diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,12213 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "eval_steps": 500,
+  "global_step": 870,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.88542175292969,
+      "epoch": 0.011494252873563218,
+      "grad_norm": 1.3213221828105846,
+      "kl": 0.0,
+      "learning_rate": 9.988505747126437e-07,
+      "loss": 0.0,
+      "reward": 1.8197596073150635,
+      "reward_std": 0.06271697580814362,
+      "rewards/accuracy_reward": 0.8243168592453003,
+      "rewards/format_reward": 0.9954427480697632,
+      "step": 1
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.89388275146484,
+      "epoch": 0.022988505747126436,
+      "grad_norm": 1.006446520082249,
+      "kl": 0.000926971435546875,
+      "learning_rate": 9.977011494252872e-07,
+      "loss": 0.0001,
+      "reward": 1.7921943664550781,
+      "reward_std": 0.06994771957397461,
+      "rewards/accuracy_reward": 0.7961006164550781,
+      "rewards/format_reward": 0.99609375,
+      "step": 2
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.96745300292969,
+      "epoch": 0.034482758620689655,
+      "grad_norm": 1.4023786331817765,
+      "kl": 0.0013885498046875,
+      "learning_rate": 9.96551724137931e-07,
+      "loss": 0.0001,
+      "reward": 1.8005397319793701,
+      "reward_std": 0.06632152944803238,
+      "rewards/accuracy_reward": 0.8037948608398438,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 3
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.48958587646484,
+      "epoch": 0.04597701149425287,
+      "grad_norm": 1.0375104386573222,
+      "kl": 0.00238037109375,
+      "learning_rate": 9.954022988505747e-07,
+      "loss": 0.0001,
+      "reward": 1.801190733909607,
+      "reward_std": 0.06357432901859283,
+      "rewards/accuracy_reward": 0.8044459819793701,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 4
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.017578125,
+      "epoch": 0.05747126436781609,
+      "grad_norm": 1.2136231517307197,
+      "kl": 0.0023956298828125,
+      "learning_rate": 9.942528735632182e-07,
+      "loss": 0.0001,
+      "reward": 1.7834184169769287,
+      "reward_std": 0.06750813126564026,
+      "rewards/accuracy_reward": 0.7886267900466919,
+      "rewards/format_reward": 0.9947916865348816,
+      "step": 5
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.845703125,
+      "epoch": 0.06896551724137931,
+      "grad_norm": 1.9568797025276756,
+      "kl": 0.003631591796875,
+      "learning_rate": 9.93103448275862e-07,
+      "loss": 0.0001,
+      "reward": 1.816709280014038,
+      "reward_std": 0.06217224523425102,
+      "rewards/accuracy_reward": 0.8199643492698669,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 6
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.53450775146484,
+      "epoch": 0.08045977011494253,
+      "grad_norm": 1.0517811222524571,
+      "kl": 0.00457763671875,
+      "learning_rate": 9.919540229885057e-07,
+      "loss": 0.0002,
+      "reward": 1.8246958255767822,
+      "reward_std": 0.05271482467651367,
+      "rewards/accuracy_reward": 0.8246957063674927,
+      "rewards/format_reward": 1.0,
+      "step": 7
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.068359375,
+      "epoch": 0.09195402298850575,
+      "grad_norm": 1.1645981671220067,
+      "kl": 0.004638671875,
+      "learning_rate": 9.908045977011493e-07,
+      "loss": 0.0002,
+      "reward": 1.8276481628417969,
+      "reward_std": 0.05832577869296074,
+      "rewards/accuracy_reward": 0.8296012878417969,
+      "rewards/format_reward": 0.998046875,
+      "step": 8
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.41471862792969,
+      "epoch": 0.10344827586206896,
+      "grad_norm": 1.323752481080967,
+      "kl": 0.0045166015625,
+      "learning_rate": 9.89655172413793e-07,
+      "loss": 0.0002,
+      "reward": 1.7714653015136719,
+      "reward_std": 0.05968251824378967,
+      "rewards/accuracy_reward": 0.7734185457229614,
+      "rewards/format_reward": 0.998046875,
+      "step": 9
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.76042175292969,
+      "epoch": 0.11494252873563218,
+      "grad_norm": 1.412974502439727,
+      "kl": 0.00445556640625,
+      "learning_rate": 9.885057471264367e-07,
+      "loss": 0.0002,
+      "reward": 1.8277686834335327,
+      "reward_std": 0.05361395701766014,
+      "rewards/accuracy_reward": 0.8284196853637695,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 10
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.87630462646484,
+      "epoch": 0.12643678160919541,
+      "grad_norm": 2.075304847500042,
+      "kl": 0.00518798828125,
+      "learning_rate": 9.873563218390805e-07,
+      "loss": 0.0002,
+      "reward": 1.8043527603149414,
+      "reward_std": 0.05421295762062073,
+      "rewards/accuracy_reward": 0.805003821849823,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 11
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.30989837646484,
+      "epoch": 0.13793103448275862,
+      "grad_norm": 1.4906362546711622,
+      "kl": 0.005401611328125,
+      "learning_rate": 9.86206896551724e-07,
+      "loss": 0.0002,
+      "reward": 1.804455280303955,
+      "reward_std": 0.054727546870708466,
+      "rewards/accuracy_reward": 0.805106520652771,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 12
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.990234375,
+      "epoch": 0.14942528735632185,
+      "grad_norm": 1.5774307568902088,
+      "kl": 0.007354736328125,
+      "learning_rate": 9.850574712643678e-07,
+      "loss": 0.0003,
+      "reward": 1.8098227977752686,
+      "reward_std": 0.05695592984557152,
+      "rewards/accuracy_reward": 0.8098229169845581,
+      "rewards/format_reward": 1.0,
+      "step": 13
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.994140625,
+      "epoch": 0.16091954022988506,
+      "grad_norm": 1.1798175513976064,
+      "kl": 0.00909423828125,
+      "learning_rate": 9.839080459770115e-07,
+      "loss": 0.0004,
+      "reward": 1.804555892944336,
+      "reward_std": 0.054406534880399704,
+      "rewards/accuracy_reward": 0.8045558929443359,
+      "rewards/format_reward": 1.0,
+      "step": 14
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.54622650146484,
+      "epoch": 0.1724137931034483,
+      "grad_norm": 1.4179734216380988,
+      "kl": 0.00946044921875,
+      "learning_rate": 9.82758620689655e-07,
+      "loss": 0.0004,
+      "reward": 1.842616081237793,
+      "reward_std": 0.04714566469192505,
+      "rewards/accuracy_reward": 0.8432672023773193,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 15
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.39974212646484,
+      "epoch": 0.1839080459770115,
+      "grad_norm": 4.716568487745124,
+      "kl": 0.01055908203125,
+      "learning_rate": 9.816091954022988e-07,
+      "loss": 0.0004,
+      "reward": 1.7991819381713867,
+      "reward_std": 0.05012369155883789,
+      "rewards/accuracy_reward": 0.7991819381713867,
+      "rewards/format_reward": 1.0,
+      "step": 16
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.38216400146484,
+      "epoch": 0.19540229885057472,
+      "grad_norm": 1.288358836642226,
+      "kl": 0.01361083984375,
+      "learning_rate": 9.804597701149425e-07,
+      "loss": 0.0006,
+      "reward": 1.8223531246185303,
+      "reward_std": 0.05100415274500847,
+      "rewards/accuracy_reward": 0.8223528861999512,
+      "rewards/format_reward": 1.0,
+      "step": 17
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.85612487792969,
+      "epoch": 0.20689655172413793,
+      "grad_norm": 1.4234083835613787,
+      "kl": 0.01214599609375,
+      "learning_rate": 9.79310344827586e-07,
+      "loss": 0.0005,
+      "reward": 1.8194246292114258,
+      "reward_std": 0.04791136831045151,
+      "rewards/accuracy_reward": 0.8200756311416626,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 18
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.123046875,
+      "epoch": 0.21839080459770116,
+      "grad_norm": 1.4919735757668362,
+      "kl": 0.0126953125,
+      "learning_rate": 9.781609195402298e-07,
+      "loss": 0.0006,
+      "reward": 1.7989381551742554,
+      "reward_std": 0.05323593318462372,
+      "rewards/accuracy_reward": 0.7989381551742554,
+      "rewards/format_reward": 1.0,
+      "step": 19
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.22917175292969,
+      "epoch": 0.22988505747126436,
+      "grad_norm": 1.149828657266046,
+      "kl": 0.012939453125,
+      "learning_rate": 9.770114942528735e-07,
+      "loss": 0.0006,
+      "reward": 1.8272727727890015,
+      "reward_std": 0.04600293189287186,
+      "rewards/accuracy_reward": 0.8272727727890015,
+      "rewards/format_reward": 1.0,
+      "step": 20
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.38411712646484,
+      "epoch": 0.2413793103448276,
+      "grad_norm": 4.856552679666189,
+      "kl": 0.01416015625,
+      "learning_rate": 9.75862068965517e-07,
+      "loss": 0.0006,
+      "reward": 1.8167914152145386,
+      "reward_std": 0.04727660119533539,
+      "rewards/accuracy_reward": 0.816791296005249,
+      "rewards/format_reward": 1.0,
+      "step": 21
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.02734375,
+      "epoch": 0.25287356321839083,
+      "grad_norm": 1.4678458148251887,
+      "kl": 0.0172119140625,
+      "learning_rate": 9.747126436781608e-07,
+      "loss": 0.0007,
+      "reward": 1.8238710165023804,
+      "reward_std": 0.04785631224513054,
+      "rewards/accuracy_reward": 0.8245220184326172,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 22
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.50521087646484,
+      "epoch": 0.26436781609195403,
+      "grad_norm": 2.388579996164326,
+      "kl": 0.0181884765625,
+      "learning_rate": 9.735632183908046e-07,
+      "loss": 0.0007,
+      "reward": 1.8233423233032227,
+      "reward_std": 0.04244537279009819,
+      "rewards/accuracy_reward": 0.8233422040939331,
+      "rewards/format_reward": 1.0,
+      "step": 23
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.33138275146484,
+      "epoch": 0.27586206896551724,
+      "grad_norm": 1.5276786773430984,
+      "kl": 0.01806640625,
+      "learning_rate": 9.72413793103448e-07,
+      "loss": 0.0008,
+      "reward": 1.837747573852539,
+      "reward_std": 0.041725195944309235,
+      "rewards/accuracy_reward": 0.8383986353874207,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 24
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.095703125,
+      "epoch": 0.28735632183908044,
+      "grad_norm": 2.321983309381386,
+      "kl": 0.0213623046875,
+      "learning_rate": 9.712643678160918e-07,
+      "loss": 0.0009,
+      "reward": 1.7901400327682495,
+      "reward_std": 0.04634283110499382,
+      "rewards/accuracy_reward": 0.7901400923728943,
+      "rewards/format_reward": 1.0,
+      "step": 25
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.55859375,
+      "epoch": 0.2988505747126437,
+      "grad_norm": 2.351843921213514,
+      "kl": 0.023681640625,
+      "learning_rate": 9.701149425287356e-07,
+      "loss": 0.001,
+      "reward": 1.804998755455017,
+      "reward_std": 0.0408855676651001,
+      "rewards/accuracy_reward": 0.8049987554550171,
+      "rewards/format_reward": 1.0,
+      "step": 26
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.91732025146484,
+      "epoch": 0.3103448275862069,
+      "grad_norm": 1.1953728712395675,
+      "kl": 0.023193359375,
+      "learning_rate": 9.689655172413793e-07,
+      "loss": 0.001,
+      "reward": 1.8027284145355225,
+      "reward_std": 0.04391753673553467,
+      "rewards/accuracy_reward": 0.8033794164657593,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 27
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.53841400146484,
+      "epoch": 0.3218390804597701,
+      "grad_norm": 1.4473686865312074,
+      "kl": 0.0211181640625,
+      "learning_rate": 9.678160919540228e-07,
+      "loss": 0.0009,
+      "reward": 1.8385992050170898,
+      "reward_std": 0.038850087672472,
+      "rewards/accuracy_reward": 0.8385992050170898,
+      "rewards/format_reward": 1.0,
+      "step": 28
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.99153900146484,
+      "epoch": 0.3333333333333333,
+      "grad_norm": 1.7266188910571618,
+      "kl": 0.02099609375,
+      "learning_rate": 9.666666666666666e-07,
+      "loss": 0.0009,
+      "reward": 1.8396878242492676,
+      "reward_std": 0.03943753242492676,
+      "rewards/accuracy_reward": 0.840338945388794,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 29
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.44140625,
+      "epoch": 0.3448275862068966,
+      "grad_norm": 1.264532198419213,
+      "kl": 0.0211181640625,
+      "learning_rate": 9.655172413793103e-07,
+      "loss": 0.0009,
+      "reward": 1.819528341293335,
+      "reward_std": 0.04290936142206192,
+      "rewards/accuracy_reward": 0.8195282816886902,
+      "rewards/format_reward": 1.0,
+      "step": 30
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.53255462646484,
+      "epoch": 0.3563218390804598,
+      "grad_norm": 1.7079785748793026,
+      "kl": 0.0220947265625,
+      "learning_rate": 9.643678160919539e-07,
+      "loss": 0.0009,
+      "reward": 1.8193325996398926,
+      "reward_std": 0.039525143802165985,
+      "rewards/accuracy_reward": 0.8193327188491821,
+      "rewards/format_reward": 1.0,
+      "step": 31
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.01237487792969,
+      "epoch": 0.367816091954023,
+      "grad_norm": 1.1050680979361267,
+      "kl": 0.018310546875,
+      "learning_rate": 9.632183908045976e-07,
+      "loss": 0.0008,
+      "reward": 1.8326280117034912,
+      "reward_std": 0.03926192969083786,
+      "rewards/accuracy_reward": 0.833279013633728,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 32
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.38542175292969,
+      "epoch": 0.3793103448275862,
+      "grad_norm": 1.9594290453169385,
+      "kl": 0.018310546875,
+      "learning_rate": 9.620689655172413e-07,
+      "loss": 0.0008,
+      "reward": 1.830482840538025,
+      "reward_std": 0.03912848234176636,
+      "rewards/accuracy_reward": 0.8304828405380249,
+      "rewards/format_reward": 1.0,
+      "step": 33
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.35482025146484,
+      "epoch": 0.39080459770114945,
+      "grad_norm": 1.1967253632472934,
+      "kl": 0.01806640625,
+      "learning_rate": 9.609195402298849e-07,
+      "loss": 0.0007,
+      "reward": 1.8504703044891357,
+      "reward_std": 0.03594374656677246,
+      "rewards/accuracy_reward": 0.8504700660705566,
+      "rewards/format_reward": 1.0,
+      "step": 34
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.44271087646484,
+      "epoch": 0.40229885057471265,
+      "grad_norm": 1.4075688908737447,
+      "kl": 0.020751953125,
+      "learning_rate": 9.597701149425286e-07,
+      "loss": 0.0009,
+      "reward": 1.8235161304473877,
+      "reward_std": 0.03952939808368683,
+      "rewards/accuracy_reward": 0.8241671919822693,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 35
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.69010925292969,
+      "epoch": 0.41379310344827586,
+      "grad_norm": 1.1584071923987709,
+      "kl": 0.0166015625,
+      "learning_rate": 9.586206896551724e-07,
+      "loss": 0.0007,
+      "reward": 1.8258438110351562,
+      "reward_std": 0.03740541636943817,
+      "rewards/accuracy_reward": 0.8258438110351562,
+      "rewards/format_reward": 1.0,
+      "step": 36
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.52083587646484,
+      "epoch": 0.42528735632183906,
+      "grad_norm": 3.559550579405242,
+      "kl": 0.017578125,
+      "learning_rate": 9.57471264367816e-07,
+      "loss": 0.0007,
+      "reward": 1.8376644849777222,
+      "reward_std": 0.03800010681152344,
+      "rewards/accuracy_reward": 0.8376644849777222,
+      "rewards/format_reward": 1.0,
+      "step": 37
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.80078125,
+      "epoch": 0.4367816091954023,
+      "grad_norm": 1.414167289320497,
+      "kl": 0.017333984375,
+      "learning_rate": 9.563218390804596e-07,
+      "loss": 0.0007,
+      "reward": 1.8298163414001465,
+      "reward_std": 0.03874580189585686,
+      "rewards/accuracy_reward": 0.8304674029350281,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 38
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.865234375,
+      "epoch": 0.4482758620689655,
+      "grad_norm": 3.3277943720668635,
+      "kl": 0.01806640625,
+      "learning_rate": 9.551724137931034e-07,
+      "loss": 0.0008,
+      "reward": 1.8412444591522217,
+      "reward_std": 0.0331571064889431,
+      "rewards/accuracy_reward": 0.8412443995475769,
+      "rewards/format_reward": 1.0,
+      "step": 39
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.65625,
+      "epoch": 0.45977011494252873,
+      "grad_norm": 1.3938070271983027,
+      "kl": 0.01904296875,
+      "learning_rate": 9.540229885057471e-07,
+      "loss": 0.0008,
+      "reward": 1.829883098602295,
+      "reward_std": 0.03403263911604881,
+      "rewards/accuracy_reward": 0.8298830986022949,
+      "rewards/format_reward": 1.0,
+      "step": 40
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.31315612792969,
+      "epoch": 0.47126436781609193,
+      "grad_norm": 2.4737739620251444,
+      "kl": 0.01806640625,
+      "learning_rate": 9.528735632183908e-07,
+      "loss": 0.0008,
+      "reward": 1.8481197357177734,
+      "reward_std": 0.03319514915347099,
+      "rewards/accuracy_reward": 0.8481197357177734,
+      "rewards/format_reward": 1.0,
+      "step": 41
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.20247650146484,
+      "epoch": 0.4827586206896552,
+      "grad_norm": 3.381834125397947,
+      "kl": 0.0181884765625,
+      "learning_rate": 9.517241379310345e-07,
+      "loss": 0.0008,
+      "reward": 1.8192237615585327,
+      "reward_std": 0.03670245409011841,
+      "rewards/accuracy_reward": 0.8198747634887695,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 42
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.3125,
+      "epoch": 0.4942528735632184,
+      "grad_norm": 1.8762875005665294,
+      "kl": 0.0189208984375,
+      "learning_rate": 9.505747126436781e-07,
+      "loss": 0.0008,
+      "reward": 1.866302251815796,
+      "reward_std": 0.029023345559835434,
+      "rewards/accuracy_reward": 0.8663021922111511,
+      "rewards/format_reward": 1.0,
+      "step": 43
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.919921875,
+      "epoch": 0.5057471264367817,
+      "grad_norm": 1.364282450282579,
+      "kl": 0.0186767578125,
+      "learning_rate": 9.494252873563218e-07,
+      "loss": 0.0008,
+      "reward": 1.8467683792114258,
+      "reward_std": 0.032513901591300964,
+      "rewards/accuracy_reward": 0.8467683792114258,
+      "rewards/format_reward": 1.0,
+      "step": 44
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.521484375,
+      "epoch": 0.5172413793103449,
+      "grad_norm": 1.26529494442379,
+      "kl": 0.0184326171875,
+      "learning_rate": 9.482758620689655e-07,
+      "loss": 0.0008,
+      "reward": 1.8102507591247559,
+      "reward_std": 0.03791056573390961,
+      "rewards/accuracy_reward": 0.8115529417991638,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 45
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.720703125,
+      "epoch": 0.5287356321839081,
+      "grad_norm": 1.9592533419058873,
+      "kl": 0.0203857421875,
+      "learning_rate": 9.471264367816092e-07,
+      "loss": 0.0009,
+      "reward": 1.8435447216033936,
+      "reward_std": 0.03397071361541748,
+      "rewards/accuracy_reward": 0.844195544719696,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 46
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.55013275146484,
+      "epoch": 0.5402298850574713,
+      "grad_norm": 1.434106375840011,
+      "kl": 0.01953125,
+      "learning_rate": 9.459770114942528e-07,
+      "loss": 0.0008,
+      "reward": 1.8275771141052246,
+      "reward_std": 0.039521463215351105,
+      "rewards/accuracy_reward": 0.8288793563842773,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 47
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.88346862792969,
+      "epoch": 0.5517241379310345,
+      "grad_norm": 1.3331417373343877,
+      "kl": 0.019775390625,
+      "learning_rate": 9.448275862068965e-07,
+      "loss": 0.0008,
+      "reward": 1.847245216369629,
+      "reward_std": 0.031749702990055084,
+      "rewards/accuracy_reward": 0.8472453951835632,
+      "rewards/format_reward": 1.0,
+      "step": 48
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.53841400146484,
+      "epoch": 0.5632183908045977,
+      "grad_norm": 1.2065807063642335,
+      "kl": 0.0191650390625,
+      "learning_rate": 9.436781609195402e-07,
+      "loss": 0.0008,
+      "reward": 1.8429206609725952,
+      "reward_std": 0.030585885047912598,
+      "rewards/accuracy_reward": 0.8435718417167664,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 49
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.23503112792969,
+      "epoch": 0.5747126436781609,
+      "grad_norm": 1.8805762744730514,
+      "kl": 0.0216064453125,
+      "learning_rate": 9.425287356321838e-07,
+      "loss": 0.0009,
+      "reward": 1.8460264205932617,
+      "reward_std": 0.030056733638048172,
+      "rewards/accuracy_reward": 0.8460264205932617,
+      "rewards/format_reward": 1.0,
+      "step": 50
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.20964050292969,
+      "epoch": 0.5862068965517241,
+      "grad_norm": 2.416272655538971,
+      "kl": 0.031494140625,
+      "learning_rate": 9.413793103448276e-07,
+      "loss": 0.0013,
+      "reward": 1.8473304510116577,
+      "reward_std": 0.028418144211173058,
+      "rewards/accuracy_reward": 0.8473303318023682,
+      "rewards/format_reward": 1.0,
+      "step": 51
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.59245300292969,
+      "epoch": 0.5977011494252874,
+      "grad_norm": 1.7021853778739644,
+      "kl": 0.0206298828125,
+      "learning_rate": 9.402298850574713e-07,
+      "loss": 0.0009,
+      "reward": 1.837215781211853,
+      "reward_std": 0.027952462434768677,
+      "rewards/accuracy_reward": 0.837215781211853,
+      "rewards/format_reward": 1.0,
+      "step": 52
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.181640625,
+      "epoch": 0.6091954022988506,
+      "grad_norm": 1.5829777832061342,
+      "kl": 0.020263671875,
+      "learning_rate": 9.390804597701148e-07,
+      "loss": 0.0009,
+      "reward": 1.8315935134887695,
+      "reward_std": 0.03717650845646858,
+      "rewards/accuracy_reward": 0.8315935134887695,
+      "rewards/format_reward": 1.0,
+      "step": 53
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.23112487792969,
+      "epoch": 0.6206896551724138,
+      "grad_norm": 1.5951363696277052,
+      "kl": 0.027099609375,
+      "learning_rate": 9.379310344827586e-07,
+      "loss": 0.0011,
+      "reward": 1.8406846523284912,
+      "reward_std": 0.031024938449263573,
+      "rewards/accuracy_reward": 0.8413355350494385,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 54
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.03190612792969,
+      "epoch": 0.632183908045977,
+      "grad_norm": 1.3835743922707338,
+      "kl": 0.018798828125,
+      "learning_rate": 9.367816091954023e-07,
+      "loss": 0.0009,
+      "reward": 1.8285382986068726,
+      "reward_std": 0.02894814871251583,
+      "rewards/accuracy_reward": 0.8285383582115173,
+      "rewards/format_reward": 1.0,
+      "step": 55
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.86849212646484,
+      "epoch": 0.6436781609195402,
+      "grad_norm": 1.3591671428361543,
+      "kl": 0.019287109375,
+      "learning_rate": 9.356321839080458e-07,
+      "loss": 0.0008,
+      "reward": 1.8349835872650146,
+      "reward_std": 0.033723484724760056,
+      "rewards/accuracy_reward": 0.8356344699859619,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 56
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.513671875,
+      "epoch": 0.6551724137931034,
+      "grad_norm": 2.359691828231059,
+      "kl": 0.02392578125,
+      "learning_rate": 9.344827586206896e-07,
+      "loss": 0.001,
+      "reward": 1.8344968557357788,
+      "reward_std": 0.03300865739583969,
+      "rewards/accuracy_reward": 0.83514803647995,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 57
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.46875,
+      "epoch": 0.6666666666666666,
+      "grad_norm": 1.4850162206462427,
+      "kl": 0.020263671875,
+      "learning_rate": 9.333333333333333e-07,
+      "loss": 0.0009,
+      "reward": 1.8346723318099976,
+      "reward_std": 0.03354714438319206,
+      "rewards/accuracy_reward": 0.8346724510192871,
+      "rewards/format_reward": 1.0,
+      "step": 58
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.7109375,
+      "epoch": 0.6781609195402298,
+      "grad_norm": 1.2483974787682879,
+      "kl": 0.02197265625,
+      "learning_rate": 9.321839080459771e-07,
+      "loss": 0.0009,
+      "reward": 1.8298664093017578,
+      "reward_std": 0.03340629115700722,
+      "rewards/accuracy_reward": 0.8305175304412842,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 59
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.80078125,
+      "epoch": 0.6896551724137931,
+      "grad_norm": 2.363012214927635,
+      "kl": 0.02978515625,
+      "learning_rate": 9.310344827586206e-07,
+      "loss": 0.0012,
+      "reward": 1.831284761428833,
+      "reward_std": 0.03303222730755806,
+      "rewards/accuracy_reward": 0.8319358229637146,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 60
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.81771087646484,
+      "epoch": 0.7011494252873564,
+      "grad_norm": 4.0916014790353445,
+      "kl": 0.0196533203125,
+      "learning_rate": 9.298850574712643e-07,
+      "loss": 0.0008,
+      "reward": 1.839867353439331,
+      "reward_std": 0.031664229929447174,
+      "rewards/accuracy_reward": 0.8398674726486206,
+      "rewards/format_reward": 1.0,
+      "step": 61
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.52214050292969,
+      "epoch": 0.7126436781609196,
+      "grad_norm": 3.2309398976005297,
+      "kl": 0.022216796875,
+      "learning_rate": 9.287356321839081e-07,
+      "loss": 0.0009,
+      "reward": 1.854928970336914,
+      "reward_std": 0.031918901950120926,
+      "rewards/accuracy_reward": 0.8549291491508484,
+      "rewards/format_reward": 1.0,
+      "step": 62
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.703125,
+      "epoch": 0.7241379310344828,
+      "grad_norm": 1.9168568902881002,
+      "kl": 0.019287109375,
+      "learning_rate": 9.275862068965516e-07,
+      "loss": 0.0008,
+      "reward": 1.8610057830810547,
+      "reward_std": 0.02559598907828331,
+      "rewards/accuracy_reward": 0.8610057234764099,
+      "rewards/format_reward": 1.0,
+      "step": 63
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.36067962646484,
+      "epoch": 0.735632183908046,
+      "grad_norm": 1.3566272735502378,
+      "kl": 0.0191650390625,
+      "learning_rate": 9.264367816091954e-07,
+      "loss": 0.0008,
+      "reward": 1.8530144691467285,
+      "reward_std": 0.0317344069480896,
+      "rewards/accuracy_reward": 0.8530145883560181,
+      "rewards/format_reward": 1.0,
+      "step": 64
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.51888275146484,
+      "epoch": 0.7471264367816092,
+      "grad_norm": 1.9418693251170596,
+      "kl": 0.01806640625,
+      "learning_rate": 9.252873563218391e-07,
+      "loss": 0.0008,
+      "reward": 1.8553061485290527,
+      "reward_std": 0.028751468285918236,
+      "rewards/accuracy_reward": 0.8553061485290527,
+      "rewards/format_reward": 1.0,
+      "step": 65
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.462890625,
+      "epoch": 0.7586206896551724,
+      "grad_norm": 1.6584510097653566,
+      "kl": 0.0179443359375,
+      "learning_rate": 9.241379310344826e-07,
+      "loss": 0.0008,
+      "reward": 1.8472495079040527,
+      "reward_std": 0.029529428109526634,
+      "rewards/accuracy_reward": 0.8472495079040527,
+      "rewards/format_reward": 1.0,
+      "step": 66
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.38671875,
+      "epoch": 0.7701149425287356,
+      "grad_norm": 1.5373869592745633,
+      "kl": 0.0185546875,
+      "learning_rate": 9.229885057471264e-07,
+      "loss": 0.0008,
+      "reward": 1.8314546346664429,
+      "reward_std": 0.03092101775109768,
+      "rewards/accuracy_reward": 0.8314546346664429,
+      "rewards/format_reward": 1.0,
+      "step": 67
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.568359375,
+      "epoch": 0.7816091954022989,
+      "grad_norm": 1.4559609132563736,
+      "kl": 0.01806640625,
+      "learning_rate": 9.218390804597701e-07,
+      "loss": 0.0008,
+      "reward": 1.8663041591644287,
+      "reward_std": 0.026553651317954063,
+      "rewards/accuracy_reward": 0.8663042783737183,
+      "rewards/format_reward": 1.0,
+      "step": 68
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.46419525146484,
+      "epoch": 0.7931034482758621,
+      "grad_norm": 1.2008629627793295,
+      "kl": 0.0186767578125,
+      "learning_rate": 9.206896551724138e-07,
+      "loss": 0.0008,
+      "reward": 1.8613271713256836,
+      "reward_std": 0.02855812758207321,
+      "rewards/accuracy_reward": 0.8613271713256836,
+      "rewards/format_reward": 1.0,
+      "step": 69
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.99153900146484,
+      "epoch": 0.8045977011494253,
+      "grad_norm": 4.9275967882281275,
+      "kl": 0.020751953125,
+      "learning_rate": 9.195402298850574e-07,
+      "loss": 0.0009,
+      "reward": 1.8591216802597046,
+      "reward_std": 0.028761819005012512,
+      "rewards/accuracy_reward": 0.8591216802597046,
+      "rewards/format_reward": 1.0,
+      "step": 70
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.91536712646484,
+      "epoch": 0.8160919540229885,
+      "grad_norm": 9.850480698430928,
+      "kl": 0.01953125,
+      "learning_rate": 9.183908045977011e-07,
+      "loss": 0.0008,
+      "reward": 1.8541549444198608,
+      "reward_std": 0.028836514800786972,
+      "rewards/accuracy_reward": 0.8548059463500977,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 71
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.697265625,
+      "epoch": 0.8275862068965517,
+      "grad_norm": 1.210498300407391,
+      "kl": 0.0201416015625,
+      "learning_rate": 9.172413793103448e-07,
+      "loss": 0.0009,
+      "reward": 1.843854308128357,
+      "reward_std": 0.030750762671232224,
+      "rewards/accuracy_reward": 0.8438543677330017,
+      "rewards/format_reward": 1.0,
+      "step": 72
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.44075775146484,
+      "epoch": 0.8390804597701149,
+      "grad_norm": 1.8321539372725062,
+      "kl": 0.0205078125,
+      "learning_rate": 9.160919540229884e-07,
+      "loss": 0.0009,
+      "reward": 1.8534529209136963,
+      "reward_std": 0.028409739956259727,
+      "rewards/accuracy_reward": 0.8534530401229858,
+      "rewards/format_reward": 1.0,
+      "step": 73
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.00065612792969,
+      "epoch": 0.8505747126436781,
+      "grad_norm": 1.1934111085412287,
+      "kl": 0.0218505859375,
+      "learning_rate": 9.149425287356322e-07,
+      "loss": 0.0009,
+      "reward": 1.8261594772338867,
+      "reward_std": 0.028989605605602264,
+      "rewards/accuracy_reward": 0.8268105387687683,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 74
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.86458587646484,
+      "epoch": 0.8620689655172413,
+      "grad_norm": 1.1973438361715183,
+      "kl": 0.02294921875,
+      "learning_rate": 9.137931034482759e-07,
+      "loss": 0.001,
+      "reward": 1.8427714109420776,
+      "reward_std": 0.028574138879776,
+      "rewards/accuracy_reward": 0.8427714109420776,
+      "rewards/format_reward": 1.0,
+      "step": 75
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.2109375,
+      "epoch": 0.8735632183908046,
+      "grad_norm": 1.3720840552975169,
+      "kl": 0.0208740234375,
+      "learning_rate": 9.126436781609194e-07,
+      "loss": 0.0009,
+      "reward": 1.8438997268676758,
+      "reward_std": 0.034771427512168884,
+      "rewards/accuracy_reward": 0.8452019095420837,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 76
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.87825775146484,
+      "epoch": 0.8850574712643678,
+      "grad_norm": 1.2733923638265987,
+      "kl": 0.021240234375,
+      "learning_rate": 9.114942528735632e-07,
+      "loss": 0.0009,
+      "reward": 1.8525419235229492,
+      "reward_std": 0.02576223388314247,
+      "rewards/accuracy_reward": 0.8525419235229492,
+      "rewards/format_reward": 1.0,
+      "step": 77
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.30599212646484,
+      "epoch": 0.896551724137931,
+      "grad_norm": 1.5683715987496294,
+      "kl": 0.022705078125,
+      "learning_rate": 9.103448275862069e-07,
+      "loss": 0.0009,
+      "reward": 1.8584070205688477,
+      "reward_std": 0.026499662548303604,
+      "rewards/accuracy_reward": 0.8590580821037292,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 78
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.30208587646484,
+      "epoch": 0.9080459770114943,
+      "grad_norm": 1.7144360008338546,
+      "kl": 0.0206298828125,
+      "learning_rate": 9.091954022988505e-07,
+      "loss": 0.0009,
+      "reward": 1.836766004562378,
+      "reward_std": 0.03153412044048309,
+      "rewards/accuracy_reward": 0.8374168276786804,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 79
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.78646087646484,
+      "epoch": 0.9195402298850575,
+      "grad_norm": 1.4249106766015625,
+      "kl": 0.021484375,
+      "learning_rate": 9.080459770114942e-07,
+      "loss": 0.0009,
+      "reward": 1.8593868017196655,
+      "reward_std": 0.030518915504217148,
+      "rewards/accuracy_reward": 0.859386682510376,
+      "rewards/format_reward": 1.0,
+      "step": 80
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.84700775146484,
+      "epoch": 0.9310344827586207,
+      "grad_norm": 1.6378679289752103,
+      "kl": 0.0205078125,
+      "learning_rate": 9.068965517241379e-07,
+      "loss": 0.0009,
+      "reward": 1.8533859252929688,
+      "reward_std": 0.02878081053495407,
+      "rewards/accuracy_reward": 0.8533859252929688,
+      "rewards/format_reward": 1.0,
+      "step": 81
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.23567962646484,
+      "epoch": 0.9425287356321839,
+      "grad_norm": 1.412691105042645,
+      "kl": 0.021484375,
+      "learning_rate": 9.057471264367816e-07,
+      "loss": 0.0009,
+      "reward": 1.847813606262207,
+      "reward_std": 0.03038616292178631,
+      "rewards/accuracy_reward": 0.8484646677970886,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 82
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.27474212646484,
+      "epoch": 0.9540229885057471,
+      "grad_norm": 2.0344799768549784,
+      "kl": 0.0244140625,
+      "learning_rate": 9.045977011494252e-07,
+      "loss": 0.001,
+      "reward": 1.8461253643035889,
+      "reward_std": 0.03104616329073906,
+      "rewards/accuracy_reward": 0.8467763662338257,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 83
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.6640625,
+      "epoch": 0.9655172413793104,
+      "grad_norm": 1.5604062393784244,
+      "kl": 0.0211181640625,
+      "learning_rate": 9.034482758620689e-07,
+      "loss": 0.0009,
+      "reward": 1.8446769714355469,
+      "reward_std": 0.025367258116602898,
+      "rewards/accuracy_reward": 0.8446769714355469,
+      "rewards/format_reward": 1.0,
+      "step": 84
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.02734375,
+      "epoch": 0.9770114942528736,
+      "grad_norm": 1.362117300400181,
+      "kl": 0.0216064453125,
+      "learning_rate": 9.022988505747126e-07,
+      "loss": 0.0009,
+      "reward": 1.850293517112732,
+      "reward_std": 0.027060410007834435,
+      "rewards/accuracy_reward": 0.8502935171127319,
+      "rewards/format_reward": 1.0,
+      "step": 85
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.21159362792969,
+      "epoch": 0.9885057471264368,
+      "grad_norm": 4.622693800475361,
+      "kl": 0.0196533203125,
+      "learning_rate": 9.011494252873562e-07,
+      "loss": 0.0008,
+      "reward": 1.8619520664215088,
+      "reward_std": 0.02936243824660778,
+      "rewards/accuracy_reward": 0.8626030683517456,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 86
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.17837524414062,
+      "epoch": 1.0,
+      "grad_norm": 1.1739865165675025,
+      "kl": 0.0206298828125,
+      "learning_rate": 9e-07,
+      "loss": 0.0009,
+      "reward": 1.865034818649292,
+      "reward_std": 0.024553239345550537,
+      "rewards/accuracy_reward": 0.8650349974632263,
+      "rewards/format_reward": 1.0,
+      "step": 87
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.279296875,
+      "epoch": 1.0114942528735633,
+      "grad_norm": 1.7212140289390774,
+      "kl": 0.022216796875,
+      "learning_rate": 8.988505747126436e-07,
+      "loss": 0.001,
+      "reward": 1.844975471496582,
+      "reward_std": 0.031139438971877098,
+      "rewards/accuracy_reward": 0.8449755907058716,
+      "rewards/format_reward": 1.0,
+      "step": 88
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.97721862792969,
+      "epoch": 1.0229885057471264,
+      "grad_norm": 2.1208124047907018,
+      "kl": 0.02294921875,
+      "learning_rate": 8.977011494252873e-07,
+      "loss": 0.001,
+      "reward": 1.8310832977294922,
+      "reward_std": 0.030721893534064293,
+      "rewards/accuracy_reward": 0.8310832381248474,
+      "rewards/format_reward": 1.0,
+      "step": 89
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.72982025146484,
+      "epoch": 1.0344827586206897,
+      "grad_norm": 1.7904458786353679,
+      "kl": 0.025390625,
+      "learning_rate": 8.96551724137931e-07,
+      "loss": 0.0011,
+      "reward": 1.8376115560531616,
+      "reward_std": 0.03200274333357811,
+      "rewards/accuracy_reward": 0.8382627367973328,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 90
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.40495300292969,
+      "epoch": 1.0459770114942528,
+      "grad_norm": 1.6518476245455913,
+      "kl": 0.0216064453125,
+      "learning_rate": 8.954022988505747e-07,
+      "loss": 0.0009,
+      "reward": 1.8409684896469116,
+      "reward_std": 0.02987781912088394,
+      "rewards/accuracy_reward": 0.8416194915771484,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 91
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.744140625,
+      "epoch": 1.0574712643678161,
+      "grad_norm": 2.410653089678698,
+      "kl": 0.021484375,
+      "learning_rate": 8.942528735632184e-07,
+      "loss": 0.0009,
+      "reward": 1.8022823333740234,
+      "reward_std": 0.03017442300915718,
+      "rewards/accuracy_reward": 0.8022822141647339,
+      "rewards/format_reward": 1.0,
+      "step": 92
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.03255462646484,
+      "epoch": 1.0689655172413792,
+      "grad_norm": 1.070833843767388,
+      "kl": 0.0230712890625,
+      "learning_rate": 8.93103448275862e-07,
+      "loss": 0.001,
+      "reward": 1.8557833433151245,
+      "reward_std": 0.030436282977461815,
+      "rewards/accuracy_reward": 0.8570854663848877,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 93
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.84635925292969,
+      "epoch": 1.0804597701149425,
+      "grad_norm": 1.2651791750173826,
+      "kl": 0.0216064453125,
+      "learning_rate": 8.919540229885057e-07,
+      "loss": 0.0009,
+      "reward": 1.8365752696990967,
+      "reward_std": 0.025262486189603806,
+      "rewards/accuracy_reward": 0.8372262716293335,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 94
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.35091400146484,
+      "epoch": 1.0919540229885056,
+      "grad_norm": 2.5454147273169463,
+      "kl": 0.0213623046875,
+      "learning_rate": 8.908045977011494e-07,
+      "loss": 0.0009,
+      "reward": 1.8451789617538452,
+      "reward_std": 0.02507897838950157,
+      "rewards/accuracy_reward": 0.8451790809631348,
+      "rewards/format_reward": 1.0,
+      "step": 95
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.16862487792969,
+      "epoch": 1.103448275862069,
+      "grad_norm": 1.5956876645494014,
+      "kl": 0.020263671875,
+      "learning_rate": 8.896551724137931e-07,
+      "loss": 0.0009,
+      "reward": 1.8393771648406982,
+      "reward_std": 0.02759551629424095,
+      "rewards/accuracy_reward": 0.8393771052360535,
+      "rewards/format_reward": 1.0,
+      "step": 96
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.232421875,
+      "epoch": 1.1149425287356323,
+      "grad_norm": 1.6538617605670678,
+      "kl": 0.0230712890625,
+      "learning_rate": 8.885057471264368e-07,
+      "loss": 0.001,
+      "reward": 1.8187679052352905,
+      "reward_std": 0.028722627088427544,
+      "rewards/accuracy_reward": 0.8187679052352905,
+      "rewards/format_reward": 1.0,
+      "step": 97
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.37109375,
+      "epoch": 1.1264367816091954,
+      "grad_norm": 2.466954207815668,
+      "kl": 0.020263671875,
+      "learning_rate": 8.873563218390804e-07,
+      "loss": 0.0009,
+      "reward": 1.8639543056488037,
+      "reward_std": 0.02820052206516266,
+      "rewards/accuracy_reward": 0.8646053075790405,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 98
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.6171875,
+      "epoch": 1.1379310344827587,
+      "grad_norm": 1.252778063732612,
+      "kl": 0.0211181640625,
+      "learning_rate": 8.862068965517241e-07,
+      "loss": 0.0009,
+      "reward": 1.8373963832855225,
+      "reward_std": 0.027982115745544434,
+      "rewards/accuracy_reward": 0.837396502494812,
+      "rewards/format_reward": 1.0,
+      "step": 99
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.939453125,
+      "epoch": 1.1494252873563218,
+      "grad_norm": 1.5031364905308602,
+      "kl": 0.021240234375,
+      "learning_rate": 8.850574712643678e-07,
+      "loss": 0.0009,
+      "reward": 1.8384751081466675,
+      "reward_std": 0.02889677882194519,
+      "rewards/accuracy_reward": 0.8384751081466675,
+      "rewards/format_reward": 1.0,
+      "step": 100
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.173828125,
+      "epoch": 1.160919540229885,
+      "grad_norm": 3.2098798594469957,
+      "kl": 0.022216796875,
+      "learning_rate": 8.839080459770114e-07,
+      "loss": 0.001,
+      "reward": 1.858123779296875,
+      "reward_std": 0.027859613299369812,
+      "rewards/accuracy_reward": 0.8581236600875854,
+      "rewards/format_reward": 1.0,
+      "step": 101
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.611328125,
+      "epoch": 1.1724137931034484,
+      "grad_norm": 1.6395969811083824,
+      "kl": 0.022705078125,
+      "learning_rate": 8.827586206896551e-07,
+      "loss": 0.001,
+      "reward": 1.8613924980163574,
+      "reward_std": 0.030063219368457794,
+      "rewards/accuracy_reward": 0.8626946806907654,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 102
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.9609375,
+      "epoch": 1.1839080459770115,
+      "grad_norm": 1.5759802737821602,
+      "kl": 0.0218505859375,
+      "learning_rate": 8.816091954022988e-07,
+      "loss": 0.001,
+      "reward": 1.87135648727417,
+      "reward_std": 0.025703629478812218,
+      "rewards/accuracy_reward": 0.8713564872741699,
+      "rewards/format_reward": 1.0,
+      "step": 103
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.47591400146484,
+      "epoch": 1.1954022988505748,
+      "grad_norm": 2.1153792944695753,
+      "kl": 0.0218505859375,
+      "learning_rate": 8.804597701149424e-07,
+      "loss": 0.001,
+      "reward": 1.8344066143035889,
+      "reward_std": 0.02677982673048973,
+      "rewards/accuracy_reward": 0.8344065546989441,
+      "rewards/format_reward": 1.0,
+      "step": 104
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.17513275146484,
+      "epoch": 1.206896551724138,
+      "grad_norm": 1.5729705960989235,
+      "kl": 0.0223388671875,
+      "learning_rate": 8.793103448275862e-07,
+      "loss": 0.0009,
+      "reward": 1.8618988990783691,
+      "reward_std": 0.025562942028045654,
+      "rewards/accuracy_reward": 0.8618988990783691,
+      "rewards/format_reward": 1.0,
+      "step": 105
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.10417175292969,
+      "epoch": 1.2183908045977012,
+      "grad_norm": 1.5280787952862458,
+      "kl": 0.021484375,
+      "learning_rate": 8.781609195402299e-07,
+      "loss": 0.0009,
+      "reward": 1.8250049352645874,
+      "reward_std": 0.030421411618590355,
+      "rewards/accuracy_reward": 0.8263069987297058,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 106
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.99674987792969,
+      "epoch": 1.2298850574712643,
+      "grad_norm": 1.3417425340508164,
+      "kl": 0.02294921875,
+      "learning_rate": 8.770114942528735e-07,
+      "loss": 0.001,
+      "reward": 1.8627097606658936,
+      "reward_std": 0.029694318771362305,
+      "rewards/accuracy_reward": 0.8633607625961304,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 107
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.16732025146484,
+      "epoch": 1.2413793103448276,
+      "grad_norm": 1.3637947595566382,
+      "kl": 0.023193359375,
+      "learning_rate": 8.758620689655172e-07,
+      "loss": 0.001,
+      "reward": 1.8729504346847534,
+      "reward_std": 0.02589436247944832,
+      "rewards/accuracy_reward": 0.8729504942893982,
+      "rewards/format_reward": 1.0,
+      "step": 108
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.79232025146484,
+      "epoch": 1.2528735632183907,
+      "grad_norm": 1.8473222859814702,
+      "kl": 0.02294921875,
+      "learning_rate": 8.747126436781609e-07,
+      "loss": 0.001,
+      "reward": 1.8625543117523193,
+      "reward_std": 0.026772310957312584,
+      "rewards/accuracy_reward": 0.8625543117523193,
+      "rewards/format_reward": 1.0,
+      "step": 109
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.390625,
+      "epoch": 1.264367816091954,
+      "grad_norm": 1.6150275987926945,
+      "kl": 0.024658203125,
+      "learning_rate": 8.735632183908046e-07,
+      "loss": 0.001,
+      "reward": 1.8422784805297852,
+      "reward_std": 0.030604541301727295,
+      "rewards/accuracy_reward": 0.8435806632041931,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 110
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.96354675292969,
+      "epoch": 1.2758620689655173,
+      "grad_norm": 1.7104109971917698,
+      "kl": 0.03125,
+      "learning_rate": 8.724137931034482e-07,
+      "loss": 0.0013,
+      "reward": 1.8350896835327148,
+      "reward_std": 0.028940599411725998,
+      "rewards/accuracy_reward": 0.8357405662536621,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 111
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.47135925292969,
+      "epoch": 1.2873563218390804,
+      "grad_norm": 1.48493737972391,
+      "kl": 0.02490234375,
+      "learning_rate": 8.712643678160919e-07,
+      "loss": 0.0011,
+      "reward": 1.8481062650680542,
+      "reward_std": 0.029237091541290283,
+      "rewards/accuracy_reward": 0.8481062650680542,
+      "rewards/format_reward": 1.0,
+      "step": 112
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.853515625,
+      "epoch": 1.2988505747126438,
+      "grad_norm": 1.8812767455198556,
+      "kl": 0.024169921875,
+      "learning_rate": 8.701149425287357e-07,
+      "loss": 0.001,
+      "reward": 1.8743027448654175,
+      "reward_std": 0.028526946902275085,
+      "rewards/accuracy_reward": 0.8743027448654175,
+      "rewards/format_reward": 1.0,
+      "step": 113
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.11653900146484,
+      "epoch": 1.3103448275862069,
+      "grad_norm": 1.4552951626639703,
+      "kl": 0.0244140625,
+      "learning_rate": 8.689655172413792e-07,
+      "loss": 0.001,
+      "reward": 1.8737125396728516,
+      "reward_std": 0.02662818133831024,
+      "rewards/accuracy_reward": 0.8737127184867859,
+      "rewards/format_reward": 1.0,
+      "step": 114
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.68620300292969,
+      "epoch": 1.3218390804597702,
+      "grad_norm": 1.2099323721036837,
+      "kl": 0.0242919921875,
+      "learning_rate": 8.67816091954023e-07,
+      "loss": 0.001,
+      "reward": 1.8564056158065796,
+      "reward_std": 0.03203584998846054,
+      "rewards/accuracy_reward": 0.8570567965507507,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 115
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.44140625,
+      "epoch": 1.3333333333333333,
+      "grad_norm": 1.512025900233657,
+      "kl": 0.0262451171875,
+      "learning_rate": 8.666666666666667e-07,
+      "loss": 0.0011,
+      "reward": 1.8566768169403076,
+      "reward_std": 0.028156783431768417,
+      "rewards/accuracy_reward": 0.857327938079834,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 116
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.33073425292969,
+      "epoch": 1.3448275862068966,
+      "grad_norm": 2.0358354675170904,
+      "kl": 0.02392578125,
+      "learning_rate": 8.655172413793102e-07,
+      "loss": 0.001,
+      "reward": 1.8653233051300049,
+      "reward_std": 0.03316108509898186,
+      "rewards/accuracy_reward": 0.8666254281997681,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 117
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.40234375,
+      "epoch": 1.3563218390804597,
+      "grad_norm": 4.5664679538349136,
+      "kl": 0.0260009765625,
+      "learning_rate": 8.64367816091954e-07,
+      "loss": 0.0011,
+      "reward": 1.8475828170776367,
+      "reward_std": 0.028350701555609703,
+      "rewards/accuracy_reward": 0.8475827574729919,
+      "rewards/format_reward": 1.0,
+      "step": 118
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.509765625,
+      "epoch": 1.367816091954023,
+      "grad_norm": 3.3498974515563305,
+      "kl": 0.02392578125,
+      "learning_rate": 8.632183908045977e-07,
+      "loss": 0.001,
+      "reward": 1.8256216049194336,
+      "reward_std": 0.03040180169045925,
+      "rewards/accuracy_reward": 0.8256216049194336,
+      "rewards/format_reward": 1.0,
+      "step": 119
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.119140625,
+      "epoch": 1.3793103448275863,
+      "grad_norm": 2.0523739279685618,
+      "kl": 0.03564453125,
+      "learning_rate": 8.620689655172412e-07,
+      "loss": 0.0015,
+      "reward": 1.8660404682159424,
+      "reward_std": 0.02924610674381256,
+      "rewards/accuracy_reward": 0.8660405874252319,
+      "rewards/format_reward": 1.0,
+      "step": 120
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.34440612792969,
+      "epoch": 1.3908045977011494,
+      "grad_norm": 4.105539506193381,
+      "kl": 0.0361328125,
+      "learning_rate": 8.60919540229885e-07,
+      "loss": 0.0015,
+      "reward": 1.8672317266464233,
+      "reward_std": 0.0260650422424078,
+      "rewards/accuracy_reward": 0.8672318458557129,
+      "rewards/format_reward": 1.0,
+      "step": 121
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.32878112792969,
+      "epoch": 1.4022988505747127,
+      "grad_norm": 1.7057843588411452,
+      "kl": 0.0303955078125,
+      "learning_rate": 8.597701149425287e-07,
+      "loss": 0.0012,
+      "reward": 1.8730971813201904,
+      "reward_std": 0.032900359481573105,
+      "rewards/accuracy_reward": 0.8737481832504272,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 122
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.82421875,
+      "epoch": 1.4137931034482758,
+      "grad_norm": 2.6319293895634828,
+      "kl": 0.024658203125,
+      "learning_rate": 8.586206896551725e-07,
+      "loss": 0.0011,
+      "reward": 1.8619554042816162,
+      "reward_std": 0.025756051763892174,
+      "rewards/accuracy_reward": 0.8619554042816162,
+      "rewards/format_reward": 1.0,
+      "step": 123
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.46810150146484,
+      "epoch": 1.4252873563218391,
+      "grad_norm": 1.8929420183153929,
+      "kl": 0.0264892578125,
+      "learning_rate": 8.57471264367816e-07,
+      "loss": 0.0011,
+      "reward": 1.8555949926376343,
+      "reward_std": 0.027035661041736603,
+      "rewards/accuracy_reward": 0.855595052242279,
+      "rewards/format_reward": 1.0,
+      "step": 124
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.79362487792969,
+      "epoch": 1.4367816091954024,
+      "grad_norm": 4.4900781638741645,
+      "kl": 0.0264892578125,
+      "learning_rate": 8.563218390804597e-07,
+      "loss": 0.0011,
+      "reward": 1.8362255096435547,
+      "reward_std": 0.0311919953674078,
+      "rewards/accuracy_reward": 0.836225688457489,
+      "rewards/format_reward": 1.0,
+      "step": 125
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.28190612792969,
+      "epoch": 1.4482758620689655,
+      "grad_norm": 2.686615982073592,
+      "kl": 0.0294189453125,
+      "learning_rate": 8.551724137931035e-07,
+      "loss": 0.0012,
+      "reward": 1.8675731420516968,
+      "reward_std": 0.02920936420559883,
+      "rewards/accuracy_reward": 0.8682241439819336,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 126
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.333984375,
+      "epoch": 1.4597701149425286,
+      "grad_norm": 1.4366785062548348,
+      "kl": 0.0245361328125,
+      "learning_rate": 8.54022988505747e-07,
+      "loss": 0.001,
+      "reward": 1.8725385665893555,
+      "reward_std": 0.028989439830183983,
+      "rewards/accuracy_reward": 0.8725385665893555,
+      "rewards/format_reward": 1.0,
+      "step": 127
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.35286712646484,
+      "epoch": 1.471264367816092,
+      "grad_norm": 3.153210014316318,
+      "kl": 0.025146484375,
+      "learning_rate": 8.528735632183908e-07,
+      "loss": 0.0011,
+      "reward": 1.8572996854782104,
+      "reward_std": 0.02787836454808712,
+      "rewards/accuracy_reward": 0.8579506874084473,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 128
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.70768737792969,
+      "epoch": 1.4827586206896552,
+      "grad_norm": 1.312276972160894,
+      "kl": 0.023681640625,
+      "learning_rate": 8.517241379310345e-07,
+      "loss": 0.001,
+      "reward": 1.8692939281463623,
+      "reward_std": 0.028887853026390076,
+      "rewards/accuracy_reward": 0.8692940473556519,
+      "rewards/format_reward": 1.0,
+      "step": 129
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.83333587646484,
+      "epoch": 1.4942528735632183,
+      "grad_norm": 1.7208602570293514,
+      "kl": 0.0279541015625,
+      "learning_rate": 8.50574712643678e-07,
+      "loss": 0.0012,
+      "reward": 1.86876380443573,
+      "reward_std": 0.026345182210206985,
+      "rewards/accuracy_reward": 0.86876380443573,
+      "rewards/format_reward": 1.0,
+      "step": 130
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.58919525146484,
+      "epoch": 1.5057471264367817,
+      "grad_norm": 1.5360698582173107,
+      "kl": 0.02490234375,
+      "learning_rate": 8.494252873563218e-07,
+      "loss": 0.0011,
+      "reward": 1.8543965816497803,
+      "reward_std": 0.025941472500562668,
+      "rewards/accuracy_reward": 0.8543965816497803,
+      "rewards/format_reward": 1.0,
+      "step": 131
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.38607025146484,
+      "epoch": 1.5172413793103448,
+      "grad_norm": 1.5741802825985347,
+      "kl": 0.0244140625,
+      "learning_rate": 8.482758620689655e-07,
+      "loss": 0.0011,
+      "reward": 1.8572925329208374,
+      "reward_std": 0.02648422122001648,
+      "rewards/accuracy_reward": 0.8572925329208374,
+      "rewards/format_reward": 1.0,
+      "step": 132
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.529296875,
+      "epoch": 1.528735632183908,
+      "grad_norm": 2.5696729376896577,
+      "kl": 0.02685546875,
+      "learning_rate": 8.471264367816092e-07,
+      "loss": 0.0012,
+      "reward": 1.8575305938720703,
+      "reward_std": 0.023815011605620384,
+      "rewards/accuracy_reward": 0.8575305938720703,
+      "rewards/format_reward": 1.0,
+      "step": 133
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.123046875,
+      "epoch": 1.5402298850574714,
+      "grad_norm": 1.6144061698095113,
+      "kl": 0.029052734375,
+      "learning_rate": 8.459770114942528e-07,
+      "loss": 0.0013,
+      "reward": 1.8380780220031738,
+      "reward_std": 0.02718261629343033,
+      "rewards/accuracy_reward": 0.8380780220031738,
+      "rewards/format_reward": 1.0,
+      "step": 134
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.32552337646484,
+      "epoch": 1.5517241379310345,
+      "grad_norm": 1.4219475281295146,
+      "kl": 0.0281982421875,
+      "learning_rate": 8.448275862068965e-07,
+      "loss": 0.0012,
+      "reward": 1.8486030101776123,
+      "reward_std": 0.02719944715499878,
+      "rewards/accuracy_reward": 0.8492540121078491,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 135
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.57747650146484,
+      "epoch": 1.5632183908045976,
+      "grad_norm": 1.2505748762461004,
+      "kl": 0.0284423828125,
+      "learning_rate": 8.436781609195402e-07,
+      "loss": 0.0012,
+      "reward": 1.8550881147384644,
+      "reward_std": 0.024348240345716476,
+      "rewards/accuracy_reward": 0.8550881147384644,
+      "rewards/format_reward": 1.0,
+      "step": 136
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.931640625,
+      "epoch": 1.5747126436781609,
+      "grad_norm": 1.4653956935392716,
+      "kl": 0.030517578125,
+      "learning_rate": 8.425287356321838e-07,
+      "loss": 0.0013,
+      "reward": 1.8590949773788452,
+      "reward_std": 0.024702435359358788,
+      "rewards/accuracy_reward": 0.8597459197044373,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 137
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.66796875,
+      "epoch": 1.5862068965517242,
+      "grad_norm": 1.5116616200702753,
+      "kl": 0.0257568359375,
+      "learning_rate": 8.413793103448276e-07,
+      "loss": 0.0011,
+      "reward": 1.865983247756958,
+      "reward_std": 0.02651612088084221,
+      "rewards/accuracy_reward": 0.8659831881523132,
+      "rewards/format_reward": 1.0,
+      "step": 138
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.427734375,
+      "epoch": 1.5977011494252875,
+      "grad_norm": 2.539622540808917,
+      "kl": 0.026123046875,
+      "learning_rate": 8.402298850574713e-07,
+      "loss": 0.0011,
+      "reward": 1.8506149053573608,
+      "reward_std": 0.024511417374014854,
+      "rewards/accuracy_reward": 0.8506147861480713,
+      "rewards/format_reward": 1.0,
+      "step": 139
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.677734375,
+      "epoch": 1.6091954022988506,
+      "grad_norm": 2.7029566355060575,
+      "kl": 0.030029296875,
+      "learning_rate": 8.390804597701148e-07,
+      "loss": 0.0013,
+      "reward": 1.8749054670333862,
+      "reward_std": 0.022635221481323242,
+      "rewards/accuracy_reward": 0.8749054670333862,
+      "rewards/format_reward": 1.0,
+      "step": 140
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.28190612792969,
+      "epoch": 1.6206896551724137,
+      "grad_norm": 1.6125529702224055,
+      "kl": 0.0306396484375,
+      "learning_rate": 8.379310344827586e-07,
+      "loss": 0.0013,
+      "reward": 1.8687496185302734,
+      "reward_std": 0.02769436687231064,
+      "rewards/accuracy_reward": 0.8687496185302734,
+      "rewards/format_reward": 1.0,
+      "step": 141
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.73177337646484,
+      "epoch": 1.632183908045977,
+      "grad_norm": 1.6175458343103775,
+      "kl": 0.03076171875,
+      "learning_rate": 8.367816091954023e-07,
+      "loss": 0.0013,
+      "reward": 1.8663349151611328,
+      "reward_std": 0.028380822390317917,
+      "rewards/accuracy_reward": 0.8663349151611328,
+      "rewards/format_reward": 1.0,
+      "step": 142
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.64128112792969,
+      "epoch": 1.6436781609195403,
+      "grad_norm": 1.5215374212801362,
+      "kl": 0.03515625,
+      "learning_rate": 8.35632183908046e-07,
+      "loss": 0.0015,
+      "reward": 1.850799560546875,
+      "reward_std": 0.026175182312726974,
+      "rewards/accuracy_reward": 0.8507994413375854,
+      "rewards/format_reward": 1.0,
+      "step": 143
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.19792175292969,
+      "epoch": 1.6551724137931034,
+      "grad_norm": 1.3135082334407917,
+      "kl": 0.035400390625,
+      "learning_rate": 8.344827586206896e-07,
+      "loss": 0.0015,
+      "reward": 1.8789458274841309,
+      "reward_std": 0.023477703332901,
+      "rewards/accuracy_reward": 0.8789458274841309,
+      "rewards/format_reward": 1.0,
+      "step": 144
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.71940612792969,
+      "epoch": 1.6666666666666665,
+      "grad_norm": 1.7597191781848593,
+      "kl": 0.0400390625,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 0.0017,
+      "reward": 1.862714171409607,
+      "reward_std": 0.02495882660150528,
+      "rewards/accuracy_reward": 0.8627142906188965,
+      "rewards/format_reward": 1.0,
+      "step": 145
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.86653900146484,
+      "epoch": 1.6781609195402298,
+      "grad_norm": 1.1271572137611656,
+      "kl": 0.041015625,
+      "learning_rate": 8.32183908045977e-07,
+      "loss": 0.0017,
+      "reward": 1.863875389099121,
+      "reward_std": 0.022430753335356712,
+      "rewards/accuracy_reward": 0.8638753890991211,
+      "rewards/format_reward": 1.0,
+      "step": 146
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.78971862792969,
+      "epoch": 1.6896551724137931,
+      "grad_norm": 1.7900229849601172,
+      "kl": 0.0361328125,
+      "learning_rate": 8.310344827586206e-07,
+      "loss": 0.0015,
+      "reward": 1.8638134002685547,
+      "reward_std": 0.024309411644935608,
+      "rewards/accuracy_reward": 0.8638134002685547,
+      "rewards/format_reward": 1.0,
+      "step": 147
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.46159362792969,
+      "epoch": 1.7011494252873565,
+      "grad_norm": 1.4211028736618636,
+      "kl": 0.03466796875,
+      "learning_rate": 8.298850574712643e-07,
+      "loss": 0.0015,
+      "reward": 1.8384380340576172,
+      "reward_std": 0.027933349832892418,
+      "rewards/accuracy_reward": 0.8390890955924988,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 148
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.04817962646484,
+      "epoch": 1.7126436781609196,
+      "grad_norm": 1.6133215111280255,
+      "kl": 0.033935546875,
+      "learning_rate": 8.28735632183908e-07,
+      "loss": 0.0014,
+      "reward": 1.8644790649414062,
+      "reward_std": 0.023334093391895294,
+      "rewards/accuracy_reward": 0.8644790649414062,
+      "rewards/format_reward": 1.0,
+      "step": 149
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.50130462646484,
+      "epoch": 1.7241379310344827,
+      "grad_norm": 1.71716415208747,
+      "kl": 0.0380859375,
+      "learning_rate": 8.275862068965517e-07,
+      "loss": 0.0016,
+      "reward": 1.863791823387146,
+      "reward_std": 0.025837119668722153,
+      "rewards/accuracy_reward": 0.8644427061080933,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 150
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.55339050292969,
+      "epoch": 1.735632183908046,
+      "grad_norm": 1.7799429692812343,
+      "kl": 0.032470703125,
+      "learning_rate": 8.264367816091954e-07,
+      "loss": 0.0014,
+      "reward": 1.8520851135253906,
+      "reward_std": 0.02567559853196144,
+      "rewards/accuracy_reward": 0.8520849943161011,
+      "rewards/format_reward": 1.0,
+      "step": 151
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.16015625,
+      "epoch": 1.7471264367816093,
+      "grad_norm": 6.845861661718817,
+      "kl": 0.033203125,
+      "learning_rate": 8.25287356321839e-07,
+      "loss": 0.0014,
+      "reward": 1.8801301717758179,
+      "reward_std": 0.02385757677257061,
+      "rewards/accuracy_reward": 0.8801302909851074,
+      "rewards/format_reward": 1.0,
+      "step": 152
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.34635925292969,
+      "epoch": 1.7586206896551724,
+      "grad_norm": 2.31131545485098,
+      "kl": 0.030029296875,
+      "learning_rate": 8.241379310344827e-07,
+      "loss": 0.0013,
+      "reward": 1.8744109869003296,
+      "reward_std": 0.02264220081269741,
+      "rewards/accuracy_reward": 0.8744109869003296,
+      "rewards/format_reward": 1.0,
+      "step": 153
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.56901550292969,
+      "epoch": 1.7701149425287355,
+      "grad_norm": 1.773846804268088,
+      "kl": 0.0302734375,
+      "learning_rate": 8.229885057471264e-07,
+      "loss": 0.0013,
+      "reward": 1.864478349685669,
+      "reward_std": 0.022709330543875694,
+      "rewards/accuracy_reward": 0.864478349685669,
+      "rewards/format_reward": 1.0,
+      "step": 154
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.39583587646484,
+      "epoch": 1.7816091954022988,
+      "grad_norm": 3.139231735878295,
+      "kl": 0.03662109375,
+      "learning_rate": 8.218390804597701e-07,
+      "loss": 0.0016,
+      "reward": 1.8694690465927124,
+      "reward_std": 0.023755362257361412,
+      "rewards/accuracy_reward": 0.8701201677322388,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 155
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.47786712646484,
+      "epoch": 1.793103448275862,
+      "grad_norm": 1.978341379497861,
+      "kl": 0.033203125,
+      "learning_rate": 8.206896551724138e-07,
+      "loss": 0.0014,
+      "reward": 1.8873121738433838,
+      "reward_std": 0.02343663200736046,
+      "rewards/accuracy_reward": 0.8873120546340942,
+      "rewards/format_reward": 1.0,
+      "step": 156
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.451171875,
+      "epoch": 1.8045977011494254,
+      "grad_norm": 1.9961053634302923,
+      "kl": 0.031494140625,
+      "learning_rate": 8.195402298850574e-07,
+      "loss": 0.0013,
+      "reward": 1.8565791845321655,
+      "reward_std": 0.02433241717517376,
+      "rewards/accuracy_reward": 0.8565791845321655,
+      "rewards/format_reward": 1.0,
+      "step": 157
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.66015625,
+      "epoch": 1.8160919540229885,
+      "grad_norm": 1.4619162689448402,
+      "kl": 0.0279541015625,
+      "learning_rate": 8.183908045977011e-07,
+      "loss": 0.0012,
+      "reward": 1.893612265586853,
+      "reward_std": 0.020092740654945374,
+      "rewards/accuracy_reward": 0.893612265586853,
+      "rewards/format_reward": 1.0,
+      "step": 158
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.15299987792969,
+      "epoch": 1.8275862068965516,
+      "grad_norm": 1.3651687933062773,
+      "kl": 0.0283203125,
+      "learning_rate": 8.172413793103448e-07,
+      "loss": 0.0012,
+      "reward": 1.8662446737289429,
+      "reward_std": 0.02392234466969967,
+      "rewards/accuracy_reward": 0.8662446737289429,
+      "rewards/format_reward": 1.0,
+      "step": 159
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.61458587646484,
+      "epoch": 1.839080459770115,
+      "grad_norm": 1.2135275799512748,
+      "kl": 0.0262451171875,
+      "learning_rate": 8.160919540229885e-07,
+      "loss": 0.0011,
+      "reward": 1.8714094161987305,
+      "reward_std": 0.024804111570119858,
+      "rewards/accuracy_reward": 0.8714094161987305,
+      "rewards/format_reward": 1.0,
+      "step": 160
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.740234375,
+      "epoch": 1.8505747126436782,
+      "grad_norm": 1.8554584998042072,
+      "kl": 0.03125,
+      "learning_rate": 8.149425287356322e-07,
+      "loss": 0.0013,
+      "reward": 1.885859489440918,
+      "reward_std": 0.0239472147077322,
+      "rewards/accuracy_reward": 0.8858596682548523,
+      "rewards/format_reward": 1.0,
+      "step": 161
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.990234375,
+      "epoch": 1.8620689655172413,
+      "grad_norm": 2.7992855237483445,
+      "kl": 0.027587890625,
+      "learning_rate": 8.137931034482758e-07,
+      "loss": 0.0012,
+      "reward": 1.8619565963745117,
+      "reward_std": 0.025602132081985474,
+      "rewards/accuracy_reward": 0.8619565963745117,
+      "rewards/format_reward": 1.0,
+      "step": 162
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.46810150146484,
+      "epoch": 1.8735632183908046,
+      "grad_norm": 1.153159461213296,
+      "kl": 0.0291748046875,
+      "learning_rate": 8.126436781609195e-07,
+      "loss": 0.0012,
+      "reward": 1.8758127689361572,
+      "reward_std": 0.02637392282485962,
+      "rewards/accuracy_reward": 0.876463770866394,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 163
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 115.09440612792969,
+      "epoch": 1.8850574712643677,
+      "grad_norm": 1.8638466478476596,
+      "kl": 0.031005859375,
+      "learning_rate": 8.114942528735632e-07,
+      "loss": 0.0013,
+      "reward": 1.8715770244598389,
+      "reward_std": 0.02399243414402008,
+      "rewards/accuracy_reward": 0.8715770244598389,
+      "rewards/format_reward": 1.0,
+      "step": 164
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 116.47265625,
+      "epoch": 1.896551724137931,
+      "grad_norm": 1.1827867594913961,
+      "kl": 0.0291748046875,
+      "learning_rate": 8.103448275862068e-07,
+      "loss": 0.0012,
+      "reward": 1.8500595092773438,
+      "reward_std": 0.030623754486441612,
+      "rewards/accuracy_reward": 0.8520126342773438,
+      "rewards/format_reward": 0.998046875,
+      "step": 165
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 116.31510925292969,
+      "epoch": 1.9080459770114944,
+      "grad_norm": 3.8800308289928274,
+      "kl": 0.037353515625,
+      "learning_rate": 8.091954022988506e-07,
+      "loss": 0.0016,
+      "reward": 1.8412752151489258,
+      "reward_std": 0.02564075216650963,
+      "rewards/accuracy_reward": 0.8412752151489258,
+      "rewards/format_reward": 1.0,
+      "step": 166
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.9921875,
+      "epoch": 1.9195402298850575,
+      "grad_norm": 1.7793621376166886,
+      "kl": 0.0322265625,
+      "learning_rate": 8.080459770114942e-07,
+      "loss": 0.0014,
+      "reward": 1.8666815757751465,
+      "reward_std": 0.01957223378121853,
+      "rewards/accuracy_reward": 0.8666815757751465,
+      "rewards/format_reward": 1.0,
+      "step": 167
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 116.001953125,
+      "epoch": 1.9310344827586206,
+      "grad_norm": 3.366934016860489,
+      "kl": 0.0306396484375,
+      "learning_rate": 8.068965517241378e-07,
+      "loss": 0.0013,
+      "reward": 1.8858134746551514,
+      "reward_std": 0.021044649183750153,
+      "rewards/accuracy_reward": 0.8858134150505066,
+      "rewards/format_reward": 1.0,
+      "step": 168
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.40495300292969,
+      "epoch": 1.9425287356321839,
+      "grad_norm": 1.4480350054291504,
+      "kl": 0.06689453125,
+      "learning_rate": 8.057471264367816e-07,
+      "loss": 0.0028,
+      "reward": 1.888282299041748,
+      "reward_std": 0.025501668453216553,
+      "rewards/accuracy_reward": 0.8889333605766296,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 169
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.05339050292969,
+      "epoch": 1.9540229885057472,
+      "grad_norm": 1.9158676691938177,
+      "kl": 0.0322265625,
+      "learning_rate": 8.045977011494253e-07,
+      "loss": 0.0014,
+      "reward": 1.8491878509521484,
+      "reward_std": 0.023884495720267296,
+      "rewards/accuracy_reward": 0.8491878509521484,
+      "rewards/format_reward": 1.0,
+      "step": 170
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.62760925292969,
+      "epoch": 1.9655172413793105,
+      "grad_norm": 2.780653624143398,
+      "kl": 0.033203125,
+      "learning_rate": 8.03448275862069e-07,
+      "loss": 0.0014,
+      "reward": 1.887178897857666,
+      "reward_std": 0.02224629744887352,
+      "rewards/accuracy_reward": 0.887178897857666,
+      "rewards/format_reward": 1.0,
+      "step": 171
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.57096862792969,
+      "epoch": 1.9770114942528736,
+      "grad_norm": 1.9318099805591311,
+      "kl": 0.031494140625,
+      "learning_rate": 8.022988505747126e-07,
+      "loss": 0.0013,
+      "reward": 1.8778488636016846,
+      "reward_std": 0.024461418390274048,
+      "rewards/accuracy_reward": 0.8784998655319214,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 172
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.85612487792969,
+      "epoch": 1.9885057471264367,
+      "grad_norm": 1.8812732346859014,
+      "kl": 0.031982421875,
+      "learning_rate": 8.011494252873563e-07,
+      "loss": 0.0013,
+      "reward": 1.896362543106079,
+      "reward_std": 0.021257508546113968,
+      "rewards/accuracy_reward": 0.8963624238967896,
+      "rewards/format_reward": 1.0,
+      "step": 173
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.70365142822266,
+      "epoch": 2.0,
+      "grad_norm": 4.371328763954834,
+      "kl": 0.033935546875,
+      "learning_rate": 8e-07,
+      "loss": 0.0014,
+      "reward": 1.8666094541549683,
+      "reward_std": 0.023954574018716812,
+      "rewards/accuracy_reward": 0.8680139780044556,
+      "rewards/format_reward": 0.9985955357551575,
+      "step": 174
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.845703125,
+      "epoch": 2.0114942528735633,
+      "grad_norm": 3.642725753887041,
+      "kl": 0.028564453125,
+      "learning_rate": 7.988505747126436e-07,
+      "loss": 0.0012,
+      "reward": 1.8493850231170654,
+      "reward_std": 0.021547168493270874,
+      "rewards/accuracy_reward": 0.8493850231170654,
+      "rewards/format_reward": 1.0,
+      "step": 175
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.64518737792969,
+      "epoch": 2.0229885057471266,
+      "grad_norm": 1.813754894659516,
+      "kl": 0.0308837890625,
+      "learning_rate": 7.977011494252873e-07,
+      "loss": 0.0013,
+      "reward": 1.8406703472137451,
+      "reward_std": 0.022422365844249725,
+      "rewards/accuracy_reward": 0.8406702876091003,
+      "rewards/format_reward": 1.0,
+      "step": 176
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.38021087646484,
+      "epoch": 2.0344827586206895,
+      "grad_norm": 1.1813724927973488,
+      "kl": 0.030029296875,
+      "learning_rate": 7.965517241379311e-07,
+      "loss": 0.0013,
+      "reward": 1.8357830047607422,
+      "reward_std": 0.02784503623843193,
+      "rewards/accuracy_reward": 0.8370852470397949,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 177
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.87890625,
+      "epoch": 2.045977011494253,
+      "grad_norm": 1.8350693828851994,
+      "kl": 0.032470703125,
+      "learning_rate": 7.954022988505746e-07,
+      "loss": 0.0014,
+      "reward": 1.8329381942749023,
+      "reward_std": 0.02206650748848915,
+      "rewards/accuracy_reward": 0.8329381942749023,
+      "rewards/format_reward": 1.0,
+      "step": 178
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.98112487792969,
+      "epoch": 2.057471264367816,
+      "grad_norm": 6.041427121899008,
+      "kl": 0.03662109375,
+      "learning_rate": 7.942528735632184e-07,
+      "loss": 0.0016,
+      "reward": 1.8575410842895508,
+      "reward_std": 0.022278612479567528,
+      "rewards/accuracy_reward": 0.8575412034988403,
+      "rewards/format_reward": 1.0,
+      "step": 179
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.130859375,
+      "epoch": 2.0689655172413794,
+      "grad_norm": 1.3127737451001564,
+      "kl": 0.037109375,
+      "learning_rate": 7.931034482758621e-07,
+      "loss": 0.0016,
+      "reward": 1.8538506031036377,
+      "reward_std": 0.026692640036344528,
+      "rewards/accuracy_reward": 0.8551527261734009,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 180
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.36849212646484,
+      "epoch": 2.0804597701149423,
+      "grad_norm": 2.4756435889269186,
+      "kl": 0.03759765625,
+      "learning_rate": 7.919540229885056e-07,
+      "loss": 0.0016,
+      "reward": 1.867905616760254,
+      "reward_std": 0.023630604147911072,
+      "rewards/accuracy_reward": 0.8685566782951355,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 181
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.97005462646484,
+      "epoch": 2.0919540229885056,
+      "grad_norm": 2.0378313336318823,
+      "kl": 0.039794921875,
+      "learning_rate": 7.908045977011494e-07,
+      "loss": 0.0017,
+      "reward": 1.8557872772216797,
+      "reward_std": 0.024489006027579308,
+      "rewards/accuracy_reward": 0.856438159942627,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 182
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.88802337646484,
+      "epoch": 2.103448275862069,
+      "grad_norm": 1.9136054549569206,
+      "kl": 0.042724609375,
+      "learning_rate": 7.896551724137931e-07,
+      "loss": 0.0018,
+      "reward": 1.8639075756072998,
+      "reward_std": 0.025589074939489365,
+      "rewards/accuracy_reward": 0.865860641002655,
+      "rewards/format_reward": 0.998046875,
+      "step": 183
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.36653900146484,
+      "epoch": 2.1149425287356323,
+      "grad_norm": 2.633698077075855,
+      "kl": 0.038330078125,
+      "learning_rate": 7.885057471264366e-07,
+      "loss": 0.0016,
+      "reward": 1.8632612228393555,
+      "reward_std": 0.030802268534898758,
+      "rewards/accuracy_reward": 0.8671674728393555,
+      "rewards/format_reward": 0.99609375,
+      "step": 184
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.41796875,
+      "epoch": 2.1264367816091956,
+      "grad_norm": 2.407318367905216,
+      "kl": 0.03662109375,
+      "learning_rate": 7.873563218390804e-07,
+      "loss": 0.0015,
+      "reward": 1.8410747051239014,
+      "reward_std": 0.02962721884250641,
+      "rewards/accuracy_reward": 0.844329833984375,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 185
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.22591400146484,
+      "epoch": 2.1379310344827585,
+      "grad_norm": 1.506301090880529,
+      "kl": 0.033935546875,
+      "learning_rate": 7.862068965517241e-07,
+      "loss": 0.0014,
+      "reward": 1.8682615756988525,
+      "reward_std": 0.03243795782327652,
+      "rewards/accuracy_reward": 0.8734698295593262,
+      "rewards/format_reward": 0.9947916865348816,
+      "step": 186
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.826171875,
+      "epoch": 2.1494252873563218,
+      "grad_norm": 3.409263140415921,
+      "kl": 0.03662109375,
+      "learning_rate": 7.850574712643679e-07,
+      "loss": 0.0015,
+      "reward": 1.875942587852478,
+      "reward_std": 0.03060537576675415,
+      "rewards/accuracy_reward": 0.8791979551315308,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 187
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.181640625,
+      "epoch": 2.160919540229885,
+      "grad_norm": 3.163868782667335,
+      "kl": 0.031005859375,
+      "learning_rate": 7.839080459770114e-07,
+      "loss": 0.0013,
+      "reward": 1.8670661449432373,
+      "reward_std": 0.03131501376628876,
+      "rewards/accuracy_reward": 0.8703212738037109,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 188
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.701171875,
+      "epoch": 2.1724137931034484,
+      "grad_norm": 2.26227947471076,
+      "kl": 0.0299072265625,
+      "learning_rate": 7.827586206896552e-07,
+      "loss": 0.0012,
+      "reward": 1.8607233762741089,
+      "reward_std": 0.030661556869745255,
+      "rewards/accuracy_reward": 0.8639785647392273,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 189
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.59505462646484,
+      "epoch": 2.1839080459770113,
+      "grad_norm": 2.1434523868036885,
+      "kl": 0.0308837890625,
+      "learning_rate": 7.816091954022989e-07,
+      "loss": 0.0013,
+      "reward": 1.8534635305404663,
+      "reward_std": 0.03587997704744339,
+      "rewards/accuracy_reward": 0.8580207824707031,
+      "rewards/format_reward": 0.9954427480697632,
+      "step": 190
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.111328125,
+      "epoch": 2.1954022988505746,
+      "grad_norm": 1.8017139066303491,
+      "kl": 0.034423828125,
+      "learning_rate": 7.804597701149424e-07,
+      "loss": 0.0015,
+      "reward": 1.8646326065063477,
+      "reward_std": 0.031285952776670456,
+      "rewards/accuracy_reward": 0.8672367930412292,
+      "rewards/format_reward": 0.9973958730697632,
+      "step": 191
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.44921875,
+      "epoch": 2.206896551724138,
+      "grad_norm": 1.4767880401798514,
+      "kl": 0.0302734375,
+      "learning_rate": 7.793103448275862e-07,
+      "loss": 0.0013,
+      "reward": 1.8693158626556396,
+      "reward_std": 0.03322295472025871,
+      "rewards/accuracy_reward": 0.8725711107254028,
+      "rewards/format_reward": 0.9967448115348816,
+      "step": 192
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.24674987792969,
+      "epoch": 2.218390804597701,
+      "grad_norm": 1.1854249053762298,
+      "kl": 0.0294189453125,
+      "learning_rate": 7.781609195402299e-07,
+      "loss": 0.0013,
+      "reward": 1.8650622367858887,
+      "reward_std": 0.03180694580078125,
+      "rewards/accuracy_reward": 0.867666482925415,
+      "rewards/format_reward": 0.9973958730697632,
+      "step": 193
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.73372650146484,
+      "epoch": 2.2298850574712645,
+      "grad_norm": 1.1139391781844512,
+      "kl": 0.033447265625,
+      "learning_rate": 7.770114942528734e-07,
+      "loss": 0.0014,
+      "reward": 1.87180757522583,
+      "reward_std": 0.03504088148474693,
+      "rewards/accuracy_reward": 0.8757138252258301,
+      "rewards/format_reward": 0.99609375,
+      "step": 194
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.34245300292969,
+      "epoch": 2.2413793103448274,
+      "grad_norm": 1.6586210582781218,
+      "kl": 0.032470703125,
+      "learning_rate": 7.758620689655172e-07,
+      "loss": 0.0014,
+      "reward": 1.8676183223724365,
+      "reward_std": 0.028974760323762894,
+      "rewards/accuracy_reward": 0.8689204454421997,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 195
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.36653900146484,
+      "epoch": 2.2528735632183907,
+      "grad_norm": 1.1693377970360763,
+      "kl": 0.0361328125,
+      "learning_rate": 7.747126436781609e-07,
+      "loss": 0.0015,
+      "reward": 1.8650755882263184,
+      "reward_std": 0.03180007264018059,
+      "rewards/accuracy_reward": 0.8676798343658447,
+      "rewards/format_reward": 0.9973958730697632,
+      "step": 196
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.658203125,
+      "epoch": 2.264367816091954,
+      "grad_norm": 5.251694511136823,
+      "kl": 0.032470703125,
+      "learning_rate": 7.735632183908046e-07,
+      "loss": 0.0014,
+      "reward": 1.8851394653320312,
+      "reward_std": 0.028069045394659042,
+      "rewards/accuracy_reward": 0.8864415884017944,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 197
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.24674987792969,
+      "epoch": 2.2758620689655173,
+      "grad_norm": 1.3160075358259606,
+      "kl": 0.0306396484375,
+      "learning_rate": 7.724137931034482e-07,
+      "loss": 0.0013,
+      "reward": 1.871217966079712,
+      "reward_std": 0.02678101509809494,
+      "rewards/accuracy_reward": 0.8718689680099487,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 198
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.22396087646484,
+      "epoch": 2.2873563218390807,
+      "grad_norm": 1.7363926773399032,
+      "kl": 0.0322265625,
+      "learning_rate": 7.712643678160919e-07,
+      "loss": 0.0014,
+      "reward": 1.8681142330169678,
+      "reward_std": 0.025049438700079918,
+      "rewards/accuracy_reward": 0.8687652349472046,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 199
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.36328125,
+      "epoch": 2.2988505747126435,
+      "grad_norm": 1.6358927960776593,
+      "kl": 0.032470703125,
+      "learning_rate": 7.701149425287356e-07,
+      "loss": 0.0014,
+      "reward": 1.8802045583724976,
+      "reward_std": 0.025107329711318016,
+      "rewards/accuracy_reward": 0.8808557391166687,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 200
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.34049987792969,
+      "epoch": 2.310344827586207,
+      "grad_norm": 1.9769472874812075,
+      "kl": 0.034912109375,
+      "learning_rate": 7.689655172413792e-07,
+      "loss": 0.0015,
+      "reward": 1.8836421966552734,
+      "reward_std": 0.023718908429145813,
+      "rewards/accuracy_reward": 0.8842934370040894,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 201
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.58203125,
+      "epoch": 2.32183908045977,
+      "grad_norm": 1.5235580660765125,
+      "kl": 0.04150390625,
+      "learning_rate": 7.67816091954023e-07,
+      "loss": 0.0017,
+      "reward": 1.872357726097107,
+      "reward_std": 0.027882186695933342,
+      "rewards/accuracy_reward": 0.8736597895622253,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 202
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.212890625,
+      "epoch": 2.3333333333333335,
+      "grad_norm": 1.6195272681997732,
+      "kl": 0.03515625,
+      "learning_rate": 7.666666666666667e-07,
+      "loss": 0.0015,
+      "reward": 1.864328145980835,
+      "reward_std": 0.023437950760126114,
+      "rewards/accuracy_reward": 0.8649791479110718,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 203
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.525390625,
+      "epoch": 2.344827586206897,
+      "grad_norm": 1.8341465818857292,
+      "kl": 0.038818359375,
+      "learning_rate": 7.655172413793102e-07,
+      "loss": 0.0016,
+      "reward": 1.8627145290374756,
+      "reward_std": 0.022373493760824203,
+      "rewards/accuracy_reward": 0.8627144694328308,
+      "rewards/format_reward": 1.0,
+      "step": 204
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.689453125,
+      "epoch": 2.3563218390804597,
+      "grad_norm": 1.1804395489400663,
+      "kl": 0.0390625,
+      "learning_rate": 7.64367816091954e-07,
+      "loss": 0.0016,
+      "reward": 1.8903453350067139,
+      "reward_std": 0.02181018702685833,
+      "rewards/accuracy_reward": 0.8903453350067139,
+      "rewards/format_reward": 1.0,
+      "step": 205
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.21875,
+      "epoch": 2.367816091954023,
+      "grad_norm": 1.54557267471986,
+      "kl": 0.03955078125,
+      "learning_rate": 7.632183908045977e-07,
+      "loss": 0.0016,
+      "reward": 1.8669939041137695,
+      "reward_std": 0.02308422140777111,
+      "rewards/accuracy_reward": 0.8669940829277039,
+      "rewards/format_reward": 1.0,
+      "step": 206
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.130859375,
+      "epoch": 2.3793103448275863,
+      "grad_norm": 1.6770832687532953,
+      "kl": 0.041259765625,
+      "learning_rate": 7.620689655172414e-07,
+      "loss": 0.0017,
+      "reward": 1.8921034336090088,
+      "reward_std": 0.024557670578360558,
+      "rewards/accuracy_reward": 0.8934054374694824,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 207
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.56510925292969,
+      "epoch": 2.3908045977011496,
+      "grad_norm": 1.53505457334918,
+      "kl": 0.04150390625,
+      "learning_rate": 7.60919540229885e-07,
+      "loss": 0.0017,
+      "reward": 1.8635462522506714,
+      "reward_std": 0.028757482767105103,
+      "rewards/accuracy_reward": 0.8635462522506714,
+      "rewards/format_reward": 1.0,
+      "step": 208
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.521484375,
+      "epoch": 2.4022988505747125,
+      "grad_norm": 1.5279469918158477,
+      "kl": 0.0458984375,
+      "learning_rate": 7.597701149425287e-07,
+      "loss": 0.0019,
+      "reward": 1.8711223602294922,
+      "reward_std": 0.022866621613502502,
+      "rewards/accuracy_reward": 0.8711225390434265,
+      "rewards/format_reward": 1.0,
+      "step": 209
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.22135925292969,
+      "epoch": 2.413793103448276,
+      "grad_norm": 1.9431939650471932,
+      "kl": 0.04296875,
+      "learning_rate": 7.586206896551724e-07,
+      "loss": 0.0018,
+      "reward": 1.8868978023529053,
+      "reward_std": 0.02192886732518673,
+      "rewards/accuracy_reward": 0.8868978023529053,
+      "rewards/format_reward": 1.0,
+      "step": 210
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.751953125,
+      "epoch": 2.425287356321839,
+      "grad_norm": 1.499449740827804,
+      "kl": 0.042236328125,
+      "learning_rate": 7.57471264367816e-07,
+      "loss": 0.0017,
+      "reward": 1.865365743637085,
+      "reward_std": 0.025590822100639343,
+      "rewards/accuracy_reward": 0.8660167455673218,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 211
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.19075775146484,
+      "epoch": 2.4367816091954024,
+      "grad_norm": 2.0546219049276684,
+      "kl": 0.042724609375,
+      "learning_rate": 7.563218390804598e-07,
+      "loss": 0.0017,
+      "reward": 1.8754686117172241,
+      "reward_std": 0.02179308794438839,
+      "rewards/accuracy_reward": 0.8754686713218689,
+      "rewards/format_reward": 1.0,
+      "step": 212
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.09700775146484,
+      "epoch": 2.4482758620689653,
+      "grad_norm": 1.7497621009886284,
+      "kl": 0.0361328125,
+      "learning_rate": 7.551724137931034e-07,
+      "loss": 0.0015,
+      "reward": 1.8816776275634766,
+      "reward_std": 0.023087939247488976,
+      "rewards/accuracy_reward": 0.8816776275634766,
+      "rewards/format_reward": 1.0,
+      "step": 213
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.85612487792969,
+      "epoch": 2.4597701149425286,
+      "grad_norm": 1.6521998481125082,
+      "kl": 0.037109375,
+      "learning_rate": 7.540229885057471e-07,
+      "loss": 0.0016,
+      "reward": 1.8775031566619873,
+      "reward_std": 0.022142350673675537,
+      "rewards/accuracy_reward": 0.8781541585922241,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 214
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.697265625,
+      "epoch": 2.471264367816092,
+      "grad_norm": 1.4603832529701488,
+      "kl": 0.037353515625,
+      "learning_rate": 7.528735632183908e-07,
+      "loss": 0.0015,
+      "reward": 1.8763501644134521,
+      "reward_std": 0.02745872735977173,
+      "rewards/accuracy_reward": 0.8770012259483337,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 215
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.99284362792969,
+      "epoch": 2.4827586206896552,
+      "grad_norm": 1.4199974173722163,
+      "kl": 0.033447265625,
+      "learning_rate": 7.517241379310344e-07,
+      "loss": 0.0014,
+      "reward": 1.879393458366394,
+      "reward_std": 0.023880278691649437,
+      "rewards/accuracy_reward": 0.8800445795059204,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 216
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.63216400146484,
+      "epoch": 2.4942528735632186,
+      "grad_norm": 1.4805312918485005,
+      "kl": 0.034423828125,
+      "learning_rate": 7.505747126436781e-07,
+      "loss": 0.0014,
+      "reward": 1.8871581554412842,
+      "reward_std": 0.023221522569656372,
+      "rewards/accuracy_reward": 0.8871581554412842,
+      "rewards/format_reward": 1.0,
+      "step": 217
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.69010925292969,
+      "epoch": 2.5057471264367814,
+      "grad_norm": 1.5882030215776515,
+      "kl": 0.031982421875,
+      "learning_rate": 7.494252873563218e-07,
+      "loss": 0.0014,
+      "reward": 1.8449172973632812,
+      "reward_std": 0.0257129929959774,
+      "rewards/accuracy_reward": 0.8449174761772156,
+      "rewards/format_reward": 1.0,
+      "step": 218
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.49089050292969,
+      "epoch": 2.5172413793103448,
+      "grad_norm": 3.140150954079358,
+      "kl": 0.0322265625,
+      "learning_rate": 7.482758620689655e-07,
+      "loss": 0.0014,
+      "reward": 1.8943170309066772,
+      "reward_std": 0.023106535896658897,
+      "rewards/accuracy_reward": 0.8949680328369141,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 219
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.779296875,
+      "epoch": 2.528735632183908,
+      "grad_norm": 1.8149238607460914,
+      "kl": 0.032958984375,
+      "learning_rate": 7.471264367816092e-07,
+      "loss": 0.0014,
+      "reward": 1.8684821128845215,
+      "reward_std": 0.023636629804968834,
+      "rewards/accuracy_reward": 0.8684821128845215,
+      "rewards/format_reward": 1.0,
+      "step": 220
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.23503112792969,
+      "epoch": 2.5402298850574714,
+      "grad_norm": 2.0933238300012715,
+      "kl": 0.048095703125,
+      "learning_rate": 7.459770114942528e-07,
+      "loss": 0.002,
+      "reward": 1.8713550567626953,
+      "reward_std": 0.023653965443372726,
+      "rewards/accuracy_reward": 0.8713551759719849,
+      "rewards/format_reward": 1.0,
+      "step": 221
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.32292175292969,
+      "epoch": 2.5517241379310347,
+      "grad_norm": 2.40931495027469,
+      "kl": 0.0311279296875,
+      "learning_rate": 7.448275862068965e-07,
+      "loss": 0.0013,
+      "reward": 1.8844565153121948,
+      "reward_std": 0.025886138901114464,
+      "rewards/accuracy_reward": 0.8844565153121948,
+      "rewards/format_reward": 1.0,
+      "step": 222
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.625,
+      "epoch": 2.5632183908045976,
+      "grad_norm": 2.3506903297912234,
+      "kl": 0.035400390625,
+      "learning_rate": 7.436781609195402e-07,
+      "loss": 0.0015,
+      "reward": 1.8940942287445068,
+      "reward_std": 0.021925464272499084,
+      "rewards/accuracy_reward": 0.8940942287445068,
+      "rewards/format_reward": 1.0,
+      "step": 223
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.16796875,
+      "epoch": 2.574712643678161,
+      "grad_norm": 4.228763111501542,
+      "kl": 0.0302734375,
+      "learning_rate": 7.425287356321839e-07,
+      "loss": 0.0013,
+      "reward": 1.8942646980285645,
+      "reward_std": 0.02186274155974388,
+      "rewards/accuracy_reward": 0.8942646980285645,
+      "rewards/format_reward": 1.0,
+      "step": 224
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.87760925292969,
+      "epoch": 2.586206896551724,
+      "grad_norm": 1.339092605600081,
+      "kl": 0.0286865234375,
+      "learning_rate": 7.413793103448276e-07,
+      "loss": 0.0012,
+      "reward": 1.881058692932129,
+      "reward_std": 0.025702446699142456,
+      "rewards/accuracy_reward": 0.8817097544670105,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 225
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.62435150146484,
+      "epoch": 2.5977011494252875,
+      "grad_norm": 1.770916862954278,
+      "kl": 0.03125,
+      "learning_rate": 7.402298850574712e-07,
+      "loss": 0.0013,
+      "reward": 1.8780031204223633,
+      "reward_std": 0.02341182343661785,
+      "rewards/accuracy_reward": 0.8780032396316528,
+      "rewards/format_reward": 1.0,
+      "step": 226
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.64453125,
+      "epoch": 2.609195402298851,
+      "grad_norm": 1.929796218203904,
+      "kl": 0.0289306640625,
+      "learning_rate": 7.390804597701149e-07,
+      "loss": 0.0012,
+      "reward": 1.8760128021240234,
+      "reward_std": 0.025408655405044556,
+      "rewards/accuracy_reward": 0.8760129809379578,
+      "rewards/format_reward": 1.0,
+      "step": 227
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.86393737792969,
+      "epoch": 2.6206896551724137,
+      "grad_norm": 1.2206148979521507,
+      "kl": 0.0294189453125,
+      "learning_rate": 7.379310344827586e-07,
+      "loss": 0.0012,
+      "reward": 1.8749685287475586,
+      "reward_std": 0.02704436704516411,
+      "rewards/accuracy_reward": 0.8749687075614929,
+      "rewards/format_reward": 1.0,
+      "step": 228
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.52474212646484,
+      "epoch": 2.632183908045977,
+      "grad_norm": 1.1892012690175122,
+      "kl": 0.03076171875,
+      "learning_rate": 7.367816091954022e-07,
+      "loss": 0.0013,
+      "reward": 1.885100245475769,
+      "reward_std": 0.02430718205869198,
+      "rewards/accuracy_reward": 0.8851003646850586,
+      "rewards/format_reward": 1.0,
+      "step": 229
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.72786712646484,
+      "epoch": 2.6436781609195403,
+      "grad_norm": 1.474891820836295,
+      "kl": 0.03076171875,
+      "learning_rate": 7.35632183908046e-07,
+      "loss": 0.0013,
+      "reward": 1.8827028274536133,
+      "reward_std": 0.026975825428962708,
+      "rewards/accuracy_reward": 0.8827028274536133,
+      "rewards/format_reward": 1.0,
+      "step": 230
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.77083587646484,
+      "epoch": 2.655172413793103,
+      "grad_norm": 1.7269926903360489,
+      "kl": 0.0380859375,
+      "learning_rate": 7.344827586206897e-07,
+      "loss": 0.0016,
+      "reward": 1.8721890449523926,
+      "reward_std": 0.02634214609861374,
+      "rewards/accuracy_reward": 0.8721892237663269,
+      "rewards/format_reward": 1.0,
+      "step": 231
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.22526550292969,
+      "epoch": 2.6666666666666665,
+      "grad_norm": 1.9439461898517982,
+      "kl": 0.0322265625,
+      "learning_rate": 7.333333333333332e-07,
+      "loss": 0.0014,
+      "reward": 1.8815977573394775,
+      "reward_std": 0.027597349137067795,
+      "rewards/accuracy_reward": 0.8822487592697144,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 232
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.474609375,
+      "epoch": 2.67816091954023,
+      "grad_norm": 1.4550598569752655,
+      "kl": 0.03173828125,
+      "learning_rate": 7.32183908045977e-07,
+      "loss": 0.0013,
+      "reward": 1.8949522972106934,
+      "reward_std": 0.023627810180187225,
+      "rewards/accuracy_reward": 0.8949524164199829,
+      "rewards/format_reward": 1.0,
+      "step": 233
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.77214050292969,
+      "epoch": 2.689655172413793,
+      "grad_norm": 1.231833875421197,
+      "kl": 0.033447265625,
+      "learning_rate": 7.310344827586207e-07,
+      "loss": 0.0014,
+      "reward": 1.8845272064208984,
+      "reward_std": 0.02502075955271721,
+      "rewards/accuracy_reward": 0.8845272064208984,
+      "rewards/format_reward": 1.0,
+      "step": 234
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.39974212646484,
+      "epoch": 2.7011494252873565,
+      "grad_norm": 1.3652986563130975,
+      "kl": 0.035400390625,
+      "learning_rate": 7.298850574712644e-07,
+      "loss": 0.0015,
+      "reward": 1.8603239059448242,
+      "reward_std": 0.0263795405626297,
+      "rewards/accuracy_reward": 0.8603239059448242,
+      "rewards/format_reward": 1.0,
+      "step": 235
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.92057800292969,
+      "epoch": 2.7126436781609193,
+      "grad_norm": 1.6201244118842026,
+      "kl": 0.038818359375,
+      "learning_rate": 7.28735632183908e-07,
+      "loss": 0.0016,
+      "reward": 1.8918843269348145,
+      "reward_std": 0.0242290198802948,
+      "rewards/accuracy_reward": 0.891884446144104,
+      "rewards/format_reward": 1.0,
+      "step": 236
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.63346862792969,
+      "epoch": 2.7241379310344827,
+      "grad_norm": 1.6072202456697025,
+      "kl": 0.037841796875,
+      "learning_rate": 7.275862068965517e-07,
+      "loss": 0.0016,
+      "reward": 1.891068458557129,
+      "reward_std": 0.025224387645721436,
+      "rewards/accuracy_reward": 0.8917193412780762,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 237
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.85807800292969,
+      "epoch": 2.735632183908046,
+      "grad_norm": 3.1474234625781308,
+      "kl": 0.0390625,
+      "learning_rate": 7.264367816091954e-07,
+      "loss": 0.0016,
+      "reward": 1.886623740196228,
+      "reward_std": 0.024830807000398636,
+      "rewards/accuracy_reward": 0.8866235613822937,
+      "rewards/format_reward": 1.0,
+      "step": 238
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.75521087646484,
+      "epoch": 2.7471264367816093,
+      "grad_norm": 1.6342951016344707,
+      "kl": 0.0419921875,
+      "learning_rate": 7.25287356321839e-07,
+      "loss": 0.0017,
+      "reward": 1.8781700134277344,
+      "reward_std": 0.02809235453605652,
+      "rewards/accuracy_reward": 0.8794721364974976,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 239
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.28255462646484,
+      "epoch": 2.7586206896551726,
+      "grad_norm": 1.1515713105906038,
+      "kl": 0.036865234375,
+      "learning_rate": 7.241379310344827e-07,
+      "loss": 0.0015,
+      "reward": 1.89352548122406,
+      "reward_std": 0.021531865000724792,
+      "rewards/accuracy_reward": 0.8935256004333496,
+      "rewards/format_reward": 1.0,
+      "step": 240
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.35482025146484,
+      "epoch": 2.7701149425287355,
+      "grad_norm": 2.045990886967949,
+      "kl": 0.045166015625,
+      "learning_rate": 7.229885057471265e-07,
+      "loss": 0.0018,
+      "reward": 1.8543777465820312,
+      "reward_std": 0.028905829414725304,
+      "rewards/accuracy_reward": 0.8556797504425049,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 241
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.45573425292969,
+      "epoch": 2.781609195402299,
+      "grad_norm": 2.388872547648165,
+      "kl": 0.03955078125,
+      "learning_rate": 7.2183908045977e-07,
+      "loss": 0.0016,
+      "reward": 1.8795366287231445,
+      "reward_std": 0.025653596967458725,
+      "rewards/accuracy_reward": 0.8801875114440918,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 242
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.02799987792969,
+      "epoch": 2.793103448275862,
+      "grad_norm": 1.313601894213018,
+      "kl": 0.041015625,
+      "learning_rate": 7.206896551724138e-07,
+      "loss": 0.0017,
+      "reward": 1.9008476734161377,
+      "reward_std": 0.02234504744410515,
+      "rewards/accuracy_reward": 0.9008476734161377,
+      "rewards/format_reward": 1.0,
+      "step": 243
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.38802337646484,
+      "epoch": 2.8045977011494254,
+      "grad_norm": 1.690388570174564,
+      "kl": 0.039794921875,
+      "learning_rate": 7.195402298850575e-07,
+      "loss": 0.0017,
+      "reward": 1.8841087818145752,
+      "reward_std": 0.02542022429406643,
+      "rewards/accuracy_reward": 0.8854107856750488,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 244
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.02409362792969,
+      "epoch": 2.8160919540229887,
+      "grad_norm": 1.4752619364083788,
+      "kl": 0.038330078125,
+      "learning_rate": 7.18390804597701e-07,
+      "loss": 0.0016,
+      "reward": 1.872473120689392,
+      "reward_std": 0.026318645104765892,
+      "rewards/accuracy_reward": 0.8724731206893921,
+      "rewards/format_reward": 1.0,
+      "step": 245
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.24674987792969,
+      "epoch": 2.8275862068965516,
+      "grad_norm": 13.670541177628872,
+      "kl": 0.036376953125,
+      "learning_rate": 7.172413793103448e-07,
+      "loss": 0.0015,
+      "reward": 1.8723421096801758,
+      "reward_std": 0.025184884667396545,
+      "rewards/accuracy_reward": 0.8723421096801758,
+      "rewards/format_reward": 1.0,
+      "step": 246
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.20638275146484,
+      "epoch": 2.839080459770115,
+      "grad_norm": 1.0940987098177108,
+      "kl": 0.03515625,
+      "learning_rate": 7.160919540229885e-07,
+      "loss": 0.0015,
+      "reward": 1.875187873840332,
+      "reward_std": 0.02431398257613182,
+      "rewards/accuracy_reward": 0.8751880526542664,
+      "rewards/format_reward": 1.0,
+      "step": 247
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.302734375,
+      "epoch": 2.8505747126436782,
+      "grad_norm": 2.079946913833384,
+      "kl": 0.052490234375,
+      "learning_rate": 7.149425287356321e-07,
+      "loss": 0.0022,
+      "reward": 1.887062430381775,
+      "reward_std": 0.024545643478631973,
+      "rewards/accuracy_reward": 0.8870624303817749,
+      "rewards/format_reward": 1.0,
+      "step": 248
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.73503112792969,
+      "epoch": 2.862068965517241,
+      "grad_norm": 1.9116339472111843,
+      "kl": 0.033935546875,
+      "learning_rate": 7.137931034482758e-07,
+      "loss": 0.0014,
+      "reward": 1.8704020977020264,
+      "reward_std": 0.02490667998790741,
+      "rewards/accuracy_reward": 0.8710530996322632,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 249
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.78255462646484,
+      "epoch": 2.873563218390805,
+      "grad_norm": 1.198578783738147,
+      "kl": 0.035400390625,
+      "learning_rate": 7.126436781609195e-07,
+      "loss": 0.0015,
+      "reward": 1.8799991607666016,
+      "reward_std": 0.022041937336325645,
+      "rewards/accuracy_reward": 0.8799993395805359,
+      "rewards/format_reward": 1.0,
+      "step": 250
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.40755462646484,
+      "epoch": 2.8850574712643677,
+      "grad_norm": 1.8122950855537878,
+      "kl": 0.037353515625,
+      "learning_rate": 7.114942528735633e-07,
+      "loss": 0.0015,
+      "reward": 1.8816900253295898,
+      "reward_std": 0.02312072180211544,
+      "rewards/accuracy_reward": 0.8816901445388794,
+      "rewards/format_reward": 1.0,
+      "step": 251
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.2578125,
+      "epoch": 2.896551724137931,
+      "grad_norm": 1.8584979960733188,
+      "kl": 0.03369140625,
+      "learning_rate": 7.103448275862068e-07,
+      "loss": 0.0014,
+      "reward": 1.8846673965454102,
+      "reward_std": 0.025826361030340195,
+      "rewards/accuracy_reward": 0.8846673965454102,
+      "rewards/format_reward": 1.0,
+      "step": 252
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.16796875,
+      "epoch": 2.9080459770114944,
+      "grad_norm": 3.6080553707797423,
+      "kl": 0.033447265625,
+      "learning_rate": 7.091954022988506e-07,
+      "loss": 0.0014,
+      "reward": 1.8778409957885742,
+      "reward_std": 0.02433183044195175,
+      "rewards/accuracy_reward": 0.8778411746025085,
+      "rewards/format_reward": 1.0,
+      "step": 253
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.78776550292969,
+      "epoch": 2.9195402298850572,
+      "grad_norm": 1.4145733069699882,
+      "kl": 0.037109375,
+      "learning_rate": 7.080459770114943e-07,
+      "loss": 0.0015,
+      "reward": 1.881535291671753,
+      "reward_std": 0.023436537012457848,
+      "rewards/accuracy_reward": 0.8815354108810425,
+      "rewards/format_reward": 1.0,
+      "step": 254
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.01237487792969,
+      "epoch": 2.9310344827586206,
+      "grad_norm": 1.3798889630920503,
+      "kl": 0.037109375,
+      "learning_rate": 7.068965517241378e-07,
+      "loss": 0.0016,
+      "reward": 1.8669624328613281,
+      "reward_std": 0.024586662650108337,
+      "rewards/accuracy_reward": 0.8669624328613281,
+      "rewards/format_reward": 1.0,
+      "step": 255
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.84049987792969,
+      "epoch": 2.942528735632184,
+      "grad_norm": 1.501972491549555,
+      "kl": 0.033935546875,
+      "learning_rate": 7.057471264367816e-07,
+      "loss": 0.0014,
+      "reward": 1.8706506490707397,
+      "reward_std": 0.02558681182563305,
+      "rewards/accuracy_reward": 0.8706506490707397,
+      "rewards/format_reward": 1.0,
+      "step": 256
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.26692962646484,
+      "epoch": 2.954022988505747,
+      "grad_norm": 1.4790939200318616,
+      "kl": 0.037841796875,
+      "learning_rate": 7.045977011494253e-07,
+      "loss": 0.0016,
+      "reward": 1.8889020681381226,
+      "reward_std": 0.02365148440003395,
+      "rewards/accuracy_reward": 0.888901948928833,
+      "rewards/format_reward": 1.0,
+      "step": 257
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.91146087646484,
+      "epoch": 2.9655172413793105,
+      "grad_norm": 3.7173905554951854,
+      "kl": 0.0380859375,
+      "learning_rate": 7.034482758620688e-07,
+      "loss": 0.0016,
+      "reward": 1.8991615772247314,
+      "reward_std": 0.01971290074288845,
+      "rewards/accuracy_reward": 0.8991615772247314,
+      "rewards/format_reward": 1.0,
+      "step": 258
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.97917175292969,
+      "epoch": 2.9770114942528734,
+      "grad_norm": 1.2829606539449347,
+      "kl": 0.03564453125,
+      "learning_rate": 7.022988505747126e-07,
+      "loss": 0.0015,
+      "reward": 1.88856840133667,
+      "reward_std": 0.02619217149913311,
+      "rewards/accuracy_reward": 0.8898705840110779,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 259
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.66471862792969,
+      "epoch": 2.9885057471264367,
+      "grad_norm": 1.470317712158937,
+      "kl": 0.03759765625,
+      "learning_rate": 7.011494252873563e-07,
+      "loss": 0.0016,
+      "reward": 1.8722280263900757,
+      "reward_std": 0.022509008646011353,
+      "rewards/accuracy_reward": 0.8722281455993652,
+      "rewards/format_reward": 1.0,
+      "step": 260
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 96.02809143066406,
+      "epoch": 3.0,
+      "grad_norm": 3.0919770267204147,
+      "kl": 0.03466796875,
+      "learning_rate": 7e-07,
+      "loss": 0.0014,
+      "reward": 1.8606176376342773,
+      "reward_std": 0.022723043337464333,
+      "rewards/accuracy_reward": 0.8606176376342773,
+      "rewards/format_reward": 1.0,
+      "step": 261
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.25260925292969,
+      "epoch": 3.0114942528735633,
+      "grad_norm": 2.7419615747949386,
+      "kl": 0.034912109375,
+      "learning_rate": 6.988505747126436e-07,
+      "loss": 0.0014,
+      "reward": 1.869284987449646,
+      "reward_std": 0.024234648793935776,
+      "rewards/accuracy_reward": 0.8692850470542908,
+      "rewards/format_reward": 1.0,
+      "step": 262
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.68359375,
+      "epoch": 3.0229885057471266,
+      "grad_norm": 1.559372722176547,
+      "kl": 0.03271484375,
+      "learning_rate": 6.977011494252873e-07,
+      "loss": 0.0013,
+      "reward": 1.8653929233551025,
+      "reward_std": 0.02791503071784973,
+      "rewards/accuracy_reward": 0.8660439252853394,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 263
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.7109375,
+      "epoch": 3.0344827586206895,
+      "grad_norm": 2.1221245906508464,
+      "kl": 0.033203125,
+      "learning_rate": 6.96551724137931e-07,
+      "loss": 0.0014,
+      "reward": 1.8776719570159912,
+      "reward_std": 0.023673653602600098,
+      "rewards/accuracy_reward": 0.8776720762252808,
+      "rewards/format_reward": 1.0,
+      "step": 264
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.60872650146484,
+      "epoch": 3.045977011494253,
+      "grad_norm": 1.6002601391217546,
+      "kl": 0.03662109375,
+      "learning_rate": 6.954022988505746e-07,
+      "loss": 0.0015,
+      "reward": 1.830482840538025,
+      "reward_std": 0.02900915965437889,
+      "rewards/accuracy_reward": 0.8304829597473145,
+      "rewards/format_reward": 1.0,
+      "step": 265
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.30794525146484,
+      "epoch": 3.057471264367816,
+      "grad_norm": 1.6434905190427025,
+      "kl": 0.035888671875,
+      "learning_rate": 6.942528735632184e-07,
+      "loss": 0.0015,
+      "reward": 1.8308700323104858,
+      "reward_std": 0.02639869973063469,
+      "rewards/accuracy_reward": 0.8315210342407227,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 266
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.29167175292969,
+      "epoch": 3.0689655172413794,
+      "grad_norm": 1.027886396009367,
+      "kl": 0.03466796875,
+      "learning_rate": 6.931034482758621e-07,
+      "loss": 0.0014,
+      "reward": 1.8851213455200195,
+      "reward_std": 0.021387819200754166,
+      "rewards/accuracy_reward": 0.8851213455200195,
+      "rewards/format_reward": 1.0,
+      "step": 267
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.84700775146484,
+      "epoch": 3.0804597701149423,
+      "grad_norm": 1.5206328556554167,
+      "kl": 0.038818359375,
+      "learning_rate": 6.919540229885057e-07,
+      "loss": 0.0016,
+      "reward": 1.8900096416473389,
+      "reward_std": 0.02050493285059929,
+      "rewards/accuracy_reward": 0.8900095820426941,
+      "rewards/format_reward": 1.0,
+      "step": 268
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.73177337646484,
+      "epoch": 3.0919540229885056,
+      "grad_norm": 1.1641802469606335,
+      "kl": 0.034423828125,
+      "learning_rate": 6.908045977011494e-07,
+      "loss": 0.0015,
+      "reward": 1.858663558959961,
+      "reward_std": 0.025883881375193596,
+      "rewards/accuracy_reward": 0.8593146800994873,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 269
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.19140625,
+      "epoch": 3.103448275862069,
+      "grad_norm": 1.332286209838367,
+      "kl": 0.03271484375,
+      "learning_rate": 6.896551724137931e-07,
+      "loss": 0.0014,
+      "reward": 1.8819489479064941,
+      "reward_std": 0.023023171350359917,
+      "rewards/accuracy_reward": 0.8819491267204285,
+      "rewards/format_reward": 1.0,
+      "step": 270
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.453125,
+      "epoch": 3.1149425287356323,
+      "grad_norm": 1.3705118964619707,
+      "kl": 0.0341796875,
+      "learning_rate": 6.885057471264368e-07,
+      "loss": 0.0015,
+      "reward": 1.858891248703003,
+      "reward_std": 0.022298548370599747,
+      "rewards/accuracy_reward": 0.8595423698425293,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 271
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.52278900146484,
+      "epoch": 3.1264367816091956,
+      "grad_norm": 1.5377438314997252,
+      "kl": 0.041015625,
+      "learning_rate": 6.873563218390804e-07,
+      "loss": 0.0017,
+      "reward": 1.884607195854187,
+      "reward_std": 0.024785785004496574,
+      "rewards/accuracy_reward": 0.8852583169937134,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 272
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.56380462646484,
+      "epoch": 3.1379310344827585,
+      "grad_norm": 1.9722987101278098,
+      "kl": 0.03515625,
+      "learning_rate": 6.862068965517241e-07,
+      "loss": 0.0015,
+      "reward": 1.8918275833129883,
+      "reward_std": 0.021715257316827774,
+      "rewards/accuracy_reward": 0.8918277621269226,
+      "rewards/format_reward": 1.0,
+      "step": 273
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.19792175292969,
+      "epoch": 3.1494252873563218,
+      "grad_norm": 1.5898347428804778,
+      "kl": 0.03125,
+      "learning_rate": 6.850574712643678e-07,
+      "loss": 0.0013,
+      "reward": 1.8759284019470215,
+      "reward_std": 0.025104787200689316,
+      "rewards/accuracy_reward": 0.8765794634819031,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 274
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.07747650146484,
+      "epoch": 3.160919540229885,
+      "grad_norm": 1.6076437000195318,
+      "kl": 0.0380859375,
+      "learning_rate": 6.839080459770114e-07,
+      "loss": 0.0016,
+      "reward": 1.8818080425262451,
+      "reward_std": 0.02497541531920433,
+      "rewards/accuracy_reward": 0.8824591040611267,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 275
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 115.35546875,
+      "epoch": 3.1724137931034484,
+      "grad_norm": 1.7685732057058994,
+      "kl": 0.039306640625,
+      "learning_rate": 6.827586206896552e-07,
+      "loss": 0.0016,
+      "reward": 1.8782709836959839,
+      "reward_std": 0.025368181988596916,
+      "rewards/accuracy_reward": 0.8789221048355103,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 276
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 115.39128112792969,
+      "epoch": 3.1839080459770113,
+      "grad_norm": 4.02053655146932,
+      "kl": 0.04150390625,
+      "learning_rate": 6.816091954022988e-07,
+      "loss": 0.0017,
+      "reward": 1.8910026550292969,
+      "reward_std": 0.022438380867242813,
+      "rewards/accuracy_reward": 0.8910026550292969,
+      "rewards/format_reward": 1.0,
+      "step": 277
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 118.00130462646484,
+      "epoch": 3.1954022988505746,
+      "grad_norm": 1.2219545665192806,
+      "kl": 0.033935546875,
+      "learning_rate": 6.804597701149425e-07,
+      "loss": 0.0014,
+      "reward": 1.8810728788375854,
+      "reward_std": 0.027784962207078934,
+      "rewards/accuracy_reward": 0.8830260038375854,
+      "rewards/format_reward": 0.998046875,
+      "step": 278
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.451171875,
+      "epoch": 3.206896551724138,
+      "grad_norm": 1.7436462195514546,
+      "kl": 0.038818359375,
+      "learning_rate": 6.793103448275862e-07,
+      "loss": 0.0016,
+      "reward": 1.8771369457244873,
+      "reward_std": 0.023900484666228294,
+      "rewards/accuracy_reward": 0.8771368861198425,
+      "rewards/format_reward": 1.0,
+      "step": 279
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.857421875,
+      "epoch": 3.218390804597701,
+      "grad_norm": 1.564180567510325,
+      "kl": 0.0390625,
+      "learning_rate": 6.781609195402298e-07,
+      "loss": 0.0016,
+      "reward": 1.8888477087020874,
+      "reward_std": 0.02220313809812069,
+      "rewards/accuracy_reward": 0.8888477683067322,
+      "rewards/format_reward": 1.0,
+      "step": 280
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.80924987792969,
+      "epoch": 3.2298850574712645,
+      "grad_norm": 4.146116713241933,
+      "kl": 0.040771484375,
+      "learning_rate": 6.770114942528736e-07,
+      "loss": 0.0017,
+      "reward": 1.8889455795288086,
+      "reward_std": 0.021025802940130234,
+      "rewards/accuracy_reward": 0.8889455795288086,
+      "rewards/format_reward": 1.0,
+      "step": 281
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 114.03450775146484,
+      "epoch": 3.2413793103448274,
+      "grad_norm": 2.0093993818021167,
+      "kl": 0.037841796875,
+      "learning_rate": 6.758620689655172e-07,
+      "loss": 0.0016,
+      "reward": 1.8997611999511719,
+      "reward_std": 0.022101037204265594,
+      "rewards/accuracy_reward": 0.8997613191604614,
+      "rewards/format_reward": 1.0,
+      "step": 282
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.27734375,
+      "epoch": 3.2528735632183907,
+      "grad_norm": 1.7252038523721087,
+      "kl": 0.043212890625,
+      "learning_rate": 6.747126436781609e-07,
+      "loss": 0.0018,
+      "reward": 1.893257975578308,
+      "reward_std": 0.024223104119300842,
+      "rewards/accuracy_reward": 0.8945600986480713,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 283
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.23177337646484,
+      "epoch": 3.264367816091954,
+      "grad_norm": 1.3254878153740217,
+      "kl": 0.03955078125,
+      "learning_rate": 6.735632183908046e-07,
+      "loss": 0.0017,
+      "reward": 1.897439956665039,
+      "reward_std": 0.021015014499425888,
+      "rewards/accuracy_reward": 0.8974398374557495,
+      "rewards/format_reward": 1.0,
+      "step": 284
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.99609375,
+      "epoch": 3.2758620689655173,
+      "grad_norm": 1.5719828379459555,
+      "kl": 0.03955078125,
+      "learning_rate": 6.724137931034482e-07,
+      "loss": 0.0017,
+      "reward": 1.8719691038131714,
+      "reward_std": 0.022564705461263657,
+      "rewards/accuracy_reward": 0.8719691038131714,
+      "rewards/format_reward": 1.0,
+      "step": 285
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.23372650146484,
+      "epoch": 3.2873563218390807,
+      "grad_norm": 1.1121764344089415,
+      "kl": 0.04052734375,
+      "learning_rate": 6.71264367816092e-07,
+      "loss": 0.0017,
+      "reward": 1.88375985622406,
+      "reward_std": 0.022753890603780746,
+      "rewards/accuracy_reward": 0.8837599158287048,
+      "rewards/format_reward": 1.0,
+      "step": 286
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.83724212646484,
+      "epoch": 3.2988505747126435,
+      "grad_norm": 4.799405333071082,
+      "kl": 0.04296875,
+      "learning_rate": 6.701149425287356e-07,
+      "loss": 0.0018,
+      "reward": 1.8786205053329468,
+      "reward_std": 0.025325840339064598,
+      "rewards/accuracy_reward": 0.8799225687980652,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 287
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.91536712646484,
+      "epoch": 3.310344827586207,
+      "grad_norm": 2.021051884647491,
+      "kl": 0.04345703125,
+      "learning_rate": 6.689655172413793e-07,
+      "loss": 0.0018,
+      "reward": 1.8967899084091187,
+      "reward_std": 0.019771821796894073,
+      "rewards/accuracy_reward": 0.8967899084091187,
+      "rewards/format_reward": 1.0,
+      "step": 288
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.49544525146484,
+      "epoch": 3.32183908045977,
+      "grad_norm": 1.9020287418215316,
+      "kl": 0.04052734375,
+      "learning_rate": 6.67816091954023e-07,
+      "loss": 0.0017,
+      "reward": 1.8606189489364624,
+      "reward_std": 0.025568749755620956,
+      "rewards/accuracy_reward": 0.8619210720062256,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 289
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.46745300292969,
+      "epoch": 3.3333333333333335,
+      "grad_norm": 1.2275084362530455,
+      "kl": 0.041015625,
+      "learning_rate": 6.666666666666666e-07,
+      "loss": 0.0017,
+      "reward": 1.8807505369186401,
+      "reward_std": 0.02470255456864834,
+      "rewards/accuracy_reward": 0.8807506561279297,
+      "rewards/format_reward": 1.0,
+      "step": 290
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.484375,
+      "epoch": 3.344827586206897,
+      "grad_norm": 1.5137610627678855,
+      "kl": 0.036376953125,
+      "learning_rate": 6.655172413793103e-07,
+      "loss": 0.0015,
+      "reward": 1.8718066215515137,
+      "reward_std": 0.025618024170398712,
+      "rewards/accuracy_reward": 0.8718067407608032,
+      "rewards/format_reward": 1.0,
+      "step": 291
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.90234375,
+      "epoch": 3.3563218390804597,
+      "grad_norm": 3.2723373427710034,
+      "kl": 0.037109375,
+      "learning_rate": 6.64367816091954e-07,
+      "loss": 0.0016,
+      "reward": 1.8626410961151123,
+      "reward_std": 0.02414146065711975,
+      "rewards/accuracy_reward": 0.8632920980453491,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 292
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.763671875,
+      "epoch": 3.367816091954023,
+      "grad_norm": 1.4200196313502904,
+      "kl": 0.042724609375,
+      "learning_rate": 6.632183908045976e-07,
+      "loss": 0.0018,
+      "reward": 1.8814753293991089,
+      "reward_std": 0.02261793240904808,
+      "rewards/accuracy_reward": 0.8814753293991089,
+      "rewards/format_reward": 1.0,
+      "step": 293
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.11849212646484,
+      "epoch": 3.3793103448275863,
+      "grad_norm": 2.2801476286690554,
+      "kl": 0.038330078125,
+      "learning_rate": 6.620689655172414e-07,
+      "loss": 0.0016,
+      "reward": 1.8842586278915405,
+      "reward_std": 0.026064470410346985,
+      "rewards/accuracy_reward": 0.8842586874961853,
+      "rewards/format_reward": 1.0,
+      "step": 294
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.70052337646484,
+      "epoch": 3.3908045977011496,
+      "grad_norm": 1.1324912574969757,
+      "kl": 0.0390625,
+      "learning_rate": 6.609195402298851e-07,
+      "loss": 0.0016,
+      "reward": 1.8832792043685913,
+      "reward_std": 0.02494307979941368,
+      "rewards/accuracy_reward": 0.8832792043685913,
+      "rewards/format_reward": 1.0,
+      "step": 295
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.57682800292969,
+      "epoch": 3.4022988505747125,
+      "grad_norm": 1.8918486699769739,
+      "kl": 0.037841796875,
+      "learning_rate": 6.597701149425286e-07,
+      "loss": 0.0016,
+      "reward": 1.8709214925765991,
+      "reward_std": 0.02543533965945244,
+      "rewards/accuracy_reward": 0.8709214925765991,
+      "rewards/format_reward": 1.0,
+      "step": 296
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.50911712646484,
+      "epoch": 3.413793103448276,
+      "grad_norm": 3.3366246111039506,
+      "kl": 0.036865234375,
+      "learning_rate": 6.586206896551724e-07,
+      "loss": 0.0015,
+      "reward": 1.9022067785263062,
+      "reward_std": 0.02121621184051037,
+      "rewards/accuracy_reward": 0.9022065997123718,
+      "rewards/format_reward": 1.0,
+      "step": 297
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.03060150146484,
+      "epoch": 3.425287356321839,
+      "grad_norm": 1.784585488176211,
+      "kl": 0.0380859375,
+      "learning_rate": 6.574712643678161e-07,
+      "loss": 0.0016,
+      "reward": 1.8803653717041016,
+      "reward_std": 0.02356140874326229,
+      "rewards/accuracy_reward": 0.8810164332389832,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 298
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.619140625,
+      "epoch": 3.4367816091954024,
+      "grad_norm": 1.799334133513421,
+      "kl": 0.039306640625,
+      "learning_rate": 6.563218390804598e-07,
+      "loss": 0.0016,
+      "reward": 1.8940705060958862,
+      "reward_std": 0.023154854774475098,
+      "rewards/accuracy_reward": 0.8940703272819519,
+      "rewards/format_reward": 1.0,
+      "step": 299
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.240234375,
+      "epoch": 3.4482758620689653,
+      "grad_norm": 1.5600328539462516,
+      "kl": 0.037841796875,
+      "learning_rate": 6.551724137931034e-07,
+      "loss": 0.0016,
+      "reward": 1.8870646953582764,
+      "reward_std": 0.021850477904081345,
+      "rewards/accuracy_reward": 0.8870646357536316,
+      "rewards/format_reward": 1.0,
+      "step": 300
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82682800292969,
+      "epoch": 3.4597701149425286,
+      "grad_norm": 1.3183868361300648,
+      "kl": 0.041259765625,
+      "learning_rate": 6.540229885057471e-07,
+      "loss": 0.0017,
+      "reward": 1.9015138149261475,
+      "reward_std": 0.02374776266515255,
+      "rewards/accuracy_reward": 0.9015137553215027,
+      "rewards/format_reward": 1.0,
+      "step": 301
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.62109375,
+      "epoch": 3.471264367816092,
+      "grad_norm": 1.7939569716400035,
+      "kl": 0.04052734375,
+      "learning_rate": 6.528735632183908e-07,
+      "loss": 0.0017,
+      "reward": 1.8765311241149902,
+      "reward_std": 0.023926347494125366,
+      "rewards/accuracy_reward": 0.8765311241149902,
+      "rewards/format_reward": 1.0,
+      "step": 302
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.55143737792969,
+      "epoch": 3.4827586206896552,
+      "grad_norm": 1.4131833430934648,
+      "kl": 0.04150390625,
+      "learning_rate": 6.517241379310344e-07,
+      "loss": 0.0017,
+      "reward": 1.8952958583831787,
+      "reward_std": 0.021189194172620773,
+      "rewards/accuracy_reward": 0.8952958583831787,
+      "rewards/format_reward": 1.0,
+      "step": 303
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.07421875,
+      "epoch": 3.4942528735632186,
+      "grad_norm": 2.76037518653071,
+      "kl": 0.039794921875,
+      "learning_rate": 6.505747126436782e-07,
+      "loss": 0.0016,
+      "reward": 1.8874156475067139,
+      "reward_std": 0.024307802319526672,
+      "rewards/accuracy_reward": 0.8874154090881348,
+      "rewards/format_reward": 1.0,
+      "step": 304
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.955078125,
+      "epoch": 3.5057471264367814,
+      "grad_norm": 2.1138139550456376,
+      "kl": 0.04150390625,
+      "learning_rate": 6.494252873563219e-07,
+      "loss": 0.0017,
+      "reward": 1.8822648525238037,
+      "reward_std": 0.023932967334985733,
+      "rewards/accuracy_reward": 0.8822647929191589,
+      "rewards/format_reward": 1.0,
+      "step": 305
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.31575775146484,
+      "epoch": 3.5172413793103448,
+      "grad_norm": 1.2027549538193985,
+      "kl": 0.044921875,
+      "learning_rate": 6.482758620689654e-07,
+      "loss": 0.0019,
+      "reward": 1.8690357208251953,
+      "reward_std": 0.02493324503302574,
+      "rewards/accuracy_reward": 0.8696866035461426,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 306
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.220703125,
+      "epoch": 3.528735632183908,
+      "grad_norm": 3.667217381252123,
+      "kl": 0.042236328125,
+      "learning_rate": 6.471264367816092e-07,
+      "loss": 0.0018,
+      "reward": 1.9012823104858398,
+      "reward_std": 0.021280398592352867,
+      "rewards/accuracy_reward": 0.9012823104858398,
+      "rewards/format_reward": 1.0,
+      "step": 307
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.03646087646484,
+      "epoch": 3.5402298850574714,
+      "grad_norm": 1.3771503535932048,
+      "kl": 0.044189453125,
+      "learning_rate": 6.459770114942529e-07,
+      "loss": 0.0018,
+      "reward": 1.8968042135238647,
+      "reward_std": 0.0209085401147604,
+      "rewards/accuracy_reward": 0.8968044519424438,
+      "rewards/format_reward": 1.0,
+      "step": 308
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.03841400146484,
+      "epoch": 3.5517241379310347,
+      "grad_norm": 2.601418631517434,
+      "kl": 0.041748046875,
+      "learning_rate": 6.448275862068964e-07,
+      "loss": 0.0018,
+      "reward": 1.8753538131713867,
+      "reward_std": 0.024225711822509766,
+      "rewards/accuracy_reward": 0.8760050535202026,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 309
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.80403900146484,
+      "epoch": 3.5632183908045976,
+      "grad_norm": 1.5004741067737122,
+      "kl": 0.0419921875,
+      "learning_rate": 6.436781609195402e-07,
+      "loss": 0.0018,
+      "reward": 1.8968862295150757,
+      "reward_std": 0.019103027880191803,
+      "rewards/accuracy_reward": 0.8968861103057861,
+      "rewards/format_reward": 1.0,
+      "step": 310
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.68685150146484,
+      "epoch": 3.574712643678161,
+      "grad_norm": 1.788362217489612,
+      "kl": 0.040771484375,
+      "learning_rate": 6.425287356321839e-07,
+      "loss": 0.0017,
+      "reward": 1.8962510824203491,
+      "reward_std": 0.02111215889453888,
+      "rewards/accuracy_reward": 0.8962510824203491,
+      "rewards/format_reward": 1.0,
+      "step": 311
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.00260925292969,
+      "epoch": 3.586206896551724,
+      "grad_norm": 1.789359418846744,
+      "kl": 0.042724609375,
+      "learning_rate": 6.413793103448275e-07,
+      "loss": 0.0018,
+      "reward": 1.8833904266357422,
+      "reward_std": 0.019987134262919426,
+      "rewards/accuracy_reward": 0.8833904266357422,
+      "rewards/format_reward": 1.0,
+      "step": 312
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.673828125,
+      "epoch": 3.5977011494252875,
+      "grad_norm": 4.249206728495501,
+      "kl": 0.03857421875,
+      "learning_rate": 6.402298850574712e-07,
+      "loss": 0.0016,
+      "reward": 1.891036868095398,
+      "reward_std": 0.020312292501330376,
+      "rewards/accuracy_reward": 0.8916880488395691,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 313
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.59375,
+      "epoch": 3.609195402298851,
+      "grad_norm": 4.978124356714563,
+      "kl": 0.03955078125,
+      "learning_rate": 6.390804597701149e-07,
+      "loss": 0.0017,
+      "reward": 1.8778576850891113,
+      "reward_std": 0.021446645259857178,
+      "rewards/accuracy_reward": 0.8785087466239929,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 314
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.49349212646484,
+      "epoch": 3.6206896551724137,
+      "grad_norm": 1.5867170646844477,
+      "kl": 0.037353515625,
+      "learning_rate": 6.379310344827587e-07,
+      "loss": 0.0016,
+      "reward": 1.898721694946289,
+      "reward_std": 0.020030900835990906,
+      "rewards/accuracy_reward": 0.8987216949462891,
+      "rewards/format_reward": 1.0,
+      "step": 315
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.83659362792969,
+      "epoch": 3.632183908045977,
+      "grad_norm": 1.873279997651989,
+      "kl": 0.041259765625,
+      "learning_rate": 6.367816091954022e-07,
+      "loss": 0.0017,
+      "reward": 1.8622578382492065,
+      "reward_std": 0.022315729409456253,
+      "rewards/accuracy_reward": 0.8629088401794434,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 316
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.33464050292969,
+      "epoch": 3.6436781609195403,
+      "grad_norm": 3.2650242705040964,
+      "kl": 0.041748046875,
+      "learning_rate": 6.35632183908046e-07,
+      "loss": 0.0017,
+      "reward": 1.886378288269043,
+      "reward_std": 0.019871540367603302,
+      "rewards/accuracy_reward": 0.8863782286643982,
+      "rewards/format_reward": 1.0,
+      "step": 317
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.236328125,
+      "epoch": 3.655172413793103,
+      "grad_norm": 2.4548899206996175,
+      "kl": 0.038818359375,
+      "learning_rate": 6.344827586206897e-07,
+      "loss": 0.0016,
+      "reward": 1.8966128826141357,
+      "reward_std": 0.023452280089259148,
+      "rewards/accuracy_reward": 0.896612823009491,
+      "rewards/format_reward": 1.0,
+      "step": 318
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.52409362792969,
+      "epoch": 3.6666666666666665,
+      "grad_norm": 1.753082960789312,
+      "kl": 0.042236328125,
+      "learning_rate": 6.333333333333332e-07,
+      "loss": 0.0018,
+      "reward": 1.8872454166412354,
+      "reward_std": 0.02262219414114952,
+      "rewards/accuracy_reward": 0.8878964185714722,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 319
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.630859375,
+      "epoch": 3.67816091954023,
+      "grad_norm": 2.176188734436804,
+      "kl": 0.038818359375,
+      "learning_rate": 6.32183908045977e-07,
+      "loss": 0.0016,
+      "reward": 1.8845866918563843,
+      "reward_std": 0.02126806601881981,
+      "rewards/accuracy_reward": 0.8845868110656738,
+      "rewards/format_reward": 1.0,
+      "step": 320
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.69140625,
+      "epoch": 3.689655172413793,
+      "grad_norm": 1.5506807446234716,
+      "kl": 0.0390625,
+      "learning_rate": 6.310344827586207e-07,
+      "loss": 0.0016,
+      "reward": 1.8919544219970703,
+      "reward_std": 0.02176390215754509,
+      "rewards/accuracy_reward": 0.8926056623458862,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 321
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.40234375,
+      "epoch": 3.7011494252873565,
+      "grad_norm": 7.014431541777074,
+      "kl": 0.03662109375,
+      "learning_rate": 6.298850574712643e-07,
+      "loss": 0.0015,
+      "reward": 1.895882487297058,
+      "reward_std": 0.020401567220687866,
+      "rewards/accuracy_reward": 0.8958825469017029,
+      "rewards/format_reward": 1.0,
+      "step": 322
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.291015625,
+      "epoch": 3.7126436781609193,
+      "grad_norm": 2.3982992480163206,
+      "kl": 0.041259765625,
+      "learning_rate": 6.28735632183908e-07,
+      "loss": 0.0017,
+      "reward": 1.8751071691513062,
+      "reward_std": 0.02190619520843029,
+      "rewards/accuracy_reward": 0.8751071691513062,
+      "rewards/format_reward": 1.0,
+      "step": 323
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.45638275146484,
+      "epoch": 3.7241379310344827,
+      "grad_norm": 1.903714474914862,
+      "kl": 0.043212890625,
+      "learning_rate": 6.275862068965517e-07,
+      "loss": 0.0018,
+      "reward": 1.883380651473999,
+      "reward_std": 0.02156060002744198,
+      "rewards/accuracy_reward": 0.8840314745903015,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 324
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.33724212646484,
+      "epoch": 3.735632183908046,
+      "grad_norm": 3.392518970641651,
+      "kl": 0.03955078125,
+      "learning_rate": 6.264367816091954e-07,
+      "loss": 0.0017,
+      "reward": 1.8731813430786133,
+      "reward_std": 0.022301897406578064,
+      "rewards/accuracy_reward": 0.8738325834274292,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 325
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.10221862792969,
+      "epoch": 3.7471264367816093,
+      "grad_norm": 2.4944444297866886,
+      "kl": 0.0400390625,
+      "learning_rate": 6.25287356321839e-07,
+      "loss": 0.0017,
+      "reward": 1.8849986791610718,
+      "reward_std": 0.025180388242006302,
+      "rewards/accuracy_reward": 0.8863007426261902,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 326
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.51692962646484,
+      "epoch": 3.7586206896551726,
+      "grad_norm": 1.5524974855638687,
+      "kl": 0.048095703125,
+      "learning_rate": 6.241379310344828e-07,
+      "loss": 0.002,
+      "reward": 1.8775360584259033,
+      "reward_std": 0.020194731652736664,
+      "rewards/accuracy_reward": 0.8781870603561401,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 327
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.32487487792969,
+      "epoch": 3.7701149425287355,
+      "grad_norm": 1.4924502520471505,
+      "kl": 0.044189453125,
+      "learning_rate": 6.229885057471264e-07,
+      "loss": 0.0018,
+      "reward": 1.8867871761322021,
+      "reward_std": 0.02044479362666607,
+      "rewards/accuracy_reward": 0.8874382972717285,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 328
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.1875,
+      "epoch": 3.781609195402299,
+      "grad_norm": 2.570142758125839,
+      "kl": 0.04248046875,
+      "learning_rate": 6.2183908045977e-07,
+      "loss": 0.0017,
+      "reward": 1.8741064071655273,
+      "reward_std": 0.019452206790447235,
+      "rewards/accuracy_reward": 0.8741063475608826,
+      "rewards/format_reward": 1.0,
+      "step": 329
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.177734375,
+      "epoch": 3.793103448275862,
+      "grad_norm": 1.638699330431456,
+      "kl": 0.048583984375,
+      "learning_rate": 6.206896551724138e-07,
+      "loss": 0.002,
+      "reward": 1.883441686630249,
+      "reward_std": 0.021484442055225372,
+      "rewards/accuracy_reward": 0.8847436904907227,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 330
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.916015625,
+      "epoch": 3.8045977011494254,
+      "grad_norm": 1.4709865788564034,
+      "kl": 0.041015625,
+      "learning_rate": 6.195402298850575e-07,
+      "loss": 0.0017,
+      "reward": 1.8892269134521484,
+      "reward_std": 0.019907724112272263,
+      "rewards/accuracy_reward": 0.8892270922660828,
+      "rewards/format_reward": 1.0,
+      "step": 331
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.96224212646484,
+      "epoch": 3.8160919540229887,
+      "grad_norm": 3.682456578081511,
+      "kl": 0.044189453125,
+      "learning_rate": 6.183908045977011e-07,
+      "loss": 0.0018,
+      "reward": 1.8904244899749756,
+      "reward_std": 0.017542677000164986,
+      "rewards/accuracy_reward": 0.8904244303703308,
+      "rewards/format_reward": 1.0,
+      "step": 332
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.03385925292969,
+      "epoch": 3.8275862068965516,
+      "grad_norm": 1.4011409988409274,
+      "kl": 0.042724609375,
+      "learning_rate": 6.172413793103448e-07,
+      "loss": 0.0018,
+      "reward": 1.8981231451034546,
+      "reward_std": 0.019530044868588448,
+      "rewards/accuracy_reward": 0.8981232047080994,
+      "rewards/format_reward": 1.0,
+      "step": 333
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.95052337646484,
+      "epoch": 3.839080459770115,
+      "grad_norm": 2.57360289310368,
+      "kl": 0.04296875,
+      "learning_rate": 6.160919540229885e-07,
+      "loss": 0.0018,
+      "reward": 1.8822907209396362,
+      "reward_std": 0.020302042365074158,
+      "rewards/accuracy_reward": 0.882290780544281,
+      "rewards/format_reward": 1.0,
+      "step": 334
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.95964050292969,
+      "epoch": 3.8505747126436782,
+      "grad_norm": 3.9665298261234136,
+      "kl": 0.04833984375,
+      "learning_rate": 6.149425287356322e-07,
+      "loss": 0.002,
+      "reward": 1.8941223621368408,
+      "reward_std": 0.02138252556324005,
+      "rewards/accuracy_reward": 0.8947733640670776,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 335
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.89974212646484,
+      "epoch": 3.862068965517241,
+      "grad_norm": 2.706717328267551,
+      "kl": 0.046875,
+      "learning_rate": 6.137931034482758e-07,
+      "loss": 0.0019,
+      "reward": 1.8608148097991943,
+      "reward_std": 0.02197238616645336,
+      "rewards/accuracy_reward": 0.8608149886131287,
+      "rewards/format_reward": 1.0,
+      "step": 336
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.17839050292969,
+      "epoch": 3.873563218390805,
+      "grad_norm": 2.9331158810141704,
+      "kl": 0.043212890625,
+      "learning_rate": 6.126436781609195e-07,
+      "loss": 0.0018,
+      "reward": 1.877394676208496,
+      "reward_std": 0.022260598838329315,
+      "rewards/accuracy_reward": 0.8773946762084961,
+      "rewards/format_reward": 1.0,
+      "step": 337
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.91732025146484,
+      "epoch": 3.8850574712643677,
+      "grad_norm": 1.7505237848600723,
+      "kl": 0.046875,
+      "learning_rate": 6.114942528735632e-07,
+      "loss": 0.002,
+      "reward": 1.8993761539459229,
+      "reward_std": 0.01834731549024582,
+      "rewards/accuracy_reward": 0.8993761539459229,
+      "rewards/format_reward": 1.0,
+      "step": 338
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.50456237792969,
+      "epoch": 3.896551724137931,
+      "grad_norm": 1.8091603678916988,
+      "kl": 0.04541015625,
+      "learning_rate": 6.103448275862068e-07,
+      "loss": 0.0019,
+      "reward": 1.8772138357162476,
+      "reward_std": 0.02404957078397274,
+      "rewards/accuracy_reward": 0.8778649568557739,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 339
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.89453125,
+      "epoch": 3.9080459770114944,
+      "grad_norm": 2.7420144960936064,
+      "kl": 0.0439453125,
+      "learning_rate": 6.091954022988506e-07,
+      "loss": 0.0018,
+      "reward": 1.8750686645507812,
+      "reward_std": 0.02069086581468582,
+      "rewards/accuracy_reward": 0.8750687837600708,
+      "rewards/format_reward": 1.0,
+      "step": 340
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.76692962646484,
+      "epoch": 3.9195402298850572,
+      "grad_norm": 1.3320968254982721,
+      "kl": 0.042724609375,
+      "learning_rate": 6.080459770114942e-07,
+      "loss": 0.0017,
+      "reward": 1.8823789358139038,
+      "reward_std": 0.020603572949767113,
+      "rewards/accuracy_reward": 0.8823789358139038,
+      "rewards/format_reward": 1.0,
+      "step": 341
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.79232025146484,
+      "epoch": 3.9310344827586206,
+      "grad_norm": 2.6292451489584363,
+      "kl": 0.042236328125,
+      "learning_rate": 6.068965517241379e-07,
+      "loss": 0.0017,
+      "reward": 1.8959369659423828,
+      "reward_std": 0.02122236043214798,
+      "rewards/accuracy_reward": 0.8965880274772644,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 342
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.26237487792969,
+      "epoch": 3.942528735632184,
+      "grad_norm": 1.1786665172082536,
+      "kl": 0.041748046875,
+      "learning_rate": 6.057471264367816e-07,
+      "loss": 0.0017,
+      "reward": 1.8724334239959717,
+      "reward_std": 0.02036610245704651,
+      "rewards/accuracy_reward": 0.8724333643913269,
+      "rewards/format_reward": 1.0,
+      "step": 343
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.0703125,
+      "epoch": 3.954022988505747,
+      "grad_norm": 1.8258343549622353,
+      "kl": 0.041015625,
+      "learning_rate": 6.045977011494252e-07,
+      "loss": 0.0017,
+      "reward": 1.8895347118377686,
+      "reward_std": 0.021330809220671654,
+      "rewards/accuracy_reward": 0.8901857137680054,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 344
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.08984375,
+      "epoch": 3.9655172413793105,
+      "grad_norm": 1.2316728903522134,
+      "kl": 0.04150390625,
+      "learning_rate": 6.03448275862069e-07,
+      "loss": 0.0017,
+      "reward": 1.8853051662445068,
+      "reward_std": 0.020341580733656883,
+      "rewards/accuracy_reward": 0.8853051662445068,
+      "rewards/format_reward": 1.0,
+      "step": 345
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.59765625,
+      "epoch": 3.9770114942528734,
+      "grad_norm": 3.7347087879528753,
+      "kl": 0.0498046875,
+      "learning_rate": 6.022988505747126e-07,
+      "loss": 0.0021,
+      "reward": 1.8948701620101929,
+      "reward_std": 0.020036697387695312,
+      "rewards/accuracy_reward": 0.8948701620101929,
+      "rewards/format_reward": 1.0,
+      "step": 346
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.96745300292969,
+      "epoch": 3.9885057471264367,
+      "grad_norm": 1.374220142289457,
+      "kl": 0.047119140625,
+      "learning_rate": 6.011494252873563e-07,
+      "loss": 0.002,
+      "reward": 1.878260612487793,
+      "reward_std": 0.020665448158979416,
+      "rewards/accuracy_reward": 0.8782605528831482,
+      "rewards/format_reward": 1.0,
+      "step": 347
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.40168762207031,
+      "epoch": 4.0,
+      "grad_norm": 1.3796700468562093,
+      "kl": 0.0361328125,
+      "learning_rate": 6e-07,
+      "loss": 0.0015,
+      "reward": 1.8900009393692017,
+      "reward_std": 0.0169126745313406,
+      "rewards/accuracy_reward": 0.890001118183136,
+      "rewards/format_reward": 1.0,
+      "step": 348
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.21484375,
+      "epoch": 4.011494252873563,
+      "grad_norm": 1.4078584124283187,
+      "kl": 0.036376953125,
+      "learning_rate": 5.988505747126437e-07,
+      "loss": 0.0015,
+      "reward": 1.861340880393982,
+      "reward_std": 0.024126818403601646,
+      "rewards/accuracy_reward": 0.8613407611846924,
+      "rewards/format_reward": 1.0,
+      "step": 349
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.14388275146484,
+      "epoch": 4.022988505747127,
+      "grad_norm": 1.9884071163588437,
+      "kl": 0.044189453125,
+      "learning_rate": 5.977011494252874e-07,
+      "loss": 0.0019,
+      "reward": 1.8884938955307007,
+      "reward_std": 0.02110714465379715,
+      "rewards/accuracy_reward": 0.8884940147399902,
+      "rewards/format_reward": 1.0,
+      "step": 350
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.16341400146484,
+      "epoch": 4.0344827586206895,
+      "grad_norm": 1.4221350160981188,
+      "kl": 0.036376953125,
+      "learning_rate": 5.96551724137931e-07,
+      "loss": 0.0015,
+      "reward": 1.8791723251342773,
+      "reward_std": 0.022562285885214806,
+      "rewards/accuracy_reward": 0.8791724443435669,
+      "rewards/format_reward": 1.0,
+      "step": 351
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.138671875,
+      "epoch": 4.045977011494253,
+      "grad_norm": 1.654901894003328,
+      "kl": 0.038818359375,
+      "learning_rate": 5.954022988505747e-07,
+      "loss": 0.0016,
+      "reward": 1.8821697235107422,
+      "reward_std": 0.02151145413517952,
+      "rewards/accuracy_reward": 0.8821697235107422,
+      "rewards/format_reward": 1.0,
+      "step": 352
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.92839050292969,
+      "epoch": 4.057471264367816,
+      "grad_norm": 1.5963537501741232,
+      "kl": 0.039306640625,
+      "learning_rate": 5.942528735632184e-07,
+      "loss": 0.0016,
+      "reward": 1.8790719509124756,
+      "reward_std": 0.021063879132270813,
+      "rewards/accuracy_reward": 0.8797230124473572,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 353
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.041015625,
+      "epoch": 4.068965517241379,
+      "grad_norm": 1.8114870192410322,
+      "kl": 0.03759765625,
+      "learning_rate": 5.93103448275862e-07,
+      "loss": 0.0016,
+      "reward": 1.865633487701416,
+      "reward_std": 0.0278116874396801,
+      "rewards/accuracy_reward": 0.8669356107711792,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 354
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82357025146484,
+      "epoch": 4.080459770114943,
+      "grad_norm": 2.8490723955207993,
+      "kl": 0.037841796875,
+      "learning_rate": 5.919540229885057e-07,
+      "loss": 0.0016,
+      "reward": 1.8946452140808105,
+      "reward_std": 0.021742399781942368,
+      "rewards/accuracy_reward": 0.8946452140808105,
+      "rewards/format_reward": 1.0,
+      "step": 355
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.27474212646484,
+      "epoch": 4.091954022988506,
+      "grad_norm": 1.5566047273728554,
+      "kl": 0.037353515625,
+      "learning_rate": 5.908045977011494e-07,
+      "loss": 0.0016,
+      "reward": 1.875108242034912,
+      "reward_std": 0.021865393966436386,
+      "rewards/accuracy_reward": 0.8751082420349121,
+      "rewards/format_reward": 1.0,
+      "step": 356
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.48763275146484,
+      "epoch": 4.103448275862069,
+      "grad_norm": 1.9776553067288296,
+      "kl": 0.03564453125,
+      "learning_rate": 5.89655172413793e-07,
+      "loss": 0.0015,
+      "reward": 1.881319284439087,
+      "reward_std": 0.024803385138511658,
+      "rewards/accuracy_reward": 0.8813192248344421,
+      "rewards/format_reward": 1.0,
+      "step": 357
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.21940612792969,
+      "epoch": 4.114942528735632,
+      "grad_norm": 1.6859022821903558,
+      "kl": 0.0361328125,
+      "learning_rate": 5.885057471264368e-07,
+      "loss": 0.0015,
+      "reward": 1.86405348777771,
+      "reward_std": 0.023439563810825348,
+      "rewards/accuracy_reward": 0.8653554916381836,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 358
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.50521087646484,
+      "epoch": 4.126436781609195,
+      "grad_norm": 2.4612604405299474,
+      "kl": 0.03955078125,
+      "learning_rate": 5.873563218390805e-07,
+      "loss": 0.0017,
+      "reward": 1.8807547092437744,
+      "reward_std": 0.02118275687098503,
+      "rewards/accuracy_reward": 0.880754828453064,
+      "rewards/format_reward": 1.0,
+      "step": 359
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.39714050292969,
+      "epoch": 4.137931034482759,
+      "grad_norm": 1.6609596832825777,
+      "kl": 0.035888671875,
+      "learning_rate": 5.86206896551724e-07,
+      "loss": 0.0015,
+      "reward": 1.8441890478134155,
+      "reward_std": 0.023664182052016258,
+      "rewards/accuracy_reward": 0.8441890478134155,
+      "rewards/format_reward": 1.0,
+      "step": 360
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.435546875,
+      "epoch": 4.149425287356322,
+      "grad_norm": 2.5111143452832625,
+      "kl": 0.034912109375,
+      "learning_rate": 5.850574712643678e-07,
+      "loss": 0.0015,
+      "reward": 1.8899774551391602,
+      "reward_std": 0.026515571400523186,
+      "rewards/accuracy_reward": 0.8906285762786865,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 361
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.841796875,
+      "epoch": 4.160919540229885,
+      "grad_norm": 3.9427924775939784,
+      "kl": 0.037353515625,
+      "learning_rate": 5.839080459770115e-07,
+      "loss": 0.0016,
+      "reward": 1.8844678401947021,
+      "reward_std": 0.02146860770881176,
+      "rewards/accuracy_reward": 0.8844677805900574,
+      "rewards/format_reward": 1.0,
+      "step": 362
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.84765625,
+      "epoch": 4.172413793103448,
+      "grad_norm": 1.6920489480855494,
+      "kl": 0.037109375,
+      "learning_rate": 5.827586206896552e-07,
+      "loss": 0.0016,
+      "reward": 1.8880507946014404,
+      "reward_std": 0.022687291726469994,
+      "rewards/accuracy_reward": 0.8880506753921509,
+      "rewards/format_reward": 1.0,
+      "step": 363
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.06315612792969,
+      "epoch": 4.183908045977011,
+      "grad_norm": 2.152778241207781,
+      "kl": 0.037353515625,
+      "learning_rate": 5.816091954022988e-07,
+      "loss": 0.0016,
+      "reward": 1.8860154151916504,
+      "reward_std": 0.02288450300693512,
+      "rewards/accuracy_reward": 0.8860155940055847,
+      "rewards/format_reward": 1.0,
+      "step": 364
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.43034362792969,
+      "epoch": 4.195402298850575,
+      "grad_norm": 1.3564134180585476,
+      "kl": 0.037353515625,
+      "learning_rate": 5.804597701149425e-07,
+      "loss": 0.0015,
+      "reward": 1.8712117671966553,
+      "reward_std": 0.02276797592639923,
+      "rewards/accuracy_reward": 0.8712117671966553,
+      "rewards/format_reward": 1.0,
+      "step": 365
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.294921875,
+      "epoch": 4.206896551724138,
+      "grad_norm": 1.1730079534850056e+29,
+      "kl": 2.937689741663549e+26,
+      "learning_rate": 5.793103448275862e-07,
+      "loss": 1.1743651528396316e+25,
+      "reward": 1.8752689361572266,
+      "reward_std": 0.025293130427598953,
+      "rewards/accuracy_reward": 0.8765711784362793,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 366
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.33203125,
+      "epoch": 4.218390804597701,
+      "grad_norm": 1.492960779675916,
+      "kl": 0.038818359375,
+      "learning_rate": 5.781609195402298e-07,
+      "loss": 0.0016,
+      "reward": 1.8979356288909912,
+      "reward_std": 0.023821339011192322,
+      "rewards/accuracy_reward": 0.8985867500305176,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 367
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.62630462646484,
+      "epoch": 4.2298850574712645,
+      "grad_norm": 1.4743009917575458,
+      "kl": 0.040283203125,
+      "learning_rate": 5.770114942528736e-07,
+      "loss": 0.0017,
+      "reward": 1.9055975675582886,
+      "reward_std": 0.019971728324890137,
+      "rewards/accuracy_reward": 0.9055975675582886,
+      "rewards/format_reward": 1.0,
+      "step": 368
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.99674987792969,
+      "epoch": 4.241379310344827,
+      "grad_norm": 3.9302261665663982,
+      "kl": 0.0400390625,
+      "learning_rate": 5.758620689655173e-07,
+      "loss": 0.0016,
+      "reward": 1.8823680877685547,
+      "reward_std": 0.022455014288425446,
+      "rewards/accuracy_reward": 0.8823680877685547,
+      "rewards/format_reward": 1.0,
+      "step": 369
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.39583587646484,
+      "epoch": 4.252873563218391,
+      "grad_norm": 2.0938326905522238,
+      "kl": 0.0400390625,
+      "learning_rate": 5.747126436781608e-07,
+      "loss": 0.0016,
+      "reward": 1.8727083206176758,
+      "reward_std": 0.02592053823173046,
+      "rewards/accuracy_reward": 0.8733593821525574,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 370
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.81706237792969,
+      "epoch": 4.264367816091954,
+      "grad_norm": 2.888553640412068,
+      "kl": 0.040771484375,
+      "learning_rate": 5.735632183908046e-07,
+      "loss": 0.0017,
+      "reward": 1.882491111755371,
+      "reward_std": 0.020577851682901382,
+      "rewards/accuracy_reward": 0.8824911117553711,
+      "rewards/format_reward": 1.0,
+      "step": 371
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.640625,
+      "epoch": 4.275862068965517,
+      "grad_norm": 1.4686082294156166,
+      "kl": 0.051513671875,
+      "learning_rate": 5.724137931034483e-07,
+      "loss": 0.0021,
+      "reward": 1.882289171218872,
+      "reward_std": 0.023612579330801964,
+      "rewards/accuracy_reward": 0.8822891712188721,
+      "rewards/format_reward": 1.0,
+      "step": 372
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.6875,
+      "epoch": 4.287356321839081,
+      "grad_norm": 2.244464531990695,
+      "kl": 0.040283203125,
+      "learning_rate": 5.712643678160918e-07,
+      "loss": 0.0017,
+      "reward": 1.886061429977417,
+      "reward_std": 0.023760950192809105,
+      "rewards/accuracy_reward": 0.8867123126983643,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 373
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.7734375,
+      "epoch": 4.2988505747126435,
+      "grad_norm": 1.5023210240131932,
+      "kl": 0.040771484375,
+      "learning_rate": 5.701149425287356e-07,
+      "loss": 0.0017,
+      "reward": 1.8881802558898926,
+      "reward_std": 0.019625011831521988,
+      "rewards/accuracy_reward": 0.8881802558898926,
+      "rewards/format_reward": 1.0,
+      "step": 374
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.12044525146484,
+      "epoch": 4.310344827586207,
+      "grad_norm": 1.757713278971671,
+      "kl": 0.038330078125,
+      "learning_rate": 5.689655172413793e-07,
+      "loss": 0.0016,
+      "reward": 1.8772163391113281,
+      "reward_std": 0.022540029138326645,
+      "rewards/accuracy_reward": 0.8772165179252625,
+      "rewards/format_reward": 1.0,
+      "step": 375
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.61458587646484,
+      "epoch": 4.32183908045977,
+      "grad_norm": 1.6452684635685983,
+      "kl": 0.038818359375,
+      "learning_rate": 5.678160919540229e-07,
+      "loss": 0.0016,
+      "reward": 1.8954042196273804,
+      "reward_std": 0.02163463830947876,
+      "rewards/accuracy_reward": 0.8967063426971436,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 376
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.05208587646484,
+      "epoch": 4.333333333333333,
+      "grad_norm": 2.2555305608969154,
+      "kl": 0.03955078125,
+      "learning_rate": 5.666666666666666e-07,
+      "loss": 0.0016,
+      "reward": 1.8833587169647217,
+      "reward_std": 0.021704208105802536,
+      "rewards/accuracy_reward": 0.8833586573600769,
+      "rewards/format_reward": 1.0,
+      "step": 377
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.78450775146484,
+      "epoch": 4.344827586206897,
+      "grad_norm": 2.212131167395634,
+      "kl": 0.04541015625,
+      "learning_rate": 5.655172413793103e-07,
+      "loss": 0.0019,
+      "reward": 1.8870363235473633,
+      "reward_std": 0.020737424492836,
+      "rewards/accuracy_reward": 0.8870364427566528,
+      "rewards/format_reward": 1.0,
+      "step": 378
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.40104675292969,
+      "epoch": 4.35632183908046,
+      "grad_norm": 1.529590167352119,
+      "kl": 0.045654296875,
+      "learning_rate": 5.643678160919541e-07,
+      "loss": 0.0019,
+      "reward": 1.8749973773956299,
+      "reward_std": 0.018837254494428635,
+      "rewards/accuracy_reward": 0.8749972581863403,
+      "rewards/format_reward": 1.0,
+      "step": 379
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.90495300292969,
+      "epoch": 4.3678160919540225,
+      "grad_norm": 2.1938200030551385,
+      "kl": 0.04248046875,
+      "learning_rate": 5.632183908045976e-07,
+      "loss": 0.0018,
+      "reward": 1.8977670669555664,
+      "reward_std": 0.02074752375483513,
+      "rewards/accuracy_reward": 0.8977672457695007,
+      "rewards/format_reward": 1.0,
+      "step": 380
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.07096862792969,
+      "epoch": 4.379310344827586,
+      "grad_norm": 3.9533429205437813,
+      "kl": 0.047119140625,
+      "learning_rate": 5.620689655172414e-07,
+      "loss": 0.0019,
+      "reward": 1.9117012023925781,
+      "reward_std": 0.02050834149122238,
+      "rewards/accuracy_reward": 0.9117013216018677,
+      "rewards/format_reward": 1.0,
+      "step": 381
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.44140625,
+      "epoch": 4.390804597701149,
+      "grad_norm": 2.4576931014163828,
+      "kl": 0.04150390625,
+      "learning_rate": 5.609195402298851e-07,
+      "loss": 0.0017,
+      "reward": 1.8892667293548584,
+      "reward_std": 0.024918202310800552,
+      "rewards/accuracy_reward": 0.8899178504943848,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 382
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.494140625,
+      "epoch": 4.402298850574713,
+      "grad_norm": 2.1902999394095324,
+      "kl": 0.04150390625,
+      "learning_rate": 5.597701149425286e-07,
+      "loss": 0.0017,
+      "reward": 1.9016425609588623,
+      "reward_std": 0.018868938088417053,
+      "rewards/accuracy_reward": 0.9016425013542175,
+      "rewards/format_reward": 1.0,
+      "step": 383
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.78515625,
+      "epoch": 4.413793103448276,
+      "grad_norm": 1.883452563889067,
+      "kl": 0.041259765625,
+      "learning_rate": 5.586206896551724e-07,
+      "loss": 0.0017,
+      "reward": 1.8980283737182617,
+      "reward_std": 0.021328266710042953,
+      "rewards/accuracy_reward": 0.8986794352531433,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 384
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.46745300292969,
+      "epoch": 4.425287356321839,
+      "grad_norm": 2.3714294651121084,
+      "kl": 0.042724609375,
+      "learning_rate": 5.574712643678161e-07,
+      "loss": 0.0017,
+      "reward": 1.8801054954528809,
+      "reward_std": 0.02180486172437668,
+      "rewards/accuracy_reward": 0.8801054954528809,
+      "rewards/format_reward": 1.0,
+      "step": 385
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.17057800292969,
+      "epoch": 4.436781609195402,
+      "grad_norm": 1.3480188133058681,
+      "kl": 0.044189453125,
+      "learning_rate": 5.563218390804598e-07,
+      "loss": 0.0018,
+      "reward": 1.9038050174713135,
+      "reward_std": 0.023260444402694702,
+      "rewards/accuracy_reward": 0.9044560194015503,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 386
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.60807800292969,
+      "epoch": 4.448275862068965,
+      "grad_norm": 1.558932018548512,
+      "kl": 0.040283203125,
+      "learning_rate": 5.551724137931034e-07,
+      "loss": 0.0016,
+      "reward": 1.883833646774292,
+      "reward_std": 0.021813951432704926,
+      "rewards/accuracy_reward": 0.883833646774292,
+      "rewards/format_reward": 1.0,
+      "step": 387
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.23372650146484,
+      "epoch": 4.459770114942529,
+      "grad_norm": 3.4341748113797994,
+      "kl": 0.042724609375,
+      "learning_rate": 5.540229885057471e-07,
+      "loss": 0.0018,
+      "reward": 1.8973731994628906,
+      "reward_std": 0.021094098687171936,
+      "rewards/accuracy_reward": 0.8973731398582458,
+      "rewards/format_reward": 1.0,
+      "step": 388
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.3828125,
+      "epoch": 4.471264367816092,
+      "grad_norm": 2.3537727105401713,
+      "kl": 0.045654296875,
+      "learning_rate": 5.528735632183908e-07,
+      "loss": 0.0019,
+      "reward": 1.9044450521469116,
+      "reward_std": 0.02275899611413479,
+      "rewards/accuracy_reward": 0.9050959944725037,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 389
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.58528900146484,
+      "epoch": 4.482758620689655,
+      "grad_norm": 2.091387625283226,
+      "kl": 0.044921875,
+      "learning_rate": 5.517241379310344e-07,
+      "loss": 0.0019,
+      "reward": 1.8750637769699097,
+      "reward_std": 0.02193189226090908,
+      "rewards/accuracy_reward": 0.8750636577606201,
+      "rewards/format_reward": 1.0,
+      "step": 390
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.68489837646484,
+      "epoch": 4.494252873563219,
+      "grad_norm": 1.563517095069176,
+      "kl": 0.0439453125,
+      "learning_rate": 5.505747126436782e-07,
+      "loss": 0.0018,
+      "reward": 1.895681381225586,
+      "reward_std": 0.021227214485406876,
+      "rewards/accuracy_reward": 0.8956813812255859,
+      "rewards/format_reward": 1.0,
+      "step": 391
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.5703125,
+      "epoch": 4.505747126436781,
+      "grad_norm": 1.4561203507859735,
+      "kl": 0.044677734375,
+      "learning_rate": 5.494252873563218e-07,
+      "loss": 0.0018,
+      "reward": 1.903104543685913,
+      "reward_std": 0.019182192161679268,
+      "rewards/accuracy_reward": 0.9031044840812683,
+      "rewards/format_reward": 1.0,
+      "step": 392
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.791015625,
+      "epoch": 4.517241379310345,
+      "grad_norm": 2.20068200415075,
+      "kl": 0.03955078125,
+      "learning_rate": 5.482758620689654e-07,
+      "loss": 0.0016,
+      "reward": 1.8813633918762207,
+      "reward_std": 0.02065703272819519,
+      "rewards/accuracy_reward": 0.8813633918762207,
+      "rewards/format_reward": 1.0,
+      "step": 393
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.048828125,
+      "epoch": 4.528735632183908,
+      "grad_norm": 2.153835030110258,
+      "kl": 0.0400390625,
+      "learning_rate": 5.471264367816092e-07,
+      "loss": 0.0017,
+      "reward": 1.8950791358947754,
+      "reward_std": 0.018896300345659256,
+      "rewards/accuracy_reward": 0.8950790762901306,
+      "rewards/format_reward": 1.0,
+      "step": 394
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.44206237792969,
+      "epoch": 4.540229885057471,
+      "grad_norm": 1.8689205868418028,
+      "kl": 0.04931640625,
+      "learning_rate": 5.459770114942529e-07,
+      "loss": 0.002,
+      "reward": 1.8724088668823242,
+      "reward_std": 0.022876422852277756,
+      "rewards/accuracy_reward": 0.8730599880218506,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 395
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.875,
+      "epoch": 4.551724137931035,
+      "grad_norm": 1.6153380302756586,
+      "kl": 0.040283203125,
+      "learning_rate": 5.448275862068966e-07,
+      "loss": 0.0017,
+      "reward": 1.8891279697418213,
+      "reward_std": 0.022592080757021904,
+      "rewards/accuracy_reward": 0.8897790312767029,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 396
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.71614837646484,
+      "epoch": 4.563218390804598,
+      "grad_norm": 1.5101980917336142,
+      "kl": 0.039306640625,
+      "learning_rate": 5.436781609195402e-07,
+      "loss": 0.0016,
+      "reward": 1.8952821493148804,
+      "reward_std": 0.02106665074825287,
+      "rewards/accuracy_reward": 0.8952820301055908,
+      "rewards/format_reward": 1.0,
+      "step": 397
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.232421875,
+      "epoch": 4.574712643678161,
+      "grad_norm": 1.4461465189423137,
+      "kl": 0.040283203125,
+      "learning_rate": 5.425287356321839e-07,
+      "loss": 0.0017,
+      "reward": 1.8775584697723389,
+      "reward_std": 0.025320343673229218,
+      "rewards/accuracy_reward": 0.8782094717025757,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 398
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.45052337646484,
+      "epoch": 4.586206896551724,
+      "grad_norm": 1.2782259292486342,
+      "kl": 0.038818359375,
+      "learning_rate": 5.413793103448276e-07,
+      "loss": 0.0016,
+      "reward": 1.8783025741577148,
+      "reward_std": 0.024571552872657776,
+      "rewards/accuracy_reward": 0.8789536356925964,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 399
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.37435150146484,
+      "epoch": 4.597701149425287,
+      "grad_norm": 4.341249687805903,
+      "kl": 0.03955078125,
+      "learning_rate": 5.402298850574712e-07,
+      "loss": 0.0016,
+      "reward": 1.8882777690887451,
+      "reward_std": 0.025847896933555603,
+      "rewards/accuracy_reward": 0.8895797729492188,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 400
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.140625,
+      "epoch": 4.609195402298851,
+      "grad_norm": 1.3663166069622814,
+      "kl": 0.03759765625,
+      "learning_rate": 5.39080459770115e-07,
+      "loss": 0.0016,
+      "reward": 1.8888124227523804,
+      "reward_std": 0.025306126102805138,
+      "rewards/accuracy_reward": 0.8901144862174988,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 401
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.66471862792969,
+      "epoch": 4.620689655172414,
+      "grad_norm": 1.6302378119468088,
+      "kl": 0.03955078125,
+      "learning_rate": 5.379310344827586e-07,
+      "loss": 0.0017,
+      "reward": 1.8746542930603027,
+      "reward_std": 0.02441917359828949,
+      "rewards/accuracy_reward": 0.8753053545951843,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 402
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.66276550292969,
+      "epoch": 4.6321839080459775,
+      "grad_norm": 1.4816501158522215,
+      "kl": 0.039794921875,
+      "learning_rate": 5.367816091954022e-07,
+      "loss": 0.0016,
+      "reward": 1.894063949584961,
+      "reward_std": 0.024088043719530106,
+      "rewards/accuracy_reward": 0.8947149515151978,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 403
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.3828125,
+      "epoch": 4.64367816091954,
+      "grad_norm": 1.6836651779382024,
+      "kl": 0.04345703125,
+      "learning_rate": 5.35632183908046e-07,
+      "loss": 0.0018,
+      "reward": 1.8845057487487793,
+      "reward_std": 0.02305610477924347,
+      "rewards/accuracy_reward": 0.8845058679580688,
+      "rewards/format_reward": 1.0,
+      "step": 404
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.69596862792969,
+      "epoch": 4.655172413793103,
+      "grad_norm": 1.1778273851354586,
+      "kl": 0.04345703125,
+      "learning_rate": 5.344827586206896e-07,
+      "loss": 0.0018,
+      "reward": 1.8848464488983154,
+      "reward_std": 0.02404012344777584,
+      "rewards/accuracy_reward": 0.8854975700378418,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 405
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.251953125,
+      "epoch": 4.666666666666667,
+      "grad_norm": 1.427234022329123,
+      "kl": 0.043212890625,
+      "learning_rate": 5.333333333333333e-07,
+      "loss": 0.0018,
+      "reward": 1.9008078575134277,
+      "reward_std": 0.020701348781585693,
+      "rewards/accuracy_reward": 0.9008078575134277,
+      "rewards/format_reward": 1.0,
+      "step": 406
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.42513275146484,
+      "epoch": 4.67816091954023,
+      "grad_norm": 2.5640372485607696,
+      "kl": 0.04296875,
+      "learning_rate": 5.32183908045977e-07,
+      "loss": 0.0018,
+      "reward": 1.8793271780014038,
+      "reward_std": 0.026321526616811752,
+      "rewards/accuracy_reward": 0.8806290626525879,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 407
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.38932800292969,
+      "epoch": 4.689655172413794,
+      "grad_norm": 2.3519843115525343,
+      "kl": 0.041259765625,
+      "learning_rate": 5.310344827586206e-07,
+      "loss": 0.0017,
+      "reward": 1.9006179571151733,
+      "reward_std": 0.0217905193567276,
+      "rewards/accuracy_reward": 0.9006179571151733,
+      "rewards/format_reward": 1.0,
+      "step": 408
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.32161712646484,
+      "epoch": 4.7011494252873565,
+      "grad_norm": 3.886735063581474,
+      "kl": 0.0458984375,
+      "learning_rate": 5.298850574712644e-07,
+      "loss": 0.0019,
+      "reward": 1.904909372329712,
+      "reward_std": 0.019571471959352493,
+      "rewards/accuracy_reward": 0.9049093127250671,
+      "rewards/format_reward": 1.0,
+      "step": 409
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.49935150146484,
+      "epoch": 4.712643678160919,
+      "grad_norm": 1.9215943999639125,
+      "kl": 0.045654296875,
+      "learning_rate": 5.28735632183908e-07,
+      "loss": 0.0019,
+      "reward": 1.896188735961914,
+      "reward_std": 0.02258872799575329,
+      "rewards/accuracy_reward": 0.8961889147758484,
+      "rewards/format_reward": 1.0,
+      "step": 410
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.84635925292969,
+      "epoch": 4.724137931034483,
+      "grad_norm": 2.323925371294666,
+      "kl": 0.039794921875,
+      "learning_rate": 5.275862068965517e-07,
+      "loss": 0.0017,
+      "reward": 1.8925457000732422,
+      "reward_std": 0.01859492063522339,
+      "rewards/accuracy_reward": 0.8925456404685974,
+      "rewards/format_reward": 1.0,
+      "step": 411
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.20638275146484,
+      "epoch": 4.735632183908046,
+      "grad_norm": 1.655400396429134,
+      "kl": 0.041748046875,
+      "learning_rate": 5.264367816091954e-07,
+      "loss": 0.0017,
+      "reward": 1.8739862442016602,
+      "reward_std": 0.021640470251441002,
+      "rewards/accuracy_reward": 0.8746374845504761,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 412
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.99153900146484,
+      "epoch": 4.747126436781609,
+      "grad_norm": 1.9363409073368856,
+      "kl": 0.049072265625,
+      "learning_rate": 5.252873563218391e-07,
+      "loss": 0.002,
+      "reward": 1.8912228345870972,
+      "reward_std": 0.021241484209895134,
+      "rewards/accuracy_reward": 0.8918737769126892,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 413
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.73763275146484,
+      "epoch": 4.758620689655173,
+      "grad_norm": 2.279760774245835,
+      "kl": 0.04052734375,
+      "learning_rate": 5.241379310344828e-07,
+      "loss": 0.0017,
+      "reward": 1.9129436016082764,
+      "reward_std": 0.01870916411280632,
+      "rewards/accuracy_reward": 0.9129435420036316,
+      "rewards/format_reward": 1.0,
+      "step": 414
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.61003112792969,
+      "epoch": 4.7701149425287355,
+      "grad_norm": 2.45904909614863,
+      "kl": 0.03515625,
+      "learning_rate": 5.229885057471264e-07,
+      "loss": 0.0015,
+      "reward": 1.9052283763885498,
+      "reward_std": 0.019063513725996017,
+      "rewards/accuracy_reward": 0.9052283763885498,
+      "rewards/format_reward": 1.0,
+      "step": 415
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.0078125,
+      "epoch": 4.781609195402299,
+      "grad_norm": 1.6845003543176094,
+      "kl": 0.041015625,
+      "learning_rate": 5.218390804597701e-07,
+      "loss": 0.0017,
+      "reward": 1.894358515739441,
+      "reward_std": 0.022152118384838104,
+      "rewards/accuracy_reward": 0.8950096964836121,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 416
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.95833587646484,
+      "epoch": 4.793103448275862,
+      "grad_norm": 4.433200172934703,
+      "kl": 0.0634765625,
+      "learning_rate": 5.206896551724138e-07,
+      "loss": 0.0026,
+      "reward": 1.8981165885925293,
+      "reward_std": 0.020231489092111588,
+      "rewards/accuracy_reward": 0.8981165885925293,
+      "rewards/format_reward": 1.0,
+      "step": 417
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.35482025146484,
+      "epoch": 4.804597701149425,
+      "grad_norm": 1.7763571926472603,
+      "kl": 0.042724609375,
+      "learning_rate": 5.195402298850574e-07,
+      "loss": 0.0018,
+      "reward": 1.906449556350708,
+      "reward_std": 0.020989403128623962,
+      "rewards/accuracy_reward": 0.9064494967460632,
+      "rewards/format_reward": 1.0,
+      "step": 418
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.24089050292969,
+      "epoch": 4.816091954022989,
+      "grad_norm": 1.4886525030955973,
+      "kl": 0.04736328125,
+      "learning_rate": 5.183908045977012e-07,
+      "loss": 0.002,
+      "reward": 1.8850057125091553,
+      "reward_std": 0.020064577460289,
+      "rewards/accuracy_reward": 0.8850056529045105,
+      "rewards/format_reward": 1.0,
+      "step": 419
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.37565612792969,
+      "epoch": 4.827586206896552,
+      "grad_norm": 2.8859840023098866,
+      "kl": 0.0390625,
+      "learning_rate": 5.172413793103448e-07,
+      "loss": 0.0016,
+      "reward": 1.8768961429595947,
+      "reward_std": 0.018807468935847282,
+      "rewards/accuracy_reward": 0.8768962621688843,
+      "rewards/format_reward": 1.0,
+      "step": 420
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.56380462646484,
+      "epoch": 4.8390804597701145,
+      "grad_norm": 1.8199908836931964,
+      "kl": 0.04150390625,
+      "learning_rate": 5.160919540229884e-07,
+      "loss": 0.0017,
+      "reward": 1.8867862224578857,
+      "reward_std": 0.022299369797110558,
+      "rewards/accuracy_reward": 0.8867861032485962,
+      "rewards/format_reward": 1.0,
+      "step": 421
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.96484375,
+      "epoch": 4.850574712643678,
+      "grad_norm": 1.415400095170082,
+      "kl": 0.042724609375,
+      "learning_rate": 5.149425287356322e-07,
+      "loss": 0.0018,
+      "reward": 1.9118112325668335,
+      "reward_std": 0.019953366369009018,
+      "rewards/accuracy_reward": 0.9118112325668335,
+      "rewards/format_reward": 1.0,
+      "step": 422
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.650390625,
+      "epoch": 4.862068965517241,
+      "grad_norm": 2.027158566984518,
+      "kl": 0.044189453125,
+      "learning_rate": 5.137931034482759e-07,
+      "loss": 0.0018,
+      "reward": 1.878554344177246,
+      "reward_std": 0.019561052322387695,
+      "rewards/accuracy_reward": 0.8785543441772461,
+      "rewards/format_reward": 1.0,
+      "step": 423
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.13932800292969,
+      "epoch": 4.873563218390805,
+      "grad_norm": 3.123515896042613,
+      "kl": 0.048095703125,
+      "learning_rate": 5.126436781609194e-07,
+      "loss": 0.002,
+      "reward": 1.8973406553268433,
+      "reward_std": 0.021455293521285057,
+      "rewards/accuracy_reward": 0.8973406553268433,
+      "rewards/format_reward": 1.0,
+      "step": 424
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.84831237792969,
+      "epoch": 4.885057471264368,
+      "grad_norm": 2.2665060638151737,
+      "kl": 0.05078125,
+      "learning_rate": 5.114942528735632e-07,
+      "loss": 0.0021,
+      "reward": 1.8799655437469482,
+      "reward_std": 0.02300919219851494,
+      "rewards/accuracy_reward": 0.8799654841423035,
+      "rewards/format_reward": 1.0,
+      "step": 425
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.94401550292969,
+      "epoch": 4.896551724137931,
+      "grad_norm": 1.6919605234703265,
+      "kl": 0.0556640625,
+      "learning_rate": 5.103448275862069e-07,
+      "loss": 0.0023,
+      "reward": 1.9173061847686768,
+      "reward_std": 0.01950909197330475,
+      "rewards/accuracy_reward": 0.9173061847686768,
+      "rewards/format_reward": 1.0,
+      "step": 426
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.74153900146484,
+      "epoch": 4.908045977011494,
+      "grad_norm": 2.1229070642282553,
+      "kl": 0.04931640625,
+      "learning_rate": 5.091954022988506e-07,
+      "loss": 0.002,
+      "reward": 1.8742494583129883,
+      "reward_std": 0.023737162351608276,
+      "rewards/accuracy_reward": 0.8742495775222778,
+      "rewards/format_reward": 1.0,
+      "step": 427
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.42643737792969,
+      "epoch": 4.919540229885057,
+      "grad_norm": 2.1832269520490892,
+      "kl": 0.052001953125,
+      "learning_rate": 5.080459770114942e-07,
+      "loss": 0.0021,
+      "reward": 1.8955520391464233,
+      "reward_std": 0.020405521616339684,
+      "rewards/accuracy_reward": 0.8955520391464233,
+      "rewards/format_reward": 1.0,
+      "step": 428
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.27214050292969,
+      "epoch": 4.931034482758621,
+      "grad_norm": 2.536754289442291,
+      "kl": 0.0546875,
+      "learning_rate": 5.068965517241379e-07,
+      "loss": 0.0023,
+      "reward": 1.896885633468628,
+      "reward_std": 0.020467374473810196,
+      "rewards/accuracy_reward": 0.8968856334686279,
+      "rewards/format_reward": 1.0,
+      "step": 429
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.57357025146484,
+      "epoch": 4.942528735632184,
+      "grad_norm": 1.3718070549122303,
+      "kl": 0.05517578125,
+      "learning_rate": 5.057471264367817e-07,
+      "loss": 0.0022,
+      "reward": 1.8929929733276367,
+      "reward_std": 0.024712808430194855,
+      "rewards/accuracy_reward": 0.8929929733276367,
+      "rewards/format_reward": 1.0,
+      "step": 430
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.10807800292969,
+      "epoch": 4.954022988505747,
+      "grad_norm": 2.668322989932513,
+      "kl": 0.05322265625,
+      "learning_rate": 5.045977011494252e-07,
+      "loss": 0.0022,
+      "reward": 1.8981032371520996,
+      "reward_std": 0.020073171705007553,
+      "rewards/accuracy_reward": 0.8981032371520996,
+      "rewards/format_reward": 1.0,
+      "step": 431
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.62630462646484,
+      "epoch": 4.9655172413793105,
+      "grad_norm": 2.4540498394359047,
+      "kl": 0.0546875,
+      "learning_rate": 5.03448275862069e-07,
+      "loss": 0.0023,
+      "reward": 1.8993358612060547,
+      "reward_std": 0.018647989258170128,
+      "rewards/accuracy_reward": 0.8993359804153442,
+      "rewards/format_reward": 1.0,
+      "step": 432
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.70964050292969,
+      "epoch": 4.977011494252873,
+      "grad_norm": 1.963612640849598,
+      "kl": 0.052734375,
+      "learning_rate": 5.022988505747127e-07,
+      "loss": 0.0022,
+      "reward": 1.8898952007293701,
+      "reward_std": 0.022081639617681503,
+      "rewards/accuracy_reward": 0.8898952007293701,
+      "rewards/format_reward": 1.0,
+      "step": 433
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.87109375,
+      "epoch": 4.988505747126437,
+      "grad_norm": 2.3000221102493157,
+      "kl": 0.046142578125,
+      "learning_rate": 5.011494252873562e-07,
+      "loss": 0.0019,
+      "reward": 1.89603853225708,
+      "reward_std": 0.020961783826351166,
+      "rewards/accuracy_reward": 0.8960385322570801,
+      "rewards/format_reward": 1.0,
+      "step": 434
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.20084381103516,
+      "epoch": 5.0,
+      "grad_norm": 1.792728503326833,
+      "kl": 0.045166015625,
+      "learning_rate": 5e-07,
+      "loss": 0.0019,
+      "reward": 1.9164615869522095,
+      "reward_std": 0.01855144090950489,
+      "rewards/accuracy_reward": 0.9164614677429199,
+      "rewards/format_reward": 1.0,
+      "step": 435
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.03125,
+      "epoch": 5.011494252873563,
+      "grad_norm": 2.411488733190351,
+      "kl": 0.042236328125,
+      "learning_rate": 4.988505747126436e-07,
+      "loss": 0.0017,
+      "reward": 1.8886754512786865,
+      "reward_std": 0.023060960695147514,
+      "rewards/accuracy_reward": 0.8886754512786865,
+      "rewards/format_reward": 1.0,
+      "step": 436
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.51302337646484,
+      "epoch": 5.022988505747127,
+      "grad_norm": 1.2846385937268994,
+      "kl": 0.04638671875,
+      "learning_rate": 4.977011494252874e-07,
+      "loss": 0.0019,
+      "reward": 1.8596938848495483,
+      "reward_std": 0.02523483708500862,
+      "rewards/accuracy_reward": 0.8603450059890747,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 437
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.48177337646484,
+      "epoch": 5.0344827586206895,
+      "grad_norm": 4.302610321521328,
+      "kl": 0.04638671875,
+      "learning_rate": 4.96551724137931e-07,
+      "loss": 0.0019,
+      "reward": 1.875889539718628,
+      "reward_std": 0.02352536842226982,
+      "rewards/accuracy_reward": 0.8758895397186279,
+      "rewards/format_reward": 1.0,
+      "step": 438
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.92578125,
+      "epoch": 5.045977011494253,
+      "grad_norm": 1.9265386376130789,
+      "kl": 0.0390625,
+      "learning_rate": 4.954022988505746e-07,
+      "loss": 0.0016,
+      "reward": 1.8992599248886108,
+      "reward_std": 0.023157542571425438,
+      "rewards/accuracy_reward": 0.8999109268188477,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 439
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.67448425292969,
+      "epoch": 5.057471264367816,
+      "grad_norm": 1.6723222924852237,
+      "kl": 0.0439453125,
+      "learning_rate": 4.942528735632184e-07,
+      "loss": 0.0018,
+      "reward": 1.8739842176437378,
+      "reward_std": 0.022657310590147972,
+      "rewards/accuracy_reward": 0.8746352195739746,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 440
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.04167175292969,
+      "epoch": 5.068965517241379,
+      "grad_norm": 1.602166960354794,
+      "kl": 0.039306640625,
+      "learning_rate": 4.93103448275862e-07,
+      "loss": 0.0016,
+      "reward": 1.8998823165893555,
+      "reward_std": 0.02184360846877098,
+      "rewards/accuracy_reward": 0.8998822569847107,
+      "rewards/format_reward": 1.0,
+      "step": 441
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.63932800292969,
+      "epoch": 5.080459770114943,
+      "grad_norm": 2.4680045505704205,
+      "kl": 0.039306640625,
+      "learning_rate": 4.919540229885058e-07,
+      "loss": 0.0016,
+      "reward": 1.8900998830795288,
+      "reward_std": 0.02164739929139614,
+      "rewards/accuracy_reward": 0.8900998830795288,
+      "rewards/format_reward": 1.0,
+      "step": 442
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.908203125,
+      "epoch": 5.091954022988506,
+      "grad_norm": 1.390816247246552,
+      "kl": 0.041748046875,
+      "learning_rate": 4.908045977011494e-07,
+      "loss": 0.0017,
+      "reward": 1.889683723449707,
+      "reward_std": 0.022919952869415283,
+      "rewards/accuracy_reward": 0.8896838426589966,
+      "rewards/format_reward": 1.0,
+      "step": 443
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.91341400146484,
+      "epoch": 5.103448275862069,
+      "grad_norm": 1.197765074052055,
+      "kl": 0.04736328125,
+      "learning_rate": 4.89655172413793e-07,
+      "loss": 0.002,
+      "reward": 1.8937995433807373,
+      "reward_std": 0.021353360265493393,
+      "rewards/accuracy_reward": 0.8937994837760925,
+      "rewards/format_reward": 1.0,
+      "step": 444
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.26692962646484,
+      "epoch": 5.114942528735632,
+      "grad_norm": 24.16259559584872,
+      "kl": 0.0478515625,
+      "learning_rate": 4.885057471264368e-07,
+      "loss": 0.002,
+      "reward": 1.8777693510055542,
+      "reward_std": 0.019756315276026726,
+      "rewards/accuracy_reward": 0.8777693510055542,
+      "rewards/format_reward": 1.0,
+      "step": 445
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.94206237792969,
+      "epoch": 5.126436781609195,
+      "grad_norm": 2.3533060150658844,
+      "kl": 0.050537109375,
+      "learning_rate": 4.873563218390804e-07,
+      "loss": 0.0021,
+      "reward": 1.8900680541992188,
+      "reward_std": 0.021740447729825974,
+      "rewards/accuracy_reward": 0.8900680541992188,
+      "rewards/format_reward": 1.0,
+      "step": 446
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.99544525146484,
+      "epoch": 5.137931034482759,
+      "grad_norm": 2.1820025718352136,
+      "kl": 0.046142578125,
+      "learning_rate": 4.86206896551724e-07,
+      "loss": 0.0019,
+      "reward": 1.8730883598327637,
+      "reward_std": 0.022952785715460777,
+      "rewards/accuracy_reward": 0.8730884790420532,
+      "rewards/format_reward": 1.0,
+      "step": 447
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.60546875,
+      "epoch": 5.149425287356322,
+      "grad_norm": 1.6285679666257735,
+      "kl": 0.0400390625,
+      "learning_rate": 4.850574712643678e-07,
+      "loss": 0.0017,
+      "reward": 1.889723777770996,
+      "reward_std": 0.02422255277633667,
+      "rewards/accuracy_reward": 0.8897239565849304,
+      "rewards/format_reward": 1.0,
+      "step": 448
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.62760925292969,
+      "epoch": 5.160919540229885,
+      "grad_norm": 1.838531888615258,
+      "kl": 0.04150390625,
+      "learning_rate": 4.839080459770114e-07,
+      "loss": 0.0017,
+      "reward": 1.8831251859664917,
+      "reward_std": 0.023560214787721634,
+      "rewards/accuracy_reward": 0.8831251859664917,
+      "rewards/format_reward": 1.0,
+      "step": 449
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.32747650146484,
+      "epoch": 5.172413793103448,
+      "grad_norm": 3.3852429919466283,
+      "kl": 0.04296875,
+      "learning_rate": 4.827586206896552e-07,
+      "loss": 0.0018,
+      "reward": 1.8916248083114624,
+      "reward_std": 0.018707800656557083,
+      "rewards/accuracy_reward": 0.891624927520752,
+      "rewards/format_reward": 1.0,
+      "step": 450
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82747650146484,
+      "epoch": 5.183908045977011,
+      "grad_norm": 1.0585987907156273,
+      "kl": 0.040771484375,
+      "learning_rate": 4.816091954022988e-07,
+      "loss": 0.0017,
+      "reward": 1.8994290828704834,
+      "reward_std": 0.020674733445048332,
+      "rewards/accuracy_reward": 0.9000802040100098,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 451
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.90950775146484,
+      "epoch": 5.195402298850575,
+      "grad_norm": 1.5208325687196538,
+      "kl": 0.041259765625,
+      "learning_rate": 4.804597701149424e-07,
+      "loss": 0.0017,
+      "reward": 1.8975791931152344,
+      "reward_std": 0.021052822470664978,
+      "rewards/accuracy_reward": 0.898230254650116,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 452
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.69140625,
+      "epoch": 5.206896551724138,
+      "grad_norm": 1.7465767723786063,
+      "kl": 0.04541015625,
+      "learning_rate": 4.793103448275862e-07,
+      "loss": 0.0019,
+      "reward": 1.901599645614624,
+      "reward_std": 0.019404159858822823,
+      "rewards/accuracy_reward": 0.9015995860099792,
+      "rewards/format_reward": 1.0,
+      "step": 453
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.00716400146484,
+      "epoch": 5.218390804597701,
+      "grad_norm": 1.7494745095643287,
+      "kl": 0.04296875,
+      "learning_rate": 4.781609195402298e-07,
+      "loss": 0.0018,
+      "reward": 1.897741675376892,
+      "reward_std": 0.018847916275262833,
+      "rewards/accuracy_reward": 0.8977417945861816,
+      "rewards/format_reward": 1.0,
+      "step": 454
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82552337646484,
+      "epoch": 5.2298850574712645,
+      "grad_norm": 149.72760374173768,
+      "kl": 0.1123046875,
+      "learning_rate": 4.770114942528736e-07,
+      "loss": 0.0045,
+      "reward": 1.9023511409759521,
+      "reward_std": 0.02475971356034279,
+      "rewards/accuracy_reward": 0.9030022621154785,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 455
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.265625,
+      "epoch": 5.241379310344827,
+      "grad_norm": 1.5032856559021295,
+      "kl": 0.04248046875,
+      "learning_rate": 4.7586206896551725e-07,
+      "loss": 0.0018,
+      "reward": 1.8960156440734863,
+      "reward_std": 0.020089447498321533,
+      "rewards/accuracy_reward": 0.8960156440734863,
+      "rewards/format_reward": 1.0,
+      "step": 456
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.44466400146484,
+      "epoch": 5.252873563218391,
+      "grad_norm": 5.380770477839403,
+      "kl": 0.043701171875,
+      "learning_rate": 4.747126436781609e-07,
+      "loss": 0.0018,
+      "reward": 1.879178762435913,
+      "reward_std": 0.024046046659350395,
+      "rewards/accuracy_reward": 0.8798297643661499,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 457
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.21419525146484,
+      "epoch": 5.264367816091954,
+      "grad_norm": 1.307828461511747,
+      "kl": 0.039794921875,
+      "learning_rate": 4.735632183908046e-07,
+      "loss": 0.0017,
+      "reward": 1.9000663757324219,
+      "reward_std": 0.0211566723883152,
+      "rewards/accuracy_reward": 0.9000665545463562,
+      "rewards/format_reward": 1.0,
+      "step": 458
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.48828125,
+      "epoch": 5.275862068965517,
+      "grad_norm": 1.6755788743402862,
+      "kl": 0.046142578125,
+      "learning_rate": 4.7241379310344827e-07,
+      "loss": 0.0019,
+      "reward": 1.9154987335205078,
+      "reward_std": 0.019508585333824158,
+      "rewards/accuracy_reward": 0.9154989123344421,
+      "rewards/format_reward": 1.0,
+      "step": 459
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.59505462646484,
+      "epoch": 5.287356321839081,
+      "grad_norm": 1.410115167386873,
+      "kl": 0.048828125,
+      "learning_rate": 4.712643678160919e-07,
+      "loss": 0.002,
+      "reward": 1.8869271278381348,
+      "reward_std": 0.021610528230667114,
+      "rewards/accuracy_reward": 0.8875781893730164,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 460
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.11653900146484,
+      "epoch": 5.2988505747126435,
+      "grad_norm": 2.472073181988844,
+      "kl": 0.046142578125,
+      "learning_rate": 4.7011494252873565e-07,
+      "loss": 0.0019,
+      "reward": 1.8838642835617065,
+      "reward_std": 0.020128531381487846,
+      "rewards/accuracy_reward": 0.8838642835617065,
+      "rewards/format_reward": 1.0,
+      "step": 461
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.26237487792969,
+      "epoch": 5.310344827586207,
+      "grad_norm": 1.982912784034594,
+      "kl": 0.04638671875,
+      "learning_rate": 4.689655172413793e-07,
+      "loss": 0.0019,
+      "reward": 1.8954026699066162,
+      "reward_std": 0.020416123792529106,
+      "rewards/accuracy_reward": 0.8960537910461426,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 462
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.216796875,
+      "epoch": 5.32183908045977,
+      "grad_norm": 1.9630010216472946,
+      "kl": 0.048828125,
+      "learning_rate": 4.678160919540229e-07,
+      "loss": 0.002,
+      "reward": 1.8689860105514526,
+      "reward_std": 0.02052994631230831,
+      "rewards/accuracy_reward": 0.8689860105514526,
+      "rewards/format_reward": 1.0,
+      "step": 463
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.13997650146484,
+      "epoch": 5.333333333333333,
+      "grad_norm": 2.293151993990854,
+      "kl": 0.05078125,
+      "learning_rate": 4.6666666666666666e-07,
+      "loss": 0.0021,
+      "reward": 1.8916696310043335,
+      "reward_std": 0.016695309430360794,
+      "rewards/accuracy_reward": 0.8916696906089783,
+      "rewards/format_reward": 1.0,
+      "step": 464
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.533203125,
+      "epoch": 5.344827586206897,
+      "grad_norm": 2.180207259559519,
+      "kl": 0.044921875,
+      "learning_rate": 4.655172413793103e-07,
+      "loss": 0.0019,
+      "reward": 1.8959643840789795,
+      "reward_std": 0.018700793385505676,
+      "rewards/accuracy_reward": 0.8959642648696899,
+      "rewards/format_reward": 1.0,
+      "step": 465
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.427734375,
+      "epoch": 5.35632183908046,
+      "grad_norm": 1.8241902873816702,
+      "kl": 0.0498046875,
+      "learning_rate": 4.6436781609195404e-07,
+      "loss": 0.0021,
+      "reward": 1.8787130117416382,
+      "reward_std": 0.020657163113355637,
+      "rewards/accuracy_reward": 0.8793638944625854,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 466
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.609375,
+      "epoch": 5.3678160919540225,
+      "grad_norm": 2.607546823999089,
+      "kl": 0.04638671875,
+      "learning_rate": 4.632183908045977e-07,
+      "loss": 0.0019,
+      "reward": 1.8970588445663452,
+      "reward_std": 0.018911032006144524,
+      "rewards/accuracy_reward": 0.8970588445663452,
+      "rewards/format_reward": 1.0,
+      "step": 467
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.97526550292969,
+      "epoch": 5.379310344827586,
+      "grad_norm": 2.615397696144289,
+      "kl": 0.04345703125,
+      "learning_rate": 4.620689655172413e-07,
+      "loss": 0.0018,
+      "reward": 1.881915807723999,
+      "reward_std": 0.01975308358669281,
+      "rewards/accuracy_reward": 0.881915807723999,
+      "rewards/format_reward": 1.0,
+      "step": 468
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.55534362792969,
+      "epoch": 5.390804597701149,
+      "grad_norm": 1.9430369662373463,
+      "kl": 0.049560546875,
+      "learning_rate": 4.6091954022988506e-07,
+      "loss": 0.0021,
+      "reward": 1.8931167125701904,
+      "reward_std": 0.021739525720477104,
+      "rewards/accuracy_reward": 0.8937677145004272,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 469
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.79362487792969,
+      "epoch": 5.402298850574713,
+      "grad_norm": 2.0918662703233295,
+      "kl": 0.0498046875,
+      "learning_rate": 4.597701149425287e-07,
+      "loss": 0.0021,
+      "reward": 1.9036979675292969,
+      "reward_std": 0.015338393859565258,
+      "rewards/accuracy_reward": 0.9036981463432312,
+      "rewards/format_reward": 1.0,
+      "step": 470
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.98828125,
+      "epoch": 5.413793103448276,
+      "grad_norm": 1.6066027943146437,
+      "kl": 0.043701171875,
+      "learning_rate": 4.586206896551724e-07,
+      "loss": 0.0018,
+      "reward": 1.9139502048492432,
+      "reward_std": 0.01869853213429451,
+      "rewards/accuracy_reward": 0.9139501452445984,
+      "rewards/format_reward": 1.0,
+      "step": 471
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.34114837646484,
+      "epoch": 5.425287356321839,
+      "grad_norm": 2.0098215318578716,
+      "kl": 0.047607421875,
+      "learning_rate": 4.574712643678161e-07,
+      "loss": 0.002,
+      "reward": 1.8916778564453125,
+      "reward_std": 0.020163239911198616,
+      "rewards/accuracy_reward": 0.8916778564453125,
+      "rewards/format_reward": 1.0,
+      "step": 472
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.19271087646484,
+      "epoch": 5.436781609195402,
+      "grad_norm": 1.9402327167023088,
+      "kl": 0.05029296875,
+      "learning_rate": 4.563218390804597e-07,
+      "loss": 0.0021,
+      "reward": 1.8789701461791992,
+      "reward_std": 0.023091118782758713,
+      "rewards/accuracy_reward": 0.8796213865280151,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 473
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.625,
+      "epoch": 5.448275862068965,
+      "grad_norm": 1.5800586554447937,
+      "kl": 0.04443359375,
+      "learning_rate": 4.5517241379310346e-07,
+      "loss": 0.0019,
+      "reward": 1.8999593257904053,
+      "reward_std": 0.019150175154209137,
+      "rewards/accuracy_reward": 0.8999593257904053,
+      "rewards/format_reward": 1.0,
+      "step": 474
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.41732025146484,
+      "epoch": 5.459770114942529,
+      "grad_norm": 1.3572231730795086,
+      "kl": 0.048583984375,
+      "learning_rate": 4.540229885057471e-07,
+      "loss": 0.002,
+      "reward": 1.9113028049468994,
+      "reward_std": 0.018931671977043152,
+      "rewards/accuracy_reward": 0.9113027453422546,
+      "rewards/format_reward": 1.0,
+      "step": 475
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.810546875,
+      "epoch": 5.471264367816092,
+      "grad_norm": 1.4707638373352834,
+      "kl": 0.04931640625,
+      "learning_rate": 4.528735632183908e-07,
+      "loss": 0.0021,
+      "reward": 1.9164535999298096,
+      "reward_std": 0.01708863489329815,
+      "rewards/accuracy_reward": 0.91645348072052,
+      "rewards/format_reward": 1.0,
+      "step": 476
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.80013275146484,
+      "epoch": 5.482758620689655,
+      "grad_norm": 1.6157878642932322,
+      "kl": 0.0458984375,
+      "learning_rate": 4.5172413793103447e-07,
+      "loss": 0.0019,
+      "reward": 1.897383689880371,
+      "reward_std": 0.019057631492614746,
+      "rewards/accuracy_reward": 0.8973836302757263,
+      "rewards/format_reward": 1.0,
+      "step": 477
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.48372650146484,
+      "epoch": 5.494252873563219,
+      "grad_norm": 2.039464641047124,
+      "kl": 0.048828125,
+      "learning_rate": 4.505747126436781e-07,
+      "loss": 0.002,
+      "reward": 1.8875281810760498,
+      "reward_std": 0.02109697088599205,
+      "rewards/accuracy_reward": 0.8875283002853394,
+      "rewards/format_reward": 1.0,
+      "step": 478
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.49870300292969,
+      "epoch": 5.505747126436781,
+      "grad_norm": 2.039169092715578,
+      "kl": 0.046142578125,
+      "learning_rate": 4.494252873563218e-07,
+      "loss": 0.0019,
+      "reward": 1.8861839771270752,
+      "reward_std": 0.02034366875886917,
+      "rewards/accuracy_reward": 0.8861838579177856,
+      "rewards/format_reward": 1.0,
+      "step": 479
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.89453125,
+      "epoch": 5.517241379310345,
+      "grad_norm": 1.4686995511345136,
+      "kl": 0.050537109375,
+      "learning_rate": 4.482758620689655e-07,
+      "loss": 0.0021,
+      "reward": 1.9063835144042969,
+      "reward_std": 0.018992867320775986,
+      "rewards/accuracy_reward": 0.9070345163345337,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 480
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.40690612792969,
+      "epoch": 5.528735632183908,
+      "grad_norm": 2.232648278432696,
+      "kl": 0.04296875,
+      "learning_rate": 4.471264367816092e-07,
+      "loss": 0.0018,
+      "reward": 1.8770596981048584,
+      "reward_std": 0.02336341142654419,
+      "rewards/accuracy_reward": 0.8770596981048584,
+      "rewards/format_reward": 1.0,
+      "step": 481
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.29427337646484,
+      "epoch": 5.540229885057471,
+      "grad_norm": 1.9103143500415654,
+      "kl": 0.04833984375,
+      "learning_rate": 4.4597701149425287e-07,
+      "loss": 0.002,
+      "reward": 1.8889049291610718,
+      "reward_std": 0.019074462354183197,
+      "rewards/accuracy_reward": 0.8889049887657166,
+      "rewards/format_reward": 1.0,
+      "step": 482
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.26171875,
+      "epoch": 5.551724137931035,
+      "grad_norm": 1.5192873928817636,
+      "kl": 0.041015625,
+      "learning_rate": 4.4482758620689656e-07,
+      "loss": 0.0017,
+      "reward": 1.882556438446045,
+      "reward_std": 0.02097918465733528,
+      "rewards/accuracy_reward": 0.8825565576553345,
+      "rewards/format_reward": 1.0,
+      "step": 483
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.24284362792969,
+      "epoch": 5.563218390804598,
+      "grad_norm": 4.569269242461156,
+      "kl": 0.047607421875,
+      "learning_rate": 4.436781609195402e-07,
+      "loss": 0.0019,
+      "reward": 1.9005216360092163,
+      "reward_std": 0.02364097349345684,
+      "rewards/accuracy_reward": 0.9011726379394531,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 484
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.662109375,
+      "epoch": 5.574712643678161,
+      "grad_norm": 1.7930715875513572,
+      "kl": 0.045654296875,
+      "learning_rate": 4.425287356321839e-07,
+      "loss": 0.0019,
+      "reward": 1.9116904735565186,
+      "reward_std": 0.016724035143852234,
+      "rewards/accuracy_reward": 0.9116904139518738,
+      "rewards/format_reward": 1.0,
+      "step": 485
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.36003112792969,
+      "epoch": 5.586206896551724,
+      "grad_norm": 4.334056364497592,
+      "kl": 0.044921875,
+      "learning_rate": 4.413793103448276e-07,
+      "loss": 0.0018,
+      "reward": 1.895902156829834,
+      "reward_std": 0.02204793691635132,
+      "rewards/accuracy_reward": 0.8959023356437683,
+      "rewards/format_reward": 1.0,
+      "step": 486
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.138671875,
+      "epoch": 5.597701149425287,
+      "grad_norm": 1.9027704497320623,
+      "kl": 0.044921875,
+      "learning_rate": 4.402298850574712e-07,
+      "loss": 0.0019,
+      "reward": 1.8900482654571533,
+      "reward_std": 0.019576409831643105,
+      "rewards/accuracy_reward": 0.8900482654571533,
+      "rewards/format_reward": 1.0,
+      "step": 487
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.41471862792969,
+      "epoch": 5.609195402298851,
+      "grad_norm": 3.118890962601994,
+      "kl": 0.044189453125,
+      "learning_rate": 4.3908045977011495e-07,
+      "loss": 0.0018,
+      "reward": 1.890057921409607,
+      "reward_std": 0.01878949999809265,
+      "rewards/accuracy_reward": 0.8907088041305542,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 488
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.119140625,
+      "epoch": 5.620689655172414,
+      "grad_norm": 1.369384680255404,
+      "kl": 0.045166015625,
+      "learning_rate": 4.379310344827586e-07,
+      "loss": 0.0019,
+      "reward": 1.8981773853302002,
+      "reward_std": 0.020252332091331482,
+      "rewards/accuracy_reward": 0.8981773853302002,
+      "rewards/format_reward": 1.0,
+      "step": 489
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.96028900146484,
+      "epoch": 5.6321839080459775,
+      "grad_norm": 1.3682700993582457,
+      "kl": 0.04150390625,
+      "learning_rate": 4.367816091954023e-07,
+      "loss": 0.0017,
+      "reward": 1.8991467952728271,
+      "reward_std": 0.019649513065814972,
+      "rewards/accuracy_reward": 0.8991466760635376,
+      "rewards/format_reward": 1.0,
+      "step": 490
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.962890625,
+      "epoch": 5.64367816091954,
+      "grad_norm": 1.7998521186414984,
+      "kl": 0.046875,
+      "learning_rate": 4.3563218390804597e-07,
+      "loss": 0.0019,
+      "reward": 1.8821454048156738,
+      "reward_std": 0.019271574914455414,
+      "rewards/accuracy_reward": 0.8821454048156738,
+      "rewards/format_reward": 1.0,
+      "step": 491
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.05208587646484,
+      "epoch": 5.655172413793103,
+      "grad_norm": 2.074675270916482,
+      "kl": 0.046875,
+      "learning_rate": 4.344827586206896e-07,
+      "loss": 0.0019,
+      "reward": 1.9152133464813232,
+      "reward_std": 0.019003426656126976,
+      "rewards/accuracy_reward": 0.9152132868766785,
+      "rewards/format_reward": 1.0,
+      "step": 492
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.97396087646484,
+      "epoch": 5.666666666666667,
+      "grad_norm": 1.6500508152691888,
+      "kl": 0.047607421875,
+      "learning_rate": 4.3333333333333335e-07,
+      "loss": 0.002,
+      "reward": 1.9029719829559326,
+      "reward_std": 0.019596107304096222,
+      "rewards/accuracy_reward": 0.9029721021652222,
+      "rewards/format_reward": 1.0,
+      "step": 493
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.75130462646484,
+      "epoch": 5.67816091954023,
+      "grad_norm": 1.7635232706402635,
+      "kl": 0.043701171875,
+      "learning_rate": 4.32183908045977e-07,
+      "loss": 0.0018,
+      "reward": 1.8977859020233154,
+      "reward_std": 0.01936359517276287,
+      "rewards/accuracy_reward": 0.8977858424186707,
+      "rewards/format_reward": 1.0,
+      "step": 494
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.8671875,
+      "epoch": 5.689655172413794,
+      "grad_norm": 7.5630569258900815,
+      "kl": 0.049560546875,
+      "learning_rate": 4.310344827586206e-07,
+      "loss": 0.0021,
+      "reward": 1.8984394073486328,
+      "reward_std": 0.020785300061106682,
+      "rewards/accuracy_reward": 0.8984395265579224,
+      "rewards/format_reward": 1.0,
+      "step": 495
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.19075775146484,
+      "epoch": 5.7011494252873565,
+      "grad_norm": 1.9789378304791254,
+      "kl": 0.04541015625,
+      "learning_rate": 4.2988505747126437e-07,
+      "loss": 0.0019,
+      "reward": 1.8898248672485352,
+      "reward_std": 0.019473157823085785,
+      "rewards/accuracy_reward": 0.8898249864578247,
+      "rewards/format_reward": 1.0,
+      "step": 496
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.52214050292969,
+      "epoch": 5.712643678160919,
+      "grad_norm": 2.230731469046121,
+      "kl": 0.046875,
+      "learning_rate": 4.28735632183908e-07,
+      "loss": 0.0019,
+      "reward": 1.8999899625778198,
+      "reward_std": 0.017196331173181534,
+      "rewards/accuracy_reward": 0.8999900221824646,
+      "rewards/format_reward": 1.0,
+      "step": 497
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.529296875,
+      "epoch": 5.724137931034483,
+      "grad_norm": 1.8172179930249694,
+      "kl": 0.04736328125,
+      "learning_rate": 4.2758620689655174e-07,
+      "loss": 0.002,
+      "reward": 1.9146230220794678,
+      "reward_std": 0.016667529940605164,
+      "rewards/accuracy_reward": 0.9146231412887573,
+      "rewards/format_reward": 1.0,
+      "step": 498
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.14323425292969,
+      "epoch": 5.735632183908046,
+      "grad_norm": 1.7799144316476456,
+      "kl": 0.051025390625,
+      "learning_rate": 4.264367816091954e-07,
+      "loss": 0.0021,
+      "reward": 1.8982200622558594,
+      "reward_std": 0.01865963265299797,
+      "rewards/accuracy_reward": 0.8982200026512146,
+      "rewards/format_reward": 1.0,
+      "step": 499
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.78060150146484,
+      "epoch": 5.747126436781609,
+      "grad_norm": 3.2056367063069513,
+      "kl": 0.045166015625,
+      "learning_rate": 4.25287356321839e-07,
+      "loss": 0.0019,
+      "reward": 1.9110320806503296,
+      "reward_std": 0.018477408215403557,
+      "rewards/accuracy_reward": 0.9110321998596191,
+      "rewards/format_reward": 1.0,
+      "step": 500
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.09375,
+      "epoch": 5.758620689655173,
+      "grad_norm": 2.1278601395571775,
+      "kl": 0.04345703125,
+      "learning_rate": 4.2413793103448276e-07,
+      "loss": 0.0018,
+      "reward": 1.896863341331482,
+      "reward_std": 0.02090194821357727,
+      "rewards/accuracy_reward": 0.8981654047966003,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 501
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.63346862792969,
+      "epoch": 5.7701149425287355,
+      "grad_norm": 1.933803516191336,
+      "kl": 0.046875,
+      "learning_rate": 4.229885057471264e-07,
+      "loss": 0.002,
+      "reward": 1.8992946147918701,
+      "reward_std": 0.01685231737792492,
+      "rewards/accuracy_reward": 0.8992947340011597,
+      "rewards/format_reward": 1.0,
+      "step": 502
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.67903900146484,
+      "epoch": 5.781609195402299,
+      "grad_norm": 2.5798280611914204,
+      "kl": 0.049072265625,
+      "learning_rate": 4.218390804597701e-07,
+      "loss": 0.002,
+      "reward": 1.8981560468673706,
+      "reward_std": 0.017263038083910942,
+      "rewards/accuracy_reward": 0.8981560468673706,
+      "rewards/format_reward": 1.0,
+      "step": 503
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.14583587646484,
+      "epoch": 5.793103448275862,
+      "grad_norm": 2.076692138986448,
+      "kl": 0.043701171875,
+      "learning_rate": 4.206896551724138e-07,
+      "loss": 0.0018,
+      "reward": 1.8787815570831299,
+      "reward_std": 0.019007613882422447,
+      "rewards/accuracy_reward": 0.8787816762924194,
+      "rewards/format_reward": 1.0,
+      "step": 504
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.9453125,
+      "epoch": 5.804597701149425,
+      "grad_norm": 2.062105078777337,
+      "kl": 0.049072265625,
+      "learning_rate": 4.195402298850574e-07,
+      "loss": 0.002,
+      "reward": 1.8786977529525757,
+      "reward_std": 0.018628563731908798,
+      "rewards/accuracy_reward": 0.8786976337432861,
+      "rewards/format_reward": 1.0,
+      "step": 505
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.208984375,
+      "epoch": 5.816091954022989,
+      "grad_norm": 3.4179007868834175,
+      "kl": 0.04443359375,
+      "learning_rate": 4.1839080459770116e-07,
+      "loss": 0.0019,
+      "reward": 1.9090993404388428,
+      "reward_std": 0.018269993364810944,
+      "rewards/accuracy_reward": 0.9090994596481323,
+      "rewards/format_reward": 1.0,
+      "step": 506
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.859375,
+      "epoch": 5.827586206896552,
+      "grad_norm": 1.642778882446284,
+      "kl": 0.04541015625,
+      "learning_rate": 4.172413793103448e-07,
+      "loss": 0.0019,
+      "reward": 1.895350694656372,
+      "reward_std": 0.017278993502259254,
+      "rewards/accuracy_reward": 0.8960016965866089,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 507
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.00521087646484,
+      "epoch": 5.8390804597701145,
+      "grad_norm": 2.6806085379649702,
+      "kl": 0.049072265625,
+      "learning_rate": 4.160919540229885e-07,
+      "loss": 0.002,
+      "reward": 1.9006847143173218,
+      "reward_std": 0.016903840005397797,
+      "rewards/accuracy_reward": 0.9006847143173218,
+      "rewards/format_reward": 1.0,
+      "step": 508
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.69466400146484,
+      "epoch": 5.850574712643678,
+      "grad_norm": 3.6854562093571386,
+      "kl": 0.05859375,
+      "learning_rate": 4.149425287356322e-07,
+      "loss": 0.0024,
+      "reward": 1.908639907836914,
+      "reward_std": 0.016876667737960815,
+      "rewards/accuracy_reward": 0.9086400866508484,
+      "rewards/format_reward": 1.0,
+      "step": 509
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.81120300292969,
+      "epoch": 5.862068965517241,
+      "grad_norm": 3.453501517404399,
+      "kl": 0.0517578125,
+      "learning_rate": 4.1379310344827586e-07,
+      "loss": 0.0022,
+      "reward": 1.903754472732544,
+      "reward_std": 0.019029080867767334,
+      "rewards/accuracy_reward": 0.9044055342674255,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 510
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.5859375,
+      "epoch": 5.873563218390805,
+      "grad_norm": 1.803202256070274,
+      "kl": 0.054443359375,
+      "learning_rate": 4.126436781609195e-07,
+      "loss": 0.0022,
+      "reward": 1.8829306364059448,
+      "reward_std": 0.01749059185385704,
+      "rewards/accuracy_reward": 0.8829306364059448,
+      "rewards/format_reward": 1.0,
+      "step": 511
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.27995300292969,
+      "epoch": 5.885057471264368,
+      "grad_norm": 1.5556838404555604,
+      "kl": 0.044677734375,
+      "learning_rate": 4.114942528735632e-07,
+      "loss": 0.0019,
+      "reward": 1.9034370183944702,
+      "reward_std": 0.018126491457223892,
+      "rewards/accuracy_reward": 0.903437077999115,
+      "rewards/format_reward": 1.0,
+      "step": 512
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.17057800292969,
+      "epoch": 5.896551724137931,
+      "grad_norm": 3.4957318575488383,
+      "kl": 0.0498046875,
+      "learning_rate": 4.103448275862069e-07,
+      "loss": 0.002,
+      "reward": 1.9079878330230713,
+      "reward_std": 0.01815355196595192,
+      "rewards/accuracy_reward": 0.9079879522323608,
+      "rewards/format_reward": 1.0,
+      "step": 513
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.59765625,
+      "epoch": 5.908045977011494,
+      "grad_norm": 2.707726805902382,
+      "kl": 0.044921875,
+      "learning_rate": 4.0919540229885057e-07,
+      "loss": 0.0019,
+      "reward": 1.884354591369629,
+      "reward_std": 0.018658233806490898,
+      "rewards/accuracy_reward": 0.8843547105789185,
+      "rewards/format_reward": 1.0,
+      "step": 514
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.71354675292969,
+      "epoch": 5.919540229885057,
+      "grad_norm": 1.4965768623048503,
+      "kl": 0.04931640625,
+      "learning_rate": 4.0804597701149426e-07,
+      "loss": 0.0021,
+      "reward": 1.8936374187469482,
+      "reward_std": 0.01746196486055851,
+      "rewards/accuracy_reward": 0.8936373591423035,
+      "rewards/format_reward": 1.0,
+      "step": 515
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.97982025146484,
+      "epoch": 5.931034482758621,
+      "grad_norm": 2.9479475900202115,
+      "kl": 0.046630859375,
+      "learning_rate": 4.068965517241379e-07,
+      "loss": 0.0019,
+      "reward": 1.902561902999878,
+      "reward_std": 0.01716381497681141,
+      "rewards/accuracy_reward": 0.9025619029998779,
+      "rewards/format_reward": 1.0,
+      "step": 516
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.25456237792969,
+      "epoch": 5.942528735632184,
+      "grad_norm": 2.5970229292314677,
+      "kl": 0.05029296875,
+      "learning_rate": 4.057471264367816e-07,
+      "loss": 0.0021,
+      "reward": 1.8918704986572266,
+      "reward_std": 0.019317103549838066,
+      "rewards/accuracy_reward": 0.8918704986572266,
+      "rewards/format_reward": 1.0,
+      "step": 517
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.76888275146484,
+      "epoch": 5.954022988505747,
+      "grad_norm": 1.9125407656929412,
+      "kl": 0.045654296875,
+      "learning_rate": 4.045977011494253e-07,
+      "loss": 0.0019,
+      "reward": 1.89088773727417,
+      "reward_std": 0.0201788991689682,
+      "rewards/accuracy_reward": 0.8915388584136963,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 518
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.34375,
+      "epoch": 5.9655172413793105,
+      "grad_norm": 1.906615415914849,
+      "kl": 0.05322265625,
+      "learning_rate": 4.034482758620689e-07,
+      "loss": 0.0022,
+      "reward": 1.8955434560775757,
+      "reward_std": 0.017683900892734528,
+      "rewards/accuracy_reward": 0.8955433368682861,
+      "rewards/format_reward": 1.0,
+      "step": 519
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.95573425292969,
+      "epoch": 5.977011494252873,
+      "grad_norm": 1.930378197428578,
+      "kl": 0.052734375,
+      "learning_rate": 4.0229885057471266e-07,
+      "loss": 0.0021,
+      "reward": 1.9113073348999023,
+      "reward_std": 0.015919553115963936,
+      "rewards/accuracy_reward": 0.9113074541091919,
+      "rewards/format_reward": 1.0,
+      "step": 520
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.345703125,
+      "epoch": 5.988505747126437,
+      "grad_norm": 2.559459936562671,
+      "kl": 0.048828125,
+      "learning_rate": 4.011494252873563e-07,
+      "loss": 0.0021,
+      "reward": 1.903090476989746,
+      "reward_std": 0.017953019589185715,
+      "rewards/accuracy_reward": 0.9030904769897461,
+      "rewards/format_reward": 1.0,
+      "step": 521
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 101.29354095458984,
+      "epoch": 6.0,
+      "grad_norm": 1.320684949997662,
+      "kl": 0.043212890625,
+      "learning_rate": 4e-07,
+      "loss": 0.0018,
+      "reward": 1.9057987928390503,
+      "reward_std": 0.014725634828209877,
+      "rewards/accuracy_reward": 0.9057987332344055,
+      "rewards/format_reward": 1.0,
+      "step": 522
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.42317962646484,
+      "epoch": 6.011494252873563,
+      "grad_norm": 1.785801287990398,
+      "kl": 0.046875,
+      "learning_rate": 3.9885057471264367e-07,
+      "loss": 0.0019,
+      "reward": 1.8730857372283936,
+      "reward_std": 0.016687028110027313,
+      "rewards/accuracy_reward": 0.8730856776237488,
+      "rewards/format_reward": 1.0,
+      "step": 523
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.64778900146484,
+      "epoch": 6.022988505747127,
+      "grad_norm": 2.818216992605221,
+      "kl": 0.05126953125,
+      "learning_rate": 3.977011494252873e-07,
+      "loss": 0.0021,
+      "reward": 1.8802497386932373,
+      "reward_std": 0.019351143389940262,
+      "rewards/accuracy_reward": 0.8802497386932373,
+      "rewards/format_reward": 1.0,
+      "step": 524
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.970703125,
+      "epoch": 6.0344827586206895,
+      "grad_norm": 1.373324008322693,
+      "kl": 0.04833984375,
+      "learning_rate": 3.9655172413793105e-07,
+      "loss": 0.002,
+      "reward": 1.8902562856674194,
+      "reward_std": 0.017598673701286316,
+      "rewards/accuracy_reward": 0.8902562856674194,
+      "rewards/format_reward": 1.0,
+      "step": 525
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.61263275146484,
+      "epoch": 6.045977011494253,
+      "grad_norm": 1.4203897246386383,
+      "kl": 0.046630859375,
+      "learning_rate": 3.954022988505747e-07,
+      "loss": 0.002,
+      "reward": 1.896735668182373,
+      "reward_std": 0.0163632333278656,
+      "rewards/accuracy_reward": 0.896735668182373,
+      "rewards/format_reward": 1.0,
+      "step": 526
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.630859375,
+      "epoch": 6.057471264367816,
+      "grad_norm": 1.9485819155646742,
+      "kl": 0.047607421875,
+      "learning_rate": 3.942528735632183e-07,
+      "loss": 0.002,
+      "reward": 1.8787517547607422,
+      "reward_std": 0.020119938999414444,
+      "rewards/accuracy_reward": 0.879402756690979,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 527
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.02214050292969,
+      "epoch": 6.068965517241379,
+      "grad_norm": 2.495312428574511,
+      "kl": 0.050537109375,
+      "learning_rate": 3.9310344827586207e-07,
+      "loss": 0.0021,
+      "reward": 1.9046882390975952,
+      "reward_std": 0.01929212175309658,
+      "rewards/accuracy_reward": 0.9046882390975952,
+      "rewards/format_reward": 1.0,
+      "step": 528
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.466796875,
+      "epoch": 6.080459770114943,
+      "grad_norm": 1.8085811940047551,
+      "kl": 0.048583984375,
+      "learning_rate": 3.919540229885057e-07,
+      "loss": 0.002,
+      "reward": 1.8941335678100586,
+      "reward_std": 0.0205199234187603,
+      "rewards/accuracy_reward": 0.894133448600769,
+      "rewards/format_reward": 1.0,
+      "step": 529
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.52018737792969,
+      "epoch": 6.091954022988506,
+      "grad_norm": 5.365446513796388,
+      "kl": 0.04638671875,
+      "learning_rate": 3.9080459770114945e-07,
+      "loss": 0.0019,
+      "reward": 1.9010367393493652,
+      "reward_std": 0.017713043838739395,
+      "rewards/accuracy_reward": 0.9010368585586548,
+      "rewards/format_reward": 1.0,
+      "step": 530
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.74089050292969,
+      "epoch": 6.103448275862069,
+      "grad_norm": 2.828191865732173,
+      "kl": 0.049072265625,
+      "learning_rate": 3.896551724137931e-07,
+      "loss": 0.002,
+      "reward": 1.8785228729248047,
+      "reward_std": 0.02002035826444626,
+      "rewards/accuracy_reward": 0.8785228133201599,
+      "rewards/format_reward": 1.0,
+      "step": 531
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.07747650146484,
+      "epoch": 6.114942528735632,
+      "grad_norm": 4.047622387425111,
+      "kl": 0.04443359375,
+      "learning_rate": 3.885057471264367e-07,
+      "loss": 0.0018,
+      "reward": 1.8791890144348145,
+      "reward_std": 0.01751401089131832,
+      "rewards/accuracy_reward": 0.8791891932487488,
+      "rewards/format_reward": 1.0,
+      "step": 532
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.36263275146484,
+      "epoch": 6.126436781609195,
+      "grad_norm": 1.3128782551208291,
+      "kl": 0.044189453125,
+      "learning_rate": 3.8735632183908046e-07,
+      "loss": 0.0019,
+      "reward": 1.9140952825546265,
+      "reward_std": 0.01683102920651436,
+      "rewards/accuracy_reward": 0.9140952825546265,
+      "rewards/format_reward": 1.0,
+      "step": 533
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.18815612792969,
+      "epoch": 6.137931034482759,
+      "grad_norm": 1.933372124731939,
+      "kl": 0.038818359375,
+      "learning_rate": 3.862068965517241e-07,
+      "loss": 0.0016,
+      "reward": 1.8693971633911133,
+      "reward_std": 0.016525980085134506,
+      "rewards/accuracy_reward": 0.8693971633911133,
+      "rewards/format_reward": 1.0,
+      "step": 534
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.75521087646484,
+      "epoch": 6.149425287356322,
+      "grad_norm": 1.6333597141209881,
+      "kl": 0.04150390625,
+      "learning_rate": 3.850574712643678e-07,
+      "loss": 0.0018,
+      "reward": 1.90386962890625,
+      "reward_std": 0.015211975201964378,
+      "rewards/accuracy_reward": 0.9038697481155396,
+      "rewards/format_reward": 1.0,
+      "step": 535
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.96549987792969,
+      "epoch": 6.160919540229885,
+      "grad_norm": 1.2717871465323818,
+      "kl": 0.0419921875,
+      "learning_rate": 3.839080459770115e-07,
+      "loss": 0.0018,
+      "reward": 1.8973326683044434,
+      "reward_std": 0.01976935565471649,
+      "rewards/accuracy_reward": 0.8979837894439697,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 536
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.41341400146484,
+      "epoch": 6.172413793103448,
+      "grad_norm": 2.481211895104965,
+      "kl": 0.04248046875,
+      "learning_rate": 3.827586206896551e-07,
+      "loss": 0.0018,
+      "reward": 1.9014673233032227,
+      "reward_std": 0.018813788890838623,
+      "rewards/accuracy_reward": 0.9014673233032227,
+      "rewards/format_reward": 1.0,
+      "step": 537
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.07292175292969,
+      "epoch": 6.183908045977011,
+      "grad_norm": 1.7579797956745593,
+      "kl": 0.03759765625,
+      "learning_rate": 3.8160919540229886e-07,
+      "loss": 0.0016,
+      "reward": 1.8949092626571655,
+      "reward_std": 0.018055422231554985,
+      "rewards/accuracy_reward": 0.8955603837966919,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 538
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.69140625,
+      "epoch": 6.195402298850575,
+      "grad_norm": 2.0716783964182257,
+      "kl": 0.045166015625,
+      "learning_rate": 3.804597701149425e-07,
+      "loss": 0.0019,
+      "reward": 1.8800392150878906,
+      "reward_std": 0.02071218751370907,
+      "rewards/accuracy_reward": 0.8806902766227722,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 539
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82942962646484,
+      "epoch": 6.206896551724138,
+      "grad_norm": 3.1245967017230347,
+      "kl": 0.0390625,
+      "learning_rate": 3.793103448275862e-07,
+      "loss": 0.0016,
+      "reward": 1.8988111019134521,
+      "reward_std": 0.01829328015446663,
+      "rewards/accuracy_reward": 0.8988110423088074,
+      "rewards/format_reward": 1.0,
+      "step": 540
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.06901550292969,
+      "epoch": 6.218390804597701,
+      "grad_norm": 4.050364995136617,
+      "kl": 0.05517578125,
+      "learning_rate": 3.781609195402299e-07,
+      "loss": 0.0023,
+      "reward": 1.8841352462768555,
+      "reward_std": 0.022105498239398003,
+      "rewards/accuracy_reward": 0.8847863674163818,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 541
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82747650146484,
+      "epoch": 6.2298850574712645,
+      "grad_norm": 1.8078253085183804,
+      "kl": 0.038330078125,
+      "learning_rate": 3.7701149425287357e-07,
+      "loss": 0.0016,
+      "reward": 1.910055160522461,
+      "reward_std": 0.021085239946842194,
+      "rewards/accuracy_reward": 0.9107062816619873,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 542
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.75325775146484,
+      "epoch": 6.241379310344827,
+      "grad_norm": 1.5902924468702986,
+      "kl": 0.048583984375,
+      "learning_rate": 3.758620689655172e-07,
+      "loss": 0.002,
+      "reward": 1.8845319747924805,
+      "reward_std": 0.022094160318374634,
+      "rewards/accuracy_reward": 0.8845321536064148,
+      "rewards/format_reward": 1.0,
+      "step": 543
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.82682800292969,
+      "epoch": 6.252873563218391,
+      "grad_norm": 2.0704586256752253,
+      "kl": 0.042236328125,
+      "learning_rate": 3.747126436781609e-07,
+      "loss": 0.0018,
+      "reward": 1.8901569843292236,
+      "reward_std": 0.020656492561101913,
+      "rewards/accuracy_reward": 0.8901569247245789,
+      "rewards/format_reward": 1.0,
+      "step": 544
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.267578125,
+      "epoch": 6.264367816091954,
+      "grad_norm": 2.4508205899480724,
+      "kl": 0.04248046875,
+      "learning_rate": 3.735632183908046e-07,
+      "loss": 0.0018,
+      "reward": 1.8985986709594727,
+      "reward_std": 0.021605072543025017,
+      "rewards/accuracy_reward": 0.8992499113082886,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 545
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.693359375,
+      "epoch": 6.275862068965517,
+      "grad_norm": 1.5807641157109527,
+      "kl": 0.0439453125,
+      "learning_rate": 3.7241379310344827e-07,
+      "loss": 0.0018,
+      "reward": 1.903763771057129,
+      "reward_std": 0.020712215453386307,
+      "rewards/accuracy_reward": 0.9037638902664185,
+      "rewards/format_reward": 1.0,
+      "step": 546
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.677734375,
+      "epoch": 6.287356321839081,
+      "grad_norm": 1.8174266396212175,
+      "kl": 0.043212890625,
+      "learning_rate": 3.7126436781609196e-07,
+      "loss": 0.0018,
+      "reward": 1.8896994590759277,
+      "reward_std": 0.02042810432612896,
+      "rewards/accuracy_reward": 0.8896996378898621,
+      "rewards/format_reward": 1.0,
+      "step": 547
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.779296875,
+      "epoch": 6.2988505747126435,
+      "grad_norm": 1.3923115045110266,
+      "kl": 0.03857421875,
+      "learning_rate": 3.701149425287356e-07,
+      "loss": 0.0016,
+      "reward": 1.8963736295700073,
+      "reward_std": 0.021433323621749878,
+      "rewards/accuracy_reward": 0.8963736295700073,
+      "rewards/format_reward": 1.0,
+      "step": 548
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.72526550292969,
+      "epoch": 6.310344827586207,
+      "grad_norm": 1.7072054434768456,
+      "kl": 0.041748046875,
+      "learning_rate": 3.689655172413793e-07,
+      "loss": 0.0017,
+      "reward": 1.91050124168396,
+      "reward_std": 0.02200714498758316,
+      "rewards/accuracy_reward": 0.9111522436141968,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 549
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.98567962646484,
+      "epoch": 6.32183908045977,
+      "grad_norm": 1.3415596828035086,
+      "kl": 0.0390625,
+      "learning_rate": 3.67816091954023e-07,
+      "loss": 0.0017,
+      "reward": 1.9038093090057373,
+      "reward_std": 0.01953584887087345,
+      "rewards/accuracy_reward": 0.9044603109359741,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 550
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.85482025146484,
+      "epoch": 6.333333333333333,
+      "grad_norm": 1.2624977576240533,
+      "kl": 0.039306640625,
+      "learning_rate": 3.666666666666666e-07,
+      "loss": 0.0017,
+      "reward": 1.915178656578064,
+      "reward_std": 0.017109807580709457,
+      "rewards/accuracy_reward": 0.9151787161827087,
+      "rewards/format_reward": 1.0,
+      "step": 551
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.31185150146484,
+      "epoch": 6.344827586206897,
+      "grad_norm": 1.7459891843430122,
+      "kl": 0.0400390625,
+      "learning_rate": 3.6551724137931036e-07,
+      "loss": 0.0017,
+      "reward": 1.8864727020263672,
+      "reward_std": 0.02323821559548378,
+      "rewards/accuracy_reward": 0.8871237635612488,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 552
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.73893737792969,
+      "epoch": 6.35632183908046,
+      "grad_norm": 2.027513640363097,
+      "kl": 0.042724609375,
+      "learning_rate": 3.64367816091954e-07,
+      "loss": 0.0018,
+      "reward": 1.9012677669525146,
+      "reward_std": 0.01830562949180603,
+      "rewards/accuracy_reward": 0.9012677073478699,
+      "rewards/format_reward": 1.0,
+      "step": 553
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.47200775146484,
+      "epoch": 6.3678160919540225,
+      "grad_norm": 3.2725385424225064,
+      "kl": 0.0419921875,
+      "learning_rate": 3.632183908045977e-07,
+      "loss": 0.0017,
+      "reward": 1.9009044170379639,
+      "reward_std": 0.019121093675494194,
+      "rewards/accuracy_reward": 0.9009043574333191,
+      "rewards/format_reward": 1.0,
+      "step": 554
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.43034362792969,
+      "epoch": 6.379310344827586,
+      "grad_norm": 2.1062592964548665,
+      "kl": 0.0419921875,
+      "learning_rate": 3.620689655172414e-07,
+      "loss": 0.0017,
+      "reward": 1.9104199409484863,
+      "reward_std": 0.020279204472899437,
+      "rewards/accuracy_reward": 0.9110710024833679,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 555
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.58919525146484,
+      "epoch": 6.390804597701149,
+      "grad_norm": 1.4371380643185012,
+      "kl": 0.04150390625,
+      "learning_rate": 3.60919540229885e-07,
+      "loss": 0.0017,
+      "reward": 1.9064548015594482,
+      "reward_std": 0.019678324460983276,
+      "rewards/accuracy_reward": 0.9077568054199219,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 556
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.373046875,
+      "epoch": 6.402298850574713,
+      "grad_norm": 1.4862995937435215,
+      "kl": 0.041259765625,
+      "learning_rate": 3.5977011494252875e-07,
+      "loss": 0.0017,
+      "reward": 1.9054267406463623,
+      "reward_std": 0.02058161422610283,
+      "rewards/accuracy_reward": 0.9054265022277832,
+      "rewards/format_reward": 1.0,
+      "step": 557
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.744140625,
+      "epoch": 6.413793103448276,
+      "grad_norm": 1.4955532746736115,
+      "kl": 0.04052734375,
+      "learning_rate": 3.586206896551724e-07,
+      "loss": 0.0017,
+      "reward": 1.9132366180419922,
+      "reward_std": 0.0166233628988266,
+      "rewards/accuracy_reward": 0.9132366180419922,
+      "rewards/format_reward": 1.0,
+      "step": 558
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.3828125,
+      "epoch": 6.425287356321839,
+      "grad_norm": 1.5190626491993644,
+      "kl": 0.038818359375,
+      "learning_rate": 3.5747126436781603e-07,
+      "loss": 0.0016,
+      "reward": 1.9011564254760742,
+      "reward_std": 0.018908878788352013,
+      "rewards/accuracy_reward": 0.9018074870109558,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 559
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.10612487792969,
+      "epoch": 6.436781609195402,
+      "grad_norm": 2.3464151581290253,
+      "kl": 0.041259765625,
+      "learning_rate": 3.5632183908045977e-07,
+      "loss": 0.0017,
+      "reward": 1.8979761600494385,
+      "reward_std": 0.019787931814789772,
+      "rewards/accuracy_reward": 0.8986272811889648,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 560
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.68685150146484,
+      "epoch": 6.448275862068965,
+      "grad_norm": 5.8754331087951925,
+      "kl": 0.04443359375,
+      "learning_rate": 3.551724137931034e-07,
+      "loss": 0.0018,
+      "reward": 1.9075603485107422,
+      "reward_std": 0.018795132637023926,
+      "rewards/accuracy_reward": 0.9075603485107422,
+      "rewards/format_reward": 1.0,
+      "step": 561
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.45833587646484,
+      "epoch": 6.459770114942529,
+      "grad_norm": 1.995883232305232,
+      "kl": 0.04296875,
+      "learning_rate": 3.5402298850574715e-07,
+      "loss": 0.0018,
+      "reward": 1.9002068042755127,
+      "reward_std": 0.019703200086951256,
+      "rewards/accuracy_reward": 0.9002067446708679,
+      "rewards/format_reward": 1.0,
+      "step": 562
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.5234375,
+      "epoch": 6.471264367816092,
+      "grad_norm": 2.589878830540706,
+      "kl": 0.044189453125,
+      "learning_rate": 3.528735632183908e-07,
+      "loss": 0.0018,
+      "reward": 1.898942470550537,
+      "reward_std": 0.023017754778265953,
+      "rewards/accuracy_reward": 0.9008955955505371,
+      "rewards/format_reward": 0.998046875,
+      "step": 563
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.181640625,
+      "epoch": 6.482758620689655,
+      "grad_norm": 2.327789501169785,
+      "kl": 0.044677734375,
+      "learning_rate": 3.517241379310344e-07,
+      "loss": 0.0019,
+      "reward": 1.9264135360717773,
+      "reward_std": 0.018227433785796165,
+      "rewards/accuracy_reward": 0.9264135360717773,
+      "rewards/format_reward": 1.0,
+      "step": 564
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.37435150146484,
+      "epoch": 6.494252873563219,
+      "grad_norm": 3.3012610405965925,
+      "kl": 0.044677734375,
+      "learning_rate": 3.5057471264367817e-07,
+      "loss": 0.0019,
+      "reward": 1.882051944732666,
+      "reward_std": 0.020881911739706993,
+      "rewards/accuracy_reward": 0.882051944732666,
+      "rewards/format_reward": 1.0,
+      "step": 565
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.64388275146484,
+      "epoch": 6.505747126436781,
+      "grad_norm": 1.5695423368611925,
+      "kl": 0.039306640625,
+      "learning_rate": 3.494252873563218e-07,
+      "loss": 0.0016,
+      "reward": 1.9002180099487305,
+      "reward_std": 0.016297608613967896,
+      "rewards/accuracy_reward": 0.9002181887626648,
+      "rewards/format_reward": 1.0,
+      "step": 566
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.037109375,
+      "epoch": 6.517241379310345,
+      "grad_norm": 1.3087146389425561,
+      "kl": 0.04248046875,
+      "learning_rate": 3.482758620689655e-07,
+      "loss": 0.0018,
+      "reward": 1.9085086584091187,
+      "reward_std": 0.017730899155139923,
+      "rewards/accuracy_reward": 0.9085086584091187,
+      "rewards/format_reward": 1.0,
+      "step": 567
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.88607025146484,
+      "epoch": 6.528735632183908,
+      "grad_norm": 2.971298257475111,
+      "kl": 0.046142578125,
+      "learning_rate": 3.471264367816092e-07,
+      "loss": 0.0019,
+      "reward": 1.8897184133529663,
+      "reward_std": 0.022931650280952454,
+      "rewards/accuracy_reward": 0.8910205364227295,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 568
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.04362487792969,
+      "epoch": 6.540229885057471,
+      "grad_norm": 2.2629176129306514,
+      "kl": 0.04833984375,
+      "learning_rate": 3.4597701149425287e-07,
+      "loss": 0.002,
+      "reward": 1.888806700706482,
+      "reward_std": 0.022193504497408867,
+      "rewards/accuracy_reward": 0.8901089429855347,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 569
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.546875,
+      "epoch": 6.551724137931035,
+      "grad_norm": 1.6844269731795065,
+      "kl": 0.0458984375,
+      "learning_rate": 3.4482758620689656e-07,
+      "loss": 0.0019,
+      "reward": 1.8828990459442139,
+      "reward_std": 0.018883569166064262,
+      "rewards/accuracy_reward": 0.8835498690605164,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 570
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.8671875,
+      "epoch": 6.563218390804598,
+      "grad_norm": 2.284457618723532,
+      "kl": 0.044921875,
+      "learning_rate": 3.436781609195402e-07,
+      "loss": 0.0018,
+      "reward": 1.884222388267517,
+      "reward_std": 0.018331531435251236,
+      "rewards/accuracy_reward": 0.8848733305931091,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 571
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.43099212646484,
+      "epoch": 6.574712643678161,
+      "grad_norm": 1.446359862116147,
+      "kl": 0.04345703125,
+      "learning_rate": 3.425287356321839e-07,
+      "loss": 0.0018,
+      "reward": 1.9134483337402344,
+      "reward_std": 0.019138170406222343,
+      "rewards/accuracy_reward": 0.9134482741355896,
+      "rewards/format_reward": 1.0,
+      "step": 572
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.00130462646484,
+      "epoch": 6.586206896551724,
+      "grad_norm": 5.703380489491454,
+      "kl": 0.05078125,
+      "learning_rate": 3.413793103448276e-07,
+      "loss": 0.0021,
+      "reward": 1.8881677389144897,
+      "reward_std": 0.018784310668706894,
+      "rewards/accuracy_reward": 0.8881677389144897,
+      "rewards/format_reward": 1.0,
+      "step": 573
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.427734375,
+      "epoch": 6.597701149425287,
+      "grad_norm": 3.5020194713310895,
+      "kl": 0.04345703125,
+      "learning_rate": 3.4022988505747127e-07,
+      "loss": 0.0018,
+      "reward": 1.8986886739730835,
+      "reward_std": 0.01682961732149124,
+      "rewards/accuracy_reward": 0.8986887335777283,
+      "rewards/format_reward": 1.0,
+      "step": 574
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.43424987792969,
+      "epoch": 6.609195402298851,
+      "grad_norm": 2.303566130374915,
+      "kl": 0.04296875,
+      "learning_rate": 3.390804597701149e-07,
+      "loss": 0.0018,
+      "reward": 1.8962440490722656,
+      "reward_std": 0.02106454037129879,
+      "rewards/accuracy_reward": 0.8975462913513184,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 575
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.39583587646484,
+      "epoch": 6.620689655172414,
+      "grad_norm": 1.6584764881239733,
+      "kl": 0.045166015625,
+      "learning_rate": 3.379310344827586e-07,
+      "loss": 0.0019,
+      "reward": 1.88081693649292,
+      "reward_std": 0.018141578882932663,
+      "rewards/accuracy_reward": 0.8808168768882751,
+      "rewards/format_reward": 1.0,
+      "step": 576
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.30143737792969,
+      "epoch": 6.6321839080459775,
+      "grad_norm": 4.350037902113492,
+      "kl": 0.048828125,
+      "learning_rate": 3.367816091954023e-07,
+      "loss": 0.002,
+      "reward": 1.8807071447372437,
+      "reward_std": 0.019819162786006927,
+      "rewards/accuracy_reward": 0.8807072043418884,
+      "rewards/format_reward": 1.0,
+      "step": 577
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.64453125,
+      "epoch": 6.64367816091954,
+      "grad_norm": 1.8183763915247184,
+      "kl": 0.04736328125,
+      "learning_rate": 3.35632183908046e-07,
+      "loss": 0.002,
+      "reward": 1.8958330154418945,
+      "reward_std": 0.017030756920576096,
+      "rewards/accuracy_reward": 0.8958329558372498,
+      "rewards/format_reward": 1.0,
+      "step": 578
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.712890625,
+      "epoch": 6.655172413793103,
+      "grad_norm": 1.6800458122792343,
+      "kl": 0.045166015625,
+      "learning_rate": 3.3448275862068966e-07,
+      "loss": 0.0019,
+      "reward": 1.899322748184204,
+      "reward_std": 0.01855292171239853,
+      "rewards/accuracy_reward": 0.8993227481842041,
+      "rewards/format_reward": 1.0,
+      "step": 579
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.41862487792969,
+      "epoch": 6.666666666666667,
+      "grad_norm": 1.5389866031610586,
+      "kl": 0.048828125,
+      "learning_rate": 3.333333333333333e-07,
+      "loss": 0.002,
+      "reward": 1.9041564464569092,
+      "reward_std": 0.019357208162546158,
+      "rewards/accuracy_reward": 0.9041563868522644,
+      "rewards/format_reward": 1.0,
+      "step": 580
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.85482025146484,
+      "epoch": 6.67816091954023,
+      "grad_norm": 1.854004391136165,
+      "kl": 0.046630859375,
+      "learning_rate": 3.32183908045977e-07,
+      "loss": 0.0019,
+      "reward": 1.896866798400879,
+      "reward_std": 0.02253865823149681,
+      "rewards/accuracy_reward": 0.8981690406799316,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 581
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.93359375,
+      "epoch": 6.689655172413794,
+      "grad_norm": 2.018641316913995,
+      "kl": 0.048828125,
+      "learning_rate": 3.310344827586207e-07,
+      "loss": 0.002,
+      "reward": 1.9180282354354858,
+      "reward_std": 0.01753108948469162,
+      "rewards/accuracy_reward": 0.9180280566215515,
+      "rewards/format_reward": 1.0,
+      "step": 582
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.14192962646484,
+      "epoch": 6.7011494252873565,
+      "grad_norm": 2.590727908860308,
+      "kl": 0.047119140625,
+      "learning_rate": 3.298850574712643e-07,
+      "loss": 0.002,
+      "reward": 1.8914936780929565,
+      "reward_std": 0.019188500940799713,
+      "rewards/accuracy_reward": 0.8921446800231934,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 583
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.40625,
+      "epoch": 6.712643678160919,
+      "grad_norm": 1.4866576141042993,
+      "kl": 0.05029296875,
+      "learning_rate": 3.2873563218390806e-07,
+      "loss": 0.0021,
+      "reward": 1.900843620300293,
+      "reward_std": 0.02220628783106804,
+      "rewards/accuracy_reward": 0.9014946818351746,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 584
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.1640625,
+      "epoch": 6.724137931034483,
+      "grad_norm": 1.7834180698321564,
+      "kl": 0.049560546875,
+      "learning_rate": 3.275862068965517e-07,
+      "loss": 0.002,
+      "reward": 1.8919929265975952,
+      "reward_std": 0.019489863887429237,
+      "rewards/accuracy_reward": 0.8919928073883057,
+      "rewards/format_reward": 1.0,
+      "step": 585
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.11263275146484,
+      "epoch": 6.735632183908046,
+      "grad_norm": 2.080295738757342,
+      "kl": 0.0517578125,
+      "learning_rate": 3.264367816091954e-07,
+      "loss": 0.0021,
+      "reward": 1.9068288803100586,
+      "reward_std": 0.01798483356833458,
+      "rewards/accuracy_reward": 0.9068288803100586,
+      "rewards/format_reward": 1.0,
+      "step": 586
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.28515625,
+      "epoch": 6.747126436781609,
+      "grad_norm": 1.363971268029249,
+      "kl": 0.05322265625,
+      "learning_rate": 3.252873563218391e-07,
+      "loss": 0.0022,
+      "reward": 1.896977186203003,
+      "reward_std": 0.01630130037665367,
+      "rewards/accuracy_reward": 0.8969771265983582,
+      "rewards/format_reward": 1.0,
+      "step": 587
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.81120300292969,
+      "epoch": 6.758620689655173,
+      "grad_norm": 1.8998231610338354,
+      "kl": 0.04541015625,
+      "learning_rate": 3.241379310344827e-07,
+      "loss": 0.0019,
+      "reward": 1.8929922580718994,
+      "reward_std": 0.018269825726747513,
+      "rewards/accuracy_reward": 0.8929921984672546,
+      "rewards/format_reward": 1.0,
+      "step": 588
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.96810150146484,
+      "epoch": 6.7701149425287355,
+      "grad_norm": 1.4654423135239003,
+      "kl": 0.055908203125,
+      "learning_rate": 3.2298850574712646e-07,
+      "loss": 0.0023,
+      "reward": 1.9026551246643066,
+      "reward_std": 0.018198613077402115,
+      "rewards/accuracy_reward": 0.9026551246643066,
+      "rewards/format_reward": 1.0,
+      "step": 589
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.50521087646484,
+      "epoch": 6.781609195402299,
+      "grad_norm": 1.720473724542019,
+      "kl": 0.05029296875,
+      "learning_rate": 3.218390804597701e-07,
+      "loss": 0.0021,
+      "reward": 1.896925687789917,
+      "reward_std": 0.016040321439504623,
+      "rewards/accuracy_reward": 0.8969255685806274,
+      "rewards/format_reward": 1.0,
+      "step": 590
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.015625,
+      "epoch": 6.793103448275862,
+      "grad_norm": 1.8599065609753198,
+      "kl": 0.046875,
+      "learning_rate": 3.2068965517241373e-07,
+      "loss": 0.002,
+      "reward": 1.896074652671814,
+      "reward_std": 0.021062329411506653,
+      "rewards/accuracy_reward": 0.898027777671814,
+      "rewards/format_reward": 0.998046875,
+      "step": 591
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.74870300292969,
+      "epoch": 6.804597701149425,
+      "grad_norm": 2.454880208710065,
+      "kl": 0.04541015625,
+      "learning_rate": 3.1954022988505747e-07,
+      "loss": 0.0019,
+      "reward": 1.9107999801635742,
+      "reward_std": 0.015166133642196655,
+      "rewards/accuracy_reward": 0.9108001589775085,
+      "rewards/format_reward": 1.0,
+      "step": 592
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.056640625,
+      "epoch": 6.816091954022989,
+      "grad_norm": 1.7036855487787057,
+      "kl": 0.045166015625,
+      "learning_rate": 3.183908045977011e-07,
+      "loss": 0.0019,
+      "reward": 1.8942689895629883,
+      "reward_std": 0.02249247394502163,
+      "rewards/accuracy_reward": 0.8949202299118042,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 593
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.830078125,
+      "epoch": 6.827586206896552,
+      "grad_norm": 53.62639568378942,
+      "kl": 0.0458984375,
+      "learning_rate": 3.1724137931034485e-07,
+      "loss": 0.0019,
+      "reward": 1.8942097425460815,
+      "reward_std": 0.019950736314058304,
+      "rewards/accuracy_reward": 0.8948608636856079,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 594
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.10286712646484,
+      "epoch": 6.8390804597701145,
+      "grad_norm": 5.944879218709257,
+      "kl": 0.040283203125,
+      "learning_rate": 3.160919540229885e-07,
+      "loss": 0.0017,
+      "reward": 1.89532470703125,
+      "reward_std": 0.016749953851103783,
+      "rewards/accuracy_reward": 0.89532470703125,
+      "rewards/format_reward": 1.0,
+      "step": 595
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.40560150146484,
+      "epoch": 6.850574712643678,
+      "grad_norm": 1.4735604817931964,
+      "kl": 0.045654296875,
+      "learning_rate": 3.149425287356321e-07,
+      "loss": 0.002,
+      "reward": 1.9018571376800537,
+      "reward_std": 0.01927165314555168,
+      "rewards/accuracy_reward": 0.9018572568893433,
+      "rewards/format_reward": 1.0,
+      "step": 596
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.650390625,
+      "epoch": 6.862068965517241,
+      "grad_norm": 2.3771678349296113,
+      "kl": 0.04638671875,
+      "learning_rate": 3.1379310344827587e-07,
+      "loss": 0.0019,
+      "reward": 1.902691125869751,
+      "reward_std": 0.019225675612688065,
+      "rewards/accuracy_reward": 0.9033421277999878,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 597
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.69661712646484,
+      "epoch": 6.873563218390805,
+      "grad_norm": 1.3517485796795934,
+      "kl": 0.041015625,
+      "learning_rate": 3.126436781609195e-07,
+      "loss": 0.0017,
+      "reward": 1.904984712600708,
+      "reward_std": 0.01752365753054619,
+      "rewards/accuracy_reward": 0.9049846529960632,
+      "rewards/format_reward": 1.0,
+      "step": 598
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.94792175292969,
+      "epoch": 6.885057471264368,
+      "grad_norm": 1.7049030907812799,
+      "kl": 0.046630859375,
+      "learning_rate": 3.114942528735632e-07,
+      "loss": 0.002,
+      "reward": 1.8980730772018433,
+      "reward_std": 0.016529429703950882,
+      "rewards/accuracy_reward": 0.8980730772018433,
+      "rewards/format_reward": 1.0,
+      "step": 599
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.34700775146484,
+      "epoch": 6.896551724137931,
+      "grad_norm": 1.8048662278319052,
+      "kl": 0.04638671875,
+      "learning_rate": 3.103448275862069e-07,
+      "loss": 0.002,
+      "reward": 1.9081945419311523,
+      "reward_std": 0.01681654527783394,
+      "rewards/accuracy_reward": 0.9081945419311523,
+      "rewards/format_reward": 1.0,
+      "step": 600
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.6796875,
+      "epoch": 6.908045977011494,
+      "grad_norm": 1.9132885179118997,
+      "kl": 0.04736328125,
+      "learning_rate": 3.091954022988506e-07,
+      "loss": 0.0019,
+      "reward": 1.8925232887268066,
+      "reward_std": 0.020483635365962982,
+      "rewards/accuracy_reward": 0.8931743502616882,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 601
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.28841400146484,
+      "epoch": 6.919540229885057,
+      "grad_norm": 2.831161118788208,
+      "kl": 0.04541015625,
+      "learning_rate": 3.0804597701149426e-07,
+      "loss": 0.0019,
+      "reward": 1.8872265815734863,
+      "reward_std": 0.01937289349734783,
+      "rewards/accuracy_reward": 0.8872265815734863,
+      "rewards/format_reward": 1.0,
+      "step": 602
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.193359375,
+      "epoch": 6.931034482758621,
+      "grad_norm": 1.9008533326371801,
+      "kl": 0.046875,
+      "learning_rate": 3.068965517241379e-07,
+      "loss": 0.002,
+      "reward": 1.9161338806152344,
+      "reward_std": 0.017956051975488663,
+      "rewards/accuracy_reward": 0.9161338806152344,
+      "rewards/format_reward": 1.0,
+      "step": 603
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.76432800292969,
+      "epoch": 6.942528735632184,
+      "grad_norm": 1.9049955094649353,
+      "kl": 0.04541015625,
+      "learning_rate": 3.057471264367816e-07,
+      "loss": 0.0019,
+      "reward": 1.8926061391830444,
+      "reward_std": 0.019360072910785675,
+      "rewards/accuracy_reward": 0.8932573199272156,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 604
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.96745300292969,
+      "epoch": 6.954022988505747,
+      "grad_norm": 1.275994764733278,
+      "kl": 0.0458984375,
+      "learning_rate": 3.045977011494253e-07,
+      "loss": 0.0019,
+      "reward": 1.9014930725097656,
+      "reward_std": 0.017896311357617378,
+      "rewards/accuracy_reward": 0.9014931917190552,
+      "rewards/format_reward": 1.0,
+      "step": 605
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.83528900146484,
+      "epoch": 6.9655172413793105,
+      "grad_norm": 1.6892176163979111,
+      "kl": 0.044677734375,
+      "learning_rate": 3.0344827586206897e-07,
+      "loss": 0.0019,
+      "reward": 1.8952045440673828,
+      "reward_std": 0.01996638998389244,
+      "rewards/accuracy_reward": 0.895204484462738,
+      "rewards/format_reward": 1.0,
+      "step": 606
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.58659362792969,
+      "epoch": 6.977011494252873,
+      "grad_norm": 1.3212776547555032,
+      "kl": 0.04541015625,
+      "learning_rate": 3.022988505747126e-07,
+      "loss": 0.0019,
+      "reward": 1.9071601629257202,
+      "reward_std": 0.020716873928904533,
+      "rewards/accuracy_reward": 0.907811164855957,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 607
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.09505462646484,
+      "epoch": 6.988505747126437,
+      "grad_norm": 1.3698382830038993,
+      "kl": 0.046142578125,
+      "learning_rate": 3.011494252873563e-07,
+      "loss": 0.0019,
+      "reward": 1.8822848796844482,
+      "reward_std": 0.02124325931072235,
+      "rewards/accuracy_reward": 0.8822848200798035,
+      "rewards/format_reward": 1.0,
+      "step": 608
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.09129333496094,
+      "epoch": 7.0,
+      "grad_norm": 1.611157838283439,
+      "kl": 0.04150390625,
+      "learning_rate": 3e-07,
+      "loss": 0.0017,
+      "reward": 1.8909047842025757,
+      "reward_std": 0.01773572526872158,
+      "rewards/accuracy_reward": 0.8909049034118652,
+      "rewards/format_reward": 1.0,
+      "step": 609
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 112.59440612792969,
+      "epoch": 7.011494252873563,
+      "grad_norm": 2.0128308488525946,
+      "kl": 0.04345703125,
+      "learning_rate": 2.988505747126437e-07,
+      "loss": 0.0019,
+      "reward": 1.8733458518981934,
+      "reward_std": 0.021298928186297417,
+      "rewards/accuracy_reward": 0.873996913433075,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 610
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 113.76628112792969,
+      "epoch": 7.022988505747127,
+      "grad_norm": 1.7741703706980398,
+      "kl": 0.04296875,
+      "learning_rate": 2.9770114942528737e-07,
+      "loss": 0.0018,
+      "reward": 1.8772882223129272,
+      "reward_std": 0.02053675800561905,
+      "rewards/accuracy_reward": 0.8772882223129272,
+      "rewards/format_reward": 1.0,
+      "step": 611
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.43229675292969,
+      "epoch": 7.0344827586206895,
+      "grad_norm": 2.61099721986832,
+      "kl": 0.042724609375,
+      "learning_rate": 2.96551724137931e-07,
+      "loss": 0.0018,
+      "reward": 1.8760509490966797,
+      "reward_std": 0.023378584533929825,
+      "rewards/accuracy_reward": 0.8767021894454956,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 612
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.71745300292969,
+      "epoch": 7.045977011494253,
+      "grad_norm": 1.681876205953948,
+      "kl": 0.043212890625,
+      "learning_rate": 2.954022988505747e-07,
+      "loss": 0.0018,
+      "reward": 1.9030117988586426,
+      "reward_std": 0.019194534048438072,
+      "rewards/accuracy_reward": 0.9030117988586426,
+      "rewards/format_reward": 1.0,
+      "step": 613
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.92253112792969,
+      "epoch": 7.057471264367816,
+      "grad_norm": 1.401507959053581,
+      "kl": 0.042236328125,
+      "learning_rate": 2.942528735632184e-07,
+      "loss": 0.0018,
+      "reward": 1.886817216873169,
+      "reward_std": 0.0201483853161335,
+      "rewards/accuracy_reward": 0.8868171572685242,
+      "rewards/format_reward": 1.0,
+      "step": 614
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.21224212646484,
+      "epoch": 7.068965517241379,
+      "grad_norm": 2.3290159563411725,
+      "kl": 0.046142578125,
+      "learning_rate": 2.93103448275862e-07,
+      "loss": 0.0019,
+      "reward": 1.9030017852783203,
+      "reward_std": 0.021339600905776024,
+      "rewards/accuracy_reward": 0.9036529064178467,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 615
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.49089050292969,
+      "epoch": 7.080459770114943,
+      "grad_norm": 2.406758547267441,
+      "kl": 0.044921875,
+      "learning_rate": 2.9195402298850576e-07,
+      "loss": 0.0019,
+      "reward": 1.8917362689971924,
+      "reward_std": 0.02221861481666565,
+      "rewards/accuracy_reward": 0.892387330532074,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 616
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.44921875,
+      "epoch": 7.091954022988506,
+      "grad_norm": 2.4685783634645224,
+      "kl": 0.047607421875,
+      "learning_rate": 2.908045977011494e-07,
+      "loss": 0.002,
+      "reward": 1.8855441808700562,
+      "reward_std": 0.022871162742376328,
+      "rewards/accuracy_reward": 0.8855441808700562,
+      "rewards/format_reward": 1.0,
+      "step": 617
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.88021087646484,
+      "epoch": 7.103448275862069,
+      "grad_norm": 6.626736993198597,
+      "kl": 0.047119140625,
+      "learning_rate": 2.896551724137931e-07,
+      "loss": 0.002,
+      "reward": 1.8934516906738281,
+      "reward_std": 0.021955883130431175,
+      "rewards/accuracy_reward": 0.8934516906738281,
+      "rewards/format_reward": 1.0,
+      "step": 618
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.73698425292969,
+      "epoch": 7.114942528735632,
+      "grad_norm": 1.4755499555332456,
+      "kl": 0.043212890625,
+      "learning_rate": 2.885057471264368e-07,
+      "loss": 0.0019,
+      "reward": 1.8936246633529663,
+      "reward_std": 0.020434757694602013,
+      "rewards/accuracy_reward": 0.8936247229576111,
+      "rewards/format_reward": 1.0,
+      "step": 619
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.470703125,
+      "epoch": 7.126436781609195,
+      "grad_norm": 2.394219832051532,
+      "kl": 0.04541015625,
+      "learning_rate": 2.873563218390804e-07,
+      "loss": 0.0019,
+      "reward": 1.9143403768539429,
+      "reward_std": 0.020542293787002563,
+      "rewards/accuracy_reward": 0.9149913787841797,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 620
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.93815612792969,
+      "epoch": 7.137931034482759,
+      "grad_norm": 1.5127458134308953,
+      "kl": 0.042724609375,
+      "learning_rate": 2.8620689655172416e-07,
+      "loss": 0.0018,
+      "reward": 1.8820905685424805,
+      "reward_std": 0.0194623414427042,
+      "rewards/accuracy_reward": 0.8827415704727173,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 621
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.171875,
+      "epoch": 7.149425287356322,
+      "grad_norm": 1.239872088979501,
+      "kl": 0.04296875,
+      "learning_rate": 2.850574712643678e-07,
+      "loss": 0.0018,
+      "reward": 1.8986458778381348,
+      "reward_std": 0.020148858428001404,
+      "rewards/accuracy_reward": 0.8986459970474243,
+      "rewards/format_reward": 1.0,
+      "step": 622
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.97265625,
+      "epoch": 7.160919540229885,
+      "grad_norm": 2.458759960422581,
+      "kl": 0.04248046875,
+      "learning_rate": 2.8390804597701143e-07,
+      "loss": 0.0018,
+      "reward": 1.9095439910888672,
+      "reward_std": 0.018151775002479553,
+      "rewards/accuracy_reward": 0.9095439314842224,
+      "rewards/format_reward": 1.0,
+      "step": 623
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.96484375,
+      "epoch": 7.172413793103448,
+      "grad_norm": 1.7316074752480677,
+      "kl": 0.0439453125,
+      "learning_rate": 2.827586206896552e-07,
+      "loss": 0.0018,
+      "reward": 1.916879653930664,
+      "reward_std": 0.016119133681058884,
+      "rewards/accuracy_reward": 0.9168797731399536,
+      "rewards/format_reward": 1.0,
+      "step": 624
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.93685150146484,
+      "epoch": 7.183908045977011,
+      "grad_norm": 7.6966596331183545,
+      "kl": 0.0439453125,
+      "learning_rate": 2.816091954022988e-07,
+      "loss": 0.0018,
+      "reward": 1.904394507408142,
+      "reward_std": 0.01645834557712078,
+      "rewards/accuracy_reward": 0.9043946266174316,
+      "rewards/format_reward": 1.0,
+      "step": 625
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.34310150146484,
+      "epoch": 7.195402298850575,
+      "grad_norm": 3.0027221313342034,
+      "kl": 0.0419921875,
+      "learning_rate": 2.8045977011494255e-07,
+      "loss": 0.0017,
+      "reward": 1.899554967880249,
+      "reward_std": 0.022434815764427185,
+      "rewards/accuracy_reward": 0.9002060294151306,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 626
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.80989837646484,
+      "epoch": 7.206896551724138,
+      "grad_norm": 1.6600118557647507,
+      "kl": 0.0400390625,
+      "learning_rate": 2.793103448275862e-07,
+      "loss": 0.0016,
+      "reward": 1.8966201543807983,
+      "reward_std": 0.019420992583036423,
+      "rewards/accuracy_reward": 0.8972713351249695,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 627
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.37174987792969,
+      "epoch": 7.218390804597701,
+      "grad_norm": 1.710315303436248,
+      "kl": 0.046875,
+      "learning_rate": 2.781609195402299e-07,
+      "loss": 0.002,
+      "reward": 1.9059438705444336,
+      "reward_std": 0.01961967721581459,
+      "rewards/accuracy_reward": 0.9059439897537231,
+      "rewards/format_reward": 1.0,
+      "step": 628
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.951171875,
+      "epoch": 7.2298850574712645,
+      "grad_norm": 1.3756891644716789,
+      "kl": 0.041748046875,
+      "learning_rate": 2.7701149425287357e-07,
+      "loss": 0.0017,
+      "reward": 1.888430118560791,
+      "reward_std": 0.0190572552382946,
+      "rewards/accuracy_reward": 0.8890811800956726,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 629
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.404296875,
+      "epoch": 7.241379310344827,
+      "grad_norm": 3970437544517841.0,
+      "kl": 11751030521856.0,
+      "learning_rate": 2.758620689655172e-07,
+      "loss": 470864691200.0,
+      "reward": 1.9054559469223022,
+      "reward_std": 0.020271051675081253,
+      "rewards/accuracy_reward": 0.9067580103874207,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 630
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.57747650146484,
+      "epoch": 7.252873563218391,
+      "grad_norm": 1.8817887294402726,
+      "kl": 0.04541015625,
+      "learning_rate": 2.747126436781609e-07,
+      "loss": 0.0019,
+      "reward": 1.8884859085083008,
+      "reward_std": 0.018849171698093414,
+      "rewards/accuracy_reward": 0.8884860277175903,
+      "rewards/format_reward": 1.0,
+      "step": 631
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.96875,
+      "epoch": 7.264367816091954,
+      "grad_norm": 1.7719505691506174,
+      "kl": 0.046142578125,
+      "learning_rate": 2.735632183908046e-07,
+      "loss": 0.0019,
+      "reward": 1.9029910564422607,
+      "reward_std": 0.021298250183463097,
+      "rewards/accuracy_reward": 0.9036421179771423,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 632
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.53841400146484,
+      "epoch": 7.275862068965517,
+      "grad_norm": 2.449085093736059,
+      "kl": 0.05078125,
+      "learning_rate": 2.724137931034483e-07,
+      "loss": 0.0021,
+      "reward": 1.8815187215805054,
+      "reward_std": 0.018999949097633362,
+      "rewards/accuracy_reward": 0.8815188407897949,
+      "rewards/format_reward": 1.0,
+      "step": 633
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.56640625,
+      "epoch": 7.287356321839081,
+      "grad_norm": 2.98961076193299,
+      "kl": 0.044677734375,
+      "learning_rate": 2.7126436781609197e-07,
+      "loss": 0.0019,
+      "reward": 1.9084595441818237,
+      "reward_std": 0.0186910443007946,
+      "rewards/accuracy_reward": 0.9084595441818237,
+      "rewards/format_reward": 1.0,
+      "step": 634
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.57357025146484,
+      "epoch": 7.2988505747126435,
+      "grad_norm": 2.134961142900107,
+      "kl": 0.04736328125,
+      "learning_rate": 2.701149425287356e-07,
+      "loss": 0.0019,
+      "reward": 1.9063668251037598,
+      "reward_std": 0.016174225136637688,
+      "rewards/accuracy_reward": 0.906366765499115,
+      "rewards/format_reward": 1.0,
+      "step": 635
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.66276550292969,
+      "epoch": 7.310344827586207,
+      "grad_norm": 1.4763178317390981,
+      "kl": 0.04052734375,
+      "learning_rate": 2.689655172413793e-07,
+      "loss": 0.0017,
+      "reward": 1.898642897605896,
+      "reward_std": 0.017671559005975723,
+      "rewards/accuracy_reward": 0.8986430168151855,
+      "rewards/format_reward": 1.0,
+      "step": 636
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.041015625,
+      "epoch": 7.32183908045977,
+      "grad_norm": 1.6822870021161793,
+      "kl": 0.048583984375,
+      "learning_rate": 2.67816091954023e-07,
+      "loss": 0.002,
+      "reward": 1.8975701332092285,
+      "reward_std": 0.018257655203342438,
+      "rewards/accuracy_reward": 0.8982211947441101,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 637
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.96549987792969,
+      "epoch": 7.333333333333333,
+      "grad_norm": 1.4931878670654746,
+      "kl": 0.044189453125,
+      "learning_rate": 2.6666666666666667e-07,
+      "loss": 0.0018,
+      "reward": 1.9090592861175537,
+      "reward_std": 0.018635626882314682,
+      "rewards/accuracy_reward": 0.9090592265129089,
+      "rewards/format_reward": 1.0,
+      "step": 638
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.53450775146484,
+      "epoch": 7.344827586206897,
+      "grad_norm": 1.96698633932992,
+      "kl": 0.044677734375,
+      "learning_rate": 2.655172413793103e-07,
+      "loss": 0.0019,
+      "reward": 1.8955191373825073,
+      "reward_std": 0.01896989531815052,
+      "rewards/accuracy_reward": 0.8961701393127441,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 639
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.40495300292969,
+      "epoch": 7.35632183908046,
+      "grad_norm": 2.0570123873234203,
+      "kl": 0.0419921875,
+      "learning_rate": 2.64367816091954e-07,
+      "loss": 0.0018,
+      "reward": 1.9056072235107422,
+      "reward_std": 0.01762087456882,
+      "rewards/accuracy_reward": 0.9062582850456238,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 640
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.19857025146484,
+      "epoch": 7.3678160919540225,
+      "grad_norm": 1.6420815311017691,
+      "kl": 0.043212890625,
+      "learning_rate": 2.632183908045977e-07,
+      "loss": 0.0018,
+      "reward": 1.8763535022735596,
+      "reward_std": 0.02036045491695404,
+      "rewards/accuracy_reward": 0.8763534426689148,
+      "rewards/format_reward": 1.0,
+      "step": 641
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.95247650146484,
+      "epoch": 7.379310344827586,
+      "grad_norm": 1.68174331233095,
+      "kl": 0.04638671875,
+      "learning_rate": 2.620689655172414e-07,
+      "loss": 0.0019,
+      "reward": 1.9021470546722412,
+      "reward_std": 0.02229691669344902,
+      "rewards/accuracy_reward": 0.9021470546722412,
+      "rewards/format_reward": 1.0,
+      "step": 642
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.66732025146484,
+      "epoch": 7.390804597701149,
+      "grad_norm": 1.6367021800993964,
+      "kl": 0.04296875,
+      "learning_rate": 2.6091954022988507e-07,
+      "loss": 0.0018,
+      "reward": 1.904337763786316,
+      "reward_std": 0.021419089287519455,
+      "rewards/accuracy_reward": 0.9056398868560791,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 643
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.353515625,
+      "epoch": 7.402298850574713,
+      "grad_norm": 1.7586162669562002,
+      "kl": 0.047607421875,
+      "learning_rate": 2.597701149425287e-07,
+      "loss": 0.002,
+      "reward": 1.8953742980957031,
+      "reward_std": 0.01957518607378006,
+      "rewards/accuracy_reward": 0.8960254192352295,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 644
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.19010925292969,
+      "epoch": 7.413793103448276,
+      "grad_norm": 2.5073706132424447,
+      "kl": 0.044921875,
+      "learning_rate": 2.586206896551724e-07,
+      "loss": 0.0018,
+      "reward": 1.8890656232833862,
+      "reward_std": 0.02112973853945732,
+      "rewards/accuracy_reward": 0.8903676867485046,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 645
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.71484375,
+      "epoch": 7.425287356321839,
+      "grad_norm": 2.802792541775325,
+      "kl": 0.045166015625,
+      "learning_rate": 2.574712643678161e-07,
+      "loss": 0.0019,
+      "reward": 1.8906625509262085,
+      "reward_std": 0.021000966429710388,
+      "rewards/accuracy_reward": 0.8906625509262085,
+      "rewards/format_reward": 1.0,
+      "step": 646
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.3203125,
+      "epoch": 7.436781609195402,
+      "grad_norm": 1.531808532330182,
+      "kl": 0.04296875,
+      "learning_rate": 2.563218390804597e-07,
+      "loss": 0.0018,
+      "reward": 1.902920126914978,
+      "reward_std": 0.018811281770467758,
+      "rewards/accuracy_reward": 0.902920126914978,
+      "rewards/format_reward": 1.0,
+      "step": 647
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.66796875,
+      "epoch": 7.448275862068965,
+      "grad_norm": 2.9559025913941355,
+      "kl": 0.0439453125,
+      "learning_rate": 2.5517241379310346e-07,
+      "loss": 0.0018,
+      "reward": 1.8963336944580078,
+      "reward_std": 0.020042497664690018,
+      "rewards/accuracy_reward": 0.8969849348068237,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 648
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.14909362792969,
+      "epoch": 7.459770114942529,
+      "grad_norm": 1.9565147315824727,
+      "kl": 0.04736328125,
+      "learning_rate": 2.540229885057471e-07,
+      "loss": 0.0019,
+      "reward": 1.9087027311325073,
+      "reward_std": 0.018100788816809654,
+      "rewards/accuracy_reward": 0.9087027311325073,
+      "rewards/format_reward": 1.0,
+      "step": 649
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.951171875,
+      "epoch": 7.471264367816092,
+      "grad_norm": 3.147083124745903,
+      "kl": 0.04443359375,
+      "learning_rate": 2.5287356321839084e-07,
+      "loss": 0.0019,
+      "reward": 1.9156570434570312,
+      "reward_std": 0.01904350332915783,
+      "rewards/accuracy_reward": 0.9156570434570312,
+      "rewards/format_reward": 1.0,
+      "step": 650
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.552734375,
+      "epoch": 7.482758620689655,
+      "grad_norm": 1.6343407556749157,
+      "kl": 0.051513671875,
+      "learning_rate": 2.517241379310345e-07,
+      "loss": 0.0022,
+      "reward": 1.9167011976242065,
+      "reward_std": 0.01607966236770153,
+      "rewards/accuracy_reward": 0.9167011976242065,
+      "rewards/format_reward": 1.0,
+      "step": 651
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.46940612792969,
+      "epoch": 7.494252873563219,
+      "grad_norm": 2.380370478966516,
+      "kl": 0.04541015625,
+      "learning_rate": 2.505747126436781e-07,
+      "loss": 0.0019,
+      "reward": 1.8996713161468506,
+      "reward_std": 0.018466008827090263,
+      "rewards/accuracy_reward": 0.9003223776817322,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 652
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.76692962646484,
+      "epoch": 7.505747126436781,
+      "grad_norm": 1.572354097669339,
+      "kl": 0.049560546875,
+      "learning_rate": 2.494252873563218e-07,
+      "loss": 0.0021,
+      "reward": 1.8812230825424194,
+      "reward_std": 0.020288746803998947,
+      "rewards/accuracy_reward": 0.8818740844726562,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 653
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.83984375,
+      "epoch": 7.517241379310345,
+      "grad_norm": 3.4037020500640507,
+      "kl": 0.057373046875,
+      "learning_rate": 2.482758620689655e-07,
+      "loss": 0.0024,
+      "reward": 1.9109630584716797,
+      "reward_std": 0.01749694161117077,
+      "rewards/accuracy_reward": 0.9109630584716797,
+      "rewards/format_reward": 1.0,
+      "step": 654
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.07292175292969,
+      "epoch": 7.528735632183908,
+      "grad_norm": 1.4740870320089872,
+      "kl": 0.049072265625,
+      "learning_rate": 2.471264367816092e-07,
+      "loss": 0.002,
+      "reward": 1.9072259664535522,
+      "reward_std": 0.016256902366876602,
+      "rewards/accuracy_reward": 0.907226026058197,
+      "rewards/format_reward": 1.0,
+      "step": 655
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.85612487792969,
+      "epoch": 7.540229885057471,
+      "grad_norm": 2.126019785623936,
+      "kl": 0.04931640625,
+      "learning_rate": 2.459770114942529e-07,
+      "loss": 0.002,
+      "reward": 1.8846591711044312,
+      "reward_std": 0.018113628029823303,
+      "rewards/accuracy_reward": 0.8846592903137207,
+      "rewards/format_reward": 1.0,
+      "step": 656
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.55794525146484,
+      "epoch": 7.551724137931035,
+      "grad_norm": 7.89243748687455,
+      "kl": 0.047607421875,
+      "learning_rate": 2.448275862068965e-07,
+      "loss": 0.002,
+      "reward": 1.8965566158294678,
+      "reward_std": 0.019315050914883614,
+      "rewards/accuracy_reward": 0.896556556224823,
+      "rewards/format_reward": 1.0,
+      "step": 657
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.54817962646484,
+      "epoch": 7.563218390804598,
+      "grad_norm": 3.4267861157909403,
+      "kl": 0.046142578125,
+      "learning_rate": 2.436781609195402e-07,
+      "loss": 0.0019,
+      "reward": 1.8947169780731201,
+      "reward_std": 0.019318781793117523,
+      "rewards/accuracy_reward": 0.8947169184684753,
+      "rewards/format_reward": 1.0,
+      "step": 658
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.568359375,
+      "epoch": 7.574712643678161,
+      "grad_norm": 1.6852921029556645,
+      "kl": 0.052490234375,
+      "learning_rate": 2.425287356321839e-07,
+      "loss": 0.0022,
+      "reward": 1.9210675954818726,
+      "reward_std": 0.015680838376283646,
+      "rewards/accuracy_reward": 0.9210675954818726,
+      "rewards/format_reward": 1.0,
+      "step": 659
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.3046875,
+      "epoch": 7.586206896551724,
+      "grad_norm": 2.206767755219908,
+      "kl": 0.04541015625,
+      "learning_rate": 2.413793103448276e-07,
+      "loss": 0.0019,
+      "reward": 1.8979721069335938,
+      "reward_std": 0.016466915607452393,
+      "rewards/accuracy_reward": 0.8979721069335938,
+      "rewards/format_reward": 1.0,
+      "step": 660
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.42643737792969,
+      "epoch": 7.597701149425287,
+      "grad_norm": 2.116528759364446,
+      "kl": 0.043701171875,
+      "learning_rate": 2.402298850574712e-07,
+      "loss": 0.0018,
+      "reward": 1.9062350988388062,
+      "reward_std": 0.015808025375008583,
+      "rewards/accuracy_reward": 0.9062352180480957,
+      "rewards/format_reward": 1.0,
+      "step": 661
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.13021087646484,
+      "epoch": 7.609195402298851,
+      "grad_norm": 2.3658556212140773,
+      "kl": 0.04736328125,
+      "learning_rate": 2.390804597701149e-07,
+      "loss": 0.002,
+      "reward": 1.9090293645858765,
+      "reward_std": 0.020903117954730988,
+      "rewards/accuracy_reward": 0.9096805453300476,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 662
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.53060150146484,
+      "epoch": 7.620689655172414,
+      "grad_norm": 1.736164910774429,
+      "kl": 0.04443359375,
+      "learning_rate": 2.3793103448275863e-07,
+      "loss": 0.0018,
+      "reward": 1.892618179321289,
+      "reward_std": 0.020376306027173996,
+      "rewards/accuracy_reward": 0.8932693004608154,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 663
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.14192962646484,
+      "epoch": 7.6321839080459775,
+      "grad_norm": 1.661775706838517,
+      "kl": 0.04345703125,
+      "learning_rate": 2.367816091954023e-07,
+      "loss": 0.0018,
+      "reward": 1.910132884979248,
+      "reward_std": 0.01860654354095459,
+      "rewards/accuracy_reward": 0.9107840061187744,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 664
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.82878112792969,
+      "epoch": 7.64367816091954,
+      "grad_norm": 1.775794567645862,
+      "kl": 0.047607421875,
+      "learning_rate": 2.3563218390804595e-07,
+      "loss": 0.002,
+      "reward": 1.8953571319580078,
+      "reward_std": 0.019160928204655647,
+      "rewards/accuracy_reward": 0.8953571319580078,
+      "rewards/format_reward": 1.0,
+      "step": 665
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.8515625,
+      "epoch": 7.655172413793103,
+      "grad_norm": 2.852042326779303,
+      "kl": 0.05224609375,
+      "learning_rate": 2.3448275862068964e-07,
+      "loss": 0.0021,
+      "reward": 1.9160139560699463,
+      "reward_std": 0.01864560693502426,
+      "rewards/accuracy_reward": 0.9160139560699463,
+      "rewards/format_reward": 1.0,
+      "step": 666
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.22135925292969,
+      "epoch": 7.666666666666667,
+      "grad_norm": 2.00298655553839,
+      "kl": 0.046142578125,
+      "learning_rate": 2.3333333333333333e-07,
+      "loss": 0.0019,
+      "reward": 1.8910598754882812,
+      "reward_std": 0.020590651780366898,
+      "rewards/accuracy_reward": 0.8910599946975708,
+      "rewards/format_reward": 1.0,
+      "step": 667
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.58268737792969,
+      "epoch": 7.67816091954023,
+      "grad_norm": 2.2015919771538255,
+      "kl": 0.044189453125,
+      "learning_rate": 2.3218390804597702e-07,
+      "loss": 0.0018,
+      "reward": 1.897012710571289,
+      "reward_std": 0.02280309796333313,
+      "rewards/accuracy_reward": 0.8983148336410522,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 668
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.41471862792969,
+      "epoch": 7.689655172413794,
+      "grad_norm": 4.620691107539993,
+      "kl": 0.048828125,
+      "learning_rate": 2.3103448275862066e-07,
+      "loss": 0.0021,
+      "reward": 1.9026827812194824,
+      "reward_std": 0.018936630338430405,
+      "rewards/accuracy_reward": 0.903333842754364,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 669
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.208984375,
+      "epoch": 7.7011494252873565,
+      "grad_norm": 1.5207850843287138,
+      "kl": 0.052001953125,
+      "learning_rate": 2.2988505747126435e-07,
+      "loss": 0.0021,
+      "reward": 1.8975833654403687,
+      "reward_std": 0.01803259551525116,
+      "rewards/accuracy_reward": 0.8975834846496582,
+      "rewards/format_reward": 1.0,
+      "step": 670
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.48698425292969,
+      "epoch": 7.712643678160919,
+      "grad_norm": 1.1439576464320695,
+      "kl": 0.04248046875,
+      "learning_rate": 2.2873563218390804e-07,
+      "loss": 0.0018,
+      "reward": 1.8865493535995483,
+      "reward_std": 0.018641415983438492,
+      "rewards/accuracy_reward": 0.8865493535995483,
+      "rewards/format_reward": 1.0,
+      "step": 671
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.46745300292969,
+      "epoch": 7.724137931034483,
+      "grad_norm": 1.5940069065520925,
+      "kl": 0.0419921875,
+      "learning_rate": 2.2758620689655173e-07,
+      "loss": 0.0017,
+      "reward": 1.874269723892212,
+      "reward_std": 0.02033749222755432,
+      "rewards/accuracy_reward": 0.8742696642875671,
+      "rewards/format_reward": 1.0,
+      "step": 672
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.087890625,
+      "epoch": 7.735632183908046,
+      "grad_norm": 2.1488145925757234,
+      "kl": 0.048828125,
+      "learning_rate": 2.264367816091954e-07,
+      "loss": 0.002,
+      "reward": 1.9152870178222656,
+      "reward_std": 0.015497724525630474,
+      "rewards/accuracy_reward": 0.9152869582176208,
+      "rewards/format_reward": 1.0,
+      "step": 673
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.20052337646484,
+      "epoch": 7.747126436781609,
+      "grad_norm": 1.5334204725557474,
+      "kl": 0.045654296875,
+      "learning_rate": 2.2528735632183905e-07,
+      "loss": 0.0019,
+      "reward": 1.9050960540771484,
+      "reward_std": 0.021023746579885483,
+      "rewards/accuracy_reward": 0.9057471752166748,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 674
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.88671875,
+      "epoch": 7.758620689655173,
+      "grad_norm": 1.6306945359970868,
+      "kl": 0.03955078125,
+      "learning_rate": 2.2413793103448274e-07,
+      "loss": 0.0017,
+      "reward": 1.908623456954956,
+      "reward_std": 0.017657993361353874,
+      "rewards/accuracy_reward": 0.9086233973503113,
+      "rewards/format_reward": 1.0,
+      "step": 675
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.23567962646484,
+      "epoch": 7.7701149425287355,
+      "grad_norm": 1.3620347321088389,
+      "kl": 0.0400390625,
+      "learning_rate": 2.2298850574712643e-07,
+      "loss": 0.0017,
+      "reward": 1.8560750484466553,
+      "reward_std": 0.021627038717269897,
+      "rewards/accuracy_reward": 0.8573770523071289,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 676
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 111.77278900146484,
+      "epoch": 7.781609195402299,
+      "grad_norm": 2.4445187153033547,
+      "kl": 0.041015625,
+      "learning_rate": 2.218390804597701e-07,
+      "loss": 0.0017,
+      "reward": 1.9019120931625366,
+      "reward_std": 0.02054600417613983,
+      "rewards/accuracy_reward": 0.9019122123718262,
+      "rewards/format_reward": 1.0,
+      "step": 677
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.171875,
+      "epoch": 7.793103448275862,
+      "grad_norm": 2.2620273599837537,
+      "kl": 0.03759765625,
+      "learning_rate": 2.206896551724138e-07,
+      "loss": 0.0016,
+      "reward": 1.899298071861267,
+      "reward_std": 0.019102804362773895,
+      "rewards/accuracy_reward": 0.8992981314659119,
+      "rewards/format_reward": 1.0,
+      "step": 678
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.884765625,
+      "epoch": 7.804597701149425,
+      "grad_norm": 2.7848035680621446,
+      "kl": 0.038330078125,
+      "learning_rate": 2.1954022988505748e-07,
+      "loss": 0.0016,
+      "reward": 1.8842277526855469,
+      "reward_std": 0.015971308574080467,
+      "rewards/accuracy_reward": 0.8842277526855469,
+      "rewards/format_reward": 1.0,
+      "step": 679
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.04232025146484,
+      "epoch": 7.816091954022989,
+      "grad_norm": 1.526409335432903,
+      "kl": 0.04736328125,
+      "learning_rate": 2.1839080459770114e-07,
+      "loss": 0.002,
+      "reward": 1.9061853885650635,
+      "reward_std": 0.018970629200339317,
+      "rewards/accuracy_reward": 0.9061853289604187,
+      "rewards/format_reward": 1.0,
+      "step": 680
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.986328125,
+      "epoch": 7.827586206896552,
+      "grad_norm": 2.4205054636705627,
+      "kl": 0.04345703125,
+      "learning_rate": 2.172413793103448e-07,
+      "loss": 0.0019,
+      "reward": 1.8783591985702515,
+      "reward_std": 0.020205635577440262,
+      "rewards/accuracy_reward": 0.8783591985702515,
+      "rewards/format_reward": 1.0,
+      "step": 681
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.24544525146484,
+      "epoch": 7.8390804597701145,
+      "grad_norm": 1.6438250497849267,
+      "kl": 0.0380859375,
+      "learning_rate": 2.160919540229885e-07,
+      "loss": 0.0016,
+      "reward": 1.9045305252075195,
+      "reward_std": 0.020377201959490776,
+      "rewards/accuracy_reward": 0.9051817655563354,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 682
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.76171875,
+      "epoch": 7.850574712643678,
+      "grad_norm": 1.2996080329358168,
+      "kl": 0.041748046875,
+      "learning_rate": 2.1494252873563218e-07,
+      "loss": 0.0017,
+      "reward": 1.907975435256958,
+      "reward_std": 0.017765961587429047,
+      "rewards/accuracy_reward": 0.9079753756523132,
+      "rewards/format_reward": 1.0,
+      "step": 683
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.03190612792969,
+      "epoch": 7.862068965517241,
+      "grad_norm": 5.191492192434162,
+      "kl": 0.0400390625,
+      "learning_rate": 2.1379310344827587e-07,
+      "loss": 0.0017,
+      "reward": 1.9033807516098022,
+      "reward_std": 0.019389839842915535,
+      "rewards/accuracy_reward": 0.9033806324005127,
+      "rewards/format_reward": 1.0,
+      "step": 684
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.60026550292969,
+      "epoch": 7.873563218390805,
+      "grad_norm": 2.4150479297071725,
+      "kl": 0.04248046875,
+      "learning_rate": 2.126436781609195e-07,
+      "loss": 0.0018,
+      "reward": 1.9296969175338745,
+      "reward_std": 0.01699325256049633,
+      "rewards/accuracy_reward": 0.9296969175338745,
+      "rewards/format_reward": 1.0,
+      "step": 685
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.80078125,
+      "epoch": 7.885057471264368,
+      "grad_norm": 1.3996427566274907,
+      "kl": 0.041015625,
+      "learning_rate": 2.114942528735632e-07,
+      "loss": 0.0017,
+      "reward": 1.9054107666015625,
+      "reward_std": 0.01808765158057213,
+      "rewards/accuracy_reward": 0.905410885810852,
+      "rewards/format_reward": 1.0,
+      "step": 686
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.173828125,
+      "epoch": 7.896551724137931,
+      "grad_norm": 1.4978085636014402,
+      "kl": 0.041259765625,
+      "learning_rate": 2.103448275862069e-07,
+      "loss": 0.0017,
+      "reward": 1.9016507863998413,
+      "reward_std": 0.018792547285556793,
+      "rewards/accuracy_reward": 0.9016507863998413,
+      "rewards/format_reward": 1.0,
+      "step": 687
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.52995300292969,
+      "epoch": 7.908045977011494,
+      "grad_norm": 3.073796889740199,
+      "kl": 0.03857421875,
+      "learning_rate": 2.0919540229885058e-07,
+      "loss": 0.0016,
+      "reward": 1.913766860961914,
+      "reward_std": 0.01984010636806488,
+      "rewards/accuracy_reward": 0.9144179224967957,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 688
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.84049987792969,
+      "epoch": 7.919540229885057,
+      "grad_norm": 1.7046597310271008,
+      "kl": 0.039306640625,
+      "learning_rate": 2.0804597701149424e-07,
+      "loss": 0.0017,
+      "reward": 1.8879507780075073,
+      "reward_std": 0.02212076261639595,
+      "rewards/accuracy_reward": 0.8892530202865601,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 689
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.87435150146484,
+      "epoch": 7.931034482758621,
+      "grad_norm": 2.469859985195517,
+      "kl": 0.044921875,
+      "learning_rate": 2.0689655172413793e-07,
+      "loss": 0.0019,
+      "reward": 1.9116500616073608,
+      "reward_std": 0.021436279639601707,
+      "rewards/accuracy_reward": 0.9129521250724792,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 690
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.86979675292969,
+      "epoch": 7.942528735632184,
+      "grad_norm": 1.7410487864129292,
+      "kl": 0.043212890625,
+      "learning_rate": 2.057471264367816e-07,
+      "loss": 0.0018,
+      "reward": 1.9140040874481201,
+      "reward_std": 0.018954966217279434,
+      "rewards/accuracy_reward": 0.9140040278434753,
+      "rewards/format_reward": 1.0,
+      "step": 691
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.541015625,
+      "epoch": 7.954022988505747,
+      "grad_norm": 1.7493728678791296,
+      "kl": 0.04443359375,
+      "learning_rate": 2.0459770114942528e-07,
+      "loss": 0.0019,
+      "reward": 1.9026793241500854,
+      "reward_std": 0.015167943201959133,
+      "rewards/accuracy_reward": 0.9026793241500854,
+      "rewards/format_reward": 1.0,
+      "step": 692
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.72982025146484,
+      "epoch": 7.9655172413793105,
+      "grad_norm": 1.7859158502563095,
+      "kl": 0.043701171875,
+      "learning_rate": 2.0344827586206895e-07,
+      "loss": 0.0018,
+      "reward": 1.8844761848449707,
+      "reward_std": 0.021765243262052536,
+      "rewards/accuracy_reward": 0.8844761848449707,
+      "rewards/format_reward": 1.0,
+      "step": 693
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.84440612792969,
+      "epoch": 7.977011494252873,
+      "grad_norm": 1.5585940385657546,
+      "kl": 0.04541015625,
+      "learning_rate": 2.0229885057471264e-07,
+      "loss": 0.0019,
+      "reward": 1.8896058797836304,
+      "reward_std": 0.022624578326940536,
+      "rewards/accuracy_reward": 0.8902568817138672,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 694
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.83268737792969,
+      "epoch": 7.988505747126437,
+      "grad_norm": 4.611642156415871,
+      "kl": 0.045654296875,
+      "learning_rate": 2.0114942528735633e-07,
+      "loss": 0.0019,
+      "reward": 1.9184387922286987,
+      "reward_std": 0.01827239617705345,
+      "rewards/accuracy_reward": 0.9184388518333435,
+      "rewards/format_reward": 1.0,
+      "step": 695
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 100.7078628540039,
+      "epoch": 8.0,
+      "grad_norm": 3.175089899536824,
+      "kl": 0.043212890625,
+      "learning_rate": 2e-07,
+      "loss": 0.0019,
+      "reward": 1.917198657989502,
+      "reward_std": 0.017983654513955116,
+      "rewards/accuracy_reward": 0.9171985983848572,
+      "rewards/format_reward": 1.0,
+      "step": 696
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.88671875,
+      "epoch": 8.011494252873563,
+      "grad_norm": 1.599375509212271,
+      "kl": 0.04736328125,
+      "learning_rate": 1.9885057471264365e-07,
+      "loss": 0.002,
+      "reward": 1.900770902633667,
+      "reward_std": 0.020885910838842392,
+      "rewards/accuracy_reward": 0.9007708430290222,
+      "rewards/format_reward": 1.0,
+      "step": 697
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 110.009765625,
+      "epoch": 8.022988505747126,
+      "grad_norm": 1.5228925864244678,
+      "kl": 0.042236328125,
+      "learning_rate": 1.9770114942528734e-07,
+      "loss": 0.0018,
+      "reward": 1.8776483535766602,
+      "reward_std": 0.020006125792860985,
+      "rewards/accuracy_reward": 0.878299355506897,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 698
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.82942962646484,
+      "epoch": 8.03448275862069,
+      "grad_norm": 1.8238558955068145,
+      "kl": 0.044189453125,
+      "learning_rate": 1.9655172413793103e-07,
+      "loss": 0.0018,
+      "reward": 1.9061753749847412,
+      "reward_std": 0.01833120919764042,
+      "rewards/accuracy_reward": 0.9061753153800964,
+      "rewards/format_reward": 1.0,
+      "step": 699
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.52864837646484,
+      "epoch": 8.045977011494253,
+      "grad_norm": 1.3341303550118901,
+      "kl": 0.048095703125,
+      "learning_rate": 1.9540229885057472e-07,
+      "loss": 0.002,
+      "reward": 1.8628864288330078,
+      "reward_std": 0.02142808958888054,
+      "rewards/accuracy_reward": 0.8628865480422974,
+      "rewards/format_reward": 1.0,
+      "step": 700
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.52799987792969,
+      "epoch": 8.057471264367816,
+      "grad_norm": 1.761222226586845,
+      "kl": 0.0439453125,
+      "learning_rate": 1.9425287356321836e-07,
+      "loss": 0.0018,
+      "reward": 1.8934277296066284,
+      "reward_std": 0.0226028673350811,
+      "rewards/accuracy_reward": 0.8940789103507996,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 701
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.29948425292969,
+      "epoch": 8.068965517241379,
+      "grad_norm": 1.5102446080782972,
+      "kl": 0.0439453125,
+      "learning_rate": 1.9310344827586205e-07,
+      "loss": 0.0018,
+      "reward": 1.8807815313339233,
+      "reward_std": 0.023064211010932922,
+      "rewards/accuracy_reward": 0.8814325332641602,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 702
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.68099212646484,
+      "epoch": 8.080459770114942,
+      "grad_norm": 2.6463573417228234,
+      "kl": 0.0458984375,
+      "learning_rate": 1.9195402298850574e-07,
+      "loss": 0.0019,
+      "reward": 1.91706120967865,
+      "reward_std": 0.018780669197440147,
+      "rewards/accuracy_reward": 0.9170613288879395,
+      "rewards/format_reward": 1.0,
+      "step": 703
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.86653900146484,
+      "epoch": 8.091954022988507,
+      "grad_norm": 1.9782556380033625,
+      "kl": 0.046875,
+      "learning_rate": 1.9080459770114943e-07,
+      "loss": 0.002,
+      "reward": 1.9006422758102417,
+      "reward_std": 0.02227187156677246,
+      "rewards/accuracy_reward": 0.9012932777404785,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 704
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.27799987792969,
+      "epoch": 8.10344827586207,
+      "grad_norm": 1.571198957982024,
+      "kl": 0.043701171875,
+      "learning_rate": 1.896551724137931e-07,
+      "loss": 0.0018,
+      "reward": 1.901580810546875,
+      "reward_std": 0.018761413171887398,
+      "rewards/accuracy_reward": 0.901580810546875,
+      "rewards/format_reward": 1.0,
+      "step": 705
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.69140625,
+      "epoch": 8.114942528735632,
+      "grad_norm": 2.093159658605368,
+      "kl": 0.044921875,
+      "learning_rate": 1.8850574712643678e-07,
+      "loss": 0.0019,
+      "reward": 1.891250491142273,
+      "reward_std": 0.019944649189710617,
+      "rewards/accuracy_reward": 0.8912505507469177,
+      "rewards/format_reward": 1.0,
+      "step": 706
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.19921875,
+      "epoch": 8.126436781609195,
+      "grad_norm": 2.529542496168213,
+      "kl": 0.041748046875,
+      "learning_rate": 1.8735632183908045e-07,
+      "loss": 0.0018,
+      "reward": 1.8948949575424194,
+      "reward_std": 0.020839963108301163,
+      "rewards/accuracy_reward": 0.8948950171470642,
+      "rewards/format_reward": 1.0,
+      "step": 707
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.47721862792969,
+      "epoch": 8.137931034482758,
+      "grad_norm": 2.451670338869928,
+      "kl": 0.05419921875,
+      "learning_rate": 1.8620689655172414e-07,
+      "loss": 0.0023,
+      "reward": 1.8975166082382202,
+      "reward_std": 0.02222670242190361,
+      "rewards/accuracy_reward": 0.898167610168457,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 708
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.72005462646484,
+      "epoch": 8.149425287356323,
+      "grad_norm": 2.6109529506496525,
+      "kl": 0.04150390625,
+      "learning_rate": 1.850574712643678e-07,
+      "loss": 0.0018,
+      "reward": 1.8816875219345093,
+      "reward_std": 0.021295679733157158,
+      "rewards/accuracy_reward": 0.881687343120575,
+      "rewards/format_reward": 1.0,
+      "step": 709
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.10482025146484,
+      "epoch": 8.160919540229886,
+      "grad_norm": 2.236575205027573,
+      "kl": 0.04638671875,
+      "learning_rate": 1.839080459770115e-07,
+      "loss": 0.0019,
+      "reward": 1.8859424591064453,
+      "reward_std": 0.01964872144162655,
+      "rewards/accuracy_reward": 0.8859424591064453,
+      "rewards/format_reward": 1.0,
+      "step": 710
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.97721862792969,
+      "epoch": 8.172413793103448,
+      "grad_norm": 2.981879850123249,
+      "kl": 0.044677734375,
+      "learning_rate": 1.8275862068965518e-07,
+      "loss": 0.0019,
+      "reward": 1.9109057188034058,
+      "reward_std": 0.01663116365671158,
+      "rewards/accuracy_reward": 0.9109057188034058,
+      "rewards/format_reward": 1.0,
+      "step": 711
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.57487487792969,
+      "epoch": 8.183908045977011,
+      "grad_norm": 2.3188147233415277,
+      "kl": 0.05029296875,
+      "learning_rate": 1.8160919540229884e-07,
+      "loss": 0.0021,
+      "reward": 1.8978303670883179,
+      "reward_std": 0.021748527884483337,
+      "rewards/accuracy_reward": 0.8978304266929626,
+      "rewards/format_reward": 1.0,
+      "step": 712
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.91927337646484,
+      "epoch": 8.195402298850574,
+      "grad_norm": 3.225354097129705,
+      "kl": 0.0439453125,
+      "learning_rate": 1.804597701149425e-07,
+      "loss": 0.0019,
+      "reward": 1.8894078731536865,
+      "reward_std": 0.018167873844504356,
+      "rewards/accuracy_reward": 0.8894079923629761,
+      "rewards/format_reward": 1.0,
+      "step": 713
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.33203125,
+      "epoch": 8.206896551724139,
+      "grad_norm": 2.1412928496025216,
+      "kl": 0.0458984375,
+      "learning_rate": 1.793103448275862e-07,
+      "loss": 0.0019,
+      "reward": 1.9061758518218994,
+      "reward_std": 0.019810751080513,
+      "rewards/accuracy_reward": 0.9061757922172546,
+      "rewards/format_reward": 1.0,
+      "step": 714
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.83919525146484,
+      "epoch": 8.218390804597702,
+      "grad_norm": 1.4510305932233432,
+      "kl": 0.049072265625,
+      "learning_rate": 1.7816091954022988e-07,
+      "loss": 0.002,
+      "reward": 1.8892874717712402,
+      "reward_std": 0.01888991892337799,
+      "rewards/accuracy_reward": 0.8892874717712402,
+      "rewards/format_reward": 1.0,
+      "step": 715
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.51302337646484,
+      "epoch": 8.229885057471265,
+      "grad_norm": 2.2029044678618663,
+      "kl": 0.0537109375,
+      "learning_rate": 1.7701149425287357e-07,
+      "loss": 0.0022,
+      "reward": 1.885032057762146,
+      "reward_std": 0.01888071931898594,
+      "rewards/accuracy_reward": 0.8850321173667908,
+      "rewards/format_reward": 1.0,
+      "step": 716
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.80534362792969,
+      "epoch": 8.241379310344827,
+      "grad_norm": 2.3323680467220145,
+      "kl": 0.04833984375,
+      "learning_rate": 1.758620689655172e-07,
+      "loss": 0.002,
+      "reward": 1.9214333295822144,
+      "reward_std": 0.021036818623542786,
+      "rewards/accuracy_reward": 0.9220844507217407,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 717
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.37435150146484,
+      "epoch": 8.25287356321839,
+      "grad_norm": 2.8730245319899317,
+      "kl": 0.051513671875,
+      "learning_rate": 1.747126436781609e-07,
+      "loss": 0.0022,
+      "reward": 1.8911713361740112,
+      "reward_std": 0.019480139017105103,
+      "rewards/accuracy_reward": 0.892473578453064,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 718
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.451171875,
+      "epoch": 8.264367816091955,
+      "grad_norm": 1.9234276428720276,
+      "kl": 0.045654296875,
+      "learning_rate": 1.735632183908046e-07,
+      "loss": 0.0019,
+      "reward": 1.9077482223510742,
+      "reward_std": 0.01729690097272396,
+      "rewards/accuracy_reward": 0.9077481627464294,
+      "rewards/format_reward": 1.0,
+      "step": 719
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 102.9140625,
+      "epoch": 8.275862068965518,
+      "grad_norm": 1.7140585268322626,
+      "kl": 0.05126953125,
+      "learning_rate": 1.7241379310344828e-07,
+      "loss": 0.0021,
+      "reward": 1.9137029647827148,
+      "reward_std": 0.018116027116775513,
+      "rewards/accuracy_reward": 0.9137029647827148,
+      "rewards/format_reward": 1.0,
+      "step": 720
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.93034362792969,
+      "epoch": 8.28735632183908,
+      "grad_norm": 2.219230959644551,
+      "kl": 0.04736328125,
+      "learning_rate": 1.7126436781609194e-07,
+      "loss": 0.0019,
+      "reward": 1.894079566001892,
+      "reward_std": 0.02045278623700142,
+      "rewards/accuracy_reward": 0.8940793871879578,
+      "rewards/format_reward": 1.0,
+      "step": 721
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.455078125,
+      "epoch": 8.298850574712644,
+      "grad_norm": 2.5331188317191087,
+      "kl": 0.04541015625,
+      "learning_rate": 1.7011494252873563e-07,
+      "loss": 0.0019,
+      "reward": 1.9165410995483398,
+      "reward_std": 0.017832912504673004,
+      "rewards/accuracy_reward": 0.9165410995483398,
+      "rewards/format_reward": 1.0,
+      "step": 722
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.32357025146484,
+      "epoch": 8.310344827586206,
+      "grad_norm": 4.02373195859424,
+      "kl": 0.0478515625,
+      "learning_rate": 1.689655172413793e-07,
+      "loss": 0.002,
+      "reward": 1.894669532775879,
+      "reward_std": 0.020446766167879105,
+      "rewards/accuracy_reward": 0.8946696519851685,
+      "rewards/format_reward": 1.0,
+      "step": 723
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.5,
+      "epoch": 8.32183908045977,
+      "grad_norm": 1.8288623944358686,
+      "kl": 0.051513671875,
+      "learning_rate": 1.67816091954023e-07,
+      "loss": 0.0021,
+      "reward": 1.9076260328292847,
+      "reward_std": 0.019866902381181717,
+      "rewards/accuracy_reward": 0.9082770347595215,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 724
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.72786712646484,
+      "epoch": 8.333333333333334,
+      "grad_norm": 5.4298852737123635,
+      "kl": 0.048095703125,
+      "learning_rate": 1.6666666666666665e-07,
+      "loss": 0.002,
+      "reward": 1.9031809568405151,
+      "reward_std": 0.02082451805472374,
+      "rewards/accuracy_reward": 0.903831958770752,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 725
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.61458587646484,
+      "epoch": 8.344827586206897,
+      "grad_norm": 2.2093808314572296,
+      "kl": 0.049072265625,
+      "learning_rate": 1.6551724137931034e-07,
+      "loss": 0.002,
+      "reward": 1.8967069387435913,
+      "reward_std": 0.018487486988306046,
+      "rewards/accuracy_reward": 0.8967069983482361,
+      "rewards/format_reward": 1.0,
+      "step": 726
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.33268737792969,
+      "epoch": 8.35632183908046,
+      "grad_norm": 1.6410241094216826,
+      "kl": 0.0478515625,
+      "learning_rate": 1.6436781609195403e-07,
+      "loss": 0.002,
+      "reward": 1.9131641387939453,
+      "reward_std": 0.016289152204990387,
+      "rewards/accuracy_reward": 0.9131642580032349,
+      "rewards/format_reward": 1.0,
+      "step": 727
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.259765625,
+      "epoch": 8.367816091954023,
+      "grad_norm": 1.7686301784709073,
+      "kl": 0.049560546875,
+      "learning_rate": 1.632183908045977e-07,
+      "loss": 0.002,
+      "reward": 1.9046128988265991,
+      "reward_std": 0.0201382078230381,
+      "rewards/accuracy_reward": 0.9046128988265991,
+      "rewards/format_reward": 1.0,
+      "step": 728
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.443359375,
+      "epoch": 8.379310344827585,
+      "grad_norm": 2.16179386604187,
+      "kl": 0.0458984375,
+      "learning_rate": 1.6206896551724136e-07,
+      "loss": 0.0019,
+      "reward": 1.8986444473266602,
+      "reward_std": 0.01659722626209259,
+      "rewards/accuracy_reward": 0.8986444473266602,
+      "rewards/format_reward": 1.0,
+      "step": 729
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.22721862792969,
+      "epoch": 8.39080459770115,
+      "grad_norm": 2.8856095934176857,
+      "kl": 0.048828125,
+      "learning_rate": 1.6091954022988505e-07,
+      "loss": 0.002,
+      "reward": 1.9096076488494873,
+      "reward_std": 0.018815144896507263,
+      "rewards/accuracy_reward": 0.9096077680587769,
+      "rewards/format_reward": 1.0,
+      "step": 730
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.619140625,
+      "epoch": 8.402298850574713,
+      "grad_norm": 2.9006809475096658,
+      "kl": 0.051513671875,
+      "learning_rate": 1.5977011494252874e-07,
+      "loss": 0.0022,
+      "reward": 1.9206702709197998,
+      "reward_std": 0.01633971370756626,
+      "rewards/accuracy_reward": 0.920670211315155,
+      "rewards/format_reward": 1.0,
+      "step": 731
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.259765625,
+      "epoch": 8.413793103448276,
+      "grad_norm": 1.9751932337603062,
+      "kl": 0.052734375,
+      "learning_rate": 1.5862068965517243e-07,
+      "loss": 0.0022,
+      "reward": 1.893907070159912,
+      "reward_std": 0.02224368415772915,
+      "rewards/accuracy_reward": 0.8939072489738464,
+      "rewards/format_reward": 1.0,
+      "step": 732
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.00846862792969,
+      "epoch": 8.425287356321839,
+      "grad_norm": 2.0957540123306364,
+      "kl": 0.0556640625,
+      "learning_rate": 1.5747126436781606e-07,
+      "loss": 0.0023,
+      "reward": 1.9105441570281982,
+      "reward_std": 0.01823435351252556,
+      "rewards/accuracy_reward": 0.9105440974235535,
+      "rewards/format_reward": 1.0,
+      "step": 733
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 103.10872650146484,
+      "epoch": 8.436781609195402,
+      "grad_norm": 1.7574530073426011,
+      "kl": 0.049072265625,
+      "learning_rate": 1.5632183908045975e-07,
+      "loss": 0.002,
+      "reward": 1.9063684940338135,
+      "reward_std": 0.017554182559251785,
+      "rewards/accuracy_reward": 0.9063684940338135,
+      "rewards/format_reward": 1.0,
+      "step": 734
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.40755462646484,
+      "epoch": 8.448275862068966,
+      "grad_norm": 1.718104369150053,
+      "kl": 0.04638671875,
+      "learning_rate": 1.5517241379310344e-07,
+      "loss": 0.0019,
+      "reward": 1.9011447429656982,
+      "reward_std": 0.02049892395734787,
+      "rewards/accuracy_reward": 0.9017957448959351,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 735
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.095703125,
+      "epoch": 8.459770114942529,
+      "grad_norm": 1.8096241235073485,
+      "kl": 0.044921875,
+      "learning_rate": 1.5402298850574713e-07,
+      "loss": 0.0019,
+      "reward": 1.9035829305648804,
+      "reward_std": 0.017755288630723953,
+      "rewards/accuracy_reward": 0.9035829305648804,
+      "rewards/format_reward": 1.0,
+      "step": 736
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.228515625,
+      "epoch": 8.471264367816092,
+      "grad_norm": 4.3156887737751966,
+      "kl": 0.04541015625,
+      "learning_rate": 1.528735632183908e-07,
+      "loss": 0.0019,
+      "reward": 1.8951524496078491,
+      "reward_std": 0.01714376173913479,
+      "rewards/accuracy_reward": 0.8951523303985596,
+      "rewards/format_reward": 1.0,
+      "step": 737
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.240234375,
+      "epoch": 8.482758620689655,
+      "grad_norm": 1.8203610430487198,
+      "kl": 0.046142578125,
+      "learning_rate": 1.5172413793103449e-07,
+      "loss": 0.0019,
+      "reward": 1.9012572765350342,
+      "reward_std": 0.020000826567411423,
+      "rewards/accuracy_reward": 0.9019083976745605,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 738
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.80078125,
+      "epoch": 8.494252873563218,
+      "grad_norm": 2.4493356638370547,
+      "kl": 0.046630859375,
+      "learning_rate": 1.5057471264367815e-07,
+      "loss": 0.002,
+      "reward": 1.8994574546813965,
+      "reward_std": 0.01900039240717888,
+      "rewards/accuracy_reward": 0.9001085758209229,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 739
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.68815612792969,
+      "epoch": 8.505747126436782,
+      "grad_norm": 1.7961165376254185,
+      "kl": 0.048583984375,
+      "learning_rate": 1.4942528735632184e-07,
+      "loss": 0.002,
+      "reward": 1.8904352188110352,
+      "reward_std": 0.018353838473558426,
+      "rewards/accuracy_reward": 0.8904353380203247,
+      "rewards/format_reward": 1.0,
+      "step": 740
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.22721862792969,
+      "epoch": 8.517241379310345,
+      "grad_norm": 2.124448648242551,
+      "kl": 0.048583984375,
+      "learning_rate": 1.482758620689655e-07,
+      "loss": 0.002,
+      "reward": 1.9015287160873413,
+      "reward_std": 0.017897652462124825,
+      "rewards/accuracy_reward": 0.9015287756919861,
+      "rewards/format_reward": 1.0,
+      "step": 741
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.91081237792969,
+      "epoch": 8.528735632183908,
+      "grad_norm": 1.5182426183899522,
+      "kl": 0.048095703125,
+      "learning_rate": 1.471264367816092e-07,
+      "loss": 0.002,
+      "reward": 1.9014829397201538,
+      "reward_std": 0.020442795008420944,
+      "rewards/accuracy_reward": 0.9021339416503906,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 742
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.91536712646484,
+      "epoch": 8.540229885057471,
+      "grad_norm": 1.5309622912667127,
+      "kl": 0.05029296875,
+      "learning_rate": 1.4597701149425288e-07,
+      "loss": 0.0021,
+      "reward": 1.9063705205917358,
+      "reward_std": 0.019261373206973076,
+      "rewards/accuracy_reward": 0.9070214629173279,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 743
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.28190612792969,
+      "epoch": 8.551724137931034,
+      "grad_norm": 1.4853282790643823,
+      "kl": 0.04296875,
+      "learning_rate": 1.4482758620689654e-07,
+      "loss": 0.0018,
+      "reward": 1.9033745527267456,
+      "reward_std": 0.01895163580775261,
+      "rewards/accuracy_reward": 0.904025673866272,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 744
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.767578125,
+      "epoch": 8.563218390804598,
+      "grad_norm": 1.5019420177279086,
+      "kl": 0.053466796875,
+      "learning_rate": 1.436781609195402e-07,
+      "loss": 0.0022,
+      "reward": 1.8998239040374756,
+      "reward_std": 0.021759074181318283,
+      "rewards/accuracy_reward": 0.9004749059677124,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 745
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.77018737792969,
+      "epoch": 8.574712643678161,
+      "grad_norm": 1.5414148798759049,
+      "kl": 0.04736328125,
+      "learning_rate": 1.425287356321839e-07,
+      "loss": 0.002,
+      "reward": 1.9003026485443115,
+      "reward_std": 0.021097760647535324,
+      "rewards/accuracy_reward": 0.9009537100791931,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 746
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.10872650146484,
+      "epoch": 8.586206896551724,
+      "grad_norm": 4.960221853934078,
+      "kl": 0.04345703125,
+      "learning_rate": 1.413793103448276e-07,
+      "loss": 0.0018,
+      "reward": 1.920248031616211,
+      "reward_std": 0.01949543133378029,
+      "rewards/accuracy_reward": 0.9208990931510925,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 747
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.27409362792969,
+      "epoch": 8.597701149425287,
+      "grad_norm": 1.464489207745014,
+      "kl": 0.044921875,
+      "learning_rate": 1.4022988505747128e-07,
+      "loss": 0.0019,
+      "reward": 1.9014743566513062,
+      "reward_std": 0.017846036702394485,
+      "rewards/accuracy_reward": 0.9014744162559509,
+      "rewards/format_reward": 1.0,
+      "step": 748
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.01302337646484,
+      "epoch": 8.60919540229885,
+      "grad_norm": 5.576587184664697,
+      "kl": 0.04443359375,
+      "learning_rate": 1.3908045977011494e-07,
+      "loss": 0.0018,
+      "reward": 1.8989577293395996,
+      "reward_std": 0.019103601574897766,
+      "rewards/accuracy_reward": 0.8989577293395996,
+      "rewards/format_reward": 1.0,
+      "step": 749
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.55208587646484,
+      "epoch": 8.620689655172415,
+      "grad_norm": 3.391688889257715,
+      "kl": 0.042236328125,
+      "learning_rate": 1.379310344827586e-07,
+      "loss": 0.0018,
+      "reward": 1.9022835493087769,
+      "reward_std": 0.020112819969654083,
+      "rewards/accuracy_reward": 0.9022834300994873,
+      "rewards/format_reward": 1.0,
+      "step": 750
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.23112487792969,
+      "epoch": 8.632183908045977,
+      "grad_norm": 1.415131831015453,
+      "kl": 0.046630859375,
+      "learning_rate": 1.367816091954023e-07,
+      "loss": 0.0019,
+      "reward": 1.8941152095794678,
+      "reward_std": 0.02364903688430786,
+      "rewards/accuracy_reward": 0.894115149974823,
+      "rewards/format_reward": 1.0,
+      "step": 751
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.39974212646484,
+      "epoch": 8.64367816091954,
+      "grad_norm": 3.686963959842639,
+      "kl": 0.043212890625,
+      "learning_rate": 1.3563218390804598e-07,
+      "loss": 0.0018,
+      "reward": 1.8956639766693115,
+      "reward_std": 0.01937917247414589,
+      "rewards/accuracy_reward": 0.8956639170646667,
+      "rewards/format_reward": 1.0,
+      "step": 752
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.33659362792969,
+      "epoch": 8.655172413793103,
+      "grad_norm": 11.793387997885002,
+      "kl": 0.04443359375,
+      "learning_rate": 1.3448275862068965e-07,
+      "loss": 0.0018,
+      "reward": 1.9092130661010742,
+      "reward_std": 0.020166244357824326,
+      "rewards/accuracy_reward": 0.9098643064498901,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 753
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.69075775146484,
+      "epoch": 8.666666666666666,
+      "grad_norm": 1.419014822654193,
+      "kl": 0.0517578125,
+      "learning_rate": 1.3333333333333334e-07,
+      "loss": 0.0021,
+      "reward": 1.8976190090179443,
+      "reward_std": 0.018913621082901955,
+      "rewards/accuracy_reward": 0.8976188898086548,
+      "rewards/format_reward": 1.0,
+      "step": 754
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.52799987792969,
+      "epoch": 8.678160919540229,
+      "grad_norm": 5.95529479625069,
+      "kl": 0.043212890625,
+      "learning_rate": 1.32183908045977e-07,
+      "loss": 0.0018,
+      "reward": 1.9074852466583252,
+      "reward_std": 0.017688315361738205,
+      "rewards/accuracy_reward": 0.9074852466583252,
+      "rewards/format_reward": 1.0,
+      "step": 755
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.25521087646484,
+      "epoch": 8.689655172413794,
+      "grad_norm": 2.7625404647260265,
+      "kl": 0.048828125,
+      "learning_rate": 1.310344827586207e-07,
+      "loss": 0.002,
+      "reward": 1.8992561101913452,
+      "reward_std": 0.021415000781416893,
+      "rewards/accuracy_reward": 0.899907112121582,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 756
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.72721862792969,
+      "epoch": 8.701149425287356,
+      "grad_norm": 3.132870047881132,
+      "kl": 0.04638671875,
+      "learning_rate": 1.2988505747126435e-07,
+      "loss": 0.0019,
+      "reward": 1.8953626155853271,
+      "reward_std": 0.01965472660958767,
+      "rewards/accuracy_reward": 0.8953627347946167,
+      "rewards/format_reward": 1.0,
+      "step": 757
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.06185150146484,
+      "epoch": 8.71264367816092,
+      "grad_norm": 1.4067186469745852,
+      "kl": 0.047119140625,
+      "learning_rate": 1.2873563218390804e-07,
+      "loss": 0.002,
+      "reward": 1.9052462577819824,
+      "reward_std": 0.019855869933962822,
+      "rewards/accuracy_reward": 0.9052464365959167,
+      "rewards/format_reward": 1.0,
+      "step": 758
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.248046875,
+      "epoch": 8.724137931034482,
+      "grad_norm": 1.6124971541846125,
+      "kl": 0.0478515625,
+      "learning_rate": 1.2758620689655173e-07,
+      "loss": 0.002,
+      "reward": 1.9074506759643555,
+      "reward_std": 0.017859049141407013,
+      "rewards/accuracy_reward": 0.9074506759643555,
+      "rewards/format_reward": 1.0,
+      "step": 759
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.123046875,
+      "epoch": 8.735632183908045,
+      "grad_norm": 1.6651646299458338,
+      "kl": 0.041259765625,
+      "learning_rate": 1.2643678160919542e-07,
+      "loss": 0.0017,
+      "reward": 1.8970210552215576,
+      "reward_std": 0.01788368821144104,
+      "rewards/accuracy_reward": 0.8970209956169128,
+      "rewards/format_reward": 1.0,
+      "step": 760
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.68099212646484,
+      "epoch": 8.74712643678161,
+      "grad_norm": 1.908276402264626,
+      "kl": 0.0478515625,
+      "learning_rate": 1.2528735632183906e-07,
+      "loss": 0.002,
+      "reward": 1.911945104598999,
+      "reward_std": 0.020722268149256706,
+      "rewards/accuracy_reward": 0.9125961065292358,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 761
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.97396087646484,
+      "epoch": 8.758620689655173,
+      "grad_norm": 1.909218993992126,
+      "kl": 0.04345703125,
+      "learning_rate": 1.2413793103448275e-07,
+      "loss": 0.0018,
+      "reward": 1.9106658697128296,
+      "reward_std": 0.017978399991989136,
+      "rewards/accuracy_reward": 0.9106658697128296,
+      "rewards/format_reward": 1.0,
+      "step": 762
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.734375,
+      "epoch": 8.770114942528735,
+      "grad_norm": 2.2467623211703014,
+      "kl": 0.045654296875,
+      "learning_rate": 1.2298850574712644e-07,
+      "loss": 0.0019,
+      "reward": 1.9027222394943237,
+      "reward_std": 0.016572915017604828,
+      "rewards/accuracy_reward": 0.9027222394943237,
+      "rewards/format_reward": 1.0,
+      "step": 763
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.62630462646484,
+      "epoch": 8.781609195402298,
+      "grad_norm": 5.599771626606617,
+      "kl": 0.04833984375,
+      "learning_rate": 1.218390804597701e-07,
+      "loss": 0.002,
+      "reward": 1.9221599102020264,
+      "reward_std": 0.016314268112182617,
+      "rewards/accuracy_reward": 0.9221599102020264,
+      "rewards/format_reward": 1.0,
+      "step": 764
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.12760925292969,
+      "epoch": 8.793103448275861,
+      "grad_norm": 2.4498609839357823,
+      "kl": 0.05078125,
+      "learning_rate": 1.206896551724138e-07,
+      "loss": 0.0021,
+      "reward": 1.9224801063537598,
+      "reward_std": 0.01809362694621086,
+      "rewards/accuracy_reward": 0.9224801063537598,
+      "rewards/format_reward": 1.0,
+      "step": 765
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.19857025146484,
+      "epoch": 8.804597701149426,
+      "grad_norm": 2.5801196829452584,
+      "kl": 0.0439453125,
+      "learning_rate": 1.1954022988505745e-07,
+      "loss": 0.0018,
+      "reward": 1.898393154144287,
+      "reward_std": 0.01924772374331951,
+      "rewards/accuracy_reward": 0.8983932733535767,
+      "rewards/format_reward": 1.0,
+      "step": 766
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.40690612792969,
+      "epoch": 8.816091954022989,
+      "grad_norm": 2.032967972913373,
+      "kl": 0.04345703125,
+      "learning_rate": 1.1839080459770114e-07,
+      "loss": 0.0018,
+      "reward": 1.9099705219268799,
+      "reward_std": 0.02100074663758278,
+      "rewards/accuracy_reward": 0.9106216430664062,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 767
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.61653900146484,
+      "epoch": 8.827586206896552,
+      "grad_norm": 2.07582370315509,
+      "kl": 0.046630859375,
+      "learning_rate": 1.1724137931034482e-07,
+      "loss": 0.002,
+      "reward": 1.9123599529266357,
+      "reward_std": 0.018274664878845215,
+      "rewards/accuracy_reward": 0.912359893321991,
+      "rewards/format_reward": 1.0,
+      "step": 768
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.04817962646484,
+      "epoch": 8.839080459770114,
+      "grad_norm": 2.1468683147084873,
+      "kl": 0.04638671875,
+      "learning_rate": 1.1609195402298851e-07,
+      "loss": 0.0019,
+      "reward": 1.905731439590454,
+      "reward_std": 0.021076953038573265,
+      "rewards/accuracy_reward": 0.9057313799858093,
+      "rewards/format_reward": 1.0,
+      "step": 769
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.3046875,
+      "epoch": 8.850574712643677,
+      "grad_norm": 1.5666594265928468,
+      "kl": 0.044677734375,
+      "learning_rate": 1.1494252873563217e-07,
+      "loss": 0.0019,
+      "reward": 1.9021600484848022,
+      "reward_std": 0.021117765456438065,
+      "rewards/accuracy_reward": 0.902160108089447,
+      "rewards/format_reward": 1.0,
+      "step": 770
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.53255462646484,
+      "epoch": 8.862068965517242,
+      "grad_norm": 3.9234683438946254,
+      "kl": 0.044677734375,
+      "learning_rate": 1.1379310344827586e-07,
+      "loss": 0.0019,
+      "reward": 1.9176257848739624,
+      "reward_std": 0.018359411507844925,
+      "rewards/accuracy_reward": 0.9176258444786072,
+      "rewards/format_reward": 1.0,
+      "step": 771
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.14974212646484,
+      "epoch": 8.873563218390805,
+      "grad_norm": 3.472231746273925,
+      "kl": 0.044189453125,
+      "learning_rate": 1.1264367816091953e-07,
+      "loss": 0.0018,
+      "reward": 1.9001665115356445,
+      "reward_std": 0.018229342997074127,
+      "rewards/accuracy_reward": 0.9001666903495789,
+      "rewards/format_reward": 1.0,
+      "step": 772
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.37825775146484,
+      "epoch": 8.885057471264368,
+      "grad_norm": 3.1028518982061635,
+      "kl": 0.046875,
+      "learning_rate": 1.1149425287356322e-07,
+      "loss": 0.0019,
+      "reward": 1.911559820175171,
+      "reward_std": 0.018318263813853264,
+      "rewards/accuracy_reward": 0.9115597605705261,
+      "rewards/format_reward": 1.0,
+      "step": 773
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.07487487792969,
+      "epoch": 8.89655172413793,
+      "grad_norm": 2.1968963906989276,
+      "kl": 0.047607421875,
+      "learning_rate": 1.103448275862069e-07,
+      "loss": 0.002,
+      "reward": 1.9014253616333008,
+      "reward_std": 0.02046075090765953,
+      "rewards/accuracy_reward": 0.9020764827728271,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 774
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.416015625,
+      "epoch": 8.908045977011493,
+      "grad_norm": 1.899936745200757,
+      "kl": 0.046630859375,
+      "learning_rate": 1.0919540229885057e-07,
+      "loss": 0.002,
+      "reward": 1.8940951824188232,
+      "reward_std": 0.017228955402970314,
+      "rewards/accuracy_reward": 0.8940951228141785,
+      "rewards/format_reward": 1.0,
+      "step": 775
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.01692962646484,
+      "epoch": 8.919540229885058,
+      "grad_norm": 2.350383361616655,
+      "kl": 0.046630859375,
+      "learning_rate": 1.0804597701149425e-07,
+      "loss": 0.002,
+      "reward": 1.8946069478988647,
+      "reward_std": 0.01980327069759369,
+      "rewards/accuracy_reward": 0.8946070075035095,
+      "rewards/format_reward": 1.0,
+      "step": 776
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.88411712646484,
+      "epoch": 8.931034482758621,
+      "grad_norm": 5.641269463386258,
+      "kl": 0.04296875,
+      "learning_rate": 1.0689655172413794e-07,
+      "loss": 0.0018,
+      "reward": 1.8930943012237549,
+      "reward_std": 0.020226947963237762,
+      "rewards/accuracy_reward": 0.8930944204330444,
+      "rewards/format_reward": 1.0,
+      "step": 777
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.46224212646484,
+      "epoch": 8.942528735632184,
+      "grad_norm": 1.590527535893143,
+      "kl": 0.04541015625,
+      "learning_rate": 1.057471264367816e-07,
+      "loss": 0.0019,
+      "reward": 1.9104032516479492,
+      "reward_std": 0.01798829436302185,
+      "rewards/accuracy_reward": 0.9104033708572388,
+      "rewards/format_reward": 1.0,
+      "step": 778
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.85612487792969,
+      "epoch": 8.954022988505747,
+      "grad_norm": 2.6107268057573365,
+      "kl": 0.048095703125,
+      "learning_rate": 1.0459770114942529e-07,
+      "loss": 0.002,
+      "reward": 1.9010858535766602,
+      "reward_std": 0.020739298313856125,
+      "rewards/accuracy_reward": 0.9010860323905945,
+      "rewards/format_reward": 1.0,
+      "step": 779
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.03385925292969,
+      "epoch": 8.96551724137931,
+      "grad_norm": 1.6092109607995282,
+      "kl": 0.053466796875,
+      "learning_rate": 1.0344827586206897e-07,
+      "loss": 0.0023,
+      "reward": 1.900530219078064,
+      "reward_std": 0.01735313981771469,
+      "rewards/accuracy_reward": 0.900530219078064,
+      "rewards/format_reward": 1.0,
+      "step": 780
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.65950775146484,
+      "epoch": 8.977011494252874,
+      "grad_norm": 1.4883968181068261,
+      "kl": 0.049560546875,
+      "learning_rate": 1.0229885057471264e-07,
+      "loss": 0.002,
+      "reward": 1.9065711498260498,
+      "reward_std": 0.016935113817453384,
+      "rewards/accuracy_reward": 0.9065712690353394,
+      "rewards/format_reward": 1.0,
+      "step": 781
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.84635925292969,
+      "epoch": 8.988505747126437,
+      "grad_norm": 1.6246833714697817,
+      "kl": 0.04931640625,
+      "learning_rate": 1.0114942528735632e-07,
+      "loss": 0.002,
+      "reward": 1.91986083984375,
+      "reward_std": 0.015720397233963013,
+      "rewards/accuracy_reward": 0.91986083984375,
+      "rewards/format_reward": 1.0,
+      "step": 782
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 99.25140380859375,
+      "epoch": 9.0,
+      "grad_norm": 2.248799852714355,
+      "kl": 0.041259765625,
+      "learning_rate": 1e-07,
+      "loss": 0.0017,
+      "reward": 1.899660587310791,
+      "reward_std": 0.019224824383854866,
+      "rewards/accuracy_reward": 0.8996607661247253,
+      "rewards/format_reward": 1.0,
+      "step": 783
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.44075775146484,
+      "epoch": 9.011494252873563,
+      "grad_norm": 1.8721483816760687,
+      "kl": 0.051513671875,
+      "learning_rate": 9.885057471264367e-08,
+      "loss": 0.0021,
+      "reward": 1.8844293355941772,
+      "reward_std": 0.025172477588057518,
+      "rewards/accuracy_reward": 0.8857313990592957,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 784
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.6640625,
+      "epoch": 9.022988505747126,
+      "grad_norm": 1.8506016770380465,
+      "kl": 0.04150390625,
+      "learning_rate": 9.770114942528736e-08,
+      "loss": 0.0017,
+      "reward": 1.911178469657898,
+      "reward_std": 0.021104078739881516,
+      "rewards/accuracy_reward": 0.911178469657898,
+      "rewards/format_reward": 1.0,
+      "step": 785
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.47526550292969,
+      "epoch": 9.03448275862069,
+      "grad_norm": 2.4400240165615634,
+      "kl": 0.04248046875,
+      "learning_rate": 9.655172413793103e-08,
+      "loss": 0.0018,
+      "reward": 1.9191011190414429,
+      "reward_std": 0.016874711960554123,
+      "rewards/accuracy_reward": 0.9191012382507324,
+      "rewards/format_reward": 1.0,
+      "step": 786
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.58268737792969,
+      "epoch": 9.045977011494253,
+      "grad_norm": 3.3949879642064476,
+      "kl": 0.04150390625,
+      "learning_rate": 9.540229885057471e-08,
+      "loss": 0.0017,
+      "reward": 1.8661808967590332,
+      "reward_std": 0.02081323228776455,
+      "rewards/accuracy_reward": 0.8661808967590332,
+      "rewards/format_reward": 1.0,
+      "step": 787
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.15234375,
+      "epoch": 9.057471264367816,
+      "grad_norm": 1.407910897768254,
+      "kl": 0.051025390625,
+      "learning_rate": 9.425287356321839e-08,
+      "loss": 0.0021,
+      "reward": 1.8680272102355957,
+      "reward_std": 0.02056352235376835,
+      "rewards/accuracy_reward": 0.8680272102355957,
+      "rewards/format_reward": 1.0,
+      "step": 788
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.87956237792969,
+      "epoch": 9.068965517241379,
+      "grad_norm": 2.6980150332453334,
+      "kl": 0.04443359375,
+      "learning_rate": 9.310344827586207e-08,
+      "loss": 0.0018,
+      "reward": 1.885411262512207,
+      "reward_std": 0.021717345342040062,
+      "rewards/accuracy_reward": 0.885411262512207,
+      "rewards/format_reward": 1.0,
+      "step": 789
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.94010925292969,
+      "epoch": 9.080459770114942,
+      "grad_norm": 1.6808051261990247,
+      "kl": 0.05615234375,
+      "learning_rate": 9.195402298850574e-08,
+      "loss": 0.0023,
+      "reward": 1.901023268699646,
+      "reward_std": 0.021362271159887314,
+      "rewards/accuracy_reward": 0.9010233879089355,
+      "rewards/format_reward": 1.0,
+      "step": 790
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.22721862792969,
+      "epoch": 9.091954022988507,
+      "grad_norm": 1.9651267648591069,
+      "kl": 0.048828125,
+      "learning_rate": 9.080459770114942e-08,
+      "loss": 0.002,
+      "reward": 1.8997548818588257,
+      "reward_std": 0.018945304676890373,
+      "rewards/accuracy_reward": 0.8997548818588257,
+      "rewards/format_reward": 1.0,
+      "step": 791
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.98046875,
+      "epoch": 9.10344827586207,
+      "grad_norm": 1.5615310862319665,
+      "kl": 0.0458984375,
+      "learning_rate": 8.96551724137931e-08,
+      "loss": 0.0019,
+      "reward": 1.9096624851226807,
+      "reward_std": 0.018772246316075325,
+      "rewards/accuracy_reward": 0.9096624851226807,
+      "rewards/format_reward": 1.0,
+      "step": 792
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.50456237792969,
+      "epoch": 9.114942528735632,
+      "grad_norm": 1.8984567348891395,
+      "kl": 0.053466796875,
+      "learning_rate": 8.850574712643679e-08,
+      "loss": 0.0022,
+      "reward": 1.8971011638641357,
+      "reward_std": 0.02134164236485958,
+      "rewards/accuracy_reward": 0.8971012830734253,
+      "rewards/format_reward": 1.0,
+      "step": 793
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.59765625,
+      "epoch": 9.126436781609195,
+      "grad_norm": 1.57123496870828,
+      "kl": 0.05126953125,
+      "learning_rate": 8.735632183908045e-08,
+      "loss": 0.0021,
+      "reward": 1.8993648290634155,
+      "reward_std": 0.02058519423007965,
+      "rewards/accuracy_reward": 0.8993649482727051,
+      "rewards/format_reward": 1.0,
+      "step": 794
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.01692962646484,
+      "epoch": 9.137931034482758,
+      "grad_norm": 2.1980038174190373,
+      "kl": 0.038818359375,
+      "learning_rate": 8.620689655172414e-08,
+      "loss": 0.0016,
+      "reward": 1.8727396726608276,
+      "reward_std": 0.021485071629285812,
+      "rewards/accuracy_reward": 0.8727396726608276,
+      "rewards/format_reward": 1.0,
+      "step": 795
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.57682800292969,
+      "epoch": 9.149425287356323,
+      "grad_norm": 1.5298307460022529,
+      "kl": 0.044921875,
+      "learning_rate": 8.505747126436782e-08,
+      "loss": 0.0019,
+      "reward": 1.9137659072875977,
+      "reward_std": 0.01766360178589821,
+      "rewards/accuracy_reward": 0.9137659072875977,
+      "rewards/format_reward": 1.0,
+      "step": 796
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.77474212646484,
+      "epoch": 9.160919540229886,
+      "grad_norm": 1.801682202262651,
+      "kl": 0.04296875,
+      "learning_rate": 8.39080459770115e-08,
+      "loss": 0.0018,
+      "reward": 1.8823485374450684,
+      "reward_std": 0.023980390280485153,
+      "rewards/accuracy_reward": 0.8829997777938843,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 797
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.94596862792969,
+      "epoch": 9.172413793103448,
+      "grad_norm": 2.2238602851803284,
+      "kl": 0.041015625,
+      "learning_rate": 8.275862068965517e-08,
+      "loss": 0.0017,
+      "reward": 1.8901784420013428,
+      "reward_std": 0.021750375628471375,
+      "rewards/accuracy_reward": 0.8908295631408691,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 798
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.58464050292969,
+      "epoch": 9.183908045977011,
+      "grad_norm": 3.448825022079825,
+      "kl": 0.042724609375,
+      "learning_rate": 8.160919540229885e-08,
+      "loss": 0.0018,
+      "reward": 1.902247428894043,
+      "reward_std": 0.020935572683811188,
+      "rewards/accuracy_reward": 0.9028984904289246,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 799
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.63216400146484,
+      "epoch": 9.195402298850574,
+      "grad_norm": 1.7683408378496788,
+      "kl": 0.041259765625,
+      "learning_rate": 8.045977011494252e-08,
+      "loss": 0.0017,
+      "reward": 1.8915985822677612,
+      "reward_std": 0.02043468877673149,
+      "rewards/accuracy_reward": 0.891598641872406,
+      "rewards/format_reward": 1.0,
+      "step": 800
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.3046875,
+      "epoch": 9.206896551724139,
+      "grad_norm": 3.4479135032415082,
+      "kl": 0.0419921875,
+      "learning_rate": 7.931034482758621e-08,
+      "loss": 0.0018,
+      "reward": 1.8980789184570312,
+      "reward_std": 0.020334240049123764,
+      "rewards/accuracy_reward": 0.8980789184570312,
+      "rewards/format_reward": 1.0,
+      "step": 801
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.453125,
+      "epoch": 9.218390804597702,
+      "grad_norm": 2.8315233473682286,
+      "kl": 0.0439453125,
+      "learning_rate": 7.816091954022988e-08,
+      "loss": 0.0018,
+      "reward": 1.9106957912445068,
+      "reward_std": 0.01965712383389473,
+      "rewards/accuracy_reward": 0.9106957912445068,
+      "rewards/format_reward": 1.0,
+      "step": 802
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.0078125,
+      "epoch": 9.229885057471265,
+      "grad_norm": 8.809973822334465,
+      "kl": 0.04443359375,
+      "learning_rate": 7.701149425287357e-08,
+      "loss": 0.0018,
+      "reward": 1.8973687887191772,
+      "reward_std": 0.02227511629462242,
+      "rewards/accuracy_reward": 0.8973686695098877,
+      "rewards/format_reward": 1.0,
+      "step": 803
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.857421875,
+      "epoch": 9.241379310344827,
+      "grad_norm": 2.290511339554376,
+      "kl": 0.048095703125,
+      "learning_rate": 7.586206896551724e-08,
+      "loss": 0.002,
+      "reward": 1.926220417022705,
+      "reward_std": 0.020036086440086365,
+      "rewards/accuracy_reward": 0.9268714785575867,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 804
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.24153900146484,
+      "epoch": 9.25287356321839,
+      "grad_norm": 10.259694200109895,
+      "kl": 0.04541015625,
+      "learning_rate": 7.471264367816092e-08,
+      "loss": 0.0019,
+      "reward": 1.9228681325912476,
+      "reward_std": 0.02166038751602173,
+      "rewards/accuracy_reward": 0.9228681325912476,
+      "rewards/format_reward": 1.0,
+      "step": 805
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.47786712646484,
+      "epoch": 9.264367816091955,
+      "grad_norm": 1.888552839892748,
+      "kl": 0.040771484375,
+      "learning_rate": 7.35632183908046e-08,
+      "loss": 0.0017,
+      "reward": 1.905164122581482,
+      "reward_std": 0.017032243311405182,
+      "rewards/accuracy_reward": 0.9051642417907715,
+      "rewards/format_reward": 1.0,
+      "step": 806
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.90560150146484,
+      "epoch": 9.275862068965518,
+      "grad_norm": 1.5511138521125212,
+      "kl": 0.052001953125,
+      "learning_rate": 7.241379310344827e-08,
+      "loss": 0.0021,
+      "reward": 1.912177562713623,
+      "reward_std": 0.018603047356009483,
+      "rewards/accuracy_reward": 0.912177562713623,
+      "rewards/format_reward": 1.0,
+      "step": 807
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.09114837646484,
+      "epoch": 9.28735632183908,
+      "grad_norm": 2.2223963738104775,
+      "kl": 0.04443359375,
+      "learning_rate": 7.126436781609195e-08,
+      "loss": 0.0018,
+      "reward": 1.9001119136810303,
+      "reward_std": 0.023462196812033653,
+      "rewards/accuracy_reward": 0.9001118540763855,
+      "rewards/format_reward": 1.0,
+      "step": 808
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.89388275146484,
+      "epoch": 9.298850574712644,
+      "grad_norm": 6.250219527341955,
+      "kl": 0.046142578125,
+      "learning_rate": 7.011494252873564e-08,
+      "loss": 0.0019,
+      "reward": 1.9060745239257812,
+      "reward_std": 0.021167151629924774,
+      "rewards/accuracy_reward": 0.9067255258560181,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 809
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.06120300292969,
+      "epoch": 9.310344827586206,
+      "grad_norm": 3.0663601968471044,
+      "kl": 0.04248046875,
+      "learning_rate": 6.89655172413793e-08,
+      "loss": 0.0018,
+      "reward": 1.904802680015564,
+      "reward_std": 0.02016662247478962,
+      "rewards/accuracy_reward": 0.9054538011550903,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 810
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.279296875,
+      "epoch": 9.32183908045977,
+      "grad_norm": 2.1266705913854707,
+      "kl": 0.06591796875,
+      "learning_rate": 6.781609195402299e-08,
+      "loss": 0.0027,
+      "reward": 1.8952583074569702,
+      "reward_std": 0.023519501090049744,
+      "rewards/accuracy_reward": 0.895909309387207,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 811
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.70573425292969,
+      "epoch": 9.333333333333334,
+      "grad_norm": 1.2920255105728329,
+      "kl": 0.041259765625,
+      "learning_rate": 6.666666666666667e-08,
+      "loss": 0.0017,
+      "reward": 1.9121431112289429,
+      "reward_std": 0.01818576082587242,
+      "rewards/accuracy_reward": 0.9121431112289429,
+      "rewards/format_reward": 1.0,
+      "step": 812
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.04948425292969,
+      "epoch": 9.344827586206897,
+      "grad_norm": 1.9698975522109214,
+      "kl": 0.044189453125,
+      "learning_rate": 6.551724137931034e-08,
+      "loss": 0.0018,
+      "reward": 1.9061200618743896,
+      "reward_std": 0.019229721277952194,
+      "rewards/accuracy_reward": 0.9067711234092712,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 813
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.99479675292969,
+      "epoch": 9.35632183908046,
+      "grad_norm": 1.5466544483838331,
+      "kl": 0.044677734375,
+      "learning_rate": 6.436781609195402e-08,
+      "loss": 0.0019,
+      "reward": 1.8965463638305664,
+      "reward_std": 0.026135286316275597,
+      "rewards/accuracy_reward": 0.8978484869003296,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 814
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.06771087646484,
+      "epoch": 9.367816091954023,
+      "grad_norm": 2.9008920163167327,
+      "kl": 0.04345703125,
+      "learning_rate": 6.321839080459771e-08,
+      "loss": 0.0018,
+      "reward": 1.8903131484985352,
+      "reward_std": 0.020935822278261185,
+      "rewards/accuracy_reward": 0.8903131484985352,
+      "rewards/format_reward": 1.0,
+      "step": 815
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.97396087646484,
+      "epoch": 9.379310344827585,
+      "grad_norm": 2.2134917020578797,
+      "kl": 0.043212890625,
+      "learning_rate": 6.206896551724137e-08,
+      "loss": 0.0018,
+      "reward": 1.894513487815857,
+      "reward_std": 0.019658654928207397,
+      "rewards/accuracy_reward": 0.8945134878158569,
+      "rewards/format_reward": 1.0,
+      "step": 816
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.47135925292969,
+      "epoch": 9.39080459770115,
+      "grad_norm": 2.0451132122580216,
+      "kl": 0.044189453125,
+      "learning_rate": 6.091954022988505e-08,
+      "loss": 0.0019,
+      "reward": 1.9205522537231445,
+      "reward_std": 0.022365611046552658,
+      "rewards/accuracy_reward": 0.9205522537231445,
+      "rewards/format_reward": 1.0,
+      "step": 817
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.357421875,
+      "epoch": 9.402298850574713,
+      "grad_norm": 1.5413854537310678,
+      "kl": 0.04345703125,
+      "learning_rate": 5.977011494252873e-08,
+      "loss": 0.0018,
+      "reward": 1.906002402305603,
+      "reward_std": 0.020225465297698975,
+      "rewards/accuracy_reward": 0.9060024619102478,
+      "rewards/format_reward": 1.0,
+      "step": 818
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.248046875,
+      "epoch": 9.413793103448276,
+      "grad_norm": 1.5942001444003233,
+      "kl": 0.040771484375,
+      "learning_rate": 5.862068965517241e-08,
+      "loss": 0.0017,
+      "reward": 1.9052398204803467,
+      "reward_std": 0.018013250082731247,
+      "rewards/accuracy_reward": 0.9052395820617676,
+      "rewards/format_reward": 1.0,
+      "step": 819
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.83528900146484,
+      "epoch": 9.425287356321839,
+      "grad_norm": 2.321727658388936,
+      "kl": 0.04248046875,
+      "learning_rate": 5.747126436781609e-08,
+      "loss": 0.0018,
+      "reward": 1.9112927913665771,
+      "reward_std": 0.020977940410375595,
+      "rewards/accuracy_reward": 0.911943793296814,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 820
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.5625,
+      "epoch": 9.436781609195402,
+      "grad_norm": 3.718818284558716,
+      "kl": 0.0458984375,
+      "learning_rate": 5.6321839080459764e-08,
+      "loss": 0.002,
+      "reward": 1.9173364639282227,
+      "reward_std": 0.01930234208703041,
+      "rewards/accuracy_reward": 0.9173363447189331,
+      "rewards/format_reward": 1.0,
+      "step": 821
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 108.20833587646484,
+      "epoch": 9.448275862068966,
+      "grad_norm": 1.7447738271778865,
+      "kl": 0.03857421875,
+      "learning_rate": 5.517241379310345e-08,
+      "loss": 0.0016,
+      "reward": 1.905761480331421,
+      "reward_std": 0.017823033034801483,
+      "rewards/accuracy_reward": 0.9057614207267761,
+      "rewards/format_reward": 1.0,
+      "step": 822
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 109.22005462646484,
+      "epoch": 9.459770114942529,
+      "grad_norm": 5.556438363038334,
+      "kl": 0.044921875,
+      "learning_rate": 5.402298850574712e-08,
+      "loss": 0.0019,
+      "reward": 1.912406325340271,
+      "reward_std": 0.017947331070899963,
+      "rewards/accuracy_reward": 0.912406325340271,
+      "rewards/format_reward": 1.0,
+      "step": 823
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.21354675292969,
+      "epoch": 9.471264367816092,
+      "grad_norm": 1.564280665225921,
+      "kl": 0.04296875,
+      "learning_rate": 5.28735632183908e-08,
+      "loss": 0.0018,
+      "reward": 1.9196040630340576,
+      "reward_std": 0.018538126721978188,
+      "rewards/accuracy_reward": 0.9196040034294128,
+      "rewards/format_reward": 1.0,
+      "step": 824
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.55729675292969,
+      "epoch": 9.482758620689655,
+      "grad_norm": 2.4144959847961758,
+      "kl": 0.0439453125,
+      "learning_rate": 5.172413793103448e-08,
+      "loss": 0.0019,
+      "reward": 1.9196009635925293,
+      "reward_std": 0.016086673364043236,
+      "rewards/accuracy_reward": 0.9196010828018188,
+      "rewards/format_reward": 1.0,
+      "step": 825
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.38021087646484,
+      "epoch": 9.494252873563218,
+      "grad_norm": 4.197005794788358,
+      "kl": 0.04345703125,
+      "learning_rate": 5.057471264367816e-08,
+      "loss": 0.0019,
+      "reward": 1.898109793663025,
+      "reward_std": 0.021654874086380005,
+      "rewards/accuracy_reward": 0.8987607955932617,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 826
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.59310150146484,
+      "epoch": 9.505747126436782,
+      "grad_norm": 1.8296996407650712,
+      "kl": 0.04345703125,
+      "learning_rate": 4.9425287356321836e-08,
+      "loss": 0.0018,
+      "reward": 1.9078954458236694,
+      "reward_std": 0.018013805150985718,
+      "rewards/accuracy_reward": 0.907895565032959,
+      "rewards/format_reward": 1.0,
+      "step": 827
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.60677337646484,
+      "epoch": 9.517241379310345,
+      "grad_norm": 1.6814305912226317,
+      "kl": 0.041259765625,
+      "learning_rate": 4.827586206896551e-08,
+      "loss": 0.0017,
+      "reward": 1.8963592052459717,
+      "reward_std": 0.022151313722133636,
+      "rewards/accuracy_reward": 0.8963589668273926,
+      "rewards/format_reward": 1.0,
+      "step": 828
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.73567962646484,
+      "epoch": 9.528735632183908,
+      "grad_norm": 1.7924163077233033,
+      "kl": 0.04296875,
+      "learning_rate": 4.7126436781609196e-08,
+      "loss": 0.0018,
+      "reward": 1.9191563129425049,
+      "reward_std": 0.01933225616812706,
+      "rewards/accuracy_reward": 0.9198073744773865,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 829
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.1171875,
+      "epoch": 9.540229885057471,
+      "grad_norm": 2.5340920833447895,
+      "kl": 0.0439453125,
+      "learning_rate": 4.597701149425287e-08,
+      "loss": 0.0018,
+      "reward": 1.89328932762146,
+      "reward_std": 0.01949036866426468,
+      "rewards/accuracy_reward": 0.8939403295516968,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 830
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.53255462646484,
+      "epoch": 9.551724137931034,
+      "grad_norm": 2.483443690558432,
+      "kl": 0.052490234375,
+      "learning_rate": 4.482758620689655e-08,
+      "loss": 0.0022,
+      "reward": 1.8969571590423584,
+      "reward_std": 0.017558250576257706,
+      "rewards/accuracy_reward": 0.8969571590423584,
+      "rewards/format_reward": 1.0,
+      "step": 831
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.533203125,
+      "epoch": 9.563218390804598,
+      "grad_norm": 1.291578178291959,
+      "kl": 0.045166015625,
+      "learning_rate": 4.3678160919540225e-08,
+      "loss": 0.0019,
+      "reward": 1.8948633670806885,
+      "reward_std": 0.021366355940699577,
+      "rewards/accuracy_reward": 0.8948633074760437,
+      "rewards/format_reward": 1.0,
+      "step": 832
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.82487487792969,
+      "epoch": 9.574712643678161,
+      "grad_norm": 2.7537847562643374,
+      "kl": 0.0419921875,
+      "learning_rate": 4.252873563218391e-08,
+      "loss": 0.0017,
+      "reward": 1.9019429683685303,
+      "reward_std": 0.018435677513480186,
+      "rewards/accuracy_reward": 0.9019430875778198,
+      "rewards/format_reward": 1.0,
+      "step": 833
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.77799987792969,
+      "epoch": 9.586206896551724,
+      "grad_norm": 2.2181518752294624,
+      "kl": 0.04443359375,
+      "learning_rate": 4.1379310344827585e-08,
+      "loss": 0.0018,
+      "reward": 1.8903684616088867,
+      "reward_std": 0.018676333129405975,
+      "rewards/accuracy_reward": 0.8903685808181763,
+      "rewards/format_reward": 1.0,
+      "step": 834
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.46224212646484,
+      "epoch": 9.597701149425287,
+      "grad_norm": 1.756566376667342,
+      "kl": 0.046875,
+      "learning_rate": 4.022988505747126e-08,
+      "loss": 0.0019,
+      "reward": 1.8650926351547241,
+      "reward_std": 0.02378205582499504,
+      "rewards/accuracy_reward": 0.8657435178756714,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 835
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.02474212646484,
+      "epoch": 9.60919540229885,
+      "grad_norm": 1.99335590706468,
+      "kl": 0.04345703125,
+      "learning_rate": 3.908045977011494e-08,
+      "loss": 0.0018,
+      "reward": 1.8908905982971191,
+      "reward_std": 0.01960897445678711,
+      "rewards/accuracy_reward": 0.8915416598320007,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 836
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.8046875,
+      "epoch": 9.620689655172415,
+      "grad_norm": 4.785304870304732,
+      "kl": 0.04345703125,
+      "learning_rate": 3.793103448275862e-08,
+      "loss": 0.0018,
+      "reward": 1.909947395324707,
+      "reward_std": 0.0192283745855093,
+      "rewards/accuracy_reward": 0.9099475741386414,
+      "rewards/format_reward": 1.0,
+      "step": 837
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.50456237792969,
+      "epoch": 9.632183908045977,
+      "grad_norm": 1.6423863350528298,
+      "kl": 0.04736328125,
+      "learning_rate": 3.67816091954023e-08,
+      "loss": 0.0019,
+      "reward": 1.9007571935653687,
+      "reward_std": 0.017737818881869316,
+      "rewards/accuracy_reward": 0.9007573127746582,
+      "rewards/format_reward": 1.0,
+      "step": 838
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.982421875,
+      "epoch": 9.64367816091954,
+      "grad_norm": 2.049656890092837,
+      "kl": 0.043701171875,
+      "learning_rate": 3.5632183908045974e-08,
+      "loss": 0.0018,
+      "reward": 1.9175758361816406,
+      "reward_std": 0.01626306213438511,
+      "rewards/accuracy_reward": 0.9175758361816406,
+      "rewards/format_reward": 1.0,
+      "step": 839
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.10417175292969,
+      "epoch": 9.655172413793103,
+      "grad_norm": 1.8848940408267045,
+      "kl": 0.045654296875,
+      "learning_rate": 3.448275862068965e-08,
+      "loss": 0.0019,
+      "reward": 1.9053196907043457,
+      "reward_std": 0.022480791434645653,
+      "rewards/accuracy_reward": 0.9059707522392273,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 840
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.12890625,
+      "epoch": 9.666666666666666,
+      "grad_norm": 3.234936396417619,
+      "kl": 0.04736328125,
+      "learning_rate": 3.3333333333333334e-08,
+      "loss": 0.0019,
+      "reward": 1.905747413635254,
+      "reward_std": 0.02039037086069584,
+      "rewards/accuracy_reward": 0.9057474136352539,
+      "rewards/format_reward": 1.0,
+      "step": 841
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.443359375,
+      "epoch": 9.678160919540229,
+      "grad_norm": 1.7322466630483822,
+      "kl": 0.04443359375,
+      "learning_rate": 3.218390804597701e-08,
+      "loss": 0.0018,
+      "reward": 1.906328558921814,
+      "reward_std": 0.022260412573814392,
+      "rewards/accuracy_reward": 0.9076308012008667,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 842
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.857421875,
+      "epoch": 9.689655172413794,
+      "grad_norm": 1.7097166944224358,
+      "kl": 0.05224609375,
+      "learning_rate": 3.103448275862069e-08,
+      "loss": 0.0022,
+      "reward": 1.9105417728424072,
+      "reward_std": 0.017679734155535698,
+      "rewards/accuracy_reward": 0.9105417728424072,
+      "rewards/format_reward": 1.0,
+      "step": 843
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.515625,
+      "epoch": 9.701149425287356,
+      "grad_norm": 1.884287958847617,
+      "kl": 0.047119140625,
+      "learning_rate": 2.9885057471264364e-08,
+      "loss": 0.002,
+      "reward": 1.9169318675994873,
+      "reward_std": 0.020296216011047363,
+      "rewards/accuracy_reward": 0.9182338714599609,
+      "rewards/format_reward": 0.9986979365348816,
+      "step": 844
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.80143737792969,
+      "epoch": 9.71264367816092,
+      "grad_norm": 1.9163910071570234,
+      "kl": 0.0458984375,
+      "learning_rate": 2.8735632183908043e-08,
+      "loss": 0.002,
+      "reward": 1.9143104553222656,
+      "reward_std": 0.019828440621495247,
+      "rewards/accuracy_reward": 0.9149616956710815,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 845
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.53190612792969,
+      "epoch": 9.724137931034482,
+      "grad_norm": 1.73873304639946,
+      "kl": 0.0439453125,
+      "learning_rate": 2.7586206896551723e-08,
+      "loss": 0.0019,
+      "reward": 1.9067649841308594,
+      "reward_std": 0.018557554110884666,
+      "rewards/accuracy_reward": 0.9067651033401489,
+      "rewards/format_reward": 1.0,
+      "step": 846
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.27604675292969,
+      "epoch": 9.735632183908045,
+      "grad_norm": 2.3263503442259807,
+      "kl": 0.04345703125,
+      "learning_rate": 2.64367816091954e-08,
+      "loss": 0.0018,
+      "reward": 1.9176081418991089,
+      "reward_std": 0.018732137978076935,
+      "rewards/accuracy_reward": 0.9176082015037537,
+      "rewards/format_reward": 1.0,
+      "step": 847
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.03125,
+      "epoch": 9.74712643678161,
+      "grad_norm": 1.495014646664977,
+      "kl": 0.04931640625,
+      "learning_rate": 2.528735632183908e-08,
+      "loss": 0.0021,
+      "reward": 1.914261817932129,
+      "reward_std": 0.01746547594666481,
+      "rewards/accuracy_reward": 0.9142619371414185,
+      "rewards/format_reward": 1.0,
+      "step": 848
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.43099212646484,
+      "epoch": 9.758620689655173,
+      "grad_norm": 2.129213121110771,
+      "kl": 0.044189453125,
+      "learning_rate": 2.4137931034482756e-08,
+      "loss": 0.0018,
+      "reward": 1.9081742763519287,
+      "reward_std": 0.02037705108523369,
+      "rewards/accuracy_reward": 0.9088253378868103,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 849
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.64453125,
+      "epoch": 9.770114942528735,
+      "grad_norm": 1.5696963468743126,
+      "kl": 0.046142578125,
+      "learning_rate": 2.2988505747126436e-08,
+      "loss": 0.0019,
+      "reward": 1.9174790382385254,
+      "reward_std": 0.017753083258867264,
+      "rewards/accuracy_reward": 0.9174790382385254,
+      "rewards/format_reward": 1.0,
+      "step": 850
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.01302337646484,
+      "epoch": 9.781609195402298,
+      "grad_norm": 2.3977935164505286,
+      "kl": 0.050537109375,
+      "learning_rate": 2.1839080459770113e-08,
+      "loss": 0.0021,
+      "reward": 1.8931522369384766,
+      "reward_std": 0.019253626465797424,
+      "rewards/accuracy_reward": 0.8931522369384766,
+      "rewards/format_reward": 1.0,
+      "step": 851
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.55339050292969,
+      "epoch": 9.793103448275861,
+      "grad_norm": 2.3423979754131383,
+      "kl": 0.04541015625,
+      "learning_rate": 2.0689655172413793e-08,
+      "loss": 0.0019,
+      "reward": 1.9183381795883179,
+      "reward_std": 0.019186435267329216,
+      "rewards/accuracy_reward": 0.9183380603790283,
+      "rewards/format_reward": 1.0,
+      "step": 852
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.85417175292969,
+      "epoch": 9.804597701149426,
+      "grad_norm": 2.073174318294833,
+      "kl": 0.046630859375,
+      "learning_rate": 1.954022988505747e-08,
+      "loss": 0.0019,
+      "reward": 1.890181303024292,
+      "reward_std": 0.018408436328172684,
+      "rewards/accuracy_reward": 0.8901811838150024,
+      "rewards/format_reward": 1.0,
+      "step": 853
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.06640625,
+      "epoch": 9.816091954022989,
+      "grad_norm": 1.9299645375268184,
+      "kl": 0.05078125,
+      "learning_rate": 1.839080459770115e-08,
+      "loss": 0.0021,
+      "reward": 1.9052647352218628,
+      "reward_std": 0.017054375261068344,
+      "rewards/accuracy_reward": 0.9052648544311523,
+      "rewards/format_reward": 1.0,
+      "step": 854
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.09896087646484,
+      "epoch": 9.827586206896552,
+      "grad_norm": 1.4804834876534272,
+      "kl": 0.045166015625,
+      "learning_rate": 1.7241379310344825e-08,
+      "loss": 0.0019,
+      "reward": 1.9028480052947998,
+      "reward_std": 0.017569242045283318,
+      "rewards/accuracy_reward": 0.9028478860855103,
+      "rewards/format_reward": 1.0,
+      "step": 855
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.41862487792969,
+      "epoch": 9.839080459770114,
+      "grad_norm": 3.920206306053486,
+      "kl": 0.049560546875,
+      "learning_rate": 1.6091954022988505e-08,
+      "loss": 0.0021,
+      "reward": 1.9126800298690796,
+      "reward_std": 0.018128346651792526,
+      "rewards/accuracy_reward": 0.913331151008606,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 856
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.01432800292969,
+      "epoch": 9.850574712643677,
+      "grad_norm": 3.684344553078877,
+      "kl": 0.046875,
+      "learning_rate": 1.4942528735632182e-08,
+      "loss": 0.0019,
+      "reward": 1.8925762176513672,
+      "reward_std": 0.021131232380867004,
+      "rewards/accuracy_reward": 0.8925762176513672,
+      "rewards/format_reward": 1.0,
+      "step": 857
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.326171875,
+      "epoch": 9.862068965517242,
+      "grad_norm": 1.3668989587295126,
+      "kl": 0.04833984375,
+      "learning_rate": 1.3793103448275862e-08,
+      "loss": 0.002,
+      "reward": 1.8880128860473633,
+      "reward_std": 0.020463991910219193,
+      "rewards/accuracy_reward": 0.8880130648612976,
+      "rewards/format_reward": 1.0,
+      "step": 858
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.76692962646484,
+      "epoch": 9.873563218390805,
+      "grad_norm": 2.38568885867465,
+      "kl": 0.045166015625,
+      "learning_rate": 1.264367816091954e-08,
+      "loss": 0.0019,
+      "reward": 1.9086008071899414,
+      "reward_std": 0.020096510648727417,
+      "rewards/accuracy_reward": 0.9086008071899414,
+      "rewards/format_reward": 1.0,
+      "step": 859
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.86198425292969,
+      "epoch": 9.885057471264368,
+      "grad_norm": 4.216482284140485,
+      "kl": 0.1513671875,
+      "learning_rate": 1.1494252873563218e-08,
+      "loss": 0.0061,
+      "reward": 1.9235996007919312,
+      "reward_std": 0.019205166026949883,
+      "rewards/accuracy_reward": 0.9242507815361023,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 860
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.77409362792969,
+      "epoch": 9.89655172413793,
+      "grad_norm": 4.8514292391747675,
+      "kl": 0.0498046875,
+      "learning_rate": 1.0344827586206896e-08,
+      "loss": 0.0021,
+      "reward": 1.9049177169799805,
+      "reward_std": 0.017066650092601776,
+      "rewards/accuracy_reward": 0.90491783618927,
+      "rewards/format_reward": 1.0,
+      "step": 861
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.80599212646484,
+      "epoch": 9.908045977011493,
+      "grad_norm": 1.450621393155795,
+      "kl": 0.043212890625,
+      "learning_rate": 9.195402298850574e-09,
+      "loss": 0.0019,
+      "reward": 1.9170316457748413,
+      "reward_std": 0.016973568126559258,
+      "rewards/accuracy_reward": 0.9170317649841309,
+      "rewards/format_reward": 1.0,
+      "step": 862
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.71614837646484,
+      "epoch": 9.919540229885058,
+      "grad_norm": 2.5531007269256762,
+      "kl": 0.04931640625,
+      "learning_rate": 8.045977011494253e-09,
+      "loss": 0.002,
+      "reward": 1.9102344512939453,
+      "reward_std": 0.0186156053096056,
+      "rewards/accuracy_reward": 0.9102343916893005,
+      "rewards/format_reward": 1.0,
+      "step": 863
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.51432800292969,
+      "epoch": 9.931034482758621,
+      "grad_norm": 1.7598054384403226,
+      "kl": 0.044921875,
+      "learning_rate": 6.896551724137931e-09,
+      "loss": 0.0019,
+      "reward": 1.8904626369476318,
+      "reward_std": 0.022165637463331223,
+      "rewards/accuracy_reward": 0.8911136388778687,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 864
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 107.57292175292969,
+      "epoch": 9.942528735632184,
+      "grad_norm": 341994.5350692805,
+      "kl": 3088.0,
+      "learning_rate": 5.747126436781609e-09,
+      "loss": 123.6946,
+      "reward": 1.9070371389389038,
+      "reward_std": 0.020475659519433975,
+      "rewards/accuracy_reward": 0.9076882600784302,
+      "rewards/format_reward": 0.9993489980697632,
+      "step": 865
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 104.62956237792969,
+      "epoch": 9.954022988505747,
+      "grad_norm": 18.068040151493758,
+      "kl": 0.052001953125,
+      "learning_rate": 4.597701149425287e-09,
+      "loss": 0.0022,
+      "reward": 1.910335898399353,
+      "reward_std": 0.018777839839458466,
+      "rewards/accuracy_reward": 0.9103360176086426,
+      "rewards/format_reward": 1.0,
+      "step": 866
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 106.146484375,
+      "epoch": 9.96551724137931,
+      "grad_norm": 4.0533200337142565,
+      "kl": 0.046142578125,
+      "learning_rate": 3.4482758620689654e-09,
+      "loss": 0.0019,
+      "reward": 1.9074934720993042,
+      "reward_std": 0.02056286484003067,
+      "rewards/accuracy_reward": 0.9074934720993042,
+      "rewards/format_reward": 1.0,
+      "step": 867
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.169921875,
+      "epoch": 9.977011494252874,
+      "grad_norm": 2.988025803486446,
+      "kl": 0.04833984375,
+      "learning_rate": 2.2988505747126436e-09,
+      "loss": 0.002,
+      "reward": 1.9082448482513428,
+      "reward_std": 0.017992818728089333,
+      "rewards/accuracy_reward": 0.908244788646698,
+      "rewards/format_reward": 1.0,
+      "step": 868
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 105.54622650146484,
+      "epoch": 9.988505747126437,
+      "grad_norm": 1.9804111140556449,
+      "kl": 0.044677734375,
+      "learning_rate": 1.1494252873563218e-09,
+      "loss": 0.0018,
+      "reward": 1.898494005203247,
+      "reward_std": 0.01871921308338642,
+      "rewards/accuracy_reward": 0.8984940052032471,
+      "rewards/format_reward": 1.0,
+      "step": 869
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 98.9199447631836,
+      "epoch": 10.0,
+      "grad_norm": 3.2087576817168033,
+      "kl": 0.041015625,
+      "learning_rate": 0.0,
+      "loss": 0.0017,
+      "reward": 1.9117414951324463,
+      "reward_std": 0.016582123935222626,
+      "rewards/accuracy_reward": 0.9117417335510254,
+      "rewards/format_reward": 1.0,
+      "step": 870
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 870,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 50,
+  "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": 48,
+  "trial_name": null,
+  "trial_params": null
+}