, + "completions/mean_terminated_length": 672.4880981445312, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 5.158600583090379, + "grad_norm": 0.13239631056785583, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 335169582.0, + "reward": 0.504464328289032, + "reward_std": 0.16927595436573029, + "rewards/verify_math_reward/mean": 0.5044642686843872, + "rewards/verify_math_reward/std": 0.5002593398094177, + "step": 552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3440.0, + "completions/mean_length": 819.6752319335938, + "completions/mean_terminated_length": 654.5146484375, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 5.167930029154519, + "grad_norm": 0.12959615886211395, + "learning_rate": 1e-06, + "loss": -0.0021, + "num_tokens": 335820411.0, + "reward": 0.637276828289032, + "reward_std": 0.16435441374778748, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3919.0, + "completions/mean_length": 824.716552734375, + "completions/mean_terminated_length": 602.4719848632812, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.1772594752186585, + "grad_norm": 0.15217341482639313, + "learning_rate": 1e-06, + "loss": -0.0235, + "num_tokens": 336410781.0, + "reward": 0.6484375, + "reward_std": 0.19336046278476715, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2336.0, + "completions/mean_length": 843.7467041015625, + "completions/mean_terminated_length": 626.9298095703125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 5.186588921282799, + "grad_norm": 0.15517523884773254, + "learning_rate": 1e-06, + "loss": -0.0351, + "num_tokens": 337020074.0, + "reward": 0.6205357313156128, + "reward_std": 0.18968309462070465, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3798.0, + "completions/mean_length": 900.78466796875, + "completions/mean_terminated_length": 659.129638671875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.1959183673469385, + "grad_norm": 0.1398380547761917, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 337662601.0, + "reward": 0.613839328289032, + "reward_std": 0.1760370135307312, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 862.8471069335938, + "completions/mean_terminated_length": 605.7530517578125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.205247813411079, + "grad_norm": 0.17168253660202026, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 338251480.0, + "reward": 0.6651785969734192, + "reward_std": 0.1845717579126358, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3851.0, + "completions/mean_length": 849.9542846679688, + "completions/mean_terminated_length": 625.28759765625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.214577259475218, + "grad_norm": 0.1361083835363388, + "learning_rate": 1e-06, + "loss": -0.0205, + "num_tokens": 338864031.0, + "reward": 0.590401828289032, + "reward_std": 0.15033742785453796, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2474.0, + "completions/mean_length": 913.5178833007812, + "completions/mean_terminated_length": 660.4530029296875, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 5.223906705539359, + "grad_norm": 0.13579238951206207, + "learning_rate": 1e-06, + "loss": -0.0275, + "num_tokens": 339506991.0, + "reward": 0.578125, + "reward_std": 0.1685623973608017, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4014.0, + "completions/mean_length": 821.2455444335938, + "completions/mean_terminated_length": 586.2152709960938, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 5.233236151603498, + "grad_norm": 0.14370384812355042, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 340081019.0, + "reward": 0.6774553656578064, + "reward_std": 0.15082457661628723, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 823.4910888671875, + "completions/mean_terminated_length": 596.9928588867188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.242565597667639, + "grad_norm": 0.14612165093421936, + "learning_rate": 1e-06, + "loss": -0.0033, + "num_tokens": 340672531.0, + "reward": 0.6328125, + "reward_std": 0.17442260682582855, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2471.0, + "completions/mean_length": 760.2210083007812, + "completions/mean_terminated_length": 579.6964721679688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 5.251895043731778, + "grad_norm": 0.14979024231433868, + "learning_rate": 1e-06, + "loss": -0.0123, + "num_tokens": 341256473.0, + "reward": 0.645089328289032, + "reward_std": 0.1756240278482437, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4056.0, + "completions/mean_length": 1067.1898193359375, + "completions/mean_terminated_length": 716.4059448242188, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 5.261224489795918, + "grad_norm": 0.1400134116411209, + "learning_rate": 1e-06, + "loss": -0.0378, + "num_tokens": 341926595.0, + "reward": 0.5145089626312256, + "reward_std": 0.17510268092155457, + "rewards/verify_math_reward/mean": 0.5145089030265808, + "rewards/verify_math_reward/std": 0.5000685453414917, + "step": 563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3840.0, + "completions/mean_length": 801.9688110351562, + "completions/mean_terminated_length": 615.51416015625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 5.270553935860058, + "grad_norm": 0.14505355060100555, + "learning_rate": 1e-06, + "loss": -0.0086, + "num_tokens": 342541015.0, + "reward": 0.590401828289032, + "reward_std": 0.17559124529361725, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 859.0647583007812, + "completions/mean_terminated_length": 626.748779296875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.279883381924198, + "grad_norm": 0.14341820776462555, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 343151633.0, + "reward": 0.5658482313156128, + "reward_std": 0.16060107946395874, + "rewards/verify_math_reward/mean": 0.5658482313156128, + "rewards/verify_math_reward/std": 0.49592188000679016, + "step": 565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3584.0, + "completions/mean_length": 1012.1328735351562, + "completions/mean_terminated_length": 701.4729614257812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 5.289212827988338, + "grad_norm": 0.13181324303150177, + "learning_rate": 1e-06, + "loss": -0.077, + "num_tokens": 343817280.0, + "reward": 0.5926339626312256, + "reward_std": 0.19418390095233917, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161848425865173, + "step": 566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3365.0, + "completions/mean_length": 874.0938110351562, + "completions/mean_terminated_length": 588.3110961914062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 5.298542274052478, + "grad_norm": 0.1548570990562439, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 344394980.0, + "reward": 0.660714328289032, + "reward_std": 0.18858583271503448, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 944.8404541015625, + "completions/mean_terminated_length": 656.9756469726562, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 5.307871720116618, + "grad_norm": 0.13820196688175201, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 345025277.0, + "reward": 0.5848214626312256, + "reward_std": 0.16499200463294983, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0457589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2487.0, + "completions/mean_length": 693.8248291015625, + "completions/mean_terminated_length": 530.6795043945312, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 5.317201166180758, + "grad_norm": 0.14032450318336487, + "learning_rate": 1e-06, + "loss": -0.0319, + "num_tokens": 345556376.0, + "reward": 0.7087053656578064, + "reward_std": 0.14229759573936462, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2866.0, + "completions/mean_length": 853.3303833007812, + "completions/mean_terminated_length": 624.7550659179688, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 5.326530612244898, + "grad_norm": 0.15670815110206604, + "learning_rate": 1e-06, + "loss": -0.0156, + "num_tokens": 346170056.0, + "reward": 0.6618303656578064, + "reward_std": 0.19674400985240936, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2937.0, + "completions/mean_length": 1003.7623291015625, + "completions/mean_terminated_length": 671.2224731445312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 5.335860058309038, + "grad_norm": 0.1319567710161209, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 346810179.0, + "reward": 0.582589328289032, + "reward_std": 0.170298770070076, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.493407279253006, + "step": 571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0513392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3537.0, + "completions/mean_length": 779.9152221679688, + "completions/mean_terminated_length": 600.4564208984375, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 5.345189504373177, + "grad_norm": 0.15103895962238312, + "learning_rate": 1e-06, + "loss": -0.0094, + "num_tokens": 347404335.0, + "reward": 0.6696428656578064, + "reward_std": 0.18340425193309784, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3306.0, + "completions/mean_length": 939.6808471679688, + "completions/mean_terminated_length": 651.3447265625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.354518950437318, + "grad_norm": 0.14508672058582306, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 348033809.0, + "reward": 0.6082589626312256, + "reward_std": 0.16563068330287933, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.4884119927883148, + "step": 573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2852.0, + "completions/mean_length": 876.7388916015625, + "completions/mean_terminated_length": 624.931396484375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.363848396501457, + "grad_norm": 0.1509408801794052, + "learning_rate": 1e-06, + "loss": -0.0054, + "num_tokens": 348644039.0, + "reward": 0.6540178656578064, + "reward_std": 0.16225893795490265, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.475953072309494, + "step": 574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2204.0, + "completions/mean_length": 848.2846069335938, + "completions/mean_terminated_length": 568.7842407226562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.373177842565598, + "grad_norm": 0.14984580874443054, + "learning_rate": 1e-06, + "loss": -0.028, + "num_tokens": 349214006.0, + "reward": 0.6417410969734192, + "reward_std": 0.17287889122962952, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975653409957886, + "step": 575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3561.0, + "completions/mean_length": 731.9207763671875, + "completions/mean_terminated_length": 566.4742431640625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 5.382507288629737, + "grad_norm": 0.15863996744155884, + "learning_rate": 1e-06, + "loss": -0.0202, + "num_tokens": 349783927.0, + "reward": 0.6908482313156128, + "reward_std": 0.1578873097896576, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3934.0, + "completions/mean_length": 912.8069458007812, + "completions/mean_terminated_length": 638.859375, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 5.391836734693878, + "grad_norm": 0.15568514168262482, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 350401082.0, + "reward": 0.5758928656578064, + "reward_std": 0.18512418866157532, + "rewards/verify_math_reward/mean": 0.5758928656578064, + "rewards/verify_math_reward/std": 0.49448272585868835, + "step": 577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3526.0, + "completions/mean_length": 922.1038208007812, + "completions/mean_terminated_length": 593.7697143554688, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 5.401166180758017, + "grad_norm": 0.15030905604362488, + "learning_rate": 1e-06, + "loss": -0.0262, + "num_tokens": 350981071.0, + "reward": 0.6104910969734192, + "reward_std": 0.1554141491651535, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3758.0, + "completions/mean_length": 817.5848388671875, + "completions/mean_terminated_length": 623.8250732421875, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 5.410495626822158, + "grad_norm": 0.16143877804279327, + "learning_rate": 1e-06, + "loss": -0.0172, + "num_tokens": 351593787.0, + "reward": 0.6573660969734192, + "reward_std": 0.1687125563621521, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3989.0, + "completions/mean_length": 869.1373291015625, + "completions/mean_terminated_length": 658.1058349609375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.419825072886297, + "grad_norm": 0.13263040781021118, + "learning_rate": 1e-06, + "loss": -0.0197, + "num_tokens": 352249254.0, + "reward": 0.590401828289032, + "reward_std": 0.1642010509967804, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3095.0, + "completions/mean_length": 902.8795166015625, + "completions/mean_terminated_length": 653.1167602539062, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 5.429154518950437, + "grad_norm": 0.15279339253902435, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 352890202.0, + "reward": 0.609375, + "reward_std": 0.15718218684196472, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3083.0, + "completions/mean_length": 726.6484985351562, + "completions/mean_terminated_length": 573.3173828125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 5.438483965014577, + "grad_norm": 0.1596021056175232, + "learning_rate": 1e-06, + "loss": -0.0161, + "num_tokens": 353465615.0, + "reward": 0.7087053656578064, + "reward_std": 0.15898343920707703, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4044.0, + "completions/mean_length": 851.7254638671875, + "completions/mean_terminated_length": 597.9614868164062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 5.447813411078717, + "grad_norm": 0.14287786185741425, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 354062873.0, + "reward": 0.5948660969734192, + "reward_std": 0.17002595961093903, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3734.0, + "completions/mean_length": 916.646240234375, + "completions/mean_terminated_length": 634.637939453125, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.457142857142857, + "grad_norm": 0.23945553600788116, + "learning_rate": 1e-06, + "loss": -0.0372, + "num_tokens": 354671044.0, + "reward": 0.6004464626312256, + "reward_std": 0.15123826265335083, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2969.0, + "completions/mean_length": 800.6484985351562, + "completions/mean_terminated_length": 572.5692138671875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 5.466472303206997, + "grad_norm": 0.15243570506572723, + "learning_rate": 1e-06, + "loss": -0.0382, + "num_tokens": 355236065.0, + "reward": 0.6852678656578064, + "reward_std": 0.1820102334022522, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3518.0, + "completions/mean_length": 921.74560546875, + "completions/mean_terminated_length": 606.2674560546875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 5.475801749271137, + "grad_norm": 0.16602863371372223, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 355826165.0, + "reward": 0.6283482313156128, + "reward_std": 0.17810675501823425, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3279.0, + "completions/mean_length": 842.0189819335938, + "completions/mean_terminated_length": 612.6463623046875, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 5.485131195335277, + "grad_norm": 0.15476739406585693, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 356428318.0, + "reward": 0.6595982313156128, + "reward_std": 0.16416895389556885, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2564.0, + "completions/mean_length": 909.90185546875, + "completions/mean_terminated_length": 610.3541259765625, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.494460641399417, + "grad_norm": 0.14812588691711426, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 357025782.0, + "reward": 0.6551339626312256, + "reward_std": 0.15698717534542084, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3983.0, + "completions/mean_length": 854.8125610351562, + "completions/mean_terminated_length": 630.4821166992188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.503790087463557, + "grad_norm": 0.13244247436523438, + "learning_rate": 1e-06, + "loss": -0.0062, + "num_tokens": 357632102.0, + "reward": 0.590401828289032, + "reward_std": 0.15570978820323944, + "rewards/verify_math_reward/mean": 0.5904017686843872, + "rewards/verify_math_reward/std": 0.49203425645828247, + "step": 589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3391.0, + "completions/mean_length": 918.7444458007812, + "completions/mean_terminated_length": 653.6529541015625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.513119533527696, + "grad_norm": 0.15337373316287994, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 358261089.0, + "reward": 0.6361607313156128, + "reward_std": 0.18904118239879608, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3689.0, + "completions/mean_length": 943.2266235351562, + "completions/mean_terminated_length": 667.7415161132812, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 5.522448979591837, + "grad_norm": 0.13773854076862335, + "learning_rate": 1e-06, + "loss": -0.042, + "num_tokens": 358907604.0, + "reward": 0.6261160969734192, + "reward_std": 0.16366903483867645, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4082.0, + "completions/mean_length": 931.71435546875, + "completions/mean_terminated_length": 634.2173461914062, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 5.531778425655976, + "grad_norm": 0.1630888730287552, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 359521884.0, + "reward": 0.629464328289032, + "reward_std": 0.17938737571239471, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3255.0, + "completions/mean_length": 859.107177734375, + "completions/mean_terminated_length": 626.7942504882812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.541107871720117, + "grad_norm": 0.13411062955856323, + "learning_rate": 1e-06, + "loss": -0.0251, + "num_tokens": 360130524.0, + "reward": 0.6696428656578064, + "reward_std": 0.1557832509279251, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3928.0, + "completions/mean_length": 898.76123046875, + "completions/mean_terminated_length": 623.6048583984375, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 5.550437317784256, + "grad_norm": 0.14365045726299286, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 360745078.0, + "reward": 0.6272321939468384, + "reward_std": 0.14613084495067596, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.041294642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3853.0, + "completions/mean_length": 726.7109985351562, + "completions/mean_terminated_length": 581.5844116210938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.559766763848397, + "grad_norm": 0.14855937659740448, + "learning_rate": 1e-06, + "loss": -0.011, + "num_tokens": 361327891.0, + "reward": 0.676339328289032, + "reward_std": 0.19223956763744354, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 804.1205444335938, + "completions/mean_terminated_length": 613.6812133789062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.569096209912536, + "grad_norm": 0.1522262692451477, + "learning_rate": 1e-06, + "loss": -0.0314, + "num_tokens": 361938511.0, + "reward": 0.6171875, + "reward_std": 0.21282006800174713, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0435267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3918.0, + "completions/mean_length": 755.6038208007812, + "completions/mean_terminated_length": 603.5904541015625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 5.578425655976677, + "grad_norm": 0.15594299137592316, + "learning_rate": 1e-06, + "loss": -0.019, + "num_tokens": 362551796.0, + "reward": 0.6506696939468384, + "reward_std": 0.18268606066703796, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2126.0, + "completions/mean_length": 858.474365234375, + "completions/mean_terminated_length": 567.0182495117188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 5.587755102040816, + "grad_norm": 0.1511084884405136, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 363111405.0, + "reward": 0.6629464626312256, + "reward_std": 0.15710552036762238, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2504.0, + "completions/mean_length": 836.2779541015625, + "completions/mean_terminated_length": 593.9484252929688, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 5.597084548104956, + "grad_norm": 0.1524042934179306, + "learning_rate": 1e-06, + "loss": -0.0349, + "num_tokens": 363700030.0, + "reward": 0.6361607313156128, + "reward_std": 0.16282165050506592, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.049107142857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3437.0, + "completions/mean_length": 812.6641235351562, + "completions/mean_terminated_length": 643.1021118164062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 5.606413994169096, + "grad_norm": 0.1423005759716034, + "learning_rate": 1e-06, + "loss": -0.009, + "num_tokens": 364331241.0, + "reward": 0.6037946939468384, + "reward_std": 0.1617264747619629, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 600 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3961.0, + "completions/mean_length": 899.755615234375, + "completions/mean_terminated_length": 645.5963745117188, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 5.615743440233236, + "grad_norm": 0.15083767473697662, + "learning_rate": 1e-06, + "loss": -0.0241, + "num_tokens": 364957470.0, + "reward": 0.6540178656578064, + "reward_std": 0.18986602127552032, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 601 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3045.0, + "completions/mean_length": 952.8170166015625, + "completions/mean_terminated_length": 627.6600952148438, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 5.625072886297376, + "grad_norm": 0.1534229964017868, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 365556466.0, + "reward": 0.609375, + "reward_std": 0.19110387563705444, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 602 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0546875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3887.0, + "completions/mean_length": 750.193115234375, + "completions/mean_terminated_length": 556.6339721679688, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 5.634402332361516, + "grad_norm": 0.14218640327453613, + "learning_rate": 1e-06, + "loss": -0.014, + "num_tokens": 366117255.0, + "reward": 0.6718750596046448, + "reward_std": 0.13809071481227875, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 603 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3414.0, + "completions/mean_length": 917.6707763671875, + "completions/mean_terminated_length": 652.48974609375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 5.643731778425656, + "grad_norm": 0.15133124589920044, + "learning_rate": 1e-06, + "loss": -0.0317, + "num_tokens": 366737096.0, + "reward": 0.5691964626312256, + "reward_std": 0.20339447259902954, + "rewards/verify_math_reward/mean": 0.5691964030265808, + "rewards/verify_math_reward/std": 0.4954652488231659, + "step": 604 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3449.0, + "completions/mean_length": 870.8861694335938, + "completions/mean_terminated_length": 606.021728515625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.653061224489796, + "grad_norm": 0.14131079614162445, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 367325426.0, + "reward": 0.6863839626312256, + "reward_std": 0.17404918372631073, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 605 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3051.0, + "completions/mean_length": 944.3013916015625, + "completions/mean_terminated_length": 618.2635498046875, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 5.662390670553936, + "grad_norm": 0.1873425543308258, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 367926440.0, + "reward": 0.5792410969734192, + "reward_std": 0.14446905255317688, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 606 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.046875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 751.1250610351562, + "completions/mean_terminated_length": 586.6229248046875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 5.671720116618076, + "grad_norm": 0.12935638427734375, + "learning_rate": 1e-06, + "loss": -0.0248, + "num_tokens": 368522600.0, + "reward": 0.7131696939468384, + "reward_std": 0.14823377132415771, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 607 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3137.0, + "completions/mean_length": 900.7902221679688, + "completions/mean_terminated_length": 655.0048217773438, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 5.681049562682215, + "grad_norm": 0.15427511930465698, + "learning_rate": 1e-06, + "loss": -0.0066, + "num_tokens": 369160092.0, + "reward": 0.6194196939468384, + "reward_std": 0.17626412212848663, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 608 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2963.0, + "completions/mean_length": 834.8772583007812, + "completions/mean_terminated_length": 654.3439331054688, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.690379008746356, + "grad_norm": 0.13125135004520416, + "learning_rate": 1e-06, + "loss": -0.0194, + "num_tokens": 369799854.0, + "reward": 0.6116071939468384, + "reward_std": 0.15357083082199097, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 609 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2962.0, + "completions/mean_length": 812.8705444335938, + "completions/mean_terminated_length": 589.8212280273438, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 5.699708454810495, + "grad_norm": 0.1430377960205078, + "learning_rate": 1e-06, + "loss": -0.0318, + "num_tokens": 370383658.0, + "reward": 0.6551339626312256, + "reward_std": 0.17107972502708435, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 610 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4091.0, + "completions/mean_length": 790.5770263671875, + "completions/mean_terminated_length": 578.5902709960938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.709037900874636, + "grad_norm": 0.15086200833320618, + "learning_rate": 1e-06, + "loss": -0.0159, + "num_tokens": 370957191.0, + "reward": 0.676339328289032, + "reward_std": 0.17235784232616425, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 611 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 952.2131958007812, + "completions/mean_terminated_length": 660.8378295898438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 5.718367346938775, + "grad_norm": 0.14042527973651886, + "learning_rate": 1e-06, + "loss": -0.056, + "num_tokens": 371589694.0, + "reward": 0.5881696939468384, + "reward_std": 0.1762627214193344, + "rewards/verify_math_reward/mean": 0.5881696343421936, + "rewards/verify_math_reward/std": 0.4924395978450775, + "step": 612 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.044642857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3154.0, + "completions/mean_length": 722.091552734375, + "completions/mean_terminated_length": 564.4322509765625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.727696793002916, + "grad_norm": 0.13101571798324585, + "learning_rate": 1e-06, + "loss": -0.0227, + "num_tokens": 372153040.0, + "reward": 0.7042410969734192, + "reward_std": 0.11648217588663101, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 613 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3687.0, + "completions/mean_length": 818.3683471679688, + "completions/mean_terminated_length": 612.3013305664062, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 5.737026239067055, + "grad_norm": 0.16602221131324768, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 372749194.0, + "reward": 0.6830357313156128, + "reward_std": 0.19554010033607483, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 614 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3717.0, + "completions/mean_length": 836.4364013671875, + "completions/mean_terminated_length": 606.6702270507812, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 5.746355685131196, + "grad_norm": 0.14646975696086884, + "learning_rate": 1e-06, + "loss": 0.0005, + "num_tokens": 373338641.0, + "reward": 0.6852678656578064, + "reward_std": 0.14271238446235657, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 615 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3454.0, + "completions/mean_length": 892.7701416015625, + "completions/mean_terminated_length": 638.055419921875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 5.755685131195335, + "grad_norm": 0.13492605090141296, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 373953427.0, + "reward": 0.6729910969734192, + "reward_std": 0.15549084544181824, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 616 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3218.0, + "completions/mean_length": 861.7142944335938, + "completions/mean_terminated_length": 617.1044311523438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 5.765014577259475, + "grad_norm": 0.14568999409675598, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 374548219.0, + "reward": 0.6629464626312256, + "reward_std": 0.17554594576358795, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 617 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2873.0, + "completions/mean_length": 909.32373046875, + "completions/mean_terminated_length": 609.7216186523438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.774344023323615, + "grad_norm": 0.13196724653244019, + "learning_rate": 1e-06, + "loss": -0.0298, + "num_tokens": 375132165.0, + "reward": 0.6729910969734192, + "reward_std": 0.1314416527748108, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 618 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3656.0, + "completions/mean_length": 882.2042846679688, + "completions/mean_terminated_length": 639.14404296875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 5.783673469387755, + "grad_norm": 0.14321132004261017, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 375765884.0, + "reward": 0.637276828289032, + "reward_std": 0.19058141112327576, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 619 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 851.6116333007812, + "completions/mean_terminated_length": 610.4220581054688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 5.793002915451895, + "grad_norm": 0.15028724074363708, + "learning_rate": 1e-06, + "loss": -0.0298, + "num_tokens": 376354840.0, + "reward": 0.6026785969734192, + "reward_std": 0.1806240826845169, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 620 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 820.5435791015625, + "completions/mean_terminated_length": 610.4786376953125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 5.802332361516035, + "grad_norm": 0.14434611797332764, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 376950695.0, + "reward": 0.6774553656578064, + "reward_std": 0.17510157823562622, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 621 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0558035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3406.0, + "completions/mean_length": 781.4732666015625, + "completions/mean_terminated_length": 585.5792236328125, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 5.811661807580175, + "grad_norm": 0.15130048990249634, + "learning_rate": 1e-06, + "loss": -0.0195, + "num_tokens": 377525167.0, + "reward": 0.6662946939468384, + "reward_std": 0.16468818485736847, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 622 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3172.0, + "completions/mean_length": 1031.622802734375, + "completions/mean_terminated_length": 668.1822509765625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 5.820991253644315, + "grad_norm": 0.14292049407958984, + "learning_rate": 1e-06, + "loss": -0.0263, + "num_tokens": 378146845.0, + "reward": 0.5446428656578064, + "reward_std": 0.1868608444929123, + "rewards/verify_math_reward/mean": 0.5446428656578064, + "rewards/verify_math_reward/std": 0.49828118085861206, + "step": 623 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3145.0, + "completions/mean_length": 904.4364013671875, + "completions/mean_terminated_length": 650.6494140625, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 5.830320699708455, + "grad_norm": 0.1270424723625183, + "learning_rate": 1e-06, + "loss": -0.0458, + "num_tokens": 378778404.0, + "reward": 0.578125, + "reward_std": 0.11712227016687393, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 624 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2395.0, + "completions/mean_length": 803.8582763671875, + "completions/mean_terminated_length": 592.7233276367188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 5.839650145772595, + "grad_norm": 0.12839479744434357, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 379362293.0, + "reward": 0.6629464626312256, + "reward_std": 0.1277952641248703, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 625 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2403.0, + "completions/mean_length": 805.5234985351562, + "completions/mean_terminated_length": 598.6488647460938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 5.848979591836734, + "grad_norm": 0.14423708617687225, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 379951530.0, + "reward": 0.613839328289032, + "reward_std": 0.1581925004720688, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 626 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3237.0, + "completions/mean_length": 831.6395263671875, + "completions/mean_terminated_length": 533.4336547851562, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 5.858309037900875, + "grad_norm": 0.16185839474201202, + "learning_rate": 1e-06, + "loss": -0.029, + "num_tokens": 380468759.0, + "reward": 0.7254464626312256, + "reward_std": 0.17664320766925812, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 627 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.056919642857142905, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2450.0, + "completions/mean_length": 836.7332763671875, + "completions/mean_terminated_length": 640.0201416015625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 5.867638483965014, + "grad_norm": 0.13921836018562317, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 381105248.0, + "reward": 0.6696428656578064, + "reward_std": 0.17633871734142303, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 628 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2003.0, + "completions/mean_length": 805.5703735351562, + "completions/mean_terminated_length": 535.341796875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 5.876967930029155, + "grad_norm": 0.15873199701309204, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 381632471.0, + "reward": 0.6953125596046448, + "reward_std": 0.17694559693336487, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 629 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 1038.2623291015625, + "completions/mean_terminated_length": 649.7949829101562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 5.886297376093294, + "grad_norm": 0.1511031836271286, + "learning_rate": 1e-06, + "loss": -0.0164, + "num_tokens": 382245162.0, + "reward": 0.5848214626312256, + "reward_std": 0.17446258664131165, + "rewards/verify_math_reward/mean": 0.5848214030265808, + "rewards/verify_math_reward/std": 0.49302801489830017, + "step": 630 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3142.0, + "completions/mean_length": 920.6551513671875, + "completions/mean_terminated_length": 600.7800903320312, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 5.895626822157435, + "grad_norm": 0.15285640954971313, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 382827405.0, + "reward": 0.6104910969734192, + "reward_std": 0.1698736995458603, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 631 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2569.0, + "completions/mean_length": 810.4631958007812, + "completions/mean_terminated_length": 574.6590576171875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 5.904956268221574, + "grad_norm": 0.153993621468544, + "learning_rate": 1e-06, + "loss": -0.0181, + "num_tokens": 383399332.0, + "reward": 0.6674107313156128, + "reward_std": 0.1521807461977005, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 632 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4017.0, + "completions/mean_length": 962.0480346679688, + "completions/mean_terminated_length": 671.5841674804688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 5.914285714285715, + "grad_norm": 0.14376863837242126, + "learning_rate": 1e-06, + "loss": -0.0351, + "num_tokens": 384035239.0, + "reward": 0.6272321939468384, + "reward_std": 0.1906474083662033, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 633 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0479910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2917.0, + "completions/mean_length": 757.3471069335938, + "completions/mean_terminated_length": 589.0445556640625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.923615160349854, + "grad_norm": 0.14942072331905365, + "learning_rate": 1e-06, + "loss": -0.0185, + "num_tokens": 384625606.0, + "reward": 0.6953125596046448, + "reward_std": 0.16086068749427795, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 634 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3056.0, + "completions/mean_length": 988.0089721679688, + "completions/mean_terminated_length": 615.0499877929688, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 5.932944606413994, + "grad_norm": 0.15287573635578156, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 385210910.0, + "reward": 0.621651828289032, + "reward_std": 0.1543617844581604, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 635 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3069.0, + "completions/mean_length": 915.6975708007812, + "completions/mean_terminated_length": 608.177490234375, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 5.942274052478134, + "grad_norm": 0.14832653105258942, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 385793799.0, + "reward": 0.6328125, + "reward_std": 0.19554080069065094, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 636 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3119.0, + "completions/mean_length": 837.419677734375, + "completions/mean_terminated_length": 620.1809692382812, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.9516034985422746, + "grad_norm": 0.12214856594800949, + "learning_rate": 1e-06, + "loss": -0.0346, + "num_tokens": 386405911.0, + "reward": 0.5915178656578064, + "reward_std": 0.1418454498052597, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 637 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3049.0, + "completions/mean_length": 771.5123291015625, + "completions/mean_terminated_length": 558.3028564453125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 5.960932944606414, + "grad_norm": 0.16159658133983612, + "learning_rate": 1e-06, + "loss": -0.0087, + "num_tokens": 386958514.0, + "reward": 0.7053571939468384, + "reward_std": 0.1906997561454773, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 638 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3366.0, + "completions/mean_length": 902.9051513671875, + "completions/mean_terminated_length": 623.8968505859375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.970262390670554, + "grad_norm": 0.1529039740562439, + "learning_rate": 1e-06, + "loss": -0.0126, + "num_tokens": 387561197.0, + "reward": 0.637276828289032, + "reward_std": 0.15534000098705292, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 639 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3818.0, + "completions/mean_length": 868.060302734375, + "completions/mean_terminated_length": 564.5787963867188, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 5.979591836734694, + "grad_norm": 0.1584150344133377, + "learning_rate": 1e-06, + "loss": -0.0192, + "num_tokens": 388113867.0, + "reward": 0.6495535969734192, + "reward_std": 0.165370374917984, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 640 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4007.0, + "completions/mean_length": 906.7176513671875, + "completions/mean_terminated_length": 598.3292846679688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 5.988921282798834, + "grad_norm": 0.1418311595916748, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 388693070.0, + "reward": 0.6886160969734192, + "reward_std": 0.16386333107948303, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 641 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.048295454545454586, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 743.8778686523438, + "completions/mean_terminated_length": 573.7701416015625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 5.998250728862974, + "grad_norm": 0.1640530824661255, + "learning_rate": 1e-06, + "loss": -0.0393, + "num_tokens": 389256557.0, + "reward": 0.6707589626312256, + "reward_std": 0.1731078028678894, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 642 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3275.0, + "completions/mean_length": 856.5614013671875, + "completions/mean_terminated_length": 573.503662109375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.0093294460641395, + "grad_norm": 0.14234739542007446, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 389823220.0, + "reward": 0.6618303656578064, + "reward_std": 0.14297270774841309, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 643 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3624.0, + "completions/mean_length": 869.7377319335938, + "completions/mean_terminated_length": 604.7789916992188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.01865889212828, + "grad_norm": 0.12194245308637619, + "learning_rate": 1e-06, + "loss": -0.0377, + "num_tokens": 390408713.0, + "reward": 0.6361607313156128, + "reward_std": 0.13891443610191345, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 644 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3388.0, + "completions/mean_length": 900.9375610351562, + "completions/mean_terminated_length": 600.5470581054688, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.0279883381924195, + "grad_norm": 0.14325366914272308, + "learning_rate": 1e-06, + "loss": -0.0295, + "num_tokens": 390978953.0, + "reward": 0.6194196939468384, + "reward_std": 0.17156830430030823, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 645 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 880.1574096679688, + "completions/mean_terminated_length": 632.7848510742188, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 6.03731778425656, + "grad_norm": 0.1419011652469635, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 391591414.0, + "reward": 0.6875000596046448, + "reward_std": 0.17964836955070496, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 646 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3961.0, + "completions/mean_length": 894.55029296875, + "completions/mean_terminated_length": 648.2848510742188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 6.0466472303206995, + "grad_norm": 0.1366526186466217, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 392215843.0, + "reward": 0.6428571939468384, + "reward_std": 0.14925476908683777, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 647 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4061.0, + "completions/mean_length": 994.7623291015625, + "completions/mean_terminated_length": 635.5902709960938, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.05597667638484, + "grad_norm": 0.18923039734363556, + "learning_rate": 1e-06, + "loss": -0.0776, + "num_tokens": 392826094.0, + "reward": 0.621651828289032, + "reward_std": 0.2201537787914276, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 648 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 814.2344360351562, + "completions/mean_terminated_length": 536.11865234375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 6.0653061224489795, + "grad_norm": 0.16814157366752625, + "learning_rate": 1e-06, + "loss": -0.0305, + "num_tokens": 393352928.0, + "reward": 0.6674107313156128, + "reward_std": 0.17258675396442413, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 649 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3583.0, + "completions/mean_length": 902.65966796875, + "completions/mean_terminated_length": 580.9717407226562, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 6.07463556851312, + "grad_norm": 0.13550487160682678, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 393916143.0, + "reward": 0.6640625, + "reward_std": 0.1503053456544876, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 650 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 837.122802734375, + "completions/mean_terminated_length": 607.4049682617188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 6.0839650145772595, + "grad_norm": 0.14408986270427704, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 394514197.0, + "reward": 0.6227678656578064, + "reward_std": 0.14635583758354187, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 651 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3004.0, + "completions/mean_length": 1014.7053833007812, + "completions/mean_terminated_length": 627.6080322265625, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 6.093294460641399, + "grad_norm": 0.16759803891181946, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 395094981.0, + "reward": 0.6116071939468384, + "reward_std": 0.19290581345558167, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 652 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 932.5089721679688, + "completions/mean_terminated_length": 635.0867309570312, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 6.1026239067055394, + "grad_norm": 0.11808391660451889, + "learning_rate": 1e-06, + "loss": -0.0355, + "num_tokens": 395705717.0, + "reward": 0.6149553656578064, + "reward_std": 0.11881474405527115, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 653 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 947.1417846679688, + "completions/mean_terminated_length": 651.0952758789062, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 6.111953352769679, + "grad_norm": 0.14583361148834229, + "learning_rate": 1e-06, + "loss": -0.0308, + "num_tokens": 396321052.0, + "reward": 0.5948660969734192, + "reward_std": 0.16961409151554108, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 654 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3367.0, + "completions/mean_length": 898.4910888671875, + "completions/mean_terminated_length": 627.5157470703125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 6.121282798833819, + "grad_norm": 0.14111021161079407, + "learning_rate": 1e-06, + "loss": -0.0317, + "num_tokens": 396923100.0, + "reward": 0.6439732313156128, + "reward_std": 0.1572902798652649, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 655 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 903.6886596679688, + "completions/mean_terminated_length": 645.6851806640625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 6.130612244897959, + "grad_norm": 0.14313118159770966, + "learning_rate": 1e-06, + "loss": -0.0327, + "num_tokens": 397548669.0, + "reward": 0.6361607313156128, + "reward_std": 0.1674344688653946, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 656 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2741.0, + "completions/mean_length": 891.372802734375, + "completions/mean_terminated_length": 615.5806274414062, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 6.139941690962099, + "grad_norm": 0.13182544708251953, + "learning_rate": 1e-06, + "loss": -0.0311, + "num_tokens": 398154883.0, + "reward": 0.6283482313156128, + "reward_std": 0.1338074505329132, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159480571747, + "step": 657 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3931.0, + "completions/mean_length": 852.302490234375, + "completions/mean_terminated_length": 615.3377075195312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 6.149271137026239, + "grad_norm": 0.16187366843223572, + "learning_rate": 1e-06, + "loss": -0.0422, + "num_tokens": 398765794.0, + "reward": 0.6473214626312256, + "reward_std": 0.18419449031352997, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 658 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 960.0089721679688, + "completions/mean_terminated_length": 592.4489135742188, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 6.158600583090379, + "grad_norm": 0.15267084538936615, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 399331946.0, + "reward": 0.6618303656578064, + "reward_std": 0.1725439727306366, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 659 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 945.9129638671875, + "completions/mean_terminated_length": 628.582275390625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.167930029154519, + "grad_norm": 0.1346728652715683, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 399929636.0, + "reward": 0.6540178656578064, + "reward_std": 0.17490653693675995, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 660 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3947.0, + "completions/mean_length": 938.6194458007812, + "completions/mean_terminated_length": 616.2791748046875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.1772594752186585, + "grad_norm": 0.16225330531597137, + "learning_rate": 1e-06, + "loss": -0.0166, + "num_tokens": 400521079.0, + "reward": 0.640625, + "reward_std": 0.1985887587070465, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 661 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3338.0, + "completions/mean_length": 874.9631958007812, + "completions/mean_terminated_length": 614.6381225585938, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 6.186588921282799, + "grad_norm": 0.14276763796806335, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 401126718.0, + "reward": 0.6696428656578064, + "reward_std": 0.17739136517047882, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 662 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3870.0, + "completions/mean_length": 965.37060546875, + "completions/mean_terminated_length": 641.5123291015625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 6.1959183673469385, + "grad_norm": 0.1537197083234787, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 401736746.0, + "reward": 0.6238839626312256, + "reward_std": 0.19227458536624908, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 663 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4052.0, + "completions/mean_length": 931.161865234375, + "completions/mean_terminated_length": 620.8836059570312, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 6.205247813411079, + "grad_norm": 0.13656635582447052, + "learning_rate": 1e-06, + "loss": -0.0112, + "num_tokens": 402331723.0, + "reward": 0.6629464626312256, + "reward_std": 0.14845947921276093, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 664 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3742.0, + "completions/mean_length": 926.0457763671875, + "completions/mean_terminated_length": 653.2376098632812, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.214577259475218, + "grad_norm": 0.14149661362171173, + "learning_rate": 1e-06, + "loss": -0.0384, + "num_tokens": 402962164.0, + "reward": 0.6629464626312256, + "reward_std": 0.16972355544567108, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 665 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3206.0, + "completions/mean_length": 900.45654296875, + "completions/mean_terminated_length": 595.7469482421875, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.223906705539359, + "grad_norm": 0.1426355391740799, + "learning_rate": 1e-06, + "loss": -0.0463, + "num_tokens": 403536453.0, + "reward": 0.6283482313156128, + "reward_std": 0.15082597732543945, + "rewards/verify_math_reward/mean": 0.6283482313156128, + "rewards/verify_math_reward/std": 0.4835159182548523, + "step": 666 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3801.0, + "completions/mean_length": 952.794677734375, + "completions/mean_terminated_length": 678.1456298828125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 6.233236151603498, + "grad_norm": 0.1378466784954071, + "learning_rate": 1e-06, + "loss": -0.0355, + "num_tokens": 404189125.0, + "reward": 0.6149553656578064, + "reward_std": 0.16863587498664856, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 667 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3476.0, + "completions/mean_length": 881.8170166015625, + "completions/mean_terminated_length": 617.8502197265625, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.242565597667639, + "grad_norm": 0.14645691215991974, + "learning_rate": 1e-06, + "loss": -0.0207, + "num_tokens": 404797473.0, + "reward": 0.6785714626312256, + "reward_std": 0.16059784591197968, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 668 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3574.0, + "completions/mean_length": 892.1272583007812, + "completions/mean_terminated_length": 641.5234985351562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.251895043731778, + "grad_norm": 0.1429297924041748, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 405415563.0, + "reward": 0.6707589626312256, + "reward_std": 0.181003138422966, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 669 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3012.0, + "completions/mean_length": 892.958740234375, + "completions/mean_terminated_length": 621.5145263671875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.261224489795918, + "grad_norm": 0.13177646696567535, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 406013766.0, + "reward": 0.6417410969734192, + "reward_std": 0.13696233928203583, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 670 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3991.0, + "completions/mean_length": 1022.7913208007812, + "completions/mean_terminated_length": 654.0062255859375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.270553935860058, + "grad_norm": 0.14521095156669617, + "learning_rate": 1e-06, + "loss": -0.0753, + "num_tokens": 406622603.0, + "reward": 0.6037946939468384, + "reward_std": 0.1802103966474533, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 671 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2827.0, + "completions/mean_length": 940.294677734375, + "completions/mean_terminated_length": 643.6044311523438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 6.279883381924198, + "grad_norm": 0.14574165642261505, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 407235387.0, + "reward": 0.6160714626312256, + "reward_std": 0.16424493491649628, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 672 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2938.0, + "completions/mean_length": 950.3303833007812, + "completions/mean_terminated_length": 624.916259765625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 6.289212827988338, + "grad_norm": 0.1456872522830963, + "learning_rate": 1e-06, + "loss": -0.0487, + "num_tokens": 407831595.0, + "reward": 0.6540178656578064, + "reward_std": 0.15879730880260468, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 673 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 1039.2991943359375, + "completions/mean_terminated_length": 702.1908569335938, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 6.298542274052478, + "grad_norm": 0.14945022761821747, + "learning_rate": 1e-06, + "loss": -0.0315, + "num_tokens": 408498487.0, + "reward": 0.582589328289032, + "reward_std": 0.17217488586902618, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 674 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3519.0, + "completions/mean_length": 935.646240234375, + "completions/mean_terminated_length": 680.2255859375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 6.307871720116618, + "grad_norm": 0.1541002094745636, + "learning_rate": 1e-06, + "loss": -0.0461, + "num_tokens": 409153746.0, + "reward": 0.6183035969734192, + "reward_std": 0.19851066172122955, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 675 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3251.0, + "completions/mean_length": 832.9063110351562, + "completions/mean_terminated_length": 598.712890625, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 6.317201166180758, + "grad_norm": 0.13816270232200623, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 409748926.0, + "reward": 0.6819196939468384, + "reward_std": 0.15420952439308167, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 676 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 1006.64404296875, + "completions/mean_terminated_length": 609.7745361328125, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 6.326530612244898, + "grad_norm": 0.14650343358516693, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 410323615.0, + "reward": 0.625, + "reward_std": 0.1820116490125656, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 677 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3043.0, + "completions/mean_length": 893.0881958007812, + "completions/mean_terminated_length": 663.214111328125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.335860058309038, + "grad_norm": 0.11925744265317917, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 410964878.0, + "reward": 0.6629464626312256, + "reward_std": 0.15146467089653015, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 678 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 812.0592041015625, + "completions/mean_terminated_length": 567.9292602539062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.345189504373177, + "grad_norm": 0.14287491142749786, + "learning_rate": 1e-06, + "loss": 0.0035, + "num_tokens": 411525475.0, + "reward": 0.7131696939468384, + "reward_std": 0.13437016308307648, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342583656311, + "step": 679 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3755.0, + "completions/mean_length": 852.8605346679688, + "completions/mean_terminated_length": 611.7637939453125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.354518950437318, + "grad_norm": 0.1437513679265976, + "learning_rate": 1e-06, + "loss": -0.0128, + "num_tokens": 412123334.0, + "reward": 0.6930803656578064, + "reward_std": 0.13606193661689758, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 680 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3154.0, + "completions/mean_length": 857.9230346679688, + "completions/mean_terminated_length": 570.7059936523438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.363848396501457, + "grad_norm": 0.1336721032857895, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 412680489.0, + "reward": 0.6540178656578064, + "reward_std": 0.155109241604805, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 681 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3161.0, + "completions/mean_length": 910.1629638671875, + "completions/mean_terminated_length": 631.788818359375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 6.373177842565598, + "grad_norm": 0.14269065856933594, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 413286707.0, + "reward": 0.6551339626312256, + "reward_std": 0.16642485558986664, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 682 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 974.6585083007812, + "completions/mean_terminated_length": 664.4392700195312, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.382507288629737, + "grad_norm": 0.14866507053375244, + "learning_rate": 1e-06, + "loss": -0.0385, + "num_tokens": 413918145.0, + "reward": 0.6227678656578064, + "reward_std": 0.15357083082199097, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 683 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 981.12841796875, + "completions/mean_terminated_length": 658.9002075195312, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 6.391836734693878, + "grad_norm": 0.15749813616275787, + "learning_rate": 1e-06, + "loss": -0.0186, + "num_tokens": 414549188.0, + "reward": 0.5859375, + "reward_std": 0.18092785775661469, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 684 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 952.71435546875, + "completions/mean_terminated_length": 661.3853759765625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 6.401166180758017, + "grad_norm": 0.12710365653038025, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 415189156.0, + "reward": 0.6350446939468384, + "reward_std": 0.15079136192798615, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 685 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3387.0, + "completions/mean_length": 863.5067138671875, + "completions/mean_terminated_length": 602.2557373046875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 6.410495626822158, + "grad_norm": 0.14502064883708954, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 415779874.0, + "reward": 0.640625, + "reward_std": 0.1710023432970047, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 686 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 989.9766235351562, + "completions/mean_terminated_length": 621.5967407226562, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 6.419825072886297, + "grad_norm": 0.1616830825805664, + "learning_rate": 1e-06, + "loss": -0.0333, + "num_tokens": 416369597.0, + "reward": 0.640625, + "reward_std": 0.1674344390630722, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 687 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3798.0, + "completions/mean_length": 972.2199096679688, + "completions/mean_terminated_length": 678.5311889648438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 6.429154518950437, + "grad_norm": 0.1500493437051773, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 417016282.0, + "reward": 0.6417410969734192, + "reward_std": 0.1824253350496292, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 688 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3898.0, + "completions/mean_length": 1091.552490234375, + "completions/mean_terminated_length": 662.3456420898438, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 6.438483965014577, + "grad_norm": 0.13980795443058014, + "learning_rate": 1e-06, + "loss": -0.0313, + "num_tokens": 417626809.0, + "reward": 0.566964328289032, + "reward_std": 0.14553210139274597, + "rewards/verify_math_reward/mean": 0.5669642686843872, + "rewards/verify_math_reward/std": 0.49577224254608154, + "step": 689 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2959.0, + "completions/mean_length": 955.247802734375, + "completions/mean_terminated_length": 608.869873046875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.447813411078717, + "grad_norm": 0.15495805442333221, + "learning_rate": 1e-06, + "loss": -0.0322, + "num_tokens": 418207175.0, + "reward": 0.6718750596046448, + "reward_std": 0.1691262423992157, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 690 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4093.0, + "completions/mean_length": 903.86279296875, + "completions/mean_terminated_length": 612.2545776367188, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 6.457142857142857, + "grad_norm": 0.14153935015201569, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 418795676.0, + "reward": 0.6863839626312256, + "reward_std": 0.1726934313774109, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 691 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0535714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3847.0, + "completions/mean_length": 865.4375610351562, + "completions/mean_terminated_length": 682.5755004882812, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.466472303206997, + "grad_norm": 0.1511821299791336, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 419461996.0, + "reward": 0.6863839626312256, + "reward_std": 0.19425947964191437, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 692 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3805.0, + "completions/mean_length": 1053.759033203125, + "completions/mean_terminated_length": 619.153076171875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.475801749271137, + "grad_norm": 0.14980465173721313, + "learning_rate": 1e-06, + "loss": -0.0588, + "num_tokens": 420038196.0, + "reward": 0.660714328289032, + "reward_std": 0.14740821719169617, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 693 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 1087.680908203125, + "completions/mean_terminated_length": 666.669189453125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.485131195335277, + "grad_norm": 0.13873988389968872, + "learning_rate": 1e-06, + "loss": -0.033, + "num_tokens": 420656022.0, + "reward": 0.6049107313156128, + "reward_std": 0.15236620604991913, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 694 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3044.0, + "completions/mean_length": 912.9375610351562, + "completions/mean_terminated_length": 579.3242797851562, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 6.494460641399417, + "grad_norm": 0.15610650181770325, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 421212630.0, + "reward": 0.6495535969734192, + "reward_std": 0.15613093972206116, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 695 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4069.0, + "completions/mean_length": 985.4364013671875, + "completions/mean_terminated_length": 607.8085327148438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 6.503790087463557, + "grad_norm": 0.14351730048656464, + "learning_rate": 1e-06, + "loss": -0.0186, + "num_tokens": 421790509.0, + "reward": 0.6261160969734192, + "reward_std": 0.13534724712371826, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 696 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3434.0, + "completions/mean_length": 831.9453735351562, + "completions/mean_terminated_length": 572.3939819335938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 6.513119533527696, + "grad_norm": 0.1457185596227646, + "learning_rate": 1e-06, + "loss": -0.0266, + "num_tokens": 422345764.0, + "reward": 0.6964285969734192, + "reward_std": 0.14861032366752625, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 697 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3618.0, + "completions/mean_length": 929.6719360351562, + "completions/mean_terminated_length": 669.63525390625, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 6.522448979591837, + "grad_norm": 0.1634896993637085, + "learning_rate": 1e-06, + "loss": -0.0387, + "num_tokens": 422982518.0, + "reward": 0.6473214626312256, + "reward_std": 0.1843799650669098, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 698 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3476.0, + "completions/mean_length": 1110.2935791015625, + "completions/mean_terminated_length": 666.265380859375, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 6.531778425655976, + "grad_norm": 0.1411629021167755, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 423595085.0, + "reward": 0.5602678656578064, + "reward_std": 0.17115317285060883, + "rewards/verify_math_reward/mean": 0.5602678656578064, + "rewards/verify_math_reward/std": 0.4966317415237427, + "step": 699 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3406.0, + "completions/mean_length": 832.7969360351562, + "completions/mean_terminated_length": 602.774169921875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 6.541107871720117, + "grad_norm": 0.1352481096982956, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 424177959.0, + "reward": 0.6640625, + "reward_std": 0.1339162141084671, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 700 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 973.8538208007812, + "completions/mean_terminated_length": 655.1107177734375, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 6.550437317784256, + "grad_norm": 0.13937661051750183, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 424796572.0, + "reward": 0.6595982313156128, + "reward_std": 0.17371399700641632, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 701 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4004.0, + "completions/mean_length": 919.122802734375, + "completions/mean_terminated_length": 573.126220703125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.559766763848397, + "grad_norm": 0.15014061331748962, + "learning_rate": 1e-06, + "loss": -0.0601, + "num_tokens": 425345554.0, + "reward": 0.7254464626312256, + "reward_std": 0.17720702290534973, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 702 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3873.0, + "completions/mean_length": 1026.7679443359375, + "completions/mean_terminated_length": 654.15771484375, + "completions/min_length": 83.0, + "completions/min_terminated_length": 83.0, + "epoch": 6.569096209912536, + "grad_norm": 0.14446890354156494, + "learning_rate": 1e-06, + "loss": -0.0528, + "num_tokens": 425961666.0, + "reward": 0.645089328289032, + "reward_std": 0.15131382644176483, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 703 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 1053.071533203125, + "completions/mean_terminated_length": 657.8360595703125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 6.578425655976677, + "grad_norm": 0.1557384729385376, + "learning_rate": 1e-06, + "loss": -0.0831, + "num_tokens": 426580882.0, + "reward": 0.59375, + "reward_std": 0.17152412235736847, + "rewards/verify_math_reward/mean": 0.59375, + "rewards/verify_math_reward/std": 0.4914066195487976, + "step": 704 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2204.0, + "completions/mean_length": 876.7288208007812, + "completions/mean_terminated_length": 620.7385864257812, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 6.587755102040816, + "grad_norm": 0.13753269612789154, + "learning_rate": 1e-06, + "loss": -0.0119, + "num_tokens": 427181935.0, + "reward": 0.691964328289032, + "reward_std": 0.14060622453689575, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 705 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3479.0, + "completions/mean_length": 897.4285888671875, + "completions/mean_terminated_length": 613.7156982421875, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 6.597084548104956, + "grad_norm": 0.15909099578857422, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 427775327.0, + "reward": 0.6428571939468384, + "reward_std": 0.18227237462997437, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 706 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2677.0, + "completions/mean_length": 974.08935546875, + "completions/mean_terminated_length": 603.8251953125, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 6.606413994169096, + "grad_norm": 0.1492423266172409, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 428344783.0, + "reward": 0.621651828289032, + "reward_std": 0.15680170059204102, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.485245943069458, + "step": 707 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3719.0, + "completions/mean_length": 902.0402221679688, + "completions/mean_terminated_length": 622.956298828125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.615743440233236, + "grad_norm": 0.14789535105228424, + "learning_rate": 1e-06, + "loss": -0.0525, + "num_tokens": 428937483.0, + "reward": 0.676339328289032, + "reward_std": 0.17194482684135437, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 708 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 884.72998046875, + "completions/mean_terminated_length": 569.8995361328125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 6.625072886297376, + "grad_norm": 0.15483340620994568, + "learning_rate": 1e-06, + "loss": -0.05, + "num_tokens": 429488153.0, + "reward": 0.7053571939468384, + "reward_std": 0.14120353758335114, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 709 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3953.0, + "completions/mean_length": 1068.734375, + "completions/mean_terminated_length": 636.267822265625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 6.634402332361516, + "grad_norm": 0.15241001546382904, + "learning_rate": 1e-06, + "loss": -0.0536, + "num_tokens": 430078747.0, + "reward": 0.6127232313156128, + "reward_std": 0.17661041021347046, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 710 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2739.0, + "completions/mean_length": 817.0335083007812, + "completions/mean_terminated_length": 585.8995971679688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 6.643731778425656, + "grad_norm": 0.1431237906217575, + "learning_rate": 1e-06, + "loss": -0.0372, + "num_tokens": 430651977.0, + "reward": 0.6852678656578064, + "reward_std": 0.1448495239019394, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 711 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 1037.8973388671875, + "completions/mean_terminated_length": 662.3408203125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 6.653061224489796, + "grad_norm": 0.1589026004076004, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 431270061.0, + "reward": 0.6261160969734192, + "reward_std": 0.16976633667945862, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 712 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3211.0, + "completions/mean_length": 964.5949096679688, + "completions/mean_terminated_length": 597.5723266601562, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 6.662390670553936, + "grad_norm": 0.14149640500545502, + "learning_rate": 1e-06, + "loss": -0.0739, + "num_tokens": 431840346.0, + "reward": 0.7098214626312256, + "reward_std": 0.15195322036743164, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 713 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3896.0, + "completions/mean_length": 930.6886596679688, + "completions/mean_terminated_length": 641.5310668945312, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 6.671720116618076, + "grad_norm": 0.149410679936409, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 432458187.0, + "reward": 0.645089328289032, + "reward_std": 0.1560874730348587, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 714 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3469.0, + "completions/mean_length": 947.26904296875, + "completions/mean_terminated_length": 655.4353637695312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.681049562682215, + "grad_norm": 0.1367138922214508, + "learning_rate": 1e-06, + "loss": -0.038, + "num_tokens": 433090844.0, + "reward": 0.6205357313156128, + "reward_std": 0.1543956845998764, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 715 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3140.0, + "completions/mean_length": 970.5145263671875, + "completions/mean_terminated_length": 642.9358520507812, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 6.690379008746356, + "grad_norm": 0.15205246210098267, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 433710841.0, + "reward": 0.6026785969734192, + "reward_std": 0.1888468712568283, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 716 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3867.0, + "completions/mean_length": 1123.083740234375, + "completions/mean_terminated_length": 632.10791015625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 6.699708454810495, + "grad_norm": 0.16829803586006165, + "learning_rate": 1e-06, + "loss": -0.0661, + "num_tokens": 434291212.0, + "reward": 0.5915178656578064, + "reward_std": 0.1829131692647934, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 717 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3726.0, + "completions/mean_length": 1159.1395263671875, + "completions/mean_terminated_length": 669.6627807617188, + "completions/min_length": 190.0, + "completions/min_terminated_length": 190.0, + "epoch": 6.709037900874636, + "grad_norm": 0.16678796708583832, + "learning_rate": 1e-06, + "loss": -0.0724, + "num_tokens": 434897105.0, + "reward": 0.6049107313156128, + "reward_std": 0.19261159002780914, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 718 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2520.0, + "completions/mean_length": 1044.89404296875, + "completions/mean_terminated_length": 644.2437133789062, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 6.718367346938775, + "grad_norm": 0.145506352186203, + "learning_rate": 1e-06, + "loss": -0.021, + "num_tokens": 435502114.0, + "reward": 0.609375, + "reward_std": 0.15326520800590515, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 719 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2719.0, + "completions/mean_length": 1008.6719360351562, + "completions/mean_terminated_length": 651.11083984375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 6.727696793002916, + "grad_norm": 0.16412486135959625, + "learning_rate": 1e-06, + "loss": -0.0313, + "num_tokens": 436114468.0, + "reward": 0.5993303656578064, + "reward_std": 0.14586910605430603, + "rewards/verify_math_reward/mean": 0.5993303656578064, + "rewards/verify_math_reward/std": 0.49030786752700806, + "step": 720 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3120.0, + "completions/mean_length": 901.427490234375, + "completions/mean_terminated_length": 601.0830688476562, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.737026239067055, + "grad_norm": 0.14590677618980408, + "learning_rate": 1e-06, + "loss": -0.0433, + "num_tokens": 436704867.0, + "reward": 0.65625, + "reward_std": 0.18516905605793, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 721 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3546.0, + "completions/mean_length": 1101.8515625, + "completions/mean_terminated_length": 674.1160278320312, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 6.746355685131196, + "grad_norm": 0.14151152968406677, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 437328086.0, + "reward": 0.5915178656578064, + "reward_std": 0.1660429686307907, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 722 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3775.0, + "completions/mean_length": 910.4230346679688, + "completions/mean_terminated_length": 619.4141845703125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 6.755685131195335, + "grad_norm": 0.13452281057834625, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 437930249.0, + "reward": 0.6640625, + "reward_std": 0.14680413901805878, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 723 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3963.0, + "completions/mean_length": 941.9810791015625, + "completions/mean_terminated_length": 658.0426025390625, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 6.765014577259475, + "grad_norm": 0.13875821232795715, + "learning_rate": 1e-06, + "loss": -0.0649, + "num_tokens": 438555760.0, + "reward": 0.7343750596046448, + "reward_std": 0.17836818099021912, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 724 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3943.0, + "completions/mean_length": 1076.5592041015625, + "completions/mean_terminated_length": 675.7484741210938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 6.774344023323615, + "grad_norm": 0.13399802148342133, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 439193261.0, + "reward": 0.6383928656578064, + "reward_std": 0.1561630368232727, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 725 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3305.0, + "completions/mean_length": 857.1607666015625, + "completions/mean_terminated_length": 574.1553344726562, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 6.783673469387755, + "grad_norm": 0.14856798946857452, + "learning_rate": 1e-06, + "loss": -0.0423, + "num_tokens": 439751413.0, + "reward": 0.6808035969734192, + "reward_std": 0.14992626011371613, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 726 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3847.0, + "completions/mean_length": 953.8739013671875, + "completions/mean_terminated_length": 633.0910034179688, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 6.793002915451895, + "grad_norm": 0.11999521404504776, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 440358428.0, + "reward": 0.6183035969734192, + "reward_std": 0.13373005390167236, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 727 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3954.0, + "completions/mean_length": 1010.42529296875, + "completions/mean_terminated_length": 609.6506958007812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.802332361516035, + "grad_norm": 0.19046589732170105, + "learning_rate": 1e-06, + "loss": -0.0594, + "num_tokens": 440922769.0, + "reward": 0.6517857313156128, + "reward_std": 0.19670899212360382, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 728 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3262.0, + "completions/mean_length": 1006.1138916015625, + "completions/mean_terminated_length": 652.5447387695312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 6.811661807580175, + "grad_norm": 0.16059322655200958, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 441544191.0, + "reward": 0.6116071939468384, + "reward_std": 0.18047182261943817, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 729 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 1020.1785888671875, + "completions/mean_terminated_length": 672.4769897460938, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 6.820991253644315, + "grad_norm": 0.15191487967967987, + "learning_rate": 1e-06, + "loss": -0.0501, + "num_tokens": 442179631.0, + "reward": 0.6149553656578064, + "reward_std": 0.19881586730480194, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 730 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3857.0, + "completions/mean_length": 944.2902221679688, + "completions/mean_terminated_length": 647.9755859375, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 6.830320699708455, + "grad_norm": 0.16478540003299713, + "learning_rate": 1e-06, + "loss": -0.0576, + "num_tokens": 442799195.0, + "reward": 0.6183035969734192, + "reward_std": 0.18088370561599731, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 731 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3638.0, + "completions/mean_length": 1061.7410888671875, + "completions/mean_terminated_length": 641.4942626953125, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 6.839650145772595, + "grad_norm": 0.1515892744064331, + "learning_rate": 1e-06, + "loss": -0.0478, + "num_tokens": 443386363.0, + "reward": 0.65625, + "reward_std": 0.1409093141555786, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 732 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2568.0, + "completions/mean_length": 1060.466552734375, + "completions/mean_terminated_length": 661.861083984375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 6.848979591836734, + "grad_norm": 0.14148367941379547, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 443993661.0, + "reward": 0.625, + "reward_std": 0.12692944705486298, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 733 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4074.0, + "completions/mean_length": 1087.958740234375, + "completions/mean_terminated_length": 688.6611938476562, + "completions/min_length": 216.0, + "completions/min_terminated_length": 216.0, + "epoch": 6.858309037900875, + "grad_norm": 0.15652503073215485, + "learning_rate": 1e-06, + "loss": -0.0304, + "num_tokens": 444626224.0, + "reward": 0.6049107313156128, + "reward_std": 0.163072407245636, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 734 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3020.0, + "completions/mean_length": 981.380615234375, + "completions/mean_terminated_length": 654.9407958984375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 6.867638483965014, + "grad_norm": 0.14108288288116455, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 445250269.0, + "reward": 0.6517857313156128, + "reward_std": 0.13354460895061493, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 735 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2872.0, + "completions/mean_length": 950.1272583007812, + "completions/mean_terminated_length": 568.2127685546875, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 6.876967930029155, + "grad_norm": 0.15446877479553223, + "learning_rate": 1e-06, + "loss": -0.0846, + "num_tokens": 445789239.0, + "reward": 0.6875000596046448, + "reward_std": 0.15267042815685272, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 736 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3761.0, + "completions/mean_length": 984.9152221679688, + "completions/mean_terminated_length": 654.6024780273438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 6.886297376093294, + "grad_norm": 0.17698132991790771, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 446415867.0, + "reward": 0.6004464626312256, + "reward_std": 0.1683780699968338, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 737 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3886.0, + "completions/mean_length": 970.7935791015625, + "completions/mean_terminated_length": 591.3880004882812, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 6.895626822157435, + "grad_norm": 0.12509697675704956, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 446979690.0, + "reward": 0.6316964626312256, + "reward_std": 0.12433614581823349, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 738 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2788.0, + "completions/mean_length": 1183.818115234375, + "completions/mean_terminated_length": 724.7919921875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 6.904956268221574, + "grad_norm": 0.1635567843914032, + "learning_rate": 1e-06, + "loss": -0.0528, + "num_tokens": 447636855.0, + "reward": 0.5412946939468384, + "reward_std": 0.19016842544078827, + "rewards/verify_math_reward/mean": 0.5412946343421936, + "rewards/verify_math_reward/std": 0.49857014417648315, + "step": 739 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3124.0, + "completions/mean_length": 1079.9832763671875, + "completions/mean_terminated_length": 644.7215576171875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 6.914285714285715, + "grad_norm": 0.15128153562545776, + "learning_rate": 1e-06, + "loss": -0.0548, + "num_tokens": 448233488.0, + "reward": 0.621651828289032, + "reward_std": 0.18043790757656097, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 740 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3537.0, + "completions/mean_length": 930.4464721679688, + "completions/mean_terminated_length": 628.5966186523438, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 6.923615160349854, + "grad_norm": 0.15063291788101196, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 448837144.0, + "reward": 0.6104910969734192, + "reward_std": 0.1652994155883789, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 741 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3951.0, + "completions/mean_length": 1076.9832763671875, + "completions/mean_terminated_length": 641.2886352539062, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 6.932944606413994, + "grad_norm": 0.1616082787513733, + "learning_rate": 1e-06, + "loss": -0.075, + "num_tokens": 449428305.0, + "reward": 0.625, + "reward_std": 0.18032237887382507, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 742 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 881.8839721679688, + "completions/mean_terminated_length": 605.275146484375, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 6.942274052478134, + "grad_norm": 0.15175040066242218, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 450007769.0, + "reward": 0.6662946939468384, + "reward_std": 0.15030281245708466, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 743 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2878.0, + "completions/mean_length": 975.55029296875, + "completions/mean_terminated_length": 618.4838256835938, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 6.9516034985422746, + "grad_norm": 0.14866061508655548, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 450594838.0, + "reward": 0.6808035969734192, + "reward_std": 0.16037032008171082, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 744 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3004.0, + "completions/mean_length": 864.8783569335938, + "completions/mean_terminated_length": 569.7089233398438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 6.960932944606414, + "grad_norm": 0.15572293102741241, + "learning_rate": 1e-06, + "loss": -0.0342, + "num_tokens": 451146817.0, + "reward": 0.7198660969734192, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 745 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2350.0, + "completions/mean_length": 936.5000610351562, + "completions/mean_terminated_length": 618.2211303710938, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 6.970262390670554, + "grad_norm": 0.15726704895496368, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 451729737.0, + "reward": 0.6852678656578064, + "reward_std": 0.18393094837665558, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 746 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2202.0, + "completions/mean_length": 997.3973388671875, + "completions/mean_terminated_length": 642.830810546875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 6.979591836734694, + "grad_norm": 0.1498706042766571, + "learning_rate": 1e-06, + "loss": -0.0218, + "num_tokens": 452339277.0, + "reward": 0.613839328289032, + "reward_std": 0.14458850026130676, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 747 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3101.0, + "completions/mean_length": 970.0022583007812, + "completions/mean_terminated_length": 638.106201171875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 6.988921282798834, + "grad_norm": 0.1454838216304779, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 452948351.0, + "reward": 0.6127232313156128, + "reward_std": 0.15721359848976135, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 748 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3443.0, + "completions/mean_length": 907.6392211914062, + "completions/mean_terminated_length": 577.8087768554688, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 6.998250728862974, + "grad_norm": 0.14799146354198456, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 453556804.0, + "reward": 0.6238839626312256, + "reward_std": 0.13470645248889923, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 749 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 917.45654296875, + "completions/mean_terminated_length": 627.0901489257812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 7.0093294460641395, + "grad_norm": 0.1317032277584076, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 454149237.0, + "reward": 0.699776828289032, + "reward_std": 0.15988247096538544, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 750 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3164.0, + "completions/mean_length": 992.5145263671875, + "completions/mean_terminated_length": 663.0086669921875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.01865889212828, + "grad_norm": 0.15026330947875977, + "learning_rate": 1e-06, + "loss": -0.0301, + "num_tokens": 454767994.0, + "reward": 0.6004464626312256, + "reward_std": 0.16544592380523682, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 751 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4083.0, + "completions/mean_length": 1095.052490234375, + "completions/mean_terminated_length": 670.7146606445312, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.0279883381924195, + "grad_norm": 0.15019945800304413, + "learning_rate": 1e-06, + "loss": -0.053, + "num_tokens": 455384865.0, + "reward": 0.6272321939468384, + "reward_std": 0.17746664583683014, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111400604248, + "step": 752 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2482.0, + "completions/mean_length": 987.6451416015625, + "completions/mean_terminated_length": 618.98876953125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 7.03731778425656, + "grad_norm": 0.1614040732383728, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 455973283.0, + "reward": 0.6674107313156128, + "reward_std": 0.17900758981704712, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 753 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2577.0, + "completions/mean_length": 884.88623046875, + "completions/mean_terminated_length": 621.1715087890625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 7.0466472303206995, + "grad_norm": 0.15174107253551483, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 456570725.0, + "reward": 0.6986607313156128, + "reward_std": 0.154775470495224, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 754 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3830.0, + "completions/mean_length": 990.380615234375, + "completions/mean_terminated_length": 613.3529663085938, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 7.05597667638484, + "grad_norm": 0.15587319433689117, + "learning_rate": 1e-06, + "loss": -0.0137, + "num_tokens": 457143722.0, + "reward": 0.6651785969734192, + "reward_std": 0.15240967273712158, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 755 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 948.474365234375, + "completions/mean_terminated_length": 635.6527709960938, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 7.0653061224489795, + "grad_norm": 0.15170101821422577, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 457744315.0, + "reward": 0.640625, + "reward_std": 0.15274599194526672, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 756 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3473.0, + "completions/mean_length": 947.349365234375, + "completions/mean_terminated_length": 600.1004028320312, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 7.07463556851312, + "grad_norm": 0.1556454300880432, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 458316044.0, + "reward": 0.65625, + "reward_std": 0.1766764223575592, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 757 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3354.0, + "completions/mean_length": 962.36279296875, + "completions/mean_terminated_length": 599.4383544921875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.0839650145772595, + "grad_norm": 0.1393265575170517, + "learning_rate": 1e-06, + "loss": -0.0452, + "num_tokens": 458890897.0, + "reward": 0.660714328289032, + "reward_std": 0.13496677577495575, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 758 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4004.0, + "completions/mean_length": 1038.4398193359375, + "completions/mean_terminated_length": 658.6423950195312, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.093294460641399, + "grad_norm": 0.1311669498682022, + "learning_rate": 1e-06, + "loss": -0.0222, + "num_tokens": 459515091.0, + "reward": 0.5625, + "reward_std": 0.1151694804430008, + "rewards/verify_math_reward/mean": 0.5625, + "rewards/verify_math_reward/std": 0.49635544419288635, + "step": 759 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4009.0, + "completions/mean_length": 909.989990234375, + "completions/mean_terminated_length": 601.9180297851562, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.1026239067055394, + "grad_norm": 0.13678602874279022, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 460089826.0, + "reward": 0.6830357313156128, + "reward_std": 0.15101072192192078, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 760 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4012.0, + "completions/mean_length": 927.91748046875, + "completions/mean_terminated_length": 634.2902221679688, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.111953352769679, + "grad_norm": 0.15244071185588837, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 460699960.0, + "reward": 0.6986607313156128, + "reward_std": 0.16157494485378265, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 761 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2406.0, + "completions/mean_length": 917.0313110351562, + "completions/mean_terminated_length": 583.8471069335938, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 7.121282798833819, + "grad_norm": 0.1366218626499176, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 461255428.0, + "reward": 0.7220982313156128, + "reward_std": 0.1292613297700882, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 762 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 954.5145263671875, + "completions/mean_terminated_length": 663.3524169921875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.130612244897959, + "grad_norm": 0.13859689235687256, + "learning_rate": 1e-06, + "loss": -0.0737, + "num_tokens": 461890369.0, + "reward": 0.6517857313156128, + "reward_std": 0.13095128536224365, + "rewards/verify_math_reward/mean": 0.6517857313156128, + "rewards/verify_math_reward/std": 0.47667041420936584, + "step": 763 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3923.0, + "completions/mean_length": 1081.4765625, + "completions/mean_terminated_length": 619.7927856445312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.139941690962099, + "grad_norm": 0.15567605197429657, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 462471204.0, + "reward": 0.5457589626312256, + "reward_std": 0.15996167063713074, + "rewards/verify_math_reward/mean": 0.5457589030265808, + "rewards/verify_math_reward/std": 0.4981798231601715, + "step": 764 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2903.0, + "completions/mean_length": 901.9364013671875, + "completions/mean_terminated_length": 588.7929077148438, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 7.149271137026239, + "grad_norm": 0.15489253401756287, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 463039995.0, + "reward": 0.6729910969734192, + "reward_std": 0.15202920138835907, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 765 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2981.0, + "completions/mean_length": 915.9185791015625, + "completions/mean_terminated_length": 695.8174438476562, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 7.158600583090379, + "grad_norm": 0.15642905235290527, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 463710370.0, + "reward": 0.6774553656578064, + "reward_std": 0.2059541493654251, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 766 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3409.0, + "completions/mean_length": 922.7366333007812, + "completions/mean_terminated_length": 607.3570556640625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.167930029154519, + "grad_norm": 0.13663925230503082, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 464291590.0, + "reward": 0.6506696939468384, + "reward_std": 0.1560867577791214, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 767 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3000.0, + "completions/mean_length": 1055.7388916015625, + "completions/mean_terminated_length": 652.1643676757812, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 7.1772594752186585, + "grad_norm": 0.15713904798030853, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 464894868.0, + "reward": 0.6171875, + "reward_std": 0.14071568846702576, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 768 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3536.0, + "completions/mean_length": 1064.099365234375, + "completions/mean_terminated_length": 670.2963256835938, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 7.186588921282799, + "grad_norm": 0.15653076767921448, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 465512317.0, + "reward": 0.6718750596046448, + "reward_std": 0.19013704359531403, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 769 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3706.0, + "completions/mean_length": 873.4989013671875, + "completions/mean_terminated_length": 646.3452758789062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 7.1959183673469385, + "grad_norm": 0.15411219000816345, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 466144420.0, + "reward": 0.6986607313156128, + "reward_std": 0.1641671359539032, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960144996643, + "step": 770 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3881.0, + "completions/mean_length": 1104.9498291015625, + "completions/mean_terminated_length": 690.6873779296875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.205247813411079, + "grad_norm": 0.13861779868602753, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 466776959.0, + "reward": 0.6361607313156128, + "reward_std": 0.14432887732982635, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 771 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3738.0, + "completions/mean_length": 1007.19873046875, + "completions/mean_terminated_length": 627.8721923828125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.214577259475218, + "grad_norm": 0.14062440395355225, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 467373529.0, + "reward": 0.6651785969734192, + "reward_std": 0.13647310435771942, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 772 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2984.0, + "completions/mean_length": 1030.2288818359375, + "completions/mean_terminated_length": 614.4651489257812, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.223906705539359, + "grad_norm": 0.1771906465291977, + "learning_rate": 1e-06, + "loss": -0.0616, + "num_tokens": 467943654.0, + "reward": 0.6584821939468384, + "reward_std": 0.17900758981704712, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 773 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3313.0, + "completions/mean_length": 999.1250610351562, + "completions/mean_terminated_length": 623.158935546875, + "completions/min_length": 101.0, + "completions/min_terminated_length": 101.0, + "epoch": 7.233236151603498, + "grad_norm": 0.1333370953798294, + "learning_rate": 1e-06, + "loss": -0.0484, + "num_tokens": 468531278.0, + "reward": 0.6071428656578064, + "reward_std": 0.14489158987998962, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 774 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3495.0, + "completions/mean_length": 1019.3125610351562, + "completions/mean_terminated_length": 624.0704956054688, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.242565597667639, + "grad_norm": 0.1312686651945114, + "learning_rate": 1e-06, + "loss": -0.0356, + "num_tokens": 469117918.0, + "reward": 0.6785714626312256, + "reward_std": 0.11144820600748062, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 775 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3048.0, + "completions/mean_length": 973.7522583007812, + "completions/mean_terminated_length": 646.512939453125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 7.251895043731778, + "grad_norm": 0.12949679791927338, + "learning_rate": 1e-06, + "loss": -0.0623, + "num_tokens": 469727560.0, + "reward": 0.6584821939468384, + "reward_std": 0.1241927519440651, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 776 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3634.0, + "completions/mean_length": 1050.515625, + "completions/mean_terminated_length": 650.603515625, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 7.261224489795918, + "grad_norm": 0.1619434654712677, + "learning_rate": 1e-06, + "loss": -0.0691, + "num_tokens": 470328110.0, + "reward": 0.6149553656578064, + "reward_std": 0.193997323513031, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 777 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3848.0, + "completions/mean_length": 1066.993408203125, + "completions/mean_terminated_length": 695.010009765625, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 7.270553935860058, + "grad_norm": 0.1315089762210846, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 470982376.0, + "reward": 0.582589328289032, + "reward_std": 0.16150076687335968, + "rewards/verify_math_reward/mean": 0.5825892686843872, + "rewards/verify_math_reward/std": 0.4934072494506836, + "step": 778 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2332.0, + "completions/mean_length": 967.7756958007812, + "completions/mean_terminated_length": 627.0779418945312, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 7.279883381924198, + "grad_norm": 0.15142987668514252, + "learning_rate": 1e-06, + "loss": -0.0379, + "num_tokens": 471580063.0, + "reward": 0.6439732313156128, + "reward_std": 0.16424313187599182, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 779 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3239.0, + "completions/mean_length": 931.7656860351562, + "completions/mean_terminated_length": 600.125732421875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.289212827988338, + "grad_norm": 0.1581142544746399, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 472152669.0, + "reward": 0.6729910969734192, + "reward_std": 0.1589067578315735, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 780 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 948.0156860351562, + "completions/mean_terminated_length": 622.362060546875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.298542274052478, + "grad_norm": 0.15665648877620697, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 472743491.0, + "reward": 0.6350446939468384, + "reward_std": 0.17066533863544464, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 781 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3151.0, + "completions/mean_length": 952.6038208007812, + "completions/mean_terminated_length": 635.9471435546875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 7.307871720116618, + "grad_norm": 0.15677344799041748, + "learning_rate": 1e-06, + "loss": -0.0356, + "num_tokens": 473347912.0, + "reward": 0.6227678656578064, + "reward_std": 0.15714125335216522, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 782 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3440.0, + "completions/mean_length": 993.1261596679688, + "completions/mean_terminated_length": 650.9256591796875, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 7.317201166180758, + "grad_norm": 0.1502755582332611, + "learning_rate": 1e-06, + "loss": -0.0216, + "num_tokens": 473965465.0, + "reward": 0.640625, + "reward_std": 0.1380895972251892, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 783 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3555.0, + "completions/mean_length": 934.021240234375, + "completions/mean_terminated_length": 636.7411499023438, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 7.326530612244898, + "grad_norm": 0.13268038630485535, + "learning_rate": 1e-06, + "loss": -0.0058, + "num_tokens": 474571924.0, + "reward": 0.6328125, + "reward_std": 0.14053022861480713, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 784 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2837.0, + "completions/mean_length": 907.6719360351562, + "completions/mean_terminated_length": 616.4117431640625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 7.335860058309038, + "grad_norm": 0.16889922320842743, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 475160222.0, + "reward": 0.6618303656578064, + "reward_std": 0.1765669286251068, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 785 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3191.0, + "completions/mean_length": 946.9006958007812, + "completions/mean_terminated_length": 603.929443359375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 7.345189504373177, + "grad_norm": 0.14676925539970398, + "learning_rate": 1e-06, + "loss": -0.0097, + "num_tokens": 475740885.0, + "reward": 0.6071428656578064, + "reward_std": 0.13764451444149017, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 786 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3170.0, + "completions/mean_length": 899.8348388671875, + "completions/mean_terminated_length": 599.3406982421875, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 7.354518950437318, + "grad_norm": 0.1353752315044403, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 476326321.0, + "reward": 0.6819196939468384, + "reward_std": 0.1465073823928833, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 787 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2171.0, + "completions/mean_length": 823.1004638671875, + "completions/mean_terminated_length": 554.3115844726562, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 7.363848396501457, + "grad_norm": 0.17552407085895538, + "learning_rate": 1e-06, + "loss": -0.036, + "num_tokens": 476875843.0, + "reward": 0.6383928656578064, + "reward_std": 0.16724900901317596, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 788 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3850.0, + "completions/mean_length": 1052.532470703125, + "completions/mean_terminated_length": 670.1871948242188, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.373177842565598, + "grad_norm": 0.12911708652973175, + "learning_rate": 1e-06, + "loss": -0.0386, + "num_tokens": 477501792.0, + "reward": 0.6160714626312256, + "reward_std": 0.1497408002614975, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 789 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3505.0, + "completions/mean_length": 1014.8281860351562, + "completions/mean_terminated_length": 619.0100708007812, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 7.382507288629737, + "grad_norm": 0.1339099258184433, + "learning_rate": 1e-06, + "loss": -0.0417, + "num_tokens": 478089830.0, + "reward": 0.6696428656578064, + "reward_std": 0.12805598974227905, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 790 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3517.0, + "completions/mean_length": 967.5435791015625, + "completions/mean_terminated_length": 626.820556640625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 7.391836734693878, + "grad_norm": 0.11781653016805649, + "learning_rate": 1e-06, + "loss": -0.035, + "num_tokens": 478682877.0, + "reward": 0.7042410969734192, + "reward_std": 0.1054396703839302, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 791 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3237.0, + "completions/mean_length": 988.2701416015625, + "completions/mean_terminated_length": 615.3424682617188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 7.401166180758017, + "grad_norm": 0.11937696486711502, + "learning_rate": 1e-06, + "loss": -0.0386, + "num_tokens": 479270775.0, + "reward": 0.6674107313156128, + "reward_std": 0.10197984427213669, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 792 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 1157.8739013671875, + "completions/mean_terminated_length": 627.5401611328125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 7.410495626822158, + "grad_norm": 0.1595282107591629, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 479841118.0, + "reward": 0.5837053656578064, + "reward_std": 0.1640053242444992, + "rewards/verify_math_reward/mean": 0.5837053656578064, + "rewards/verify_math_reward/std": 0.49321895837783813, + "step": 793 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3319.0, + "completions/mean_length": 955.9989013671875, + "completions/mean_terminated_length": 652.3757934570312, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 7.419825072886297, + "grad_norm": 0.14234353601932526, + "learning_rate": 1e-06, + "loss": -0.0148, + "num_tokens": 480459845.0, + "reward": 0.6640625, + "reward_std": 0.15097682178020477, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 794 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3701.0, + "completions/mean_length": 878.4453735351562, + "completions/mean_terminated_length": 571.636962890625, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.429154518950437, + "grad_norm": 0.1421203911304474, + "learning_rate": 1e-06, + "loss": -0.0547, + "num_tokens": 481013668.0, + "reward": 0.660714328289032, + "reward_std": 0.12639741599559784, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 795 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2733.0, + "completions/mean_length": 976.62841796875, + "completions/mean_terminated_length": 611.0162353515625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.438483965014577, + "grad_norm": 0.1589864194393158, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 481585231.0, + "reward": 0.6774553656578064, + "reward_std": 0.18196998536586761, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 796 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2672.0, + "completions/mean_length": 912.03466796875, + "completions/mean_terminated_length": 604.1603393554688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.447813411078717, + "grad_norm": 0.1368117779493332, + "learning_rate": 1e-06, + "loss": -0.0614, + "num_tokens": 482161718.0, + "reward": 0.6986607313156128, + "reward_std": 0.14304757118225098, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 797 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3448.0, + "completions/mean_length": 939.6920166015625, + "completions/mean_terminated_length": 604.5778198242188, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 7.457142857142857, + "grad_norm": 0.15140080451965332, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 482736066.0, + "reward": 0.645089328289032, + "reward_std": 0.1773899644613266, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 798 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3427.0, + "completions/mean_length": 1114.0413818359375, + "completions/mean_terminated_length": 630.5849609375, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.466472303206997, + "grad_norm": 0.13179807364940643, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 483319215.0, + "reward": 0.5870535969734192, + "reward_std": 0.13891443610191345, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263834953308105, + "step": 799 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 946.8973388671875, + "completions/mean_terminated_length": 590.9118041992188, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 7.475801749271137, + "grad_norm": 0.17107652127742767, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 483882635.0, + "reward": 0.668526828289032, + "reward_std": 0.19941289722919464, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 800 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3279.0, + "completions/mean_length": 1010.6574096679688, + "completions/mean_terminated_length": 670.3903198242188, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 7.485131195335277, + "grad_norm": 0.16185325384140015, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 484522256.0, + "reward": 0.5915178656578064, + "reward_std": 0.17577597498893738, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 801 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3480.0, + "completions/mean_length": 1066.9140625, + "completions/mean_terminated_length": 647.3837280273438, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 7.494460641399417, + "grad_norm": 0.15170329809188843, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 485129419.0, + "reward": 0.6227678656578064, + "reward_std": 0.14835071563720703, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 802 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2649.0, + "completions/mean_length": 894.8594360351562, + "completions/mean_terminated_length": 623.5762939453125, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 7.503790087463557, + "grad_norm": 0.15010656416416168, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 485730149.0, + "reward": 0.7142857313156128, + "reward_std": 0.16740237176418304, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 803 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2709.0, + "completions/mean_length": 890.7076416015625, + "completions/mean_terminated_length": 593.6317138671875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.513119533527696, + "grad_norm": 0.15986177325248718, + "learning_rate": 1e-06, + "loss": -0.0484, + "num_tokens": 486300647.0, + "reward": 0.676339328289032, + "reward_std": 0.16661594808101654, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 804 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1450892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3606.0, + "completions/mean_length": 1117.96435546875, + "completions/mean_terminated_length": 612.5535278320312, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 7.522448979591837, + "grad_norm": 0.15698200464248657, + "learning_rate": 1e-06, + "loss": -0.0648, + "num_tokens": 486865943.0, + "reward": 0.645089328289032, + "reward_std": 0.1582231968641281, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 805 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3951.0, + "completions/mean_length": 948.5022583007812, + "completions/mean_terminated_length": 618.6165161132812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 7.531778425655976, + "grad_norm": 0.14912272989749908, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 487468873.0, + "reward": 0.6573660969734192, + "reward_std": 0.14699704945087433, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485536336898804, + "step": 806 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3478.0, + "completions/mean_length": 1049.2623291015625, + "completions/mean_terminated_length": 569.0271606445312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.541107871720117, + "grad_norm": 0.16505460441112518, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 488005620.0, + "reward": 0.6484375, + "reward_std": 0.15401659905910492, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 807 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3173.0, + "completions/mean_length": 1151.89404296875, + "completions/mean_terminated_length": 692.2335815429688, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.550437317784256, + "grad_norm": 0.13680331408977509, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 488626677.0, + "reward": 0.606026828289032, + "reward_std": 0.16442997753620148, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 808 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2841.0, + "completions/mean_length": 1084.669677734375, + "completions/mean_terminated_length": 654.4795532226562, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 7.559766763848397, + "grad_norm": 0.14948244392871857, + "learning_rate": 1e-06, + "loss": -0.0506, + "num_tokens": 489246829.0, + "reward": 0.578125, + "reward_std": 0.15240715444087982, + "rewards/verify_math_reward/mean": 0.578125, + "rewards/verify_math_reward/std": 0.4941346049308777, + "step": 809 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3462.0, + "completions/mean_length": 1108.203125, + "completions/mean_terminated_length": 637.2584228515625, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 7.569096209912536, + "grad_norm": 0.16127030551433563, + "learning_rate": 1e-06, + "loss": -0.076, + "num_tokens": 489815867.0, + "reward": 0.640625, + "reward_std": 0.18644899129867554, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 810 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2358.0, + "completions/mean_length": 928.7053833007812, + "completions/mean_terminated_length": 596.7447509765625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 7.578425655976677, + "grad_norm": 0.16200365126132965, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 490377411.0, + "reward": 0.6495535969734192, + "reward_std": 0.1679650843143463, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 811 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3646.0, + "completions/mean_length": 956.8906860351562, + "completions/mean_terminated_length": 649.1348266601562, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 7.587755102040816, + "grad_norm": 0.12454048544168472, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 490991617.0, + "reward": 0.606026828289032, + "reward_std": 0.14545612037181854, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 812 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3928.0, + "completions/mean_length": 1027.6195068359375, + "completions/mean_terminated_length": 624.7007446289062, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 7.597084548104956, + "grad_norm": 0.14836883544921875, + "learning_rate": 1e-06, + "loss": -0.0584, + "num_tokens": 491577972.0, + "reward": 0.5926339626312256, + "reward_std": 0.18663373589515686, + "rewards/verify_math_reward/mean": 0.5926339030265808, + "rewards/verify_math_reward/std": 0.49161845445632935, + "step": 813 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2695.0, + "completions/mean_length": 1121.5546875, + "completions/mean_terminated_length": 621.2894287109375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 7.606413994169096, + "grad_norm": 0.18633154034614563, + "learning_rate": 1e-06, + "loss": -0.0179, + "num_tokens": 492141813.0, + "reward": 0.609375, + "reward_std": 0.17071063816547394, + "rewards/verify_math_reward/mean": 0.609375, + "rewards/verify_math_reward/std": 0.48816296458244324, + "step": 814 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 956.44873046875, + "completions/mean_terminated_length": 627.3958129882812, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 7.615743440233236, + "grad_norm": 0.16072827577590942, + "learning_rate": 1e-06, + "loss": -0.0588, + "num_tokens": 492739071.0, + "reward": 0.6774553656578064, + "reward_std": 0.18648287653923035, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 815 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3915.0, + "completions/mean_length": 886.0926513671875, + "completions/mean_terminated_length": 614.0665893554688, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.625072886297376, + "grad_norm": 0.15345019102096558, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 493325890.0, + "reward": 0.6741071939468384, + "reward_std": 0.164046972990036, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 816 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3502.0, + "completions/mean_length": 981.4688110351562, + "completions/mean_terminated_length": 620.7571411132812, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 7.634402332361516, + "grad_norm": 0.1588490605354309, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 493919134.0, + "reward": 0.676339328289032, + "reward_std": 0.18077494204044342, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 817 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3265.0, + "completions/mean_length": 918.2210083007812, + "completions/mean_terminated_length": 648.917724609375, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 7.643731778425656, + "grad_norm": 0.16136573255062103, + "learning_rate": 1e-06, + "loss": -0.0724, + "num_tokens": 494541580.0, + "reward": 0.668526828289032, + "reward_std": 0.19088450074195862, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 818 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3742.0, + "completions/mean_length": 996.8616333007812, + "completions/mean_terminated_length": 611.8995971679688, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 7.653061224489796, + "grad_norm": 0.15670543909072876, + "learning_rate": 1e-06, + "loss": -0.0598, + "num_tokens": 495118752.0, + "reward": 0.637276828289032, + "reward_std": 0.17498251795768738, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 819 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3475.0, + "completions/mean_length": 1005.9107666015625, + "completions/mean_terminated_length": 595.7218627929688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 7.662390670553936, + "grad_norm": 0.17109698057174683, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 495677152.0, + "reward": 0.6729910969734192, + "reward_std": 0.17333421111106873, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 820 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 954.2801513671875, + "completions/mean_terminated_length": 590.419677734375, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 7.671720116618076, + "grad_norm": 0.14786724746227264, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 496238051.0, + "reward": 0.699776828289032, + "reward_std": 0.14669284224510193, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 821 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 866.919677734375, + "completions/mean_terminated_length": 571.9367065429688, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 7.681049562682215, + "grad_norm": 0.14048032462596893, + "learning_rate": 1e-06, + "loss": -0.0696, + "num_tokens": 496801675.0, + "reward": 0.6986607313156128, + "reward_std": 0.1385781466960907, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 822 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3929.0, + "completions/mean_length": 1072.91748046875, + "completions/mean_terminated_length": 654.218505859375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 7.690379008746356, + "grad_norm": 0.15230326354503632, + "learning_rate": 1e-06, + "loss": -0.0728, + "num_tokens": 497405929.0, + "reward": 0.6104910969734192, + "reward_std": 0.16943003237247467, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 823 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 959.2701416015625, + "completions/mean_terminated_length": 569.6386108398438, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.699708454810495, + "grad_norm": 0.14827117323875427, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 497951419.0, + "reward": 0.6774553656578064, + "reward_std": 0.1420711725950241, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 824 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3848.0, + "completions/mean_length": 955.5826416015625, + "completions/mean_terminated_length": 617.861572265625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 7.709037900874636, + "grad_norm": 0.14377804100513458, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 498546669.0, + "reward": 0.6484375, + "reward_std": 0.17081871628761292, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 825 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3461.0, + "completions/mean_length": 1082.0413818359375, + "completions/mean_terminated_length": 633.8115844726562, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 7.718367346938775, + "grad_norm": 0.16568532586097717, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 499140690.0, + "reward": 0.6707589626312256, + "reward_std": 0.18674391508102417, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 826 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3760.0, + "completions/mean_length": 1054.2801513671875, + "completions/mean_terminated_length": 592.9395751953125, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.727696793002916, + "grad_norm": 0.13595949113368988, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 499691653.0, + "reward": 0.6305803656578064, + "reward_std": 0.12351170182228088, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 827 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3704.0, + "completions/mean_length": 943.8839721679688, + "completions/mean_terminated_length": 604.90478515625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 7.737026239067055, + "grad_norm": 0.16388928890228271, + "learning_rate": 1e-06, + "loss": -0.0605, + "num_tokens": 500274757.0, + "reward": 0.652901828289032, + "reward_std": 0.15686701238155365, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 828 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2996.0, + "completions/mean_length": 952.0256958007812, + "completions/mean_terminated_length": 600.9615478515625, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 7.746355685131196, + "grad_norm": 0.16453956067562103, + "learning_rate": 1e-06, + "loss": -0.0666, + "num_tokens": 500847628.0, + "reward": 0.6383928656578064, + "reward_std": 0.180777445435524, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 829 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2568.0, + "completions/mean_length": 947.3672485351562, + "completions/mean_terminated_length": 600.1202392578125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 7.755685131195335, + "grad_norm": 0.13221710920333862, + "learning_rate": 1e-06, + "loss": -0.0525, + "num_tokens": 501424757.0, + "reward": 0.6718750596046448, + "reward_std": 0.1544366478919983, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 830 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2947.0, + "completions/mean_length": 926.2813110351562, + "completions/mean_terminated_length": 585.4091186523438, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 7.765014577259475, + "grad_norm": 0.15517419576644897, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 501984689.0, + "reward": 0.6752232313156128, + "reward_std": 0.15251775085926056, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 831 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3754.0, + "completions/mean_length": 1088.4866943359375, + "completions/mean_terminated_length": 650.0511474609375, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.774344023323615, + "grad_norm": 0.14194026589393616, + "learning_rate": 1e-06, + "loss": -0.073, + "num_tokens": 502577317.0, + "reward": 0.6651785969734192, + "reward_std": 0.16360554099082947, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 832 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3930.0, + "completions/mean_length": 881.419677734375, + "completions/mean_terminated_length": 592.0291748046875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 7.783673469387755, + "grad_norm": 0.14261199533939362, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 503153469.0, + "reward": 0.6551339626312256, + "reward_std": 0.1510535031557083, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 833 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2533.0, + "completions/mean_length": 853.0892944335938, + "completions/mean_terminated_length": 582.5199584960938, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 7.793002915451895, + "grad_norm": 0.14982765913009644, + "learning_rate": 1e-06, + "loss": -0.0257, + "num_tokens": 503732093.0, + "reward": 0.6830357313156128, + "reward_std": 0.1487216204404831, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 834 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2622.0, + "completions/mean_length": 928.6406860351562, + "completions/mean_terminated_length": 605.2816772460938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 7.802332361516035, + "grad_norm": 0.14122828841209412, + "learning_rate": 1e-06, + "loss": -0.0432, + "num_tokens": 504311267.0, + "reward": 0.6729910969734192, + "reward_std": 0.13437014818191528, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 835 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3527.0, + "completions/mean_length": 897.83935546875, + "completions/mean_terminated_length": 639.3630981445312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.811661807580175, + "grad_norm": 0.14737960696220398, + "learning_rate": 1e-06, + "loss": -0.0332, + "num_tokens": 504925507.0, + "reward": 0.6908482313156128, + "reward_std": 0.17051450908184052, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 836 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3887.0, + "completions/mean_length": 1123.3671875, + "completions/mean_terminated_length": 698.705322265625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 7.820991253644315, + "grad_norm": 0.16875937581062317, + "learning_rate": 1e-06, + "loss": -0.0858, + "num_tokens": 505573692.0, + "reward": 0.6328125, + "reward_std": 0.1921621859073639, + "rewards/verify_math_reward/mean": 0.6328125, + "rewards/verify_math_reward/std": 0.48230743408203125, + "step": 837 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4075.0, + "completions/mean_length": 1088.829345703125, + "completions/mean_terminated_length": 654.8441772460938, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 7.830320699708455, + "grad_norm": 0.14596500992774963, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 506188211.0, + "reward": 0.598214328289032, + "reward_std": 0.1403786838054657, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053287506103516, + "step": 838 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3951.0, + "completions/mean_length": 911.4944458007812, + "completions/mean_terminated_length": 594.99755859375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 7.839650145772595, + "grad_norm": 0.16681191325187683, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 506752766.0, + "reward": 0.7064732313156128, + "reward_std": 0.17607979476451874, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 839 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3154.0, + "completions/mean_length": 891.169677734375, + "completions/mean_terminated_length": 585.5745849609375, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 7.848979591836734, + "grad_norm": 0.1402122974395752, + "learning_rate": 1e-06, + "loss": -0.0465, + "num_tokens": 507311742.0, + "reward": 0.6595982313156128, + "reward_std": 0.13185353577136993, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 840 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2925.0, + "completions/mean_length": 894.12060546875, + "completions/mean_terminated_length": 601.6224365234375, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 7.858309037900875, + "grad_norm": 0.16580727696418762, + "learning_rate": 1e-06, + "loss": -0.0348, + "num_tokens": 507900634.0, + "reward": 0.7053571939468384, + "reward_std": 0.1437968611717224, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 841 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2631.0, + "completions/mean_length": 980.0725708007812, + "completions/mean_terminated_length": 614.8641357421875, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 7.867638483965014, + "grad_norm": 0.16850878298282623, + "learning_rate": 1e-06, + "loss": -0.0354, + "num_tokens": 508474059.0, + "reward": 0.6629464626312256, + "reward_std": 0.1270802915096283, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 842 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3808.0, + "completions/mean_length": 956.6886596679688, + "completions/mean_terminated_length": 627.660888671875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 7.876967930029155, + "grad_norm": 0.15846849977970123, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 509081212.0, + "reward": 0.6819196939468384, + "reward_std": 0.1423000991344452, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 843 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2426.0, + "completions/mean_length": 929.9810791015625, + "completions/mean_terminated_length": 611.0454711914062, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 7.886297376093294, + "grad_norm": 0.1680334061384201, + "learning_rate": 1e-06, + "loss": -0.0267, + "num_tokens": 509670635.0, + "reward": 0.6875000596046448, + "reward_std": 0.16912764310836792, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 844 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 939.5313110351562, + "completions/mean_terminated_length": 587.0719604492188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 7.895626822157435, + "grad_norm": 0.17341989278793335, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 510231503.0, + "reward": 0.660714328289032, + "reward_std": 0.16634997725486755, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 845 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3538.0, + "completions/mean_length": 1057.036865234375, + "completions/mean_terminated_length": 675.2575073242188, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 7.904956268221574, + "grad_norm": 0.14785552024841309, + "learning_rate": 1e-06, + "loss": -0.0637, + "num_tokens": 510865960.0, + "reward": 0.6417410969734192, + "reward_std": 0.1853192150592804, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 846 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3074.0, + "completions/mean_length": 885.6998291015625, + "completions/mean_terminated_length": 609.4193725585938, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 7.914285714285715, + "grad_norm": 0.15614314377307892, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 511449739.0, + "reward": 0.6930803656578064, + "reward_std": 0.17389945685863495, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 847 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2769.0, + "completions/mean_length": 889.2131958007812, + "completions/mean_terminated_length": 625.8538818359375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 7.923615160349854, + "grad_norm": 0.1569957733154297, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 512060186.0, + "reward": 0.6819196939468384, + "reward_std": 0.18186120688915253, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 848 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3231.0, + "completions/mean_length": 1074.532470703125, + "completions/mean_terminated_length": 620.73046875, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 7.932944606413994, + "grad_norm": 0.17628583312034607, + "learning_rate": 1e-06, + "loss": -0.0748, + "num_tokens": 512641135.0, + "reward": 0.6339285969734192, + "reward_std": 0.1797250658273697, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 849 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 1061.40625, + "completions/mean_terminated_length": 641.113037109375, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 7.942274052478134, + "grad_norm": 0.13701649010181427, + "learning_rate": 1e-06, + "loss": -0.0163, + "num_tokens": 513232811.0, + "reward": 0.6361607313156128, + "reward_std": 0.13850146532058716, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 850 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2761.0, + "completions/mean_length": 1012.3660888671875, + "completions/mean_terminated_length": 624.974853515625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 7.9516034985422746, + "grad_norm": 0.13524264097213745, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 513826019.0, + "reward": 0.625, + "reward_std": 0.14522789418697357, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 851 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3675.0, + "completions/mean_length": 1089.794677734375, + "completions/mean_terminated_length": 695.0404052734375, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 7.960932944606414, + "grad_norm": 0.1434006541967392, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 514472227.0, + "reward": 0.625, + "reward_std": 0.14579172432422638, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 852 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3150.0, + "completions/mean_length": 914.6339721679688, + "completions/mean_terminated_length": 619.775634765625, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 7.970262390670554, + "grad_norm": 0.15225620567798615, + "learning_rate": 1e-06, + "loss": -0.0328, + "num_tokens": 515071659.0, + "reward": 0.6551339626312256, + "reward_std": 0.16311517357826233, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 853 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3347.0, + "completions/mean_length": 840.1975708007812, + "completions/mean_terminated_length": 577.062744140625, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 7.979591836734694, + "grad_norm": 0.14436179399490356, + "learning_rate": 1e-06, + "loss": -0.0283, + "num_tokens": 515630284.0, + "reward": 0.6741071939468384, + "reward_std": 0.12388966977596283, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 854 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4005.0, + "completions/mean_length": 1140.739990234375, + "completions/mean_terminated_length": 657.1519165039062, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 7.988921282798834, + "grad_norm": 0.15935580432415009, + "learning_rate": 1e-06, + "loss": -0.0915, + "num_tokens": 516236259.0, + "reward": 0.6037946939468384, + "reward_std": 0.18130375444889069, + "rewards/verify_math_reward/mean": 0.6037946343421936, + "rewards/verify_math_reward/std": 0.48938122391700745, + "step": 855 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.12215909090909094, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3603.0, + "completions/mean_length": 1079.3494873046875, + "completions/mean_terminated_length": 659.556640625, + "completions/min_length": 197.0, + "completions/min_terminated_length": 197.0, + "epoch": 7.998250728862974, + "grad_norm": 0.13951878249645233, + "learning_rate": 1e-06, + "loss": -0.0402, + "num_tokens": 516816035.0, + "reward": 0.6897321939468384, + "reward_std": 0.12268754839897156, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 856 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3059.0, + "completions/mean_length": 1011.02685546875, + "completions/mean_terminated_length": 666.5508422851562, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.00932944606414, + "grad_norm": 0.13932685554027557, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 517458219.0, + "reward": 0.6127232313156128, + "reward_std": 0.16044770181179047, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 857 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3908.0, + "completions/mean_length": 890.1239013671875, + "completions/mean_terminated_length": 571.5030517578125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 8.018658892128279, + "grad_norm": 0.17174571752548218, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 518001610.0, + "reward": 0.7008928656578064, + "reward_std": 0.1770561784505844, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 858 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3267.0, + "completions/mean_length": 1002.9766235351562, + "completions/mean_terminated_length": 631.813720703125, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 8.02798833819242, + "grad_norm": 0.1499703824520111, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 518598589.0, + "reward": 0.6785714626312256, + "reward_std": 0.18280188739299774, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 859 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2968.0, + "completions/mean_length": 1118.2779541015625, + "completions/mean_terminated_length": 666.6439819335938, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 8.03731778425656, + "grad_norm": 0.14569251239299774, + "learning_rate": 1e-06, + "loss": -0.0815, + "num_tokens": 519218070.0, + "reward": 0.5725446939468384, + "reward_std": 0.15518662333488464, + "rewards/verify_math_reward/mean": 0.5725446343421936, + "rewards/verify_math_reward/std": 0.49498558044433594, + "step": 860 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 1133.5101318359375, + "completions/mean_terminated_length": 731.7528686523438, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 8.0466472303207, + "grad_norm": 0.13986298441886902, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 519879759.0, + "reward": 0.6383928656578064, + "reward_std": 0.16961227357387543, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 861 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3120.0, + "completions/mean_length": 964.02685546875, + "completions/mean_terminated_length": 618.6171264648438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.055976676384839, + "grad_norm": 0.17931996285915375, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 520466519.0, + "reward": 0.6819196939468384, + "reward_std": 0.16616523265838623, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 862 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3990.0, + "completions/mean_length": 1042.2757568359375, + "completions/mean_terminated_length": 662.954833984375, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.06530612244898, + "grad_norm": 0.12542754411697388, + "learning_rate": 1e-06, + "loss": -0.0538, + "num_tokens": 521082614.0, + "reward": 0.6584821939468384, + "reward_std": 0.13590925931930542, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 863 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2792.0, + "completions/mean_length": 937.2422485351562, + "completions/mean_terminated_length": 597.548828125, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 8.07463556851312, + "grad_norm": 0.13399145007133484, + "learning_rate": 1e-06, + "loss": -0.0628, + "num_tokens": 521657367.0, + "reward": 0.6540178656578064, + "reward_std": 0.15465529263019562, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 864 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2475.0, + "completions/mean_length": 808.9207763671875, + "completions/mean_terminated_length": 593.9512329101562, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.08396501457726, + "grad_norm": 0.1413942575454712, + "learning_rate": 1e-06, + "loss": -0.0339, + "num_tokens": 522247872.0, + "reward": 0.6774553656578064, + "reward_std": 0.14496897161006927, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 865 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2676.0, + "completions/mean_length": 923.2199096679688, + "completions/mean_terminated_length": 607.8883666992188, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 8.093294460641399, + "grad_norm": 0.12779447436332703, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 522831477.0, + "reward": 0.6774553656578064, + "reward_std": 0.12422778457403183, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 866 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3592.0, + "completions/mean_length": 1027.513427734375, + "completions/mean_terminated_length": 637.6804809570312, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 8.102623906705539, + "grad_norm": 0.15007537603378296, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 523429225.0, + "reward": 0.6417410969734192, + "reward_std": 0.15033742785453796, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 867 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4036.0, + "completions/mean_length": 1025.3828125, + "completions/mean_terminated_length": 674.0186157226562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 8.11195335276968, + "grad_norm": 0.17137961089611053, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 524068352.0, + "reward": 0.6395089626312256, + "reward_std": 0.1593964546918869, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 868 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2846.0, + "completions/mean_length": 963.9844360351562, + "completions/mean_terminated_length": 605.5945434570312, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 8.12128279883382, + "grad_norm": 0.16211485862731934, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 524637394.0, + "reward": 0.652901828289032, + "reward_std": 0.16142450273036957, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 869 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3241.0, + "completions/mean_length": 942.6730346679688, + "completions/mean_terminated_length": 581.8445434570312, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 8.130612244897959, + "grad_norm": 0.15985849499702454, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 525182429.0, + "reward": 0.7053571939468384, + "reward_std": 0.18054740130901337, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 870 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2955.0, + "completions/mean_length": 986.5256958007812, + "completions/mean_terminated_length": 643.5972900390625, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 8.139941690962099, + "grad_norm": 0.14217232167720795, + "learning_rate": 1e-06, + "loss": -0.0844, + "num_tokens": 525786812.0, + "reward": 0.6908482313156128, + "reward_std": 0.16668446362018585, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 871 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2696.0, + "completions/mean_length": 1080.0826416015625, + "completions/mean_terminated_length": 600.1889038085938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.14927113702624, + "grad_norm": 0.1697220504283905, + "learning_rate": 1e-06, + "loss": -0.0738, + "num_tokens": 526344566.0, + "reward": 0.6473214626312256, + "reward_std": 0.1651107668876648, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 872 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 856.2433471679688, + "completions/mean_terminated_length": 611.2196655273438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 8.15860058309038, + "grad_norm": 0.1487022340297699, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 526938784.0, + "reward": 0.7354910969734192, + "reward_std": 0.14666074514389038, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 873 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3888.0, + "completions/mean_length": 1024.505615234375, + "completions/mean_terminated_length": 625.5598754882812, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 8.167930029154519, + "grad_norm": 0.15943771600723267, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 527521397.0, + "reward": 0.6573660969734192, + "reward_std": 0.1556737869977951, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 874 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3264.0, + "completions/mean_length": 1122.5279541015625, + "completions/mean_terminated_length": 689.0549926757812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 8.177259475218658, + "grad_norm": 0.15239205956459045, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 528162518.0, + "reward": 0.5892857313156128, + "reward_std": 0.15913361310958862, + "rewards/verify_math_reward/mean": 0.5892857313156128, + "rewards/verify_math_reward/std": 0.49223825335502625, + "step": 875 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4073.0, + "completions/mean_length": 892.51123046875, + "completions/mean_terminated_length": 629.4226684570312, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.186588921282798, + "grad_norm": 0.13224801421165466, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 528766608.0, + "reward": 0.7209821939468384, + "reward_std": 0.1366906315088272, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 876 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 976.4375610351562, + "completions/mean_terminated_length": 615.1431884765625, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 8.19591836734694, + "grad_norm": 0.1724257469177246, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 529353984.0, + "reward": 0.699776828289032, + "reward_std": 0.19708172976970673, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 877 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3157.0, + "completions/mean_length": 933.1785888671875, + "completions/mean_terminated_length": 618.8367919921875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 8.205247813411079, + "grad_norm": 0.14418606460094452, + "learning_rate": 1e-06, + "loss": -0.032, + "num_tokens": 529939464.0, + "reward": 0.6339285969734192, + "reward_std": 0.14842739701271057, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 878 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 1099.208740234375, + "completions/mean_terminated_length": 701.4046020507812, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 8.214577259475218, + "grad_norm": 0.13870897889137268, + "learning_rate": 1e-06, + "loss": -0.0283, + "num_tokens": 530593155.0, + "reward": 0.5502232313156128, + "reward_std": 0.14522859454154968, + "rewards/verify_math_reward/mean": 0.5502232313156128, + "rewards/verify_math_reward/std": 0.49774909019470215, + "step": 879 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1895.0, + "completions/mean_length": 900.935302734375, + "completions/mean_terminated_length": 561.7061767578125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.223906705539358, + "grad_norm": 0.1580512821674347, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 531126513.0, + "reward": 0.7053571939468384, + "reward_std": 0.13650910556316376, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 880 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3835.0, + "completions/mean_length": 1087.5457763671875, + "completions/mean_terminated_length": 713.84814453125, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 8.2332361516035, + "grad_norm": 0.13706760108470917, + "learning_rate": 1e-06, + "loss": -0.0669, + "num_tokens": 531777570.0, + "reward": 0.6674107313156128, + "reward_std": 0.16777077317237854, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 881 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3957.0, + "completions/mean_length": 1109.6171875, + "completions/mean_terminated_length": 700.3159790039062, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 8.242565597667639, + "grad_norm": 0.14511780440807343, + "learning_rate": 1e-06, + "loss": -0.0545, + "num_tokens": 532425203.0, + "reward": 0.606026828289032, + "reward_std": 0.16134671866893768, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 882 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2641.0, + "completions/mean_length": 959.849365234375, + "completions/mean_terminated_length": 669.1817016601562, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 8.251895043731778, + "grad_norm": 0.14673292636871338, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 533065284.0, + "reward": 0.598214328289032, + "reward_std": 0.1599598526954651, + "rewards/verify_math_reward/mean": 0.5982142686843872, + "rewards/verify_math_reward/std": 0.49053290486335754, + "step": 883 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3967.0, + "completions/mean_length": 970.5748291015625, + "completions/mean_terminated_length": 577.9334106445312, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 8.261224489795918, + "grad_norm": 0.15665669739246368, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 533615551.0, + "reward": 0.6930803656578064, + "reward_std": 0.12633030116558075, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147334575653076, + "step": 884 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3463.0, + "completions/mean_length": 1066.8382568359375, + "completions/mean_terminated_length": 602.9124755859375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.270553935860057, + "grad_norm": 0.15283460915088654, + "learning_rate": 1e-06, + "loss": -0.0633, + "num_tokens": 534173054.0, + "reward": 0.6718750596046448, + "reward_std": 0.146736741065979, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 885 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3185.0, + "completions/mean_length": 1029.6429443359375, + "completions/mean_terminated_length": 653.0726928710938, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.279883381924199, + "grad_norm": 0.1382666528224945, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 534788510.0, + "reward": 0.6674107313156128, + "reward_std": 0.14222271740436554, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 886 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2515.0, + "completions/mean_length": 974.20654296875, + "completions/mean_terminated_length": 621.3080444335938, + "completions/min_length": 195.0, + "completions/min_terminated_length": 195.0, + "epoch": 8.289212827988338, + "grad_norm": 0.14796394109725952, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 535388391.0, + "reward": 0.6462053656578064, + "reward_std": 0.15974049270153046, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 887 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2782.0, + "completions/mean_length": 965.3850708007812, + "completions/mean_terminated_length": 637.268798828125, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 8.298542274052478, + "grad_norm": 0.14198701083660126, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 535992056.0, + "reward": 0.6863839626312256, + "reward_std": 0.14628097414970398, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.4642214775085449, + "step": 888 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3928.0, + "completions/mean_length": 1059.2265625, + "completions/mean_terminated_length": 651.7607421875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.307871720116617, + "grad_norm": 0.15404768288135529, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 536582691.0, + "reward": 0.6473214626312256, + "reward_std": 0.17333492636680603, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807058691978455, + "step": 889 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3369.0, + "completions/mean_length": 1057.2890625, + "completions/mean_terminated_length": 662.6015014648438, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 8.317201166180759, + "grad_norm": 0.14827096462249756, + "learning_rate": 1e-06, + "loss": -0.0752, + "num_tokens": 537195238.0, + "reward": 0.65625, + "reward_std": 0.1644253432750702, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 890 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3633.0, + "completions/mean_length": 1000.193115234375, + "completions/mean_terminated_length": 628.6962280273438, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 8.326530612244898, + "grad_norm": 0.1479225754737854, + "learning_rate": 1e-06, + "loss": -0.0847, + "num_tokens": 537789555.0, + "reward": 0.6540178656578064, + "reward_std": 0.16781170666217804, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 891 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 980.4531860351562, + "completions/mean_terminated_length": 602.2202758789062, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.335860058309038, + "grad_norm": 0.1485033631324768, + "learning_rate": 1e-06, + "loss": -0.0288, + "num_tokens": 538373777.0, + "reward": 0.6439732313156128, + "reward_std": 0.13624738156795502, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 892 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2951.0, + "completions/mean_length": 966.2467041015625, + "completions/mean_terminated_length": 590.6762084960938, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 8.345189504373177, + "grad_norm": 0.16735050082206726, + "learning_rate": 1e-06, + "loss": -0.0834, + "num_tokens": 538934814.0, + "reward": 0.6930803656578064, + "reward_std": 0.20016105473041534, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 893 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3791.0, + "completions/mean_length": 1043.130615234375, + "completions/mean_terminated_length": 650.9483642578125, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 8.354518950437317, + "grad_norm": 0.1467769593000412, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 539537531.0, + "reward": 0.6674107313156128, + "reward_std": 0.17148950695991516, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 894 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3525.0, + "completions/mean_length": 1145.438720703125, + "completions/mean_terminated_length": 671.5142211914062, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.363848396501458, + "grad_norm": 0.14385691285133362, + "learning_rate": 1e-06, + "loss": -0.0524, + "num_tokens": 540146740.0, + "reward": 0.6082589626312256, + "reward_std": 0.13940230011940002, + "rewards/verify_math_reward/mean": 0.6082589030265808, + "rewards/verify_math_reward/std": 0.48841196298599243, + "step": 895 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3657.0, + "completions/mean_length": 1022.89404296875, + "completions/mean_terminated_length": 662.7045288085938, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.373177842565598, + "grad_norm": 0.16795089840888977, + "learning_rate": 1e-06, + "loss": -0.0299, + "num_tokens": 540773573.0, + "reward": 0.6428571939468384, + "reward_std": 0.15405797958374023, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 896 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2947.0, + "completions/mean_length": 948.6451416015625, + "completions/mean_terminated_length": 623.056640625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.382507288629737, + "grad_norm": 0.18909591436386108, + "learning_rate": 1e-06, + "loss": -0.0576, + "num_tokens": 541372095.0, + "reward": 0.6953125596046448, + "reward_std": 0.1716078370809555, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 897 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4077.0, + "completions/mean_length": 1079.368408203125, + "completions/mean_terminated_length": 687.5485229492188, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 8.391836734693877, + "grad_norm": 0.1608656495809555, + "learning_rate": 1e-06, + "loss": -0.0366, + "num_tokens": 542007641.0, + "reward": 0.6049107313156128, + "reward_std": 0.16751113533973694, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 898 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3159.0, + "completions/mean_length": 955.7835083007812, + "completions/mean_terminated_length": 565.7189331054688, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.401166180758018, + "grad_norm": 0.15515096485614777, + "learning_rate": 1e-06, + "loss": -0.0765, + "num_tokens": 542550167.0, + "reward": 0.7220982313156128, + "reward_std": 0.15199598670005798, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821488857269287, + "step": 899 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 1041.946533203125, + "completions/mean_terminated_length": 618.9580688476562, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 8.410495626822158, + "grad_norm": 0.15280771255493164, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 543124791.0, + "reward": 0.660714328289032, + "reward_std": 0.15428578853607178, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 900 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2845.0, + "completions/mean_length": 1075.680908203125, + "completions/mean_terminated_length": 657.3646240234375, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 8.419825072886297, + "grad_norm": 0.13621953129768372, + "learning_rate": 1e-06, + "loss": -0.0749, + "num_tokens": 543726569.0, + "reward": 0.6484375, + "reward_std": 0.1771642565727234, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 901 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 1058.76123046875, + "completions/mean_terminated_length": 659.9318237304688, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 8.429154518950437, + "grad_norm": 0.15349024534225464, + "learning_rate": 1e-06, + "loss": -0.0392, + "num_tokens": 544336787.0, + "reward": 0.6506696939468384, + "reward_std": 0.15018658339977264, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 902 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3523.0, + "completions/mean_length": 1085.1942138671875, + "completions/mean_terminated_length": 655.0790405273438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 8.438483965014576, + "grad_norm": 0.14632688462734222, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 544941233.0, + "reward": 0.6339285969734192, + "reward_std": 0.16833347082138062, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 903 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2572.0, + "completions/mean_length": 985.5234985351562, + "completions/mean_terminated_length": 620.953857421875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 8.447813411078718, + "grad_norm": 0.14520664513111115, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 545524070.0, + "reward": 0.6863839626312256, + "reward_std": 0.14053022861480713, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 904 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2956.0, + "completions/mean_length": 1054.626220703125, + "completions/mean_terminated_length": 655.2537841796875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 8.457142857142857, + "grad_norm": 0.16137388348579407, + "learning_rate": 1e-06, + "loss": -0.0627, + "num_tokens": 546131503.0, + "reward": 0.6495535969734192, + "reward_std": 0.173787459731102, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 905 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3783.0, + "completions/mean_length": 922.6082763671875, + "completions/mean_terminated_length": 620.0110473632812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.466472303206997, + "grad_norm": 0.13741783797740936, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 546728960.0, + "reward": 0.6975446939468384, + "reward_std": 0.13437014818191528, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 906 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3915.0, + "completions/mean_length": 994.5513916015625, + "completions/mean_terminated_length": 613.6716918945312, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 8.475801749271136, + "grad_norm": 0.15759779512882233, + "learning_rate": 1e-06, + "loss": -0.0758, + "num_tokens": 547301062.0, + "reward": 0.7031250596046448, + "reward_std": 0.1476346254348755, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 907 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4003.0, + "completions/mean_length": 989.3359985351562, + "completions/mean_terminated_length": 603.4391479492188, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 8.485131195335278, + "grad_norm": 0.1323668658733368, + "learning_rate": 1e-06, + "loss": -0.0483, + "num_tokens": 547873571.0, + "reward": 0.6863839626312256, + "reward_std": 0.13962869346141815, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 908 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 1045.9788818359375, + "completions/mean_terminated_length": 623.5488891601562, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.494460641399417, + "grad_norm": 0.16331632435321808, + "learning_rate": 1e-06, + "loss": -0.087, + "num_tokens": 548451664.0, + "reward": 0.6540178656578064, + "reward_std": 0.1713361293077469, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 909 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3817.0, + "completions/mean_length": 1083.09716796875, + "completions/mean_terminated_length": 643.8759765625, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 8.503790087463557, + "grad_norm": 0.14266420900821686, + "learning_rate": 1e-06, + "loss": -0.042, + "num_tokens": 549048767.0, + "reward": 0.6417410969734192, + "reward_std": 0.1524856686592102, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 910 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3673.0, + "completions/mean_length": 1160.0, + "completions/mean_terminated_length": 634.6105346679688, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 8.513119533527696, + "grad_norm": 0.1584021896123886, + "learning_rate": 1e-06, + "loss": -0.0671, + "num_tokens": 549627743.0, + "reward": 0.6462053656578064, + "reward_std": 0.172020822763443, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 911 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3202.0, + "completions/mean_length": 1042.6976318359375, + "completions/mean_terminated_length": 619.8131713867188, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 8.522448979591836, + "grad_norm": 0.1526673585176468, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 550215288.0, + "reward": 0.6651785969734192, + "reward_std": 0.152593731880188, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 912 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3151.0, + "completions/mean_length": 919.583740234375, + "completions/mean_terminated_length": 564.8970336914062, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 8.531778425655977, + "grad_norm": 0.15874874591827393, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 550760323.0, + "reward": 0.7366071939468384, + "reward_std": 0.1521807461977005, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.4407196640968323, + "step": 913 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2391.0, + "completions/mean_length": 977.4375610351562, + "completions/mean_terminated_length": 576.8161010742188, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 8.541107871720117, + "grad_norm": 0.1279672384262085, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 551298499.0, + "reward": 0.6863839626312256, + "reward_std": 0.10761747509241104, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 914 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3814.0, + "completions/mean_length": 990.1819458007812, + "completions/mean_terminated_length": 639.0894165039062, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.550437317784256, + "grad_norm": 0.14278873801231384, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 551898102.0, + "reward": 0.6696428656578064, + "reward_std": 0.13508442044258118, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 915 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3437.0, + "completions/mean_length": 927.779052734375, + "completions/mean_terminated_length": 595.7213134765625, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 8.559766763848396, + "grad_norm": 0.13501763343811035, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 552465536.0, + "reward": 0.6808035969734192, + "reward_std": 0.12290509790182114, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 916 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1618303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3809.0, + "completions/mean_length": 1239.4007568359375, + "completions/mean_terminated_length": 687.8601684570312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 8.569096209912537, + "grad_norm": 0.16332502663135529, + "learning_rate": 1e-06, + "loss": -0.0709, + "num_tokens": 553058463.0, + "reward": 0.5948660969734192, + "reward_std": 0.1736719310283661, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 917 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3247.0, + "completions/mean_length": 901.2913208007812, + "completions/mean_terminated_length": 600.93408203125, + "completions/min_length": 184.0, + "completions/min_terminated_length": 184.0, + "epoch": 8.578425655976677, + "grad_norm": 0.16154848039150238, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 553636180.0, + "reward": 0.7165178656578064, + "reward_std": 0.1675432324409485, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 918 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 1032.0413818359375, + "completions/mean_terminated_length": 672.9239501953125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 8.587755102040816, + "grad_norm": 0.1495819091796875, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 554257025.0, + "reward": 0.6886160969734192, + "reward_std": 0.17641358077526093, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 919 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3748.0, + "completions/mean_length": 1161.114990234375, + "completions/mean_terminated_length": 654.0405883789062, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 8.597084548104956, + "grad_norm": 0.13234129548072815, + "learning_rate": 1e-06, + "loss": -0.0768, + "num_tokens": 554849768.0, + "reward": 0.6473214626312256, + "reward_std": 0.1488724797964096, + "rewards/verify_math_reward/mean": 0.6473214030265808, + "rewards/verify_math_reward/std": 0.47807061672210693, + "step": 920 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2974.0, + "completions/mean_length": 886.1417846679688, + "completions/mean_terminated_length": 588.6427001953125, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 8.606413994169095, + "grad_norm": 0.13144135475158691, + "learning_rate": 1e-06, + "loss": -0.0466, + "num_tokens": 555423847.0, + "reward": 0.6729910969734192, + "reward_std": 0.12377272546291351, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 921 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3778.0, + "completions/mean_length": 880.8147583007812, + "completions/mean_terminated_length": 591.3698120117188, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 8.615743440233237, + "grad_norm": 0.15162374079227448, + "learning_rate": 1e-06, + "loss": -0.0691, + "num_tokens": 556000081.0, + "reward": 0.7098214626312256, + "reward_std": 0.17209568619728088, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 922 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3887.0, + "completions/mean_length": 1095.3616943359375, + "completions/mean_terminated_length": 631.3453369140625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.625072886297376, + "grad_norm": 0.17637977004051208, + "learning_rate": 1e-06, + "loss": -0.073, + "num_tokens": 556579301.0, + "reward": 0.6595982313156128, + "reward_std": 0.18115399777889252, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 923 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3240.0, + "completions/mean_length": 1066.2232666015625, + "completions/mean_terminated_length": 606.694091796875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 8.634402332361516, + "grad_norm": 0.15288929641246796, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 557151101.0, + "reward": 0.6875000596046448, + "reward_std": 0.1478992998600006, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 924 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3440.0, + "completions/mean_length": 1001.25341796875, + "completions/mean_terminated_length": 634.2109985351562, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 8.643731778425655, + "grad_norm": 0.135414257645607, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 557742960.0, + "reward": 0.6462053656578064, + "reward_std": 0.13286243379116058, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 925 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 1092.5357666015625, + "completions/mean_terminated_length": 601.0597534179688, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 8.653061224489797, + "grad_norm": 0.16886992752552032, + "learning_rate": 1e-06, + "loss": -0.0453, + "num_tokens": 558302144.0, + "reward": 0.6183035969734192, + "reward_std": 0.14358071982860565, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 926 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3347.0, + "completions/mean_length": 993.3828735351562, + "completions/mean_terminated_length": 616.7196655273438, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 8.662390670553936, + "grad_norm": 0.1313415765762329, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 558885919.0, + "reward": 0.6785714626312256, + "reward_std": 0.11460563540458679, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 927 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2960.0, + "completions/mean_length": 1058.7254638671875, + "completions/mean_terminated_length": 629.2509765625, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 8.671720116618076, + "grad_norm": 0.1457333117723465, + "learning_rate": 1e-06, + "loss": -0.0664, + "num_tokens": 559478993.0, + "reward": 0.6116071939468384, + "reward_std": 0.12892180681228638, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 928 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2272.0, + "completions/mean_length": 1036.3695068359375, + "completions/mean_terminated_length": 621.4385375976562, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 8.681049562682215, + "grad_norm": 0.13202376663684845, + "learning_rate": 1e-06, + "loss": -0.0579, + "num_tokens": 560060044.0, + "reward": 0.6808035969734192, + "reward_std": 0.11727311462163925, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 929 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2738.0, + "completions/mean_length": 1036.28125, + "completions/mean_terminated_length": 643.2191162109375, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.690379008746355, + "grad_norm": 0.15420278906822205, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 560661512.0, + "reward": 0.6160714626312256, + "reward_std": 0.15390601754188538, + "rewards/verify_math_reward/mean": 0.6160714030265808, + "rewards/verify_math_reward/std": 0.486612468957901, + "step": 930 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 1060.3326416015625, + "completions/mean_terminated_length": 683.2546997070312, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 8.699708454810496, + "grad_norm": 0.1317526251077652, + "learning_rate": 1e-06, + "loss": -0.0689, + "num_tokens": 561295074.0, + "reward": 0.6495535969734192, + "reward_std": 0.139630526304245, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 931 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2959.0, + "completions/mean_length": 990.0234985351562, + "completions/mean_terminated_length": 599.8253784179688, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 8.709037900874636, + "grad_norm": 0.15937726199626923, + "learning_rate": 1e-06, + "loss": -0.0523, + "num_tokens": 561856855.0, + "reward": 0.652901828289032, + "reward_std": 0.15830057859420776, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 932 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3629.0, + "completions/mean_length": 981.9230346679688, + "completions/mean_terminated_length": 616.9314575195312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 8.718367346938775, + "grad_norm": 0.1484292596578598, + "learning_rate": 1e-06, + "loss": -0.0648, + "num_tokens": 562450514.0, + "reward": 0.6718750596046448, + "reward_std": 0.17348577082157135, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 933 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3364.0, + "completions/mean_length": 1076.6317138671875, + "completions/mean_terminated_length": 614.2059326171875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 8.727696793002915, + "grad_norm": 0.14339610934257507, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 563016672.0, + "reward": 0.629464328289032, + "reward_std": 0.14124813675880432, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 934 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3372.0, + "completions/mean_length": 1001.9810791015625, + "completions/mean_terminated_length": 604.5125732421875, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 8.737026239067056, + "grad_norm": 0.13736626505851746, + "learning_rate": 1e-06, + "loss": -0.0504, + "num_tokens": 563577575.0, + "reward": 0.6674107313156128, + "reward_std": 0.15360291302204132, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 935 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3566.0, + "completions/mean_length": 951.6707763671875, + "completions/mean_terminated_length": 574.3512573242188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 8.746355685131196, + "grad_norm": 0.16429510712623596, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 564139832.0, + "reward": 0.6640625, + "reward_std": 0.11840036511421204, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 936 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2883.0, + "completions/mean_length": 989.630615234375, + "completions/mean_terminated_length": 590.5755615234375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 8.755685131195335, + "grad_norm": 0.14231492578983307, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 564696685.0, + "reward": 0.6551339626312256, + "reward_std": 0.11355367302894592, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 937 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3315.0, + "completions/mean_length": 839.708740234375, + "completions/mean_terminated_length": 593.4345703125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 8.765014577259475, + "grad_norm": 0.15080682933330536, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 565272048.0, + "reward": 0.715401828289032, + "reward_std": 0.14496898651123047, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 938 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3488.0, + "completions/mean_length": 1076.8873291015625, + "completions/mean_terminated_length": 627.8910522460938, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 8.774344023323614, + "grad_norm": 0.16189219057559967, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 565847451.0, + "reward": 0.6071428656578064, + "reward_std": 0.1530490666627884, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 939 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2843.0, + "completions/mean_length": 1000.9832763671875, + "completions/mean_terminated_length": 616.5332641601562, + "completions/min_length": 210.0, + "completions/min_terminated_length": 210.0, + "epoch": 8.783673469387756, + "grad_norm": 0.17406950891017914, + "learning_rate": 1e-06, + "loss": -0.0805, + "num_tokens": 566426276.0, + "reward": 0.6897321939468384, + "reward_std": 0.15390713512897491, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 940 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3325.0, + "completions/mean_length": 1090.3070068359375, + "completions/mean_terminated_length": 660.9221801757812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 8.793002915451895, + "grad_norm": 0.1804381161928177, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 567034839.0, + "reward": 0.6071428656578064, + "reward_std": 0.16622982919216156, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865824937820435, + "step": 941 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2670.0, + "completions/mean_length": 1041.782470703125, + "completions/mean_terminated_length": 623.1840209960938, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.802332361516035, + "grad_norm": 0.1717265546321869, + "learning_rate": 1e-06, + "loss": -0.0409, + "num_tokens": 567609916.0, + "reward": 0.6205357313156128, + "reward_std": 0.15428438782691956, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 942 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2995.0, + "completions/mean_length": 1010.8906860351562, + "completions/mean_terminated_length": 627.6712646484375, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 8.811661807580174, + "grad_norm": 0.14452873170375824, + "learning_rate": 1e-06, + "loss": -0.0447, + "num_tokens": 568191546.0, + "reward": 0.676339328289032, + "reward_std": 0.1295548528432846, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 943 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 1058.891845703125, + "completions/mean_terminated_length": 611.686279296875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 8.820991253644316, + "grad_norm": 0.15940111875534058, + "learning_rate": 1e-06, + "loss": -0.0732, + "num_tokens": 568745809.0, + "reward": 0.5915178656578064, + "reward_std": 0.15954771637916565, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 944 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 1088.5, + "completions/mean_terminated_length": 632.349609375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 8.830320699708455, + "grad_norm": 0.15225429832935333, + "learning_rate": 1e-06, + "loss": -0.0644, + "num_tokens": 569326577.0, + "reward": 0.6127232313156128, + "reward_std": 0.13914015889167786, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 945 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3833.0, + "completions/mean_length": 1053.805908203125, + "completions/mean_terminated_length": 662.9949340820312, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "epoch": 8.839650145772595, + "grad_norm": 0.145501971244812, + "learning_rate": 1e-06, + "loss": -0.0939, + "num_tokens": 569943579.0, + "reward": 0.6651785969734192, + "reward_std": 0.16461333632469177, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219160199165344, + "step": 946 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4038.0, + "completions/mean_length": 940.4944458007812, + "completions/mean_terminated_length": 631.1311645507812, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 8.848979591836734, + "grad_norm": 0.15688036382198334, + "learning_rate": 1e-06, + "loss": -0.0353, + "num_tokens": 570559670.0, + "reward": 0.6718750596046448, + "reward_std": 0.16416826844215393, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 947 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2896.0, + "completions/mean_length": 973.0938110351562, + "completions/mean_terminated_length": 607.0673828125, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 8.858309037900874, + "grad_norm": 0.12668636441230774, + "learning_rate": 1e-06, + "loss": -0.0318, + "num_tokens": 571135250.0, + "reward": 0.6573660969734192, + "reward_std": 0.11265214532613754, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 948 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2580.0, + "completions/mean_length": 988.6160888671875, + "completions/mean_terminated_length": 602.6298828125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 8.867638483965015, + "grad_norm": 0.1370144933462143, + "learning_rate": 1e-06, + "loss": -0.0709, + "num_tokens": 571715330.0, + "reward": 0.6484375, + "reward_std": 0.12681996822357178, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 949 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3445.0, + "completions/mean_length": 957.7745971679688, + "completions/mean_terminated_length": 615.9876098632812, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 8.876967930029155, + "grad_norm": 0.14461150765419006, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 572313184.0, + "reward": 0.7031250596046448, + "reward_std": 0.13978135585784912, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 950 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2239.0, + "completions/mean_length": 814.2767944335938, + "completions/mean_terminated_length": 587.1408081054688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 8.886297376093294, + "grad_norm": 0.13550324738025665, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 572904224.0, + "reward": 0.699776828289032, + "reward_std": 0.11930190026760101, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 951 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3672.0, + "completions/mean_length": 1096.6820068359375, + "completions/mean_terminated_length": 663.8301391601562, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 8.895626822157434, + "grad_norm": 0.1618662327528, + "learning_rate": 1e-06, + "loss": -0.038, + "num_tokens": 573518075.0, + "reward": 0.6149553656578064, + "reward_std": 0.14755979180335999, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 952 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4027.0, + "completions/mean_length": 941.6395263671875, + "completions/mean_terminated_length": 576.3150634765625, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 8.904956268221575, + "grad_norm": 0.14199386537075043, + "learning_rate": 1e-06, + "loss": -0.0484, + "num_tokens": 574062040.0, + "reward": 0.6618303656578064, + "reward_std": 0.14120285212993622, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 953 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3893.0, + "completions/mean_length": 1111.421875, + "completions/mean_terminated_length": 645.442626953125, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 8.914285714285715, + "grad_norm": 0.1356530487537384, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 574656658.0, + "reward": 0.6015625, + "reward_std": 0.13365407288074493, + "rewards/verify_math_reward/mean": 0.6015625, + "rewards/verify_math_reward/std": 0.48984986543655396, + "step": 954 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3961.0, + "completions/mean_length": 894.9063110351562, + "completions/mean_terminated_length": 563.7586059570312, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 8.923615160349854, + "grad_norm": 0.15970854461193085, + "learning_rate": 1e-06, + "loss": -0.0709, + "num_tokens": 575194918.0, + "reward": 0.7109375596046448, + "reward_std": 0.17874614894390106, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 955 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 923.58154296875, + "completions/mean_terminated_length": 608.285888671875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 8.932944606413994, + "grad_norm": 0.14851707220077515, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 575775175.0, + "reward": 0.668526828289032, + "reward_std": 0.1468462347984314, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 956 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2820.0, + "completions/mean_length": 842.3984985351562, + "completions/mean_terminated_length": 566.6694946289062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 8.942274052478133, + "grad_norm": 0.13951444625854492, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 576338212.0, + "reward": 0.6651785969734192, + "reward_std": 0.10900939255952835, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 957 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2772.0, + "completions/mean_length": 1081.107177734375, + "completions/mean_terminated_length": 628.292724609375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 8.951603498542275, + "grad_norm": 0.14739088714122772, + "learning_rate": 1e-06, + "loss": -0.0955, + "num_tokens": 576913956.0, + "reward": 0.6004464626312256, + "reward_std": 0.14612942934036255, + "rewards/verify_math_reward/mean": 0.6004464030265808, + "rewards/verify_math_reward/std": 0.49008017778396606, + "step": 958 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3608.0, + "completions/mean_length": 984.8538208007812, + "completions/mean_terminated_length": 633.1589965820312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 8.960932944606414, + "grad_norm": 0.15532374382019043, + "learning_rate": 1e-06, + "loss": -0.075, + "num_tokens": 577514209.0, + "reward": 0.676339328289032, + "reward_std": 0.16710561513900757, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 959 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3354.0, + "completions/mean_length": 916.591552734375, + "completions/mean_terminated_length": 600.6011962890625, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 8.970262390670554, + "grad_norm": 0.15074577927589417, + "learning_rate": 1e-06, + "loss": -0.0322, + "num_tokens": 578094619.0, + "reward": 0.6718750596046448, + "reward_std": 0.149967223405838, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 960 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3248.0, + "completions/mean_length": 953.0324096679688, + "completions/mean_terminated_length": 589.0274047851562, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 8.979591836734693, + "grad_norm": 0.15810492634773254, + "learning_rate": 1e-06, + "loss": -0.0268, + "num_tokens": 578654120.0, + "reward": 0.6640625, + "reward_std": 0.1635284423828125, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 961 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3638.0, + "completions/mean_length": 946.5714721679688, + "completions/mean_terminated_length": 620.7684326171875, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 8.988921282798835, + "grad_norm": 0.1356414407491684, + "learning_rate": 1e-06, + "loss": -0.0289, + "num_tokens": 579238560.0, + "reward": 0.7053571939468384, + "reward_std": 0.12805598974227905, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 962 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.13920454545454541, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2054.0, + "completions/mean_length": 1050.5284423828125, + "completions/mean_terminated_length": 558.0264282226562, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 8.998250728862974, + "grad_norm": 0.18298207223415375, + "learning_rate": 1e-06, + "loss": -0.065, + "num_tokens": 579789177.0, + "reward": 0.6205357313156128, + "reward_std": 0.16687063872814178, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 963 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2618.0, + "completions/mean_length": 972.9933471679688, + "completions/mean_terminated_length": 585.0664672851562, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.00932944606414, + "grad_norm": 0.15523745119571686, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 580351363.0, + "reward": 0.6651785969734192, + "reward_std": 0.14560768008232117, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 964 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4073.0, + "completions/mean_length": 887.1551513671875, + "completions/mean_terminated_length": 572.5625, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 9.018658892128279, + "grad_norm": 0.15685448050498962, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 580909758.0, + "reward": 0.7165178656578064, + "reward_std": 0.14154870808124542, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 965 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3945.0, + "completions/mean_length": 972.3917846679688, + "completions/mean_terminated_length": 593.1802368164062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 9.02798833819242, + "grad_norm": 0.178028404712677, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 581468021.0, + "reward": 0.6863839626312256, + "reward_std": 0.16153329610824585, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 966 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3448.0, + "completions/mean_length": 1097.497802734375, + "completions/mean_terminated_length": 647.1450805664062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 9.03731778425656, + "grad_norm": 0.15089108049869537, + "learning_rate": 1e-06, + "loss": -0.0612, + "num_tokens": 582057803.0, + "reward": 0.652901828289032, + "reward_std": 0.1355287879705429, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631320357322693, + "step": 967 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3528.0, + "completions/mean_length": 998.1819458007812, + "completions/mean_terminated_length": 600.2254028320312, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.0466472303207, + "grad_norm": 0.14580699801445007, + "learning_rate": 1e-06, + "loss": -0.0609, + "num_tokens": 582621286.0, + "reward": 0.6674107313156128, + "reward_std": 0.1343708485364914, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140389680862427, + "step": 968 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3477.0, + "completions/mean_length": 984.7433471679688, + "completions/mean_terminated_length": 628.7288208007812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 9.055976676384839, + "grad_norm": 0.13815997540950775, + "learning_rate": 1e-06, + "loss": -0.0416, + "num_tokens": 583227464.0, + "reward": 0.6584821939468384, + "reward_std": 0.123772032558918, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 969 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3871.0, + "completions/mean_length": 997.9085083007812, + "completions/mean_terminated_length": 617.4411010742188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.06530612244898, + "grad_norm": 0.145284503698349, + "learning_rate": 1e-06, + "loss": -0.0444, + "num_tokens": 583809254.0, + "reward": 0.6640625, + "reward_std": 0.13031119108200073, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 970 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2427.0, + "completions/mean_length": 934.224365234375, + "completions/mean_terminated_length": 545.9360961914062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.07463556851312, + "grad_norm": 0.12844471633434296, + "learning_rate": 1e-06, + "loss": -0.0511, + "num_tokens": 584332151.0, + "reward": 0.6808035969734192, + "reward_std": 0.123467817902565, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 971 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3587.0, + "completions/mean_length": 975.00341796875, + "completions/mean_terminated_length": 604.84765625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 9.08396501457726, + "grad_norm": 0.15664070844650269, + "learning_rate": 1e-06, + "loss": -0.077, + "num_tokens": 584904626.0, + "reward": 0.6808035969734192, + "reward_std": 0.171934574842453, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 972 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 899.5781860351562, + "completions/mean_terminated_length": 586.2034301757812, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 9.093294460641399, + "grad_norm": 0.1798233985900879, + "learning_rate": 1e-06, + "loss": -0.0615, + "num_tokens": 585475168.0, + "reward": 0.6551339626312256, + "reward_std": 0.18013733625411987, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 973 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3997.0, + "completions/mean_length": 990.6551513671875, + "completions/mean_terminated_length": 622.3557739257812, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 9.102623906705539, + "grad_norm": 0.13798296451568604, + "learning_rate": 1e-06, + "loss": -0.0406, + "num_tokens": 586065595.0, + "reward": 0.6506696939468384, + "reward_std": 0.11802057921886444, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 974 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3365.0, + "completions/mean_length": 993.857177734375, + "completions/mean_terminated_length": 550.69384765625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.11195335276968, + "grad_norm": 0.16385027766227722, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 586580067.0, + "reward": 0.660714328289032, + "reward_std": 0.13061749935150146, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 975 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3586.0, + "completions/mean_length": 1059.05810546875, + "completions/mean_terminated_length": 647.2040405273438, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 9.12128279883382, + "grad_norm": 0.14419133961200714, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 587181055.0, + "reward": 0.640625, + "reward_std": 0.14789538085460663, + "rewards/verify_math_reward/mean": 0.640625, + "rewards/verify_math_reward/std": 0.48008525371551514, + "step": 976 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4039.0, + "completions/mean_length": 993.294677734375, + "completions/mean_terminated_length": 633.95263671875, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 9.130612244897959, + "grad_norm": 0.15172715485095978, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 587784871.0, + "reward": 0.6506696939468384, + "reward_std": 0.17325752973556519, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 977 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3393.0, + "completions/mean_length": 911.1183471679688, + "completions/mean_terminated_length": 624.4014282226562, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 9.139941690962099, + "grad_norm": 0.1590283066034317, + "learning_rate": 1e-06, + "loss": -0.0376, + "num_tokens": 588378329.0, + "reward": 0.7087053656578064, + "reward_std": 0.14920946955680847, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461273193359375, + "step": 978 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3390.0, + "completions/mean_length": 1045.560302734375, + "completions/mean_terminated_length": 640.6346435546875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 9.14927113702624, + "grad_norm": 0.15851905941963196, + "learning_rate": 1e-06, + "loss": -0.0494, + "num_tokens": 588984543.0, + "reward": 0.6439732313156128, + "reward_std": 0.15067513287067413, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 979 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3693.0, + "completions/mean_length": 1075.0101318359375, + "completions/mean_terminated_length": 553.0588989257812, + "completions/min_length": 183.0, + "completions/min_terminated_length": 183.0, + "epoch": 9.15860058309038, + "grad_norm": 0.15591056644916534, + "learning_rate": 1e-06, + "loss": -0.0323, + "num_tokens": 589490032.0, + "reward": 0.676339328289032, + "reward_std": 0.10780616104602814, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 980 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3340.0, + "completions/mean_length": 951.4553833007812, + "completions/mean_terminated_length": 617.5901489257812, + "completions/min_length": 176.0, + "completions/min_terminated_length": 176.0, + "epoch": 9.167930029154519, + "grad_norm": 0.15840359032154083, + "learning_rate": 1e-06, + "loss": -0.0271, + "num_tokens": 590083664.0, + "reward": 0.6305803656578064, + "reward_std": 0.15300628542900085, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 981 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4037.0, + "completions/mean_length": 880.911865234375, + "completions/mean_terminated_length": 591.4756469726562, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 9.177259475218658, + "grad_norm": 0.15357515215873718, + "learning_rate": 1e-06, + "loss": -0.0498, + "num_tokens": 590655937.0, + "reward": 0.6774553656578064, + "reward_std": 0.16781355440616608, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 982 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 945.779052734375, + "completions/mean_terminated_length": 615.6078491210938, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 9.186588921282798, + "grad_norm": 0.13610170781612396, + "learning_rate": 1e-06, + "loss": -0.0827, + "num_tokens": 591243235.0, + "reward": 0.6774553656578064, + "reward_std": 0.15773534774780273, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 983 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3086.0, + "completions/mean_length": 1023.2701416015625, + "completions/mean_terminated_length": 619.7802734375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 9.19591836734694, + "grad_norm": 0.16177140176296234, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 591826621.0, + "reward": 0.6484375, + "reward_std": 0.15044620633125305, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 984 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 1004.0803833007812, + "completions/mean_terminated_length": 645.987548828125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 9.205247813411079, + "grad_norm": 0.16494090855121613, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 592430781.0, + "reward": 0.621651828289032, + "reward_std": 0.15886442363262177, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 985 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3912.0, + "completions/mean_length": 1041.2054443359375, + "completions/mean_terminated_length": 600.3473510742188, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 9.214577259475218, + "grad_norm": 0.18257379531860352, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 592994245.0, + "reward": 0.6808035969734192, + "reward_std": 0.16529759764671326, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 986 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2958.0, + "completions/mean_length": 966.20654296875, + "completions/mean_terminated_length": 555.2235107421875, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 9.223906705539358, + "grad_norm": 0.1816340833902359, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 593528862.0, + "reward": 0.637276828289032, + "reward_std": 0.16638389229774475, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 987 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3838.0, + "completions/mean_length": 878.8092041015625, + "completions/mean_terminated_length": 589.1837158203125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 9.2332361516035, + "grad_norm": 0.16462019085884094, + "learning_rate": 1e-06, + "loss": -0.0173, + "num_tokens": 594104987.0, + "reward": 0.6506696939468384, + "reward_std": 0.15113018453121185, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 988 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3998.0, + "completions/mean_length": 1006.1194458007812, + "completions/mean_terminated_length": 673.8331298828125, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.242565597667639, + "grad_norm": 0.14079514145851135, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 594741406.0, + "reward": 0.7008928656578064, + "reward_std": 0.1510860174894333, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 989 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3574.0, + "completions/mean_length": 1008.3270263671875, + "completions/mean_terminated_length": 633.4780883789062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.251895043731778, + "grad_norm": 0.15933242440223694, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 595333539.0, + "reward": 0.6618303656578064, + "reward_std": 0.16033892333507538, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 990 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3410.0, + "completions/mean_length": 902.9185791015625, + "completions/mean_terminated_length": 563.9000244140625, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 9.261224489795918, + "grad_norm": 0.14286759495735168, + "learning_rate": 1e-06, + "loss": -0.0744, + "num_tokens": 595872562.0, + "reward": 0.715401828289032, + "reward_std": 0.13868620991706848, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 991 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2865.0, + "completions/mean_length": 930.0324096679688, + "completions/mean_terminated_length": 632.3773193359375, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 9.270553935860057, + "grad_norm": 0.15137651562690735, + "learning_rate": 1e-06, + "loss": -0.0533, + "num_tokens": 596475951.0, + "reward": 0.6975446939468384, + "reward_std": 0.16093626618385315, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957791805267334, + "step": 992 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2948.0, + "completions/mean_length": 1091.102783203125, + "completions/mean_terminated_length": 653.048583984375, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 9.279883381924199, + "grad_norm": 0.1477823704481125, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 597079171.0, + "reward": 0.5859375, + "reward_std": 0.17743414640426636, + "rewards/verify_math_reward/mean": 0.5859375, + "rewards/verify_math_reward/std": 0.4928344786167145, + "step": 993 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3348.0, + "completions/mean_length": 944.5167846679688, + "completions/mean_terminated_length": 605.6056518554688, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 9.289212827988338, + "grad_norm": 0.15418857336044312, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 597657330.0, + "reward": 0.6462053656578064, + "reward_std": 0.1562725007534027, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 994 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3086.0, + "completions/mean_length": 887.7857666015625, + "completions/mean_terminated_length": 581.8679809570312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 9.298542274052478, + "grad_norm": 0.13969087600708008, + "learning_rate": 1e-06, + "loss": -0.0529, + "num_tokens": 598215602.0, + "reward": 0.699776828289032, + "reward_std": 0.14139717817306519, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 995 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2832.0, + "completions/mean_length": 1043.700927734375, + "completions/mean_terminated_length": 629.7642822265625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.307871720116617, + "grad_norm": 0.17577214539051056, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 598805894.0, + "reward": 0.6171875, + "reward_std": 0.16912440955638885, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 996 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 968.3326416015625, + "completions/mean_terminated_length": 619.08935546875, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 9.317201166180759, + "grad_norm": 0.15717150270938873, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 599390224.0, + "reward": 0.6729910969734192, + "reward_std": 0.14199700951576233, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 997 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3780.0, + "completions/mean_length": 914.9989013671875, + "completions/mean_terminated_length": 568.55322265625, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.326530612244898, + "grad_norm": 0.155115008354187, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 599939559.0, + "reward": 0.6953125596046448, + "reward_std": 0.13861344754695892, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 998 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2181.0, + "completions/mean_length": 875.2053833007812, + "completions/mean_terminated_length": 580.9793090820312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 9.335860058309038, + "grad_norm": 0.14764153957366943, + "learning_rate": 1e-06, + "loss": -0.0399, + "num_tokens": 600498183.0, + "reward": 0.6897321939468384, + "reward_std": 0.14815779030323029, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 999 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3025.0, + "completions/mean_length": 1081.6707763671875, + "completions/mean_terminated_length": 597.5038452148438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 9.345189504373177, + "grad_norm": 0.1988637000322342, + "learning_rate": 1e-06, + "loss": -0.045, + "num_tokens": 601049440.0, + "reward": 0.5915178656578064, + "reward_std": 0.1542869210243225, + "rewards/verify_math_reward/mean": 0.5915178656578064, + "rewards/verify_math_reward/std": 0.49182769656181335, + "step": 1000 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 1033.6529541015625, + "completions/mean_terminated_length": 605.0801391601562, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 9.354518950437317, + "grad_norm": 0.1638064831495285, + "learning_rate": 1e-06, + "loss": -0.0534, + "num_tokens": 601605289.0, + "reward": 0.6707589626312256, + "reward_std": 0.14016184210777283, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1001 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 1007.7277221679688, + "completions/mean_terminated_length": 671.3811645507812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.363848396501458, + "grad_norm": 0.14289306104183197, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 602243333.0, + "reward": 0.668526828289032, + "reward_std": 0.1544705480337143, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1002 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3500.0, + "completions/mean_length": 923.786865234375, + "completions/mean_terminated_length": 586.9839477539062, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 9.373177842565598, + "grad_norm": 0.1334981769323349, + "learning_rate": 1e-06, + "loss": -0.0433, + "num_tokens": 602804630.0, + "reward": 0.707589328289032, + "reward_std": 0.13324038684368134, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1003 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2325.0, + "completions/mean_length": 769.4553833007812, + "completions/mean_terminated_length": 556.114013671875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 9.382507288629737, + "grad_norm": 0.19102612137794495, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 603362878.0, + "reward": 0.707589328289032, + "reward_std": 0.17882463335990906, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1004 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2919.0, + "completions/mean_length": 863.1652221679688, + "completions/mean_terminated_length": 572.1314086914062, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 9.391836734693877, + "grad_norm": 0.14407780766487122, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 603923746.0, + "reward": 0.7176339626312256, + "reward_std": 0.10870447009801865, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1005 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4073.0, + "completions/mean_length": 919.1920166015625, + "completions/mean_terminated_length": 546.847900390625, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 9.401166180758018, + "grad_norm": 0.15340860188007355, + "learning_rate": 1e-06, + "loss": -0.0269, + "num_tokens": 604459990.0, + "reward": 0.660714328289032, + "reward_std": 0.11930259317159653, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1006 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2443.0, + "completions/mean_length": 853.9074096679688, + "completions/mean_terminated_length": 566.3341674804688, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 9.410495626822158, + "grad_norm": 0.13971778750419617, + "learning_rate": 1e-06, + "loss": -0.0244, + "num_tokens": 605015131.0, + "reward": 0.7187500596046448, + "reward_std": 0.11362994462251663, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1007 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3859.0, + "completions/mean_length": 930.2623291015625, + "completions/mean_terminated_length": 624.1505737304688, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 9.419825072886297, + "grad_norm": 0.164055734872818, + "learning_rate": 1e-06, + "loss": -0.0388, + "num_tokens": 605610910.0, + "reward": 0.6953125596046448, + "reward_std": 0.14910070598125458, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1008 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 871.5938110351562, + "completions/mean_terminated_length": 533.6473388671875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.429154518950437, + "grad_norm": 0.16293291747570038, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 606123970.0, + "reward": 0.7187500596046448, + "reward_std": 0.12050722539424896, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1009 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3698.0, + "completions/mean_length": 1076.46435546875, + "completions/mean_terminated_length": 692.8502807617188, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 9.438483965014576, + "grad_norm": 0.17608752846717834, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 606774402.0, + "reward": 0.5714285969734192, + "reward_std": 0.19599542021751404, + "rewards/verify_math_reward/mean": 0.5714285969734192, + "rewards/verify_math_reward/std": 0.49514803290367126, + "step": 1010 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3659.0, + "completions/mean_length": 1127.67529296875, + "completions/mean_terminated_length": 632.9544677734375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 9.447813411078718, + "grad_norm": 0.1741783320903778, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 607347447.0, + "reward": 0.6205357313156128, + "reward_std": 0.18295523524284363, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1011 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2633.0, + "completions/mean_length": 992.4230346679688, + "completions/mean_terminated_length": 632.9800415039062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.457142857142857, + "grad_norm": 0.15500663220882416, + "learning_rate": 1e-06, + "loss": -0.0523, + "num_tokens": 607951898.0, + "reward": 0.6908482313156128, + "reward_std": 0.15259191393852234, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1012 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3389.0, + "completions/mean_length": 961.2779541015625, + "completions/mean_terminated_length": 624.1693115234375, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 9.466472303206997, + "grad_norm": 0.14826615154743195, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 608548507.0, + "reward": 0.7008928656578064, + "reward_std": 0.12895894050598145, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1013 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3428.0, + "completions/mean_length": 914.1975708007812, + "completions/mean_terminated_length": 627.7578735351562, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 9.475801749271136, + "grad_norm": 0.12536974251270294, + "learning_rate": 1e-06, + "loss": -0.0029, + "num_tokens": 609155756.0, + "reward": 0.7187500596046448, + "reward_std": 0.10498502850532532, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1014 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3230.0, + "completions/mean_length": 1088.0804443359375, + "completions/mean_terminated_length": 591.3237915039062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.485131195335278, + "grad_norm": 0.17755606770515442, + "learning_rate": 1e-06, + "loss": -0.058, + "num_tokens": 609693284.0, + "reward": 0.65625, + "reward_std": 0.14744143187999725, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1015 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3253.0, + "completions/mean_length": 1079.6004638671875, + "completions/mean_terminated_length": 608.6529541015625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 9.494460641399417, + "grad_norm": 0.17347857356071472, + "learning_rate": 1e-06, + "loss": -0.0708, + "num_tokens": 610249870.0, + "reward": 0.645089328289032, + "reward_std": 0.1585274040699005, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1016 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 1072.146240234375, + "completions/mean_terminated_length": 617.9859008789062, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 9.503790087463557, + "grad_norm": 0.15390904247760773, + "learning_rate": 1e-06, + "loss": -0.0446, + "num_tokens": 610825953.0, + "reward": 0.6718750596046448, + "reward_std": 0.1424848586320877, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1017 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2796.0, + "completions/mean_length": 980.2199096679688, + "completions/mean_terminated_length": 579.9558715820312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 9.513119533527696, + "grad_norm": 0.17296528816223145, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 611378278.0, + "reward": 0.6696428656578064, + "reward_std": 0.16581936180591583, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1018 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2654.0, + "completions/mean_length": 939.2109985351562, + "completions/mean_terminated_length": 612.6465454101562, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 9.522448979591836, + "grad_norm": 0.16253484785556793, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 611960899.0, + "reward": 0.6752232313156128, + "reward_std": 0.16288763284683228, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1019 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3488.0, + "completions/mean_length": 1053.6038818359375, + "completions/mean_terminated_length": 596.6585693359375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 9.531778425655977, + "grad_norm": 0.16379345953464508, + "learning_rate": 1e-06, + "loss": -0.0479, + "num_tokens": 612504984.0, + "reward": 0.660714328289032, + "reward_std": 0.1202450841665268, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1020 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3004.0, + "completions/mean_length": 1090.860595703125, + "completions/mean_terminated_length": 608.169677734375, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 9.541107871720117, + "grad_norm": 0.15253886580467224, + "learning_rate": 1e-06, + "loss": -0.0678, + "num_tokens": 613062131.0, + "reward": 0.613839328289032, + "reward_std": 0.12952165305614471, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1021 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3964.0, + "completions/mean_length": 1005.841552734375, + "completions/mean_terminated_length": 647.95263671875, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 9.550437317784256, + "grad_norm": 0.16293002665042877, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 613661829.0, + "reward": 0.6127232313156128, + "reward_std": 0.14966455101966858, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 1022 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3766.0, + "completions/mean_length": 957.3516235351562, + "completions/mean_terminated_length": 653.8592529296875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 9.559766763848396, + "grad_norm": 0.1554853916168213, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 614283856.0, + "reward": 0.6339285969734192, + "reward_std": 0.16288836300373077, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1023 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3267.0, + "completions/mean_length": 971.9397583007812, + "completions/mean_terminated_length": 570.612060546875, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 9.569096209912537, + "grad_norm": 0.16789597272872925, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 614838962.0, + "reward": 0.6417410969734192, + "reward_std": 0.15890537202358246, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1024 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3255.0, + "completions/mean_length": 1086.685302734375, + "completions/mean_terminated_length": 639.1461791992188, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.578425655976677, + "grad_norm": 0.1663643717765808, + "learning_rate": 1e-06, + "loss": -0.0723, + "num_tokens": 615423048.0, + "reward": 0.6428571939468384, + "reward_std": 0.15417560935020447, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1025 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2953.0, + "completions/mean_length": 1089.6138916015625, + "completions/mean_terminated_length": 629.1763305664062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 9.587755102040816, + "grad_norm": 0.14654986560344696, + "learning_rate": 1e-06, + "loss": -0.0277, + "num_tokens": 616004894.0, + "reward": 0.637276828289032, + "reward_std": 0.12238264083862305, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1026 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2489.0, + "completions/mean_length": 1066.9609375, + "completions/mean_terminated_length": 566.7164916992188, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 9.597084548104956, + "grad_norm": 0.1519494354724884, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 616528155.0, + "reward": 0.6796875596046448, + "reward_std": 0.12400025874376297, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1027 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3623.0, + "completions/mean_length": 854.0636596679688, + "completions/mean_terminated_length": 536.2267456054688, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 9.606413994169095, + "grad_norm": 0.16887415945529938, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 617054892.0, + "reward": 0.7522321939468384, + "reward_std": 0.130574032664299, + "rewards/verify_math_reward/mean": 0.7522321343421936, + "rewards/verify_math_reward/std": 0.4319573938846588, + "step": 1028 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2931.0, + "completions/mean_length": 1104.26123046875, + "completions/mean_terminated_length": 578.1548461914062, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 9.615743440233237, + "grad_norm": 0.14007607102394104, + "learning_rate": 1e-06, + "loss": -0.0821, + "num_tokens": 617580982.0, + "reward": 0.6584821939468384, + "reward_std": 0.11434461921453476, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1029 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 1028.2567138671875, + "completions/mean_terminated_length": 621.0341796875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 9.625072886297376, + "grad_norm": 0.15491919219493866, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 618158372.0, + "reward": 0.6584821939468384, + "reward_std": 0.17040391266345978, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1030 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3393.0, + "completions/mean_length": 1032.325927734375, + "completions/mean_terminated_length": 590.1864624023438, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 9.634402332361516, + "grad_norm": 0.13611069321632385, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 618706120.0, + "reward": 0.6629464626312256, + "reward_std": 0.09232490509748459, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1031 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2673.0, + "completions/mean_length": 1077.8046875, + "completions/mean_terminated_length": 597.5485229492188, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 9.643731778425655, + "grad_norm": 0.17063143849372864, + "learning_rate": 1e-06, + "loss": -0.0969, + "num_tokens": 619259441.0, + "reward": 0.6484375, + "reward_std": 0.14011837542057037, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1032 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2556.0, + "completions/mean_length": 971.0324096679688, + "completions/mean_terminated_length": 630.6893310546875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 9.653061224489797, + "grad_norm": 0.1838199645280838, + "learning_rate": 1e-06, + "loss": -0.0804, + "num_tokens": 619866606.0, + "reward": 0.6595982313156128, + "reward_std": 0.20425845682621002, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1033 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3895.0, + "completions/mean_length": 887.2522583007812, + "completions/mean_terminated_length": 585.5751342773438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.662390670553936, + "grad_norm": 0.16632024943828583, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 620428760.0, + "reward": 0.7187500596046448, + "reward_std": 0.1262579709291458, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1034 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3400.0, + "completions/mean_length": 983.7969360351562, + "completions/mean_terminated_length": 661.8447875976562, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.671720116618076, + "grad_norm": 0.1719866842031479, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 621045266.0, + "reward": 0.6796875596046448, + "reward_std": 0.1623249351978302, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1035 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3319.0, + "completions/mean_length": 984.2667846679688, + "completions/mean_terminated_length": 628.19775390625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 9.681049562682215, + "grad_norm": 0.1803593635559082, + "learning_rate": 1e-06, + "loss": -0.0542, + "num_tokens": 621646633.0, + "reward": 0.65625, + "reward_std": 0.1555236428976059, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1036 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3900.0, + "completions/mean_length": 912.6172485351562, + "completions/mean_terminated_length": 570.275634765625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.690379008746355, + "grad_norm": 0.18360450863838196, + "learning_rate": 1e-06, + "loss": -0.0242, + "num_tokens": 622194098.0, + "reward": 0.6796875596046448, + "reward_std": 0.1615767478942871, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1037 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1629464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2721.0, + "completions/mean_length": 1131.9320068359375, + "completions/mean_terminated_length": 554.9266357421875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 9.699708454810496, + "grad_norm": 0.17833773791790009, + "learning_rate": 1e-06, + "loss": -0.047, + "num_tokens": 622697237.0, + "reward": 0.6149553656578064, + "reward_std": 0.13429346680641174, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1038 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 959.6663208007812, + "completions/mean_terminated_length": 613.7756958007812, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 9.709037900874636, + "grad_norm": 0.15967030823230743, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 623287602.0, + "reward": 0.6640625, + "reward_std": 0.15867894887924194, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1039 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3141.0, + "completions/mean_length": 990.7511596679688, + "completions/mean_terminated_length": 591.8400268554688, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 9.718367346938775, + "grad_norm": 0.16621524095535278, + "learning_rate": 1e-06, + "loss": -0.036, + "num_tokens": 623851635.0, + "reward": 0.65625, + "reward_std": 0.14853617548942566, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1040 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 972.9609985351562, + "completions/mean_terminated_length": 598.1962280273438, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 9.727696793002915, + "grad_norm": 0.16122128069400787, + "learning_rate": 1e-06, + "loss": -0.0617, + "num_tokens": 624425792.0, + "reward": 0.668526828289032, + "reward_std": 0.1457592099905014, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1041 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3645.0, + "completions/mean_length": 1064.47998046875, + "completions/mean_terminated_length": 653.3612060546875, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 9.737026239067056, + "grad_norm": 0.14171285927295685, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 625034038.0, + "reward": 0.6127232313156128, + "reward_std": 0.13752757012844086, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 1042 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 1004.466552734375, + "completions/mean_terminated_length": 589.6531982421875, + "completions/min_length": 88.0, + "completions/min_terminated_length": 88.0, + "epoch": 9.746355685131196, + "grad_norm": 0.19457073509693146, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 625590008.0, + "reward": 0.6796875596046448, + "reward_std": 0.1418447494506836, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1043 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2513.0, + "completions/mean_length": 1002.94091796875, + "completions/mean_terminated_length": 601.1942138671875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.755685131195335, + "grad_norm": 0.17160455882549286, + "learning_rate": 1e-06, + "loss": -0.0726, + "num_tokens": 626159571.0, + "reward": 0.6707589626312256, + "reward_std": 0.1532338410615921, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1044 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2469.0, + "completions/mean_length": 1099.8382568359375, + "completions/mean_terminated_length": 618.58935546875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 9.765014577259475, + "grad_norm": 0.15683704614639282, + "learning_rate": 1e-06, + "loss": -0.085, + "num_tokens": 626723858.0, + "reward": 0.6696428656578064, + "reward_std": 0.14135439693927765, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1045 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 1113.2388916015625, + "completions/mean_terminated_length": 643.087890625, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 9.774344023323614, + "grad_norm": 0.1665021777153015, + "learning_rate": 1e-06, + "loss": -0.0315, + "num_tokens": 627303152.0, + "reward": 0.6049107313156128, + "reward_std": 0.14150774478912354, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 1046 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3625.0, + "completions/mean_length": 946.2835083007812, + "completions/mean_terminated_length": 568.3175048828125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 9.783673469387756, + "grad_norm": 0.16322743892669678, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 627845998.0, + "reward": 0.676339328289032, + "reward_std": 0.1272720992565155, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1047 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2702.0, + "completions/mean_length": 862.5301513671875, + "completions/mean_terminated_length": 541.1668701171875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 9.793002915451895, + "grad_norm": 0.1671653389930725, + "learning_rate": 1e-06, + "loss": -0.0401, + "num_tokens": 628384145.0, + "reward": 0.7455357313156128, + "reward_std": 0.14361140131950378, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1048 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3800.0, + "completions/mean_length": 1083.396240234375, + "completions/mean_terminated_length": 617.5296020507812, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 9.802332361516035, + "grad_norm": 0.161100834608078, + "learning_rate": 1e-06, + "loss": -0.0192, + "num_tokens": 628955988.0, + "reward": 0.645089328289032, + "reward_std": 0.11336751282215118, + "rewards/verify_math_reward/mean": 0.6450892686843872, + "rewards/verify_math_reward/std": 0.4787535071372986, + "step": 1049 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3315.0, + "completions/mean_length": 1001.599365234375, + "completions/mean_terminated_length": 586.4012451171875, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 9.811661807580174, + "grad_norm": 0.14529836177825928, + "learning_rate": 1e-06, + "loss": -0.0541, + "num_tokens": 629519917.0, + "reward": 0.6484375, + "reward_std": 0.12816546857357025, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1050 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 1090.8114013671875, + "completions/mean_terminated_length": 608.1126708984375, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 9.820991253644316, + "grad_norm": 0.14906223118305206, + "learning_rate": 1e-06, + "loss": -0.0717, + "num_tokens": 630075284.0, + "reward": 0.6964285969734192, + "reward_std": 0.12651507556438446, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1051 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3209.0, + "completions/mean_length": 1045.35498046875, + "completions/mean_terminated_length": 569.0606689453125, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 9.830320699708455, + "grad_norm": 0.14279192686080933, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 630603450.0, + "reward": 0.6662946939468384, + "reward_std": 0.10138323158025742, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1052 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2816.0, + "completions/mean_length": 995.1350708007812, + "completions/mean_terminated_length": 561.1717529296875, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 9.839650145772595, + "grad_norm": 0.17934824526309967, + "learning_rate": 1e-06, + "loss": -0.0875, + "num_tokens": 631139811.0, + "reward": 0.6741071939468384, + "reward_std": 0.1666516810655594, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 1053 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3284.0, + "completions/mean_length": 970.7835083007812, + "completions/mean_terminated_length": 586.9849853515625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 9.848979591836734, + "grad_norm": 0.1426946371793747, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 631693105.0, + "reward": 0.6729910969734192, + "reward_std": 0.11381329596042633, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1054 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3501.0, + "completions/mean_length": 851.888427734375, + "completions/mean_terminated_length": 555.5322875976562, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.858309037900874, + "grad_norm": 0.1521221250295639, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 632246597.0, + "reward": 0.7366071939468384, + "reward_std": 0.13410983979701996, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1055 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4045.0, + "completions/mean_length": 1026.2935791015625, + "completions/mean_terminated_length": 631.9483642578125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 9.867638483965015, + "grad_norm": 0.1641969233751297, + "learning_rate": 1e-06, + "loss": -0.0561, + "num_tokens": 632832932.0, + "reward": 0.6395089626312256, + "reward_std": 0.15454724431037903, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 1056 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2137.0, + "completions/mean_length": 876.4989013671875, + "completions/mean_terminated_length": 543.447021484375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 9.876967930029155, + "grad_norm": 0.15613830089569092, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 633363379.0, + "reward": 0.6897321939468384, + "reward_std": 0.1460181325674057, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1057 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2311.0, + "completions/mean_length": 924.87841796875, + "completions/mean_terminated_length": 508.46844482421875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 9.886297376093294, + "grad_norm": 0.1545645147562027, + "learning_rate": 1e-06, + "loss": -0.0777, + "num_tokens": 633849238.0, + "reward": 0.7064732313156128, + "reward_std": 0.12625475227832794, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1058 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3749.0, + "completions/mean_length": 891.880615234375, + "completions/mean_terminated_length": 599.1778564453125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 9.895626822157434, + "grad_norm": 0.14035220444202423, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 634438515.0, + "reward": 0.6674107313156128, + "reward_std": 0.13804751634597778, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1059 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3062.0, + "completions/mean_length": 948.6629638671875, + "completions/mean_terminated_length": 575.3832397460938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 9.904956268221575, + "grad_norm": 0.15624937415122986, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 634996061.0, + "reward": 0.6618303656578064, + "reward_std": 0.1328292191028595, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1060 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3317.0, + "completions/mean_length": 941.3660888671875, + "completions/mean_terminated_length": 623.577392578125, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 9.914285714285715, + "grad_norm": 0.15813840925693512, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 635604485.0, + "reward": 0.6953125596046448, + "reward_std": 0.16416893899440765, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1061 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3316.0, + "completions/mean_length": 1057.9453125, + "completions/mean_terminated_length": 565.394287109375, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 9.923615160349854, + "grad_norm": 0.1818017214536667, + "learning_rate": 1e-06, + "loss": -0.0858, + "num_tokens": 636131932.0, + "reward": 0.6540178656578064, + "reward_std": 0.15293143689632416, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1062 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2635.0, + "completions/mean_length": 1010.4676513671875, + "completions/mean_terminated_length": 578.650146484375, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 9.932944606413994, + "grad_norm": 0.16857455670833588, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 636677983.0, + "reward": 0.6674107313156128, + "reward_std": 0.14661657810211182, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1063 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4084.0, + "completions/mean_length": 1129.599365234375, + "completions/mean_terminated_length": 621.62744140625, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 9.942274052478133, + "grad_norm": 0.18123479187488556, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 637248304.0, + "reward": 0.6462053656578064, + "reward_std": 0.15180166065692902, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1064 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3926.0, + "completions/mean_length": 983.0045166015625, + "completions/mean_terminated_length": 618.1396484375, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 9.951603498542275, + "grad_norm": 0.15925569832324982, + "learning_rate": 1e-06, + "loss": -0.0552, + "num_tokens": 637830204.0, + "reward": 0.6506696939468384, + "reward_std": 0.1494358777999878, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1065 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3720.0, + "completions/mean_length": 1131.5413818359375, + "completions/mean_terminated_length": 623.9019775390625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 9.960932944606414, + "grad_norm": 0.18021221458911896, + "learning_rate": 1e-06, + "loss": -0.0442, + "num_tokens": 638396305.0, + "reward": 0.6316964626312256, + "reward_std": 0.1694614291191101, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1066 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3845.0, + "completions/mean_length": 915.1730346679688, + "completions/mean_terminated_length": 594.7456665039062, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 9.970262390670554, + "grad_norm": 0.15773488581180573, + "learning_rate": 1e-06, + "loss": -0.0589, + "num_tokens": 638961300.0, + "reward": 0.7444196939468384, + "reward_std": 0.14902400970458984, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1067 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2367.0, + "completions/mean_length": 1010.0223388671875, + "completions/mean_terminated_length": 560.1483154296875, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 9.979591836734693, + "grad_norm": 0.13540399074554443, + "learning_rate": 1e-06, + "loss": -0.0744, + "num_tokens": 639484784.0, + "reward": 0.7020089626312256, + "reward_std": 0.10528742522001266, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763102173805237, + "step": 1068 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3573.0, + "completions/mean_length": 1036.1429443359375, + "completions/mean_terminated_length": 599.0203857421875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 9.988921282798835, + "grad_norm": 0.16124646365642548, + "learning_rate": 1e-06, + "loss": -0.0744, + "num_tokens": 640055528.0, + "reward": 0.6350446939468384, + "reward_std": 0.1471058428287506, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1069 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.12784090909090906, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2963.0, + "completions/mean_length": 1138.1279296875, + "completions/mean_terminated_length": 704.5634765625, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 9.998250728862974, + "grad_norm": 0.20521681010723114, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 640587899.0, + "reward": 0.6361607313156128, + "reward_std": 0.15367960929870605, + "rewards/verify_math_reward/mean": 0.6361607313156128, + "rewards/verify_math_reward/std": 0.4813718795776367, + "step": 1070 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4010.0, + "completions/mean_length": 1087.6585693359375, + "completions/mean_terminated_length": 649.102294921875, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.00932944606414, + "grad_norm": 0.1542617231607437, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 641189913.0, + "reward": 0.6506696939468384, + "reward_std": 0.1329379826784134, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1071 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3401.0, + "completions/mean_length": 1000.9420166015625, + "completions/mean_terminated_length": 598.9356689453125, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 10.018658892128279, + "grad_norm": 0.14231081306934357, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 641760725.0, + "reward": 0.7053571939468384, + "reward_std": 0.15634846687316895, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1072 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1674107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3491.0, + "completions/mean_length": 1176.2254638671875, + "completions/mean_terminated_length": 589.139404296875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 10.02798833819242, + "grad_norm": 0.17777127027511597, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 642284815.0, + "reward": 0.6305803656578064, + "reward_std": 0.1293700933456421, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1073 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3855.0, + "completions/mean_length": 1045.1015625, + "completions/mean_terminated_length": 661.8228759765625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 10.03731778425656, + "grad_norm": 0.14099112153053284, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 642892194.0, + "reward": 0.6573660969734192, + "reward_std": 0.14173486828804016, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1074 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3407.0, + "completions/mean_length": 960.7154541015625, + "completions/mean_terminated_length": 593.2381591796875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.0466472303207, + "grad_norm": 0.15719178318977356, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 643453019.0, + "reward": 0.6930803656578064, + "reward_std": 0.13711319863796234, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1075 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3348.0, + "completions/mean_length": 1066.216552734375, + "completions/mean_terminated_length": 593.1793823242188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.055976676384839, + "grad_norm": 0.15955573320388794, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 643994165.0, + "reward": 0.6830357313156128, + "reward_std": 0.15431898832321167, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1076 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3873.0, + "completions/mean_length": 1052.8671875, + "completions/mean_terminated_length": 618.1339111328125, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.06530612244898, + "grad_norm": 0.135609969496727, + "learning_rate": 1e-06, + "loss": -0.0417, + "num_tokens": 644583118.0, + "reward": 0.652901828289032, + "reward_std": 0.10968156903982162, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 1077 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3183.0, + "completions/mean_length": 1035.337158203125, + "completions/mean_terminated_length": 620.2661743164062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 10.07463556851312, + "grad_norm": 0.14457030594348907, + "learning_rate": 1e-06, + "loss": -0.0626, + "num_tokens": 645158396.0, + "reward": 0.6897321939468384, + "reward_std": 0.13436973094940186, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1078 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2183.0, + "completions/mean_length": 1063.07373046875, + "completions/mean_terminated_length": 552.9725952148438, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.08396501457726, + "grad_norm": 0.17858922481536865, + "learning_rate": 1e-06, + "loss": -0.0763, + "num_tokens": 645679950.0, + "reward": 0.6640625, + "reward_std": 0.15067623555660248, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1079 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3655.0, + "completions/mean_length": 916.6406860351562, + "completions/mean_terminated_length": 570.373779296875, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 10.093294460641399, + "grad_norm": 0.17197096347808838, + "learning_rate": 1e-06, + "loss": -0.06, + "num_tokens": 646226452.0, + "reward": 0.7042410969734192, + "reward_std": 0.16727899014949799, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1080 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3485.0, + "completions/mean_length": 1130.05810546875, + "completions/mean_terminated_length": 622.1647338867188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.102623906705539, + "grad_norm": 0.15414667129516602, + "learning_rate": 1e-06, + "loss": -0.0796, + "num_tokens": 646783112.0, + "reward": 0.6104910969734192, + "reward_std": 0.15766116976737976, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791125416755676, + "step": 1081 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3344.0, + "completions/mean_length": 1002.7589721679688, + "completions/mean_terminated_length": 618.5294799804688, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.11195335276968, + "grad_norm": 0.1677163541316986, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 647373880.0, + "reward": 0.6729910969734192, + "reward_std": 0.1473347693681717, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1082 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3704.0, + "completions/mean_length": 1027.419677734375, + "completions/mean_terminated_length": 593.5184936523438, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.12128279883382, + "grad_norm": 0.15662840008735657, + "learning_rate": 1e-06, + "loss": -0.0431, + "num_tokens": 647921952.0, + "reward": 0.6741071939468384, + "reward_std": 0.1445143222808838, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692258834839, + "step": 1083 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4008.0, + "completions/mean_length": 1005.4844360351562, + "completions/mean_terminated_length": 608.4659423828125, + "completions/min_length": 84.0, + "completions/min_terminated_length": 84.0, + "epoch": 10.130612244897959, + "grad_norm": 0.1450105905532837, + "learning_rate": 1e-06, + "loss": -0.0865, + "num_tokens": 648490458.0, + "reward": 0.637276828289032, + "reward_std": 0.13745200634002686, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1084 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2900.0, + "completions/mean_length": 1062.3460693359375, + "completions/mean_terminated_length": 575.0751342773438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 10.139941690962099, + "grad_norm": 0.16196465492248535, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 649029728.0, + "reward": 0.6875000596046448, + "reward_std": 0.10694669187068939, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1085 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3723.0, + "completions/mean_length": 950.396240234375, + "completions/mean_terminated_length": 612.117431640625, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 10.14927113702624, + "grad_norm": 0.15659761428833008, + "learning_rate": 1e-06, + "loss": -0.0503, + "num_tokens": 649621331.0, + "reward": 0.6729910969734192, + "reward_std": 0.15413284301757812, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1086 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2389.0, + "completions/mean_length": 931.1886596679688, + "completions/mean_terminated_length": 590.84423828125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 10.15860058309038, + "grad_norm": 0.15611431002616882, + "learning_rate": 1e-06, + "loss": -0.0759, + "num_tokens": 650192036.0, + "reward": 0.738839328289032, + "reward_std": 0.1438724249601364, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1087 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3513.0, + "completions/mean_length": 1146.32373046875, + "completions/mean_terminated_length": 676.9702758789062, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 10.167930029154519, + "grad_norm": 0.16885007917881012, + "learning_rate": 1e-06, + "loss": -0.0899, + "num_tokens": 650807894.0, + "reward": 0.6439732313156128, + "reward_std": 0.17878074944019318, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1088 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3530.0, + "completions/mean_length": 1063.1707763671875, + "completions/mean_terminated_length": 580.5860595703125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 10.177259475218658, + "grad_norm": 0.17859436571598053, + "learning_rate": 1e-06, + "loss": -0.0735, + "num_tokens": 651337183.0, + "reward": 0.6785714626312256, + "reward_std": 0.15713873505592346, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1089 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2721.0, + "completions/mean_length": 950.9676513671875, + "completions/mean_terminated_length": 537.9835815429688, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 10.186588921282798, + "grad_norm": 0.14909125864505768, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 651855794.0, + "reward": 0.6819196939468384, + "reward_std": 0.1301603466272354, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1090 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2778.0, + "completions/mean_length": 991.0067138671875, + "completions/mean_terminated_length": 596.5358276367188, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.19591836734694, + "grad_norm": 0.1429533213376999, + "learning_rate": 1e-06, + "loss": -0.0525, + "num_tokens": 652429760.0, + "reward": 0.6975446939468384, + "reward_std": 0.14165818691253662, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1091 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3358.0, + "completions/mean_length": 1049.046875, + "completions/mean_terminated_length": 586.91259765625, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 10.205247813411079, + "grad_norm": 0.13117572665214539, + "learning_rate": 1e-06, + "loss": -0.0638, + "num_tokens": 652979698.0, + "reward": 0.6975446939468384, + "reward_std": 0.0988972932100296, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1092 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3779.0, + "completions/mean_length": 953.2433471679688, + "completions/mean_terminated_length": 632.3960571289062, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.214577259475218, + "grad_norm": 0.12957538664340973, + "learning_rate": 1e-06, + "loss": -0.0197, + "num_tokens": 653586212.0, + "reward": 0.6729910969734192, + "reward_std": 0.12693016231060028, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.4693825840950012, + "step": 1093 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 1091.7890625, + "completions/mean_terminated_length": 675.7039184570312, + "completions/min_length": 195.0, + "completions/min_terminated_length": 195.0, + "epoch": 10.223906705539358, + "grad_norm": 0.14532352983951569, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 654207143.0, + "reward": 0.6149553656578064, + "reward_std": 0.14244982600212097, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1094 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3427.0, + "completions/mean_length": 950.9263916015625, + "completions/mean_terminated_length": 604.0718994140625, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 10.2332361516035, + "grad_norm": 0.161069855093956, + "learning_rate": 1e-06, + "loss": -0.0356, + "num_tokens": 654790053.0, + "reward": 0.6540178656578064, + "reward_std": 0.15225742757320404, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1095 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3768.0, + "completions/mean_length": 1132.7388916015625, + "completions/mean_terminated_length": 602.4710693359375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.242565597667639, + "grad_norm": 0.2017541378736496, + "learning_rate": 1e-06, + "loss": -0.0943, + "num_tokens": 655351443.0, + "reward": 0.6116071939468384, + "reward_std": 0.19918353855609894, + "rewards/verify_math_reward/mean": 0.6116071343421936, + "rewards/verify_math_reward/std": 0.48765692114830017, + "step": 1096 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3878.0, + "completions/mean_length": 1035.774658203125, + "completions/mean_terminated_length": 594.1328125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 10.251895043731778, + "grad_norm": 0.13807076215744019, + "learning_rate": 1e-06, + "loss": -0.0381, + "num_tokens": 655906361.0, + "reward": 0.6339285969734192, + "reward_std": 0.1182820275425911, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1097 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3819.0, + "completions/mean_length": 1041.732177734375, + "completions/mean_terminated_length": 623.1268920898438, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 10.261224489795918, + "grad_norm": 0.15223170816898346, + "learning_rate": 1e-06, + "loss": -0.0564, + "num_tokens": 656480185.0, + "reward": 0.65625, + "reward_std": 0.1367020159959793, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1098 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 1003.11279296875, + "completions/mean_terminated_length": 596.9760131835938, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.270553935860057, + "grad_norm": 0.170912966132164, + "learning_rate": 1e-06, + "loss": -0.0807, + "num_tokens": 657050430.0, + "reward": 0.6707589626312256, + "reward_std": 0.155858114361763, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1099 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3088.0, + "completions/mean_length": 1038.03466796875, + "completions/mean_terminated_length": 601.182373046875, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 10.279883381924199, + "grad_norm": 0.14294150471687317, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 657612541.0, + "reward": 0.6808035969734192, + "reward_std": 0.12828493118286133, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.46642565727233887, + "step": 1100 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2602.0, + "completions/mean_length": 913.2188110351562, + "completions/mean_terminated_length": 613.98291015625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 10.289212827988338, + "grad_norm": 0.1634213775396347, + "learning_rate": 1e-06, + "loss": -0.0526, + "num_tokens": 658199929.0, + "reward": 0.7176339626312256, + "reward_std": 0.16521771252155304, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1101 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2329.0, + "completions/mean_length": 1010.7522583007812, + "completions/mean_terminated_length": 605.6187133789062, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 10.298542274052478, + "grad_norm": 0.14884184300899506, + "learning_rate": 1e-06, + "loss": -0.0753, + "num_tokens": 658769579.0, + "reward": 0.668526828289032, + "reward_std": 0.12005008012056351, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1102 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3329.0, + "completions/mean_length": 1022.1763916015625, + "completions/mean_terminated_length": 609.7392578125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 10.307871720116617, + "grad_norm": 0.17236760258674622, + "learning_rate": 1e-06, + "loss": -0.1071, + "num_tokens": 659337761.0, + "reward": 0.7142857313156128, + "reward_std": 0.18039767444133759, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1103 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3994.0, + "completions/mean_length": 1079.430908203125, + "completions/mean_terminated_length": 608.456787109375, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 10.317201166180759, + "grad_norm": 0.16065767407417297, + "learning_rate": 1e-06, + "loss": -0.0648, + "num_tokens": 659902907.0, + "reward": 0.7008928656578064, + "reward_std": 0.14176876842975616, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1104 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3165.0, + "completions/mean_length": 1116.1976318359375, + "completions/mean_terminated_length": 624.0845336914062, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.326530612244898, + "grad_norm": 0.16376863420009613, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 660478772.0, + "reward": 0.6506696939468384, + "reward_std": 0.1612725555896759, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1105 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3569.0, + "completions/mean_length": 992.3873291015625, + "completions/mean_terminated_length": 558.0394287109375, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 10.335860058309038, + "grad_norm": 0.15680444240570068, + "learning_rate": 1e-06, + "loss": -0.0845, + "num_tokens": 661005375.0, + "reward": 0.637276828289032, + "reward_std": 0.1473347544670105, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1106 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 938.7277221679688, + "completions/mean_terminated_length": 624.9374389648438, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.345189504373177, + "grad_norm": 0.16408175230026245, + "learning_rate": 1e-06, + "loss": -0.029, + "num_tokens": 661595835.0, + "reward": 0.6930803656578064, + "reward_std": 0.1510867178440094, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1107 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3986.0, + "completions/mean_length": 1047.3817138671875, + "completions/mean_terminated_length": 584.994873046875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 10.354518950437317, + "grad_norm": 0.1591675579547882, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 662144457.0, + "reward": 0.676339328289032, + "reward_std": 0.12456367164850235, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1108 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4006.0, + "completions/mean_length": 1095.6842041015625, + "completions/mean_terminated_length": 658.2979736328125, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 10.363848396501458, + "grad_norm": 0.1713409572839737, + "learning_rate": 1e-06, + "loss": -0.0721, + "num_tokens": 662742718.0, + "reward": 0.6428571939468384, + "reward_std": 0.1583433598279953, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1109 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3252.0, + "completions/mean_length": 1098.9788818359375, + "completions/mean_terminated_length": 648.8485717773438, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 10.373177842565598, + "grad_norm": 0.1908992975950241, + "learning_rate": 1e-06, + "loss": -0.0856, + "num_tokens": 663331555.0, + "reward": 0.7008928656578064, + "reward_std": 0.14473934471607208, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.4581226110458374, + "step": 1110 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2726.0, + "completions/mean_length": 937.3605346679688, + "completions/mean_terminated_length": 549.4573974609375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 10.382507288629737, + "grad_norm": 0.18250666558742523, + "learning_rate": 1e-06, + "loss": -0.082, + "num_tokens": 663870742.0, + "reward": 0.6819196939468384, + "reward_std": 0.16360512375831604, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1111 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3887.0, + "completions/mean_length": 970.544677734375, + "completions/mean_terminated_length": 608.56787109375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 10.391836734693877, + "grad_norm": 0.15100763738155365, + "learning_rate": 1e-06, + "loss": -0.0338, + "num_tokens": 664436438.0, + "reward": 0.723214328289032, + "reward_std": 0.11881405115127563, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1112 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2507.0, + "completions/mean_length": 907.5725708007812, + "completions/mean_terminated_length": 569.0481567382812, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 10.401166180758018, + "grad_norm": 0.1526399403810501, + "learning_rate": 1e-06, + "loss": -0.0435, + "num_tokens": 664989471.0, + "reward": 0.7165178656578064, + "reward_std": 0.11930260062217712, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1113 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4039.0, + "completions/mean_length": 913.7489013671875, + "completions/mean_terminated_length": 588.86962890625, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.410495626822158, + "grad_norm": 0.19045433402061462, + "learning_rate": 1e-06, + "loss": -0.0419, + "num_tokens": 665548622.0, + "reward": 0.738839328289032, + "reward_std": 0.16442787647247314, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1114 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2890.0, + "completions/mean_length": 996.4788208007812, + "completions/mean_terminated_length": 633.1932983398438, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 10.419825072886297, + "grad_norm": 0.17184984683990479, + "learning_rate": 1e-06, + "loss": -0.0511, + "num_tokens": 666139563.0, + "reward": 0.6886160969734192, + "reward_std": 0.17735858261585236, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1115 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2909.0, + "completions/mean_length": 845.7522583007812, + "completions/mean_terminated_length": 566.033935546875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 10.429154518950437, + "grad_norm": 0.16167055070400238, + "learning_rate": 1e-06, + "loss": -0.0522, + "num_tokens": 666687669.0, + "reward": 0.7589285969734192, + "reward_std": 0.12343572080135345, + "rewards/verify_math_reward/mean": 0.7589285969734192, + "rewards/verify_math_reward/std": 0.4279724657535553, + "step": 1116 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3058.0, + "completions/mean_length": 995.31591796875, + "completions/mean_terminated_length": 592.5787963867188, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 10.438483965014576, + "grad_norm": 0.16173385083675385, + "learning_rate": 1e-06, + "loss": -0.0659, + "num_tokens": 667245016.0, + "reward": 0.6908482313156128, + "reward_std": 0.13665924966335297, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1117 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3224.0, + "completions/mean_length": 1032.4320068359375, + "completions/mean_terminated_length": 630.1452026367188, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 10.447813411078718, + "grad_norm": 0.15724092721939087, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 667840891.0, + "reward": 0.676339328289032, + "reward_std": 0.11817464977502823, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1118 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 968.7824096679688, + "completions/mean_terminated_length": 580.33251953125, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.457142857142857, + "grad_norm": 0.1605788618326187, + "learning_rate": 1e-06, + "loss": -0.0746, + "num_tokens": 668399280.0, + "reward": 0.6852678656578064, + "reward_std": 0.1488724648952484, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1119 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3397.0, + "completions/mean_length": 971.958740234375, + "completions/mean_terminated_length": 583.9033813476562, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.466472303206997, + "grad_norm": 0.16461539268493652, + "learning_rate": 1e-06, + "loss": -0.0483, + "num_tokens": 668950571.0, + "reward": 0.7109375596046448, + "reward_std": 0.12910795211791992, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1120 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3703.0, + "completions/mean_length": 1040.4554443359375, + "completions/mean_terminated_length": 630.4708862304688, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.475801749271136, + "grad_norm": 0.15067552030086517, + "learning_rate": 1e-06, + "loss": -0.0433, + "num_tokens": 669534091.0, + "reward": 0.6484375, + "reward_std": 0.13850216567516327, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1121 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1110.50341796875, + "completions/mean_terminated_length": 653.26513671875, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 10.485131195335278, + "grad_norm": 0.15959958732128143, + "learning_rate": 1e-06, + "loss": -0.0743, + "num_tokens": 670124910.0, + "reward": 0.6595982313156128, + "reward_std": 0.1651000678539276, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1122 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3825.0, + "completions/mean_length": 913.5547485351562, + "completions/mean_terminated_length": 592.96435546875, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.494460641399417, + "grad_norm": 0.16357047855854034, + "learning_rate": 1e-06, + "loss": -0.046, + "num_tokens": 670693639.0, + "reward": 0.6819196939468384, + "reward_std": 0.13369613885879517, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1123 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3675.0, + "completions/mean_length": 1019.4185791015625, + "completions/mean_terminated_length": 632.913330078125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 10.503790087463557, + "grad_norm": 0.12714940309524536, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 671297502.0, + "reward": 0.6618303656578064, + "reward_std": 0.10983241349458694, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1124 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4043.0, + "completions/mean_length": 797.7957763671875, + "completions/mean_terminated_length": 582.0986938476562, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 10.513119533527696, + "grad_norm": 0.13335968554019928, + "learning_rate": 1e-06, + "loss": -0.0358, + "num_tokens": 671875903.0, + "reward": 0.7455357313156128, + "reward_std": 0.10926789045333862, + "rewards/verify_math_reward/mean": 0.7455357313156128, + "rewards/verify_math_reward/std": 0.4358029067516327, + "step": 1125 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3414.0, + "completions/mean_length": 1176.50341796875, + "completions/mean_terminated_length": 667.6002807617188, + "completions/min_length": 180.0, + "completions/min_terminated_length": 180.0, + "epoch": 10.522448979591836, + "grad_norm": 0.1646053045988083, + "learning_rate": 1e-06, + "loss": -0.0695, + "num_tokens": 672479482.0, + "reward": 0.559151828289032, + "reward_std": 0.15683633089065552, + "rewards/verify_math_reward/mean": 0.5591517686843872, + "rewards/verify_math_reward/std": 0.496766060590744, + "step": 1126 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3404.0, + "completions/mean_length": 1037.4810791015625, + "completions/mean_terminated_length": 657.5646362304688, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 10.531778425655977, + "grad_norm": 0.16377611458301544, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 673088849.0, + "reward": 0.65625, + "reward_std": 0.16367970407009125, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1127 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2473.0, + "completions/mean_length": 1054.341552734375, + "completions/mean_terminated_length": 542.7718505859375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 10.541107871720117, + "grad_norm": 0.18253152072429657, + "learning_rate": 1e-06, + "loss": -0.0806, + "num_tokens": 673599419.0, + "reward": 0.6875000596046448, + "reward_std": 0.14075706899166107, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1128 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2459.0, + "completions/mean_length": 881.62841796875, + "completions/mean_terminated_length": 570.81396484375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 10.550437317784256, + "grad_norm": 0.16114667057991028, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 674154350.0, + "reward": 0.7265625596046448, + "reward_std": 0.1346297711133957, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1129 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3397.0, + "completions/mean_length": 1132.216552734375, + "completions/mean_terminated_length": 730.2838745117188, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 10.559766763848396, + "grad_norm": 0.1421719491481781, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 674829560.0, + "reward": 0.5792410969734192, + "reward_std": 0.14838463068008423, + "rewards/verify_math_reward/mean": 0.5792410969734192, + "rewards/verify_math_reward/std": 0.49395665526390076, + "step": 1130 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2863.0, + "completions/mean_length": 1135.680908203125, + "completions/mean_terminated_length": 619.661865234375, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 10.569096209912537, + "grad_norm": 0.1621275693178177, + "learning_rate": 1e-06, + "loss": -0.0549, + "num_tokens": 675390770.0, + "reward": 0.6339285969734192, + "reward_std": 0.15462210774421692, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 1131 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3329.0, + "completions/mean_length": 1011.8404541015625, + "completions/mean_terminated_length": 593.5830078125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 10.578425655976677, + "grad_norm": 0.14411039650440216, + "learning_rate": 1e-06, + "loss": -0.0374, + "num_tokens": 675953083.0, + "reward": 0.6428571939468384, + "reward_std": 0.12613528966903687, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1132 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3040.0, + "completions/mean_length": 856.5413208007812, + "completions/mean_terminated_length": 556.2987670898438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 10.587755102040816, + "grad_norm": 0.1527402102947235, + "learning_rate": 1e-06, + "loss": -0.0397, + "num_tokens": 676504768.0, + "reward": 0.7098214626312256, + "reward_std": 0.13662534952163696, + "rewards/verify_math_reward/mean": 0.7098214030265808, + "rewards/verify_math_reward/std": 0.454098105430603, + "step": 1133 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3657.0, + "completions/mean_length": 1003.4319458007812, + "completions/mean_terminated_length": 675.085205078125, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 10.597084548104956, + "grad_norm": 0.2829720377922058, + "learning_rate": 1e-06, + "loss": -0.0507, + "num_tokens": 677134315.0, + "reward": 0.6383928656578064, + "reward_std": 0.133991077542305, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1134 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3831.0, + "completions/mean_length": 1006.8136596679688, + "completions/mean_terminated_length": 605.5700073242188, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 10.606413994169095, + "grad_norm": 0.1647486388683319, + "learning_rate": 1e-06, + "loss": -0.0625, + "num_tokens": 677697212.0, + "reward": 0.6830357313156128, + "reward_std": 0.1359538733959198, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1135 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3789.0, + "completions/mean_length": 981.1629638671875, + "completions/mean_terminated_length": 654.7003784179688, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.615743440233237, + "grad_norm": 0.1767268031835556, + "learning_rate": 1e-06, + "loss": -0.0551, + "num_tokens": 678312286.0, + "reward": 0.637276828289032, + "reward_std": 0.17979854345321655, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1136 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2833.0, + "completions/mean_length": 1021.74560546875, + "completions/mean_terminated_length": 618.0555419921875, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 10.625072886297376, + "grad_norm": 0.15197418630123138, + "learning_rate": 1e-06, + "loss": -0.0373, + "num_tokens": 678897970.0, + "reward": 0.6796875596046448, + "reward_std": 0.1529296189546585, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1137 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 908.6797485351562, + "completions/mean_terminated_length": 583.2828979492188, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 10.634402332361516, + "grad_norm": 0.1528334617614746, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 679464499.0, + "reward": 0.6886160969734192, + "reward_std": 0.15582603216171265, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1138 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 983.0670166015625, + "completions/mean_terminated_length": 626.8606567382812, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 10.643731778425655, + "grad_norm": 0.1498451828956604, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 680060471.0, + "reward": 0.6707589626312256, + "reward_std": 0.13936908543109894, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1139 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3787.0, + "completions/mean_length": 1030.6763916015625, + "completions/mean_terminated_length": 597.2356567382812, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.653061224489797, + "grad_norm": 0.17636732757091522, + "learning_rate": 1e-06, + "loss": -0.0846, + "num_tokens": 680617181.0, + "reward": 0.6752232313156128, + "reward_std": 0.17829473316669464, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1140 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3332.0, + "completions/mean_length": 917.3582763671875, + "completions/mean_terminated_length": 549.2216796875, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 10.662390670553936, + "grad_norm": 0.15136787295341492, + "learning_rate": 1e-06, + "loss": -0.0255, + "num_tokens": 681149422.0, + "reward": 0.65625, + "reward_std": 0.1269715279340744, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1141 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 1116.4207763671875, + "completions/mean_terminated_length": 633.3502197265625, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 10.671720116618076, + "grad_norm": 0.16474510729312897, + "learning_rate": 1e-06, + "loss": -0.0819, + "num_tokens": 681726839.0, + "reward": 0.6205357313156128, + "reward_std": 0.15233299136161804, + "rewards/verify_math_reward/mean": 0.6205357313156128, + "rewards/verify_math_reward/std": 0.4855247139930725, + "step": 1142 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 1035.0201416015625, + "completions/mean_terminated_length": 633.0732421875, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 10.681049562682215, + "grad_norm": 0.16252972185611725, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 682316929.0, + "reward": 0.6674107313156128, + "reward_std": 0.13820087909698486, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1143 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3644.0, + "completions/mean_length": 1051.094970703125, + "completions/mean_terminated_length": 593.7728271484375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 10.690379008746355, + "grad_norm": 0.17688776552677155, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 682865790.0, + "reward": 0.6975446939468384, + "reward_std": 0.1557493507862091, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1144 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3560.0, + "completions/mean_length": 1163.4710693359375, + "completions/mean_terminated_length": 674.7161865234375, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 10.699708454810496, + "grad_norm": 0.1655026227235794, + "learning_rate": 1e-06, + "loss": -0.0697, + "num_tokens": 683477412.0, + "reward": 0.6261160969734192, + "reward_std": 0.17649024724960327, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1145 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2952.0, + "completions/mean_length": 1000.8303833007812, + "completions/mean_terminated_length": 633.73779296875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 10.709037900874636, + "grad_norm": 0.13796032965183258, + "learning_rate": 1e-06, + "loss": -0.0616, + "num_tokens": 684078644.0, + "reward": 0.625, + "reward_std": 0.15285363793373108, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1146 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3456.0, + "completions/mean_length": 1069.7210693359375, + "completions/mean_terminated_length": 668.0025634765625, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 10.718367346938775, + "grad_norm": 0.13414829969406128, + "learning_rate": 1e-06, + "loss": -0.0482, + "num_tokens": 684695722.0, + "reward": 0.6651785969734192, + "reward_std": 0.11821813136339188, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1147 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2998.0, + "completions/mean_length": 1174.7445068359375, + "completions/mean_terminated_length": 647.4558715820312, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 10.727696793002915, + "grad_norm": 0.1551147997379303, + "learning_rate": 1e-06, + "loss": -0.0657, + "num_tokens": 685273309.0, + "reward": 0.6183035969734192, + "reward_std": 0.14977401494979858, + "rewards/verify_math_reward/mean": 0.6183035969734192, + "rewards/verify_math_reward/std": 0.4860740303993225, + "step": 1148 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3917.0, + "completions/mean_length": 1058.0223388671875, + "completions/mean_terminated_length": 579.16796875, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 10.737026239067056, + "grad_norm": 0.187620609998703, + "learning_rate": 1e-06, + "loss": -0.092, + "num_tokens": 685801049.0, + "reward": 0.6629464626312256, + "reward_std": 0.1698412001132965, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1149 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3799.0, + "completions/mean_length": 967.864990234375, + "completions/mean_terminated_length": 601.2257080078125, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 10.746355685131196, + "grad_norm": 0.16204749047756195, + "learning_rate": 1e-06, + "loss": -0.066, + "num_tokens": 686371344.0, + "reward": 0.6752232313156128, + "reward_std": 0.1654042750597, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1150 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2468.0, + "completions/mean_length": 1024.2857666015625, + "completions/mean_terminated_length": 594.4020385742188, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 10.755685131195335, + "grad_norm": 0.15140943229198456, + "learning_rate": 1e-06, + "loss": -0.0593, + "num_tokens": 686931760.0, + "reward": 0.6875000596046448, + "reward_std": 0.1396651417016983, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1151 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4051.0, + "completions/mean_length": 1098.919677734375, + "completions/mean_terminated_length": 630.9884033203125, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 10.765014577259475, + "grad_norm": 0.15443748235702515, + "learning_rate": 1e-06, + "loss": -0.0502, + "num_tokens": 687507448.0, + "reward": 0.6729910969734192, + "reward_std": 0.1479395478963852, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1152 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0725446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2606.0, + "completions/mean_length": 784.9754638671875, + "completions/mean_terminated_length": 525.9903564453125, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 10.774344023323614, + "grad_norm": 0.15131253004074097, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 688028522.0, + "reward": 0.7444196939468384, + "reward_std": 0.12467285990715027, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1153 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1685267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3310.0, + "completions/mean_length": 1201.69873046875, + "completions/mean_terminated_length": 615.0684814453125, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 10.783673469387756, + "grad_norm": 0.15732906758785248, + "learning_rate": 1e-06, + "loss": -0.0934, + "num_tokens": 688580428.0, + "reward": 0.6417410969734192, + "reward_std": 0.1321566104888916, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975659370422363, + "step": 1154 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3608.0, + "completions/mean_length": 864.677490234375, + "completions/mean_terminated_length": 603.5211181640625, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.793002915451895, + "grad_norm": 0.12360245734453201, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 689159811.0, + "reward": 0.7142857313156128, + "reward_std": 0.12482328712940216, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1155 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3683.0, + "completions/mean_length": 1050.9832763671875, + "completions/mean_terminated_length": 589.1427001953125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 10.802332361516035, + "grad_norm": 0.16396304965019226, + "learning_rate": 1e-06, + "loss": -0.0316, + "num_tokens": 689703260.0, + "reward": 0.668526828289032, + "reward_std": 0.1173504963517189, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1156 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3320.0, + "completions/mean_length": 966.6964721679688, + "completions/mean_terminated_length": 595.5555419921875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 10.811661807580174, + "grad_norm": 0.18031202256679535, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 690270436.0, + "reward": 0.7131696939468384, + "reward_std": 0.15048938989639282, + "rewards/verify_math_reward/mean": 0.7131696343421936, + "rewards/verify_math_reward/std": 0.4525342881679535, + "step": 1157 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 975.5803833007812, + "completions/mean_terminated_length": 631.4448852539062, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 10.820991253644316, + "grad_norm": 0.15054838359355927, + "learning_rate": 1e-06, + "loss": -0.0782, + "num_tokens": 690865604.0, + "reward": 0.668526828289032, + "reward_std": 0.15831056237220764, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056781768799, + "step": 1158 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2717.0, + "completions/mean_length": 1046.4710693359375, + "completions/mean_terminated_length": 619.692138671875, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 10.830320699708455, + "grad_norm": 0.16395598649978638, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 691436554.0, + "reward": 0.6674107313156128, + "reward_std": 0.15398269891738892, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1159 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3587.0, + "completions/mean_length": 905.161865234375, + "completions/mean_terminated_length": 562.0185546875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.839650145772595, + "grad_norm": 0.16268029808998108, + "learning_rate": 1e-06, + "loss": -0.0783, + "num_tokens": 691974971.0, + "reward": 0.715401828289032, + "reward_std": 0.14823377132415771, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1160 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3829.0, + "completions/mean_length": 991.0256958007812, + "completions/mean_terminated_length": 596.5572509765625, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.848979591836734, + "grad_norm": 0.17570465803146362, + "learning_rate": 1e-06, + "loss": -0.0314, + "num_tokens": 692531754.0, + "reward": 0.6383928656578064, + "reward_std": 0.154469832777977, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341992855072, + "step": 1161 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4076.0, + "completions/mean_length": 1057.5145263671875, + "completions/mean_terminated_length": 684.3671875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 10.858309037900874, + "grad_norm": 0.16581018269062042, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 693167999.0, + "reward": 0.6819196939468384, + "reward_std": 0.16856282949447632, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1162 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3331.0, + "completions/mean_length": 935.6864013671875, + "completions/mean_terminated_length": 600.1469116210938, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.867638483965015, + "grad_norm": 0.15151207149028778, + "learning_rate": 1e-06, + "loss": -0.0481, + "num_tokens": 693740886.0, + "reward": 0.6662946939468384, + "reward_std": 0.13639962673187256, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1163 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 1088.7310791015625, + "completions/mean_terminated_length": 596.6324462890625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 10.876967930029155, + "grad_norm": 0.1692202389240265, + "learning_rate": 1e-06, + "loss": -0.0681, + "num_tokens": 694290349.0, + "reward": 0.660714328289032, + "reward_std": 0.14905934035778046, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1164 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1350446428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3327.0, + "completions/mean_length": 1087.6942138671875, + "completions/mean_terminated_length": 618.0103759765625, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 10.886297376093294, + "grad_norm": 0.17742066085338593, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 694863347.0, + "reward": 0.6272321939468384, + "reward_std": 0.1226440817117691, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1165 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3511.0, + "completions/mean_length": 969.6451416015625, + "completions/mean_terminated_length": 624.8550415039062, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 10.895626822157434, + "grad_norm": 0.16102254390716553, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 695457653.0, + "reward": 0.6752232313156128, + "reward_std": 0.1586022824048996, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1166 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3738.0, + "completions/mean_length": 889.5491333007812, + "completions/mean_terminated_length": 583.799560546875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.904956268221575, + "grad_norm": 0.16921687126159668, + "learning_rate": 1e-06, + "loss": -0.0681, + "num_tokens": 696026377.0, + "reward": 0.6964285969734192, + "reward_std": 0.15319104492664337, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1167 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 936.7053833007812, + "completions/mean_terminated_length": 601.2741088867188, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 10.914285714285715, + "grad_norm": 0.18251405656337738, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 696604641.0, + "reward": 0.7087053656578064, + "reward_std": 0.16506798565387726, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1168 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 973.3136596679688, + "completions/mean_terminated_length": 602.95751953125, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 10.923615160349854, + "grad_norm": 0.14678320288658142, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 697171330.0, + "reward": 0.707589328289032, + "reward_std": 0.1447407603263855, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1169 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3230.0, + "completions/mean_length": 1044.9554443359375, + "completions/mean_terminated_length": 626.7918701171875, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 10.932944606413994, + "grad_norm": 0.1669425070285797, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 697755474.0, + "reward": 0.6484375, + "reward_std": 0.12039776146411896, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1170 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1529017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2720.0, + "completions/mean_length": 1140.5457763671875, + "completions/mean_terminated_length": 607.0843505859375, + "completions/min_length": 185.0, + "completions/min_terminated_length": 185.0, + "epoch": 10.942274052478133, + "grad_norm": 0.15743738412857056, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 698307931.0, + "reward": 0.5870535969734192, + "reward_std": 0.14053094387054443, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263834953308105, + "step": 1171 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3536.0, + "completions/mean_length": 982.122802734375, + "completions/mean_terminated_length": 647.255859375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 10.951603498542275, + "grad_norm": 0.1489836871623993, + "learning_rate": 1e-06, + "loss": -0.041, + "num_tokens": 698926809.0, + "reward": 0.6618303656578064, + "reward_std": 0.13298118114471436, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1172 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2894.0, + "completions/mean_length": 1051.3973388671875, + "completions/mean_terminated_length": 580.5824584960938, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 10.960932944606414, + "grad_norm": 0.1498366743326187, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 699468549.0, + "reward": 0.6774553656578064, + "reward_std": 0.1379726529121399, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1173 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1484375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4035.0, + "completions/mean_length": 1150.0703125, + "completions/mean_terminated_length": 636.5596313476562, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 10.970262390670554, + "grad_norm": 0.15563355386257172, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 700050964.0, + "reward": 0.6629464626312256, + "reward_std": 0.14838533103466034, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1174 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3368.0, + "completions/mean_length": 889.6395263671875, + "completions/mean_terminated_length": 596.7320556640625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 10.979591836734693, + "grad_norm": 0.16050517559051514, + "learning_rate": 1e-06, + "loss": -0.0535, + "num_tokens": 700629441.0, + "reward": 0.731026828289032, + "reward_std": 0.14725808799266815, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1175 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3690.0, + "completions/mean_length": 1048.3070068359375, + "completions/mean_terminated_length": 617.3592529296875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 10.988921282798835, + "grad_norm": 0.16893556714057922, + "learning_rate": 1e-06, + "loss": -0.092, + "num_tokens": 701211340.0, + "reward": 0.6261160969734192, + "reward_std": 0.17867198586463928, + "rewards/verify_math_reward/mean": 0.6261160969734192, + "rewards/verify_math_reward/std": 0.48410359025001526, + "step": 1176 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.11647727272727271, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 1694.0, + "completions/mean_length": 1036.84375, + "completions/mean_terminated_length": 633.546630859375, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 10.998250728862974, + "grad_norm": 0.15162187814712524, + "learning_rate": 1e-06, + "loss": -0.0475, + "num_tokens": 701804960.0, + "reward": 0.6149553656578064, + "reward_std": 0.1310618817806244, + "rewards/verify_math_reward/mean": 0.6149553656578064, + "rewards/verify_math_reward/std": 0.4868776500225067, + "step": 1177 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4046.0, + "completions/mean_length": 1020.2020263671875, + "completions/mean_terminated_length": 680.9876098632812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 11.00932944606414, + "grad_norm": 0.17165175080299377, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 702446725.0, + "reward": 0.6774553656578064, + "reward_std": 0.17525240778923035, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1178 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3556.0, + "completions/mean_length": 921.9933471679688, + "completions/mean_terminated_length": 532.2030029296875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 11.018658892128279, + "grad_norm": 0.20352905988693237, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 702957407.0, + "reward": 0.715401828289032, + "reward_std": 0.1432783305644989, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1179 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 913.435302734375, + "completions/mean_terminated_length": 592.8328857421875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.02798833819242, + "grad_norm": 0.21608437597751617, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 703536645.0, + "reward": 0.6886160969734192, + "reward_std": 0.16552191972732544, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1180 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2931.0, + "completions/mean_length": 1015.05810546875, + "completions/mean_terminated_length": 632.3563232421875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.03731778425656, + "grad_norm": 0.17518115043640137, + "learning_rate": 1e-06, + "loss": -0.0686, + "num_tokens": 704131033.0, + "reward": 0.6752232313156128, + "reward_std": 0.1881735622882843, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1181 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3830.0, + "completions/mean_length": 1024.4520263671875, + "completions/mean_terminated_length": 655.8662109375, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 11.0466472303207, + "grad_norm": 0.17708617448806763, + "learning_rate": 1e-06, + "loss": -0.0651, + "num_tokens": 704761574.0, + "reward": 0.6127232313156128, + "reward_std": 0.15559709072113037, + "rewards/verify_math_reward/mean": 0.6127232313156128, + "rewards/verify_math_reward/std": 0.4873998463153839, + "step": 1182 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3168.0, + "completions/mean_length": 938.4598388671875, + "completions/mean_terminated_length": 585.880859375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.055976676384839, + "grad_norm": 0.17927464842796326, + "learning_rate": 1e-06, + "loss": -0.0987, + "num_tokens": 705318434.0, + "reward": 0.7265625596046448, + "reward_std": 0.1446644812822342, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1183 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3868.0, + "completions/mean_length": 876.9944458007812, + "completions/mean_terminated_length": 539.614013671875, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 11.06530612244898, + "grad_norm": 0.1659204363822937, + "learning_rate": 1e-06, + "loss": -0.0307, + "num_tokens": 705840589.0, + "reward": 0.715401828289032, + "reward_std": 0.1288815587759018, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1184 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3338.0, + "completions/mean_length": 1027.6741943359375, + "completions/mean_terminated_length": 611.5640258789062, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.07463556851312, + "grad_norm": 0.14054636657238007, + "learning_rate": 1e-06, + "loss": -0.0496, + "num_tokens": 706415521.0, + "reward": 0.6506696939468384, + "reward_std": 0.1422245353460312, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1185 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2627.0, + "completions/mean_length": 848.0647583007812, + "completions/mean_terminated_length": 581.3260498046875, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 11.08396501457726, + "grad_norm": 0.15509085357189178, + "learning_rate": 1e-06, + "loss": -0.0426, + "num_tokens": 706979779.0, + "reward": 0.7220982313156128, + "reward_std": 0.1411721557378769, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1186 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3046.0, + "completions/mean_length": 892.4051513671875, + "completions/mean_terminated_length": 616.7018432617188, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.093294460641399, + "grad_norm": 0.1560228317975998, + "learning_rate": 1e-06, + "loss": -0.0498, + "num_tokens": 707576582.0, + "reward": 0.6975446939468384, + "reward_std": 0.15808051824569702, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957791805267334, + "step": 1187 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3597.0, + "completions/mean_length": 1105.786865234375, + "completions/mean_terminated_length": 656.6790771484375, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 11.102623906705539, + "grad_norm": 0.1628333032131195, + "learning_rate": 1e-06, + "loss": -0.0509, + "num_tokens": 708168247.0, + "reward": 0.6495535969734192, + "reward_std": 0.1441766321659088, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1188 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2452.0, + "completions/mean_length": 1100.5023193359375, + "completions/mean_terminated_length": 605.797119140625, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.11195335276968, + "grad_norm": 0.18043670058250427, + "learning_rate": 1e-06, + "loss": -0.0642, + "num_tokens": 708739617.0, + "reward": 0.6071428656578064, + "reward_std": 0.14564089477062225, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 1189 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1573660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3841.0, + "completions/mean_length": 1123.8817138671875, + "completions/mean_terminated_length": 568.8238525390625, + "completions/min_length": 175.0, + "completions/min_terminated_length": 175.0, + "epoch": 11.12128279883382, + "grad_norm": 0.16629146039485931, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 709261255.0, + "reward": 0.6316964626312256, + "reward_std": 0.12009535729885101, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1190 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3995.0, + "completions/mean_length": 1060.3148193359375, + "completions/mean_terminated_length": 687.5112915039062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.130612244897959, + "grad_norm": 0.15539808571338654, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 709902841.0, + "reward": 0.6618303656578064, + "reward_std": 0.16848503053188324, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1191 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2600.0, + "completions/mean_length": 1105.2176513671875, + "completions/mean_terminated_length": 602.2047119140625, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.139941690962099, + "grad_norm": 0.15172308683395386, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 710453500.0, + "reward": 0.6696428656578064, + "reward_std": 0.1259191483259201, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1192 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2918.0, + "completions/mean_length": 1015.0982666015625, + "completions/mean_terminated_length": 588.391357421875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 11.14927113702624, + "grad_norm": 0.17290890216827393, + "learning_rate": 1e-06, + "loss": -0.0856, + "num_tokens": 711011356.0, + "reward": 0.6785714626312256, + "reward_std": 0.1471807062625885, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1193 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3223.0, + "completions/mean_length": 1098.9320068359375, + "completions/mean_terminated_length": 613.0259399414062, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 11.15860058309038, + "grad_norm": 0.169038787484169, + "learning_rate": 1e-06, + "loss": -0.0839, + "num_tokens": 711575199.0, + "reward": 0.6785714626312256, + "reward_std": 0.18825092911720276, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1194 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3499.0, + "completions/mean_length": 950.8326416015625, + "completions/mean_terminated_length": 586.5728149414062, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 11.167930029154519, + "grad_norm": 0.16237430274486542, + "learning_rate": 1e-06, + "loss": -0.0544, + "num_tokens": 712143641.0, + "reward": 0.6718750596046448, + "reward_std": 0.14312425255775452, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1195 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3027.0, + "completions/mean_length": 888.3984985351562, + "completions/mean_terminated_length": 595.3776245117188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.177259475218658, + "grad_norm": 0.13331414759159088, + "learning_rate": 1e-06, + "loss": -0.0562, + "num_tokens": 712711902.0, + "reward": 0.6875000596046448, + "reward_std": 0.12726575136184692, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1196 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0669642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3703.0, + "completions/mean_length": 842.044677734375, + "completions/mean_terminated_length": 608.5071411132812, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 11.186588921282798, + "grad_norm": 0.1372644603252411, + "learning_rate": 1e-06, + "loss": -0.0371, + "num_tokens": 713305966.0, + "reward": 0.7444196939468384, + "reward_std": 0.12418459355831146, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1197 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4060.0, + "completions/mean_length": 1070.2445068359375, + "completions/mean_terminated_length": 646.7926025390625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.19591836734694, + "grad_norm": 0.1399068981409073, + "learning_rate": 1e-06, + "loss": -0.0411, + "num_tokens": 713904097.0, + "reward": 0.6819196939468384, + "reward_std": 0.131517231464386, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1198 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3031.0, + "completions/mean_length": 820.0089721679688, + "completions/mean_terminated_length": 568.0096435546875, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 11.205247813411079, + "grad_norm": 0.15427544713020325, + "learning_rate": 1e-06, + "loss": -0.0554, + "num_tokens": 714448353.0, + "reward": 0.746651828289032, + "reward_std": 0.11242572218179703, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1199 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4057.0, + "completions/mean_length": 823.9688110351562, + "completions/mean_terminated_length": 584.9341430664062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.214577259475218, + "grad_norm": 0.153758704662323, + "learning_rate": 1e-06, + "loss": -0.043, + "num_tokens": 715022357.0, + "reward": 0.7611607313156128, + "reward_std": 0.1414310783147812, + "rewards/verify_math_reward/mean": 0.7611607313156128, + "rewards/verify_math_reward/std": 0.4266124963760376, + "step": 1200 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3870.0, + "completions/mean_length": 1096.4921875, + "completions/mean_terminated_length": 681.0584106445312, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.223906705539358, + "grad_norm": 0.14793723821640015, + "learning_rate": 1e-06, + "loss": -0.0695, + "num_tokens": 715657806.0, + "reward": 0.625, + "reward_std": 0.14150916039943695, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1201 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3950.0, + "completions/mean_length": 1082.2489013671875, + "completions/mean_terminated_length": 703.6369018554688, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 11.2332361516035, + "grad_norm": 0.16762900352478027, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 716317781.0, + "reward": 0.6395089626312256, + "reward_std": 0.1716410517692566, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111421108246, + "step": 1202 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3839.0, + "completions/mean_length": 961.7176513671875, + "completions/mean_terminated_length": 585.603759765625, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 11.242565597667639, + "grad_norm": 0.1698824018239975, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 716869896.0, + "reward": 0.7142857313156128, + "reward_std": 0.12944495677947998, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1203 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 980.1663208007812, + "completions/mean_terminated_length": 597.52001953125, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.251895043731778, + "grad_norm": 0.17149057984352112, + "learning_rate": 1e-06, + "loss": -0.0981, + "num_tokens": 717444101.0, + "reward": 0.6785714626312256, + "reward_std": 0.19099396467208862, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1204 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3745.0, + "completions/mean_length": 942.8426513671875, + "completions/mean_terminated_length": 603.7515258789062, + "completions/min_length": 189.0, + "completions/min_terminated_length": 189.0, + "epoch": 11.261224489795918, + "grad_norm": 0.14584524929523468, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 718034664.0, + "reward": 0.6462053656578064, + "reward_std": 0.11468162387609482, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1205 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3418.0, + "completions/mean_length": 1093.130615234375, + "completions/mean_terminated_length": 619.8101196289062, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 11.270553935860057, + "grad_norm": 0.1828000545501709, + "learning_rate": 1e-06, + "loss": -0.0607, + "num_tokens": 718600733.0, + "reward": 0.6272321939468384, + "reward_std": 0.14474186301231384, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1206 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2929.0, + "completions/mean_length": 1004.2935791015625, + "completions/mean_terminated_length": 646.2254028320312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.279883381924199, + "grad_norm": 0.147318035364151, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 719218860.0, + "reward": 0.6785714626312256, + "reward_std": 0.14237496256828308, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1207 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3344.0, + "completions/mean_length": 988.6428833007812, + "completions/mean_terminated_length": 628.7621459960938, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.289212827988338, + "grad_norm": 0.15306153893470764, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 719806932.0, + "reward": 0.7220982313156128, + "reward_std": 0.13493287563323975, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1208 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3811.0, + "completions/mean_length": 955.2332763671875, + "completions/mean_terminated_length": 626.052978515625, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 11.298542274052478, + "grad_norm": 0.17227299511432648, + "learning_rate": 1e-06, + "loss": -0.0383, + "num_tokens": 720407205.0, + "reward": 0.6863839626312256, + "reward_std": 0.15962284803390503, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1209 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3394.0, + "completions/mean_length": 952.4944458007812, + "completions/mean_terminated_length": 588.4271240234375, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 11.307871720116617, + "grad_norm": 0.14727525413036346, + "learning_rate": 1e-06, + "loss": -0.0436, + "num_tokens": 720972152.0, + "reward": 0.6975446939468384, + "reward_std": 0.10844233632087708, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1210 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3596.0, + "completions/mean_length": 962.927490234375, + "completions/mean_terminated_length": 600.0684814453125, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 11.317201166180759, + "grad_norm": 0.1725059300661087, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 721541335.0, + "reward": 0.6785714626312256, + "reward_std": 0.1760779768228531, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1211 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 898.7589721679688, + "completions/mean_terminated_length": 615.1640625, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 11.326530612244898, + "grad_norm": 0.16215938329696655, + "learning_rate": 1e-06, + "loss": -0.0394, + "num_tokens": 722125327.0, + "reward": 0.6908482313156128, + "reward_std": 0.14846271276474, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1212 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2771.0, + "completions/mean_length": 813.8627319335938, + "completions/mean_terminated_length": 540.0205688476562, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 11.335860058309038, + "grad_norm": 0.16653843224048615, + "learning_rate": 1e-06, + "loss": -0.0206, + "num_tokens": 722663300.0, + "reward": 0.7142857313156128, + "reward_std": 0.13278645277023315, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1213 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3393.0, + "completions/mean_length": 957.818115234375, + "completions/mean_terminated_length": 607.4007568359375, + "completions/min_length": 187.0, + "completions/min_terminated_length": 187.0, + "epoch": 11.345189504373177, + "grad_norm": 0.18848681449890137, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 723247585.0, + "reward": 0.6796875596046448, + "reward_std": 0.13139817118644714, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1214 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4041.0, + "completions/mean_length": 1083.7098388671875, + "completions/mean_terminated_length": 604.3932495117188, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 11.354518950437317, + "grad_norm": 0.16831067204475403, + "learning_rate": 1e-06, + "loss": -0.0699, + "num_tokens": 723808629.0, + "reward": 0.6629464626312256, + "reward_std": 0.13534656167030334, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1215 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2446.0, + "completions/mean_length": 1016.0201416015625, + "completions/mean_terminated_length": 557.9718017578125, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 11.363848396501458, + "grad_norm": 0.15949741005897522, + "learning_rate": 1e-06, + "loss": -0.0511, + "num_tokens": 724326551.0, + "reward": 0.6707589626312256, + "reward_std": 0.11205482482910156, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1216 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2720.0, + "completions/mean_length": 1077.0023193359375, + "completions/mean_terminated_length": 619.1079711914062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 11.373177842565598, + "grad_norm": 0.1645112782716751, + "learning_rate": 1e-06, + "loss": -0.0548, + "num_tokens": 724895905.0, + "reward": 0.652901828289032, + "reward_std": 0.12520237267017365, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 1217 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3039.0, + "completions/mean_length": 971.6707763671875, + "completions/mean_terminated_length": 614.1604614257812, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 11.382507288629737, + "grad_norm": 0.1719530075788498, + "learning_rate": 1e-06, + "loss": -0.0488, + "num_tokens": 725481146.0, + "reward": 0.6662946939468384, + "reward_std": 0.15383115410804749, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1218 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3835.0, + "completions/mean_length": 1056.41748046875, + "completions/mean_terminated_length": 626.6165771484375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 11.391836734693877, + "grad_norm": 0.13213633000850677, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 726058072.0, + "reward": 0.668526828289032, + "reward_std": 0.10457134991884232, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1219 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3626.0, + "completions/mean_length": 1138.9320068359375, + "completions/mean_terminated_length": 672.8307495117188, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.401166180758018, + "grad_norm": 0.16146205365657806, + "learning_rate": 1e-06, + "loss": -0.0595, + "num_tokens": 726667451.0, + "reward": 0.6417410969734192, + "reward_std": 0.1444358229637146, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1220 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3948.0, + "completions/mean_length": 959.8214721679688, + "completions/mean_terminated_length": 579.0838623046875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 11.410495626822158, + "grad_norm": 0.15617051720619202, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 727218587.0, + "reward": 0.6852678656578064, + "reward_std": 0.10585196316242218, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1221 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4030.0, + "completions/mean_length": 1110.2723388671875, + "completions/mean_terminated_length": 635.1824340820312, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 11.419825072886297, + "grad_norm": 0.18704861402511597, + "learning_rate": 1e-06, + "loss": -0.051, + "num_tokens": 727798167.0, + "reward": 0.6238839626312256, + "reward_std": 0.1656332015991211, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 1222 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2473.0, + "completions/mean_length": 927.2879638671875, + "completions/mean_terminated_length": 646.2235717773438, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 11.429154518950437, + "grad_norm": 0.1620962917804718, + "learning_rate": 1e-06, + "loss": -0.044, + "num_tokens": 728417841.0, + "reward": 0.6886160969734192, + "reward_std": 0.1571405529975891, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1223 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4095.0, + "completions/mean_length": 1023.1975708007812, + "completions/mean_terminated_length": 606.4803466796875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 11.438483965014576, + "grad_norm": 0.17840290069580078, + "learning_rate": 1e-06, + "loss": -0.0412, + "num_tokens": 728993634.0, + "reward": 0.6674107313156128, + "reward_std": 0.13508442044258118, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1224 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3755.0, + "completions/mean_length": 973.536865234375, + "completions/mean_terminated_length": 611.9065551757812, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 11.447813411078718, + "grad_norm": 0.15685313940048218, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 729575483.0, + "reward": 0.7198660969734192, + "reward_std": 0.13598524034023285, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1225 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2823.0, + "completions/mean_length": 933.3292846679688, + "completions/mean_terminated_length": 571.4315795898438, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 11.457142857142857, + "grad_norm": 0.15781155228614807, + "learning_rate": 1e-06, + "loss": -0.0636, + "num_tokens": 730124650.0, + "reward": 0.7087053656578064, + "reward_std": 0.14928363263607025, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1226 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 1037.5926513671875, + "completions/mean_terminated_length": 622.82763671875, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.466472303206997, + "grad_norm": 0.18427269160747528, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 730722901.0, + "reward": 0.6305803656578064, + "reward_std": 0.16558972001075745, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1227 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3253.0, + "completions/mean_length": 1113.4263916015625, + "completions/mean_terminated_length": 691.6866455078125, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 11.475801749271136, + "grad_norm": 0.1259167641401291, + "learning_rate": 1e-06, + "loss": -0.0737, + "num_tokens": 731349355.0, + "reward": 0.6484375, + "reward_std": 0.11366245150566101, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1228 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1495535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3910.0, + "completions/mean_length": 1209.51904296875, + "completions/mean_terminated_length": 701.9225463867188, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 11.485131195335278, + "grad_norm": 0.15576831996440887, + "learning_rate": 1e-06, + "loss": -0.0575, + "num_tokens": 731974308.0, + "reward": 0.5736607313156128, + "reward_std": 0.17667387425899506, + "rewards/verify_math_reward/mean": 0.5736607313156128, + "rewards/verify_math_reward/std": 0.4948205351829529, + "step": 1229 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3979.0, + "completions/mean_length": 902.7723388671875, + "completions/mean_terminated_length": 598.2836303710938, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.494460641399417, + "grad_norm": 0.15066871047019958, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 732543776.0, + "reward": 0.7187500596046448, + "reward_std": 0.13109326362609863, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1230 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 955.4085083007812, + "completions/mean_terminated_length": 604.7221069335938, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 11.503790087463557, + "grad_norm": 0.15418845415115356, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 733113494.0, + "reward": 0.7087053656578064, + "reward_std": 0.1364317238330841, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1231 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3604.0, + "completions/mean_length": 902.3035888671875, + "completions/mean_terminated_length": 635.8403930664062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.513119533527696, + "grad_norm": 0.17068997025489807, + "learning_rate": 1e-06, + "loss": -0.039, + "num_tokens": 733732726.0, + "reward": 0.7321428656578064, + "reward_std": 0.1443609744310379, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1232 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3653.0, + "completions/mean_length": 999.3951416015625, + "completions/mean_terminated_length": 627.802490234375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 11.522448979591836, + "grad_norm": 0.16228055953979492, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 734316168.0, + "reward": 0.7477678656578064, + "reward_std": 0.15954798460006714, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1233 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2866.0, + "completions/mean_length": 1087.43310546875, + "completions/mean_terminated_length": 640.005126953125, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 11.531778425655977, + "grad_norm": 0.15073533356189728, + "learning_rate": 1e-06, + "loss": -0.0913, + "num_tokens": 734900220.0, + "reward": 0.65625, + "reward_std": 0.1583433598279953, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1234 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3666.0, + "completions/mean_length": 1051.71875, + "completions/mean_terminated_length": 603.4570922851562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.541107871720117, + "grad_norm": 0.18844175338745117, + "learning_rate": 1e-06, + "loss": -0.0673, + "num_tokens": 735459312.0, + "reward": 0.6640625, + "reward_std": 0.17720521986484528, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1235 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3674.0, + "completions/mean_length": 1063.9676513671875, + "completions/mean_terminated_length": 657.1380004882812, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 11.550437317784256, + "grad_norm": 0.1428215205669403, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 736066603.0, + "reward": 0.7008928656578064, + "reward_std": 0.12433585524559021, + "rewards/verify_math_reward/mean": 0.7008928656578064, + "rewards/verify_math_reward/std": 0.458122581243515, + "step": 1236 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 1219.5982666015625, + "completions/mean_terminated_length": 722.6282958984375, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 11.559766763848396, + "grad_norm": 0.13630394637584686, + "learning_rate": 1e-06, + "loss": -0.0764, + "num_tokens": 736719835.0, + "reward": 0.6026785969734192, + "reward_std": 0.12573368847370148, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 1237 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4049.0, + "completions/mean_length": 921.8538208007812, + "completions/mean_terminated_length": 627.6646118164062, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 11.569096209912537, + "grad_norm": 0.15348593890666962, + "learning_rate": 1e-06, + "loss": -0.0261, + "num_tokens": 737333888.0, + "reward": 0.731026828289032, + "reward_std": 0.138881653547287, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1238 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3959.0, + "completions/mean_length": 1063.9296875, + "completions/mean_terminated_length": 639.5941772460938, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.578425655976677, + "grad_norm": 0.16186900436878204, + "learning_rate": 1e-06, + "loss": -0.072, + "num_tokens": 737919713.0, + "reward": 0.7042410969734192, + "reward_std": 0.11753343045711517, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1239 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3553.0, + "completions/mean_length": 1004.6942138671875, + "completions/mean_terminated_length": 585.4676513671875, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 11.587755102040816, + "grad_norm": 0.1367913782596588, + "learning_rate": 1e-06, + "loss": -0.0478, + "num_tokens": 738478927.0, + "reward": 0.6852678656578064, + "reward_std": 0.11930260062217712, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1240 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 1005.34716796875, + "completions/mean_terminated_length": 608.31103515625, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.597084548104956, + "grad_norm": 0.13521602749824524, + "learning_rate": 1e-06, + "loss": -0.0493, + "num_tokens": 739047910.0, + "reward": 0.6964285969734192, + "reward_std": 0.08575044572353363, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1241 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 1109.2857666015625, + "completions/mean_terminated_length": 708.5367431640625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.606413994169095, + "grad_norm": 0.14619646966457367, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 739702086.0, + "reward": 0.6104910969734192, + "reward_std": 0.13252612948417664, + "rewards/verify_math_reward/mean": 0.6104910969734192, + "rewards/verify_math_reward/std": 0.48791128396987915, + "step": 1242 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3213.0, + "completions/mean_length": 951.0413208007812, + "completions/mean_terminated_length": 604.1995239257812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.615743440233237, + "grad_norm": 0.1499680131673813, + "learning_rate": 1e-06, + "loss": -0.0495, + "num_tokens": 740284443.0, + "reward": 0.6875000596046448, + "reward_std": 0.13290660083293915, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1243 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3033.0, + "completions/mean_length": 925.661865234375, + "completions/mean_terminated_length": 580.37744140625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.625072886297376, + "grad_norm": 0.14974988996982574, + "learning_rate": 1e-06, + "loss": -0.0445, + "num_tokens": 740838740.0, + "reward": 0.7176339626312256, + "reward_std": 0.13568103313446045, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1244 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3986.0, + "completions/mean_length": 1023.2188110351562, + "completions/mean_terminated_length": 632.8402709960938, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 11.634402332361516, + "grad_norm": 0.15908028185367584, + "learning_rate": 1e-06, + "loss": -0.0714, + "num_tokens": 741431512.0, + "reward": 0.6662946939468384, + "reward_std": 0.12941217422485352, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1245 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4013.0, + "completions/mean_length": 1098.4866943359375, + "completions/mean_terminated_length": 652.7025756835938, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 11.643731778425655, + "grad_norm": 0.17007242143154144, + "learning_rate": 1e-06, + "loss": -0.0752, + "num_tokens": 742034860.0, + "reward": 0.6696428656578064, + "reward_std": 0.17299722135066986, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1246 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 1011.1016235351562, + "completions/mean_terminated_length": 623.551513671875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 11.653061224489797, + "grad_norm": 0.13325351476669312, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 742615151.0, + "reward": 0.7120535969734192, + "reward_std": 0.12102647125720978, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1247 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3799.0, + "completions/mean_length": 1029.3795166015625, + "completions/mean_terminated_length": 648.4567260742188, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 11.662390670553936, + "grad_norm": 0.1285713016986847, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 743221595.0, + "reward": 0.6975446939468384, + "reward_std": 0.11388886719942093, + "rewards/verify_math_reward/mean": 0.6975446343421936, + "rewards/verify_math_reward/std": 0.45957788825035095, + "step": 1248 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4001.0, + "completions/mean_length": 957.1484985351562, + "completions/mean_terminated_length": 636.6998901367188, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.671720116618076, + "grad_norm": 0.1597519963979721, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 743824344.0, + "reward": 0.691964328289032, + "reward_std": 0.16134601831436157, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1249 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 970.7567138671875, + "completions/mean_terminated_length": 621.7841186523438, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 11.681049562682215, + "grad_norm": 0.14935170114040375, + "learning_rate": 1e-06, + "loss": -0.1032, + "num_tokens": 744414494.0, + "reward": 0.6953125596046448, + "reward_std": 0.1605985462665558, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1250 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1473214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3850.0, + "completions/mean_length": 1144.048095703125, + "completions/mean_terminated_length": 634.02490234375, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 11.690379008746355, + "grad_norm": 0.13916262984275818, + "learning_rate": 1e-06, + "loss": -0.0718, + "num_tokens": 744978161.0, + "reward": 0.6808035969734192, + "reward_std": 0.11719895154237747, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1251 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2912.0, + "completions/mean_length": 1064.610595703125, + "completions/mean_terminated_length": 670.8739013671875, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 11.699708454810496, + "grad_norm": 0.15602712333202362, + "learning_rate": 1e-06, + "loss": -0.0734, + "num_tokens": 745593828.0, + "reward": 0.6462053656578064, + "reward_std": 0.18121929466724396, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1252 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2870.0, + "completions/mean_length": 983.0770263671875, + "completions/mean_terminated_length": 626.8718872070312, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 11.709037900874636, + "grad_norm": 0.16188177466392517, + "learning_rate": 1e-06, + "loss": -0.0762, + "num_tokens": 746187145.0, + "reward": 0.6819196939468384, + "reward_std": 0.16807454824447632, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1253 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3322.0, + "completions/mean_length": 1079.6920166015625, + "completions/mean_terminated_length": 674.97216796875, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 11.718367346938775, + "grad_norm": 0.14454908668994904, + "learning_rate": 1e-06, + "loss": -0.061, + "num_tokens": 746805533.0, + "reward": 0.668526828289032, + "reward_std": 0.15901413559913635, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1254 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3152.0, + "completions/mean_length": 902.1585083007812, + "completions/mean_terminated_length": 558.6921997070312, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 11.727696793002915, + "grad_norm": 0.1539558470249176, + "learning_rate": 1e-06, + "loss": -0.0459, + "num_tokens": 747348075.0, + "reward": 0.731026828289032, + "reward_std": 0.11483247578144073, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1255 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3290.0, + "completions/mean_length": 918.3370971679688, + "completions/mean_terminated_length": 632.2700805664062, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 11.737026239067056, + "grad_norm": 0.1389485001564026, + "learning_rate": 1e-06, + "loss": -0.0713, + "num_tokens": 747950185.0, + "reward": 0.7299107313156128, + "reward_std": 0.14898376166820526, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1256 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3386.0, + "completions/mean_length": 1034.2254638671875, + "completions/mean_terminated_length": 623.4050903320312, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.746355685131196, + "grad_norm": 0.15202346444129944, + "learning_rate": 1e-06, + "loss": -0.0663, + "num_tokens": 748529603.0, + "reward": 0.6595982313156128, + "reward_std": 0.1341412365436554, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1257 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3335.0, + "completions/mean_length": 1097.1138916015625, + "completions/mean_terminated_length": 655.5364990234375, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 11.755685131195335, + "grad_norm": 0.17325426638126373, + "learning_rate": 1e-06, + "loss": -0.0816, + "num_tokens": 749127305.0, + "reward": 0.6729910969734192, + "reward_std": 0.1627788543701172, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1258 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4073.0, + "completions/mean_length": 886.3761596679688, + "completions/mean_terminated_length": 614.3741455078125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 11.765014577259475, + "grad_norm": 0.14746074378490448, + "learning_rate": 1e-06, + "loss": -0.0267, + "num_tokens": 749709746.0, + "reward": 0.7276785969734192, + "reward_std": 0.12343572080135345, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1259 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4023.0, + "completions/mean_length": 1078.524658203125, + "completions/mean_terminated_length": 729.0535278320312, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 11.774344023323614, + "grad_norm": 0.1544639617204666, + "learning_rate": 1e-06, + "loss": -0.0672, + "num_tokens": 750389472.0, + "reward": 0.6651785969734192, + "reward_std": 0.17761889100074768, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1260 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1607142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2627.0, + "completions/mean_length": 1214.0770263671875, + "completions/mean_terminated_length": 662.2194213867188, + "completions/min_length": 196.0, + "completions/min_terminated_length": 196.0, + "epoch": 11.783673469387756, + "grad_norm": 0.1464691162109375, + "learning_rate": 1e-06, + "loss": -0.0756, + "num_tokens": 750984869.0, + "reward": 0.613839328289032, + "reward_std": 0.15375375747680664, + "rewards/verify_math_reward/mean": 0.6138392686843872, + "rewards/verify_math_reward/std": 0.48714008927345276, + "step": 1261 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3653.0, + "completions/mean_length": 1033.49560546875, + "completions/mean_terminated_length": 653.0840454101562, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.793002915451895, + "grad_norm": 0.14695730805397034, + "learning_rate": 1e-06, + "loss": -0.0531, + "num_tokens": 751598801.0, + "reward": 0.6819196939468384, + "reward_std": 0.12723183631896973, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1262 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 1031.7645263671875, + "completions/mean_terminated_length": 646.810302734375, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 11.802332361516035, + "grad_norm": 0.17821983993053436, + "learning_rate": 1e-06, + "loss": -0.0627, + "num_tokens": 752201502.0, + "reward": 0.6551339626312256, + "reward_std": 0.17998537421226501, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 1263 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3092.0, + "completions/mean_length": 929.9297485351562, + "completions/mean_terminated_length": 619.5306396484375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 11.811661807580174, + "grad_norm": 0.17771141231060028, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 752794879.0, + "reward": 0.7500000596046448, + "reward_std": 0.15402477979660034, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1264 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2688.0, + "completions/mean_length": 1074.3270263671875, + "completions/mean_terminated_length": 620.4942626953125, + "completions/min_length": 102.0, + "completions/min_terminated_length": 102.0, + "epoch": 11.820991253644316, + "grad_norm": 0.1463192105293274, + "learning_rate": 1e-06, + "loss": -0.1116, + "num_tokens": 753358812.0, + "reward": 0.6953125596046448, + "reward_std": 0.1529296189546585, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1265 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1595982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3828.0, + "completions/mean_length": 1236.8951416015625, + "completions/mean_terminated_length": 693.9309692382812, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.830320699708455, + "grad_norm": 0.15511496365070343, + "learning_rate": 1e-06, + "loss": -0.0704, + "num_tokens": 753975590.0, + "reward": 0.6171875, + "reward_std": 0.14263640344142914, + "rewards/verify_math_reward/mean": 0.6171875, + "rewards/verify_math_reward/std": 0.4863446056842804, + "step": 1266 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3498.0, + "completions/mean_length": 1056.6685791015625, + "completions/mean_terminated_length": 622.4783325195312, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 11.839650145772595, + "grad_norm": 0.17394381761550903, + "learning_rate": 1e-06, + "loss": -0.068, + "num_tokens": 754546453.0, + "reward": 0.6629464626312256, + "reward_std": 0.162291020154953, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1267 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2888.0, + "completions/mean_length": 1039.255615234375, + "completions/mean_terminated_length": 620.3109130859375, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 11.848979591836734, + "grad_norm": 0.16789045929908752, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 755125314.0, + "reward": 0.6238839626312256, + "reward_std": 0.1317012757062912, + "rewards/verify_math_reward/mean": 0.6238839030265808, + "rewards/verify_math_reward/std": 0.4846802353858948, + "step": 1268 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3631.0, + "completions/mean_length": 990.1842041015625, + "completions/mean_terminated_length": 634.7922973632812, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 11.858309037900874, + "grad_norm": 0.13596922159194946, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 755737839.0, + "reward": 0.660714328289032, + "reward_std": 0.1086730808019638, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1269 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3026.0, + "completions/mean_length": 1142.2601318359375, + "completions/mean_terminated_length": 694.2634887695312, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 11.867638483965015, + "grad_norm": 0.16801749169826508, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 756372184.0, + "reward": 0.6462053656578064, + "reward_std": 0.16063174605369568, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1270 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3482.0, + "completions/mean_length": 1080.5848388671875, + "completions/mean_terminated_length": 614.2835083007812, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 11.876967930029155, + "grad_norm": 0.15226885676383972, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 756935972.0, + "reward": 0.6629464626312256, + "reward_std": 0.1476692408323288, + "rewards/verify_math_reward/mean": 0.6629464030265808, + "rewards/verify_math_reward/std": 0.47296738624572754, + "step": 1271 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3660.0, + "completions/mean_length": 1040.4765625, + "completions/mean_terminated_length": 639.2462158203125, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 11.886297376093294, + "grad_norm": 0.1780855357646942, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 757536103.0, + "reward": 0.6863839626312256, + "reward_std": 0.17389805614948273, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1272 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3582.0, + "completions/mean_length": 1103.6328125, + "completions/mean_terminated_length": 658.6141357421875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 11.895626822157434, + "grad_norm": 0.15448881685733795, + "learning_rate": 1e-06, + "loss": -0.0719, + "num_tokens": 758136334.0, + "reward": 0.6194196939468384, + "reward_std": 0.15225812792778015, + "rewards/verify_math_reward/mean": 0.6194196343421936, + "rewards/verify_math_reward/std": 0.48580074310302734, + "step": 1273 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3571.0, + "completions/mean_length": 941.8638916015625, + "completions/mean_terminated_length": 594.0099487304688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 11.904956268221575, + "grad_norm": 0.13812041282653809, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 758701884.0, + "reward": 0.7031250596046448, + "reward_std": 0.12715445458889008, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1274 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0591517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3987.0, + "completions/mean_length": 771.6975708007812, + "completions/mean_terminated_length": 562.6963500976562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 11.914285714285715, + "grad_norm": 0.15619826316833496, + "learning_rate": 1e-06, + "loss": -0.0335, + "num_tokens": 759256861.0, + "reward": 0.7600446939468384, + "reward_std": 0.11445268988609314, + "rewards/verify_math_reward/mean": 0.7600446343421936, + "rewards/verify_math_reward/std": 0.42729446291923523, + "step": 1275 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 1004.7723388671875, + "completions/mean_terminated_length": 607.6624145507812, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 11.923615160349854, + "grad_norm": 0.16423825919628143, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 759828201.0, + "reward": 0.6908482313156128, + "reward_std": 0.15000112354755402, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1276 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3125.0, + "completions/mean_length": 952.0870971679688, + "completions/mean_terminated_length": 557.1231079101562, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 11.932944606413994, + "grad_norm": 0.15073250234127045, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 760367271.0, + "reward": 0.6897321939468384, + "reward_std": 0.1238475888967514, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1277 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4078.0, + "completions/mean_length": 1097.114990234375, + "completions/mean_terminated_length": 703.3219604492188, + "completions/min_length": 204.0, + "completions/min_terminated_length": 204.0, + "epoch": 11.942274052478133, + "grad_norm": 0.14329920709133148, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 761004142.0, + "reward": 0.6863839626312256, + "reward_std": 0.12669983506202698, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1278 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3652.0, + "completions/mean_length": 1045.125, + "completions/mean_terminated_length": 653.198974609375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 11.951603498542275, + "grad_norm": 0.1631573885679245, + "learning_rate": 1e-06, + "loss": -0.0393, + "num_tokens": 761614686.0, + "reward": 0.676339328289032, + "reward_std": 0.14489272236824036, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1279 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3160.0, + "completions/mean_length": 938.0859985351562, + "completions/mean_terminated_length": 611.4051513671875, + "completions/min_length": 79.0, + "completions/min_terminated_length": 79.0, + "epoch": 11.960932944606414, + "grad_norm": 0.1288825422525406, + "learning_rate": 1e-06, + "loss": -0.0243, + "num_tokens": 762202323.0, + "reward": 0.6584821939468384, + "reward_std": 0.11095966398715973, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1280 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2668.0, + "completions/mean_length": 960.0156860351562, + "completions/mean_terminated_length": 652.5662231445312, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 11.970262390670554, + "grad_norm": 0.17056560516357422, + "learning_rate": 1e-06, + "loss": -0.0643, + "num_tokens": 762827601.0, + "reward": 0.6875000596046448, + "reward_std": 0.204411119222641, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1281 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3916.0, + "completions/mean_length": 923.2288208007812, + "completions/mean_terminated_length": 620.6907348632812, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 11.979591836734693, + "grad_norm": 0.13725991547107697, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 763414270.0, + "reward": 0.7511160969734192, + "reward_std": 0.1294453889131546, + "rewards/verify_math_reward/mean": 0.7511160969734192, + "rewards/verify_math_reward/std": 0.43260788917541504, + "step": 1282 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1316964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3685.0, + "completions/mean_length": 1124.609375, + "completions/mean_terminated_length": 673.9357299804688, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 11.988921282798835, + "grad_norm": 0.14854204654693604, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 764042680.0, + "reward": 0.6227678656578064, + "reward_std": 0.14458850026130676, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644899368286, + "step": 1283 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.16761363636363635, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4042.0, + "completions/mean_length": 1268.0455322265625, + "completions/mean_terminated_length": 698.5938110351562, + "completions/min_length": 215.0, + "completions/min_terminated_length": 215.0, + "epoch": 11.998250728862974, + "grad_norm": 0.1653452068567276, + "learning_rate": 1e-06, + "loss": -0.0641, + "num_tokens": 764643965.0, + "reward": 0.7031250596046448, + "reward_std": 0.15011171996593475, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1284 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 1057.899658203125, + "completions/mean_terminated_length": 667.6145629882812, + "completions/min_length": 188.0, + "completions/min_terminated_length": 188.0, + "epoch": 12.00932944606414, + "grad_norm": 0.14917266368865967, + "learning_rate": 1e-06, + "loss": -0.0492, + "num_tokens": 765266555.0, + "reward": 0.6462053656578064, + "reward_std": 0.1424509435892105, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1285 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 1019.388427734375, + "completions/mean_terminated_length": 650.1950073242188, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 12.018658892128279, + "grad_norm": 0.16054093837738037, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 765880119.0, + "reward": 0.6383928656578064, + "reward_std": 0.14373156428337097, + "rewards/verify_math_reward/mean": 0.6383928656578064, + "rewards/verify_math_reward/std": 0.4807341694831848, + "step": 1286 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3362.0, + "completions/mean_length": 1008.6719360351562, + "completions/mean_terminated_length": 638.1925048828125, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 12.02798833819242, + "grad_norm": 0.16151976585388184, + "learning_rate": 1e-06, + "loss": -0.0546, + "num_tokens": 766481601.0, + "reward": 0.676339328289032, + "reward_std": 0.1589481383562088, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1287 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3978.0, + "completions/mean_length": 1003.4308471679688, + "completions/mean_terminated_length": 649.5546875, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.03731778425656, + "grad_norm": 0.13827915489673615, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 767089795.0, + "reward": 0.6953125596046448, + "reward_std": 0.11681988090276718, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1288 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2951.0, + "completions/mean_length": 953.0234985351562, + "completions/mean_terminated_length": 674.2418212890625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 12.0466472303207, + "grad_norm": 0.15015266835689545, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 767740128.0, + "reward": 0.676339328289032, + "reward_std": 0.16915972530841827, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1289 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4022.0, + "completions/mean_length": 903.6652221679688, + "completions/mean_terminated_length": 607.7902221679688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.055976676384839, + "grad_norm": 0.1472347229719162, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 768315092.0, + "reward": 0.7488839626312256, + "reward_std": 0.1295951008796692, + "rewards/verify_math_reward/mean": 0.7488839030265808, + "rewards/verify_math_reward/std": 0.43389734625816345, + "step": 1290 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3620.0, + "completions/mean_length": 988.9866333007812, + "completions/mean_terminated_length": 663.343994140625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 12.06530612244898, + "grad_norm": 0.15924403071403503, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 768945408.0, + "reward": 0.6986607313156128, + "reward_std": 0.13493216037750244, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1291 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3074.0, + "completions/mean_length": 1049.0357666015625, + "completions/mean_terminated_length": 666.251220703125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 12.07463556851312, + "grad_norm": 0.16694320738315582, + "learning_rate": 1e-06, + "loss": -0.0621, + "num_tokens": 769563224.0, + "reward": 0.6618303656578064, + "reward_std": 0.16735847294330597, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1292 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 1038.19873046875, + "completions/mean_terminated_length": 645.3828735351562, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 12.08396501457726, + "grad_norm": 0.18636220693588257, + "learning_rate": 1e-06, + "loss": -0.0208, + "num_tokens": 770178442.0, + "reward": 0.6439732313156128, + "reward_std": 0.16724829375743866, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1293 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3792.0, + "completions/mean_length": 1080.849365234375, + "completions/mean_terminated_length": 658.8816528320312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.093294460641399, + "grad_norm": 0.15357685089111328, + "learning_rate": 1e-06, + "loss": -0.0893, + "num_tokens": 770798195.0, + "reward": 0.6651785969734192, + "reward_std": 0.15657459199428558, + "rewards/verify_math_reward/mean": 0.6651785969734192, + "rewards/verify_math_reward/std": 0.47219157218933105, + "step": 1294 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3169.0, + "completions/mean_length": 1008.982177734375, + "completions/mean_terminated_length": 621.1658325195312, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 12.102623906705539, + "grad_norm": 0.16940230131149292, + "learning_rate": 1e-06, + "loss": -0.0518, + "num_tokens": 771385307.0, + "reward": 0.6741071939468384, + "reward_std": 0.14838533103466034, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.4689692556858063, + "step": 1295 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 949.8850708007812, + "completions/mean_terminated_length": 641.4424438476562, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.11195335276968, + "grad_norm": 0.12281103432178497, + "learning_rate": 1e-06, + "loss": -0.0521, + "num_tokens": 772003996.0, + "reward": 0.6696428656578064, + "reward_std": 0.11178591102361679, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1296 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3357.0, + "completions/mean_length": 982.138427734375, + "completions/mean_terminated_length": 643.0049438476562, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 12.12128279883382, + "grad_norm": 0.14556923508644104, + "learning_rate": 1e-06, + "loss": -0.0758, + "num_tokens": 772623560.0, + "reward": 0.6886160969734192, + "reward_std": 0.13891513645648956, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1297 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3314.0, + "completions/mean_length": 1049.982177734375, + "completions/mean_terminated_length": 680.1902465820312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.130612244897959, + "grad_norm": 0.1514204889535904, + "learning_rate": 1e-06, + "loss": -0.0588, + "num_tokens": 773261472.0, + "reward": 0.637276828289032, + "reward_std": 0.15672755241394043, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1298 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3660.0, + "completions/mean_length": 1008.6574096679688, + "completions/mean_terminated_length": 680.865478515625, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 12.139941690962099, + "grad_norm": 0.13329489529132843, + "learning_rate": 1e-06, + "loss": -0.0535, + "num_tokens": 773896621.0, + "reward": 0.6830357313156128, + "reward_std": 0.1205839142203331, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1299 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3771.0, + "completions/mean_length": 838.1908569335938, + "completions/mean_terminated_length": 557.8218383789062, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 12.14927113702624, + "grad_norm": 0.15283958613872528, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 774440376.0, + "reward": 0.7979910969734192, + "reward_std": 0.09514645487070084, + "rewards/verify_math_reward/mean": 0.7979910969734192, + "rewards/verify_math_reward/std": 0.40172311663627625, + "step": 1300 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3495.0, + "completions/mean_length": 966.9375610351562, + "completions/mean_terminated_length": 617.5384521484375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 12.15860058309038, + "grad_norm": 0.12828518450260162, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 775033200.0, + "reward": 0.6819196939468384, + "reward_std": 0.10761747509241104, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599096059799194, + "step": 1301 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3349.0, + "completions/mean_length": 1072.46875, + "completions/mean_terminated_length": 675.4393920898438, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.167930029154519, + "grad_norm": 0.15174058079719543, + "learning_rate": 1e-06, + "loss": -0.0876, + "num_tokens": 775653172.0, + "reward": 0.6875000596046448, + "reward_std": 0.17795519530773163, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1302 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3710.0, + "completions/mean_length": 1168.921875, + "completions/mean_terminated_length": 689.9454345703125, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 12.177259475218658, + "grad_norm": 0.15555673837661743, + "learning_rate": 1e-06, + "loss": -0.0297, + "num_tokens": 776275990.0, + "reward": 0.6417410969734192, + "reward_std": 0.1197904497385025, + "rewards/verify_math_reward/mean": 0.6417410969734192, + "rewards/verify_math_reward/std": 0.47975656390190125, + "step": 1303 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3272.0, + "completions/mean_length": 918.6295166015625, + "completions/mean_terminated_length": 598.5503540039062, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.186588921282798, + "grad_norm": 0.1464909166097641, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 776848634.0, + "reward": 0.7366071939468384, + "reward_std": 0.09397321194410324, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1304 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3310.0, + "completions/mean_length": 851.2767944335938, + "completions/mean_terminated_length": 559.1727294921875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 12.19591836734694, + "grad_norm": 0.1588182896375656, + "learning_rate": 1e-06, + "loss": -0.0302, + "num_tokens": 777392426.0, + "reward": 0.754464328289032, + "reward_std": 0.12805600464344025, + "rewards/verify_math_reward/mean": 0.7544642686843872, + "rewards/verify_math_reward/std": 0.43064478039741516, + "step": 1305 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1517857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4086.0, + "completions/mean_length": 1188.5648193359375, + "completions/mean_terminated_length": 668.286865234375, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.205247813411079, + "grad_norm": 0.15582671761512756, + "learning_rate": 1e-06, + "loss": -0.0469, + "num_tokens": 777992220.0, + "reward": 0.6071428656578064, + "reward_std": 0.13267767429351807, + "rewards/verify_math_reward/mean": 0.6071428656578064, + "rewards/verify_math_reward/std": 0.48865827918052673, + "step": 1306 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3874.0, + "completions/mean_length": 985.22216796875, + "completions/mean_terminated_length": 637.86474609375, + "completions/min_length": 168.0, + "completions/min_terminated_length": 168.0, + "epoch": 12.214577259475218, + "grad_norm": 0.16907796263694763, + "learning_rate": 1e-06, + "loss": -0.0447, + "num_tokens": 778589531.0, + "reward": 0.676339328289032, + "reward_std": 0.15744182467460632, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1307 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3609.0, + "completions/mean_length": 1081.1116943359375, + "completions/mean_terminated_length": 672.2484130859375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 12.223906705539358, + "grad_norm": 0.17167146503925323, + "learning_rate": 1e-06, + "loss": -0.0851, + "num_tokens": 779216519.0, + "reward": 0.6339285969734192, + "reward_std": 0.1906241774559021, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 1308 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 987.9230346679688, + "completions/mean_terminated_length": 632.2723999023438, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.2332361516035, + "grad_norm": 0.15194860100746155, + "learning_rate": 1e-06, + "loss": -0.0443, + "num_tokens": 779808842.0, + "reward": 0.6774553656578064, + "reward_std": 0.1300847828388214, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1309 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.140625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 1139.3795166015625, + "completions/mean_terminated_length": 655.56884765625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.242565597667639, + "grad_norm": 0.1756047159433365, + "learning_rate": 1e-06, + "loss": -0.0842, + "num_tokens": 780404390.0, + "reward": 0.6618303656578064, + "reward_std": 0.16976381838321686, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1310 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2885.0, + "completions/mean_length": 914.2332763671875, + "completions/mean_terminated_length": 610.8374633789062, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 12.251895043731778, + "grad_norm": 0.12352645397186279, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 780990535.0, + "reward": 0.738839328289032, + "reward_std": 0.10652484744787216, + "rewards/verify_math_reward/mean": 0.7388392686843872, + "rewards/verify_math_reward/std": 0.439512699842453, + "step": 1311 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3014.0, + "completions/mean_length": 1047.0045166015625, + "completions/mean_terminated_length": 646.6312866210938, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 12.261224489795918, + "grad_norm": 0.15764611959457397, + "learning_rate": 1e-06, + "loss": -0.0655, + "num_tokens": 781588835.0, + "reward": 0.6897321939468384, + "reward_std": 0.15015266835689545, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1312 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2969.0, + "completions/mean_length": 1011.0123291015625, + "completions/mean_terminated_length": 592.642578125, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 12.270553935860057, + "grad_norm": 0.1534235179424286, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 782149750.0, + "reward": 0.6930803656578064, + "reward_std": 0.13616958260536194, + "rewards/verify_math_reward/mean": 0.6930803656578064, + "rewards/verify_math_reward/std": 0.46147337555885315, + "step": 1313 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3865.0, + "completions/mean_length": 1072.6585693359375, + "completions/mean_terminated_length": 658.2918701171875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 12.279883381924199, + "grad_norm": 0.15974879264831543, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 782769940.0, + "reward": 0.625, + "reward_std": 0.15417702496051788, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1314 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3825.0, + "completions/mean_length": 935.22998046875, + "completions/mean_terminated_length": 595.3201293945312, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.289212827988338, + "grad_norm": 0.16423502564430237, + "learning_rate": 1e-06, + "loss": -0.0415, + "num_tokens": 783341490.0, + "reward": 0.691964328289032, + "reward_std": 0.14661657810211182, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1315 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3647.0, + "completions/mean_length": 974.294677734375, + "completions/mean_terminated_length": 680.801025390625, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.298542274052478, + "grad_norm": 0.15298013389110565, + "learning_rate": 1e-06, + "loss": -0.0478, + "num_tokens": 783991714.0, + "reward": 0.6886160969734192, + "reward_std": 0.1390673965215683, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1316 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1238839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4024.0, + "completions/mean_length": 1073.90625, + "completions/mean_terminated_length": 646.578369140625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.307871720116617, + "grad_norm": 0.19653767347335815, + "learning_rate": 1e-06, + "loss": -0.0918, + "num_tokens": 784592294.0, + "reward": 0.6718750596046448, + "reward_std": 0.1755894273519516, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1317 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3439.0, + "completions/mean_length": 1049.2054443359375, + "completions/mean_terminated_length": 687.8501586914062, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.317201166180759, + "grad_norm": 0.1472688913345337, + "learning_rate": 1e-06, + "loss": -0.0452, + "num_tokens": 785227606.0, + "reward": 0.6462053656578064, + "reward_std": 0.13238415122032166, + "rewards/verify_math_reward/mean": 0.6462053656578064, + "rewards/verify_math_reward/std": 0.478413462638855, + "step": 1318 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3811.0, + "completions/mean_length": 979.72998046875, + "completions/mean_terminated_length": 648.86669921875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 12.326530612244898, + "grad_norm": 0.14581452310085297, + "learning_rate": 1e-06, + "loss": -0.0635, + "num_tokens": 785837876.0, + "reward": 0.7042410969734192, + "reward_std": 0.1526353806257248, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1319 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2732.0, + "completions/mean_length": 1018.1183471679688, + "completions/mean_terminated_length": 661.6513061523438, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 12.335860058309038, + "grad_norm": 0.16291938722133636, + "learning_rate": 1e-06, + "loss": -0.0454, + "num_tokens": 786453958.0, + "reward": 0.6796875596046448, + "reward_std": 0.14617221057415009, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1320 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3888.0, + "completions/mean_length": 845.0904541015625, + "completions/mean_terminated_length": 548.11328125, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.345189504373177, + "grad_norm": 0.15102356672286987, + "learning_rate": 1e-06, + "loss": -0.0511, + "num_tokens": 786990543.0, + "reward": 0.6953125596046448, + "reward_std": 0.11542978882789612, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1321 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3656.0, + "completions/mean_length": 973.28466796875, + "completions/mean_terminated_length": 611.6251220703125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 12.354518950437317, + "grad_norm": 0.1624811887741089, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 787565590.0, + "reward": 0.7109375596046448, + "reward_std": 0.14692038297653198, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1322 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3564.0, + "completions/mean_length": 865.8861694335938, + "completions/mean_terminated_length": 600.611083984375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.363848396501458, + "grad_norm": 0.13667301833629608, + "learning_rate": 1e-06, + "loss": -0.0248, + "num_tokens": 788150288.0, + "reward": 0.7410714626312256, + "reward_std": 0.11983321607112885, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1323 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4019.0, + "completions/mean_length": 993.4933471679688, + "completions/mean_terminated_length": 664.0913696289062, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.373177842565598, + "grad_norm": 0.14774277806282043, + "learning_rate": 1e-06, + "loss": -0.049, + "num_tokens": 788773298.0, + "reward": 0.6595982313156128, + "reward_std": 0.13591109216213226, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1324 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3933.0, + "completions/mean_length": 1012.3281860351562, + "completions/mean_terminated_length": 646.5992431640625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 12.382507288629737, + "grad_norm": 0.14986541867256165, + "learning_rate": 1e-06, + "loss": -0.0685, + "num_tokens": 789364976.0, + "reward": 0.6852678656578064, + "reward_std": 0.1265917420387268, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1325 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3267.0, + "completions/mean_length": 937.2098388671875, + "completions/mean_terminated_length": 623.2687377929688, + "completions/min_length": 105.0, + "completions/min_terminated_length": 105.0, + "epoch": 12.391836734693877, + "grad_norm": 0.11901839077472687, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 789958460.0, + "reward": 0.7500000596046448, + "reward_std": 0.10791100561618805, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1326 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3697.0, + "completions/mean_length": 1145.1585693359375, + "completions/mean_terminated_length": 693.227783203125, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 12.401166180758018, + "grad_norm": 0.12237332761287689, + "learning_rate": 1e-06, + "loss": -0.054, + "num_tokens": 790596850.0, + "reward": 0.629464328289032, + "reward_std": 0.11471301317214966, + "rewards/verify_math_reward/mean": 0.6294642686843872, + "rewards/verify_math_reward/std": 0.4832179844379425, + "step": 1327 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3038.0, + "completions/mean_length": 887.46435546875, + "completions/mean_terminated_length": 623.9613647460938, + "completions/min_length": 163.0, + "completions/min_terminated_length": 163.0, + "epoch": 12.410495626822158, + "grad_norm": 0.1623559445142746, + "learning_rate": 1e-06, + "loss": -0.0597, + "num_tokens": 791207858.0, + "reward": 0.7187500596046448, + "reward_std": 0.15706203877925873, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1328 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3717.0, + "completions/mean_length": 989.3906860351562, + "completions/mean_terminated_length": 646.7781982421875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 12.419825072886297, + "grad_norm": 0.19060024619102478, + "learning_rate": 1e-06, + "loss": -0.0733, + "num_tokens": 791809104.0, + "reward": 0.7064732313156128, + "reward_std": 0.13760174810886383, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1329 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1305803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3980.0, + "completions/mean_length": 1093.9420166015625, + "completions/mean_terminated_length": 643.0552368164062, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 12.429154518950437, + "grad_norm": 0.14512689411640167, + "learning_rate": 1e-06, + "loss": -0.0644, + "num_tokens": 792402300.0, + "reward": 0.6551339626312256, + "reward_std": 0.13809029757976532, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 1330 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3379.0, + "completions/mean_length": 964.4219360351562, + "completions/mean_terminated_length": 644.7158813476562, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 12.438483965014576, + "grad_norm": 0.15456648170948029, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 793006342.0, + "reward": 0.676339328289032, + "reward_std": 0.13470715284347534, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1331 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3642.0, + "completions/mean_length": 1029.1317138671875, + "completions/mean_terminated_length": 652.4987182617188, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 12.447813411078718, + "grad_norm": 0.16465410590171814, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 793618308.0, + "reward": 0.6707589626312256, + "reward_std": 0.16773684322834015, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1332 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3292.0, + "completions/mean_length": 1079.8359375, + "completions/mean_terminated_length": 696.6503295898438, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 12.457142857142857, + "grad_norm": 0.16468089818954468, + "learning_rate": 1e-06, + "loss": -0.0452, + "num_tokens": 794265297.0, + "reward": 0.6049107313156128, + "reward_std": 0.1518009752035141, + "rewards/verify_math_reward/mean": 0.6049107313156128, + "rewards/verify_math_reward/std": 0.48914292454719543, + "step": 1333 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3082.0, + "completions/mean_length": 1030.423095703125, + "completions/mean_terminated_length": 619.0924072265625, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 12.466472303206997, + "grad_norm": 0.14571166038513184, + "learning_rate": 1e-06, + "loss": -0.0653, + "num_tokens": 794845756.0, + "reward": 0.6774553656578064, + "reward_std": 0.1477861851453781, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1334 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3937.0, + "completions/mean_length": 1032.193115234375, + "completions/mean_terminated_length": 655.9360961914062, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 12.475801749271136, + "grad_norm": 0.14784695208072662, + "learning_rate": 1e-06, + "loss": -0.0775, + "num_tokens": 795460169.0, + "reward": 0.6718750596046448, + "reward_std": 0.14507704973220825, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1335 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3144.0, + "completions/mean_length": 942.9944458007812, + "completions/mean_terminated_length": 595.2651977539062, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 12.485131195335278, + "grad_norm": 0.15652582049369812, + "learning_rate": 1e-06, + "loss": -0.0714, + "num_tokens": 796021180.0, + "reward": 0.7276785969734192, + "reward_std": 0.1357916295528412, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1336 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3973.0, + "completions/mean_length": 999.1004638671875, + "completions/mean_terminated_length": 644.7288208007812, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 12.494460641399417, + "grad_norm": 0.14160537719726562, + "learning_rate": 1e-06, + "loss": -0.0615, + "num_tokens": 796623638.0, + "reward": 0.6986607313156128, + "reward_std": 0.1447739452123642, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1337 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3740.0, + "completions/mean_length": 1010.1551513671875, + "completions/mean_terminated_length": 639.853759765625, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.503790087463557, + "grad_norm": 0.15588660538196564, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 797221425.0, + "reward": 0.660714328289032, + "reward_std": 0.13722378015518188, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1338 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2740.0, + "completions/mean_length": 889.58935546875, + "completions/mean_terminated_length": 592.4097290039062, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.513119533527696, + "grad_norm": 0.14510086178779602, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 797787425.0, + "reward": 0.7187500596046448, + "reward_std": 0.16209529340267181, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1339 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3453.0, + "completions/mean_length": 947.3426513671875, + "completions/mean_terminated_length": 591.407470703125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.522448979591836, + "grad_norm": 0.15744498372077942, + "learning_rate": 1e-06, + "loss": -0.0508, + "num_tokens": 798369292.0, + "reward": 0.6908482313156128, + "reward_std": 0.11704740673303604, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1340 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2992.0, + "completions/mean_length": 861.200927734375, + "completions/mean_terminated_length": 578.5485229492188, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 12.531778425655977, + "grad_norm": 0.1440727561712265, + "learning_rate": 1e-06, + "loss": -0.0329, + "num_tokens": 798937856.0, + "reward": 0.7410714626312256, + "reward_std": 0.13342584669589996, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1341 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3445.0, + "completions/mean_length": 983.8381958007812, + "completions/mean_terminated_length": 584.0390014648438, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.541107871720117, + "grad_norm": 0.16874141991138458, + "learning_rate": 1e-06, + "loss": -0.0467, + "num_tokens": 799503071.0, + "reward": 0.676339328289032, + "reward_std": 0.15202990174293518, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1342 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3402.0, + "completions/mean_length": 970.9442138671875, + "completions/mean_terminated_length": 617.677001953125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 12.550437317784256, + "grad_norm": 0.14087730646133423, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 800089085.0, + "reward": 0.6819196939468384, + "reward_std": 0.1222303956747055, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1343 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3863.0, + "completions/mean_length": 992.7500610351562, + "completions/mean_terminated_length": 646.2332153320312, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 12.559766763848396, + "grad_norm": 0.15147152543067932, + "learning_rate": 1e-06, + "loss": -0.0462, + "num_tokens": 800696861.0, + "reward": 0.7243303656578064, + "reward_std": 0.14034590125083923, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1344 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3217.0, + "completions/mean_length": 893.81591796875, + "completions/mean_terminated_length": 622.4443359375, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 12.569096209912537, + "grad_norm": 0.14602839946746826, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 801291464.0, + "reward": 0.6796875596046448, + "reward_std": 0.12174323946237564, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1345 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3396.0, + "completions/mean_length": 1119.8951416015625, + "completions/mean_terminated_length": 681.6722412109375, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 12.578425655976677, + "grad_norm": 0.16214600205421448, + "learning_rate": 1e-06, + "loss": -0.0688, + "num_tokens": 801917698.0, + "reward": 0.6618303656578064, + "reward_std": 0.15228882431983948, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1346 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3782.0, + "completions/mean_length": 943.4576416015625, + "completions/mean_terminated_length": 634.3848266601562, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 12.587755102040816, + "grad_norm": 0.13340987265110016, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 802534276.0, + "reward": 0.6908482313156128, + "reward_std": 0.1303892731666565, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1347 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3862.0, + "completions/mean_length": 1031.860595703125, + "completions/mean_terminated_length": 664.1637573242188, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.597084548104956, + "grad_norm": 0.15837465226650238, + "learning_rate": 1e-06, + "loss": -0.0325, + "num_tokens": 803155439.0, + "reward": 0.6540178656578064, + "reward_std": 0.1432330161333084, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1348 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 999.0234985351562, + "completions/mean_terminated_length": 614.3299560546875, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.606413994169095, + "grad_norm": 0.16334226727485657, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 803730268.0, + "reward": 0.7299107313156128, + "reward_std": 0.1442175954580307, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1349 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3904.0, + "completions/mean_length": 1002.771240234375, + "completions/mean_terminated_length": 601.0025024414062, + "completions/min_length": 82.0, + "completions/min_terminated_length": 82.0, + "epoch": 12.615743440233237, + "grad_norm": 0.16276484727859497, + "learning_rate": 1e-06, + "loss": -0.0883, + "num_tokens": 804294591.0, + "reward": 0.6986607313156128, + "reward_std": 0.14312425255775452, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1350 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3762.0, + "completions/mean_length": 1086.8851318359375, + "completions/mean_terminated_length": 608.07373046875, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 12.625072886297376, + "grad_norm": 0.17980749905109406, + "learning_rate": 1e-06, + "loss": -0.0153, + "num_tokens": 804861432.0, + "reward": 0.6707589626312256, + "reward_std": 0.12550798058509827, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1351 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3732.0, + "completions/mean_length": 991.7500610351562, + "completions/mean_terminated_length": 606.153076171875, + "completions/min_length": 92.0, + "completions/min_terminated_length": 92.0, + "epoch": 12.634402332361516, + "grad_norm": 0.3175767660140991, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 805436472.0, + "reward": 0.7287946939468384, + "reward_std": 0.13508622348308563, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1352 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3016.0, + "completions/mean_length": 848.1194458007812, + "completions/mean_terminated_length": 648.0130615234375, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 12.643731778425655, + "grad_norm": 0.16018827259540558, + "learning_rate": 1e-06, + "loss": -0.0169, + "num_tokens": 806060531.0, + "reward": 0.7109375596046448, + "reward_std": 0.13395969569683075, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1353 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1294642857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3601.0, + "completions/mean_length": 1118.18310546875, + "completions/mean_terminated_length": 675.3282470703125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.653061224489797, + "grad_norm": 0.1592092365026474, + "learning_rate": 1e-06, + "loss": -0.0804, + "num_tokens": 806685863.0, + "reward": 0.6584821939468384, + "reward_std": 0.1529735028743744, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1354 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3878.0, + "completions/mean_length": 1150.318115234375, + "completions/mean_terminated_length": 677.1774291992188, + "completions/min_length": 151.0, + "completions/min_terminated_length": 151.0, + "epoch": 12.662390670553936, + "grad_norm": 0.15997087955474854, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 807292580.0, + "reward": 0.6595982313156128, + "reward_std": 0.14477324485778809, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1355 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3767.0, + "completions/mean_length": 1049.462158203125, + "completions/mean_terminated_length": 640.6860961914062, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 12.671720116618076, + "grad_norm": 0.1335633099079132, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 807893314.0, + "reward": 0.7142857313156128, + "reward_std": 0.10761930048465729, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1356 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3540.0, + "completions/mean_length": 1031.857177734375, + "completions/mean_terminated_length": 603.0330810546875, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 12.681049562682215, + "grad_norm": 0.16285446286201477, + "learning_rate": 1e-06, + "loss": -0.0788, + "num_tokens": 808466170.0, + "reward": 0.691964328289032, + "reward_std": 0.12982586026191711, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1357 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3775.0, + "completions/mean_length": 1027.227783203125, + "completions/mean_terminated_length": 624.257568359375, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.690379008746355, + "grad_norm": 0.18423277139663696, + "learning_rate": 1e-06, + "loss": -0.0748, + "num_tokens": 809050238.0, + "reward": 0.6595982313156128, + "reward_std": 0.15706273913383484, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1358 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3827.0, + "completions/mean_length": 1007.6607666015625, + "completions/mean_terminated_length": 649.9825439453125, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 12.699708454810496, + "grad_norm": 0.1600826382637024, + "learning_rate": 1e-06, + "loss": -0.0632, + "num_tokens": 809661230.0, + "reward": 0.6584821939468384, + "reward_std": 0.15033671259880066, + "rewards/verify_math_reward/mean": 0.6584821343421936, + "rewards/verify_math_reward/std": 0.4744836091995239, + "step": 1359 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 924.7332763671875, + "completions/mean_terminated_length": 622.3386840820312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.709037900874636, + "grad_norm": 0.144283264875412, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 810251999.0, + "reward": 0.6696428656578064, + "reward_std": 0.13342836499214172, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1360 + }, + { + "clip_ratio/high_max": 0.044763242709450424, + "clip_ratio/high_mean": 0.02001080778427422, + "clip_ratio/low_mean": 0.0020710817734652665, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.022081889619585127, + "completions/clipped_ratio": 0.1428571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3147.0, + "completions/mean_length": 1087.747802734375, + "completions/mean_terminated_length": 586.3724365234375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 12.718367346938775, + "grad_norm": 0.19869470596313477, + "learning_rate": 1e-06, + "loss": -0.0822, + "num_tokens": 810793957.0, + "reward": 0.6875000596046448, + "reward_std": 0.1593957394361496, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1361 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3465.0, + "completions/mean_length": 1020.489990234375, + "completions/mean_terminated_length": 616.6350708007812, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.727696793002915, + "grad_norm": 0.14233970642089844, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 811368244.0, + "reward": 0.7299107313156128, + "reward_std": 0.13192838430404663, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1362 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3954.0, + "completions/mean_length": 899.7277221679688, + "completions/mean_terminated_length": 611.9854125976562, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 12.737026239067056, + "grad_norm": 0.1616661548614502, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 811960384.0, + "reward": 0.754464328289032, + "reward_std": 0.14045536518096924, + "rewards/verify_math_reward/mean": 0.7544642686843872, + "rewards/verify_math_reward/std": 0.4306447505950928, + "step": 1363 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3245.0, + "completions/mean_length": 836.7545166015625, + "completions/mean_terminated_length": 573.3413696289062, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 12.746355685131196, + "grad_norm": 0.1772317737340927, + "learning_rate": 1e-06, + "loss": -0.03, + "num_tokens": 812525732.0, + "reward": 0.7767857313156128, + "reward_std": 0.13016103208065033, + "rewards/verify_math_reward/mean": 0.7767857313156128, + "rewards/verify_math_reward/std": 0.41663339734077454, + "step": 1364 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3460.0, + "completions/mean_length": 1024.688720703125, + "completions/mean_terminated_length": 660.4257202148438, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 12.755685131195335, + "grad_norm": 0.15694238245487213, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 813150549.0, + "reward": 0.691964328289032, + "reward_std": 0.14286211133003235, + "rewards/verify_math_reward/mean": 0.6919642686843872, + "rewards/verify_math_reward/std": 0.4619392454624176, + "step": 1365 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3661.0, + "completions/mean_length": 1028.212158203125, + "completions/mean_terminated_length": 672.9140625, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 12.765014577259475, + "grad_norm": 0.14862021803855896, + "learning_rate": 1e-06, + "loss": -0.0585, + "num_tokens": 813785067.0, + "reward": 0.6830357313156128, + "reward_std": 0.14207187294960022, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1366 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2939.0, + "completions/mean_length": 877.7656860351562, + "completions/mean_terminated_length": 605.033935546875, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 12.774344023323614, + "grad_norm": 0.12555168569087982, + "learning_rate": 1e-06, + "loss": -0.029, + "num_tokens": 814370041.0, + "reward": 0.7689732313156128, + "reward_std": 0.10634120553731918, + "rewards/verify_math_reward/mean": 0.7689732313156128, + "rewards/verify_math_reward/std": 0.42172491550445557, + "step": 1367 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3003.0, + "completions/mean_length": 1074.454345703125, + "completions/mean_terminated_length": 655.9682006835938, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 12.783673469387756, + "grad_norm": 0.17224238812923431, + "learning_rate": 1e-06, + "loss": -0.0574, + "num_tokens": 814981904.0, + "reward": 0.637276828289032, + "reward_std": 0.149361714720726, + "rewards/verify_math_reward/mean": 0.6372767686843872, + "rewards/verify_math_reward/std": 0.481054425239563, + "step": 1368 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3102.0, + "completions/mean_length": 1024.899658203125, + "completions/mean_terminated_length": 608.4131469726562, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 12.793002915451895, + "grad_norm": 0.15515033900737762, + "learning_rate": 1e-06, + "loss": -0.0513, + "num_tokens": 815551670.0, + "reward": 0.6897321939468384, + "reward_std": 0.14150916039943695, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1369 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4066.0, + "completions/mean_length": 933.6652221679688, + "completions/mean_terminated_length": 640.5707397460938, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 12.802332361516035, + "grad_norm": 0.14375588297843933, + "learning_rate": 1e-06, + "loss": -0.0408, + "num_tokens": 816163802.0, + "reward": 0.7265625596046448, + "reward_std": 0.13760173320770264, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1370 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2610.0, + "completions/mean_length": 873.7745971679688, + "completions/mean_terminated_length": 566.5208129882812, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.811661807580174, + "grad_norm": 0.13700838387012482, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 816707880.0, + "reward": 0.7031250596046448, + "reward_std": 0.11840105801820755, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1371 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 931.5391235351562, + "completions/mean_terminated_length": 599.8754272460938, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 12.820991253644316, + "grad_norm": 0.13432341814041138, + "learning_rate": 1e-06, + "loss": -0.0472, + "num_tokens": 817274451.0, + "reward": 0.7254464626312256, + "reward_std": 0.10626382380723953, + "rewards/verify_math_reward/mean": 0.7254464030265808, + "rewards/verify_math_reward/std": 0.4465382993221283, + "step": 1372 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2784.0, + "completions/mean_length": 952.02685546875, + "completions/mean_terminated_length": 626.7881469726562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 12.830320699708455, + "grad_norm": 0.16752497851848602, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 817885251.0, + "reward": 0.6752232313156128, + "reward_std": 0.12151430547237396, + "rewards/verify_math_reward/mean": 0.6752232313156128, + "rewards/verify_math_reward/std": 0.46855294704437256, + "step": 1373 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2790.0, + "completions/mean_length": 948.10498046875, + "completions/mean_terminated_length": 600.9392700195312, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 12.839650145772595, + "grad_norm": 0.13814741373062134, + "learning_rate": 1e-06, + "loss": -0.0499, + "num_tokens": 818457761.0, + "reward": 0.6908482313156128, + "reward_std": 0.107620008289814, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1374 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3300.0, + "completions/mean_length": 933.9576416015625, + "completions/mean_terminated_length": 580.8759155273438, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 12.848979591836734, + "grad_norm": 0.16600094735622406, + "learning_rate": 1e-06, + "loss": -0.0817, + "num_tokens": 819013907.0, + "reward": 0.7042410969734192, + "reward_std": 0.1651446521282196, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1375 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.052455357142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4020.0, + "completions/mean_length": 813.4620971679688, + "completions/mean_terminated_length": 631.7432250976562, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 12.858309037900874, + "grad_norm": 0.14070355892181396, + "learning_rate": 1e-06, + "loss": -0.024, + "num_tokens": 819637337.0, + "reward": 0.7399553656578064, + "reward_std": 0.09653403609991074, + "rewards/verify_math_reward/mean": 0.7399553656578064, + "rewards/verify_math_reward/std": 0.43890365958213806, + "step": 1376 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3696.0, + "completions/mean_length": 874.4107666015625, + "completions/mean_terminated_length": 592.91259765625, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 12.867638483965015, + "grad_norm": 0.15715956687927246, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 820223433.0, + "reward": 0.731026828289032, + "reward_std": 0.13632294535636902, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1377 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3231.0, + "completions/mean_length": 962.0949096679688, + "completions/mean_terminated_length": 607.8273315429688, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 12.876967930029155, + "grad_norm": 0.149428591132164, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 820789486.0, + "reward": 0.6830357313156128, + "reward_std": 0.12043054401874542, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1378 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4047.0, + "completions/mean_length": 976.0078735351562, + "completions/mean_terminated_length": 640.4832763671875, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 12.886297376093294, + "grad_norm": 0.13740494847297668, + "learning_rate": 1e-06, + "loss": -0.0429, + "num_tokens": 821395133.0, + "reward": 0.6741071939468384, + "reward_std": 0.1324155330657959, + "rewards/verify_math_reward/mean": 0.6741071343421936, + "rewards/verify_math_reward/std": 0.46896928548812866, + "step": 1379 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3382.0, + "completions/mean_length": 960.3538208007812, + "completions/mean_terminated_length": 631.710205078125, + "completions/min_length": 191.0, + "completions/min_terminated_length": 191.0, + "epoch": 12.895626822157434, + "grad_norm": 0.15093958377838135, + "learning_rate": 1e-06, + "loss": -0.031, + "num_tokens": 821990906.0, + "reward": 0.7354910969734192, + "reward_std": 0.13177794218063354, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1380 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3692.0, + "completions/mean_length": 1004.1551513671875, + "completions/mean_terminated_length": 633.1337280273438, + "completions/min_length": 193.0, + "completions/min_terminated_length": 193.0, + "epoch": 12.904956268221575, + "grad_norm": 0.1666753888130188, + "learning_rate": 1e-06, + "loss": -0.063, + "num_tokens": 822588709.0, + "reward": 0.6886160969734192, + "reward_std": 0.16803108155727386, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1381 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4061.0, + "completions/mean_length": 1121.8873291015625, + "completions/mean_terminated_length": 692.6730346679688, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 12.914285714285715, + "grad_norm": 0.1525610238313675, + "learning_rate": 1e-06, + "loss": -0.0679, + "num_tokens": 823217000.0, + "reward": 0.6305803656578064, + "reward_std": 0.16311518847942352, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1382 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4064.0, + "completions/mean_length": 1127.591552734375, + "completions/mean_terminated_length": 655.2574462890625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 12.923615160349854, + "grad_norm": 0.13580629229545593, + "learning_rate": 1e-06, + "loss": -0.0732, + "num_tokens": 823822730.0, + "reward": 0.6495535969734192, + "reward_std": 0.13113926351070404, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1383 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3598.0, + "completions/mean_length": 1093.421875, + "completions/mean_terminated_length": 606.62255859375, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 12.932944606413994, + "grad_norm": 0.146683469414711, + "learning_rate": 1e-06, + "loss": -0.0332, + "num_tokens": 824380700.0, + "reward": 0.6852678656578064, + "reward_std": 0.11175382882356644, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1384 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3434.0, + "completions/mean_length": 991.0748291015625, + "completions/mean_terminated_length": 596.612548828125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 12.942274052478133, + "grad_norm": 0.14774489402770996, + "learning_rate": 1e-06, + "loss": -0.0709, + "num_tokens": 824940527.0, + "reward": 0.6908482313156128, + "reward_std": 0.12989819049835205, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1385 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2578.0, + "completions/mean_length": 1005.1160888671875, + "completions/mean_terminated_length": 629.8773803710938, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 12.951603498542275, + "grad_norm": 0.16044920682907104, + "learning_rate": 1e-06, + "loss": -0.0566, + "num_tokens": 825543711.0, + "reward": 0.6819196939468384, + "reward_std": 0.1305733323097229, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1386 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3962.0, + "completions/mean_length": 1091.2388916015625, + "completions/mean_terminated_length": 718.0, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 12.960932944606414, + "grad_norm": 0.14671629667282104, + "learning_rate": 1e-06, + "loss": -0.0497, + "num_tokens": 826203421.0, + "reward": 0.6495535969734192, + "reward_std": 0.14440304040908813, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1387 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3843.0, + "completions/mean_length": 939.0703735351562, + "completions/mean_terminated_length": 625.3140869140625, + "completions/min_length": 99.0, + "completions/min_terminated_length": 99.0, + "epoch": 12.970262390670554, + "grad_norm": 0.17984363436698914, + "learning_rate": 1e-06, + "loss": -0.0746, + "num_tokens": 826805996.0, + "reward": 0.715401828289032, + "reward_std": 0.1908092051744461, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1388 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3645.0, + "completions/mean_length": 910.4442138671875, + "completions/mean_terminated_length": 585.2275390625, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 12.979591836734693, + "grad_norm": 0.17228901386260986, + "learning_rate": 1e-06, + "loss": -0.0114, + "num_tokens": 827369682.0, + "reward": 0.6897321939468384, + "reward_std": 0.12490109354257584, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1389 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3901.0, + "completions/mean_length": 1046.32373046875, + "completions/mean_terminated_length": 610.6555786132812, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 12.988921282798835, + "grad_norm": 0.15714819729328156, + "learning_rate": 1e-06, + "loss": -0.085, + "num_tokens": 827930628.0, + "reward": 0.6718750596046448, + "reward_std": 0.13688497245311737, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1390 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3909.0, + "completions/mean_length": 929.755615234375, + "completions/mean_terminated_length": 593.58642578125, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 12.998250728862974, + "grad_norm": 0.1533220112323761, + "learning_rate": 1e-06, + "loss": -0.0457, + "num_tokens": 828507937.0, + "reward": 0.6941964626312256, + "reward_std": 0.12233986705541611, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1391 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.060267857142857095, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3688.0, + "completions/mean_length": 826.6495971679688, + "completions/mean_terminated_length": 616.9762573242188, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 13.00932944606414, + "grad_norm": 0.1632736474275589, + "learning_rate": 1e-06, + "loss": -0.0428, + "num_tokens": 829117855.0, + "reward": 0.7243303656578064, + "reward_std": 0.15762796998023987, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1392 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3512.0, + "completions/mean_length": 879.5558471679688, + "completions/mean_terminated_length": 577.1550903320312, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 13.018658892128279, + "grad_norm": 0.14047789573669434, + "learning_rate": 1e-06, + "loss": -0.0645, + "num_tokens": 829686993.0, + "reward": 0.7187500596046448, + "reward_std": 0.12752537429332733, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1393 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3976.0, + "completions/mean_length": 1028.6842041015625, + "completions/mean_terminated_length": 621.5183715820312, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 13.02798833819242, + "grad_norm": 0.14009352028369904, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 830271614.0, + "reward": 0.65625, + "reward_std": 0.13211244344711304, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1394 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3496.0, + "completions/mean_length": 991.107177734375, + "completions/mean_terminated_length": 596.6490478515625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.03731778425656, + "grad_norm": 0.16221703588962555, + "learning_rate": 1e-06, + "loss": -0.0716, + "num_tokens": 830850326.0, + "reward": 0.6540178656578064, + "reward_std": 0.13403315842151642, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1395 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2624.0, + "completions/mean_length": 858.443115234375, + "completions/mean_terminated_length": 554.0574340820312, + "completions/min_length": 93.0, + "completions/min_terminated_length": 93.0, + "epoch": 13.0466472303207, + "grad_norm": 0.1568409651517868, + "learning_rate": 1e-06, + "loss": -0.0438, + "num_tokens": 831398675.0, + "reward": 0.7276785969734192, + "reward_std": 0.1302698254585266, + "rewards/verify_math_reward/mean": 0.7276785969734192, + "rewards/verify_math_reward/std": 0.4454030692577362, + "step": 1396 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3687.0, + "completions/mean_length": 961.7969360351562, + "completions/mean_terminated_length": 607.4956665039062, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 13.055976676384839, + "grad_norm": 0.14957326650619507, + "learning_rate": 1e-06, + "loss": -0.0437, + "num_tokens": 831978341.0, + "reward": 0.6573660969734192, + "reward_std": 0.1219707801938057, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485533356666565, + "step": 1397 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4032.0, + "completions/mean_length": 1021.6160888671875, + "completions/mean_terminated_length": 617.9091186523438, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 13.06530612244898, + "grad_norm": 0.1544959843158722, + "learning_rate": 1e-06, + "loss": -0.0731, + "num_tokens": 832579173.0, + "reward": 0.6026785969734192, + "reward_std": 0.1418864130973816, + "rewards/verify_math_reward/mean": 0.6026785969734192, + "rewards/verify_math_reward/std": 0.48961687088012695, + "step": 1398 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3442.0, + "completions/mean_length": 899.35498046875, + "completions/mean_terminated_length": 603.0805053710938, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.07463556851312, + "grad_norm": 0.16523684561252594, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 833161611.0, + "reward": 0.6819196939468384, + "reward_std": 0.14864817261695862, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1399 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4062.0, + "completions/mean_length": 859.0748291015625, + "completions/mean_terminated_length": 639.1644897460938, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 13.08396501457726, + "grad_norm": 0.11886455863714218, + "learning_rate": 1e-06, + "loss": -0.0421, + "num_tokens": 833785710.0, + "reward": 0.7756696939468384, + "reward_std": 0.1016007736325264, + "rewards/verify_math_reward/mean": 0.7756696343421936, + "rewards/verify_math_reward/std": 0.41737356781959534, + "step": 1400 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2811.0, + "completions/mean_length": 887.872802734375, + "completions/mean_terminated_length": 611.7794189453125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.093294460641399, + "grad_norm": 0.15652163326740265, + "learning_rate": 1e-06, + "loss": -0.034, + "num_tokens": 834375548.0, + "reward": 0.7187500596046448, + "reward_std": 0.13816627860069275, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1401 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3886.0, + "completions/mean_length": 838.021240234375, + "completions/mean_terminated_length": 595.8213500976562, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.102623906705539, + "grad_norm": 0.13710899651050568, + "learning_rate": 1e-06, + "loss": -0.0558, + "num_tokens": 834946815.0, + "reward": 0.7566964626312256, + "reward_std": 0.11678708344697952, + "rewards/verify_math_reward/mean": 0.7566964030265808, + "rewards/verify_math_reward/std": 0.4293164908885956, + "step": 1402 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3294.0, + "completions/mean_length": 825.3438110351562, + "completions/mean_terminated_length": 573.7548217773438, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.11195335276968, + "grad_norm": 0.15589973330497742, + "learning_rate": 1e-06, + "loss": -0.0476, + "num_tokens": 835497803.0, + "reward": 0.754464328289032, + "reward_std": 0.11866319924592972, + "rewards/verify_math_reward/mean": 0.7544642686843872, + "rewards/verify_math_reward/std": 0.4306447505950928, + "step": 1403 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3312.0, + "completions/mean_length": 998.5870971679688, + "completions/mean_terminated_length": 626.8974609375, + "completions/min_length": 117.0, + "completions/min_terminated_length": 117.0, + "epoch": 13.12128279883382, + "grad_norm": 0.23238998651504517, + "learning_rate": 1e-06, + "loss": -0.064, + "num_tokens": 836095961.0, + "reward": 0.6785714626312256, + "reward_std": 0.13485801219940186, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1404 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3213.0, + "completions/mean_length": 1045.578125, + "completions/mean_terminated_length": 605.3511962890625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 13.130612244897959, + "grad_norm": 0.17254333198070526, + "learning_rate": 1e-06, + "loss": -0.049, + "num_tokens": 836667151.0, + "reward": 0.6395089626312256, + "reward_std": 0.12982404232025146, + "rewards/verify_math_reward/mean": 0.6395089030265808, + "rewards/verify_math_reward/std": 0.4804111123085022, + "step": 1405 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2894.0, + "completions/mean_length": 872.3125610351562, + "completions/mean_terminated_length": 577.8222045898438, + "completions/min_length": 98.0, + "completions/min_terminated_length": 98.0, + "epoch": 13.139941690962099, + "grad_norm": 0.18169696629047394, + "learning_rate": 1e-06, + "loss": -0.0343, + "num_tokens": 837220111.0, + "reward": 0.7187500596046448, + "reward_std": 0.07289712876081467, + "rewards/verify_math_reward/mean": 0.71875, + "rewards/verify_math_reward/std": 0.4498603343963623, + "step": 1406 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3636.0, + "completions/mean_length": 928.2645263671875, + "completions/mean_terminated_length": 651.4721069335938, + "completions/min_length": 172.0, + "completions/min_terminated_length": 172.0, + "epoch": 13.14927113702624, + "grad_norm": 0.13723637163639069, + "learning_rate": 1e-06, + "loss": -0.0573, + "num_tokens": 837847012.0, + "reward": 0.7299107313156128, + "reward_std": 0.14045538008213043, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1407 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3401.0, + "completions/mean_length": 917.5223388671875, + "completions/mean_terminated_length": 601.62451171875, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 13.15860058309038, + "grad_norm": 0.14928336441516876, + "learning_rate": 1e-06, + "loss": -0.0577, + "num_tokens": 838426848.0, + "reward": 0.7265625596046448, + "reward_std": 0.1219707801938057, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1408 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1417410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3822.0, + "completions/mean_length": 1123.40625, + "completions/mean_terminated_length": 632.4837646484375, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.167930029154519, + "grad_norm": 0.1406225562095642, + "learning_rate": 1e-06, + "loss": -0.0395, + "num_tokens": 839003572.0, + "reward": 0.5948660969734192, + "reward_std": 0.11787225306034088, + "rewards/verify_math_reward/mean": 0.5948660969734192, + "rewards/verify_math_reward/std": 0.49119213223457336, + "step": 1409 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3461.0, + "completions/mean_length": 898.3672485351562, + "completions/mean_terminated_length": 635.7596435546875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.177259475218658, + "grad_norm": 0.1319359987974167, + "learning_rate": 1e-06, + "loss": -0.0272, + "num_tokens": 839618461.0, + "reward": 0.7299107313156128, + "reward_std": 0.11475647985935211, + "rewards/verify_math_reward/mean": 0.7299107313156128, + "rewards/verify_math_reward/std": 0.44425368309020996, + "step": 1410 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2531.0, + "completions/mean_length": 922.4910888671875, + "completions/mean_terminated_length": 576.8613891601562, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 13.186588921282798, + "grad_norm": 0.23846296966075897, + "learning_rate": 1e-06, + "loss": -0.056, + "num_tokens": 840167165.0, + "reward": 0.7176339626312256, + "reward_std": 0.12181880325078964, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1411 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2675.0, + "completions/mean_length": 930.5123291015625, + "completions/mean_terminated_length": 607.3444213867188, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.19591836734694, + "grad_norm": 0.14456422626972198, + "learning_rate": 1e-06, + "loss": -0.0463, + "num_tokens": 840752856.0, + "reward": 0.7087053656578064, + "reward_std": 0.1512710452079773, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1412 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3285.0, + "completions/mean_length": 930.2042846679688, + "completions/mean_terminated_length": 624.0869140625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 13.205247813411079, + "grad_norm": 0.19330009818077087, + "learning_rate": 1e-06, + "loss": -0.0678, + "num_tokens": 841355823.0, + "reward": 0.6986607313156128, + "reward_std": 0.18272772431373596, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960144996643, + "step": 1413 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3836.0, + "completions/mean_length": 1008.7332763671875, + "completions/mean_terminated_length": 676.72802734375, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 13.214577259475218, + "grad_norm": 0.1682826429605484, + "learning_rate": 1e-06, + "loss": -0.0619, + "num_tokens": 841996440.0, + "reward": 0.6796875596046448, + "reward_std": 0.16822652518749237, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1414 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0613839285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3950.0, + "completions/mean_length": 791.9163208007812, + "completions/mean_terminated_length": 575.834716796875, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.223906705539358, + "grad_norm": 0.13416460156440735, + "learning_rate": 1e-06, + "loss": -0.0418, + "num_tokens": 842565869.0, + "reward": 0.7343750596046448, + "reward_std": 0.12125540524721146, + "rewards/verify_math_reward/mean": 0.734375, + "rewards/verify_math_reward/std": 0.44191211462020874, + "step": 1415 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3180.0, + "completions/mean_length": 858.8438110351562, + "completions/mean_terminated_length": 588.7545776367188, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 13.2332361516035, + "grad_norm": 0.14242775738239288, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 843138265.0, + "reward": 0.7042410969734192, + "reward_std": 0.12171115726232529, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1416 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2674.0, + "completions/mean_length": 927.5725708007812, + "completions/mean_terminated_length": 573.7778930664062, + "completions/min_length": 115.0, + "completions/min_terminated_length": 115.0, + "epoch": 13.242565597667639, + "grad_norm": 0.14249292016029358, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 843691602.0, + "reward": 0.6796875596046448, + "reward_std": 0.11152419447898865, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1417 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2532.0, + "completions/mean_length": 949.638427734375, + "completions/mean_terminated_length": 589.60693359375, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.251895043731778, + "grad_norm": 0.16333496570587158, + "learning_rate": 1e-06, + "loss": -0.0805, + "num_tokens": 844246910.0, + "reward": 0.7511160969734192, + "reward_std": 0.13842660188674927, + "rewards/verify_math_reward/mean": 0.7511160969734192, + "rewards/verify_math_reward/std": 0.43260788917541504, + "step": 1418 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3844.0, + "completions/mean_length": 925.0324096679688, + "completions/mean_terminated_length": 570.9541015625, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 13.261224489795918, + "grad_norm": 0.17214609682559967, + "learning_rate": 1e-06, + "loss": -0.0888, + "num_tokens": 844791115.0, + "reward": 0.7500000596046448, + "reward_std": 0.15195433795452118, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1419 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0747767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3317.0, + "completions/mean_length": 858.2935791015625, + "completions/mean_terminated_length": 596.6212158203125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 13.270553935860057, + "grad_norm": 0.16037645936012268, + "learning_rate": 1e-06, + "loss": -0.0496, + "num_tokens": 845377314.0, + "reward": 0.6863839626312256, + "reward_std": 0.13793805241584778, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422141790390015, + "step": 1420 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3612.0, + "completions/mean_length": 987.9788208007812, + "completions/mean_terminated_length": 623.697021484375, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.279883381924199, + "grad_norm": 0.14235125482082367, + "learning_rate": 1e-06, + "loss": -0.036, + "num_tokens": 845963743.0, + "reward": 0.699776828289032, + "reward_std": 0.11039625108242035, + "rewards/verify_math_reward/mean": 0.6997767686843872, + "rewards/verify_math_reward/std": 0.4586109220981598, + "step": 1421 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1372767857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3372.0, + "completions/mean_length": 1091.1942138671875, + "completions/mean_terminated_length": 613.0685424804688, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.289212827988338, + "grad_norm": 0.1667003035545349, + "learning_rate": 1e-06, + "loss": -0.0608, + "num_tokens": 846529957.0, + "reward": 0.6484375, + "reward_std": 0.13816697895526886, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1422 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2764.0, + "completions/mean_length": 1032.35498046875, + "completions/mean_terminated_length": 625.6763916015625, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 13.298542274052478, + "grad_norm": 0.1497279405593872, + "learning_rate": 1e-06, + "loss": -0.0767, + "num_tokens": 847120651.0, + "reward": 0.6350446939468384, + "reward_std": 0.1537569761276245, + "rewards/verify_math_reward/mean": 0.6350446343421936, + "rewards/verify_math_reward/std": 0.481686532497406, + "step": 1423 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3786.0, + "completions/mean_length": 1071.6529541015625, + "completions/mean_terminated_length": 678.8309936523438, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.307871720116617, + "grad_norm": 0.14686422049999237, + "learning_rate": 1e-06, + "loss": -0.0587, + "num_tokens": 847752020.0, + "reward": 0.6339285969734192, + "reward_std": 0.155793234705925, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1424 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1395089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3958.0, + "completions/mean_length": 1132.51123046875, + "completions/mean_terminated_length": 652.04931640625, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.317201166180759, + "grad_norm": 0.16825416684150696, + "learning_rate": 1e-06, + "loss": -0.0581, + "num_tokens": 848348678.0, + "reward": 0.6339285969734192, + "reward_std": 0.1474096179008484, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1425 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4048.0, + "completions/mean_length": 1094.0692138671875, + "completions/mean_terminated_length": 656.4475708007812, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.326530612244898, + "grad_norm": 0.14060723781585693, + "learning_rate": 1e-06, + "loss": -0.0413, + "num_tokens": 848955428.0, + "reward": 0.6272321939468384, + "reward_std": 0.12580150365829468, + "rewards/verify_math_reward/mean": 0.6272321343421936, + "rewards/verify_math_reward/std": 0.4838111698627472, + "step": 1426 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3664.0, + "completions/mean_length": 993.2578735351562, + "completions/mean_terminated_length": 633.9115600585938, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 13.335860058309038, + "grad_norm": 0.16338886320590973, + "learning_rate": 1e-06, + "loss": -0.052, + "num_tokens": 849556027.0, + "reward": 0.6484375, + "reward_std": 0.14583592116832733, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1427 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3247.0, + "completions/mean_length": 1054.42529296875, + "completions/mean_terminated_length": 718.9851684570312, + "completions/min_length": 147.0, + "completions/min_terminated_length": 147.0, + "epoch": 13.345189504373177, + "grad_norm": 0.14065629243850708, + "learning_rate": 1e-06, + "loss": -0.0241, + "num_tokens": 850232464.0, + "reward": 0.6595982313156128, + "reward_std": 0.1428193300962448, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1428 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4071.0, + "completions/mean_length": 1020.7991333007812, + "completions/mean_terminated_length": 594.8817749023438, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.354518950437317, + "grad_norm": 0.1557115614414215, + "learning_rate": 1e-06, + "loss": -0.065, + "num_tokens": 850787292.0, + "reward": 0.7165178656578064, + "reward_std": 0.13812169432640076, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1429 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1216517857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3524.0, + "completions/mean_length": 1084.28125, + "completions/mean_terminated_length": 667.15625, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 13.363848396501458, + "grad_norm": 0.1528123915195465, + "learning_rate": 1e-06, + "loss": -0.0784, + "num_tokens": 851415728.0, + "reward": 0.6897321939468384, + "reward_std": 0.15936000645160675, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1430 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3202.0, + "completions/mean_length": 1054.029052734375, + "completions/mean_terminated_length": 637.109130859375, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 13.373177842565598, + "grad_norm": 0.14892137050628662, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 852005378.0, + "reward": 0.6573660969734192, + "reward_std": 0.11997338384389877, + "rewards/verify_math_reward/mean": 0.6573660969734192, + "rewards/verify_math_reward/std": 0.47485536336898804, + "step": 1431 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4039.0, + "completions/mean_length": 1063.009033203125, + "completions/mean_terminated_length": 677.685546875, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 13.382507288629737, + "grad_norm": 0.1488819569349289, + "learning_rate": 1e-06, + "loss": -0.0715, + "num_tokens": 852627586.0, + "reward": 0.6941964626312256, + "reward_std": 0.1514671891927719, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1432 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3492.0, + "completions/mean_length": 894.239990234375, + "completions/mean_terminated_length": 584.6450805664062, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 13.391836734693877, + "grad_norm": 0.15225546061992645, + "learning_rate": 1e-06, + "loss": -0.0445, + "num_tokens": 853189865.0, + "reward": 0.6830357313156128, + "reward_std": 0.14011767506599426, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1433 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1383928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2406.0, + "completions/mean_length": 1085.810302734375, + "completions/mean_terminated_length": 602.3082885742188, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.401166180758018, + "grad_norm": 0.1438121348619461, + "learning_rate": 1e-06, + "loss": -0.0748, + "num_tokens": 853751103.0, + "reward": 0.707589328289032, + "reward_std": 0.09840943664312363, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1434 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3750.0, + "completions/mean_length": 1068.7645263671875, + "completions/mean_terminated_length": 631.8837890625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.410495626822158, + "grad_norm": 0.1605885773897171, + "learning_rate": 1e-06, + "loss": -0.0836, + "num_tokens": 854343068.0, + "reward": 0.652901828289032, + "reward_std": 0.15026073157787323, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 1435 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3708.0, + "completions/mean_length": 910.2188110351562, + "completions/mean_terminated_length": 610.7008666992188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 13.419825072886297, + "grad_norm": 0.1605662703514099, + "learning_rate": 1e-06, + "loss": -0.0543, + "num_tokens": 854927712.0, + "reward": 0.7879464626312256, + "reward_std": 0.15296350419521332, + "rewards/verify_math_reward/mean": 0.7879464030265808, + "rewards/verify_math_reward/std": 0.40899088978767395, + "step": 1436 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3265.0, + "completions/mean_length": 973.99560546875, + "completions/mean_terminated_length": 603.7203369140625, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.429154518950437, + "grad_norm": 0.163425475358963, + "learning_rate": 1e-06, + "loss": -0.069, + "num_tokens": 855500516.0, + "reward": 0.6897321939468384, + "reward_std": 0.14478282630443573, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.4628615975379944, + "step": 1437 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2800.0, + "completions/mean_length": 929.0402221679688, + "completions/mean_terminated_length": 601.4236450195312, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.438483965014576, + "grad_norm": 0.12399692088365555, + "learning_rate": 1e-06, + "loss": -0.0664, + "num_tokens": 856071400.0, + "reward": 0.7087053656578064, + "reward_std": 0.10457022488117218, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1438 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3001.0, + "completions/mean_length": 845.7232666015625, + "completions/mean_terminated_length": 561.7184448242188, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 13.447813411078718, + "grad_norm": 0.17261086404323578, + "learning_rate": 1e-06, + "loss": -0.0083, + "num_tokens": 856618584.0, + "reward": 0.7366071939468384, + "reward_std": 0.11922702938318253, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1439 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3761.0, + "completions/mean_length": 864.271240234375, + "completions/mean_terminated_length": 594.6348266601562, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "epoch": 13.457142857142857, + "grad_norm": 0.15159589052200317, + "learning_rate": 1e-06, + "loss": -0.0296, + "num_tokens": 857199651.0, + "reward": 0.7142857313156128, + "reward_std": 0.13005226850509644, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1440 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3543.0, + "completions/mean_length": 1053.7734375, + "completions/mean_terminated_length": 667.2767333984375, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 13.466472303206997, + "grad_norm": 0.186456099152565, + "learning_rate": 1e-06, + "loss": -0.0721, + "num_tokens": 857821640.0, + "reward": 0.65625, + "reward_std": 0.1810370534658432, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1441 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3691.0, + "completions/mean_length": 1015.8035888671875, + "completions/mean_terminated_length": 628.84423828125, + "completions/min_length": 200.0, + "completions/min_terminated_length": 200.0, + "epoch": 13.475801749271136, + "grad_norm": 0.17106060683727264, + "learning_rate": 1e-06, + "loss": -0.0215, + "num_tokens": 858413056.0, + "reward": 0.6808035969734192, + "reward_std": 0.12843577563762665, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1442 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4063.0, + "completions/mean_length": 892.716552734375, + "completions/mean_terminated_length": 650.4513549804688, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 13.485131195335278, + "grad_norm": 0.12468821555376053, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 859033426.0, + "reward": 0.7109375596046448, + "reward_std": 0.12238264828920364, + "rewards/verify_math_reward/mean": 0.7109375, + "rewards/verify_math_reward/std": 0.45358020067214966, + "step": 1443 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3419.0, + "completions/mean_length": 966.1685791015625, + "completions/mean_terminated_length": 620.9950561523438, + "completions/min_length": 129.0, + "completions/min_terminated_length": 129.0, + "epoch": 13.494460641399417, + "grad_norm": 0.16026073694229126, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 859624673.0, + "reward": 0.7220982313156128, + "reward_std": 0.1411721557378769, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1444 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3975.0, + "completions/mean_length": 915.6350708007812, + "completions/mean_terminated_length": 670.9916381835938, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 13.503790087463557, + "grad_norm": 0.15779909491539001, + "learning_rate": 1e-06, + "loss": -0.0256, + "num_tokens": 860265330.0, + "reward": 0.7366071939468384, + "reward_std": 0.12437821924686432, + "rewards/verify_math_reward/mean": 0.7366071343421936, + "rewards/verify_math_reward/std": 0.44071969389915466, + "step": 1445 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3593.0, + "completions/mean_length": 985.2734985351562, + "completions/mean_terminated_length": 667.6961669921875, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 13.513119533527696, + "grad_norm": 0.1605423092842102, + "learning_rate": 1e-06, + "loss": -0.0334, + "num_tokens": 860903607.0, + "reward": 0.6662946939468384, + "reward_std": 0.1523348093032837, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1446 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3397.0, + "completions/mean_length": 1098.388427734375, + "completions/mean_terminated_length": 656.9987182617188, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 13.522448979591836, + "grad_norm": 0.17029817402362823, + "learning_rate": 1e-06, + "loss": -0.0491, + "num_tokens": 861518339.0, + "reward": 0.6305803656578064, + "reward_std": 0.16506797075271606, + "rewards/verify_math_reward/mean": 0.6305803656578064, + "rewards/verify_math_reward/std": 0.4829172194004059, + "step": 1447 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1328125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3551.0, + "completions/mean_length": 1054.7623291015625, + "completions/mean_terminated_length": 588.9871215820312, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 13.531778425655977, + "grad_norm": 0.15654556453227997, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 862069470.0, + "reward": 0.6729910969734192, + "reward_std": 0.12410903722047806, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1448 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1082589285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3611.0, + "completions/mean_length": 1044.3092041015625, + "completions/mean_terminated_length": 673.8285522460938, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 13.541107871720117, + "grad_norm": 0.1570572406053543, + "learning_rate": 1e-06, + "loss": -0.07, + "num_tokens": 862699747.0, + "reward": 0.6886160969734192, + "reward_std": 0.15608815848827362, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331802010536194, + "step": 1449 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3337.0, + "completions/mean_length": 908.3281860351562, + "completions/mean_terminated_length": 604.3692016601562, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 13.550437317784256, + "grad_norm": 0.15788523852825165, + "learning_rate": 1e-06, + "loss": -0.0398, + "num_tokens": 863287681.0, + "reward": 0.660714328289032, + "reward_std": 0.10303223878145218, + "rewards/verify_math_reward/mean": 0.6607142686843872, + "rewards/verify_math_reward/std": 0.4737313687801361, + "step": 1450 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0848214285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2636.0, + "completions/mean_length": 912.950927734375, + "completions/mean_terminated_length": 617.9365844726562, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.559766763848396, + "grad_norm": 0.16280454397201538, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 863877293.0, + "reward": 0.7321428656578064, + "reward_std": 0.16915832459926605, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1451 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3650.0, + "completions/mean_length": 1136.6898193359375, + "completions/mean_terminated_length": 670.2351684570312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 13.569096209912537, + "grad_norm": 0.1474280208349228, + "learning_rate": 1e-06, + "loss": -0.0678, + "num_tokens": 864480399.0, + "reward": 0.6339285969734192, + "reward_std": 0.11678457260131836, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199838399887085, + "step": 1452 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3226.0, + "completions/mean_length": 1057.013427734375, + "completions/mean_terminated_length": 640.5025024414062, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 13.578425655976677, + "grad_norm": 0.1564379334449768, + "learning_rate": 1e-06, + "loss": -0.0835, + "num_tokens": 865072915.0, + "reward": 0.6908482313156128, + "reward_std": 0.14920946955680847, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1453 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3815.0, + "completions/mean_length": 977.0748291015625, + "completions/mean_terminated_length": 662.88330078125, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 13.587755102040816, + "grad_norm": 0.15317924320697784, + "learning_rate": 1e-06, + "loss": -0.0567, + "num_tokens": 865701670.0, + "reward": 0.7064732313156128, + "reward_std": 0.1304224729537964, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1454 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3483.0, + "completions/mean_length": 996.0045166015625, + "completions/mean_terminated_length": 645.5701904296875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 13.597084548104956, + "grad_norm": 0.1633215993642807, + "learning_rate": 1e-06, + "loss": -0.0281, + "num_tokens": 866312442.0, + "reward": 0.6618303656578064, + "reward_std": 0.15796427428722382, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1455 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3427.0, + "completions/mean_length": 1022.060302734375, + "completions/mean_terminated_length": 678.8163452148438, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 13.606413994169095, + "grad_norm": 0.14086061716079712, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 866951320.0, + "reward": 0.7354910969734192, + "reward_std": 0.14676952362060547, + "rewards/verify_math_reward/mean": 0.7354910969734192, + "rewards/verify_math_reward/std": 0.44131770730018616, + "step": 1456 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3705.0, + "completions/mean_length": 1011.2623291015625, + "completions/mean_terminated_length": 614.9861450195312, + "completions/min_length": 106.0, + "completions/min_terminated_length": 106.0, + "epoch": 13.615743440233237, + "grad_norm": 0.1431649923324585, + "learning_rate": 1e-06, + "loss": -0.0486, + "num_tokens": 867520067.0, + "reward": 0.731026828289032, + "reward_std": 0.10867238789796829, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1457 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3228.0, + "completions/mean_length": 917.7723388671875, + "completions/mean_terminated_length": 601.8994140625, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.625072886297376, + "grad_norm": 0.1757386475801468, + "learning_rate": 1e-06, + "loss": -0.0228, + "num_tokens": 868100927.0, + "reward": 0.7053571939468384, + "reward_std": 0.17415978014469147, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1458 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2742.0, + "completions/mean_length": 972.0457763671875, + "completions/mean_terminated_length": 648.8780517578125, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.634402332361516, + "grad_norm": 0.1653522402048111, + "learning_rate": 1e-06, + "loss": -0.013, + "num_tokens": 868736184.0, + "reward": 0.6729910969734192, + "reward_std": 0.11475831270217896, + "rewards/verify_math_reward/mean": 0.6729910969734192, + "rewards/verify_math_reward/std": 0.46938255429267883, + "step": 1459 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1037946428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3802.0, + "completions/mean_length": 1012.4241333007812, + "completions/mean_terminated_length": 655.297607421875, + "completions/min_length": 167.0, + "completions/min_terminated_length": 167.0, + "epoch": 13.643731778425655, + "grad_norm": 0.16010451316833496, + "learning_rate": 1e-06, + "loss": -0.0746, + "num_tokens": 869352692.0, + "reward": 0.7020089626312256, + "reward_std": 0.16356277465820312, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1460 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3336.0, + "completions/mean_length": 1074.751220703125, + "completions/mean_terminated_length": 643.1441040039062, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 13.653061224489797, + "grad_norm": 0.16032300889492035, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 869947533.0, + "reward": 0.6316964626312256, + "reward_std": 0.13872897624969482, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1461 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3307.0, + "completions/mean_length": 945.7076416015625, + "completions/mean_terminated_length": 602.6064453125, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 13.662390670553936, + "grad_norm": 0.14728675782680511, + "learning_rate": 1e-06, + "loss": -0.0676, + "num_tokens": 870517895.0, + "reward": 0.7287946939468384, + "reward_std": 0.12482398748397827, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1462 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3459.0, + "completions/mean_length": 902.8013916015625, + "completions/mean_terminated_length": 585.4404907226562, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 13.671720116618076, + "grad_norm": 0.14219607412815094, + "learning_rate": 1e-06, + "loss": -0.0514, + "num_tokens": 871077213.0, + "reward": 0.7053571939468384, + "reward_std": 0.1251595914363861, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613667368888855, + "step": 1463 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3482.0, + "completions/mean_length": 976.5480346679688, + "completions/mean_terminated_length": 602.2137451171875, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 13.681049562682215, + "grad_norm": 0.18661396205425262, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 871645888.0, + "reward": 0.7020089626312256, + "reward_std": 0.16724969446659088, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763102173805237, + "step": 1464 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3807.0, + "completions/mean_length": 978.65185546875, + "completions/mean_terminated_length": 604.5700073242188, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 13.690379008746355, + "grad_norm": 0.1835251897573471, + "learning_rate": 1e-06, + "loss": -0.0592, + "num_tokens": 872212392.0, + "reward": 0.7142857313156128, + "reward_std": 0.12903308868408203, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1465 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3338.0, + "completions/mean_length": 1081.33154296875, + "completions/mean_terminated_length": 702.604248046875, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 13.699708454810496, + "grad_norm": 0.15788906812667847, + "learning_rate": 1e-06, + "loss": -0.0742, + "num_tokens": 872876169.0, + "reward": 0.621651828289032, + "reward_std": 0.1579628735780716, + "rewards/verify_math_reward/mean": 0.6216517686843872, + "rewards/verify_math_reward/std": 0.4852459728717804, + "step": 1466 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0647321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3220.0, + "completions/mean_length": 798.4631958007812, + "completions/mean_terminated_length": 570.2327270507812, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 13.709037900874636, + "grad_norm": 0.16740481555461884, + "learning_rate": 1e-06, + "loss": -0.0313, + "num_tokens": 873449800.0, + "reward": 0.7421875596046448, + "reward_std": 0.1370365023612976, + "rewards/verify_math_reward/mean": 0.7421875, + "rewards/verify_math_reward/std": 0.43767455220222473, + "step": 1467 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3885.0, + "completions/mean_length": 1040.068115234375, + "completions/mean_terminated_length": 664.7781982421875, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 13.718367346938775, + "grad_norm": 0.1412292867898941, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 874073141.0, + "reward": 0.6484375, + "reward_std": 0.12384689599275589, + "rewards/verify_math_reward/mean": 0.6484375, + "rewards/verify_math_reward/std": 0.4777248501777649, + "step": 1468 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3963.0, + "completions/mean_length": 983.60498046875, + "completions/mean_terminated_length": 644.6311645507812, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.727696793002915, + "grad_norm": 0.1517496556043625, + "learning_rate": 1e-06, + "loss": -0.0367, + "num_tokens": 874679787.0, + "reward": 0.6495535969734192, + "reward_std": 0.13794013857841492, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1469 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3746.0, + "completions/mean_length": 948.3605346679688, + "completions/mean_terminated_length": 627.0147705078125, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 13.737026239067056, + "grad_norm": 0.15627431869506836, + "learning_rate": 1e-06, + "loss": -0.0391, + "num_tokens": 875281398.0, + "reward": 0.6640625, + "reward_std": 0.1344454288482666, + "rewards/verify_math_reward/mean": 0.6640625, + "rewards/verify_math_reward/std": 0.4725809693336487, + "step": 1470 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2342.0, + "completions/mean_length": 819.0000610351562, + "completions/mean_terminated_length": 566.923095703125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.746355685131196, + "grad_norm": 0.15194284915924072, + "learning_rate": 1e-06, + "loss": -0.0515, + "num_tokens": 875840822.0, + "reward": 0.7678571939468384, + "reward_std": 0.12189479172229767, + "rewards/verify_math_reward/mean": 0.7678571343421936, + "rewards/verify_math_reward/std": 0.422435462474823, + "step": 1471 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3322.0, + "completions/mean_length": 1038.6138916015625, + "completions/mean_terminated_length": 632.76611328125, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 13.755685131195335, + "grad_norm": 0.16416384279727936, + "learning_rate": 1e-06, + "loss": -0.0865, + "num_tokens": 876432356.0, + "reward": 0.6964285969734192, + "reward_std": 0.15308116376399994, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565433502197, + "step": 1472 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4079.0, + "completions/mean_length": 985.5100708007812, + "completions/mean_terminated_length": 655.260498046875, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 13.765014577259475, + "grad_norm": 0.16349850594997406, + "learning_rate": 1e-06, + "loss": -0.0664, + "num_tokens": 877059805.0, + "reward": 0.6941964626312256, + "reward_std": 0.1497408002614975, + "rewards/verify_math_reward/mean": 0.6941964030265808, + "rewards/verify_math_reward/std": 0.4610042870044708, + "step": 1473 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2843.0, + "completions/mean_length": 976.6998291015625, + "completions/mean_terminated_length": 666.6834106445312, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 13.774344023323614, + "grad_norm": 0.161646768450737, + "learning_rate": 1e-06, + "loss": -0.0324, + "num_tokens": 877687936.0, + "reward": 0.6964285969734192, + "reward_std": 0.15744292736053467, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1474 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3996.0, + "completions/mean_length": 881.3717041015625, + "completions/mean_terminated_length": 591.9768676757812, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 13.783673469387756, + "grad_norm": 0.1437355875968933, + "learning_rate": 1e-06, + "loss": -0.0468, + "num_tokens": 878264893.0, + "reward": 0.7287946939468384, + "reward_std": 0.11329153925180435, + "rewards/verify_math_reward/mean": 0.7287946343421936, + "rewards/verify_math_reward/std": 0.44483017921447754, + "step": 1475 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3344.0, + "completions/mean_length": 946.1663208007812, + "completions/mean_terminated_length": 624.5965576171875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 13.793002915451895, + "grad_norm": 0.14176473021507263, + "learning_rate": 1e-06, + "loss": -0.0516, + "num_tokens": 878856338.0, + "reward": 0.7142857313156128, + "reward_std": 0.13808958232402802, + "rewards/verify_math_reward/mean": 0.7142857313156128, + "rewards/verify_math_reward/std": 0.4520062506198883, + "step": 1476 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3973.0, + "completions/mean_length": 1051.5457763671875, + "completions/mean_terminated_length": 703.1753540039062, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 13.802332361516035, + "grad_norm": 0.13618813455104828, + "learning_rate": 1e-06, + "loss": -0.0426, + "num_tokens": 879514187.0, + "reward": 0.6551339626312256, + "reward_std": 0.1220463290810585, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900800228119, + "step": 1477 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3613.0, + "completions/mean_length": 898.9375610351562, + "completions/mean_terminated_length": 594.0831298828125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 13.811661807580174, + "grad_norm": 0.14762429893016815, + "learning_rate": 1e-06, + "loss": -0.0355, + "num_tokens": 880085603.0, + "reward": 0.7120535969734192, + "reward_std": 0.12110203504562378, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1478 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2666.0, + "completions/mean_length": 927.2422485351562, + "completions/mean_terminated_length": 612.3104248046875, + "completions/min_length": 186.0, + "completions/min_terminated_length": 186.0, + "epoch": 13.820991253644316, + "grad_norm": 0.15905015170574188, + "learning_rate": 1e-06, + "loss": -0.0393, + "num_tokens": 880669020.0, + "reward": 0.6897321939468384, + "reward_std": 0.13395646214485168, + "rewards/verify_math_reward/mean": 0.6897321343421936, + "rewards/verify_math_reward/std": 0.462861567735672, + "step": 1479 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3506.0, + "completions/mean_length": 874.7344360351562, + "completions/mean_terminated_length": 605.9710083007812, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 13.830320699708455, + "grad_norm": 0.13879813253879547, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 881253030.0, + "reward": 0.7444196939468384, + "reward_std": 0.10660082846879959, + "rewards/verify_math_reward/mean": 0.7444196343421936, + "rewards/verify_math_reward/std": 0.43643057346343994, + "step": 1480 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3411.0, + "completions/mean_length": 1185.94091796875, + "completions/mean_terminated_length": 727.2493896484375, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 13.839650145772595, + "grad_norm": 0.16253475844860077, + "learning_rate": 1e-06, + "loss": -0.0563, + "num_tokens": 881916729.0, + "reward": 0.5870535969734192, + "reward_std": 0.1605992615222931, + "rewards/verify_math_reward/mean": 0.5870535969734192, + "rewards/verify_math_reward/std": 0.49263834953308105, + "step": 1481 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3404.0, + "completions/mean_length": 1024.84716796875, + "completions/mean_terminated_length": 612.7683715820312, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 13.848979591836734, + "grad_norm": 0.15938550233840942, + "learning_rate": 1e-06, + "loss": -0.0295, + "num_tokens": 882496352.0, + "reward": 0.6227678656578064, + "reward_std": 0.14042216539382935, + "rewards/verify_math_reward/mean": 0.6227678656578064, + "rewards/verify_math_reward/std": 0.4849644601345062, + "step": 1482 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0837053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3874.0, + "completions/mean_length": 874.0469360351562, + "completions/mean_terminated_length": 579.7150268554688, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.858309037900874, + "grad_norm": 0.14581868052482605, + "learning_rate": 1e-06, + "loss": -0.0017, + "num_tokens": 883052866.0, + "reward": 0.7087053656578064, + "reward_std": 0.09743304550647736, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1483 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3739.0, + "completions/mean_length": 1041.9007568359375, + "completions/mean_terminated_length": 614.482177734375, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 13.867638483965015, + "grad_norm": 0.13536785542964935, + "learning_rate": 1e-06, + "loss": -0.0556, + "num_tokens": 883627241.0, + "reward": 0.7053571939468384, + "reward_std": 0.11291245371103287, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1484 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3783.0, + "completions/mean_length": 867.2098388671875, + "completions/mean_terminated_length": 602.04345703125, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.876967930029155, + "grad_norm": 0.17088043689727783, + "learning_rate": 1e-06, + "loss": -0.0403, + "num_tokens": 884206349.0, + "reward": 0.7176339626312256, + "reward_std": 0.1381234973669052, + "rewards/verify_math_reward/mean": 0.7176339030265808, + "rewards/verify_math_reward/std": 0.4504019320011139, + "step": 1485 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3833.0, + "completions/mean_length": 930.98779296875, + "completions/mean_terminated_length": 662.766357421875, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 13.886297376093294, + "grad_norm": 0.12953662872314453, + "learning_rate": 1e-06, + "loss": -0.0171, + "num_tokens": 884833874.0, + "reward": 0.6964285969734192, + "reward_std": 0.10821662098169327, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1486 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3235.0, + "completions/mean_length": 1033.15966796875, + "completions/mean_terminated_length": 626.587890625, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 13.895626822157434, + "grad_norm": 0.14354275166988373, + "learning_rate": 1e-06, + "loss": -0.0606, + "num_tokens": 885421297.0, + "reward": 0.6618303656578064, + "reward_std": 0.14222201704978943, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1487 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0680803571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3465.0, + "completions/mean_length": 901.1842041015625, + "completions/mean_terminated_length": 667.7904052734375, + "completions/min_length": 126.0, + "completions/min_terminated_length": 126.0, + "epoch": 13.904956268221575, + "grad_norm": 0.15612778067588806, + "learning_rate": 1e-06, + "loss": -0.0624, + "num_tokens": 886075854.0, + "reward": 0.707589328289032, + "reward_std": 0.14905792474746704, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1488 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3168.0, + "completions/mean_length": 985.55029296875, + "completions/mean_terminated_length": 633.9341430664062, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 13.914285714285715, + "grad_norm": 0.15508072078227997, + "learning_rate": 1e-06, + "loss": -0.073, + "num_tokens": 886671987.0, + "reward": 0.6316964626312256, + "reward_std": 0.15044689178466797, + "rewards/verify_math_reward/mean": 0.6316964030265808, + "rewards/verify_math_reward/std": 0.4826137125492096, + "step": 1489 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3904.0, + "completions/mean_length": 1040.9029541015625, + "completions/mean_terminated_length": 712.3572387695312, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 13.923615160349854, + "grad_norm": 0.16549524664878845, + "learning_rate": 1e-06, + "loss": -0.0473, + "num_tokens": 887327908.0, + "reward": 0.6551339626312256, + "reward_std": 0.17097096145153046, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 1490 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3733.0, + "completions/mean_length": 965.0402221679688, + "completions/mean_terminated_length": 641.1477661132812, + "completions/min_length": 130.0, + "completions/min_terminated_length": 130.0, + "epoch": 13.932944606413994, + "grad_norm": 0.1526726633310318, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 887941968.0, + "reward": 0.6863839626312256, + "reward_std": 0.13098520040512085, + "rewards/verify_math_reward/mean": 0.6863839030265808, + "rewards/verify_math_reward/std": 0.46422144770622253, + "step": 1491 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3309.0, + "completions/mean_length": 993.091552734375, + "completions/mean_terminated_length": 684.7042846679688, + "completions/min_length": 181.0, + "completions/min_terminated_length": 181.0, + "epoch": 13.942274052478133, + "grad_norm": 0.15932045876979828, + "learning_rate": 1e-06, + "loss": -0.0652, + "num_tokens": 888590994.0, + "reward": 0.6618303656578064, + "reward_std": 0.18603669106960297, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1492 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0580357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3744.0, + "completions/mean_length": 761.107177734375, + "completions/mean_terminated_length": 555.6398315429688, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 13.951603498542275, + "grad_norm": 0.14150334894657135, + "learning_rate": 1e-06, + "loss": -0.0466, + "num_tokens": 889145178.0, + "reward": 0.7823660969734192, + "reward_std": 0.10716605931520462, + "rewards/verify_math_reward/mean": 0.7823660969734192, + "rewards/verify_math_reward/std": 0.41286757588386536, + "step": 1493 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3533.0, + "completions/mean_length": 915.6094360351562, + "completions/mean_terminated_length": 612.3447875976562, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 13.960932944606414, + "grad_norm": 0.14166150987148285, + "learning_rate": 1e-06, + "loss": -0.0347, + "num_tokens": 889728700.0, + "reward": 0.7500000596046448, + "reward_std": 0.11370522528886795, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1494 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2884.0, + "completions/mean_length": 991.9386596679688, + "completions/mean_terminated_length": 628.1209716796875, + "completions/min_length": 164.0, + "completions/min_terminated_length": 164.0, + "epoch": 13.970262390670554, + "grad_norm": 0.1634577065706253, + "learning_rate": 1e-06, + "loss": -0.0994, + "num_tokens": 890322373.0, + "reward": 0.7042410969734192, + "reward_std": 0.15680170059204102, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1495 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3959.0, + "completions/mean_length": 888.6842041015625, + "completions/mean_terminated_length": 569.9202270507812, + "completions/min_length": 177.0, + "completions/min_terminated_length": 177.0, + "epoch": 13.979591836734693, + "grad_norm": 0.1585603803396225, + "learning_rate": 1e-06, + "loss": -0.0474, + "num_tokens": 890871394.0, + "reward": 0.7165178656578064, + "reward_std": 0.12414928525686264, + "rewards/verify_math_reward/mean": 0.7165178656578064, + "rewards/verify_math_reward/std": 0.4509401023387909, + "step": 1496 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2570.0, + "completions/mean_length": 920.7355346679688, + "completions/mean_terminated_length": 574.9146118164062, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 13.988921282798835, + "grad_norm": 0.15268854796886444, + "learning_rate": 1e-06, + "loss": -0.0306, + "num_tokens": 891428677.0, + "reward": 0.723214328289032, + "reward_std": 0.11873877793550491, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1497 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.08522727272727271, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3741.0, + "completions/mean_length": 864.4375, + "completions/mean_terminated_length": 563.3602294921875, + "completions/min_length": 171.0, + "completions/min_terminated_length": 171.0, + "epoch": 13.998250728862974, + "grad_norm": 0.1742684543132782, + "learning_rate": 1e-06, + "loss": -0.0314, + "num_tokens": 892004860.0, + "reward": 0.7410714626312256, + "reward_std": 0.11655956506729126, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1498 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3092.0, + "completions/mean_length": 878.5324096679688, + "completions/mean_terminated_length": 605.8656616210938, + "completions/min_length": 81.0, + "completions/min_terminated_length": 81.0, + "epoch": 14.00932944606414, + "grad_norm": 0.15107493102550507, + "learning_rate": 1e-06, + "loss": -0.0209, + "num_tokens": 892593793.0, + "reward": 0.707589328289032, + "reward_std": 0.12028081715106964, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512402057647705, + "step": 1499 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3161.0, + "completions/mean_length": 983.1942138671875, + "completions/mean_terminated_length": 614.0099487304688, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 14.018658892128279, + "grad_norm": 0.16565755009651184, + "learning_rate": 1e-06, + "loss": -0.015, + "num_tokens": 893175295.0, + "reward": 0.7042410969734192, + "reward_std": 0.12253600358963013, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1500 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4047.0, + "completions/mean_length": 936.474365234375, + "completions/mean_terminated_length": 622.4601440429688, + "completions/min_length": 112.0, + "completions/min_terminated_length": 112.0, + "epoch": 14.02798833819242, + "grad_norm": 0.1617438644170761, + "learning_rate": 1e-06, + "loss": -0.0738, + "num_tokens": 893776296.0, + "reward": 0.6696428656578064, + "reward_std": 0.1515413373708725, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1501 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3832.0, + "completions/mean_length": 1010.55029296875, + "completions/mean_terminated_length": 670.2713623046875, + "completions/min_length": 156.0, + "completions/min_terminated_length": 156.0, + "epoch": 14.03731778425656, + "grad_norm": 0.17312908172607422, + "learning_rate": 1e-06, + "loss": -0.0557, + "num_tokens": 894408517.0, + "reward": 0.6540178656578064, + "reward_std": 0.16104431450366974, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1502 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2873.0, + "completions/mean_length": 878.97998046875, + "completions/mean_terminated_length": 537.4197998046875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.0466472303207, + "grad_norm": 0.1437922865152359, + "learning_rate": 1e-06, + "loss": -0.057, + "num_tokens": 894931171.0, + "reward": 0.746651828289032, + "reward_std": 0.11632992327213287, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1503 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2687.0, + "completions/mean_length": 1066.296875, + "completions/mean_terminated_length": 651.058349609375, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 14.055976676384839, + "grad_norm": 0.17059540748596191, + "learning_rate": 1e-06, + "loss": -0.0974, + "num_tokens": 895538741.0, + "reward": 0.6819196939468384, + "reward_std": 0.15361109375953674, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1504 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3915.0, + "completions/mean_length": 985.9498291015625, + "completions/mean_terminated_length": 621.43017578125, + "completions/min_length": 123.0, + "completions/min_terminated_length": 123.0, + "epoch": 14.06530612244898, + "grad_norm": 0.1539929062128067, + "learning_rate": 1e-06, + "loss": -0.0441, + "num_tokens": 896125040.0, + "reward": 0.7031250596046448, + "reward_std": 0.10660012811422348, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1505 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3431.0, + "completions/mean_length": 919.38623046875, + "completions/mean_terminated_length": 637.6209106445312, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.07463556851312, + "grad_norm": 0.16501371562480927, + "learning_rate": 1e-06, + "loss": -0.0455, + "num_tokens": 896741450.0, + "reward": 0.6953125596046448, + "reward_std": 0.1525944322347641, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1506 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3075.0, + "completions/mean_length": 852.9520263671875, + "completions/mean_terminated_length": 603.48681640625, + "completions/min_length": 131.0, + "completions/min_terminated_length": 131.0, + "epoch": 14.08396501457726, + "grad_norm": 0.1404428780078888, + "learning_rate": 1e-06, + "loss": -0.0568, + "num_tokens": 897344671.0, + "reward": 0.723214328289032, + "reward_std": 0.1261366903781891, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.4476594924926758, + "step": 1507 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0825892857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4040.0, + "completions/mean_length": 927.0826416015625, + "completions/mean_terminated_length": 641.8029174804688, + "completions/min_length": 169.0, + "completions/min_terminated_length": 169.0, + "epoch": 14.093294460641399, + "grad_norm": 0.13648349046707153, + "learning_rate": 1e-06, + "loss": -0.0588, + "num_tokens": 897957937.0, + "reward": 0.7332589626312256, + "reward_std": 0.12343642115592957, + "rewards/verify_math_reward/mean": 0.7332589030265808, + "rewards/verify_math_reward/std": 0.4425028860569, + "step": 1508 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3263.0, + "completions/mean_length": 1075.1976318359375, + "completions/mean_terminated_length": 652.4389038085938, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 14.102623906705539, + "grad_norm": 0.16810107231140137, + "learning_rate": 1e-06, + "loss": -0.0344, + "num_tokens": 898569162.0, + "reward": 0.6439732313156128, + "reward_std": 0.14027173817157745, + "rewards/verify_math_reward/mean": 0.6439732313156128, + "rewards/verify_math_reward/std": 0.47909072041511536, + "step": 1509 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4009.0, + "completions/mean_length": 993.7589721679688, + "completions/mean_terminated_length": 677.0479736328125, + "completions/min_length": 153.0, + "completions/min_terminated_length": 153.0, + "epoch": 14.11195335276968, + "grad_norm": 0.14482687413692474, + "learning_rate": 1e-06, + "loss": -0.0618, + "num_tokens": 899211450.0, + "reward": 0.7087053656578064, + "reward_std": 0.15293073654174805, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1510 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0803571428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3888.0, + "completions/mean_length": 888.90185546875, + "completions/mean_terminated_length": 608.669921875, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.12128279883382, + "grad_norm": 0.16684040427207947, + "learning_rate": 1e-06, + "loss": -0.0371, + "num_tokens": 899795746.0, + "reward": 0.746651828289032, + "reward_std": 0.12707847356796265, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1511 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0870535714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3289.0, + "completions/mean_length": 931.7098388671875, + "completions/mean_terminated_length": 629.98046875, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.130612244897959, + "grad_norm": 0.13450437784194946, + "learning_rate": 1e-06, + "loss": -0.0591, + "num_tokens": 900399478.0, + "reward": 0.723214328289032, + "reward_std": 0.12654967606067657, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1512 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1071428571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 1068.579345703125, + "completions/mean_terminated_length": 705.2887573242188, + "completions/min_length": 182.0, + "completions/min_terminated_length": 182.0, + "epoch": 14.139941690962099, + "grad_norm": 0.13896602392196655, + "learning_rate": 1e-06, + "loss": -0.0353, + "num_tokens": 901058645.0, + "reward": 0.6551339626312256, + "reward_std": 0.1284339427947998, + "rewards/verify_math_reward/mean": 0.6551339030265808, + "rewards/verify_math_reward/std": 0.4755900502204895, + "step": 1513 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3493.0, + "completions/mean_length": 863.1908569335938, + "completions/mean_terminated_length": 584.9733276367188, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 14.14927113702624, + "grad_norm": 0.16177017986774445, + "learning_rate": 1e-06, + "loss": -0.0907, + "num_tokens": 901626840.0, + "reward": 0.7243303656578064, + "reward_std": 0.15352804958820343, + "rewards/verify_math_reward/mean": 0.7243303656578064, + "rewards/verify_math_reward/std": 0.4471006691455841, + "step": 1514 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3423.0, + "completions/mean_length": 878.9230346679688, + "completions/mean_terminated_length": 606.2893676757812, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 14.15860058309038, + "grad_norm": 0.1621823012828827, + "learning_rate": 1e-06, + "loss": -0.0375, + "num_tokens": 902217371.0, + "reward": 0.6774553656578064, + "reward_std": 0.1337614506483078, + "rewards/verify_math_reward/mean": 0.6774553656578064, + "rewards/verify_math_reward/std": 0.4677111804485321, + "step": 1515 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1116071428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4061.0, + "completions/mean_length": 1045.094970703125, + "completions/mean_terminated_length": 661.8153076171875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.167930029154519, + "grad_norm": 0.26154232025146484, + "learning_rate": 1e-06, + "loss": -0.0571, + "num_tokens": 902834552.0, + "reward": 0.6662946939468384, + "reward_std": 0.15315966308116913, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179922461509705, + "step": 1516 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3901.0, + "completions/mean_length": 1004.8203735351562, + "completions/mean_terminated_length": 659.6513671875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 14.177259475218658, + "grad_norm": 0.1650465428829193, + "learning_rate": 1e-06, + "loss": -0.0519, + "num_tokens": 903457495.0, + "reward": 0.6540178656578064, + "reward_std": 0.1546206921339035, + "rewards/verify_math_reward/mean": 0.6540178656578064, + "rewards/verify_math_reward/std": 0.4759531021118164, + "step": 1517 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3266.0, + "completions/mean_length": 880.9766235351562, + "completions/mean_terminated_length": 604.2896728515625, + "completions/min_length": 140.0, + "completions/min_terminated_length": 140.0, + "epoch": 14.186588921282798, + "grad_norm": 0.15514257550239563, + "learning_rate": 1e-06, + "loss": -0.0276, + "num_tokens": 904063298.0, + "reward": 0.6662946939468384, + "reward_std": 0.13203756511211395, + "rewards/verify_math_reward/mean": 0.6662946343421936, + "rewards/verify_math_reward/std": 0.47179925441741943, + "step": 1518 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0658482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3987.0, + "completions/mean_length": 799.2232666015625, + "completions/mean_terminated_length": 566.8339233398438, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 14.19591836734694, + "grad_norm": 0.16251403093338013, + "learning_rate": 1e-06, + "loss": -0.0629, + "num_tokens": 904625034.0, + "reward": 0.746651828289032, + "reward_std": 0.14597928524017334, + "rewards/verify_math_reward/mean": 0.7466517686843872, + "rewards/verify_math_reward/std": 0.435171514749527, + "step": 1519 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3805.0, + "completions/mean_length": 896.6585083007812, + "completions/mean_terminated_length": 582.99755859375, + "completions/min_length": 121.0, + "completions/min_terminated_length": 121.0, + "epoch": 14.205247813411079, + "grad_norm": 0.14653554558753967, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 905190192.0, + "reward": 0.723214328289032, + "reward_std": 0.11809796839952469, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1520 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2729.0, + "completions/mean_length": 1002.7813110351562, + "completions/mean_terminated_length": 618.5545654296875, + "completions/min_length": 154.0, + "completions/min_terminated_length": 154.0, + "epoch": 14.214577259475218, + "grad_norm": 0.16323019564151764, + "learning_rate": 1e-06, + "loss": -0.057, + "num_tokens": 905766588.0, + "reward": 0.6595982313156128, + "reward_std": 0.1577032506465912, + "rewards/verify_math_reward/mean": 0.6595982313156128, + "rewards/verify_math_reward/std": 0.4741089344024658, + "step": 1521 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0736607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3548.0, + "completions/mean_length": 884.9308471679688, + "completions/mean_terminated_length": 629.5927734375, + "completions/min_length": 133.0, + "completions/min_terminated_length": 133.0, + "epoch": 14.223906705539358, + "grad_norm": 0.15244054794311523, + "learning_rate": 1e-06, + "loss": -0.0223, + "num_tokens": 906370886.0, + "reward": 0.6819196939468384, + "reward_std": 0.1209939569234848, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1522 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2828.0, + "completions/mean_length": 989.4475708007812, + "completions/mean_terminated_length": 541.1200561523438, + "completions/min_length": 137.0, + "completions/min_terminated_length": 137.0, + "epoch": 14.2332361516035, + "grad_norm": 0.16388195753097534, + "learning_rate": 1e-06, + "loss": -0.0729, + "num_tokens": 906885775.0, + "reward": 0.6964285969734192, + "reward_std": 0.13455379009246826, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600565731525421, + "step": 1523 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0814732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3409.0, + "completions/mean_length": 863.3772583007812, + "completions/mean_terminated_length": 576.6439819335938, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 14.242565597667639, + "grad_norm": 0.13973724842071533, + "learning_rate": 1e-06, + "loss": -0.027, + "num_tokens": 907435497.0, + "reward": 0.7689732313156128, + "reward_std": 0.11324945837259293, + "rewards/verify_math_reward/mean": 0.7689732313156128, + "rewards/verify_math_reward/std": 0.42172491550445557, + "step": 1524 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3467.0, + "completions/mean_length": 990.982177734375, + "completions/mean_terminated_length": 583.2525024414062, + "completions/min_length": 161.0, + "completions/min_terminated_length": 161.0, + "epoch": 14.251895043731778, + "grad_norm": 0.14989370107650757, + "learning_rate": 1e-06, + "loss": -0.0464, + "num_tokens": 907985913.0, + "reward": 0.6964285969734192, + "reward_std": 0.10228294879198074, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1525 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2809.0, + "completions/mean_length": 862.4297485351562, + "completions/mean_terminated_length": 588.3983154296875, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 14.261224489795918, + "grad_norm": 0.1450241059064865, + "learning_rate": 1e-06, + "loss": -0.0803, + "num_tokens": 908551898.0, + "reward": 0.7633928656578064, + "reward_std": 0.11276272684335709, + "rewards/verify_math_reward/mean": 0.7633928656578064, + "rewards/verify_math_reward/std": 0.42523646354675293, + "step": 1526 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2447.0, + "completions/mean_length": 856.466552734375, + "completions/mean_terminated_length": 590.4178466796875, + "completions/min_length": 166.0, + "completions/min_terminated_length": 166.0, + "epoch": 14.270553935860057, + "grad_norm": 0.14993762969970703, + "learning_rate": 1e-06, + "loss": -0.0485, + "num_tokens": 909126420.0, + "reward": 0.6796875596046448, + "reward_std": 0.1602187603712082, + "rewards/verify_math_reward/mean": 0.6796875, + "rewards/verify_math_reward/std": 0.4668572247028351, + "step": 1527 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3649.0, + "completions/mean_length": 899.7355346679688, + "completions/mean_terminated_length": 551.62744140625, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 14.279883381924199, + "grad_norm": 0.1586182564496994, + "learning_rate": 1e-06, + "loss": -0.0578, + "num_tokens": 909649199.0, + "reward": 0.707589328289032, + "reward_std": 0.11205439269542694, + "rewards/verify_math_reward/mean": 0.7075892686843872, + "rewards/verify_math_reward/std": 0.45512405037879944, + "step": 1528 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0982142857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3979.0, + "completions/mean_length": 956.6395263671875, + "completions/mean_terminated_length": 614.7289428710938, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 14.289212827988338, + "grad_norm": 0.15270592272281647, + "learning_rate": 1e-06, + "loss": -0.0362, + "num_tokens": 910229156.0, + "reward": 0.6808035969734192, + "reward_std": 0.12779638171195984, + "rewards/verify_math_reward/mean": 0.6808035969734192, + "rewards/verify_math_reward/std": 0.4664256274700165, + "step": 1529 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2980.0, + "completions/mean_length": 948.02685546875, + "completions/mean_terminated_length": 635.1607055664062, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 14.298542274052478, + "grad_norm": 0.16676433384418488, + "learning_rate": 1e-06, + "loss": -0.0444, + "num_tokens": 910839804.0, + "reward": 0.6785714626312256, + "reward_std": 0.16675932705402374, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1530 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0926339285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3263.0, + "completions/mean_length": 941.036865234375, + "completions/mean_terminated_length": 618.9434204101562, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.307871720116617, + "grad_norm": 0.18637950718402863, + "learning_rate": 1e-06, + "loss": -0.0321, + "num_tokens": 911438693.0, + "reward": 0.6819196939468384, + "reward_std": 0.14958783984184265, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599096059799194, + "step": 1531 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3303.0, + "completions/mean_length": 890.6629638671875, + "completions/mean_terminated_length": 576.4142456054688, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.317201166180759, + "grad_norm": 0.15555548667907715, + "learning_rate": 1e-06, + "loss": -0.0404, + "num_tokens": 912000047.0, + "reward": 0.7120535969734192, + "reward_std": 0.13380561769008636, + "rewards/verify_math_reward/mean": 0.7120535969734192, + "rewards/verify_math_reward/std": 0.4530589282512665, + "step": 1532 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1138392857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3966.0, + "completions/mean_length": 1022.4922485351562, + "completions/mean_terminated_length": 627.65869140625, + "completions/min_length": 113.0, + "completions/min_terminated_length": 113.0, + "epoch": 14.326530612244898, + "grad_norm": 0.16348303854465485, + "learning_rate": 1e-06, + "loss": -0.0674, + "num_tokens": 912589064.0, + "reward": 0.6696428656578064, + "reward_std": 0.13688749074935913, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1533 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0636160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3849.0, + "completions/mean_length": 843.2980346679688, + "completions/mean_terminated_length": 622.3158569335938, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 14.335860058309038, + "grad_norm": 0.14868338406085968, + "learning_rate": 1e-06, + "loss": -0.0204, + "num_tokens": 913194251.0, + "reward": 0.7633928656578064, + "reward_std": 0.10002705454826355, + "rewards/verify_math_reward/mean": 0.7633928656578064, + "rewards/verify_math_reward/std": 0.42523646354675293, + "step": 1534 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2953.0, + "completions/mean_length": 910.29248046875, + "completions/mean_terminated_length": 576.4019775390625, + "completions/min_length": 107.0, + "completions/min_terminated_length": 107.0, + "epoch": 14.345189504373177, + "grad_norm": 0.15315856039524078, + "learning_rate": 1e-06, + "loss": -0.0337, + "num_tokens": 913753553.0, + "reward": 0.684151828289032, + "reward_std": 0.11832619458436966, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1535 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2811.0, + "completions/mean_length": 1024.7723388671875, + "completions/mean_terminated_length": 586.0255126953125, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 14.354518950437317, + "grad_norm": 0.1642405390739441, + "learning_rate": 1e-06, + "loss": -0.0456, + "num_tokens": 914298325.0, + "reward": 0.6707589626312256, + "reward_std": 0.13921935856342316, + "rewards/verify_math_reward/mean": 0.6707589030265808, + "rewards/verify_math_reward/std": 0.4702001214027405, + "step": 1536 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4002.0, + "completions/mean_length": 880.12841796875, + "completions/mean_terminated_length": 641.0587768554688, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 14.363848396501458, + "grad_norm": 0.15526075661182404, + "learning_rate": 1e-06, + "loss": -0.029, + "num_tokens": 914918104.0, + "reward": 0.7220982313156128, + "reward_std": 0.13665924966335297, + "rewards/verify_math_reward/mean": 0.7220982313156128, + "rewards/verify_math_reward/std": 0.44821491837501526, + "step": 1537 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3880.0, + "completions/mean_length": 991.5402221679688, + "completions/mean_terminated_length": 666.1652221679688, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.373177842565598, + "grad_norm": 0.1703079640865326, + "learning_rate": 1e-06, + "loss": -0.0365, + "num_tokens": 915540820.0, + "reward": 0.6618303656578064, + "reward_std": 0.15094473958015442, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1538 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3570.0, + "completions/mean_length": 920.513427734375, + "completions/mean_terminated_length": 583.3629760742188, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 14.382507288629737, + "grad_norm": 0.15420016646385193, + "learning_rate": 1e-06, + "loss": -0.057, + "num_tokens": 916098376.0, + "reward": 0.7578125596046448, + "reward_std": 0.11532101780176163, + "rewards/verify_math_reward/mean": 0.7578125, + "rewards/verify_math_reward/std": 0.428646445274353, + "step": 1539 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3660.0, + "completions/mean_length": 1028.390625, + "completions/mean_terminated_length": 647.3450317382812, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 14.391836734693877, + "grad_norm": 0.12327955663204193, + "learning_rate": 1e-06, + "loss": -0.0565, + "num_tokens": 916708006.0, + "reward": 0.668526828289032, + "reward_std": 0.10453673452138901, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1540 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1026785714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3658.0, + "completions/mean_length": 986.4006958007812, + "completions/mean_terminated_length": 630.5758666992188, + "completions/min_length": 109.0, + "completions/min_terminated_length": 109.0, + "epoch": 14.401166180758018, + "grad_norm": 0.16874539852142334, + "learning_rate": 1e-06, + "loss": -0.0405, + "num_tokens": 917303501.0, + "reward": 0.684151828289032, + "reward_std": 0.13482661545276642, + "rewards/verify_math_reward/mean": 0.6841517686843872, + "rewards/verify_math_reward/std": 0.4651124179363251, + "step": 1541 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3522.0, + "completions/mean_length": 828.1027221679688, + "completions/mean_terminated_length": 576.7260131835938, + "completions/min_length": 90.0, + "completions/min_terminated_length": 90.0, + "epoch": 14.410495626822158, + "grad_norm": 0.16123774647712708, + "learning_rate": 1e-06, + "loss": -0.0537, + "num_tokens": 917868241.0, + "reward": 0.731026828289032, + "reward_std": 0.14913278818130493, + "rewards/verify_math_reward/mean": 0.7310267686843872, + "rewards/verify_math_reward/std": 0.44367367029190063, + "step": 1542 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3425.0, + "completions/mean_length": 880.1719360351562, + "completions/mean_terminated_length": 603.415771484375, + "completions/min_length": 178.0, + "completions/min_terminated_length": 178.0, + "epoch": 14.419825072886297, + "grad_norm": 0.13473300635814667, + "learning_rate": 1e-06, + "loss": -0.0532, + "num_tokens": 918457595.0, + "reward": 0.723214328289032, + "reward_std": 0.10359494388103485, + "rewards/verify_math_reward/mean": 0.7232142686843872, + "rewards/verify_math_reward/std": 0.44765952229499817, + "step": 1543 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3970.0, + "completions/mean_length": 1020.9855346679688, + "completions/mean_terminated_length": 621.5825805664062, + "completions/min_length": 148.0, + "completions/min_terminated_length": 148.0, + "epoch": 14.429154518950437, + "grad_norm": 0.1620435267686844, + "learning_rate": 1e-06, + "loss": -0.057, + "num_tokens": 919042142.0, + "reward": 0.676339328289032, + "reward_std": 0.1463569700717926, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335687637329, + "step": 1544 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2642.0, + "completions/mean_length": 949.4464721679688, + "completions/mean_terminated_length": 576.2596435546875, + "completions/min_length": 91.0, + "completions/min_terminated_length": 91.0, + "epoch": 14.438483965014576, + "grad_norm": 0.16333381831645966, + "learning_rate": 1e-06, + "loss": -0.0658, + "num_tokens": 919599006.0, + "reward": 0.6886160969734192, + "reward_std": 0.12369714677333832, + "rewards/verify_math_reward/mean": 0.6886160969734192, + "rewards/verify_math_reward/std": 0.46331799030303955, + "step": 1545 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2451.0, + "completions/mean_length": 957.68310546875, + "completions/mean_terminated_length": 558.9786376953125, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 14.447813411078718, + "grad_norm": 0.17028574645519257, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 920132234.0, + "reward": 0.6830357313156128, + "reward_std": 0.13290520012378693, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1546 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3790.0, + "completions/mean_length": 923.9475708007812, + "completions/mean_terminated_length": 582.824462890625, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 14.457142857142857, + "grad_norm": 0.1495293527841568, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 920688547.0, + "reward": 0.6819196939468384, + "reward_std": 0.13336236774921417, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1547 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3666.0, + "completions/mean_length": 1012.6864013671875, + "completions/mean_terminated_length": 563.2007446289062, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 14.466472303206997, + "grad_norm": 0.18091271817684174, + "learning_rate": 1e-06, + "loss": -0.0732, + "num_tokens": 921223202.0, + "reward": 0.6819196939468384, + "reward_std": 0.14774522185325623, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1548 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3586.0, + "completions/mean_length": 916.755615234375, + "completions/mean_terminated_length": 605.0650024414062, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.475801749271136, + "grad_norm": 0.15860570967197418, + "learning_rate": 1e-06, + "loss": -0.0363, + "num_tokens": 921805679.0, + "reward": 0.6819196939468384, + "reward_std": 0.11952830851078033, + "rewards/verify_math_reward/mean": 0.6819196343421936, + "rewards/verify_math_reward/std": 0.46599099040031433, + "step": 1549 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0859375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3645.0, + "completions/mean_length": 868.1361694335938, + "completions/mean_terminated_length": 564.6618041992188, + "completions/min_length": 152.0, + "completions/min_terminated_length": 152.0, + "epoch": 14.485131195335278, + "grad_norm": 0.17823606729507446, + "learning_rate": 1e-06, + "loss": -0.0505, + "num_tokens": 922351481.0, + "reward": 0.7500000596046448, + "reward_std": 0.14612692594528198, + "rewards/verify_math_reward/mean": 0.75, + "rewards/verify_math_reward/std": 0.43325456976890564, + "step": 1550 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0881696428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3290.0, + "completions/mean_length": 911.8203735351562, + "completions/mean_terminated_length": 603.9253540039062, + "completions/min_length": 142.0, + "completions/min_terminated_length": 142.0, + "epoch": 14.494460641399417, + "grad_norm": 0.14829711616039276, + "learning_rate": 1e-06, + "loss": -0.0449, + "num_tokens": 922931096.0, + "reward": 0.7064732313156128, + "reward_std": 0.10893340408802032, + "rewards/verify_math_reward/mean": 0.7064732313156128, + "rewards/verify_math_reward/std": 0.4556320011615753, + "step": 1551 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1339285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 1130.977783203125, + "completions/mean_terminated_length": 672.4690551757812, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 14.503790087463557, + "grad_norm": 0.11391159892082214, + "learning_rate": 1e-06, + "loss": -0.0771, + "num_tokens": 923551908.0, + "reward": 0.676339328289032, + "reward_std": 0.12546519935131073, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1552 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3899.0, + "completions/mean_length": 943.6652221679688, + "completions/mean_terminated_length": 626.1080932617188, + "completions/min_length": 100.0, + "completions/min_terminated_length": 100.0, + "epoch": 14.513119533527696, + "grad_norm": 0.13433076441287994, + "learning_rate": 1e-06, + "loss": -0.0649, + "num_tokens": 924146400.0, + "reward": 0.715401828289032, + "reward_std": 0.1283177137374878, + "rewards/verify_math_reward/mean": 0.7154017686843872, + "rewards/verify_math_reward/std": 0.4514748752117157, + "step": 1553 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3327.0, + "completions/mean_length": 995.4453735351562, + "completions/mean_terminated_length": 614.6754150390625, + "completions/min_length": 118.0, + "completions/min_terminated_length": 118.0, + "epoch": 14.522448979591836, + "grad_norm": 0.159366175532341, + "learning_rate": 1e-06, + "loss": -0.0679, + "num_tokens": 924730687.0, + "reward": 0.7020089626312256, + "reward_std": 0.13144053518772125, + "rewards/verify_math_reward/mean": 0.7020089030265808, + "rewards/verify_math_reward/std": 0.45763099193573, + "step": 1554 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2786.0, + "completions/mean_length": 1072.859375, + "completions/mean_terminated_length": 658.520263671875, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 14.531778425655977, + "grad_norm": 0.14054180681705475, + "learning_rate": 1e-06, + "loss": -0.0602, + "num_tokens": 925360601.0, + "reward": 0.6618303656578064, + "reward_std": 0.11681987345218658, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1555 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1015625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3188.0, + "completions/mean_length": 932.3248291015625, + "completions/mean_terminated_length": 574.69189453125, + "completions/min_length": 138.0, + "completions/min_terminated_length": 138.0, + "epoch": 14.541107871720117, + "grad_norm": 0.15395845472812653, + "learning_rate": 1e-06, + "loss": -0.0539, + "num_tokens": 925918316.0, + "reward": 0.6986607313156128, + "reward_std": 0.1454133540391922, + "rewards/verify_math_reward/mean": 0.6986607313156128, + "rewards/verify_math_reward/std": 0.4590960443019867, + "step": 1556 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4034.0, + "completions/mean_length": 963.0301513671875, + "completions/mean_terminated_length": 613.1947631835938, + "completions/min_length": 165.0, + "completions/min_terminated_length": 165.0, + "epoch": 14.550437317784256, + "grad_norm": 0.16525380313396454, + "learning_rate": 1e-06, + "loss": -0.0838, + "num_tokens": 926494687.0, + "reward": 0.7042410969734192, + "reward_std": 0.17115388810634613, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663803815841675, + "step": 1557 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1272321428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3869.0, + "completions/mean_length": 1136.3426513671875, + "completions/mean_terminated_length": 704.8836059570312, + "completions/min_length": 120.0, + "completions/min_terminated_length": 120.0, + "epoch": 14.559766763848396, + "grad_norm": 0.15317338705062866, + "learning_rate": 1e-06, + "loss": -0.0702, + "num_tokens": 927130714.0, + "reward": 0.625, + "reward_std": 0.14391589164733887, + "rewards/verify_math_reward/mean": 0.625, + "rewards/verify_math_reward/std": 0.48439329862594604, + "step": 1558 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1261160714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2434.0, + "completions/mean_length": 1044.2410888671875, + "completions/mean_terminated_length": 603.8211669921875, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.569096209912537, + "grad_norm": 0.1727406084537506, + "learning_rate": 1e-06, + "loss": -0.0414, + "num_tokens": 927690586.0, + "reward": 0.6618303656578064, + "reward_std": 0.1457182765007019, + "rewards/verify_math_reward/mean": 0.6618303656578064, + "rewards/verify_math_reward/std": 0.4733508229255676, + "step": 1559 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3752.0, + "completions/mean_length": 975.5145263671875, + "completions/mean_terminated_length": 627.0731811523438, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 14.578425655976677, + "grad_norm": 0.15712745487689972, + "learning_rate": 1e-06, + "loss": 0.0016, + "num_tokens": 928293279.0, + "reward": 0.7053571939468384, + "reward_std": 0.09153327345848083, + "rewards/verify_math_reward/mean": 0.7053571343421936, + "rewards/verify_math_reward/std": 0.45613664388656616, + "step": 1560 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4012.0, + "completions/mean_length": 983.9464721679688, + "completions/mean_terminated_length": 566.3797607421875, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.587755102040816, + "grad_norm": 0.15441763401031494, + "learning_rate": 1e-06, + "loss": -0.0477, + "num_tokens": 928827655.0, + "reward": 0.7265625596046448, + "reward_std": 0.11960500478744507, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1561 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0904017857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3651.0, + "completions/mean_length": 964.6183471679688, + "completions/mean_terminated_length": 653.4012451171875, + "completions/min_length": 170.0, + "completions/min_terminated_length": 170.0, + "epoch": 14.597084548104956, + "grad_norm": 0.14298835396766663, + "learning_rate": 1e-06, + "loss": -0.0721, + "num_tokens": 929452769.0, + "reward": 0.6785714626312256, + "reward_std": 0.12839369475841522, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1562 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1104910714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3297.0, + "completions/mean_length": 958.4922485351562, + "completions/mean_terminated_length": 568.7640991210938, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 14.606413994169095, + "grad_norm": 0.14158286154270172, + "learning_rate": 1e-06, + "loss": -0.0688, + "num_tokens": 929993098.0, + "reward": 0.7321428656578064, + "reward_std": 0.12084353715181351, + "rewards/verify_math_reward/mean": 0.7321428656578064, + "rewards/verify_math_reward/std": 0.4430900514125824, + "step": 1563 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1049107142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3360.0, + "completions/mean_length": 967.6797485351562, + "completions/mean_terminated_length": 601.0187377929688, + "completions/min_length": 122.0, + "completions/min_terminated_length": 122.0, + "epoch": 14.615743440233237, + "grad_norm": 0.15492652356624603, + "learning_rate": 1e-06, + "loss": -0.0396, + "num_tokens": 930559499.0, + "reward": 0.676339328289032, + "reward_std": 0.12636421620845795, + "rewards/verify_math_reward/mean": 0.6763392686843872, + "rewards/verify_math_reward/std": 0.4681335985660553, + "step": 1564 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0714285714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3361.0, + "completions/mean_length": 812.2678833007812, + "completions/mean_terminated_length": 559.673095703125, + "completions/min_length": 135.0, + "completions/min_terminated_length": 135.0, + "epoch": 14.625072886297376, + "grad_norm": 0.16751259565353394, + "learning_rate": 1e-06, + "loss": -0.0254, + "num_tokens": 931106067.0, + "reward": 0.7555803656578064, + "reward_std": 0.12174436450004578, + "rewards/verify_math_reward/mean": 0.7555803656578064, + "rewards/verify_math_reward/std": 0.42998260259628296, + "step": 1565 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0792410714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4085.0, + "completions/mean_length": 891.0803833007812, + "completions/mean_terminated_length": 615.2630615234375, + "completions/min_length": 116.0, + "completions/min_terminated_length": 116.0, + "epoch": 14.634402332361516, + "grad_norm": 0.1823289841413498, + "learning_rate": 1e-06, + "loss": -0.0411, + "num_tokens": 931698323.0, + "reward": 0.7611607313156128, + "reward_std": 0.14661797881126404, + "rewards/verify_math_reward/mean": 0.7611607313156128, + "rewards/verify_math_reward/std": 0.4266124963760376, + "step": 1566 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0770089285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 844.0636596679688, + "completions/mean_terminated_length": 572.7412719726562, + "completions/min_length": 124.0, + "completions/min_terminated_length": 124.0, + "epoch": 14.643731778425655, + "grad_norm": 0.15263529121875763, + "learning_rate": 1e-06, + "loss": -0.0527, + "num_tokens": 932252492.0, + "reward": 0.7477678656578064, + "reward_std": 0.12941035628318787, + "rewards/verify_math_reward/mean": 0.7477678656578064, + "rewards/verify_math_reward/std": 0.4345363676548004, + "step": 1567 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0691964285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3480.0, + "completions/mean_length": 807.8147583007812, + "completions/mean_terminated_length": 563.3693237304688, + "completions/min_length": 139.0, + "completions/min_terminated_length": 139.0, + "epoch": 14.653061224489797, + "grad_norm": 0.15244770050048828, + "learning_rate": 1e-06, + "loss": -0.0623, + "num_tokens": 932817598.0, + "reward": 0.7611607313156128, + "reward_std": 0.1267005205154419, + "rewards/verify_math_reward/mean": 0.7611607313156128, + "rewards/verify_math_reward/std": 0.4266124963760376, + "step": 1568 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3796.0, + "completions/mean_length": 1008.8192138671875, + "completions/mean_terminated_length": 616.611328125, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 14.662390670553936, + "grad_norm": 0.18300633132457733, + "learning_rate": 1e-06, + "loss": -0.0805, + "num_tokens": 933397372.0, + "reward": 0.6953125596046448, + "reward_std": 0.14027062058448792, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1569 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1361607142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3764.0, + "completions/mean_length": 1153.9732666015625, + "completions/mean_terminated_length": 690.242919921875, + "completions/min_length": 111.0, + "completions/min_terminated_length": 111.0, + "epoch": 14.671720116618076, + "grad_norm": 0.2032817304134369, + "learning_rate": 1e-06, + "loss": -0.0661, + "num_tokens": 934020580.0, + "reward": 0.606026828289032, + "reward_std": 0.15718290209770203, + "rewards/verify_math_reward/mean": 0.6060267686843872, + "rewards/verify_math_reward/std": 0.48890194296836853, + "step": 1570 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1060267857142857, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3834.0, + "completions/mean_length": 1006.7623291015625, + "completions/mean_terminated_length": 640.373291015625, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 14.681049562682215, + "grad_norm": 0.14900560677051544, + "learning_rate": 1e-06, + "loss": -0.0604, + "num_tokens": 934628135.0, + "reward": 0.6718750596046448, + "reward_std": 0.14548751711845398, + "rewards/verify_math_reward/mean": 0.671875, + "rewards/verify_math_reward/std": 0.46979284286499023, + "step": 1571 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3981.0, + "completions/mean_length": 964.37060546875, + "completions/mean_terminated_length": 640.4088745117188, + "completions/min_length": 162.0, + "completions/min_terminated_length": 162.0, + "epoch": 14.690379008746355, + "grad_norm": 0.17270837724208832, + "learning_rate": 1e-06, + "loss": -0.0658, + "num_tokens": 935232995.0, + "reward": 0.6830357313156128, + "reward_std": 0.13654978573322296, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1572 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1439732142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2589.0, + "completions/mean_length": 1109.3148193359375, + "completions/mean_terminated_length": 606.9908447265625, + "completions/min_length": 136.0, + "completions/min_terminated_length": 136.0, + "epoch": 14.699708454810496, + "grad_norm": 0.16209962964057922, + "learning_rate": 1e-06, + "loss": -0.0831, + "num_tokens": 935782845.0, + "reward": 0.6495535969734192, + "reward_std": 0.1443595588207245, + "rewards/verify_math_reward/mean": 0.6495535969734192, + "rewards/verify_math_reward/std": 0.477376252412796, + "step": 1573 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0892857142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3669.0, + "completions/mean_length": 918.2701416015625, + "completions/mean_terminated_length": 606.7279663085938, + "completions/min_length": 108.0, + "completions/min_terminated_length": 108.0, + "epoch": 14.709037900874636, + "grad_norm": 0.17579883337020874, + "learning_rate": 1e-06, + "loss": -0.0662, + "num_tokens": 936365383.0, + "reward": 0.7433035969734192, + "reward_std": 0.15071649849414825, + "rewards/verify_math_reward/mean": 0.7433035969734192, + "rewards/verify_math_reward/std": 0.43705442547798157, + "step": 1574 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0703125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3769.0, + "completions/mean_length": 895.583740234375, + "completions/mean_terminated_length": 653.535400390625, + "completions/min_length": 143.0, + "completions/min_terminated_length": 143.0, + "epoch": 14.718367346938775, + "grad_norm": 0.16020067036151886, + "learning_rate": 1e-06, + "loss": -0.0233, + "num_tokens": 937000074.0, + "reward": 0.6953125596046448, + "reward_std": 0.12858732044696808, + "rewards/verify_math_reward/mean": 0.6953125, + "rewards/verify_math_reward/std": 0.4605320394039154, + "step": 1575 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1283482142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3978.0, + "completions/mean_length": 1088.4320068359375, + "completions/mean_terminated_length": 645.576171875, + "completions/min_length": 173.0, + "completions/min_terminated_length": 173.0, + "epoch": 14.727696793002915, + "grad_norm": 0.16138674318790436, + "learning_rate": 1e-06, + "loss": -0.0401, + "num_tokens": 937606637.0, + "reward": 0.6428571939468384, + "reward_std": 0.14556418359279633, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1576 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1227678571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3525.0, + "completions/mean_length": 1069.888427734375, + "completions/mean_terminated_length": 646.3867797851562, + "completions/min_length": 160.0, + "completions/min_terminated_length": 160.0, + "epoch": 14.737026239067056, + "grad_norm": 0.17441169917583466, + "learning_rate": 1e-06, + "loss": -0.0684, + "num_tokens": 938205657.0, + "reward": 0.6830357313156128, + "reward_std": 0.16288946568965912, + "rewards/verify_math_reward/mean": 0.6830357313156128, + "rewards/verify_math_reward/std": 0.46555325388908386, + "step": 1577 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1194196428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4054.0, + "completions/mean_length": 1019.6953735351562, + "completions/mean_terminated_length": 602.503173828125, + "completions/min_length": 132.0, + "completions/min_terminated_length": 132.0, + "epoch": 14.746355685131196, + "grad_norm": 0.18642739951610565, + "learning_rate": 1e-06, + "loss": -0.0742, + "num_tokens": 938781864.0, + "reward": 0.652901828289032, + "reward_std": 0.12459757179021835, + "rewards/verify_math_reward/mean": 0.6529017686843872, + "rewards/verify_math_reward/std": 0.47631317377090454, + "step": 1578 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0758928571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3948.0, + "completions/mean_length": 880.4319458007812, + "completions/mean_terminated_length": 616.3514404296875, + "completions/min_length": 146.0, + "completions/min_terminated_length": 146.0, + "epoch": 14.755685131195335, + "grad_norm": 0.17653703689575195, + "learning_rate": 1e-06, + "loss": -0.0528, + "num_tokens": 939378123.0, + "reward": 0.7265625596046448, + "reward_std": 0.13978207111358643, + "rewards/verify_math_reward/mean": 0.7265625, + "rewards/verify_math_reward/std": 0.4459724426269531, + "step": 1579 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0948660714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3172.0, + "completions/mean_length": 967.8281860351562, + "completions/mean_terminated_length": 639.9678955078125, + "completions/min_length": 155.0, + "completions/min_terminated_length": 155.0, + "epoch": 14.765014577259475, + "grad_norm": 0.17044095695018768, + "learning_rate": 1e-06, + "loss": -0.0295, + "num_tokens": 939985801.0, + "reward": 0.7031250596046448, + "reward_std": 0.15417702496051788, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1580 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 4022.0, + "completions/mean_length": 954.9453735351562, + "completions/mean_terminated_length": 630.0086059570312, + "completions/min_length": 114.0, + "completions/min_terminated_length": 114.0, + "epoch": 14.774344023323614, + "grad_norm": 0.15684381127357483, + "learning_rate": 1e-06, + "loss": -0.0391, + "num_tokens": 940585656.0, + "reward": 0.6674107313156128, + "reward_std": 0.13680967688560486, + "rewards/verify_math_reward/mean": 0.6674107313156128, + "rewards/verify_math_reward/std": 0.47140392661094666, + "step": 1581 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1171875, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3810.0, + "completions/mean_length": 1020.5156860351562, + "completions/mean_terminated_length": 612.2655029296875, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 14.783673469387756, + "grad_norm": 0.1657339632511139, + "learning_rate": 1e-06, + "loss": -0.0582, + "num_tokens": 941162230.0, + "reward": 0.6908482313156128, + "reward_std": 0.13447892665863037, + "rewards/verify_math_reward/mean": 0.6908482313156128, + "rewards/verify_math_reward/std": 0.46240198612213135, + "step": 1582 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0970982142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3643.0, + "completions/mean_length": 960.8504638671875, + "completions/mean_terminated_length": 623.6959228515625, + "completions/min_length": 119.0, + "completions/min_terminated_length": 119.0, + "epoch": 14.793002915451895, + "grad_norm": 0.17091085016727448, + "learning_rate": 1e-06, + "loss": -0.0677, + "num_tokens": 941754632.0, + "reward": 0.6875000596046448, + "reward_std": 0.15282154083251953, + "rewards/verify_math_reward/mean": 0.6875, + "rewards/verify_math_reward/std": 0.4637712836265564, + "step": 1583 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3244.0, + "completions/mean_length": 990.8359985351562, + "completions/mean_terminated_length": 648.3829345703125, + "completions/min_length": 149.0, + "completions/min_terminated_length": 149.0, + "epoch": 14.802332361516035, + "grad_norm": 0.1544399857521057, + "learning_rate": 1e-06, + "loss": -0.0359, + "num_tokens": 942361837.0, + "reward": 0.7209821939468384, + "reward_std": 0.1324501484632492, + "rewards/verify_math_reward/mean": 0.7209821343421936, + "rewards/verify_math_reward/std": 0.448766827583313, + "step": 1584 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0959821428571429, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3253.0, + "completions/mean_length": 931.27685546875, + "completions/mean_terminated_length": 595.2691650390625, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.811661807580174, + "grad_norm": 0.1635824590921402, + "learning_rate": 1e-06, + "loss": -0.0517, + "num_tokens": 942924653.0, + "reward": 0.7031250596046448, + "reward_std": 0.14218950271606445, + "rewards/verify_math_reward/mean": 0.703125, + "rewards/verify_math_reward/std": 0.4571361541748047, + "step": 1585 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1149553571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2439.0, + "completions/mean_length": 1017.0279541015625, + "completions/mean_terminated_length": 617.1109619140625, + "completions/min_length": 110.0, + "completions/min_terminated_length": 110.0, + "epoch": 14.820991253644316, + "grad_norm": 0.16883839666843414, + "learning_rate": 1e-06, + "loss": -0.0444, + "num_tokens": 943497710.0, + "reward": 0.6852678656578064, + "reward_std": 0.13782110810279846, + "rewards/verify_math_reward/mean": 0.6852678656578064, + "rewards/verify_math_reward/std": 0.46466848254203796, + "step": 1586 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1183035714285714, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3935.0, + "completions/mean_length": 1086.2801513671875, + "completions/mean_terminated_length": 682.4443359375, + "completions/min_length": 128.0, + "completions/min_terminated_length": 128.0, + "epoch": 14.830320699708455, + "grad_norm": 0.15634258091449738, + "learning_rate": 1e-06, + "loss": -0.0634, + "num_tokens": 944126705.0, + "reward": 0.65625, + "reward_std": 0.13962799310684204, + "rewards/verify_math_reward/mean": 0.65625, + "rewards/verify_math_reward/std": 0.4752241373062134, + "step": 1587 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3366.0, + "completions/mean_length": 926.1864013671875, + "completions/mean_terminated_length": 576.6034545898438, + "completions/min_length": 141.0, + "completions/min_terminated_length": 141.0, + "epoch": 14.839650145772595, + "grad_norm": 0.1745196133852005, + "learning_rate": 1e-06, + "loss": -0.0461, + "num_tokens": 944683048.0, + "reward": 0.7611607313156128, + "reward_std": 0.10893270373344421, + "rewards/verify_math_reward/mean": 0.7611607313156128, + "rewards/verify_math_reward/std": 0.4266124963760376, + "step": 1588 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1205357142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3000.0, + "completions/mean_length": 1021.1652221679688, + "completions/mean_terminated_length": 599.7410888671875, + "completions/min_length": 125.0, + "completions/min_terminated_length": 125.0, + "epoch": 14.848979591836734, + "grad_norm": 0.15581004321575165, + "learning_rate": 1e-06, + "loss": -0.0611, + "num_tokens": 945254908.0, + "reward": 0.7042410969734192, + "reward_std": 0.12215371429920197, + "rewards/verify_math_reward/mean": 0.7042410969734192, + "rewards/verify_math_reward/std": 0.45663806796073914, + "step": 1589 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1127232142857143, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3458.0, + "completions/mean_length": 985.8504638671875, + "completions/mean_terminated_length": 590.7245483398438, + "completions/min_length": 179.0, + "completions/min_terminated_length": 179.0, + "epoch": 14.858309037900874, + "grad_norm": 0.2129804491996765, + "learning_rate": 1e-06, + "loss": -0.0572, + "num_tokens": 945807670.0, + "reward": 0.6964285969734192, + "reward_std": 0.1686376929283142, + "rewards/verify_math_reward/mean": 0.6964285969734192, + "rewards/verify_math_reward/std": 0.4600566029548645, + "step": 1590 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2799.0, + "completions/mean_length": 1041.5279541015625, + "completions/mean_terminated_length": 605.1747436523438, + "completions/min_length": 127.0, + "completions/min_terminated_length": 127.0, + "epoch": 14.867638483965015, + "grad_norm": 0.1739165186882019, + "learning_rate": 1e-06, + "loss": -0.0424, + "num_tokens": 946365687.0, + "reward": 0.6785714626312256, + "reward_std": 0.13519318401813507, + "rewards/verify_math_reward/mean": 0.6785714030265808, + "rewards/verify_math_reward/std": 0.46728572249412537, + "step": 1591 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.09375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 2802.0, + "completions/mean_length": 926.7288208007812, + "completions/mean_terminated_length": 598.8731689453125, + "completions/min_length": 150.0, + "completions/min_terminated_length": 150.0, + "epoch": 14.876967930029155, + "grad_norm": 0.13142170011997223, + "learning_rate": 1e-06, + "loss": -0.0361, + "num_tokens": 946938780.0, + "reward": 0.7198660969734192, + "reward_std": 0.09953668713569641, + "rewards/verify_math_reward/mean": 0.7198660969734192, + "rewards/verify_math_reward/std": 0.44931527972221375, + "step": 1592 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0993303571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3870.0, + "completions/mean_length": 949.661865234375, + "completions/mean_terminated_length": 602.6679077148438, + "completions/min_length": 104.0, + "completions/min_terminated_length": 104.0, + "epoch": 14.886297376093294, + "grad_norm": 0.15690341591835022, + "learning_rate": 1e-06, + "loss": -0.0357, + "num_tokens": 947512269.0, + "reward": 0.6696428656578064, + "reward_std": 0.13583257794380188, + "rewards/verify_math_reward/mean": 0.6696428656578064, + "rewards/verify_math_reward/std": 0.47060438990592957, + "step": 1593 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1004464285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3780.0, + "completions/mean_length": 1029.04248046875, + "completions/mean_terminated_length": 686.5781860351562, + "completions/min_length": 159.0, + "completions/min_terminated_length": 159.0, + "epoch": 14.895626822157434, + "grad_norm": 0.15457525849342346, + "learning_rate": 1e-06, + "loss": -0.0555, + "num_tokens": 948150987.0, + "reward": 0.6506696939468384, + "reward_std": 0.16153216361999512, + "rewards/verify_math_reward/mean": 0.6506696343421936, + "rewards/verify_math_reward/std": 0.47702476382255554, + "step": 1594 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.0915178571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3388.0, + "completions/mean_length": 892.72998046875, + "completions/mean_terminated_length": 570.041748046875, + "completions/min_length": 158.0, + "completions/min_terminated_length": 158.0, + "epoch": 14.904956268221575, + "grad_norm": 0.14987978339195251, + "learning_rate": 1e-06, + "loss": -0.0538, + "num_tokens": 948714185.0, + "reward": 0.7410714626312256, + "reward_std": 0.11986782401800156, + "rewards/verify_math_reward/mean": 0.7410714030265808, + "rewards/verify_math_reward/std": 0.43829095363616943, + "step": 1595 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1160714285714286, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3577.0, + "completions/mean_length": 1003.3013916015625, + "completions/mean_terminated_length": 597.1893920898438, + "completions/min_length": 145.0, + "completions/min_terminated_length": 145.0, + "epoch": 14.914285714285715, + "grad_norm": 0.15120556950569153, + "learning_rate": 1e-06, + "loss": -0.0721, + "num_tokens": 949277503.0, + "reward": 0.7087053656578064, + "reward_std": 0.12658992409706116, + "rewards/verify_math_reward/mean": 0.7087053656578064, + "rewards/verify_math_reward/std": 0.45461276173591614, + "step": 1596 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.1462053571428571, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3591.0, + "completions/mean_length": 1145.8359375, + "completions/mean_terminated_length": 640.6444702148438, + "completions/min_length": 134.0, + "completions/min_terminated_length": 134.0, + "epoch": 14.923615160349854, + "grad_norm": 0.20244941115379333, + "learning_rate": 1e-06, + "loss": -0.1195, + "num_tokens": 949861252.0, + "reward": 0.6339285969734192, + "reward_std": 0.16281278431415558, + "rewards/verify_math_reward/mean": 0.6339285969734192, + "rewards/verify_math_reward/std": 0.48199835419654846, + "step": 1597 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.15625, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3281.0, + "completions/mean_length": 1179.7701416015625, + "completions/mean_terminated_length": 639.7274780273438, + "completions/min_length": 174.0, + "completions/min_terminated_length": 174.0, + "epoch": 14.932944606413994, + "grad_norm": 0.20390333235263824, + "learning_rate": 1e-06, + "loss": -0.1134, + "num_tokens": 950434358.0, + "reward": 0.6428571939468384, + "reward_std": 0.15980760753154755, + "rewards/verify_math_reward/mean": 0.6428571343421936, + "rewards/verify_math_reward/std": 0.4794250428676605, + "step": 1598 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.109375, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3866.0, + "completions/mean_length": 990.8192138671875, + "completions/mean_terminated_length": 609.481201171875, + "completions/min_length": 144.0, + "completions/min_terminated_length": 144.0, + "epoch": 14.942274052478133, + "grad_norm": 0.17664200067520142, + "learning_rate": 1e-06, + "loss": -0.0613, + "num_tokens": 951015436.0, + "reward": 0.668526828289032, + "reward_std": 0.14109547436237335, + "rewards/verify_math_reward/mean": 0.6685267686843872, + "rewards/verify_math_reward/std": 0.4710056483745575, + "step": 1599 + }, + { + "clip_ratio/high_max": 0.0, + "clip_ratio/high_mean": 0.0, + "clip_ratio/low_mean": 0.0, + "clip_ratio/low_min": 0.0, + "clip_ratio/region_mean": 0.0, + "completions/clipped_ratio": 0.078125, + "completions/max_length": 4096.0, + "completions/max_terminated_length": 3008.0, + "completions/mean_length": 864.458740234375, + "completions/mean_terminated_length": 590.5993041992188, + "completions/min_length": 157.0, + "completions/min_terminated_length": 157.0, + "epoch": 14.951603498542275, + "grad_norm": 0.1586715281009674, + "learning_rate": 1e-06, + "loss": -0.0236, + "num_tokens": 951598343.0, + "reward": 0.7555803656578064, + "reward_std": 0.11809726804494858, + "rewards/verify_math_reward/mean": 0.7555803656578064, + "rewards/verify_math_reward/std": 0.42998257279396057, + "step": 1600 + }, + { + "epoch": 14.951603498542275, + "step": 1600, + "total_flos": 0.0, + "train_loss": -0.007817977162558236, + "train_runtime": 31764.4075, + "train_samples_per_second": 45.132, + "train_steps_per_second": 0.05 + } + ], + "logging_steps": 1, + "max_steps": 1600, + "num_input_tokens_seen": 951598343, + "num_train_epochs": 15, + "save_steps": 80, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": true + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 8, + "trial_name": null, + "trial_params": null +}