diff --git "a/measurement.json" "b/measurement.json" --- "a/measurement.json" +++ "b/measurement.json" @@ -2,7 +2,7 @@ "measurement": { "model.layers.0.self_attn": [ { - "accuracy": 0.891254186630249, + "accuracy": 0.8913079688423559, "total_bits": 320757760, "q_proj": { "group_size": { @@ -66,7 +66,7 @@ } }, { - "accuracy": 0.8952332044902601, + "accuracy": 0.8953711484607897, "total_bits": 329080832, "q_proj": { "group_size": { @@ -130,7 +130,7 @@ } }, { - "accuracy": 0.9228272877241436, + "accuracy": 0.9224703813854017, "total_bits": 336024576, "q_proj": { "group_size": { @@ -194,7 +194,7 @@ } }, { - "accuracy": 0.9273298288646498, + "accuracy": 0.9271842429512426, "total_bits": 401557504, "q_proj": { "group_size": { @@ -258,7 +258,7 @@ } }, { - "accuracy": 0.9251988561529862, + "accuracy": 0.9250362923270777, "total_bits": 475279360, "q_proj": { "group_size": { @@ -322,7 +322,7 @@ } }, { - "accuracy": 0.9346032393606085, + "accuracy": 0.9345189646670693, "total_bits": 475479040, "q_proj": { "group_size": { @@ -386,7 +386,7 @@ } }, { - "accuracy": 0.9266297252554643, + "accuracy": 0.9264321954626786, "total_bits": 609759232, "q_proj": { "group_size": { @@ -438,7 +438,7 @@ } }, { - "accuracy": 0.9360905697471217, + "accuracy": 0.9359589313205919, "total_bits": 610024448, "q_proj": { "group_size": { @@ -490,7 +490,7 @@ } }, { - "accuracy": 0.958151356170052, + "accuracy": 0.9578703767374942, "total_bits": 615020544, "q_proj": { "group_size": { @@ -542,7 +542,7 @@ } }, { - "accuracy": 0.958288057854301, + "accuracy": 0.9579981534104598, "total_bits": 623951872, "q_proj": { "group_size": { @@ -594,7 +594,7 @@ } }, { - "accuracy": 0.9648365158783763, + "accuracy": 0.9648009726875707, "total_bits": 626473984, "q_proj": { "group_size": { @@ -658,7 +658,7 @@ } }, { - "accuracy": 0.9780206178364, + "accuracy": 0.977967790867153, "total_bits": 630355968, "q_proj": { "group_size": { @@ -722,7 +722,7 @@ } }, { - "accuracy": 0.9653937063719097, + "accuracy": 0.965396385443838, "total_bits": 637362176, "q_proj": { "group_size": { @@ -783,7 +783,7 @@ } }, { - "accuracy": 0.978671067639401, + "accuracy": 0.9785756280547694, "total_bits": 646823936, "q_proj": { "group_size": { @@ -844,7 +844,7 @@ } }, { - "accuracy": 0.9774164400602642, + "accuracy": 0.9773460108982889, "total_bits": 784740352, "q_proj": { "group_size": { @@ -905,7 +905,7 @@ } }, { - "accuracy": 0.9901037098545777, + "accuracy": 0.9901049442981419, "total_bits": 797818880, "q_proj": { "group_size": { @@ -966,7 +966,7 @@ } }, { - "accuracy": 0.9775329498868239, + "accuracy": 0.9774822247655768, "total_bits": 911749120, "q_proj": { "group_size": { @@ -1018,7 +1018,7 @@ } }, { - "accuracy": 0.9961450362676069, + "accuracy": 0.9961438488803411, "total_bits": 942718976, "q_proj": { "group_size": { @@ -1070,7 +1070,7 @@ } }, { - "accuracy": 0.9937102347612381, + "accuracy": 0.9936837032437325, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -1124,7 +1124,7 @@ ], "model.layers.0.mlp": [ { - "accuracy": 0.9678724659116644, + "accuracy": 0.9677835922492177, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -1176,7 +1176,7 @@ } }, { - "accuracy": 0.9690336177223607, + "accuracy": 0.9696905487462094, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -1228,7 +1228,7 @@ } }, { - "accuracy": 0.9725152191362882, + "accuracy": 0.9717150832477369, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -1277,7 +1277,7 @@ } }, { - "accuracy": 0.9738709973661523, + "accuracy": 0.9730074248815838, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -1326,7 +1326,7 @@ } }, { - "accuracy": 0.9852925758612784, + "accuracy": 0.985478655288094, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -1378,7 +1378,7 @@ } }, { - "accuracy": 0.990421828470732, + "accuracy": 0.9900852010438317, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -1430,7 +1430,7 @@ } }, { - "accuracy": 0.9917013739284716, + "accuracy": 0.9913254066517478, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -1479,7 +1479,7 @@ } }, { - "accuracy": 0.9922692030668259, + "accuracy": 0.9923633501717919, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -1522,7 +1522,7 @@ } }, { - "accuracy": 0.994015393288512, + "accuracy": 0.9940901841772231, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -1565,7 +1565,7 @@ } }, { - "accuracy": 0.992849297429386, + "accuracy": 0.9930738116565504, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -1617,7 +1617,7 @@ } }, { - "accuracy": 0.9952566043326729, + "accuracy": 0.9953186045351782, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -1669,7 +1669,7 @@ } }, { - "accuracy": 0.9957317109954985, + "accuracy": 0.9958462668092627, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -1721,7 +1721,7 @@ } }, { - "accuracy": 0.9971250864235979, + "accuracy": 0.9972028175466939, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -1773,7 +1773,7 @@ } }, { - "accuracy": 0.9971372049890066, + "accuracy": 0.9970906725839564, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -1816,7 +1816,7 @@ } }, { - "accuracy": 0.9974520906413856, + "accuracy": 0.9974287811078524, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -1865,7 +1865,7 @@ } }, { - "accuracy": 0.9976318639359976, + "accuracy": 0.9976073738775755, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -1911,7 +1911,7 @@ } }, { - "accuracy": 0.9982359754411798, + "accuracy": 0.9982407394992677, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -1953,7 +1953,7 @@ ], "model.layers.1.self_attn": [ { - "accuracy": 0.9442159125679418, + "accuracy": 0.9426944130345395, "total_bits": 320757760, "q_proj": { "group_size": { @@ -2017,7 +2017,7 @@ } }, { - "accuracy": 0.9497413447028712, + "accuracy": 0.9482677924005609, "total_bits": 329080832, "q_proj": { "group_size": { @@ -2081,7 +2081,7 @@ } }, { - "accuracy": 0.9570459880326924, + "accuracy": 0.9556067303607338, "total_bits": 336024576, "q_proj": { "group_size": { @@ -2145,7 +2145,7 @@ } }, { - "accuracy": 0.9664155746761122, + "accuracy": 0.9654562661522313, "total_bits": 401557504, "q_proj": { "group_size": { @@ -2209,7 +2209,7 @@ } }, { - "accuracy": 0.9675916715672142, + "accuracy": 0.96658645491851, "total_bits": 475279360, "q_proj": { "group_size": { @@ -2273,7 +2273,7 @@ } }, { - "accuracy": 0.9672454815161855, + "accuracy": 0.9666246363991186, "total_bits": 475479040, "q_proj": { "group_size": { @@ -2337,7 +2337,7 @@ } }, { - "accuracy": 0.9710804663206402, + "accuracy": 0.9701981136673375, "total_bits": 609759232, "q_proj": { "group_size": { @@ -2389,7 +2389,7 @@ } }, { - "accuracy": 0.9708868704344097, + "accuracy": 0.9702920443133304, "total_bits": 610024448, "q_proj": { "group_size": { @@ -2441,7 +2441,7 @@ } }, { - "accuracy": 0.9637129965581392, + "accuracy": 0.9678510709812767, "total_bits": 615020544, "q_proj": { "group_size": { @@ -2493,7 +2493,7 @@ } }, { - "accuracy": 0.9641244631064565, + "accuracy": 0.9682768646039461, "total_bits": 623951872, "q_proj": { "group_size": { @@ -2545,7 +2545,7 @@ } }, { - "accuracy": 0.9750094978432906, + "accuracy": 0.9751672399671454, "total_bits": 626473984, "q_proj": { "group_size": { @@ -2609,7 +2609,7 @@ } }, { - "accuracy": 0.9771487257982555, + "accuracy": 0.9775517363297311, "total_bits": 630355968, "q_proj": { "group_size": { @@ -2673,7 +2673,7 @@ } }, { - "accuracy": 0.9756261376958144, + "accuracy": 0.9757643608670485, "total_bits": 637362176, "q_proj": { "group_size": { @@ -2734,7 +2734,7 @@ } }, { - "accuracy": 0.9778303315764979, + "accuracy": 0.9782377512831437, "total_bits": 646823936, "q_proj": { "group_size": { @@ -2795,7 +2795,7 @@ } }, { - "accuracy": 0.9820476409636045, + "accuracy": 0.981913392481051, "total_bits": 784740352, "q_proj": { "group_size": { @@ -2856,7 +2856,7 @@ } }, { - "accuracy": 0.986437135621121, + "accuracy": 0.9860767220195971, "total_bits": 797818880, "q_proj": { "group_size": { @@ -2917,7 +2917,7 @@ } }, { - "accuracy": 0.9823713741804424, + "accuracy": 0.9823048208889208, "total_bits": 911749120, "q_proj": { "group_size": { @@ -2969,7 +2969,7 @@ } }, { - "accuracy": 0.9959118338791948, + "accuracy": 0.995870193760646, "total_bits": 942718976, "q_proj": { "group_size": { @@ -3021,7 +3021,7 @@ } }, { - "accuracy": 0.995777892438989, + "accuracy": 0.9957150350275793, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -3075,7 +3075,7 @@ ], "model.layers.1.mlp": [ { - "accuracy": 0.9612327374910054, + "accuracy": 0.9611366955857528, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -3127,7 +3127,7 @@ } }, { - "accuracy": 0.9625783186209829, + "accuracy": 0.9624664281543932, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -3179,7 +3179,7 @@ } }, { - "accuracy": 0.9687596779120596, + "accuracy": 0.9685962702098646, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -3228,7 +3228,7 @@ } }, { - "accuracy": 0.9707322591229489, + "accuracy": 0.9705850011424014, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -3277,7 +3277,7 @@ } }, { - "accuracy": 0.9809978792541906, + "accuracy": 0.9808641578021803, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -3329,7 +3329,7 @@ } }, { - "accuracy": 0.9834107066455641, + "accuracy": 0.9833953929574866, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -3381,7 +3381,7 @@ } }, { - "accuracy": 0.9861036288110834, + "accuracy": 0.9861325684346651, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -3430,7 +3430,7 @@ } }, { - "accuracy": 0.9902443807376059, + "accuracy": 0.9902534241739073, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -3473,7 +3473,7 @@ } }, { - "accuracy": 0.9913441005506014, + "accuracy": 0.9913561179449684, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -3516,7 +3516,7 @@ } }, { - "accuracy": 0.9903671812070044, + "accuracy": 0.9903732953887237, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -3568,7 +3568,7 @@ } }, { - "accuracy": 0.992041350979554, + "accuracy": 0.9920450072539481, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -3620,7 +3620,7 @@ } }, { - "accuracy": 0.9949665955807033, + "accuracy": 0.9949860019903434, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -3672,7 +3672,7 @@ } }, { - "accuracy": 0.9959477339135973, + "accuracy": 0.995948794248857, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -3724,7 +3724,7 @@ } }, { - "accuracy": 0.9971439436470207, + "accuracy": 0.9971517200924849, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -3767,7 +3767,7 @@ } }, { - "accuracy": 0.9974367165643918, + "accuracy": 0.9974574529026684, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -3816,7 +3816,7 @@ } }, { - "accuracy": 0.9979243368694657, + "accuracy": 0.9979582054442481, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -3862,7 +3862,7 @@ } }, { - "accuracy": 0.999223201780727, + "accuracy": 0.9992253149516488, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -3904,7 +3904,7 @@ ], "model.layers.2.self_attn": [ { - "accuracy": 0.9511175375235708, + "accuracy": 0.9492432914282146, "total_bits": 320757760, "q_proj": { "group_size": { @@ -3968,7 +3968,7 @@ } }, { - "accuracy": 0.9534524334104437, + "accuracy": 0.9520342914681685, "total_bits": 329080832, "q_proj": { "group_size": { @@ -4032,7 +4032,7 @@ } }, { - "accuracy": 0.9615737105670729, + "accuracy": 0.9590781425174913, "total_bits": 336024576, "q_proj": { "group_size": { @@ -4096,7 +4096,7 @@ } }, { - "accuracy": 0.9699305891990662, + "accuracy": 0.9682425009576898, "total_bits": 401557504, "q_proj": { "group_size": { @@ -4160,7 +4160,7 @@ } }, { - "accuracy": 0.9709137144841646, + "accuracy": 0.969519201077913, "total_bits": 475279360, "q_proj": { "group_size": { @@ -4224,7 +4224,7 @@ } }, { - "accuracy": 0.9715015919584977, + "accuracy": 0.9707638527217665, "total_bits": 475479040, "q_proj": { "group_size": { @@ -4288,7 +4288,7 @@ } }, { - "accuracy": 0.9750262357686695, + "accuracy": 0.9744096351297278, "total_bits": 609759232, "q_proj": { "group_size": { @@ -4340,7 +4340,7 @@ } }, { - "accuracy": 0.9757627405618367, + "accuracy": 0.9756679550597542, "total_bits": 610024448, "q_proj": { "group_size": { @@ -4392,7 +4392,7 @@ } }, { - "accuracy": 0.9771439637008467, + "accuracy": 0.9756332102574801, "total_bits": 615020544, "q_proj": { "group_size": { @@ -4444,7 +4444,7 @@ } }, { - "accuracy": 0.9774530561346757, + "accuracy": 0.9759448860820971, "total_bits": 623951872, "q_proj": { "group_size": { @@ -4496,7 +4496,7 @@ } }, { - "accuracy": 0.9783729534400137, + "accuracy": 0.9788673300492136, "total_bits": 626473984, "q_proj": { "group_size": { @@ -4560,7 +4560,7 @@ } }, { - "accuracy": 0.9798371446760077, + "accuracy": 0.9795829387087571, "total_bits": 630355968, "q_proj": { "group_size": { @@ -4624,7 +4624,7 @@ } }, { - "accuracy": 0.9792431106692866, + "accuracy": 0.9796319948999506, "total_bits": 637362176, "q_proj": { "group_size": { @@ -4685,7 +4685,7 @@ } }, { - "accuracy": 0.980660722443932, + "accuracy": 0.9804158116641798, "total_bits": 646823936, "q_proj": { "group_size": { @@ -4746,7 +4746,7 @@ } }, { - "accuracy": 0.9842103500115243, + "accuracy": 0.9841269238999015, "total_bits": 784740352, "q_proj": { "group_size": { @@ -4807,7 +4807,7 @@ } }, { - "accuracy": 0.9857045898312017, + "accuracy": 0.9855421389404097, "total_bits": 797818880, "q_proj": { "group_size": { @@ -4868,7 +4868,7 @@ } }, { - "accuracy": 0.9845278984621951, + "accuracy": 0.9844953433463448, "total_bits": 911749120, "q_proj": { "group_size": { @@ -4920,7 +4920,7 @@ } }, { - "accuracy": 0.9960261787238874, + "accuracy": 0.9959946266914669, "total_bits": 942718976, "q_proj": { "group_size": { @@ -4972,7 +4972,7 @@ } }, { - "accuracy": 0.995498704282861, + "accuracy": 0.9954696023150494, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -5026,7 +5026,7 @@ ], "model.layers.2.mlp": [ { - "accuracy": 0.9515779802673742, + "accuracy": 0.9520457826162639, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -5078,7 +5078,7 @@ } }, { - "accuracy": 0.9529434630745336, + "accuracy": 0.9533393414396989, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -5130,7 +5130,7 @@ } }, { - "accuracy": 0.9602926844044736, + "accuracy": 0.9607522424898649, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -5179,7 +5179,7 @@ } }, { - "accuracy": 0.9624017539777254, + "accuracy": 0.962889037634197, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -5228,7 +5228,7 @@ } }, { - "accuracy": 0.9753169401695854, + "accuracy": 0.9754583804230941, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -5280,7 +5280,7 @@ } }, { - "accuracy": 0.9787832733831907, + "accuracy": 0.9789498871878574, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -5332,7 +5332,7 @@ } }, { - "accuracy": 0.9817603544185036, + "accuracy": 0.9819650446113787, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -5381,7 +5381,7 @@ } }, { - "accuracy": 0.9877299242898038, + "accuracy": 0.9878207147121429, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -5424,7 +5424,7 @@ } }, { - "accuracy": 0.9892617354267522, + "accuracy": 0.9893497573701959, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -5467,7 +5467,7 @@ } }, { - "accuracy": 0.9872116501394071, + "accuracy": 0.98728231223006, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -5519,7 +5519,7 @@ } }, { - "accuracy": 0.989829777886993, + "accuracy": 0.9899042578119981, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -5571,7 +5571,7 @@ } }, { - "accuracy": 0.9933448951495322, + "accuracy": 0.9933898025437405, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -5623,7 +5623,7 @@ } }, { - "accuracy": 0.9949317892131052, + "accuracy": 0.9949642262960735, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -5675,7 +5675,7 @@ } }, { - "accuracy": 0.9963868527035964, + "accuracy": 0.9964100067552767, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -5718,7 +5718,7 @@ } }, { - "accuracy": 0.996529007036435, + "accuracy": 0.9965528057594049, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -5767,7 +5767,7 @@ } }, { - "accuracy": 0.9970070172689462, + "accuracy": 0.9970356159304318, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -5813,7 +5813,7 @@ } }, { - "accuracy": 0.9990123218219531, + "accuracy": 0.9990187570648758, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -5855,7 +5855,7 @@ ], "model.layers.3.self_attn": [ { - "accuracy": 0.9425999365354839, + "accuracy": 0.942145027612385, "total_bits": 320757760, "q_proj": { "group_size": { @@ -5919,7 +5919,7 @@ } }, { - "accuracy": 0.9501617770445974, + "accuracy": 0.9492981653464467, "total_bits": 329080832, "q_proj": { "group_size": { @@ -5983,7 +5983,7 @@ } }, { - "accuracy": 0.9577044405435261, + "accuracy": 0.9566133461500469, "total_bits": 336024576, "q_proj": { "group_size": { @@ -6047,7 +6047,7 @@ } }, { - "accuracy": 0.965823233127594, + "accuracy": 0.9650772935465762, "total_bits": 401557504, "q_proj": { "group_size": { @@ -6111,7 +6111,7 @@ } }, { - "accuracy": 0.9663137134752775, + "accuracy": 0.9654968569153234, "total_bits": 475279360, "q_proj": { "group_size": { @@ -6175,7 +6175,7 @@ } }, { - "accuracy": 0.9669951890644274, + "accuracy": 0.9671572980127836, "total_bits": 475479040, "q_proj": { "group_size": { @@ -6239,7 +6239,7 @@ } }, { - "accuracy": 0.9707100360017074, + "accuracy": 0.9694834131943552, "total_bits": 609759232, "q_proj": { "group_size": { @@ -6291,7 +6291,7 @@ } }, { - "accuracy": 0.9712476322525426, + "accuracy": 0.971659054881648, "total_bits": 610024448, "q_proj": { "group_size": { @@ -6343,7 +6343,7 @@ } }, { - "accuracy": 0.9726966556749845, + "accuracy": 0.9725218446631181, "total_bits": 615020544, "q_proj": { "group_size": { @@ -6395,7 +6395,7 @@ } }, { - "accuracy": 0.972788531529276, + "accuracy": 0.9727619980510912, "total_bits": 623951872, "q_proj": { "group_size": { @@ -6447,7 +6447,7 @@ } }, { - "accuracy": 0.97900023272163, + "accuracy": 0.9785469895914981, "total_bits": 626473984, "q_proj": { "group_size": { @@ -6511,7 +6511,7 @@ } }, { - "accuracy": 0.9803975349978397, + "accuracy": 0.9799953118750924, "total_bits": 630355968, "q_proj": { "group_size": { @@ -6575,7 +6575,7 @@ } }, { - "accuracy": 0.9800364892733725, + "accuracy": 0.9797634921575847, "total_bits": 637362176, "q_proj": { "group_size": { @@ -6636,7 +6636,7 @@ } }, { - "accuracy": 0.9814623562913192, + "accuracy": 0.9809917995804235, "total_bits": 646823936, "q_proj": { "group_size": { @@ -6697,7 +6697,7 @@ } }, { - "accuracy": 0.9859066887905723, + "accuracy": 0.9858301925031763, "total_bits": 784740352, "q_proj": { "group_size": { @@ -6758,7 +6758,7 @@ } }, { - "accuracy": 0.9889583556275618, + "accuracy": 0.9889305390809712, "total_bits": 797818880, "q_proj": { "group_size": { @@ -6819,7 +6819,7 @@ } }, { - "accuracy": 0.9862525321935353, + "accuracy": 0.9862193057411596, "total_bits": 911749120, "q_proj": { "group_size": { @@ -6871,7 +6871,7 @@ } }, { - "accuracy": 0.9966635653062871, + "accuracy": 0.9965524206820288, "total_bits": 942718976, "q_proj": { "group_size": { @@ -6923,7 +6923,7 @@ } }, { - "accuracy": 0.9958450676579225, + "accuracy": 0.995772558999689, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -6977,7 +6977,7 @@ ], "model.layers.3.mlp": [ { - "accuracy": 0.8722025093279386, + "accuracy": 0.8793219767118755, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -7029,7 +7029,7 @@ } }, { - "accuracy": 0.8727699079011616, + "accuracy": 0.8801141663601524, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -7081,7 +7081,7 @@ } }, { - "accuracy": 0.8715165916242098, + "accuracy": 0.8789716645290977, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -7130,7 +7130,7 @@ } }, { - "accuracy": 0.8716030246333072, + "accuracy": 0.8790680609251323, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -7179,7 +7179,7 @@ } }, { - "accuracy": 0.9521443592874628, + "accuracy": 0.9491347607813383, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -7231,7 +7231,7 @@ } }, { - "accuracy": 0.9278651789615029, + "accuracy": 0.9270698773233514, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -7283,7 +7283,7 @@ } }, { - "accuracy": 0.9281053166640432, + "accuracy": 0.9272474866164357, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -7332,7 +7332,7 @@ } }, { - "accuracy": 0.9427242216311003, + "accuracy": 0.9413776146738153, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -7375,7 +7375,7 @@ } }, { - "accuracy": 0.9290015948446173, + "accuracy": 0.9282888864216051, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -7418,7 +7418,7 @@ } }, { - "accuracy": 0.9889214611367175, + "accuracy": 0.9888658154951898, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -7470,7 +7470,7 @@ } }, { - "accuracy": 0.9760843264429193, + "accuracy": 0.9764353444701747, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -7522,7 +7522,7 @@ } }, { - "accuracy": 0.9583742555819059, + "accuracy": 0.9590443937402022, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -7574,7 +7574,7 @@ } }, { - "accuracy": 0.9589786121719762, + "accuracy": 0.9589199674756903, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -7626,7 +7626,7 @@ } }, { - "accuracy": 0.959250952068128, + "accuracy": 0.9589597174995824, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -7669,7 +7669,7 @@ } }, { - "accuracy": 0.9982182250211113, + "accuracy": 0.9979138415502874, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -7718,7 +7718,7 @@ } }, { - "accuracy": 0.9982662140146682, + "accuracy": 0.9979441760009841, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -7764,7 +7764,7 @@ } }, { - "accuracy": 0.9983859280810544, + "accuracy": 0.9984178141151604, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -7806,7 +7806,7 @@ ], "model.layers.4.self_attn": [ { - "accuracy": 0.996596623408167, + "accuracy": 0.9967076617636179, "total_bits": 320757760, "q_proj": { "group_size": { @@ -7870,7 +7870,7 @@ } }, { - "accuracy": 0.9968075201307472, + "accuracy": 0.9968344714296492, "total_bits": 329080832, "q_proj": { "group_size": { @@ -7934,7 +7934,7 @@ } }, { - "accuracy": 0.9971539176216251, + "accuracy": 0.9972560841001963, "total_bits": 336024576, "q_proj": { "group_size": { @@ -7998,7 +7998,7 @@ } }, { - "accuracy": 0.997461390142378, + "accuracy": 0.9975254704293451, "total_bits": 401557504, "q_proj": { "group_size": { @@ -8062,7 +8062,7 @@ } }, { - "accuracy": 0.9974974111506814, + "accuracy": 0.9975679871675215, "total_bits": 475279360, "q_proj": { "group_size": { @@ -8126,7 +8126,7 @@ } }, { - "accuracy": 0.9975956858772981, + "accuracy": 0.9975430424276152, "total_bits": 475479040, "q_proj": { "group_size": { @@ -8190,7 +8190,7 @@ } }, { - "accuracy": 0.9981733576247567, + "accuracy": 0.998152835980842, "total_bits": 609759232, "q_proj": { "group_size": { @@ -8242,7 +8242,7 @@ } }, { - "accuracy": 0.998212596891742, + "accuracy": 0.9982336522325089, "total_bits": 610024448, "q_proj": { "group_size": { @@ -8294,7 +8294,7 @@ } }, { - "accuracy": 0.9983239909143824, + "accuracy": 0.9983393357772576, "total_bits": 615020544, "q_proj": { "group_size": { @@ -8346,7 +8346,7 @@ } }, { - "accuracy": 0.998335201489298, + "accuracy": 0.9983492230898455, "total_bits": 623951872, "q_proj": { "group_size": { @@ -8398,7 +8398,7 @@ } }, { - "accuracy": 0.9982773863563412, + "accuracy": 0.9983042339353185, "total_bits": 626473984, "q_proj": { "group_size": { @@ -8462,7 +8462,7 @@ } }, { - "accuracy": 0.9983736161926859, + "accuracy": 0.998389504163673, "total_bits": 630355968, "q_proj": { "group_size": { @@ -8526,7 +8526,7 @@ } }, { - "accuracy": 0.9989196890474934, + "accuracy": 0.9989347542194944, "total_bits": 637362176, "q_proj": { "group_size": { @@ -8587,7 +8587,7 @@ } }, { - "accuracy": 0.9990101504678789, + "accuracy": 0.9990375650752532, "total_bits": 646823936, "q_proj": { "group_size": { @@ -8648,7 +8648,7 @@ } }, { - "accuracy": 0.9993791671371773, + "accuracy": 0.9993939807050323, "total_bits": 784740352, "q_proj": { "group_size": { @@ -8709,7 +8709,7 @@ } }, { - "accuracy": 0.99947380343158, + "accuracy": 0.9994880955568269, "total_bits": 797818880, "q_proj": { "group_size": { @@ -8770,7 +8770,7 @@ } }, { - "accuracy": 0.9994037400833086, + "accuracy": 0.9994162290210002, "total_bits": 911749120, "q_proj": { "group_size": { @@ -8822,7 +8822,7 @@ } }, { - "accuracy": 0.9998376835995403, + "accuracy": 0.9998446839469436, "total_bits": 942718976, "q_proj": { "group_size": { @@ -8874,7 +8874,7 @@ } }, { - "accuracy": 0.9998492780386617, + "accuracy": 0.9998521383876275, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -8928,7 +8928,7 @@ ], "model.layers.4.mlp": [ { - "accuracy": 0.994694937608744, + "accuracy": 0.9947486817836761, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -8980,7 +8980,7 @@ } }, { - "accuracy": 0.994842595175693, + "accuracy": 0.9948915670576849, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -9032,7 +9032,7 @@ } }, { - "accuracy": 0.9955659140097467, + "accuracy": 0.995609168943606, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -9081,7 +9081,7 @@ } }, { - "accuracy": 0.9957820145707381, + "accuracy": 0.995823761742366, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -9130,7 +9130,7 @@ } }, { - "accuracy": 0.9972479345374986, + "accuracy": 0.9972776425512213, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -9182,7 +9182,7 @@ } }, { - "accuracy": 0.997611795208956, + "accuracy": 0.9976370338546602, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -9234,7 +9234,7 @@ } }, { - "accuracy": 0.9979161439757598, + "accuracy": 0.9979400085775476, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -9283,7 +9283,7 @@ } }, { - "accuracy": 0.9985973871264019, + "accuracy": 0.9986105899473554, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -9326,7 +9326,7 @@ } }, { - "accuracy": 0.998780930120694, + "accuracy": 0.9987936880635587, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -9369,7 +9369,7 @@ } }, { - "accuracy": 0.998596390709281, + "accuracy": 0.9986107396452051, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -9421,7 +9421,7 @@ } }, { - "accuracy": 0.9988548380175704, + "accuracy": 0.9988651113878739, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -9473,7 +9473,7 @@ } }, { - "accuracy": 0.9992720522476655, + "accuracy": 0.9992786154739166, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -9525,7 +9525,7 @@ } }, { - "accuracy": 0.9994133810855841, + "accuracy": 0.9994178080049, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -9577,7 +9577,7 @@ } }, { - "accuracy": 0.9996016605904228, + "accuracy": 0.9996053293396375, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -9620,7 +9620,7 @@ } }, { - "accuracy": 0.9996183153094822, + "accuracy": 0.999621723435427, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -9669,7 +9669,7 @@ } }, { - "accuracy": 0.9996691107455837, + "accuracy": 0.9996722449646577, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -9715,7 +9715,7 @@ } }, { - "accuracy": 0.9998784999510175, + "accuracy": 0.999879395417673, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -9757,7 +9757,7 @@ ], "model.layers.5.self_attn": [ { - "accuracy": 0.9962007458272734, + "accuracy": 0.9962098916110239, "total_bits": 320757760, "q_proj": { "group_size": { @@ -9821,7 +9821,7 @@ } }, { - "accuracy": 0.99641147961742, + "accuracy": 0.9964749699360446, "total_bits": 329080832, "q_proj": { "group_size": { @@ -9885,7 +9885,7 @@ } }, { - "accuracy": 0.9967718838076842, + "accuracy": 0.996835154726317, "total_bits": 336024576, "q_proj": { "group_size": { @@ -9949,7 +9949,7 @@ } }, { - "accuracy": 0.9975172085197348, + "accuracy": 0.9975610520494612, "total_bits": 401557504, "q_proj": { "group_size": { @@ -10013,7 +10013,7 @@ } }, { - "accuracy": 0.9975869287001459, + "accuracy": 0.9976268110698775, "total_bits": 475279360, "q_proj": { "group_size": { @@ -10077,7 +10077,7 @@ } }, { - "accuracy": 0.997627241438941, + "accuracy": 0.9976449802910027, "total_bits": 475479040, "q_proj": { "group_size": { @@ -10141,7 +10141,7 @@ } }, { - "accuracy": 0.9980781748890877, + "accuracy": 0.9981090659765821, "total_bits": 609759232, "q_proj": { "group_size": { @@ -10193,7 +10193,7 @@ } }, { - "accuracy": 0.9981155095523909, + "accuracy": 0.9981805658654163, "total_bits": 610024448, "q_proj": { "group_size": { @@ -10245,7 +10245,7 @@ } }, { - "accuracy": 0.9986656150339466, + "accuracy": 0.9986678678542376, "total_bits": 615020544, "q_proj": { "group_size": { @@ -10297,7 +10297,7 @@ } }, { - "accuracy": 0.9987380231093419, + "accuracy": 0.9987403344559042, "total_bits": 623951872, "q_proj": { "group_size": { @@ -10349,7 +10349,7 @@ } }, { - "accuracy": 0.998228382319212, + "accuracy": 0.9982575389899706, "total_bits": 626473984, "q_proj": { "group_size": { @@ -10413,7 +10413,7 @@ } }, { - "accuracy": 0.9987086756645065, + "accuracy": 0.9987127411326295, "total_bits": 630355968, "q_proj": { "group_size": { @@ -10477,7 +10477,7 @@ } }, { - "accuracy": 0.9985158341495615, + "accuracy": 0.9985207380040696, "total_bits": 637362176, "q_proj": { "group_size": { @@ -10538,7 +10538,7 @@ } }, { - "accuracy": 0.9991244713922864, + "accuracy": 0.9991205387601727, "total_bits": 646823936, "q_proj": { "group_size": { @@ -10599,7 +10599,7 @@ } }, { - "accuracy": 0.9988383460593852, + "accuracy": 0.9988506195184431, "total_bits": 784740352, "q_proj": { "group_size": { @@ -10660,7 +10660,7 @@ } }, { - "accuracy": 0.9995539788842985, + "accuracy": 0.9995617592021039, "total_bits": 797818880, "q_proj": { "group_size": { @@ -10721,7 +10721,7 @@ } }, { - "accuracy": 0.9988687332523497, + "accuracy": 0.9988815948171051, "total_bits": 911749120, "q_proj": { "group_size": { @@ -10773,7 +10773,7 @@ } }, { - "accuracy": 0.999823316551843, + "accuracy": 0.9998254804675909, "total_bits": 942718976, "q_proj": { "group_size": { @@ -10825,7 +10825,7 @@ } }, { - "accuracy": 0.9992312639951706, + "accuracy": 0.9992367904633284, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -10879,7 +10879,7 @@ ], "model.layers.5.mlp": [ { - "accuracy": 0.9940182060787552, + "accuracy": 0.9940763998188471, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -10931,7 +10931,7 @@ } }, { - "accuracy": 0.9941802616966399, + "accuracy": 0.9942348564141675, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -10983,7 +10983,7 @@ } }, { - "accuracy": 0.9950554919870276, + "accuracy": 0.9951036760681554, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -11032,7 +11032,7 @@ } }, { - "accuracy": 0.9953264852887705, + "accuracy": 0.9953726894761387, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -11081,7 +11081,7 @@ } }, { - "accuracy": 0.9968628732389525, + "accuracy": 0.996892218527041, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -11133,7 +11133,7 @@ } }, { - "accuracy": 0.9973026508171308, + "accuracy": 0.997328795688717, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -11185,7 +11185,7 @@ } }, { - "accuracy": 0.9976877121156768, + "accuracy": 0.9977120292422018, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -11234,7 +11234,7 @@ } }, { - "accuracy": 0.9983344442750278, + "accuracy": 0.9983520639178, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -11277,7 +11277,7 @@ } }, { - "accuracy": 0.9986240245205792, + "accuracy": 0.9986368707920376, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -11320,7 +11320,7 @@ } }, { - "accuracy": 0.998309123869005, + "accuracy": 0.9983234993721309, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -11372,7 +11372,7 @@ } }, { - "accuracy": 0.9986801868990848, + "accuracy": 0.998709158285668, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -11424,7 +11424,7 @@ } }, { - "accuracy": 0.9989751206808969, + "accuracy": 0.9989698508656338, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -11476,7 +11476,7 @@ } }, { - "accuracy": 0.9993447800608057, + "accuracy": 0.9993536914943865, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -11528,7 +11528,7 @@ } }, { - "accuracy": 0.9993525424874142, + "accuracy": 0.9993383121235591, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -11571,7 +11571,7 @@ } }, { - "accuracy": 0.9994447115612658, + "accuracy": 0.999447800022991, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -11620,7 +11620,7 @@ } }, { - "accuracy": 0.9994942018468129, + "accuracy": 0.9994966229914051, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -11666,7 +11666,7 @@ } }, { - "accuracy": 0.9998259799138299, + "accuracy": 0.9998231498451021, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -11708,7 +11708,7 @@ ], "model.layers.6.self_attn": [ { - "accuracy": 0.996207807409136, + "accuracy": 0.9961960727446958, "total_bits": 320757760, "q_proj": { "group_size": { @@ -11772,7 +11772,7 @@ } }, { - "accuracy": 0.9966101858176684, + "accuracy": 0.9966047821860564, "total_bits": 329080832, "q_proj": { "group_size": { @@ -11836,7 +11836,7 @@ } }, { - "accuracy": 0.9971668584958503, + "accuracy": 0.9971812792906636, "total_bits": 336024576, "q_proj": { "group_size": { @@ -11900,7 +11900,7 @@ } }, { - "accuracy": 0.9980172983517772, + "accuracy": 0.9980327428171509, "total_bits": 401557504, "q_proj": { "group_size": { @@ -11964,7 +11964,7 @@ } }, { - "accuracy": 0.9981802229426409, + "accuracy": 0.998192086815834, "total_bits": 475279360, "q_proj": { "group_size": { @@ -12028,7 +12028,7 @@ } }, { - "accuracy": 0.9981882295718318, + "accuracy": 0.9981996266073302, "total_bits": 475479040, "q_proj": { "group_size": { @@ -12092,7 +12092,7 @@ } }, { - "accuracy": 0.998793363571167, + "accuracy": 0.9987943998881077, "total_bits": 609759232, "q_proj": { "group_size": { @@ -12144,7 +12144,7 @@ } }, { - "accuracy": 0.9988138379627153, + "accuracy": 0.9988225550439797, "total_bits": 610024448, "q_proj": { "group_size": { @@ -12196,7 +12196,7 @@ } }, { - "accuracy": 0.9988799689239577, + "accuracy": 0.9988955262264139, "total_bits": 615020544, "q_proj": { "group_size": { @@ -12248,7 +12248,7 @@ } }, { - "accuracy": 0.9989080408490018, + "accuracy": 0.9989256020635366, "total_bits": 623951872, "q_proj": { "group_size": { @@ -12300,7 +12300,7 @@ } }, { - "accuracy": 0.9989084399452335, + "accuracy": 0.9989010812224526, "total_bits": 626473984, "q_proj": { "group_size": { @@ -12364,7 +12364,7 @@ } }, { - "accuracy": 0.9989766478538513, + "accuracy": 0.9989790592930818, "total_bits": 630355968, "q_proj": { "group_size": { @@ -12428,7 +12428,7 @@ } }, { - "accuracy": 0.9992043576938542, + "accuracy": 0.999208791475547, "total_bits": 637362176, "q_proj": { "group_size": { @@ -12489,7 +12489,7 @@ } }, { - "accuracy": 0.9992613621350181, + "accuracy": 0.9992624727126799, "total_bits": 646823936, "q_proj": { "group_size": { @@ -12550,7 +12550,7 @@ } }, { - "accuracy": 0.9995591619885281, + "accuracy": 0.9995601437005558, "total_bits": 784740352, "q_proj": { "group_size": { @@ -12611,7 +12611,7 @@ } }, { - "accuracy": 0.9996200683526695, + "accuracy": 0.9996221511575737, "total_bits": 797818880, "q_proj": { "group_size": { @@ -12672,7 +12672,7 @@ } }, { - "accuracy": 0.9996430145968732, + "accuracy": 0.9996435542247797, "total_bits": 911749120, "q_proj": { "group_size": { @@ -12724,7 +12724,7 @@ } }, { - "accuracy": 0.9998280815902705, + "accuracy": 0.999828881988498, "total_bits": 942718976, "q_proj": { "group_size": { @@ -12776,7 +12776,7 @@ } }, { - "accuracy": 0.9999035694268777, + "accuracy": 0.9999047451052129, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -12830,7 +12830,7 @@ ], "model.layers.6.mlp": [ { - "accuracy": 0.9931766422171342, + "accuracy": 0.9932317404370559, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -12882,7 +12882,7 @@ } }, { - "accuracy": 0.9933360909160814, + "accuracy": 0.9933933647055375, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -12934,7 +12934,7 @@ } }, { - "accuracy": 0.9943272102820245, + "accuracy": 0.9943669046226301, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -12983,7 +12983,7 @@ } }, { - "accuracy": 0.994624503349003, + "accuracy": 0.9946611539313668, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -13032,7 +13032,7 @@ } }, { - "accuracy": 0.9965119569709426, + "accuracy": 0.996550022771484, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -13084,7 +13084,7 @@ } }, { - "accuracy": 0.9969714491775161, + "accuracy": 0.9968922506821783, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -13136,7 +13136,7 @@ } }, { - "accuracy": 0.9974078455645787, + "accuracy": 0.9973084316834023, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -13185,7 +13185,7 @@ } }, { - "accuracy": 0.9982912985509947, + "accuracy": 0.9981391094624996, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -13228,7 +13228,7 @@ } }, { - "accuracy": 0.9984421872190739, + "accuracy": 0.9982764544455629, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -13271,7 +13271,7 @@ } }, { - "accuracy": 0.9983304933105644, + "accuracy": 0.9983435155529725, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -13323,7 +13323,7 @@ } }, { - "accuracy": 0.998522547710883, + "accuracy": 0.9985220422478098, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -13375,7 +13375,7 @@ } }, { - "accuracy": 0.9991355948150158, + "accuracy": 0.9991445082582926, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -13427,7 +13427,7 @@ } }, { - "accuracy": 0.999217535614183, + "accuracy": 0.9992623229658133, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -13479,7 +13479,7 @@ } }, { - "accuracy": 0.9995120765739366, + "accuracy": 0.9995166817680001, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -13522,7 +13522,7 @@ } }, { - "accuracy": 0.9995583918827929, + "accuracy": 0.9995617358700225, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -13571,7 +13571,7 @@ } }, { - "accuracy": 0.9996411723182782, + "accuracy": 0.9996437431607199, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -13617,7 +13617,7 @@ } }, { - "accuracy": 0.9998534829233234, + "accuracy": 0.9998546256438682, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -13659,7 +13659,7 @@ ], "model.layers.7.self_attn": [ { - "accuracy": 0.996611154393146, + "accuracy": 0.9966092164579191, "total_bits": 320757760, "q_proj": { "group_size": { @@ -13723,7 +13723,7 @@ } }, { - "accuracy": 0.9969208764010354, + "accuracy": 0.9968956844194939, "total_bits": 329080832, "q_proj": { "group_size": { @@ -13787,7 +13787,7 @@ } }, { - "accuracy": 0.9974951836231508, + "accuracy": 0.9975194274202773, "total_bits": 336024576, "q_proj": { "group_size": { @@ -13851,7 +13851,7 @@ } }, { - "accuracy": 0.998122817591617, + "accuracy": 0.9981430235661959, "total_bits": 401557504, "q_proj": { "group_size": { @@ -13915,7 +13915,7 @@ } }, { - "accuracy": 0.998290934648953, + "accuracy": 0.9983101922430491, "total_bits": 475279360, "q_proj": { "group_size": { @@ -13979,7 +13979,7 @@ } }, { - "accuracy": 0.9983092697435304, + "accuracy": 0.9983449070469329, "total_bits": 475479040, "q_proj": { "group_size": { @@ -14043,7 +14043,7 @@ } }, { - "accuracy": 0.9988098412163948, + "accuracy": 0.9988129462458586, "total_bits": 609759232, "q_proj": { "group_size": { @@ -14095,7 +14095,7 @@ } }, { - "accuracy": 0.9988487519715962, + "accuracy": 0.9988599712715337, "total_bits": 610024448, "q_proj": { "group_size": { @@ -14147,7 +14147,7 @@ } }, { - "accuracy": 0.9989099242773495, + "accuracy": 0.9989263969228456, "total_bits": 615020544, "q_proj": { "group_size": { @@ -14199,7 +14199,7 @@ } }, { - "accuracy": 0.9989313406771735, + "accuracy": 0.9989472919780957, "total_bits": 623951872, "q_proj": { "group_size": { @@ -14251,7 +14251,7 @@ } }, { - "accuracy": 0.9989460107723349, + "accuracy": 0.9989413064169256, "total_bits": 626473984, "q_proj": { "group_size": { @@ -14315,7 +14315,7 @@ } }, { - "accuracy": 0.9989892843326456, + "accuracy": 0.9990105677984262, "total_bits": 630355968, "q_proj": { "group_size": { @@ -14379,7 +14379,7 @@ } }, { - "accuracy": 0.9992350631050373, + "accuracy": 0.9992521332184735, "total_bits": 637362176, "q_proj": { "group_size": { @@ -14440,7 +14440,7 @@ } }, { - "accuracy": 0.9992939952858969, + "accuracy": 0.9993006920343951, "total_bits": 646823936, "q_proj": { "group_size": { @@ -14501,7 +14501,7 @@ } }, { - "accuracy": 0.9995863391576629, + "accuracy": 0.9995826370033779, "total_bits": 784740352, "q_proj": { "group_size": { @@ -14562,7 +14562,7 @@ } }, { - "accuracy": 0.9996509935314718, + "accuracy": 0.9996557628343764, "total_bits": 797818880, "q_proj": { "group_size": { @@ -14623,7 +14623,7 @@ } }, { - "accuracy": 0.9996432583583029, + "accuracy": 0.9996383065161737, "total_bits": 911749120, "q_proj": { "group_size": { @@ -14675,7 +14675,7 @@ } }, { - "accuracy": 0.9998473333145835, + "accuracy": 0.9998493494931608, "total_bits": 942718976, "q_proj": { "group_size": { @@ -14727,7 +14727,7 @@ } }, { - "accuracy": 0.9999025650261166, + "accuracy": 0.9999039103154486, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -14781,7 +14781,7 @@ ], "model.layers.7.mlp": [ { - "accuracy": 0.9922762748442198, + "accuracy": 0.9923360739883623, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -14833,7 +14833,7 @@ } }, { - "accuracy": 0.9924961022640529, + "accuracy": 0.9925532042980194, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -14885,7 +14885,7 @@ } }, { - "accuracy": 0.9935847838458262, + "accuracy": 0.993632468738054, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -14934,7 +14934,7 @@ } }, { - "accuracy": 0.9939163299767595, + "accuracy": 0.9939610714975157, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -14983,7 +14983,7 @@ } }, { - "accuracy": 0.9962378477579669, + "accuracy": 0.9962657286148322, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -15035,7 +15035,7 @@ } }, { - "accuracy": 0.9965670477402838, + "accuracy": 0.9965961375518849, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -15087,7 +15087,7 @@ } }, { - "accuracy": 0.997049139126351, + "accuracy": 0.9970748057883037, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -15136,7 +15136,7 @@ } }, { - "accuracy": 0.998086436798698, + "accuracy": 0.9981012756102964, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -15179,7 +15179,7 @@ } }, { - "accuracy": 0.9982565810021601, + "accuracy": 0.9982723452542958, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -15222,7 +15222,7 @@ } }, { - "accuracy": 0.998106975304453, + "accuracy": 0.9981202528272805, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -15274,7 +15274,7 @@ } }, { - "accuracy": 0.998343457516871, + "accuracy": 0.9983565638724127, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -15326,7 +15326,7 @@ } }, { - "accuracy": 0.999037885058083, + "accuracy": 0.9990442952043131, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -15378,7 +15378,7 @@ } }, { - "accuracy": 0.9991807854783378, + "accuracy": 0.9991866449188245, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -15430,7 +15430,7 @@ } }, { - "accuracy": 0.9994829886739975, + "accuracy": 0.9994865132887897, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -15473,7 +15473,7 @@ } }, { - "accuracy": 0.9995096185685772, + "accuracy": 0.9995124123892501, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -15522,7 +15522,7 @@ } }, { - "accuracy": 0.9996034321375191, + "accuracy": 0.9996054120802957, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -15568,7 +15568,7 @@ } }, { - "accuracy": 0.999861117599434, + "accuracy": 0.9998619523585627, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -15610,7 +15610,7 @@ ], "model.layers.8.self_attn": [ { - "accuracy": 0.9948774160523164, + "accuracy": 0.9950947812513301, "total_bits": 320757760, "q_proj": { "group_size": { @@ -15674,7 +15674,7 @@ } }, { - "accuracy": 0.9952819163077756, + "accuracy": 0.9953793534322789, "total_bits": 329080832, "q_proj": { "group_size": { @@ -15738,7 +15738,7 @@ } }, { - "accuracy": 0.9960418327858573, + "accuracy": 0.9960658660060481, "total_bits": 336024576, "q_proj": { "group_size": { @@ -15802,7 +15802,7 @@ } }, { - "accuracy": 0.996886504323859, + "accuracy": 0.9968959849916006, "total_bits": 401557504, "q_proj": { "group_size": { @@ -15866,7 +15866,7 @@ } }, { - "accuracy": 0.9970055940119844, + "accuracy": 0.9970095934052217, "total_bits": 475279360, "q_proj": { "group_size": { @@ -15930,7 +15930,7 @@ } }, { - "accuracy": 0.9970139653274888, + "accuracy": 0.9968864519737268, "total_bits": 475479040, "q_proj": { "group_size": { @@ -15994,7 +15994,7 @@ } }, { - "accuracy": 0.997854873930153, + "accuracy": 0.9978407048865369, "total_bits": 609759232, "q_proj": { "group_size": { @@ -16046,7 +16046,7 @@ } }, { - "accuracy": 0.9978736466483066, + "accuracy": 0.9978831371194438, "total_bits": 610024448, "q_proj": { "group_size": { @@ -16098,7 +16098,7 @@ } }, { - "accuracy": 0.9980307574334898, + "accuracy": 0.9979855528003291, "total_bits": 615020544, "q_proj": { "group_size": { @@ -16150,7 +16150,7 @@ } }, { - "accuracy": 0.9980530983915454, + "accuracy": 0.9980073488856617, "total_bits": 623951872, "q_proj": { "group_size": { @@ -16202,7 +16202,7 @@ } }, { - "accuracy": 0.9979243370655336, + "accuracy": 0.9979349041455671, "total_bits": 626473984, "q_proj": { "group_size": { @@ -16266,7 +16266,7 @@ } }, { - "accuracy": 0.9980539373661342, + "accuracy": 0.9980407414075575, "total_bits": 630355968, "q_proj": { "group_size": { @@ -16330,7 +16330,7 @@ } }, { - "accuracy": 0.9986467902597628, + "accuracy": 0.9986624246169078, "total_bits": 637362176, "q_proj": { "group_size": { @@ -16391,7 +16391,7 @@ } }, { - "accuracy": 0.9987475105414265, + "accuracy": 0.9987859115200607, "total_bits": 646823936, "q_proj": { "group_size": { @@ -16452,7 +16452,7 @@ } }, { - "accuracy": 0.9992625568258134, + "accuracy": 0.99925675831343, "total_bits": 784740352, "q_proj": { "group_size": { @@ -16513,7 +16513,7 @@ } }, { - "accuracy": 0.9993577986759575, + "accuracy": 0.9993705992145758, "total_bits": 797818880, "q_proj": { "group_size": { @@ -16574,7 +16574,7 @@ } }, { - "accuracy": 0.9993324039111796, + "accuracy": 0.9993268893051305, "total_bits": 911749120, "q_proj": { "group_size": { @@ -16626,7 +16626,7 @@ } }, { - "accuracy": 0.9997747827223257, + "accuracy": 0.9997756936293292, "total_bits": 942718976, "q_proj": { "group_size": { @@ -16678,7 +16678,7 @@ } }, { - "accuracy": 0.9998291633459494, + "accuracy": 0.999829723683529, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -16732,7 +16732,7 @@ ], "model.layers.8.mlp": [ { - "accuracy": 0.9916414014602962, + "accuracy": 0.9917126399906058, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -16784,7 +16784,7 @@ } }, { - "accuracy": 0.9918600062006399, + "accuracy": 0.9919314243291554, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -16836,7 +16836,7 @@ } }, { - "accuracy": 0.9930947877858814, + "accuracy": 0.993153336801027, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -16885,7 +16885,7 @@ } }, { - "accuracy": 0.9934752458020261, + "accuracy": 0.9935299958053389, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -16934,7 +16934,7 @@ } }, { - "accuracy": 0.9954490732205542, + "accuracy": 0.9954804009512851, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -16986,7 +16986,7 @@ } }, { - "accuracy": 0.9962883172066588, + "accuracy": 0.99632097545423, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -17038,7 +17038,7 @@ } }, { - "accuracy": 0.9968383027926871, + "accuracy": 0.9968657656327674, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -17087,7 +17087,7 @@ } }, { - "accuracy": 0.9977687216903034, + "accuracy": 0.9977901048566166, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -17130,7 +17130,7 @@ } }, { - "accuracy": 0.9981181515674842, + "accuracy": 0.9981356196497616, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -17173,7 +17173,7 @@ } }, { - "accuracy": 0.9975965252440226, + "accuracy": 0.9976174011826515, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -17225,7 +17225,7 @@ } }, { - "accuracy": 0.9982166609873897, + "accuracy": 0.9982323324994037, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -17277,7 +17277,7 @@ } }, { - "accuracy": 0.9987300439296585, + "accuracy": 0.9987444360985568, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -17329,7 +17329,7 @@ } }, { - "accuracy": 0.9991111864188784, + "accuracy": 0.9991189496297586, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -17381,7 +17381,7 @@ } }, { - "accuracy": 0.9993077591551762, + "accuracy": 0.9993154117347378, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -17424,7 +17424,7 @@ } }, { - "accuracy": 0.9993301606001822, + "accuracy": 0.9993372915410682, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -17473,7 +17473,7 @@ } }, { - "accuracy": 0.9994121470831727, + "accuracy": 0.9994186129617063, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -17519,7 +17519,7 @@ } }, { - "accuracy": 0.9998134470324179, + "accuracy": 0.9998155968435305, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -17561,7 +17561,7 @@ ], "model.layers.9.self_attn": [ { - "accuracy": 0.99367661068314, + "accuracy": 0.993885811222227, "total_bits": 320757760, "q_proj": { "group_size": { @@ -17625,7 +17625,7 @@ } }, { - "accuracy": 0.9940337973989939, + "accuracy": 0.9942029200886425, "total_bits": 329080832, "q_proj": { "group_size": { @@ -17689,7 +17689,7 @@ } }, { - "accuracy": 0.9948633705314837, + "accuracy": 0.9948337125150781, "total_bits": 336024576, "q_proj": { "group_size": { @@ -17753,7 +17753,7 @@ } }, { - "accuracy": 0.9958636486216595, + "accuracy": 0.9958468714826986, "total_bits": 401557504, "q_proj": { "group_size": { @@ -17817,7 +17817,7 @@ } }, { - "accuracy": 0.9959689135614195, + "accuracy": 0.9959477735193152, "total_bits": 475279360, "q_proj": { "group_size": { @@ -17881,7 +17881,7 @@ } }, { - "accuracy": 0.9960401952266693, + "accuracy": 0.9960239317856336, "total_bits": 475479040, "q_proj": { "group_size": { @@ -17945,7 +17945,7 @@ } }, { - "accuracy": 0.9968821506359076, + "accuracy": 0.9969091846754676, "total_bits": 609759232, "q_proj": { "group_size": { @@ -17997,7 +17997,7 @@ } }, { - "accuracy": 0.9970097112420359, + "accuracy": 0.9970640945983561, "total_bits": 610024448, "q_proj": { "group_size": { @@ -18049,7 +18049,7 @@ } }, { - "accuracy": 0.9971614987834504, + "accuracy": 0.9972578657692984, "total_bits": 615020544, "q_proj": { "group_size": { @@ -18101,7 +18101,7 @@ } }, { - "accuracy": 0.9971879026607463, + "accuracy": 0.9972820640786698, "total_bits": 623951872, "q_proj": { "group_size": { @@ -18153,7 +18153,7 @@ } }, { - "accuracy": 0.997099052330381, + "accuracy": 0.9971548187497415, "total_bits": 626473984, "q_proj": { "group_size": { @@ -18217,7 +18217,7 @@ } }, { - "accuracy": 0.9971843030499784, + "accuracy": 0.997298289090395, "total_bits": 630355968, "q_proj": { "group_size": { @@ -18281,7 +18281,7 @@ } }, { - "accuracy": 0.9980685058000841, + "accuracy": 0.9981724233611634, "total_bits": 637362176, "q_proj": { "group_size": { @@ -18342,7 +18342,7 @@ } }, { - "accuracy": 0.9981891479539243, + "accuracy": 0.9982493094315654, "total_bits": 646823936, "q_proj": { "group_size": { @@ -18403,7 +18403,7 @@ } }, { - "accuracy": 0.9989304523915052, + "accuracy": 0.9989301515252966, "total_bits": 784740352, "q_proj": { "group_size": { @@ -18464,7 +18464,7 @@ } }, { - "accuracy": 0.999078535033684, + "accuracy": 0.9990960391924569, "total_bits": 797818880, "q_proj": { "group_size": { @@ -18525,7 +18525,7 @@ } }, { - "accuracy": 0.9990054355248025, + "accuracy": 0.9990065645818648, "total_bits": 911749120, "q_proj": { "group_size": { @@ -18577,7 +18577,7 @@ } }, { - "accuracy": 0.9996963666850015, + "accuracy": 0.9997025538039835, "total_bits": 942718976, "q_proj": { "group_size": { @@ -18629,7 +18629,7 @@ } }, { - "accuracy": 0.9997399343226693, + "accuracy": 0.9997420995251128, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -18683,7 +18683,7 @@ ], "model.layers.9.mlp": [ { - "accuracy": 0.9905079022834176, + "accuracy": 0.9905876023204703, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -18735,7 +18735,7 @@ } }, { - "accuracy": 0.9907229597631254, + "accuracy": 0.9908057843383989, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -18787,7 +18787,7 @@ } }, { - "accuracy": 0.9921448967958751, + "accuracy": 0.9922125237552744, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -18836,7 +18836,7 @@ } }, { - "accuracy": 0.9925836167837444, + "accuracy": 0.992646408708472, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -18885,7 +18885,7 @@ } }, { - "accuracy": 0.9946720027609876, + "accuracy": 0.9947123182447333, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -18937,7 +18937,7 @@ } }, { - "accuracy": 0.9957814444052545, + "accuracy": 0.995817062101866, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -18989,7 +18989,7 @@ } }, { - "accuracy": 0.9964260678542288, + "accuracy": 0.9964563023102911, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -19038,7 +19038,7 @@ } }, { - "accuracy": 0.9971779475086614, + "accuracy": 0.9972007992236238, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -19081,7 +19081,7 @@ } }, { - "accuracy": 0.9977819537253756, + "accuracy": 0.9977994429829874, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -19124,7 +19124,7 @@ } }, { - "accuracy": 0.9967337009546003, + "accuracy": 0.9967580372957807, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -19176,7 +19176,7 @@ } }, { - "accuracy": 0.9979114179548464, + "accuracy": 0.997927107897244, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -19228,7 +19228,7 @@ } }, { - "accuracy": 0.9978288196419415, + "accuracy": 0.9978441321536115, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -19280,7 +19280,7 @@ } }, { - "accuracy": 0.9989347761791003, + "accuracy": 0.9989418286438051, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -19332,7 +19332,7 @@ } }, { - "accuracy": 0.99849335251278, + "accuracy": 0.9985027868104609, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -19375,7 +19375,7 @@ } }, { - "accuracy": 0.9985270616843512, + "accuracy": 0.9985390989796111, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -19424,7 +19424,7 @@ } }, { - "accuracy": 0.9985755756497383, + "accuracy": 0.9985870808166893, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -19470,7 +19470,7 @@ } }, { - "accuracy": 0.9993665698719653, + "accuracy": 0.9993738010525703, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -19512,7 +19512,7 @@ ], "model.layers.10.self_attn": [ { - "accuracy": 0.9925632994425925, + "accuracy": 0.992557102128079, "total_bits": 320757760, "q_proj": { "group_size": { @@ -19576,7 +19576,7 @@ } }, { - "accuracy": 0.9928731314445797, + "accuracy": 0.9929411874005669, "total_bits": 329080832, "q_proj": { "group_size": { @@ -19640,7 +19640,7 @@ } }, { - "accuracy": 0.9938257050357366, + "accuracy": 0.9939839212517989, "total_bits": 336024576, "q_proj": { "group_size": { @@ -19704,7 +19704,7 @@ } }, { - "accuracy": 0.9946374406940058, + "accuracy": 0.9948562873821509, "total_bits": 401557504, "q_proj": { "group_size": { @@ -19768,7 +19768,7 @@ } }, { - "accuracy": 0.994878737157897, + "accuracy": 0.9950894885941556, "total_bits": 475279360, "q_proj": { "group_size": { @@ -19832,7 +19832,7 @@ } }, { - "accuracy": 0.9950146580997267, + "accuracy": 0.9949907703619254, "total_bits": 475479040, "q_proj": { "group_size": { @@ -19896,7 +19896,7 @@ } }, { - "accuracy": 0.996351024822185, + "accuracy": 0.9963577201491908, "total_bits": 609759232, "q_proj": { "group_size": { @@ -19948,7 +19948,7 @@ } }, { - "accuracy": 0.9965442850401527, + "accuracy": 0.9964847886248639, "total_bits": 610024448, "q_proj": { "group_size": { @@ -20000,7 +20000,7 @@ } }, { - "accuracy": 0.9966494386133394, + "accuracy": 0.996738993415707, "total_bits": 615020544, "q_proj": { "group_size": { @@ -20052,7 +20052,7 @@ } }, { - "accuracy": 0.9966718056484273, + "accuracy": 0.9967624786260881, "total_bits": 623951872, "q_proj": { "group_size": { @@ -20104,7 +20104,7 @@ } }, { - "accuracy": 0.9965653780259585, + "accuracy": 0.9966205342819816, "total_bits": 626473984, "q_proj": { "group_size": { @@ -20168,7 +20168,7 @@ } }, { - "accuracy": 0.9968076558097413, + "accuracy": 0.9968052486840048, "total_bits": 630355968, "q_proj": { "group_size": { @@ -20232,7 +20232,7 @@ } }, { - "accuracy": 0.9979067834976473, + "accuracy": 0.9979136590110628, "total_bits": 637362176, "q_proj": { "group_size": { @@ -20293,7 +20293,7 @@ } }, { - "accuracy": 0.9980666447234782, + "accuracy": 0.9980803508507577, "total_bits": 646823936, "q_proj": { "group_size": { @@ -20354,7 +20354,7 @@ } }, { - "accuracy": 0.998860811324496, + "accuracy": 0.9988739949307943, "total_bits": 784740352, "q_proj": { "group_size": { @@ -20415,7 +20415,7 @@ } }, { - "accuracy": 0.999007496688711, + "accuracy": 0.999053506278678, "total_bits": 797818880, "q_proj": { "group_size": { @@ -20476,7 +20476,7 @@ } }, { - "accuracy": 0.9989433708159547, + "accuracy": 0.9989569059720165, "total_bits": 911749120, "q_proj": { "group_size": { @@ -20528,7 +20528,7 @@ } }, { - "accuracy": 0.9996681862608775, + "accuracy": 0.9996639552379125, "total_bits": 942718976, "q_proj": { "group_size": { @@ -20580,7 +20580,7 @@ } }, { - "accuracy": 0.9997258163792523, + "accuracy": 0.999728736345117, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -20634,7 +20634,7 @@ ], "model.layers.10.mlp": [ { - "accuracy": 0.9890653977268621, + "accuracy": 0.9891637673503474, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -20686,7 +20686,7 @@ } }, { - "accuracy": 0.9893526027077123, + "accuracy": 0.9894464172815022, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -20738,7 +20738,7 @@ } }, { - "accuracy": 0.9911234018049742, + "accuracy": 0.9911979766268479, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -20787,7 +20787,7 @@ } }, { - "accuracy": 0.9916554516867587, + "accuracy": 0.9917244432788146, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -20836,7 +20836,7 @@ } }, { - "accuracy": 0.9945794763533693, + "accuracy": 0.9946250084199404, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -20888,7 +20888,7 @@ } }, { - "accuracy": 0.995022004764331, + "accuracy": 0.9950658341771678, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -20940,7 +20940,7 @@ } }, { - "accuracy": 0.995781550674062, + "accuracy": 0.9958176934405377, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -20989,7 +20989,7 @@ } }, { - "accuracy": 0.9972204081714153, + "accuracy": 0.997244491977127, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -21032,7 +21032,7 @@ } }, { - "accuracy": 0.9974778825907331, + "accuracy": 0.9974990500823447, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -21075,7 +21075,7 @@ } }, { - "accuracy": 0.9972591831495887, + "accuracy": 0.9972836943833452, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -21127,7 +21127,7 @@ } }, { - "accuracy": 0.9976079271812188, + "accuracy": 0.9976292579973999, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -21179,7 +21179,7 @@ } }, { - "accuracy": 0.9986091293394566, + "accuracy": 0.9986210421315933, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -21231,7 +21231,7 @@ } }, { - "accuracy": 0.9988122340291739, + "accuracy": 0.9988226730768618, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -21283,7 +21283,7 @@ } }, { - "accuracy": 0.9992565329333669, + "accuracy": 0.9992630435644012, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -21326,7 +21326,7 @@ } }, { - "accuracy": 0.9992922431250152, + "accuracy": 0.9992981465827477, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -21375,7 +21375,7 @@ } }, { - "accuracy": 0.9994425452559402, + "accuracy": 0.9994467543438077, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -21421,7 +21421,7 @@ } }, { - "accuracy": 0.9998026900606132, + "accuracy": 0.9998043577284798, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -21463,7 +21463,7 @@ ], "model.layers.11.self_attn": [ { - "accuracy": 0.9927921718672702, + "accuracy": 0.9928815858928781, "total_bits": 320757760, "q_proj": { "group_size": { @@ -21527,7 +21527,7 @@ } }, { - "accuracy": 0.9930834903528816, + "accuracy": 0.9931743043033701, "total_bits": 329080832, "q_proj": { "group_size": { @@ -21591,7 +21591,7 @@ } }, { - "accuracy": 0.9939535432740262, + "accuracy": 0.9939732308450498, "total_bits": 336024576, "q_proj": { "group_size": { @@ -21655,7 +21655,7 @@ } }, { - "accuracy": 0.9949855514262852, + "accuracy": 0.9949889967316076, "total_bits": 401557504, "q_proj": { "group_size": { @@ -21719,7 +21719,7 @@ } }, { - "accuracy": 0.9953266758667795, + "accuracy": 0.995335533038566, "total_bits": 475279360, "q_proj": { "group_size": { @@ -21783,7 +21783,7 @@ } }, { - "accuracy": 0.9953772700146625, + "accuracy": 0.9954819985126194, "total_bits": 475479040, "q_proj": { "group_size": { @@ -21847,7 +21847,7 @@ } }, { - "accuracy": 0.9965916844575029, + "accuracy": 0.9966621328341333, "total_bits": 609759232, "q_proj": { "group_size": { @@ -21899,7 +21899,7 @@ } }, { - "accuracy": 0.9966645832908781, + "accuracy": 0.9967162028739327, "total_bits": 610024448, "q_proj": { "group_size": { @@ -21951,7 +21951,7 @@ } }, { - "accuracy": 0.9967724626001558, + "accuracy": 0.9967889079922124, "total_bits": 615020544, "q_proj": { "group_size": { @@ -22003,7 +22003,7 @@ } }, { - "accuracy": 0.9967999072059205, + "accuracy": 0.9968107256450152, "total_bits": 623951872, "q_proj": { "group_size": { @@ -22055,7 +22055,7 @@ } }, { - "accuracy": 0.9967029337820253, + "accuracy": 0.9967217466941005, "total_bits": 626473984, "q_proj": { "group_size": { @@ -22119,7 +22119,7 @@ } }, { - "accuracy": 0.9968336683354879, + "accuracy": 0.9968090763217524, "total_bits": 630355968, "q_proj": { "group_size": { @@ -22183,7 +22183,7 @@ } }, { - "accuracy": 0.9976443597360661, + "accuracy": 0.9975678607037193, "total_bits": 637362176, "q_proj": { "group_size": { @@ -22244,7 +22244,7 @@ } }, { - "accuracy": 0.9976459355338624, + "accuracy": 0.997695726783652, "total_bits": 646823936, "q_proj": { "group_size": { @@ -22305,7 +22305,7 @@ } }, { - "accuracy": 0.9984135402268485, + "accuracy": 0.9984387382864952, "total_bits": 784740352, "q_proj": { "group_size": { @@ -22366,7 +22366,7 @@ } }, { - "accuracy": 0.9984892486153465, + "accuracy": 0.9984951686898345, "total_bits": 797818880, "q_proj": { "group_size": { @@ -22427,7 +22427,7 @@ } }, { - "accuracy": 0.9984811155223533, + "accuracy": 0.9985082124016786, "total_bits": 911749120, "q_proj": { "group_size": { @@ -22479,7 +22479,7 @@ } }, { - "accuracy": 0.9993997096623245, + "accuracy": 0.9994026034286148, "total_bits": 942718976, "q_proj": { "group_size": { @@ -22531,7 +22531,7 @@ } }, { - "accuracy": 0.9994411038136796, + "accuracy": 0.999446891493311, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -22585,7 +22585,7 @@ ], "model.layers.11.mlp": [ { - "accuracy": 0.9871353585469095, + "accuracy": 0.9872385835961291, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -22637,7 +22637,7 @@ } }, { - "accuracy": 0.9874652035926518, + "accuracy": 0.9875653041036505, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -22689,7 +22689,7 @@ } }, { - "accuracy": 0.9893654114321658, + "accuracy": 0.9894522640265917, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -22738,7 +22738,7 @@ } }, { - "accuracy": 0.9899384669567409, + "accuracy": 0.9900198139642414, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -22787,7 +22787,7 @@ } }, { - "accuracy": 0.9933747076674512, + "accuracy": 0.99342711034574, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -22839,7 +22839,7 @@ } }, { - "accuracy": 0.9942187470825095, + "accuracy": 0.9942652622335836, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -22891,7 +22891,7 @@ } }, { - "accuracy": 0.9950555010061515, + "accuracy": 0.9950940059988123, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -22940,7 +22940,7 @@ } }, { - "accuracy": 0.9966970635087866, + "accuracy": 0.9967234905220961, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -22983,7 +22983,7 @@ } }, { - "accuracy": 0.9970754216376104, + "accuracy": 0.9970992278111609, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -23026,7 +23026,7 @@ } }, { - "accuracy": 0.9966285456168024, + "accuracy": 0.9966547547986633, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -23078,7 +23078,7 @@ } }, { - "accuracy": 0.9972234862415414, + "accuracy": 0.9972468447920523, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -23130,7 +23130,7 @@ } }, { - "accuracy": 0.9982667369277853, + "accuracy": 0.9982804685438934, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -23182,7 +23182,7 @@ } }, { - "accuracy": 0.9986202906032926, + "accuracy": 0.998631624210822, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -23234,7 +23234,7 @@ } }, { - "accuracy": 0.9990742141871076, + "accuracy": 0.9990812518486851, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -23277,7 +23277,7 @@ } }, { - "accuracy": 0.9991118788327041, + "accuracy": 0.9991190388406578, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -23326,7 +23326,7 @@ } }, { - "accuracy": 0.9992574477372201, + "accuracy": 0.9992630303298172, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -23372,7 +23372,7 @@ } }, { - "accuracy": 0.9997536544197876, + "accuracy": 0.9997555771842599, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -23414,7 +23414,7 @@ ], "model.layers.12.self_attn": [ { - "accuracy": 0.9898167527035663, + "accuracy": 0.9897953758114263, "total_bits": 320757760, "q_proj": { "group_size": { @@ -23478,7 +23478,7 @@ } }, { - "accuracy": 0.9899916241043493, + "accuracy": 0.989654372397222, "total_bits": 329080832, "q_proj": { "group_size": { @@ -23542,7 +23542,7 @@ } }, { - "accuracy": 0.9918214568966314, + "accuracy": 0.9917115067180834, "total_bits": 336024576, "q_proj": { "group_size": { @@ -23606,7 +23606,7 @@ } }, { - "accuracy": 0.9928167242752878, + "accuracy": 0.992663888554824, "total_bits": 401557504, "q_proj": { "group_size": { @@ -23670,7 +23670,7 @@ } }, { - "accuracy": 0.9930261593115958, + "accuracy": 0.9928704821749738, "total_bits": 475279360, "q_proj": { "group_size": { @@ -23734,7 +23734,7 @@ } }, { - "accuracy": 0.9930841609051353, + "accuracy": 0.9929192372058567, "total_bits": 475479040, "q_proj": { "group_size": { @@ -23798,7 +23798,7 @@ } }, { - "accuracy": 0.9953185049326796, + "accuracy": 0.9953464555897211, "total_bits": 609759232, "q_proj": { "group_size": { @@ -23850,7 +23850,7 @@ } }, { - "accuracy": 0.9955663681030273, + "accuracy": 0.9955683248607736, "total_bits": 610024448, "q_proj": { "group_size": { @@ -23902,7 +23902,7 @@ } }, { - "accuracy": 0.9959602320664808, + "accuracy": 0.9960169560815159, "total_bits": 615020544, "q_proj": { "group_size": { @@ -23954,7 +23954,7 @@ } }, { - "accuracy": 0.9959933185263684, + "accuracy": 0.9960443722574335, "total_bits": 623951872, "q_proj": { "group_size": { @@ -24006,7 +24006,7 @@ } }, { - "accuracy": 0.9956397798500563, + "accuracy": 0.9957943856716156, "total_bits": 626473984, "q_proj": { "group_size": { @@ -24070,7 +24070,7 @@ } }, { - "accuracy": 0.9960517134321364, + "accuracy": 0.9960526137759811, "total_bits": 630355968, "q_proj": { "group_size": { @@ -24134,7 +24134,7 @@ } }, { - "accuracy": 0.997583421241296, + "accuracy": 0.9975772560819199, "total_bits": 637362176, "q_proj": { "group_size": { @@ -24195,7 +24195,7 @@ } }, { - "accuracy": 0.9977563772546617, + "accuracy": 0.9977612834619848, "total_bits": 646823936, "q_proj": { "group_size": { @@ -24256,7 +24256,7 @@ } }, { - "accuracy": 0.99867605074848, + "accuracy": 0.9986907335880556, "total_bits": 784740352, "q_proj": { "group_size": { @@ -24317,7 +24317,7 @@ } }, { - "accuracy": 0.9988962203068169, + "accuracy": 0.9988828225943603, "total_bits": 797818880, "q_proj": { "group_size": { @@ -24378,7 +24378,7 @@ } }, { - "accuracy": 0.9987720803014541, + "accuracy": 0.9987858283872667, "total_bits": 911749120, "q_proj": { "group_size": { @@ -24430,7 +24430,7 @@ } }, { - "accuracy": 0.9996114436438993, + "accuracy": 0.9996146540598649, "total_bits": 942718976, "q_proj": { "group_size": { @@ -24482,7 +24482,7 @@ } }, { - "accuracy": 0.999690115917474, + "accuracy": 0.9996917817349497, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -24536,7 +24536,7 @@ ], "model.layers.12.mlp": [ { - "accuracy": 0.9861031911875072, + "accuracy": 0.9862075338238164, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -24588,7 +24588,7 @@ } }, { - "accuracy": 0.9864495631895567, + "accuracy": 0.9865482229935495, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -24640,7 +24640,7 @@ } }, { - "accuracy": 0.988624568832548, + "accuracy": 0.988704651594162, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -24689,7 +24689,7 @@ } }, { - "accuracy": 0.9892844565604862, + "accuracy": 0.989358440041542, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -24738,7 +24738,7 @@ } }, { - "accuracy": 0.9931681383597223, + "accuracy": 0.9932207041665128, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -24790,7 +24790,7 @@ } }, { - "accuracy": 0.9937302254532513, + "accuracy": 0.9937779209331462, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -24842,7 +24842,7 @@ } }, { - "accuracy": 0.9946804830902501, + "accuracy": 0.9947197127499079, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -24891,7 +24891,7 @@ } }, { - "accuracy": 0.9965048800957831, + "accuracy": 0.9965322264715245, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -24934,7 +24934,7 @@ } }, { - "accuracy": 0.9968252907458105, + "accuracy": 0.9968491884830751, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -24977,7 +24977,7 @@ } }, { - "accuracy": 0.9965364729103289, + "accuracy": 0.9965642443613002, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -25029,7 +25029,7 @@ } }, { - "accuracy": 0.9969865967960734, + "accuracy": 0.9970094014547373, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -25081,7 +25081,7 @@ } }, { - "accuracy": 0.9982362497401865, + "accuracy": 0.9982500882133057, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -25133,7 +25133,7 @@ } }, { - "accuracy": 0.9985009804368019, + "accuracy": 0.9985130016544932, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -25185,7 +25185,7 @@ } }, { - "accuracy": 0.9990543274110869, + "accuracy": 0.9990619971956077, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -25228,7 +25228,7 @@ } }, { - "accuracy": 0.9990994323437151, + "accuracy": 0.9991069900754251, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -25277,7 +25277,7 @@ } }, { - "accuracy": 0.9992827256455233, + "accuracy": 0.9992881549130145, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -25323,7 +25323,7 @@ } }, { - "accuracy": 0.9997427096394332, + "accuracy": 0.9997446544860539, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -25365,7 +25365,7 @@ ], "model.layers.13.self_attn": [ { - "accuracy": 0.9863003586467943, + "accuracy": 0.9864929585080398, "total_bits": 320757760, "q_proj": { "group_size": { @@ -25429,7 +25429,7 @@ } }, { - "accuracy": 0.9866981961225209, + "accuracy": 0.9869350249830046, "total_bits": 329080832, "q_proj": { "group_size": { @@ -25493,7 +25493,7 @@ } }, { - "accuracy": 0.9881218342404616, + "accuracy": 0.9882042235449741, "total_bits": 336024576, "q_proj": { "group_size": { @@ -25557,7 +25557,7 @@ } }, { - "accuracy": 0.990382161579634, + "accuracy": 0.9904358990882572, "total_bits": 401557504, "q_proj": { "group_size": { @@ -25621,7 +25621,7 @@ } }, { - "accuracy": 0.991610262739031, + "accuracy": 0.9915859550237656, "total_bits": 475279360, "q_proj": { "group_size": { @@ -25685,7 +25685,7 @@ } }, { - "accuracy": 0.9916592569727647, + "accuracy": 0.9915655357273001, "total_bits": 475479040, "q_proj": { "group_size": { @@ -25749,7 +25749,7 @@ } }, { - "accuracy": 0.9945553301980621, + "accuracy": 0.9944440051913261, "total_bits": 609759232, "q_proj": { "group_size": { @@ -25801,7 +25801,7 @@ } }, { - "accuracy": 0.9947441988869717, + "accuracy": 0.9949021574697996, "total_bits": 610024448, "q_proj": { "group_size": { @@ -25853,7 +25853,7 @@ } }, { - "accuracy": 0.9950469324463292, + "accuracy": 0.9951630383729935, "total_bits": 615020544, "q_proj": { "group_size": { @@ -25905,7 +25905,7 @@ } }, { - "accuracy": 0.9951449487554399, + "accuracy": 0.9952442375452895, "total_bits": 623951872, "q_proj": { "group_size": { @@ -25957,7 +25957,7 @@ } }, { - "accuracy": 0.9948465545710764, + "accuracy": 0.9949888257603896, "total_bits": 626473984, "q_proj": { "group_size": { @@ -26021,7 +26021,7 @@ } }, { - "accuracy": 0.9952640455020102, + "accuracy": 0.995261304864758, "total_bits": 630355968, "q_proj": { "group_size": { @@ -26085,7 +26085,7 @@ } }, { - "accuracy": 0.9968082371510958, + "accuracy": 0.9967758812402424, "total_bits": 637362176, "q_proj": { "group_size": { @@ -26146,7 +26146,7 @@ } }, { - "accuracy": 0.9969947620442039, + "accuracy": 0.9970574067219308, "total_bits": 646823936, "q_proj": { "group_size": { @@ -26207,7 +26207,7 @@ } }, { - "accuracy": 0.9982578781874556, + "accuracy": 0.9982592771319967, "total_bits": 784740352, "q_proj": { "group_size": { @@ -26268,7 +26268,7 @@ } }, { - "accuracy": 0.9984774811095313, + "accuracy": 0.9985162071687611, "total_bits": 797818880, "q_proj": { "group_size": { @@ -26329,7 +26329,7 @@ } }, { - "accuracy": 0.9985350281196205, + "accuracy": 0.9985318672108022, "total_bits": 911749120, "q_proj": { "group_size": { @@ -26381,7 +26381,7 @@ } }, { - "accuracy": 0.9993929364473412, + "accuracy": 0.9994098225020265, "total_bits": 942718976, "q_proj": { "group_size": { @@ -26433,7 +26433,7 @@ } }, { - "accuracy": 0.9996192598666408, + "accuracy": 0.9996145207581943, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -26487,7 +26487,7 @@ ], "model.layers.13.mlp": [ { - "accuracy": 0.9851991691087422, + "accuracy": 0.9853108086084065, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -26539,7 +26539,7 @@ } }, { - "accuracy": 0.9855290212129292, + "accuracy": 0.9856348178888622, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -26591,7 +26591,7 @@ } }, { - "accuracy": 0.9881703437943208, + "accuracy": 0.9882560740960272, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -26640,7 +26640,7 @@ } }, { - "accuracy": 0.9889714655123258, + "accuracy": 0.9890499797306562, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -26689,7 +26689,7 @@ } }, { - "accuracy": 0.9926948547363281, + "accuracy": 0.992751272885423, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -26741,7 +26741,7 @@ } }, { - "accuracy": 0.9932831494431746, + "accuracy": 0.9933325303228278, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -26793,7 +26793,7 @@ } }, { - "accuracy": 0.9944302690656561, + "accuracy": 0.9944704237737154, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -26842,7 +26842,7 @@ } }, { - "accuracy": 0.9962632848243964, + "accuracy": 0.99629166879152, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -26885,7 +26885,7 @@ } }, { - "accuracy": 0.9966198107913921, + "accuracy": 0.9966453278535291, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -26928,7 +26928,7 @@ } }, { - "accuracy": 0.9963056598054735, + "accuracy": 0.9963337390830642, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -26980,7 +26980,7 @@ } }, { - "accuracy": 0.9967768729517335, + "accuracy": 0.9968006654005301, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -27032,7 +27032,7 @@ } }, { - "accuracy": 0.9981228097489006, + "accuracy": 0.9981374860202011, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -27084,7 +27084,7 @@ } }, { - "accuracy": 0.9983963529138189, + "accuracy": 0.9984084915761885, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -27136,7 +27136,7 @@ } }, { - "accuracy": 0.9990010022332794, + "accuracy": 0.9990086367255763, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -27179,7 +27179,7 @@ } }, { - "accuracy": 0.9990452372125889, + "accuracy": 0.9990528986642235, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -27228,7 +27228,7 @@ } }, { - "accuracy": 0.999270739720056, + "accuracy": 0.9992767074390462, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -27274,7 +27274,7 @@ } }, { - "accuracy": 0.9997328612218169, + "accuracy": 0.9997349119696178, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -27316,7 +27316,7 @@ ], "model.layers.14.self_attn": [ { - "accuracy": 0.9843504570032421, + "accuracy": 0.984514992488058, "total_bits": 320757760, "q_proj": { "group_size": { @@ -27380,7 +27380,7 @@ } }, { - "accuracy": 0.9848180397560722, + "accuracy": 0.984938100764626, "total_bits": 329080832, "q_proj": { "group_size": { @@ -27444,7 +27444,7 @@ } }, { - "accuracy": 0.985895596052471, + "accuracy": 0.9858693370693609, "total_bits": 336024576, "q_proj": { "group_size": { @@ -27508,7 +27508,7 @@ } }, { - "accuracy": 0.9881298243999481, + "accuracy": 0.9880493659722177, "total_bits": 401557504, "q_proj": { "group_size": { @@ -27572,7 +27572,7 @@ } }, { - "accuracy": 0.9903234415932706, + "accuracy": 0.9902088477423316, "total_bits": 475279360, "q_proj": { "group_size": { @@ -27636,7 +27636,7 @@ } }, { - "accuracy": 0.9907533346038115, + "accuracy": 0.9904723512498956, "total_bits": 475479040, "q_proj": { "group_size": { @@ -27700,7 +27700,7 @@ } }, { - "accuracy": 0.9939921306152093, + "accuracy": 0.9940126667681494, "total_bits": 609759232, "q_proj": { "group_size": { @@ -27752,7 +27752,7 @@ } }, { - "accuracy": 0.9940372768201327, + "accuracy": 0.9941352339167344, "total_bits": 610024448, "q_proj": { "group_size": { @@ -27804,7 +27804,7 @@ } }, { - "accuracy": 0.9944445914343784, + "accuracy": 0.9945806555057827, "total_bits": 615020544, "q_proj": { "group_size": { @@ -27856,7 +27856,7 @@ } }, { - "accuracy": 0.9945611722375217, + "accuracy": 0.9946646329603697, "total_bits": 623951872, "q_proj": { "group_size": { @@ -27908,7 +27908,7 @@ } }, { - "accuracy": 0.9943394578601185, + "accuracy": 0.9942714536660596, "total_bits": 626473984, "q_proj": { "group_size": { @@ -27972,7 +27972,7 @@ } }, { - "accuracy": 0.9946683335461115, + "accuracy": 0.9946511278026983, "total_bits": 630355968, "q_proj": { "group_size": { @@ -28036,7 +28036,7 @@ } }, { - "accuracy": 0.9963634139613101, + "accuracy": 0.9963231702384195, "total_bits": 637362176, "q_proj": { "group_size": { @@ -28097,7 +28097,7 @@ } }, { - "accuracy": 0.9966878836092196, + "accuracy": 0.9967003990160791, "total_bits": 646823936, "q_proj": { "group_size": { @@ -28158,7 +28158,7 @@ } }, { - "accuracy": 0.9980671970467818, + "accuracy": 0.9980880449476995, "total_bits": 784740352, "q_proj": { "group_size": { @@ -28219,7 +28219,7 @@ } }, { - "accuracy": 0.998313349328543, + "accuracy": 0.9983328231855443, "total_bits": 797818880, "q_proj": { "group_size": { @@ -28280,7 +28280,7 @@ } }, { - "accuracy": 0.9983648358795204, + "accuracy": 0.9983832818505011, "total_bits": 911749120, "q_proj": { "group_size": { @@ -28332,7 +28332,7 @@ } }, { - "accuracy": 0.999311258036055, + "accuracy": 0.9993251724364726, "total_bits": 942718976, "q_proj": { "group_size": { @@ -28384,7 +28384,7 @@ } }, { - "accuracy": 0.9995527065996277, + "accuracy": 0.9995632869142451, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -28438,7 +28438,7 @@ ], "model.layers.14.mlp": [ { - "accuracy": 0.9823967704647466, + "accuracy": 0.982513521846972, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -28490,7 +28490,7 @@ } }, { - "accuracy": 0.9828004742923536, + "accuracy": 0.9829121043807582, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -28542,7 +28542,7 @@ } }, { - "accuracy": 0.985889921062871, + "accuracy": 0.9859807224650132, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -28591,7 +28591,7 @@ } }, { - "accuracy": 0.9868271742996416, + "accuracy": 0.9869100467154854, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -28640,7 +28640,7 @@ } }, { - "accuracy": 0.9912931973996916, + "accuracy": 0.9913491481228879, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -28692,7 +28692,7 @@ } }, { - "accuracy": 0.9919838850435457, + "accuracy": 0.9920360979280973, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -28744,7 +28744,7 @@ } }, { - "accuracy": 0.9933211936762458, + "accuracy": 0.9933646548735467, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -28793,7 +28793,7 @@ } }, { - "accuracy": 0.9955386021419576, + "accuracy": 0.9955689142409124, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -28836,7 +28836,7 @@ } }, { - "accuracy": 0.9959609030108703, + "accuracy": 0.9959886807360148, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -28879,7 +28879,7 @@ } }, { - "accuracy": 0.9955926341445822, + "accuracy": 0.9956214890668267, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -28931,7 +28931,7 @@ } }, { - "accuracy": 0.99614966190175, + "accuracy": 0.99617558404019, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -28983,7 +28983,7 @@ } }, { - "accuracy": 0.9977613681633222, + "accuracy": 0.9977757377844108, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -29035,7 +29035,7 @@ } }, { - "accuracy": 0.9980850031501368, + "accuracy": 0.9980979344170344, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -29087,7 +29087,7 @@ } }, { - "accuracy": 0.9988080790560496, + "accuracy": 0.9988160860773764, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -29130,7 +29130,7 @@ } }, { - "accuracy": 0.9988615944197303, + "accuracy": 0.998868968925978, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -29179,7 +29179,7 @@ } }, { - "accuracy": 0.9991243849263379, + "accuracy": 0.9991298783570528, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -29225,7 +29225,7 @@ } }, { - "accuracy": 0.9996825820815406, + "accuracy": 0.9996844474226236, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -29267,7 +29267,7 @@ ], "model.layers.15.self_attn": [ { - "accuracy": 0.9824525713920593, + "accuracy": 0.9824147083257374, "total_bits": 320757760, "q_proj": { "group_size": { @@ -29331,7 +29331,7 @@ } }, { - "accuracy": 0.984559493629556, + "accuracy": 0.9846857848920321, "total_bits": 329080832, "q_proj": { "group_size": { @@ -29395,7 +29395,7 @@ } }, { - "accuracy": 0.9868912743894678, + "accuracy": 0.9870848036126086, "total_bits": 336024576, "q_proj": { "group_size": { @@ -29459,7 +29459,7 @@ } }, { - "accuracy": 0.989514924977955, + "accuracy": 0.9897027548990751, "total_bits": 401557504, "q_proj": { "group_size": { @@ -29523,7 +29523,7 @@ } }, { - "accuracy": 0.9904752656033164, + "accuracy": 0.9906354436748906, "total_bits": 475279360, "q_proj": { "group_size": { @@ -29587,7 +29587,7 @@ } }, { - "accuracy": 0.9908929200548875, + "accuracy": 0.9907586456913697, "total_bits": 475479040, "q_proj": { "group_size": { @@ -29651,7 +29651,7 @@ } }, { - "accuracy": 0.9935646633568563, + "accuracy": 0.9935928355706366, "total_bits": 609759232, "q_proj": { "group_size": { @@ -29703,7 +29703,7 @@ } }, { - "accuracy": 0.9938069523165101, + "accuracy": 0.9939393428595442, "total_bits": 610024448, "q_proj": { "group_size": { @@ -29755,7 +29755,7 @@ } }, { - "accuracy": 0.9942700686423402, + "accuracy": 0.9941330226628404, "total_bits": 615020544, "q_proj": { "group_size": { @@ -29807,7 +29807,7 @@ } }, { - "accuracy": 0.9943623350639093, + "accuracy": 0.9942420234805659, "total_bits": 623951872, "q_proj": { "group_size": { @@ -29859,7 +29859,7 @@ } }, { - "accuracy": 0.9947521996341253, + "accuracy": 0.9948941445664355, "total_bits": 626473984, "q_proj": { "group_size": { @@ -29923,7 +29923,7 @@ } }, { - "accuracy": 0.9951281798513312, + "accuracy": 0.9951615255129965, "total_bits": 630355968, "q_proj": { "group_size": { @@ -29987,7 +29987,7 @@ } }, { - "accuracy": 0.9963866542828711, + "accuracy": 0.9964533252151389, "total_bits": 637362176, "q_proj": { "group_size": { @@ -30048,7 +30048,7 @@ } }, { - "accuracy": 0.9967845517553782, + "accuracy": 0.9966594400374513, "total_bits": 646823936, "q_proj": { "group_size": { @@ -30109,7 +30109,7 @@ } }, { - "accuracy": 0.9980922606038419, + "accuracy": 0.998086095248398, "total_bits": 784740352, "q_proj": { "group_size": { @@ -30170,7 +30170,7 @@ } }, { - "accuracy": 0.9983362271205375, + "accuracy": 0.9983459983609224, "total_bits": 797818880, "q_proj": { "group_size": { @@ -30231,7 +30231,7 @@ } }, { - "accuracy": 0.9984019279087845, + "accuracy": 0.9983898130686659, "total_bits": 911749120, "q_proj": { "group_size": { @@ -30283,7 +30283,7 @@ } }, { - "accuracy": 0.9993230001510758, + "accuracy": 0.9993294880872494, "total_bits": 942718976, "q_proj": { "group_size": { @@ -30335,7 +30335,7 @@ } }, { - "accuracy": 0.9995743254395691, + "accuracy": 0.9995760127509895, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -30389,7 +30389,7 @@ ], "model.layers.15.mlp": [ { - "accuracy": 0.9793524036281988, + "accuracy": 0.9794865736835882, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -30441,7 +30441,7 @@ } }, { - "accuracy": 0.9798343667858526, + "accuracy": 0.979961926999845, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -30493,7 +30493,7 @@ } }, { - "accuracy": 0.9834856532121959, + "accuracy": 0.983589093936117, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -30542,7 +30542,7 @@ } }, { - "accuracy": 0.9845627044376574, + "accuracy": 0.9846592827847129, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -30591,7 +30591,7 @@ } }, { - "accuracy": 0.9898150900476858, + "accuracy": 0.9898805383004641, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -30643,7 +30643,7 @@ } }, { - "accuracy": 0.9906108646016372, + "accuracy": 0.9906724707076424, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -30695,7 +30695,7 @@ } }, { - "accuracy": 0.9921695605704659, + "accuracy": 0.9922199774729578, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -30744,7 +30744,7 @@ } }, { - "accuracy": 0.9947993535744516, + "accuracy": 0.9948324247410423, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -30787,7 +30787,7 @@ } }, { - "accuracy": 0.9952749190361876, + "accuracy": 0.9953051005539141, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -30830,7 +30830,7 @@ } }, { - "accuracy": 0.9948449032871347, + "accuracy": 0.9948773748780552, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -30882,7 +30882,7 @@ } }, { - "accuracy": 0.9954874762579015, + "accuracy": 0.9955156284727549, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -30934,7 +30934,7 @@ } }, { - "accuracy": 0.9973817075553694, + "accuracy": 0.9973978559045416, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -30986,7 +30986,7 @@ } }, { - "accuracy": 0.9977552300613177, + "accuracy": 0.9977698098672064, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -31038,7 +31038,7 @@ } }, { - "accuracy": 0.9986074975642719, + "accuracy": 0.9986164662986994, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -31081,7 +31081,7 @@ } }, { - "accuracy": 0.9986677907015148, + "accuracy": 0.9986755146007789, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -31130,7 +31130,7 @@ } }, { - "accuracy": 0.9989705651410317, + "accuracy": 0.9989761815063263, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -31176,7 +31176,7 @@ } }, { - "accuracy": 0.9996302668495398, + "accuracy": 0.9996326718675462, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -31218,7 +31218,7 @@ ], "model.layers.16.self_attn": [ { - "accuracy": 0.9801881187840512, + "accuracy": 0.9803797900676727, "total_bits": 320757760, "q_proj": { "group_size": { @@ -31282,7 +31282,7 @@ } }, { - "accuracy": 0.9831292441016749, + "accuracy": 0.9833359592839291, "total_bits": 329080832, "q_proj": { "group_size": { @@ -31346,7 +31346,7 @@ } }, { - "accuracy": 0.9848053314183888, + "accuracy": 0.9847795649578697, "total_bits": 336024576, "q_proj": { "group_size": { @@ -31410,7 +31410,7 @@ } }, { - "accuracy": 0.9881028180059633, + "accuracy": 0.9879834495092693, "total_bits": 401557504, "q_proj": { "group_size": { @@ -31474,7 +31474,7 @@ } }, { - "accuracy": 0.989562996908238, + "accuracy": 0.9894756047349227, "total_bits": 475279360, "q_proj": { "group_size": { @@ -31538,7 +31538,7 @@ } }, { - "accuracy": 0.9895634533543336, + "accuracy": 0.9897570696316267, "total_bits": 475479040, "q_proj": { "group_size": { @@ -31602,7 +31602,7 @@ } }, { - "accuracy": 0.9935869354950754, + "accuracy": 0.9935936453311067, "total_bits": 609759232, "q_proj": { "group_size": { @@ -31654,7 +31654,7 @@ } }, { - "accuracy": 0.9938013090899116, + "accuracy": 0.9938549085667259, "total_bits": 610024448, "q_proj": { "group_size": { @@ -31706,7 +31706,7 @@ } }, { - "accuracy": 0.9943626679872212, + "accuracy": 0.9943937235756924, "total_bits": 615020544, "q_proj": { "group_size": { @@ -31758,7 +31758,7 @@ } }, { - "accuracy": 0.994471928790996, + "accuracy": 0.9944947040394733, "total_bits": 623951872, "q_proj": { "group_size": { @@ -31810,7 +31810,7 @@ } }, { - "accuracy": 0.9944174948491549, + "accuracy": 0.994548030981892, "total_bits": 626473984, "q_proj": { "group_size": { @@ -31874,7 +31874,7 @@ } }, { - "accuracy": 0.994869528632415, + "accuracy": 0.9948121877877336, "total_bits": 630355968, "q_proj": { "group_size": { @@ -31938,7 +31938,7 @@ } }, { - "accuracy": 0.9960502009642752, + "accuracy": 0.9960859912006479, "total_bits": 637362176, "q_proj": { "group_size": { @@ -31999,7 +31999,7 @@ } }, { - "accuracy": 0.9964677503234461, + "accuracy": 0.9964391616614241, "total_bits": 646823936, "q_proj": { "group_size": { @@ -32060,7 +32060,7 @@ } }, { - "accuracy": 0.9979117251932621, + "accuracy": 0.9979385008153162, "total_bits": 784740352, "q_proj": { "group_size": { @@ -32121,7 +32121,7 @@ } }, { - "accuracy": 0.9982501635033834, + "accuracy": 0.9982566111966183, "total_bits": 797818880, "q_proj": { "group_size": { @@ -32182,7 +32182,7 @@ } }, { - "accuracy": 0.9983268962486794, + "accuracy": 0.998372650558227, "total_bits": 911749120, "q_proj": { "group_size": { @@ -32234,7 +32234,7 @@ } }, { - "accuracy": 0.9992394734075979, + "accuracy": 0.9992396319285035, "total_bits": 942718976, "q_proj": { "group_size": { @@ -32286,7 +32286,7 @@ } }, { - "accuracy": 0.9995766217869363, + "accuracy": 0.9995765613980199, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -32340,7 +32340,7 @@ ], "model.layers.16.mlp": [ { - "accuracy": 0.9753150343894958, + "accuracy": 0.9754637322927776, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -32392,7 +32392,7 @@ } }, { - "accuracy": 0.9758835481969934, + "accuracy": 0.9760256095936424, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -32444,7 +32444,7 @@ } }, { - "accuracy": 0.9802227443770358, + "accuracy": 0.9803409497988852, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -32493,7 +32493,7 @@ } }, { - "accuracy": 0.9815049892977664, + "accuracy": 0.9816161315692099, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -32542,7 +32542,7 @@ } }, { - "accuracy": 0.9877431973030693, + "accuracy": 0.9878182544520027, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -32594,7 +32594,7 @@ } }, { - "accuracy": 0.9887041418175948, + "accuracy": 0.9887723146300567, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -32646,7 +32646,7 @@ } }, { - "accuracy": 0.9905570070994528, + "accuracy": 0.9906137107234252, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -32695,7 +32695,7 @@ } }, { - "accuracy": 0.9937269711180737, + "accuracy": 0.9937631644700703, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -32738,7 +32738,7 @@ } }, { - "accuracy": 0.9943001960453234, + "accuracy": 0.9943349835904021, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -32781,7 +32781,7 @@ } }, { - "accuracy": 0.9937864067523103, + "accuracy": 0.9938235353482398, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -32833,7 +32833,7 @@ } }, { - "accuracy": 0.9945560780010725, + "accuracy": 0.9945905937960273, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -32885,7 +32885,7 @@ } }, { - "accuracy": 0.9968393311688775, + "accuracy": 0.9968576070509458, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -32937,7 +32937,7 @@ } }, { - "accuracy": 0.9972904524520824, + "accuracy": 0.9973071060682598, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -32989,7 +32989,7 @@ } }, { - "accuracy": 0.9983178400679639, + "accuracy": 0.9983271046688682, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -33032,7 +33032,7 @@ } }, { - "accuracy": 0.9983907385092032, + "accuracy": 0.998400464163799, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -33081,7 +33081,7 @@ } }, { - "accuracy": 0.9987593983349047, + "accuracy": 0.9987663593339292, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -33127,7 +33127,7 @@ } }, { - "accuracy": 0.9995538266375661, + "accuracy": 0.9995563989995342, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -33169,7 +33169,7 @@ ], "model.layers.17.self_attn": [ { - "accuracy": 0.9736203457179823, + "accuracy": 0.9741273961569133, "total_bits": 320757760, "q_proj": { "group_size": { @@ -33233,7 +33233,7 @@ } }, { - "accuracy": 0.9779170453548431, + "accuracy": 0.9781463695199866, "total_bits": 329080832, "q_proj": { "group_size": { @@ -33297,7 +33297,7 @@ } }, { - "accuracy": 0.9804213470534274, + "accuracy": 0.9806286896529951, "total_bits": 336024576, "q_proj": { "group_size": { @@ -33361,7 +33361,7 @@ } }, { - "accuracy": 0.9863786571904233, + "accuracy": 0.9865448114119078, "total_bits": 401557504, "q_proj": { "group_size": { @@ -33425,7 +33425,7 @@ } }, { - "accuracy": 0.9882540059717078, + "accuracy": 0.9883713432048497, "total_bits": 475279360, "q_proj": { "group_size": { @@ -33489,7 +33489,7 @@ } }, { - "accuracy": 0.9884157572921953, + "accuracy": 0.988589703252441, "total_bits": 475479040, "q_proj": { "group_size": { @@ -33553,7 +33553,7 @@ } }, { - "accuracy": 0.9920703329538044, + "accuracy": 0.9921399135338632, "total_bits": 609759232, "q_proj": { "group_size": { @@ -33605,7 +33605,7 @@ } }, { - "accuracy": 0.9922442561701724, + "accuracy": 0.9924012163752004, "total_bits": 610024448, "q_proj": { "group_size": { @@ -33657,7 +33657,7 @@ } }, { - "accuracy": 0.9928446498356367, + "accuracy": 0.9928476120296278, "total_bits": 615020544, "q_proj": { "group_size": { @@ -33709,7 +33709,7 @@ } }, { - "accuracy": 0.9931093791597768, + "accuracy": 0.9931200852519587, "total_bits": 623951872, "q_proj": { "group_size": { @@ -33761,7 +33761,7 @@ } }, { - "accuracy": 0.9941317862585971, + "accuracy": 0.9941871413274815, "total_bits": 626473984, "q_proj": { "group_size": { @@ -33825,7 +33825,7 @@ } }, { - "accuracy": 0.9945700192137769, + "accuracy": 0.9945951468850437, "total_bits": 630355968, "q_proj": { "group_size": { @@ -33889,7 +33889,7 @@ } }, { - "accuracy": 0.9947705464927774, + "accuracy": 0.9948291731508154, "total_bits": 637362176, "q_proj": { "group_size": { @@ -33950,7 +33950,7 @@ } }, { - "accuracy": 0.9952156614316138, + "accuracy": 0.9952291971758792, "total_bits": 646823936, "q_proj": { "group_size": { @@ -34011,7 +34011,7 @@ } }, { - "accuracy": 0.9971760415325039, + "accuracy": 0.9971973199985529, "total_bits": 784740352, "q_proj": { "group_size": { @@ -34072,7 +34072,7 @@ } }, { - "accuracy": 0.9976018477035197, + "accuracy": 0.9976028523554927, "total_bits": 797818880, "q_proj": { "group_size": { @@ -34133,7 +34133,7 @@ } }, { - "accuracy": 0.9978938892875847, + "accuracy": 0.9979093982593009, "total_bits": 911749120, "q_proj": { "group_size": { @@ -34185,7 +34185,7 @@ } }, { - "accuracy": 0.9988842452631185, + "accuracy": 0.9988939397429165, "total_bits": 942718976, "q_proj": { "group_size": { @@ -34237,7 +34237,7 @@ } }, { - "accuracy": 0.9994544779979869, + "accuracy": 0.9994630680272454, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -34291,7 +34291,7 @@ ], "model.layers.17.mlp": [ { - "accuracy": 0.9712188526203758, + "accuracy": 0.9713734702060097, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -34343,7 +34343,7 @@ } }, { - "accuracy": 0.9718795167772394, + "accuracy": 0.9720371836110165, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -34395,7 +34395,7 @@ } }, { - "accuracy": 0.976847557645095, + "accuracy": 0.9769792478335532, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -34444,7 +34444,7 @@ } }, { - "accuracy": 0.9783040834100623, + "accuracy": 0.9784279904867473, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -34493,7 +34493,7 @@ } }, { - "accuracy": 0.9857219818391298, + "accuracy": 0.9858058032236601, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -34545,7 +34545,7 @@ } }, { - "accuracy": 0.9868270613645252, + "accuracy": 0.9869018644094467, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -34597,7 +34597,7 @@ } }, { - "accuracy": 0.9889355325385144, + "accuracy": 0.9889984217129255, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -34646,7 +34646,7 @@ } }, { - "accuracy": 0.9927025994187907, + "accuracy": 0.9927468495933633, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -34689,7 +34689,7 @@ } }, { - "accuracy": 0.9933551753822126, + "accuracy": 0.9933920463449076, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -34732,7 +34732,7 @@ } }, { - "accuracy": 0.9927656799554825, + "accuracy": 0.9928078729855386, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -34784,7 +34784,7 @@ } }, { - "accuracy": 0.993651021468012, + "accuracy": 0.9936860646856459, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -34836,7 +34836,7 @@ } }, { - "accuracy": 0.9963220859828749, + "accuracy": 0.9963443620424521, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -34888,7 +34888,7 @@ } }, { - "accuracy": 0.9968407738365626, + "accuracy": 0.9968583442662892, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -34940,7 +34940,7 @@ } }, { - "accuracy": 0.9980455387971903, + "accuracy": 0.9980572511098886, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -34983,7 +34983,7 @@ } }, { - "accuracy": 0.9981275679249513, + "accuracy": 0.9981387641869093, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -35032,7 +35032,7 @@ } }, { - "accuracy": 0.9985435538386044, + "accuracy": 0.9985528589275322, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -35078,7 +35078,7 @@ } }, { - "accuracy": 0.9994840158737803, + "accuracy": 0.9994872551117289, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -35120,7 +35120,7 @@ ], "model.layers.18.self_attn": [ { - "accuracy": 0.9716378168055886, + "accuracy": 0.9716777550546747, "total_bits": 320757760, "q_proj": { "group_size": { @@ -35184,7 +35184,7 @@ } }, { - "accuracy": 0.9757674320747978, + "accuracy": 0.9758694532670473, "total_bits": 329080832, "q_proj": { "group_size": { @@ -35248,7 +35248,7 @@ } }, { - "accuracy": 0.9791416682695088, + "accuracy": 0.9792861907105697, "total_bits": 336024576, "q_proj": { "group_size": { @@ -35312,7 +35312,7 @@ } }, { - "accuracy": 0.9856245925551966, + "accuracy": 0.9856744471349215, "total_bits": 401557504, "q_proj": { "group_size": { @@ -35376,7 +35376,7 @@ } }, { - "accuracy": 0.9875024478686484, + "accuracy": 0.9874892673994365, "total_bits": 475279360, "q_proj": { "group_size": { @@ -35440,7 +35440,7 @@ } }, { - "accuracy": 0.9877277283292067, + "accuracy": 0.9877497318543886, "total_bits": 475479040, "q_proj": { "group_size": { @@ -35504,7 +35504,7 @@ } }, { - "accuracy": 0.9919087620157945, + "accuracy": 0.9918321339707625, "total_bits": 609759232, "q_proj": { "group_size": { @@ -35556,7 +35556,7 @@ } }, { - "accuracy": 0.9921502714094362, + "accuracy": 0.9921730952827554, "total_bits": 610024448, "q_proj": { "group_size": { @@ -35608,7 +35608,7 @@ } }, { - "accuracy": 0.9926899404902207, + "accuracy": 0.9928006412167298, "total_bits": 615020544, "q_proj": { "group_size": { @@ -35660,7 +35660,7 @@ } }, { - "accuracy": 0.9930640913938221, + "accuracy": 0.9930984001410635, "total_bits": 623951872, "q_proj": { "group_size": { @@ -35712,7 +35712,7 @@ } }, { - "accuracy": 0.9937935926412281, + "accuracy": 0.9938601800485661, "total_bits": 626473984, "q_proj": { "group_size": { @@ -35776,7 +35776,7 @@ } }, { - "accuracy": 0.9942181420169378, + "accuracy": 0.9942674323132163, "total_bits": 630355968, "q_proj": { "group_size": { @@ -35840,7 +35840,7 @@ } }, { - "accuracy": 0.9945547725809248, + "accuracy": 0.9946191334410718, "total_bits": 637362176, "q_proj": { "group_size": { @@ -35901,7 +35901,7 @@ } }, { - "accuracy": 0.9950165878001013, + "accuracy": 0.9950485351054292, "total_bits": 646823936, "q_proj": { "group_size": { @@ -35962,7 +35962,7 @@ } }, { - "accuracy": 0.9970570671323099, + "accuracy": 0.9970986162753481, "total_bits": 784740352, "q_proj": { "group_size": { @@ -36023,7 +36023,7 @@ } }, { - "accuracy": 0.9974930686386008, + "accuracy": 0.9975272664114049, "total_bits": 797818880, "q_proj": { "group_size": { @@ -36084,7 +36084,7 @@ } }, { - "accuracy": 0.9979022382513473, + "accuracy": 0.9979424694258916, "total_bits": 911749120, "q_proj": { "group_size": { @@ -36136,7 +36136,7 @@ } }, { - "accuracy": 0.998814924276973, + "accuracy": 0.9988226778805256, "total_bits": 942718976, "q_proj": { "group_size": { @@ -36188,7 +36188,7 @@ } }, { - "accuracy": 0.9994503862567639, + "accuracy": 0.9994619520086991, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -36242,7 +36242,7 @@ ], "model.layers.18.mlp": [ { - "accuracy": 0.964525370221389, + "accuracy": 0.9646681766760976, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -36294,7 +36294,7 @@ } }, { - "accuracy": 0.9653103822156003, + "accuracy": 0.9654584307419627, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -36346,7 +36346,7 @@ } }, { - "accuracy": 0.9700868286584553, + "accuracy": 0.9701737196821916, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -36395,7 +36395,7 @@ } }, { - "accuracy": 0.9714746443848861, + "accuracy": 0.9715520174879777, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -36444,7 +36444,7 @@ } }, { - "accuracy": 0.9828015393332431, + "accuracy": 0.9827913390962701, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -36496,7 +36496,7 @@ } }, { - "accuracy": 0.9849078467017726, + "accuracy": 0.9849941636386671, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -36548,7 +36548,7 @@ } }, { - "accuracy": 0.9871182661307486, + "accuracy": 0.9871893297684821, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -36597,7 +36597,7 @@ } }, { - "accuracy": 0.9916471917378274, + "accuracy": 0.9916955334575552, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -36640,7 +36640,7 @@ } }, { - "accuracy": 0.9922995449681031, + "accuracy": 0.9923447213674846, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -36683,7 +36683,7 @@ } }, { - "accuracy": 0.9917216638201162, + "accuracy": 0.9917682840635902, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -36735,7 +36735,7 @@ } }, { - "accuracy": 0.9926933387392446, + "accuracy": 0.9927375693070261, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -36787,7 +36787,7 @@ } }, { - "accuracy": 0.9957802883888546, + "accuracy": 0.9958057956475961, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -36839,7 +36839,7 @@ } }, { - "accuracy": 0.9963691756129265, + "accuracy": 0.9963909658162218, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -36891,7 +36891,7 @@ } }, { - "accuracy": 0.9977286361942166, + "accuracy": 0.997742653677338, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -36934,7 +36934,7 @@ } }, { - "accuracy": 0.9978502316302374, + "accuracy": 0.9978637622767373, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -36983,7 +36983,7 @@ } }, { - "accuracy": 0.9982870642684007, + "accuracy": 0.9982975642932089, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -37029,7 +37029,7 @@ } }, { - "accuracy": 0.9993793972718873, + "accuracy": 0.999383513766684, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -37071,7 +37071,7 @@ ], "model.layers.19.self_attn": [ { - "accuracy": 0.9755127869154278, + "accuracy": 0.9755469560623169, "total_bits": 320757760, "q_proj": { "group_size": { @@ -37135,7 +37135,7 @@ } }, { - "accuracy": 0.9775320652284121, + "accuracy": 0.9778150994526712, "total_bits": 329080832, "q_proj": { "group_size": { @@ -37199,7 +37199,7 @@ } }, { - "accuracy": 0.9803245726384615, + "accuracy": 0.9806819291491258, "total_bits": 336024576, "q_proj": { "group_size": { @@ -37263,7 +37263,7 @@ } }, { - "accuracy": 0.9852223882549688, + "accuracy": 0.985606660968379, "total_bits": 401557504, "q_proj": { "group_size": { @@ -37327,7 +37327,7 @@ } }, { - "accuracy": 0.9869777419065174, + "accuracy": 0.9872808558376212, "total_bits": 475279360, "q_proj": { "group_size": { @@ -37391,7 +37391,7 @@ } }, { - "accuracy": 0.9875111266186363, + "accuracy": 0.9876446033778944, "total_bits": 475479040, "q_proj": { "group_size": { @@ -37455,7 +37455,7 @@ } }, { - "accuracy": 0.9917858603753542, + "accuracy": 0.9919387658959941, "total_bits": 609759232, "q_proj": { "group_size": { @@ -37507,7 +37507,7 @@ } }, { - "accuracy": 0.9922473501217993, + "accuracy": 0.9922318231118353, "total_bits": 610024448, "q_proj": { "group_size": { @@ -37559,7 +37559,7 @@ } }, { - "accuracy": 0.992727496122059, + "accuracy": 0.9927876097591299, "total_bits": 615020544, "q_proj": { "group_size": { @@ -37611,7 +37611,7 @@ } }, { - "accuracy": 0.9929983200211274, + "accuracy": 0.9930746343575025, "total_bits": 623951872, "q_proj": { "group_size": { @@ -37663,7 +37663,7 @@ } }, { - "accuracy": 0.9934035806279433, + "accuracy": 0.9934205091313312, "total_bits": 626473984, "q_proj": { "group_size": { @@ -37727,7 +37727,7 @@ } }, { - "accuracy": 0.9937662595981046, + "accuracy": 0.9939091221282357, "total_bits": 630355968, "q_proj": { "group_size": { @@ -37791,7 +37791,7 @@ } }, { - "accuracy": 0.9949815061531568, + "accuracy": 0.9950423401437307, "total_bits": 637362176, "q_proj": { "group_size": { @@ -37852,7 +37852,7 @@ } }, { - "accuracy": 0.9953521580288285, + "accuracy": 0.9954208363043634, "total_bits": 646823936, "q_proj": { "group_size": { @@ -37913,7 +37913,7 @@ } }, { - "accuracy": 0.9972557074537403, + "accuracy": 0.9972811939292833, "total_bits": 784740352, "q_proj": { "group_size": { @@ -37974,7 +37974,7 @@ } }, { - "accuracy": 0.9976938137490499, + "accuracy": 0.9977000167495326, "total_bits": 797818880, "q_proj": { "group_size": { @@ -38035,7 +38035,7 @@ } }, { - "accuracy": 0.9978761865120185, + "accuracy": 0.9978962226917869, "total_bits": 911749120, "q_proj": { "group_size": { @@ -38087,7 +38087,7 @@ } }, { - "accuracy": 0.998943293663232, + "accuracy": 0.9989494604891852, "total_bits": 942718976, "q_proj": { "group_size": { @@ -38139,7 +38139,7 @@ } }, { - "accuracy": 0.999454867535908, + "accuracy": 0.9994576915020221, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -38193,7 +38193,7 @@ ], "model.layers.19.mlp": [ { - "accuracy": 0.9638730632631403, + "accuracy": 0.9640721992442483, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -38245,7 +38245,7 @@ } }, { - "accuracy": 0.9648721908268175, + "accuracy": 0.9650518894195557, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -38297,7 +38297,7 @@ } }, { - "accuracy": 0.9705943998537565, + "accuracy": 0.9707440922134801, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -38346,7 +38346,7 @@ } }, { - "accuracy": 0.9721735213932238, + "accuracy": 0.9723140785568639, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -38395,7 +38395,7 @@ } }, { - "accuracy": 0.9822877488638225, + "accuracy": 0.9823814677564722, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -38447,7 +38447,7 @@ } }, { - "accuracy": 0.9836467928008029, + "accuracy": 0.9837378859519958, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -38499,7 +38499,7 @@ } }, { - "accuracy": 0.9859193641888467, + "accuracy": 0.985998627386595, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -38548,7 +38548,7 @@ } }, { - "accuracy": 0.9909956368960833, + "accuracy": 0.9910474301953065, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -38591,7 +38591,7 @@ } }, { - "accuracy": 0.9917139889378297, + "accuracy": 0.9917593833647276, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -38634,7 +38634,7 @@ } }, { - "accuracy": 0.9910229483717367, + "accuracy": 0.991074377768918, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -38686,7 +38686,7 @@ } }, { - "accuracy": 0.9921089057859621, + "accuracy": 0.9921528171551856, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -38738,7 +38738,7 @@ } }, { - "accuracy": 0.9954366452599827, + "accuracy": 0.9954626795492674, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -38790,7 +38790,7 @@ } }, { - "accuracy": 0.9960733899160436, + "accuracy": 0.9960954252042269, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -38842,7 +38842,7 @@ } }, { - "accuracy": 0.9975653763272261, + "accuracy": 0.9975794351806766, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -38885,7 +38885,7 @@ } }, { - "accuracy": 0.9976728007589516, + "accuracy": 0.997685789277679, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -38934,7 +38934,7 @@ } }, { - "accuracy": 0.9981181135303095, + "accuracy": 0.9981279590804326, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -38980,7 +38980,7 @@ } }, { - "accuracy": 0.9993598660160052, + "accuracy": 0.9993637749729188, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -39022,7 +39022,7 @@ ], "model.layers.20.self_attn": [ { - "accuracy": 0.9788577917375063, + "accuracy": 0.9792891238865099, "total_bits": 320757760, "q_proj": { "group_size": { @@ -39086,7 +39086,7 @@ } }, { - "accuracy": 0.9795936502908406, + "accuracy": 0.9799838505293194, "total_bits": 329080832, "q_proj": { "group_size": { @@ -39150,7 +39150,7 @@ } }, { - "accuracy": 0.9834668040275574, + "accuracy": 0.9836689672972027, "total_bits": 336024576, "q_proj": { "group_size": { @@ -39214,7 +39214,7 @@ } }, { - "accuracy": 0.9872075068323236, + "accuracy": 0.987395955543769, "total_bits": 401557504, "q_proj": { "group_size": { @@ -39278,7 +39278,7 @@ } }, { - "accuracy": 0.988563785427495, + "accuracy": 0.9887089768522664, "total_bits": 475279360, "q_proj": { "group_size": { @@ -39342,7 +39342,7 @@ } }, { - "accuracy": 0.9891500347538998, + "accuracy": 0.9891776740550995, "total_bits": 475479040, "q_proj": { "group_size": { @@ -39406,7 +39406,7 @@ } }, { - "accuracy": 0.9910190607372084, + "accuracy": 0.9910986768571954, "total_bits": 609759232, "q_proj": { "group_size": { @@ -39458,7 +39458,7 @@ } }, { - "accuracy": 0.9916288029206427, + "accuracy": 0.9916353131595411, "total_bits": 610024448, "q_proj": { "group_size": { @@ -39510,7 +39510,7 @@ } }, { - "accuracy": 0.9919651040905401, + "accuracy": 0.9920247079510438, "total_bits": 615020544, "q_proj": { "group_size": { @@ -39562,7 +39562,7 @@ } }, { - "accuracy": 0.9920761240156073, + "accuracy": 0.992149271463093, "total_bits": 623951872, "q_proj": { "group_size": { @@ -39614,7 +39614,7 @@ } }, { - "accuracy": 0.9941177501490241, + "accuracy": 0.9942162248649096, "total_bits": 626473984, "q_proj": { "group_size": { @@ -39678,7 +39678,7 @@ } }, { - "accuracy": 0.9944217036429205, + "accuracy": 0.9945300292027625, "total_bits": 630355968, "q_proj": { "group_size": { @@ -39742,7 +39742,7 @@ } }, { - "accuracy": 0.9946274859340567, + "accuracy": 0.9946998146019483, "total_bits": 637362176, "q_proj": { "group_size": { @@ -39803,7 +39803,7 @@ } }, { - "accuracy": 0.9949189318638099, + "accuracy": 0.9950166987745386, "total_bits": 646823936, "q_proj": { "group_size": { @@ -39864,7 +39864,7 @@ } }, { - "accuracy": 0.9970433596325549, + "accuracy": 0.9970882785947699, "total_bits": 784740352, "q_proj": { "group_size": { @@ -39925,7 +39925,7 @@ } }, { - "accuracy": 0.9974729487378347, + "accuracy": 0.9975208300901087, "total_bits": 797818880, "q_proj": { "group_size": { @@ -39986,7 +39986,7 @@ } }, { - "accuracy": 0.997406294667407, + "accuracy": 0.9974515375338102, "total_bits": 911749120, "q_proj": { "group_size": { @@ -40038,7 +40038,7 @@ } }, { - "accuracy": 0.998958008363843, + "accuracy": 0.9989738071239308, "total_bits": 942718976, "q_proj": { "group_size": { @@ -40090,7 +40090,7 @@ } }, { - "accuracy": 0.9993222945516831, + "accuracy": 0.9993280315967766, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -40144,7 +40144,7 @@ ], "model.layers.20.mlp": [ { - "accuracy": 0.9636889602008619, + "accuracy": 0.9638865778320714, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -40196,7 +40196,7 @@ } }, { - "accuracy": 0.9646058960964805, + "accuracy": 0.9647944851925498, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -40248,7 +40248,7 @@ } }, { - "accuracy": 0.9702938700977125, + "accuracy": 0.9704568135110956, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -40297,7 +40297,7 @@ } }, { - "accuracy": 0.971888360224272, + "accuracy": 0.9720407222446642, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -40346,7 +40346,7 @@ } }, { - "accuracy": 0.9820300513192227, + "accuracy": 0.9821287346513647, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -40398,7 +40398,7 @@ } }, { - "accuracy": 0.983422117797952, + "accuracy": 0.9835108126464643, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -40450,7 +40450,7 @@ } }, { - "accuracy": 0.9857326542076311, + "accuracy": 0.985809525376872, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -40499,7 +40499,7 @@ } }, { - "accuracy": 0.990852628099291, + "accuracy": 0.9909011666712008, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -40542,7 +40542,7 @@ } }, { - "accuracy": 0.9916105168430429, + "accuracy": 0.9916578358725497, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -40585,7 +40585,7 @@ } }, { - "accuracy": 0.9908938266729054, + "accuracy": 0.9909407614093078, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -40637,7 +40637,7 @@ } }, { - "accuracy": 0.9919972835402739, + "accuracy": 0.9920404608312406, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -40689,7 +40689,7 @@ } }, { - "accuracy": 0.9953723620427283, + "accuracy": 0.995397367367619, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -40741,7 +40741,7 @@ } }, { - "accuracy": 0.9960178505433234, + "accuracy": 0.9960397713278469, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -40793,7 +40793,7 @@ } }, { - "accuracy": 0.9975383462090242, + "accuracy": 0.9975519041089635, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -40836,7 +40836,7 @@ } }, { - "accuracy": 0.9976425531663393, + "accuracy": 0.9976551860178772, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -40885,7 +40885,7 @@ } }, { - "accuracy": 0.9980946628278807, + "accuracy": 0.9981048785542187, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -40931,7 +40931,7 @@ } }, { - "accuracy": 0.9993525329781207, + "accuracy": 0.9993559678428268, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -40973,7 +40973,7 @@ ], "model.layers.21.self_attn": [ { - "accuracy": 0.9733156248142845, + "accuracy": 0.9733950119269522, "total_bits": 320757760, "q_proj": { "group_size": { @@ -41037,7 +41037,7 @@ } }, { - "accuracy": 0.9755077409116846, + "accuracy": 0.9755143287934755, "total_bits": 329080832, "q_proj": { "group_size": { @@ -41101,7 +41101,7 @@ } }, { - "accuracy": 0.9781834702742728, + "accuracy": 0.9783172042746293, "total_bits": 336024576, "q_proj": { "group_size": { @@ -41165,7 +41165,7 @@ } }, { - "accuracy": 0.9830043896248466, + "accuracy": 0.9832822790271357, "total_bits": 401557504, "q_proj": { "group_size": { @@ -41229,7 +41229,7 @@ } }, { - "accuracy": 0.9857717517175173, + "accuracy": 0.986018485144565, "total_bits": 475279360, "q_proj": { "group_size": { @@ -41293,7 +41293,7 @@ } }, { - "accuracy": 0.9860107914397591, + "accuracy": 0.9863700898070085, "total_bits": 475479040, "q_proj": { "group_size": { @@ -41357,7 +41357,7 @@ } }, { - "accuracy": 0.9910837160913568, + "accuracy": 0.9914763554146415, "total_bits": 609759232, "q_proj": { "group_size": { @@ -41409,7 +41409,7 @@ } }, { - "accuracy": 0.9915915664873625, + "accuracy": 0.9918616735621503, "total_bits": 610024448, "q_proj": { "group_size": { @@ -41461,7 +41461,7 @@ } }, { - "accuracy": 0.9922488763144142, + "accuracy": 0.9923772458967409, "total_bits": 615020544, "q_proj": { "group_size": { @@ -41513,7 +41513,7 @@ } }, { - "accuracy": 0.9925035991166767, + "accuracy": 0.9925474744094046, "total_bits": 623951872, "q_proj": { "group_size": { @@ -41565,7 +41565,7 @@ } }, { - "accuracy": 0.992689197784976, + "accuracy": 0.9928911743979705, "total_bits": 626473984, "q_proj": { "group_size": { @@ -41629,7 +41629,7 @@ } }, { - "accuracy": 0.9931197864444632, + "accuracy": 0.9933485882846933, "total_bits": 630355968, "q_proj": { "group_size": { @@ -41693,7 +41693,7 @@ } }, { - "accuracy": 0.9944828611455465, + "accuracy": 0.9946199420251345, "total_bits": 637362176, "q_proj": { "group_size": { @@ -41754,7 +41754,7 @@ } }, { - "accuracy": 0.9950494476054844, + "accuracy": 0.9950503804966023, "total_bits": 646823936, "q_proj": { "group_size": { @@ -41815,7 +41815,7 @@ } }, { - "accuracy": 0.9970526218806443, + "accuracy": 0.996987407732951, "total_bits": 784740352, "q_proj": { "group_size": { @@ -41876,7 +41876,7 @@ } }, { - "accuracy": 0.9974870142575941, + "accuracy": 0.9974926006244985, "total_bits": 797818880, "q_proj": { "group_size": { @@ -41937,7 +41937,7 @@ } }, { - "accuracy": 0.9977661392798549, + "accuracy": 0.9976836293935776, "total_bits": 911749120, "q_proj": { "group_size": { @@ -41989,7 +41989,7 @@ } }, { - "accuracy": 0.9988588283916837, + "accuracy": 0.9988621225286471, "total_bits": 942718976, "q_proj": { "group_size": { @@ -42041,7 +42041,7 @@ } }, { - "accuracy": 0.9994009548406068, + "accuracy": 0.9994244356884768, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -42095,7 +42095,7 @@ ], "model.layers.21.mlp": [ { - "accuracy": 0.9616459764932331, + "accuracy": 0.9618283039645145, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -42147,7 +42147,7 @@ } }, { - "accuracy": 0.9626245655511555, + "accuracy": 0.9627999977061623, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -42199,7 +42199,7 @@ } }, { - "accuracy": 0.9684790906153227, + "accuracy": 0.9686283249604075, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -42248,7 +42248,7 @@ } }, { - "accuracy": 0.9700951011557328, + "accuracy": 0.9702389365748355, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -42297,7 +42297,7 @@ } }, { - "accuracy": 0.9809990180166144, + "accuracy": 0.9810954206868222, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -42349,7 +42349,7 @@ } }, { - "accuracy": 0.9824735130134382, + "accuracy": 0.9825591884161297, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -42401,7 +42401,7 @@ } }, { - "accuracy": 0.984824255893105, + "accuracy": 0.9848974033405906, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -42450,7 +42450,7 @@ } }, { - "accuracy": 0.9903360722880614, + "accuracy": 0.9903865982043115, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -42493,7 +42493,7 @@ } }, { - "accuracy": 0.9911203423612996, + "accuracy": 0.9911650238852752, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -42536,7 +42536,7 @@ } }, { - "accuracy": 0.9903604074528343, + "accuracy": 0.990408549967565, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -42588,7 +42588,7 @@ } }, { - "accuracy": 0.9915206989175395, + "accuracy": 0.9915644197087539, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -42640,7 +42640,7 @@ } }, { - "accuracy": 0.9950946263576809, + "accuracy": 0.995118404689588, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -42692,7 +42692,7 @@ } }, { - "accuracy": 0.9957777524465009, + "accuracy": 0.9957985262337484, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -42744,7 +42744,7 @@ } }, { - "accuracy": 0.9973871492241558, + "accuracy": 0.9974003059691504, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -42787,7 +42787,7 @@ } }, { - "accuracy": 0.9974970823447955, + "accuracy": 0.9975086946628595, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -42836,7 +42836,7 @@ } }, { - "accuracy": 0.9979551020813616, + "accuracy": 0.9979641990442025, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -42882,7 +42882,7 @@ } }, { - "accuracy": 0.9993056652969435, + "accuracy": 0.999309469406542, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -42924,7 +42924,7 @@ ], "model.layers.22.self_attn": [ { - "accuracy": 0.9784525648543709, + "accuracy": 0.978420426971034, "total_bits": 320757760, "q_proj": { "group_size": { @@ -42988,7 +42988,7 @@ } }, { - "accuracy": 0.9791736492985174, + "accuracy": 0.9788977786114341, "total_bits": 329080832, "q_proj": { "group_size": { @@ -43052,7 +43052,7 @@ } }, { - "accuracy": 0.9824593976924294, + "accuracy": 0.982678948264373, "total_bits": 336024576, "q_proj": { "group_size": { @@ -43116,7 +43116,7 @@ } }, { - "accuracy": 0.98565589754205, + "accuracy": 0.9858789569453189, "total_bits": 401557504, "q_proj": { "group_size": { @@ -43180,7 +43180,7 @@ } }, { - "accuracy": 0.9870079661670484, + "accuracy": 0.9872076966260609, "total_bits": 475279360, "q_proj": { "group_size": { @@ -43244,7 +43244,7 @@ } }, { - "accuracy": 0.9876601743070703, + "accuracy": 0.9878985395557002, "total_bits": 475479040, "q_proj": { "group_size": { @@ -43308,7 +43308,7 @@ } }, { - "accuracy": 0.9889768377730721, + "accuracy": 0.9892485502519106, "total_bits": 609759232, "q_proj": { "group_size": { @@ -43360,7 +43360,7 @@ } }, { - "accuracy": 0.9897909972228502, + "accuracy": 0.9899120205327084, "total_bits": 610024448, "q_proj": { "group_size": { @@ -43412,7 +43412,7 @@ } }, { - "accuracy": 0.99030680797602, + "accuracy": 0.9904349258071498, "total_bits": 615020544, "q_proj": { "group_size": { @@ -43464,7 +43464,7 @@ } }, { - "accuracy": 0.990453953021451, + "accuracy": 0.9906028909118552, "total_bits": 623951872, "q_proj": { "group_size": { @@ -43516,7 +43516,7 @@ } }, { - "accuracy": 0.9934468022302577, + "accuracy": 0.993488110601902, "total_bits": 626473984, "q_proj": { "group_size": { @@ -43580,7 +43580,7 @@ } }, { - "accuracy": 0.9938278053151933, + "accuracy": 0.9938707136009869, "total_bits": 630355968, "q_proj": { "group_size": { @@ -43644,7 +43644,7 @@ } }, { - "accuracy": 0.9938991026658761, + "accuracy": 0.993876767001654, "total_bits": 637362176, "q_proj": { "group_size": { @@ -43705,7 +43705,7 @@ } }, { - "accuracy": 0.9942796042090968, + "accuracy": 0.9942743676273447, "total_bits": 646823936, "q_proj": { "group_size": { @@ -43766,7 +43766,7 @@ } }, { - "accuracy": 0.9966416429532202, + "accuracy": 0.9966516494750977, "total_bits": 784740352, "q_proj": { "group_size": { @@ -43827,7 +43827,7 @@ } }, { - "accuracy": 0.9971604241352332, + "accuracy": 0.9971919310720343, "total_bits": 797818880, "q_proj": { "group_size": { @@ -43888,7 +43888,7 @@ } }, { - "accuracy": 0.99698741439926, + "accuracy": 0.9969936426925031, "total_bits": 911749120, "q_proj": { "group_size": { @@ -43940,7 +43940,7 @@ } }, { - "accuracy": 0.9989495730321658, + "accuracy": 0.9989588332215422, "total_bits": 942718976, "q_proj": { "group_size": { @@ -43992,7 +43992,7 @@ } }, { - "accuracy": 0.999240324881516, + "accuracy": 0.9992361320672851, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -44046,7 +44046,7 @@ ], "model.layers.22.mlp": [ { - "accuracy": 0.960502919397856, + "accuracy": 0.9606696555488988, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -44098,7 +44098,7 @@ } }, { - "accuracy": 0.961547503345891, + "accuracy": 0.9617112247567428, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -44150,7 +44150,7 @@ } }, { - "accuracy": 0.9673889028398615, + "accuracy": 0.967525911958594, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -44199,7 +44199,7 @@ } }, { - "accuracy": 0.968980139807651, + "accuracy": 0.9691068498711837, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -44248,7 +44248,7 @@ } }, { - "accuracy": 0.9804544746875763, + "accuracy": 0.9805419696004767, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -44300,7 +44300,7 @@ } }, { - "accuracy": 0.9819841855450681, + "accuracy": 0.9820594568001596, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -44352,7 +44352,7 @@ } }, { - "accuracy": 0.9843009694626457, + "accuracy": 0.984363035151833, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -44401,7 +44401,7 @@ } }, { - "accuracy": 0.9900770265805093, + "accuracy": 0.9901229072558252, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -44444,7 +44444,7 @@ } }, { - "accuracy": 0.9908768275850698, + "accuracy": 0.9909159282320424, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -44487,7 +44487,7 @@ } }, { - "accuracy": 0.9900737091114646, + "accuracy": 0.9901211747997686, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -44539,7 +44539,7 @@ } }, { - "accuracy": 0.9912905771481363, + "accuracy": 0.9913274175242374, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -44591,7 +44591,7 @@ } }, { - "accuracy": 0.9949506011448408, + "accuracy": 0.9949769460056958, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -44643,7 +44643,7 @@ } }, { - "accuracy": 0.9956653184796634, + "accuracy": 0.9956840033593931, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -44695,7 +44695,7 @@ } }, { - "accuracy": 0.9973143143089194, + "accuracy": 0.9973280635711393, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -44738,7 +44738,7 @@ } }, { - "accuracy": 0.9974219865704838, + "accuracy": 0.9974356519156381, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -44787,7 +44787,7 @@ } }, { - "accuracy": 0.9978653474857933, + "accuracy": 0.9978769719600677, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -44833,7 +44833,7 @@ } }, { - "accuracy": 0.9992947799006575, + "accuracy": 0.9992983766684407, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -44875,7 +44875,7 @@ ], "model.layers.23.self_attn": [ { - "accuracy": 0.9780251101443642, + "accuracy": 0.9779581396203292, "total_bits": 320757760, "q_proj": { "group_size": { @@ -44939,7 +44939,7 @@ } }, { - "accuracy": 0.978527883165761, + "accuracy": 0.9784011197717566, "total_bits": 329080832, "q_proj": { "group_size": { @@ -45003,7 +45003,7 @@ } }, { - "accuracy": 0.9815480724761361, + "accuracy": 0.9817277051900563, "total_bits": 336024576, "q_proj": { "group_size": { @@ -45067,7 +45067,7 @@ } }, { - "accuracy": 0.985390482764495, + "accuracy": 0.98554737944352, "total_bits": 401557504, "q_proj": { "group_size": { @@ -45131,7 +45131,7 @@ } }, { - "accuracy": 0.9874509290644997, + "accuracy": 0.987572904480131, "total_bits": 475279360, "q_proj": { "group_size": { @@ -45195,7 +45195,7 @@ } }, { - "accuracy": 0.9877755163531554, + "accuracy": 0.9880059549683019, "total_bits": 475479040, "q_proj": { "group_size": { @@ -45259,7 +45259,7 @@ } }, { - "accuracy": 0.9902387182963522, + "accuracy": 0.990292848725068, "total_bits": 609759232, "q_proj": { "group_size": { @@ -45311,7 +45311,7 @@ } }, { - "accuracy": 0.9905369901343396, + "accuracy": 0.9908109942549154, "total_bits": 610024448, "q_proj": { "group_size": { @@ -45363,7 +45363,7 @@ } }, { - "accuracy": 0.9915478260893571, + "accuracy": 0.9913683392499623, "total_bits": 615020544, "q_proj": { "group_size": { @@ -45415,7 +45415,7 @@ } }, { - "accuracy": 0.9917051023558566, + "accuracy": 0.9915535002946854, "total_bits": 623951872, "q_proj": { "group_size": { @@ -45467,7 +45467,7 @@ } }, { - "accuracy": 0.9936908279594622, + "accuracy": 0.9937422044183079, "total_bits": 626473984, "q_proj": { "group_size": { @@ -45531,7 +45531,7 @@ } }, { - "accuracy": 0.994200447280156, + "accuracy": 0.9942037514165828, "total_bits": 630355968, "q_proj": { "group_size": { @@ -45595,7 +45595,7 @@ } }, { - "accuracy": 0.9941740310505817, + "accuracy": 0.9941865844946158, "total_bits": 637362176, "q_proj": { "group_size": { @@ -45656,7 +45656,7 @@ } }, { - "accuracy": 0.9947095191792438, + "accuracy": 0.9947438338085225, "total_bits": 646823936, "q_proj": { "group_size": { @@ -45717,7 +45717,7 @@ } }, { - "accuracy": 0.9967798204406312, + "accuracy": 0.9968157998825374, "total_bits": 784740352, "q_proj": { "group_size": { @@ -45778,7 +45778,7 @@ } }, { - "accuracy": 0.9973031198115725, + "accuracy": 0.9973510195942301, "total_bits": 797818880, "q_proj": { "group_size": { @@ -45839,7 +45839,7 @@ } }, { - "accuracy": 0.997262904518529, + "accuracy": 0.9973011801117345, "total_bits": 911749120, "q_proj": { "group_size": { @@ -45891,7 +45891,7 @@ } }, { - "accuracy": 0.9989049700333884, + "accuracy": 0.9989184606820345, "total_bits": 942718976, "q_proj": { "group_size": { @@ -45943,7 +45943,7 @@ } }, { - "accuracy": 0.9992734513882744, + "accuracy": 0.9992943483551866, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -45997,7 +45997,7 @@ ], "model.layers.23.mlp": [ { - "accuracy": 0.9596339934750607, + "accuracy": 0.9597936209879423, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -46049,7 +46049,7 @@ } }, { - "accuracy": 0.9607023596763611, + "accuracy": 0.9608514559896368, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -46101,7 +46101,7 @@ } }, { - "accuracy": 0.96653813123703, + "accuracy": 0.9666632194268077, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -46150,7 +46150,7 @@ } }, { - "accuracy": 0.9681309430222762, + "accuracy": 0.9682470309106928, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -46199,7 +46199,7 @@ } }, { - "accuracy": 0.9800157625424234, + "accuracy": 0.9800928156626852, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -46251,7 +46251,7 @@ } }, { - "accuracy": 0.9815764740893715, + "accuracy": 0.9816523664876035, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -46303,7 +46303,7 @@ } }, { - "accuracy": 0.9839132365427519, + "accuracy": 0.9839770840971094, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -46352,7 +46352,7 @@ } }, { - "accuracy": 0.989879332090679, + "accuracy": 0.989916268932192, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -46395,7 +46395,7 @@ } }, { - "accuracy": 0.9906957447528839, + "accuracy": 0.9907316016523462, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -46438,7 +46438,7 @@ } }, { - "accuracy": 0.9898689036306582, + "accuracy": 0.9899080317271384, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -46490,7 +46490,7 @@ } }, { - "accuracy": 0.9911125427798221, + "accuracy": 0.9911469236800545, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -46542,7 +46542,7 @@ } }, { - "accuracy": 0.9948536341911868, + "accuracy": 0.9948727057168358, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -46594,7 +46594,7 @@ } }, { - "accuracy": 0.9955801336388839, + "accuracy": 0.9955963449258554, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -46646,7 +46646,7 @@ } }, { - "accuracy": 0.9972658561248529, + "accuracy": 0.9972752007214647, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -46689,7 +46689,7 @@ } }, { - "accuracy": 0.9973733538859769, + "accuracy": 0.997383790188714, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -46738,7 +46738,7 @@ } }, { - "accuracy": 0.9978170743898341, + "accuracy": 0.9978249071067885, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -46784,7 +46784,7 @@ } }, { - "accuracy": 0.9992816254104439, + "accuracy": 0.9992844516313389, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -46826,7 +46826,7 @@ ], "model.layers.24.self_attn": [ { - "accuracy": 0.9819054666318392, + "accuracy": 0.9821461752841347, "total_bits": 320757760, "q_proj": { "group_size": { @@ -46890,7 +46890,7 @@ } }, { - "accuracy": 0.9826154379468215, + "accuracy": 0.9828545404107947, "total_bits": 329080832, "q_proj": { "group_size": { @@ -46954,7 +46954,7 @@ } }, { - "accuracy": 0.9856235118288743, + "accuracy": 0.9857999698111886, "total_bits": 336024576, "q_proj": { "group_size": { @@ -47018,7 +47018,7 @@ } }, { - "accuracy": 0.9889042989203805, + "accuracy": 0.9890704390249754, "total_bits": 401557504, "q_proj": { "group_size": { @@ -47082,7 +47082,7 @@ } }, { - "accuracy": 0.9900922453717181, + "accuracy": 0.9902317257303941, "total_bits": 475279360, "q_proj": { "group_size": { @@ -47146,7 +47146,7 @@ } }, { - "accuracy": 0.9902266930592688, + "accuracy": 0.9904180349488008, "total_bits": 475479040, "q_proj": { "group_size": { @@ -47210,7 +47210,7 @@ } }, { - "accuracy": 0.9921271204948425, + "accuracy": 0.9922698046031752, "total_bits": 609759232, "q_proj": { "group_size": { @@ -47262,7 +47262,7 @@ } }, { - "accuracy": 0.9923141386948133, + "accuracy": 0.9924428447296745, "total_bits": 610024448, "q_proj": { "group_size": { @@ -47314,7 +47314,7 @@ } }, { - "accuracy": 0.9928566852682515, + "accuracy": 0.9929597589530443, "total_bits": 615020544, "q_proj": { "group_size": { @@ -47366,7 +47366,7 @@ } }, { - "accuracy": 0.9930416094629388, + "accuracy": 0.993130228237102, "total_bits": 623951872, "q_proj": { "group_size": { @@ -47418,7 +47418,7 @@ } }, { - "accuracy": 0.994548525857298, + "accuracy": 0.9945981851533839, "total_bits": 626473984, "q_proj": { "group_size": { @@ -47482,7 +47482,7 @@ } }, { - "accuracy": 0.9950096587601461, + "accuracy": 0.9949950234670388, "total_bits": 630355968, "q_proj": { "group_size": { @@ -47546,7 +47546,7 @@ } }, { - "accuracy": 0.9949891535859359, + "accuracy": 0.9950069404746357, "total_bits": 637362176, "q_proj": { "group_size": { @@ -47607,7 +47607,7 @@ } }, { - "accuracy": 0.9954321117777574, + "accuracy": 0.9954441613272617, "total_bits": 646823936, "q_proj": { "group_size": { @@ -47668,7 +47668,7 @@ } }, { - "accuracy": 0.9972089883920393, + "accuracy": 0.9972471990867665, "total_bits": 784740352, "q_proj": { "group_size": { @@ -47729,7 +47729,7 @@ } }, { - "accuracy": 0.9977350825149762, + "accuracy": 0.9977346001879165, "total_bits": 797818880, "q_proj": { "group_size": { @@ -47790,7 +47790,7 @@ } }, { - "accuracy": 0.9975116794046602, + "accuracy": 0.9975443741208628, "total_bits": 911749120, "q_proj": { "group_size": { @@ -47842,7 +47842,7 @@ } }, { - "accuracy": 0.999042879005796, + "accuracy": 0.9990524328068683, "total_bits": 942718976, "q_proj": { "group_size": { @@ -47894,7 +47894,7 @@ } }, { - "accuracy": 0.9993767490316379, + "accuracy": 0.9993792316435199, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -47948,7 +47948,7 @@ ], "model.layers.24.mlp": [ { - "accuracy": 0.959129622108058, + "accuracy": 0.9592599272727966, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -48000,7 +48000,7 @@ } }, { - "accuracy": 0.9601536518649051, + "accuracy": 0.9602935439661929, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -48052,7 +48052,7 @@ } }, { - "accuracy": 0.9659579992294312, + "accuracy": 0.9660659181444269, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -48101,7 +48101,7 @@ } }, { - "accuracy": 0.9675554727253161, + "accuracy": 0.9676558657696372, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -48150,7 +48150,7 @@ } }, { - "accuracy": 0.9796085734116403, + "accuracy": 0.9796816722342843, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -48202,7 +48202,7 @@ } }, { - "accuracy": 0.9812330478116086, + "accuracy": 0.9812985862556257, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -48254,7 +48254,7 @@ } }, { - "accuracy": 0.9835928835366902, + "accuracy": 0.9836476539310656, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -48303,7 +48303,7 @@ } }, { - "accuracy": 0.989642532248246, + "accuracy": 0.9896777272224426, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -48346,7 +48346,7 @@ } }, { - "accuracy": 0.9904880727592268, + "accuracy": 0.9905194389192682, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -48389,7 +48389,7 @@ } }, { - "accuracy": 0.9896393645750848, + "accuracy": 0.9896738591947054, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -48441,7 +48441,7 @@ } }, { - "accuracy": 0.9909119425635589, + "accuracy": 0.9909423534807406, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -48493,7 +48493,7 @@ } }, { - "accuracy": 0.9947243258357048, + "accuracy": 0.994740695153412, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -48545,7 +48545,7 @@ } }, { - "accuracy": 0.9954751192739135, + "accuracy": 0.9954898867167925, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -48597,7 +48597,7 @@ } }, { - "accuracy": 0.9971947634690687, + "accuracy": 0.9972033063439947, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -48640,7 +48640,7 @@ } }, { - "accuracy": 0.9973074127184717, + "accuracy": 0.9973155385569522, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -48689,7 +48689,7 @@ } }, { - "accuracy": 0.9977602841038453, + "accuracy": 0.9977662475093415, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -48735,7 +48735,7 @@ } }, { - "accuracy": 0.9992559484059089, + "accuracy": 0.9992580176576188, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -48777,7 +48777,7 @@ ], "model.layers.25.self_attn": [ { - "accuracy": 0.9777047994889712, + "accuracy": 0.9773751371785214, "total_bits": 320757760, "q_proj": { "group_size": { @@ -48841,7 +48841,7 @@ } }, { - "accuracy": 0.9779140808080372, + "accuracy": 0.9783948032479537, "total_bits": 329080832, "q_proj": { "group_size": { @@ -48905,7 +48905,7 @@ } }, { - "accuracy": 0.979844364680742, + "accuracy": 0.9800848286402853, "total_bits": 336024576, "q_proj": { "group_size": { @@ -48969,7 +48969,7 @@ } }, { - "accuracy": 0.9825152174422616, + "accuracy": 0.9829127270924417, "total_bits": 401557504, "q_proj": { "group_size": { @@ -49033,7 +49033,7 @@ } }, { - "accuracy": 0.9847685914290579, + "accuracy": 0.9849554676758615, "total_bits": 475279360, "q_proj": { "group_size": { @@ -49097,7 +49097,7 @@ } }, { - "accuracy": 0.9854450006234018, + "accuracy": 0.984764354793649, "total_bits": 475479040, "q_proj": { "group_size": { @@ -49161,7 +49161,7 @@ } }, { - "accuracy": 0.9905102566668862, + "accuracy": 0.9905638686920467, "total_bits": 609759232, "q_proj": { "group_size": { @@ -49213,7 +49213,7 @@ } }, { - "accuracy": 0.9906619991126814, + "accuracy": 0.9907821063932619, "total_bits": 610024448, "q_proj": { "group_size": { @@ -49265,7 +49265,7 @@ } }, { - "accuracy": 0.9910774670149151, + "accuracy": 0.9914219128458124, "total_bits": 615020544, "q_proj": { "group_size": { @@ -49317,7 +49317,7 @@ } }, { - "accuracy": 0.9911947022927435, + "accuracy": 0.9915334080394945, "total_bits": 623951872, "q_proj": { "group_size": { @@ -49369,7 +49369,7 @@ } }, { - "accuracy": 0.9910949382342791, + "accuracy": 0.9907150331296419, "total_bits": 626473984, "q_proj": { "group_size": { @@ -49433,7 +49433,7 @@ } }, { - "accuracy": 0.9916555058015021, + "accuracy": 0.9916954236595255, "total_bits": 630355968, "q_proj": { "group_size": { @@ -49497,7 +49497,7 @@ } }, { - "accuracy": 0.9943335107282588, + "accuracy": 0.9942814562665788, "total_bits": 637362176, "q_proj": { "group_size": { @@ -49558,7 +49558,7 @@ } }, { - "accuracy": 0.9947180257816064, + "accuracy": 0.9947328343987465, "total_bits": 646823936, "q_proj": { "group_size": { @@ -49619,7 +49619,7 @@ } }, { - "accuracy": 0.9968179436890703, + "accuracy": 0.9968384788616708, "total_bits": 784740352, "q_proj": { "group_size": { @@ -49680,7 +49680,7 @@ } }, { - "accuracy": 0.9973038731044844, + "accuracy": 0.9972734808137542, "total_bits": 797818880, "q_proj": { "group_size": { @@ -49741,7 +49741,7 @@ } }, { - "accuracy": 0.9972090485848879, + "accuracy": 0.9972179037959952, "total_bits": 911749120, "q_proj": { "group_size": { @@ -49793,7 +49793,7 @@ } }, { - "accuracy": 0.9989444636004535, + "accuracy": 0.9989458488202408, "total_bits": 942718976, "q_proj": { "group_size": { @@ -49845,7 +49845,7 @@ } }, { - "accuracy": 0.9992760197308502, + "accuracy": 0.9992611746450788, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -49899,7 +49899,7 @@ ], "model.layers.25.mlp": [ { - "accuracy": 0.9577326429517645, + "accuracy": 0.9578741443784613, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -49951,7 +49951,7 @@ } }, { - "accuracy": 0.9588356676854586, + "accuracy": 0.9589715003967285, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -50003,7 +50003,7 @@ } }, { - "accuracy": 0.9647704927544845, + "accuracy": 0.9648777905263399, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -50052,7 +50052,7 @@ } }, { - "accuracy": 0.9664005292089362, + "accuracy": 0.9665015873156095, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -50101,7 +50101,7 @@ } }, { - "accuracy": 0.9789462481674395, + "accuracy": 0.9790142680469313, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -50153,7 +50153,7 @@ } }, { - "accuracy": 0.9806254019862727, + "accuracy": 0.9806903349725824, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -50205,7 +50205,7 @@ } }, { - "accuracy": 0.9830102889161361, + "accuracy": 0.9830655973208579, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -50254,7 +50254,7 @@ } }, { - "accuracy": 0.9893159427140888, + "accuracy": 0.9893509800496855, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -50297,7 +50297,7 @@ } }, { - "accuracy": 0.9901776211826425, + "accuracy": 0.9902087991174898, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -50340,7 +50340,7 @@ } }, { - "accuracy": 0.9893088536827188, + "accuracy": 0.9893457913085034, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -50392,7 +50392,7 @@ } }, { - "accuracy": 0.9906270872605475, + "accuracy": 0.9906558904208635, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -50444,7 +50444,7 @@ } }, { - "accuracy": 0.994558362584365, + "accuracy": 0.9945763486780619, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -50496,7 +50496,7 @@ } }, { - "accuracy": 0.9953305415417019, + "accuracy": 0.9953456740630301, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -50548,7 +50548,7 @@ } }, { - "accuracy": 0.9970980831666997, + "accuracy": 0.9971073654137159, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -50591,7 +50591,7 @@ } }, { - "accuracy": 0.9972204128770452, + "accuracy": 0.997230199802863, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -50640,7 +50640,7 @@ } }, { - "accuracy": 0.997672574496583, + "accuracy": 0.997680390351697, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -50686,7 +50686,7 @@ } }, { - "accuracy": 0.9992297356948256, + "accuracy": 0.9992328942018119, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -50728,7 +50728,7 @@ ], "model.layers.26.self_attn": [ { - "accuracy": 0.9762562811374664, + "accuracy": 0.9755339936206215, "total_bits": 320757760, "q_proj": { "group_size": { @@ -50792,7 +50792,7 @@ } }, { - "accuracy": 0.9780023897949018, + "accuracy": 0.975313192919681, "total_bits": 329080832, "q_proj": { "group_size": { @@ -50856,7 +50856,7 @@ } }, { - "accuracy": 0.9799747592524478, + "accuracy": 0.9794920306456717, "total_bits": 336024576, "q_proj": { "group_size": { @@ -50920,7 +50920,7 @@ } }, { - "accuracy": 0.9832164221688321, + "accuracy": 0.9825396569151628, "total_bits": 401557504, "q_proj": { "group_size": { @@ -50984,7 +50984,7 @@ } }, { - "accuracy": 0.9859308082806436, + "accuracy": 0.9861054389100326, "total_bits": 475279360, "q_proj": { "group_size": { @@ -51048,7 +51048,7 @@ } }, { - "accuracy": 0.9860382974147797, + "accuracy": 0.9863721618526861, "total_bits": 475479040, "q_proj": { "group_size": { @@ -51112,7 +51112,7 @@ } }, { - "accuracy": 0.9903558178951866, + "accuracy": 0.9907310361924925, "total_bits": 609759232, "q_proj": { "group_size": { @@ -51164,7 +51164,7 @@ } }, { - "accuracy": 0.9912044021644091, + "accuracy": 0.991318691717951, "total_bits": 610024448, "q_proj": { "group_size": { @@ -51216,7 +51216,7 @@ } }, { - "accuracy": 0.99175552474825, + "accuracy": 0.9919482375446119, "total_bits": 615020544, "q_proj": { "group_size": { @@ -51268,7 +51268,7 @@ } }, { - "accuracy": 0.9918872157209798, + "accuracy": 0.9921927687368894, "total_bits": 623951872, "q_proj": { "group_size": { @@ -51320,7 +51320,7 @@ } }, { - "accuracy": 0.9912414809590891, + "accuracy": 0.9912879741505572, "total_bits": 626473984, "q_proj": { "group_size": { @@ -51384,7 +51384,7 @@ } }, { - "accuracy": 0.9920002465185366, + "accuracy": 0.9918787863693739, "total_bits": 630355968, "q_proj": { "group_size": { @@ -51448,7 +51448,7 @@ } }, { - "accuracy": 0.9942601158430702, + "accuracy": 0.9943384830104677, "total_bits": 637362176, "q_proj": { "group_size": { @@ -51509,7 +51509,7 @@ } }, { - "accuracy": 0.9947692685221371, + "accuracy": 0.9948039733265576, "total_bits": 646823936, "q_proj": { "group_size": { @@ -51570,7 +51570,7 @@ } }, { - "accuracy": 0.9968793552957083, + "accuracy": 0.996965599491408, "total_bits": 784740352, "q_proj": { "group_size": { @@ -51631,7 +51631,7 @@ } }, { - "accuracy": 0.9973425363239489, + "accuracy": 0.997417443088795, "total_bits": 797818880, "q_proj": { "group_size": { @@ -51692,7 +51692,7 @@ } }, { - "accuracy": 0.9973303752118036, + "accuracy": 0.9974187465482637, "total_bits": 911749120, "q_proj": { "group_size": { @@ -51744,7 +51744,7 @@ } }, { - "accuracy": 0.9989650242618824, + "accuracy": 0.998971039429307, "total_bits": 942718976, "q_proj": { "group_size": { @@ -51796,7 +51796,7 @@ } }, { - "accuracy": 0.9993348348101503, + "accuracy": 0.9993537935967508, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -51850,7 +51850,7 @@ ], "model.layers.26.mlp": [ { - "accuracy": 0.956966619742544, + "accuracy": 0.9571049087925961, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -51902,7 +51902,7 @@ } }, { - "accuracy": 0.958136426775079, + "accuracy": 0.9582706062417281, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -51954,7 +51954,7 @@ } }, { - "accuracy": 0.9640763935289884, + "accuracy": 0.9641995869184795, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -52003,7 +52003,7 @@ } }, { - "accuracy": 0.9657345445532548, + "accuracy": 0.9658494434858623, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -52052,7 +52052,7 @@ } }, { - "accuracy": 0.9786440588926014, + "accuracy": 0.9787149131298065, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -52104,7 +52104,7 @@ } }, { - "accuracy": 0.980297011764426, + "accuracy": 0.9803646551935297, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -52156,7 +52156,7 @@ } }, { - "accuracy": 0.9826924157770056, + "accuracy": 0.9827495468290228, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -52205,7 +52205,7 @@ } }, { - "accuracy": 0.9891314067338642, + "accuracy": 0.9891691733347742, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -52248,7 +52248,7 @@ } }, { - "accuracy": 0.989985026027027, + "accuracy": 0.9900179654359818, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -52291,7 +52291,7 @@ } }, { - "accuracy": 0.9891671107003563, + "accuracy": 0.9892055251096424, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -52343,7 +52343,7 @@ } }, { - "accuracy": 0.9904659068898151, + "accuracy": 0.9904995129296654, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -52395,7 +52395,7 @@ } }, { - "accuracy": 0.9944941558335957, + "accuracy": 0.9945111757046298, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -52447,7 +52447,7 @@ } }, { - "accuracy": 0.9952543224943312, + "accuracy": 0.9952713976565161, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -52499,7 +52499,7 @@ } }, { - "accuracy": 0.9970609637859621, + "accuracy": 0.9970705742506605, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -52542,7 +52542,7 @@ } }, { - "accuracy": 0.9971929800353552, + "accuracy": 0.9972024791334805, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -52591,7 +52591,7 @@ } }, { - "accuracy": 0.9976543621405175, + "accuracy": 0.9976617909575763, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -52637,7 +52637,7 @@ } }, { - "accuracy": 0.9992294672778562, + "accuracy": 0.999232502311076, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -52679,7 +52679,7 @@ ], "model.layers.27.self_attn": [ { - "accuracy": 0.9716228905477022, + "accuracy": 0.9715040075151544, "total_bits": 320757760, "q_proj": { "group_size": { @@ -52743,7 +52743,7 @@ } }, { - "accuracy": 0.9731114789059288, + "accuracy": 0.9734404243920979, "total_bits": 329080832, "q_proj": { "group_size": { @@ -52807,7 +52807,7 @@ } }, { - "accuracy": 0.9758435017184207, + "accuracy": 0.9768642045949635, "total_bits": 336024576, "q_proj": { "group_size": { @@ -52871,7 +52871,7 @@ } }, { - "accuracy": 0.9809489124699643, + "accuracy": 0.9821196606284693, "total_bits": 401557504, "q_proj": { "group_size": { @@ -52935,7 +52935,7 @@ } }, { - "accuracy": 0.9852091356327659, + "accuracy": 0.9867233508511594, "total_bits": 475279360, "q_proj": { "group_size": { @@ -52999,7 +52999,7 @@ } }, { - "accuracy": 0.9855987492360567, + "accuracy": 0.9863792548054143, "total_bits": 475479040, "q_proj": { "group_size": { @@ -53063,7 +53063,7 @@ } }, { - "accuracy": 0.9919850049834502, + "accuracy": 0.9919860363006592, "total_bits": 609759232, "q_proj": { "group_size": { @@ -53115,7 +53115,7 @@ } }, { - "accuracy": 0.9921549715493855, + "accuracy": 0.992021918296814, "total_bits": 610024448, "q_proj": { "group_size": { @@ -53167,7 +53167,7 @@ } }, { - "accuracy": 0.9927500580486498, + "accuracy": 0.9928608544562992, "total_bits": 615020544, "q_proj": { "group_size": { @@ -53219,7 +53219,7 @@ } }, { - "accuracy": 0.9930723842821623, + "accuracy": 0.9931252920313886, "total_bits": 623951872, "q_proj": { "group_size": { @@ -53271,7 +53271,7 @@ } }, { - "accuracy": 0.9925303600336376, + "accuracy": 0.9926454118992153, "total_bits": 626473984, "q_proj": { "group_size": { @@ -53335,7 +53335,7 @@ } }, { - "accuracy": 0.9930765330791473, + "accuracy": 0.9931428252082122, "total_bits": 630355968, "q_proj": { "group_size": { @@ -53399,7 +53399,7 @@ } }, { - "accuracy": 0.9942750562178461, + "accuracy": 0.9942665798099417, "total_bits": 637362176, "q_proj": { "group_size": { @@ -53460,7 +53460,7 @@ } }, { - "accuracy": 0.9948549352978405, + "accuracy": 0.9948947468870565, "total_bits": 646823936, "q_proj": { "group_size": { @@ -53521,7 +53521,7 @@ } }, { - "accuracy": 0.9969240505444376, + "accuracy": 0.9969784484097832, "total_bits": 784740352, "q_proj": { "group_size": { @@ -53582,7 +53582,7 @@ } }, { - "accuracy": 0.9973763219620052, + "accuracy": 0.9974176956242636, "total_bits": 797818880, "q_proj": { "group_size": { @@ -53643,7 +53643,7 @@ } }, { - "accuracy": 0.9978732484343805, + "accuracy": 0.9979173005803635, "total_bits": 911749120, "q_proj": { "group_size": { @@ -53695,7 +53695,7 @@ } }, { - "accuracy": 0.9986914217864212, + "accuracy": 0.9987049607658073, "total_bits": 942718976, "q_proj": { "group_size": { @@ -53747,7 +53747,7 @@ } }, { - "accuracy": 0.9994318234783254, + "accuracy": 0.9994328627853017, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -53801,7 +53801,7 @@ ], "model.layers.27.mlp": [ { - "accuracy": 0.955314648778815, + "accuracy": 0.9554841424289503, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -53853,7 +53853,7 @@ } }, { - "accuracy": 0.9565011702085796, + "accuracy": 0.9566748675547148, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -53905,7 +53905,7 @@ } }, { - "accuracy": 0.9630346486442968, + "accuracy": 0.9631753971702174, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -53954,7 +53954,7 @@ } }, { - "accuracy": 0.9648257305747584, + "accuracy": 0.9649573690012881, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -54003,7 +54003,7 @@ } }, { - "accuracy": 0.977874933104766, + "accuracy": 0.9779576565089979, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -54055,7 +54055,7 @@ } }, { - "accuracy": 0.9795821105179033, + "accuracy": 0.9796598835995323, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -54107,7 +54107,7 @@ } }, { - "accuracy": 0.9821990869547191, + "accuracy": 0.9822632952740318, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -54156,7 +54156,7 @@ } }, { - "accuracy": 0.9887343315701735, + "accuracy": 0.9887767073355223, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -54199,7 +54199,7 @@ } }, { - "accuracy": 0.989635274598473, + "accuracy": 0.9896711032641562, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -54242,7 +54242,7 @@ } }, { - "accuracy": 0.9887806302622745, + "accuracy": 0.9888228578002829, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -54294,7 +54294,7 @@ } }, { - "accuracy": 0.9901290684938431, + "accuracy": 0.9901654994801471, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -54346,7 +54346,7 @@ } }, { - "accuracy": 0.9942953296397862, + "accuracy": 0.9943169025998366, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -54398,7 +54398,7 @@ } }, { - "accuracy": 0.9950852143137079, + "accuracy": 0.9951038850765479, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -54450,7 +54450,7 @@ } }, { - "accuracy": 0.9969553527863402, + "accuracy": 0.9969662976892371, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -54493,7 +54493,7 @@ } }, { - "accuracy": 0.9970925162105184, + "accuracy": 0.9971038277604078, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -54542,7 +54542,7 @@ } }, { - "accuracy": 0.9976041428744793, + "accuracy": 0.9976127555495814, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -54588,7 +54588,7 @@ } }, { - "accuracy": 0.9991989880800247, + "accuracy": 0.9992019290496644, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -54630,7 +54630,7 @@ ], "model.layers.28.self_attn": [ { - "accuracy": 0.9738610888782301, + "accuracy": 0.97349866126713, "total_bits": 320757760, "q_proj": { "group_size": { @@ -54694,7 +54694,7 @@ } }, { - "accuracy": 0.9746474388398623, + "accuracy": 0.9742599098305953, "total_bits": 329080832, "q_proj": { "group_size": { @@ -54758,7 +54758,7 @@ } }, { - "accuracy": 0.9776925821053354, + "accuracy": 0.9778298986585516, "total_bits": 336024576, "q_proj": { "group_size": { @@ -54822,7 +54822,7 @@ } }, { - "accuracy": 0.9827634456910586, + "accuracy": 0.9828355971135592, "total_bits": 401557504, "q_proj": { "group_size": { @@ -54886,7 +54886,7 @@ } }, { - "accuracy": 0.9861497831972021, + "accuracy": 0.9863933371870142, "total_bits": 475279360, "q_proj": { "group_size": { @@ -54950,7 +54950,7 @@ } }, { - "accuracy": 0.9865390830918362, + "accuracy": 0.9867562306554694, "total_bits": 475479040, "q_proj": { "group_size": { @@ -55014,7 +55014,7 @@ } }, { - "accuracy": 0.9900580668135693, + "accuracy": 0.9901941308849737, "total_bits": 609759232, "q_proj": { "group_size": { @@ -55066,7 +55066,7 @@ } }, { - "accuracy": 0.9904427771505556, + "accuracy": 0.9907213143612209, "total_bits": 610024448, "q_proj": { "group_size": { @@ -55118,7 +55118,7 @@ } }, { - "accuracy": 0.9911647054709887, + "accuracy": 0.9911831672254362, "total_bits": 615020544, "q_proj": { "group_size": { @@ -55170,7 +55170,7 @@ } }, { - "accuracy": 0.9913590189657713, + "accuracy": 0.9914898676307577, "total_bits": 623951872, "q_proj": { "group_size": { @@ -55222,7 +55222,7 @@ } }, { - "accuracy": 0.9929419152046505, + "accuracy": 0.9930622373756609, "total_bits": 626473984, "q_proj": { "group_size": { @@ -55286,7 +55286,7 @@ } }, { - "accuracy": 0.9933965896305285, + "accuracy": 0.9933393668187293, "total_bits": 630355968, "q_proj": { "group_size": { @@ -55350,7 +55350,7 @@ } }, { - "accuracy": 0.9933423345026217, + "accuracy": 0.9935926822455305, "total_bits": 637362176, "q_proj": { "group_size": { @@ -55411,7 +55411,7 @@ } }, { - "accuracy": 0.9940297944765342, + "accuracy": 0.9938472705452066, "total_bits": 646823936, "q_proj": { "group_size": { @@ -55472,7 +55472,7 @@ } }, { - "accuracy": 0.9963218181541091, + "accuracy": 0.9963922837847158, "total_bits": 784740352, "q_proj": { "group_size": { @@ -55533,7 +55533,7 @@ } }, { - "accuracy": 0.9969229276635145, + "accuracy": 0.9969111869209691, "total_bits": 797818880, "q_proj": { "group_size": { @@ -55594,7 +55594,7 @@ } }, { - "accuracy": 0.9970948700057832, + "accuracy": 0.9971610341025027, "total_bits": 911749120, "q_proj": { "group_size": { @@ -55646,7 +55646,7 @@ } }, { - "accuracy": 0.9986296170636227, + "accuracy": 0.9986109396344737, "total_bits": 942718976, "q_proj": { "group_size": { @@ -55698,7 +55698,7 @@ } }, { - "accuracy": 0.9992697850654, + "accuracy": 0.9992560782518826, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -55752,7 +55752,7 @@ ], "model.layers.28.mlp": [ { - "accuracy": 0.9538879802352503, + "accuracy": 0.9540708284629018, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -55804,7 +55804,7 @@ } }, { - "accuracy": 0.9551119961236653, + "accuracy": 0.9552905559539795, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -55856,7 +55856,7 @@ } }, { - "accuracy": 0.9619037885414927, + "accuracy": 0.9620546384861595, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -55905,7 +55905,7 @@ } }, { - "accuracy": 0.9637731219592848, + "accuracy": 0.9639102691098264, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -55954,7 +55954,7 @@ } }, { - "accuracy": 0.9771044693495098, + "accuracy": 0.97719197367367, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -56006,7 +56006,7 @@ } }, { - "accuracy": 0.9788941003774342, + "accuracy": 0.9789723110826392, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -56058,7 +56058,7 @@ } }, { - "accuracy": 0.9816239115438963, + "accuracy": 0.9816875442078239, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -56107,7 +56107,7 @@ } }, { - "accuracy": 0.9883157579522384, + "accuracy": 0.9883560491235632, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -56150,7 +56150,7 @@ } }, { - "accuracy": 0.9892408588999196, + "accuracy": 0.9892808920458743, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -56193,7 +56193,7 @@ } }, { - "accuracy": 0.988361227669214, + "accuracy": 0.988401925877521, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -56245,7 +56245,7 @@ } }, { - "accuracy": 0.9897560806650865, + "accuracy": 0.989792799479083, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -56297,7 +56297,7 @@ } }, { - "accuracy": 0.9940680002695635, + "accuracy": 0.9940873764847454, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -56349,7 +56349,7 @@ } }, { - "accuracy": 0.9948926869975893, + "accuracy": 0.9949112432567697, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -56401,7 +56401,7 @@ } }, { - "accuracy": 0.9968279284474096, + "accuracy": 0.9968367777764797, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -56444,7 +56444,7 @@ } }, { - "accuracy": 0.9969696210403192, + "accuracy": 0.9969807198565257, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -56493,7 +56493,7 @@ } }, { - "accuracy": 0.9975027773333224, + "accuracy": 0.9975105975019304, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -56539,7 +56539,7 @@ } }, { - "accuracy": 0.9991582506581357, + "accuracy": 0.9991615357759752, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -56581,7 +56581,7 @@ ], "model.layers.29.self_attn": [ { - "accuracy": 0.9722018618332712, + "accuracy": 0.9718894770270899, "total_bits": 320757760, "q_proj": { "group_size": { @@ -56645,7 +56645,7 @@ } }, { - "accuracy": 0.9723685446538424, + "accuracy": 0.9729943746014645, "total_bits": 329080832, "q_proj": { "group_size": { @@ -56709,7 +56709,7 @@ } }, { - "accuracy": 0.9759551929800134, + "accuracy": 0.9762259643328818, "total_bits": 336024576, "q_proj": { "group_size": { @@ -56773,7 +56773,7 @@ } }, { - "accuracy": 0.9814347954172837, + "accuracy": 0.9816555537675556, "total_bits": 401557504, "q_proj": { "group_size": { @@ -56837,7 +56837,7 @@ } }, { - "accuracy": 0.9853763611693132, + "accuracy": 0.9854564196185062, "total_bits": 475279360, "q_proj": { "group_size": { @@ -56901,7 +56901,7 @@ } }, { - "accuracy": 0.9857880598620364, + "accuracy": 0.9858189413422033, "total_bits": 475479040, "q_proj": { "group_size": { @@ -56965,7 +56965,7 @@ } }, { - "accuracy": 0.989635598502661, + "accuracy": 0.9896991017617678, "total_bits": 609759232, "q_proj": { "group_size": { @@ -57017,7 +57017,7 @@ } }, { - "accuracy": 0.9900632390850469, + "accuracy": 0.9901268341039356, "total_bits": 610024448, "q_proj": { "group_size": { @@ -57069,7 +57069,7 @@ } }, { - "accuracy": 0.9905973547383359, + "accuracy": 0.990648333963595, "total_bits": 615020544, "q_proj": { "group_size": { @@ -57121,7 +57121,7 @@ } }, { - "accuracy": 0.9909579989157224, + "accuracy": 0.9909488143105256, "total_bits": 623951872, "q_proj": { "group_size": { @@ -57173,7 +57173,7 @@ } }, { - "accuracy": 0.992737385787462, + "accuracy": 0.9928041076973865, "total_bits": 626473984, "q_proj": { "group_size": { @@ -57237,7 +57237,7 @@ } }, { - "accuracy": 0.9931919127702713, + "accuracy": 0.99323476145142, "total_bits": 630355968, "q_proj": { "group_size": { @@ -57301,7 +57301,7 @@ } }, { - "accuracy": 0.9933621428514782, + "accuracy": 0.9934079074545911, "total_bits": 637362176, "q_proj": { "group_size": { @@ -57362,7 +57362,7 @@ } }, { - "accuracy": 0.9938238286658337, + "accuracy": 0.9938744165395436, "total_bits": 646823936, "q_proj": { "group_size": { @@ -57423,7 +57423,7 @@ } }, { - "accuracy": 0.9963624093093371, + "accuracy": 0.9963676764776832, "total_bits": 784740352, "q_proj": { "group_size": { @@ -57484,7 +57484,7 @@ } }, { - "accuracy": 0.9968292185742604, + "accuracy": 0.9968631220491309, "total_bits": 797818880, "q_proj": { "group_size": { @@ -57545,7 +57545,7 @@ } }, { - "accuracy": 0.9972952682720987, + "accuracy": 0.9972974275679964, "total_bits": 911749120, "q_proj": { "group_size": { @@ -57597,7 +57597,7 @@ } }, { - "accuracy": 0.9985190889749088, + "accuracy": 0.9984992661170269, "total_bits": 942718976, "q_proj": { "group_size": { @@ -57649,7 +57649,7 @@ } }, { - "accuracy": 0.9992202207151997, + "accuracy": 0.9992290339677742, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -57703,7 +57703,7 @@ ], "model.layers.29.mlp": [ { - "accuracy": 0.9530642628669739, + "accuracy": 0.95323688105533, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -57755,7 +57755,7 @@ } }, { - "accuracy": 0.9543444577016329, + "accuracy": 0.9545184656193382, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -57807,7 +57807,7 @@ } }, { - "accuracy": 0.9611886551505641, + "accuracy": 0.9613357688251295, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -57856,7 +57856,7 @@ } }, { - "accuracy": 0.9630904072209409, + "accuracy": 0.9632255466360795, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -57905,7 +57905,7 @@ } }, { - "accuracy": 0.9767542402995261, + "accuracy": 0.9768453099225697, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -57957,7 +57957,7 @@ } }, { - "accuracy": 0.9785469315553966, + "accuracy": 0.9786290149939688, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -58009,7 +58009,7 @@ } }, { - "accuracy": 0.9812845509303244, + "accuracy": 0.9813536719272011, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -58058,7 +58058,7 @@ } }, { - "accuracy": 0.9881277186305899, + "accuracy": 0.9881730573742014, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -58101,7 +58101,7 @@ } }, { - "accuracy": 0.9890538054077249, + "accuracy": 0.9890951984807065, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -58144,7 +58144,7 @@ } }, { - "accuracy": 0.988194603669016, + "accuracy": 0.9882376162629378, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -58196,7 +58196,7 @@ } }, { - "accuracy": 0.9895999596307152, + "accuracy": 0.9896394469236073, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -58248,7 +58248,7 @@ } }, { - "accuracy": 0.9939872308781272, + "accuracy": 0.9940104433580449, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -58300,7 +58300,7 @@ } }, { - "accuracy": 0.9948174776999574, + "accuracy": 0.9948366037324855, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -58352,7 +58352,7 @@ } }, { - "accuracy": 0.9967755240045095, + "accuracy": 0.996788144895905, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -58395,7 +58395,7 @@ } }, { - "accuracy": 0.9969249118707681, + "accuracy": 0.9969367331972248, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -58444,7 +58444,7 @@ } }, { - "accuracy": 0.9974512793123722, + "accuracy": 0.9974606958659071, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -58490,7 +58490,7 @@ } }, { - "accuracy": 0.9991157179404246, + "accuracy": 0.9991185640622127, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -58532,7 +58532,7 @@ ], "model.layers.30.self_attn": [ { - "accuracy": 0.97216198318883, + "accuracy": 0.9723838661846361, "total_bits": 320757760, "q_proj": { "group_size": { @@ -58596,7 +58596,7 @@ } }, { - "accuracy": 0.9728711655265406, + "accuracy": 0.9730965055917439, "total_bits": 329080832, "q_proj": { "group_size": { @@ -58660,7 +58660,7 @@ } }, { - "accuracy": 0.9767379321550068, + "accuracy": 0.9769104985814345, "total_bits": 336024576, "q_proj": { "group_size": { @@ -58724,7 +58724,7 @@ } }, { - "accuracy": 0.9829811657729902, + "accuracy": 0.983072629100398, "total_bits": 401557504, "q_proj": { "group_size": { @@ -58788,7 +58788,7 @@ } }, { - "accuracy": 0.9856587820931485, + "accuracy": 0.9857571423053741, "total_bits": 475279360, "q_proj": { "group_size": { @@ -58852,7 +58852,7 @@ } }, { - "accuracy": 0.985993808821628, + "accuracy": 0.9860321408823917, "total_bits": 475479040, "q_proj": { "group_size": { @@ -58916,7 +58916,7 @@ } }, { - "accuracy": 0.9898090354706112, + "accuracy": 0.990000138941564, "total_bits": 609759232, "q_proj": { "group_size": { @@ -58968,7 +58968,7 @@ } }, { - "accuracy": 0.9903921610430667, + "accuracy": 0.9903166137243572, "total_bits": 610024448, "q_proj": { "group_size": { @@ -59020,7 +59020,7 @@ } }, { - "accuracy": 0.9911830440947884, + "accuracy": 0.9912324979117042, "total_bits": 615020544, "q_proj": { "group_size": { @@ -59072,7 +59072,7 @@ } }, { - "accuracy": 0.9913870292274576, + "accuracy": 0.99152698563902, "total_bits": 623951872, "q_proj": { "group_size": { @@ -59124,7 +59124,7 @@ } }, { - "accuracy": 0.9928084964814939, + "accuracy": 0.9928562076468217, "total_bits": 626473984, "q_proj": { "group_size": { @@ -59188,7 +59188,7 @@ } }, { - "accuracy": 0.9931579326328478, + "accuracy": 0.9932233353978709, "total_bits": 630355968, "q_proj": { "group_size": { @@ -59252,7 +59252,7 @@ } }, { - "accuracy": 0.9935108591851435, + "accuracy": 0.9935766191858995, "total_bits": 637362176, "q_proj": { "group_size": { @@ -59313,7 +59313,7 @@ } }, { - "accuracy": 0.993921370882737, + "accuracy": 0.9939654277343499, "total_bits": 646823936, "q_proj": { "group_size": { @@ -59374,7 +59374,7 @@ } }, { - "accuracy": 0.9964265858656481, + "accuracy": 0.9964266239028228, "total_bits": 784740352, "q_proj": { "group_size": { @@ -59435,7 +59435,7 @@ } }, { - "accuracy": 0.9968979609640021, + "accuracy": 0.9969561741148171, "total_bits": 797818880, "q_proj": { "group_size": { @@ -59496,7 +59496,7 @@ } }, { - "accuracy": 0.9973428702276004, + "accuracy": 0.9973415438281862, "total_bits": 911749120, "q_proj": { "group_size": { @@ -59548,7 +59548,7 @@ } }, { - "accuracy": 0.9985481778061703, + "accuracy": 0.9985559372917602, "total_bits": 942718976, "q_proj": { "group_size": { @@ -59600,7 +59600,7 @@ } }, { - "accuracy": 0.999270358662072, + "accuracy": 0.9992730001869955, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -59654,7 +59654,7 @@ ], "model.layers.30.mlp": [ { - "accuracy": 0.9536752669434798, + "accuracy": 0.9538773235521818, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -59706,7 +59706,7 @@ } }, { - "accuracy": 0.9549426091344733, + "accuracy": 0.9551317158498263, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -59758,7 +59758,7 @@ } }, { - "accuracy": 0.9617465859965274, + "accuracy": 0.9619051061178508, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -59807,7 +59807,7 @@ } }, { - "accuracy": 0.9636587720168264, + "accuracy": 0.9638069177928724, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -59856,7 +59856,7 @@ } }, { - "accuracy": 0.9770454952591344, + "accuracy": 0.9771389773017481, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -59908,7 +59908,7 @@ } }, { - "accuracy": 0.9788132356969934, + "accuracy": 0.9789061201246161, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -59960,7 +59960,7 @@ } }, { - "accuracy": 0.9815552830696106, + "accuracy": 0.9816355862115559, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -60009,7 +60009,7 @@ } }, { - "accuracy": 0.9882500218717676, + "accuracy": 0.9882980208647879, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -60052,7 +60052,7 @@ } }, { - "accuracy": 0.9891761392354965, + "accuracy": 0.9892258150012869, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -60095,7 +60095,7 @@ } }, { - "accuracy": 0.9883322935355338, + "accuracy": 0.9883830806142405, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -60147,7 +60147,7 @@ } }, { - "accuracy": 0.9897188818768451, + "accuracy": 0.9897648794086356, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -60199,7 +60199,7 @@ } }, { - "accuracy": 0.9940555628977323, + "accuracy": 0.9940790604603919, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -60251,7 +60251,7 @@ } }, { - "accuracy": 0.9948725233736792, + "accuracy": 0.9948966134535638, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -60303,7 +60303,7 @@ } }, { - "accuracy": 0.9968131796309823, + "accuracy": 0.9968252834912977, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -60346,7 +60346,7 @@ } }, { - "accuracy": 0.9969575089451513, + "accuracy": 0.9969697351518431, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -60395,7 +60395,7 @@ } }, { - "accuracy": 0.9975041096147738, + "accuracy": 0.9975136841990446, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -60441,7 +60441,7 @@ } }, { - "accuracy": 0.9991234844844592, + "accuracy": 0.999126856558417, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -60483,7 +60483,7 @@ ], "model.layers.31.self_attn": [ { - "accuracy": 0.9611565564808092, + "accuracy": 0.9616820937708804, "total_bits": 320757760, "q_proj": { "group_size": { @@ -60547,7 +60547,7 @@ } }, { - "accuracy": 0.9638263269474632, + "accuracy": 0.9634760963289362, "total_bits": 329080832, "q_proj": { "group_size": { @@ -60611,7 +60611,7 @@ } }, { - "accuracy": 0.9670832784552323, + "accuracy": 0.9674990898684451, "total_bits": 336024576, "q_proj": { "group_size": { @@ -60675,7 +60675,7 @@ } }, { - "accuracy": 0.9742032822809721, + "accuracy": 0.9745928441223345, "total_bits": 401557504, "q_proj": { "group_size": { @@ -60739,7 +60739,7 @@ } }, { - "accuracy": 0.9807676092574471, + "accuracy": 0.9810753119619269, "total_bits": 475279360, "q_proj": { "group_size": { @@ -60803,7 +60803,7 @@ } }, { - "accuracy": 0.9808652338228727, + "accuracy": 0.9811817313495436, "total_bits": 475479040, "q_proj": { "group_size": { @@ -60867,7 +60867,7 @@ } }, { - "accuracy": 0.989666536450386, + "accuracy": 0.989535106640113, "total_bits": 609759232, "q_proj": { "group_size": { @@ -60919,7 +60919,7 @@ } }, { - "accuracy": 0.9899778279818987, + "accuracy": 0.9898071838052649, "total_bits": 610024448, "q_proj": { "group_size": { @@ -60971,7 +60971,7 @@ } }, { - "accuracy": 0.9903793664355027, + "accuracy": 0.9906579883475053, "total_bits": 615020544, "q_proj": { "group_size": { @@ -61023,7 +61023,7 @@ } }, { - "accuracy": 0.9908281447071778, + "accuracy": 0.9910202661627218, "total_bits": 623951872, "q_proj": { "group_size": { @@ -61075,7 +61075,7 @@ } }, { - "accuracy": 0.9904571622610092, + "accuracy": 0.990480553946997, "total_bits": 626473984, "q_proj": { "group_size": { @@ -61139,7 +61139,7 @@ } }, { - "accuracy": 0.9910551866418437, + "accuracy": 0.9911425192105142, "total_bits": 630355968, "q_proj": { "group_size": { @@ -61203,7 +61203,7 @@ } }, { - "accuracy": 0.992141129154908, + "accuracy": 0.9922091764839072, "total_bits": 637362176, "q_proj": { "group_size": { @@ -61264,7 +61264,7 @@ } }, { - "accuracy": 0.99290915460963, + "accuracy": 0.9929546596188294, "total_bits": 646823936, "q_proj": { "group_size": { @@ -61325,7 +61325,7 @@ } }, { - "accuracy": 0.9958019123265618, + "accuracy": 0.9958668194319072, "total_bits": 784740352, "q_proj": { "group_size": { @@ -61386,7 +61386,7 @@ } }, { - "accuracy": 0.9964247212598198, + "accuracy": 0.996466443726891, "total_bits": 797818880, "q_proj": { "group_size": { @@ -61447,7 +61447,7 @@ } }, { - "accuracy": 0.9972922021621152, + "accuracy": 0.9973548870337637, "total_bits": 911749120, "q_proj": { "group_size": { @@ -61499,7 +61499,7 @@ } }, { - "accuracy": 0.998205172780313, + "accuracy": 0.9982165470719337, "total_bits": 942718976, "q_proj": { "group_size": { @@ -61551,7 +61551,7 @@ } }, { - "accuracy": 0.9992867060691902, + "accuracy": 0.9992924422319782, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -61605,7 +61605,7 @@ ], "model.layers.31.mlp": [ { - "accuracy": 0.9527285600963392, + "accuracy": 0.9529176448520861, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -61657,7 +61657,7 @@ } }, { - "accuracy": 0.9540125476686578, + "accuracy": 0.9541927607435929, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -61709,7 +61709,7 @@ } }, { - "accuracy": 0.9610225181830556, + "accuracy": 0.9611732583296926, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -61758,7 +61758,7 @@ } }, { - "accuracy": 0.9629739052370975, + "accuracy": 0.963117517923054, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -61807,7 +61807,7 @@ } }, { - "accuracy": 0.9765667178128895, + "accuracy": 0.9766563192794198, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -61859,7 +61859,7 @@ } }, { - "accuracy": 0.9783722334786465, + "accuracy": 0.9784554070548007, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -61911,7 +61911,7 @@ } }, { - "accuracy": 0.981181684293245, + "accuracy": 0.9812531643792203, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -61960,7 +61960,7 @@ } }, { - "accuracy": 0.9880287357066807, + "accuracy": 0.9880760445406562, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -62003,7 +62003,7 @@ } }, { - "accuracy": 0.9889677198309648, + "accuracy": 0.989011362195015, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -62046,7 +62046,7 @@ } }, { - "accuracy": 0.9880915880203247, + "accuracy": 0.9881385000128495, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -62098,7 +62098,7 @@ } }, { - "accuracy": 0.9895056564556924, + "accuracy": 0.9895456519566084, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -62150,7 +62150,7 @@ } }, { - "accuracy": 0.9939357124661145, + "accuracy": 0.9939591284645232, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -62202,7 +62202,7 @@ } }, { - "accuracy": 0.9947701876885012, + "accuracy": 0.9947904387586995, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -62254,7 +62254,7 @@ } }, { - "accuracy": 0.9967580553340284, + "accuracy": 0.996771689308317, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -62297,7 +62297,7 @@ } }, { - "accuracy": 0.9969041023991609, + "accuracy": 0.996916273118634, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -62346,7 +62346,7 @@ } }, { - "accuracy": 0.9974513561709931, + "accuracy": 0.9974603533352676, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -62392,7 +62392,7 @@ } }, { - "accuracy": 0.9991434378255355, + "accuracy": 0.9991467292175481, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -62434,7 +62434,7 @@ ], "model.layers.32.self_attn": [ { - "accuracy": 0.9695752074843959, + "accuracy": 0.97015380859375, "total_bits": 320757760, "q_proj": { "group_size": { @@ -62498,7 +62498,7 @@ } }, { - "accuracy": 0.9705520397738406, + "accuracy": 0.9709460547095851, "total_bits": 329080832, "q_proj": { "group_size": { @@ -62562,7 +62562,7 @@ } }, { - "accuracy": 0.9745561574634752, + "accuracy": 0.9759863756204906, "total_bits": 336024576, "q_proj": { "group_size": { @@ -62626,7 +62626,7 @@ } }, { - "accuracy": 0.9786667949274966, + "accuracy": 0.9803256690502167, "total_bits": 401557504, "q_proj": { "group_size": { @@ -62690,7 +62690,7 @@ } }, { - "accuracy": 0.9810742814289896, + "accuracy": 0.9827089874367965, "total_bits": 475279360, "q_proj": { "group_size": { @@ -62754,7 +62754,7 @@ } }, { - "accuracy": 0.9827468614829215, + "accuracy": 0.9830549108354669, "total_bits": 475479040, "q_proj": { "group_size": { @@ -62818,7 +62818,7 @@ } }, { - "accuracy": 0.9909223067133051, + "accuracy": 0.9908861235568398, "total_bits": 609759232, "q_proj": { "group_size": { @@ -62870,7 +62870,7 @@ } }, { - "accuracy": 0.9910852979672583, + "accuracy": 0.9911015269003416, "total_bits": 610024448, "q_proj": { "group_size": { @@ -62922,7 +62922,7 @@ } }, { - "accuracy": 0.9916673247751436, + "accuracy": 0.9915319194919184, "total_bits": 615020544, "q_proj": { "group_size": { @@ -62974,7 +62974,7 @@ } }, { - "accuracy": 0.992014251257244, + "accuracy": 0.9919223910883853, "total_bits": 623951872, "q_proj": { "group_size": { @@ -63026,7 +63026,7 @@ } }, { - "accuracy": 0.9913814624673442, + "accuracy": 0.9913428347361716, "total_bits": 626473984, "q_proj": { "group_size": { @@ -63090,7 +63090,7 @@ } }, { - "accuracy": 0.9917120784521103, + "accuracy": 0.9916816785147315, "total_bits": 630355968, "q_proj": { "group_size": { @@ -63154,7 +63154,7 @@ } }, { - "accuracy": 0.9939732947631886, + "accuracy": 0.9939817715632288, "total_bits": 637362176, "q_proj": { "group_size": { @@ -63215,7 +63215,7 @@ } }, { - "accuracy": 0.9945107945486119, + "accuracy": 0.9944828380095331, "total_bits": 646823936, "q_proj": { "group_size": { @@ -63276,7 +63276,7 @@ } }, { - "accuracy": 0.9967429247733793, + "accuracy": 0.9967676554071275, "total_bits": 784740352, "q_proj": { "group_size": { @@ -63337,7 +63337,7 @@ } }, { - "accuracy": 0.9972259668927443, + "accuracy": 0.9972544283066925, "total_bits": 797818880, "q_proj": { "group_size": { @@ -63398,7 +63398,7 @@ } }, { - "accuracy": 0.9974437405012155, + "accuracy": 0.9974569850846341, "total_bits": 911749120, "q_proj": { "group_size": { @@ -63450,7 +63450,7 @@ } }, { - "accuracy": 0.998739921046715, + "accuracy": 0.998743010194678, "total_bits": 942718976, "q_proj": { "group_size": { @@ -63502,7 +63502,7 @@ } }, { - "accuracy": 0.9993312641193992, + "accuracy": 0.9993222984240243, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -63556,7 +63556,7 @@ ], "model.layers.32.mlp": [ { - "accuracy": 0.9523673998682123, + "accuracy": 0.9525714234301919, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -63608,7 +63608,7 @@ } }, { - "accuracy": 0.9536214314009014, + "accuracy": 0.9538206363979139, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -63660,7 +63660,7 @@ } }, { - "accuracy": 0.9605679010090075, + "accuracy": 0.9607362809934115, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -63709,7 +63709,7 @@ } }, { - "accuracy": 0.9625543669650429, + "accuracy": 0.9627109515039545, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -63758,7 +63758,7 @@ } }, { - "accuracy": 0.9762617804502186, + "accuracy": 0.9763601422309875, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -63810,7 +63810,7 @@ } }, { - "accuracy": 0.9781031436041782, + "accuracy": 0.9781965989815561, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -63862,7 +63862,7 @@ } }, { - "accuracy": 0.9809633211085671, + "accuracy": 0.9810395256469124, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -63911,7 +63911,7 @@ } }, { - "accuracy": 0.9878308043668145, + "accuracy": 0.9878796652743691, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -63954,7 +63954,7 @@ } }, { - "accuracy": 0.9888192266225815, + "accuracy": 0.9888659888192227, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -63997,7 +63997,7 @@ } }, { - "accuracy": 0.9879264502148879, + "accuracy": 0.9879766357572455, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -64049,7 +64049,7 @@ } }, { - "accuracy": 0.9893739427390852, + "accuracy": 0.9894197034208398, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -64101,7 +64101,7 @@ } }, { - "accuracy": 0.9938437287744722, + "accuracy": 0.9938674459331914, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -64153,7 +64153,7 @@ } }, { - "accuracy": 0.9946977649080125, + "accuracy": 0.9947221855583944, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -64205,7 +64205,7 @@ } }, { - "accuracy": 0.9966987767501881, + "accuracy": 0.9967131930354395, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -64248,7 +64248,7 @@ } }, { - "accuracy": 0.996853843135269, + "accuracy": 0.9968663689337278, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -64297,7 +64297,7 @@ } }, { - "accuracy": 0.9974177217012957, + "accuracy": 0.9974276262678599, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -64343,7 +64343,7 @@ } }, { - "accuracy": 0.9991035756507987, + "accuracy": 0.9991078602248117, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -64385,7 +64385,7 @@ ], "model.layers.33.self_attn": [ { - "accuracy": 0.9637240045949033, + "accuracy": 0.9646448116553457, "total_bits": 320757760, "q_proj": { "group_size": { @@ -64449,7 +64449,7 @@ } }, { - "accuracy": 0.9660809322407371, + "accuracy": 0.9658031212656122, "total_bits": 329080832, "q_proj": { "group_size": { @@ -64513,7 +64513,7 @@ } }, { - "accuracy": 0.969701421888251, + "accuracy": 0.9700648219961869, "total_bits": 336024576, "q_proj": { "group_size": { @@ -64577,7 +64577,7 @@ } }, { - "accuracy": 0.9765537851735165, + "accuracy": 0.9768458275418532, "total_bits": 401557504, "q_proj": { "group_size": { @@ -64641,7 +64641,7 @@ } }, { - "accuracy": 0.981755898187035, + "accuracy": 0.982065520788494, "total_bits": 475279360, "q_proj": { "group_size": { @@ -64705,7 +64705,7 @@ } }, { - "accuracy": 0.9822340905666351, + "accuracy": 0.9823171449334998, "total_bits": 475479040, "q_proj": { "group_size": { @@ -64769,7 +64769,7 @@ } }, { - "accuracy": 0.9902266507086, + "accuracy": 0.9901187843398044, "total_bits": 609759232, "q_proj": { "group_size": { @@ -64821,7 +64821,7 @@ } }, { - "accuracy": 0.9900844316733511, + "accuracy": 0.990274512454083, "total_bits": 610024448, "q_proj": { "group_size": { @@ -64873,7 +64873,7 @@ } }, { - "accuracy": 0.9911646944911856, + "accuracy": 0.9911272235606846, "total_bits": 615020544, "q_proj": { "group_size": { @@ -64925,7 +64925,7 @@ } }, { - "accuracy": 0.991605741413016, + "accuracy": 0.9915947670999327, "total_bits": 623951872, "q_proj": { "group_size": { @@ -64977,7 +64977,7 @@ } }, { - "accuracy": 0.9908133839306078, + "accuracy": 0.9908775593105116, "total_bits": 626473984, "q_proj": { "group_size": { @@ -65041,7 +65041,7 @@ } }, { - "accuracy": 0.9914343835491883, + "accuracy": 0.9916160679177234, "total_bits": 630355968, "q_proj": { "group_size": { @@ -65105,7 +65105,7 @@ } }, { - "accuracy": 0.9925628586819297, + "accuracy": 0.9927015045755788, "total_bits": 637362176, "q_proj": { "group_size": { @@ -65166,7 +65166,7 @@ } }, { - "accuracy": 0.993209960429292, + "accuracy": 0.9932732017416703, "total_bits": 646823936, "q_proj": { "group_size": { @@ -65227,7 +65227,7 @@ } }, { - "accuracy": 0.9960117767515936, + "accuracy": 0.9959997095559773, "total_bits": 784740352, "q_proj": { "group_size": { @@ -65288,7 +65288,7 @@ } }, { - "accuracy": 0.9966176799253413, + "accuracy": 0.9966030113006893, "total_bits": 797818880, "q_proj": { "group_size": { @@ -65349,7 +65349,7 @@ } }, { - "accuracy": 0.9974503517150879, + "accuracy": 0.9974330744068873, "total_bits": 911749120, "q_proj": { "group_size": { @@ -65401,7 +65401,7 @@ } }, { - "accuracy": 0.9982650134908525, + "accuracy": 0.9982742543675398, "total_bits": 942718976, "q_proj": { "group_size": { @@ -65453,7 +65453,7 @@ } }, { - "accuracy": 0.9993222631808174, + "accuracy": 0.9993285524511808, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -65507,7 +65507,7 @@ ], "model.layers.33.mlp": [ { - "accuracy": 0.9502856856898257, + "accuracy": 0.9504946972194471, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -65559,7 +65559,7 @@ } }, { - "accuracy": 0.951587830719195, + "accuracy": 0.9517873807957298, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -65611,7 +65611,7 @@ } }, { - "accuracy": 0.9584029411014757, + "accuracy": 0.9585737052716707, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -65660,7 +65660,7 @@ } }, { - "accuracy": 0.9603217808823836, + "accuracy": 0.9604815777979399, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -65709,7 +65709,7 @@ } }, { - "accuracy": 0.9710175959687484, + "accuracy": 0.9711489834283528, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -65761,7 +65761,7 @@ } }, { - "accuracy": 0.9764610591687655, + "accuracy": 0.9765641579502508, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -65813,7 +65813,7 @@ } }, { - "accuracy": 0.9791316311610373, + "accuracy": 0.9792215886868929, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -65862,7 +65862,7 @@ } }, { - "accuracy": 0.9833612881208721, + "accuracy": 0.9834349092684294, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -65905,7 +65905,7 @@ } }, { - "accuracy": 0.9865789068372626, + "accuracy": 0.9866377460329157, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -65948,7 +65948,7 @@ } }, { - "accuracy": 0.9808955678814336, + "accuracy": 0.9809822989137549, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -66000,7 +66000,7 @@ } }, { - "accuracy": 0.9870665755711103, + "accuracy": 0.9871247206863604, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -66052,7 +66052,7 @@ } }, { - "accuracy": 0.9859351390286496, + "accuracy": 0.9860014209621831, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -66104,7 +66104,7 @@ } }, { - "accuracy": 0.9918142431660703, + "accuracy": 0.9918498585098668, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -66156,7 +66156,7 @@ } }, { - "accuracy": 0.9895097793717134, + "accuracy": 0.9895634227677396, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -66199,7 +66199,7 @@ } }, { - "accuracy": 0.9895600323614321, + "accuracy": 0.9896113198054465, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -66248,7 +66248,7 @@ } }, { - "accuracy": 0.9897197351643914, + "accuracy": 0.9897701669680444, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -66294,7 +66294,7 @@ } }, { - "accuracy": 0.9944728040381482, + "accuracy": 0.9945017401325075, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -66336,7 +66336,7 @@ ], "model.layers.34.self_attn": [ { - "accuracy": 0.9591638728191978, + "accuracy": 0.9606420805579737, "total_bits": 320757760, "q_proj": { "group_size": { @@ -66400,7 +66400,7 @@ } }, { - "accuracy": 0.9606904669811851, + "accuracy": 0.962296790198276, "total_bits": 329080832, "q_proj": { "group_size": { @@ -66464,7 +66464,7 @@ } }, { - "accuracy": 0.9652948222662273, + "accuracy": 0.9693868317102131, "total_bits": 336024576, "q_proj": { "group_size": { @@ -66528,7 +66528,7 @@ } }, { - "accuracy": 0.9713765853329709, + "accuracy": 0.9762046227329656, "total_bits": 401557504, "q_proj": { "group_size": { @@ -66592,7 +66592,7 @@ } }, { - "accuracy": 0.9786824301669472, + "accuracy": 0.9826631310739016, "total_bits": 475279360, "q_proj": { "group_size": { @@ -66656,7 +66656,7 @@ } }, { - "accuracy": 0.9797062152310422, + "accuracy": 0.9812220617344505, "total_bits": 475479040, "q_proj": { "group_size": { @@ -66720,7 +66720,7 @@ } }, { - "accuracy": 0.9895731900867663, + "accuracy": 0.9895154175005461, "total_bits": 609759232, "q_proj": { "group_size": { @@ -66772,7 +66772,7 @@ } }, { - "accuracy": 0.9892431544630151, + "accuracy": 0.9899336554502186, "total_bits": 610024448, "q_proj": { "group_size": { @@ -66824,7 +66824,7 @@ } }, { - "accuracy": 0.9896507569049534, + "accuracy": 0.9894614862768274, "total_bits": 615020544, "q_proj": { "group_size": { @@ -66876,7 +66876,7 @@ } }, { - "accuracy": 0.9901015836941568, + "accuracy": 0.9899362404095499, "total_bits": 623951872, "q_proj": { "group_size": { @@ -66928,7 +66928,7 @@ } }, { - "accuracy": 0.9892259475431944, + "accuracy": 0.9898488333350733, "total_bits": 626473984, "q_proj": { "group_size": { @@ -66992,7 +66992,7 @@ } }, { - "accuracy": 0.9904273246463976, + "accuracy": 0.9911221987322757, "total_bits": 630355968, "q_proj": { "group_size": { @@ -67056,7 +67056,7 @@ } }, { - "accuracy": 0.9926021153989592, + "accuracy": 0.9926276528521588, "total_bits": 637362176, "q_proj": { "group_size": { @@ -67117,7 +67117,7 @@ } }, { - "accuracy": 0.9931527768310747, + "accuracy": 0.993188708236343, "total_bits": 646823936, "q_proj": { "group_size": { @@ -67178,7 +67178,7 @@ } }, { - "accuracy": 0.9959588835113927, + "accuracy": 0.9959735337056612, "total_bits": 784740352, "q_proj": { "group_size": { @@ -67239,7 +67239,7 @@ } }, { - "accuracy": 0.9965371030725931, + "accuracy": 0.9965946615526551, "total_bits": 797818880, "q_proj": { "group_size": { @@ -67300,7 +67300,7 @@ } }, { - "accuracy": 0.99727407235064, + "accuracy": 0.9972991486521143, "total_bits": 911749120, "q_proj": { "group_size": { @@ -67352,7 +67352,7 @@ } }, { - "accuracy": 0.9982170146939001, + "accuracy": 0.9982005106775385, "total_bits": 942718976, "q_proj": { "group_size": { @@ -67404,7 +67404,7 @@ } }, { - "accuracy": 0.9992509187248192, + "accuracy": 0.9992745639266152, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -67458,7 +67458,7 @@ ], "model.layers.34.mlp": [ { - "accuracy": 0.9506695490134389, + "accuracy": 0.9508890729201467, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -67510,7 +67510,7 @@ } }, { - "accuracy": 0.9520909221548783, + "accuracy": 0.9523131533672935, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -67562,7 +67562,7 @@ } }, { - "accuracy": 0.9597480767651608, + "accuracy": 0.9599199859719527, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -67611,7 +67611,7 @@ } }, { - "accuracy": 0.9618483719072843, + "accuracy": 0.9620102957675332, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -67660,7 +67660,7 @@ } }, { - "accuracy": 0.9757344660006071, + "accuracy": 0.9758406783405104, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -67712,7 +67712,7 @@ } }, { - "accuracy": 0.9775964696156351, + "accuracy": 0.9776934667637474, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -67764,7 +67764,7 @@ } }, { - "accuracy": 0.9805634194298795, + "accuracy": 0.9806451075955441, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -67813,7 +67813,7 @@ } }, { - "accuracy": 0.9875921983467905, + "accuracy": 0.9876479137884943, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -67856,7 +67856,7 @@ } }, { - "accuracy": 0.9885425050007669, + "accuracy": 0.9885951539403514, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -67899,7 +67899,7 @@ } }, { - "accuracy": 0.9876685001348194, + "accuracy": 0.9877244579164606, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -67951,7 +67951,7 @@ } }, { - "accuracy": 0.9891327415641985, + "accuracy": 0.9891789186941949, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -68003,7 +68003,7 @@ } }, { - "accuracy": 0.9937183351108902, + "accuracy": 0.9937453928746676, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -68055,7 +68055,7 @@ } }, { - "accuracy": 0.9945810252898618, + "accuracy": 0.9946058094501495, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -68107,7 +68107,7 @@ } }, { - "accuracy": 0.9966290099056143, + "accuracy": 0.9966448184690977, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -68150,7 +68150,7 @@ } }, { - "accuracy": 0.9967804347213945, + "accuracy": 0.9967934873543287, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -68199,7 +68199,7 @@ } }, { - "accuracy": 0.997366838157177, + "accuracy": 0.9973767970345522, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -68245,7 +68245,7 @@ } }, { - "accuracy": 0.9990676545390957, + "accuracy": 0.9990718232369736, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -68287,7 +68287,7 @@ ], "model.layers.35.self_attn": [ { - "accuracy": 0.9622690301192434, + "accuracy": 0.9639981043966193, "total_bits": 320757760, "q_proj": { "group_size": { @@ -68351,7 +68351,7 @@ } }, { - "accuracy": 0.9643836648840654, + "accuracy": 0.9654189222737363, "total_bits": 329080832, "q_proj": { "group_size": { @@ -68415,7 +68415,7 @@ } }, { - "accuracy": 0.9696788034941021, + "accuracy": 0.9701757274175945, "total_bits": 336024576, "q_proj": { "group_size": { @@ -68479,7 +68479,7 @@ } }, { - "accuracy": 0.9764469861984253, + "accuracy": 0.9770782417372653, "total_bits": 401557504, "q_proj": { "group_size": { @@ -68543,7 +68543,7 @@ } }, { - "accuracy": 0.9815615917506971, + "accuracy": 0.9818405116859236, "total_bits": 475279360, "q_proj": { "group_size": { @@ -68607,7 +68607,7 @@ } }, { - "accuracy": 0.9817503314269217, + "accuracy": 0.9822609832412318, "total_bits": 475479040, "q_proj": { "group_size": { @@ -68671,7 +68671,7 @@ } }, { - "accuracy": 0.9898328663487184, + "accuracy": 0.989889985636661, "total_bits": 609759232, "q_proj": { "group_size": { @@ -68723,7 +68723,7 @@ } }, { - "accuracy": 0.9898613448205748, + "accuracy": 0.9900501597868768, "total_bits": 610024448, "q_proj": { "group_size": { @@ -68775,7 +68775,7 @@ } }, { - "accuracy": 0.9906993649507824, + "accuracy": 0.9907666084013487, "total_bits": 615020544, "q_proj": { "group_size": { @@ -68827,7 +68827,7 @@ } }, { - "accuracy": 0.991099236042876, + "accuracy": 0.9912050264446359, "total_bits": 623951872, "q_proj": { "group_size": { @@ -68879,7 +68879,7 @@ } }, { - "accuracy": 0.990699116336672, + "accuracy": 0.9906368498739443, "total_bits": 626473984, "q_proj": { "group_size": { @@ -68943,7 +68943,7 @@ } }, { - "accuracy": 0.9912638013300142, + "accuracy": 0.9914031977716246, "total_bits": 630355968, "q_proj": { "group_size": { @@ -69007,7 +69007,7 @@ } }, { - "accuracy": 0.9925472689302344, + "accuracy": 0.9925993390773472, "total_bits": 637362176, "q_proj": { "group_size": { @@ -69068,7 +69068,7 @@ } }, { - "accuracy": 0.9931446407970629, + "accuracy": 0.9932105564757397, "total_bits": 646823936, "q_proj": { "group_size": { @@ -69129,7 +69129,7 @@ } }, { - "accuracy": 0.9959367682275019, + "accuracy": 0.9959353220305944, "total_bits": 784740352, "q_proj": { "group_size": { @@ -69190,7 +69190,7 @@ } }, { - "accuracy": 0.9965385916201692, + "accuracy": 0.9965898829855417, "total_bits": 797818880, "q_proj": { "group_size": { @@ -69251,7 +69251,7 @@ } }, { - "accuracy": 0.9972952659192839, + "accuracy": 0.9972614477339544, "total_bits": 911749120, "q_proj": { "group_size": { @@ -69303,7 +69303,7 @@ } }, { - "accuracy": 0.9982508507214094, + "accuracy": 0.998261813858622, "total_bits": 942718976, "q_proj": { "group_size": { @@ -69355,7 +69355,7 @@ } }, { - "accuracy": 0.9992756054883725, + "accuracy": 0.9992750693406713, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -69409,7 +69409,7 @@ ], "model.layers.35.mlp": [ { - "accuracy": 0.9518215562167921, + "accuracy": 0.9520704244312487, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -69461,7 +69461,7 @@ } }, { - "accuracy": 0.9532042961371572, + "accuracy": 0.9534336140281275, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -69513,7 +69513,7 @@ } }, { - "accuracy": 0.9604451374003762, + "accuracy": 0.9606398061702126, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -69562,7 +69562,7 @@ } }, { - "accuracy": 0.9624165923971879, + "accuracy": 0.9625988383042186, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -69611,7 +69611,7 @@ } }, { - "accuracy": 0.9763268075491253, + "accuracy": 0.9764426852527418, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -69663,7 +69663,7 @@ } }, { - "accuracy": 0.9781390240317897, + "accuracy": 0.978248080140666, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -69715,7 +69715,7 @@ } }, { - "accuracy": 0.9809238298943168, + "accuracy": 0.981013906629462, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -69764,7 +69764,7 @@ } }, { - "accuracy": 0.987910826739512, + "accuracy": 0.987968595404374, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -69807,7 +69807,7 @@ } }, { - "accuracy": 0.9888384451991633, + "accuracy": 0.9888933410770014, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -69850,7 +69850,7 @@ } }, { - "accuracy": 0.987971546618562, + "accuracy": 0.9880280831926748, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -69902,7 +69902,7 @@ } }, { - "accuracy": 0.9893946600587744, + "accuracy": 0.9894478642626813, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -69954,7 +69954,7 @@ } }, { - "accuracy": 0.9938710629940033, + "accuracy": 0.9938996877325209, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -70006,7 +70006,7 @@ } }, { - "accuracy": 0.9947121888399124, + "accuracy": 0.9947376102209091, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -70058,7 +70058,7 @@ } }, { - "accuracy": 0.9967148809840805, + "accuracy": 0.9967301299697474, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -70101,7 +70101,7 @@ } }, { - "accuracy": 0.9968561353652101, + "accuracy": 0.9968703338190129, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -70150,7 +70150,7 @@ } }, { - "accuracy": 0.9974012574867198, + "accuracy": 0.9974118516241249, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -70196,7 +70196,7 @@ } }, { - "accuracy": 0.9990812494958702, + "accuracy": 0.9990868233144283, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -70238,7 +70238,7 @@ ], "model.layers.36.self_attn": [ { - "accuracy": 0.9763172676688746, + "accuracy": 0.9764744953105324, "total_bits": 320757760, "q_proj": { "group_size": { @@ -70302,7 +70302,7 @@ } }, { - "accuracy": 0.9773656506287424, + "accuracy": 0.9774881742502514, "total_bits": 329080832, "q_proj": { "group_size": { @@ -70366,7 +70366,7 @@ } }, { - "accuracy": 0.9803655398519415, + "accuracy": 0.9804825704348715, "total_bits": 336024576, "q_proj": { "group_size": { @@ -70430,7 +70430,7 @@ } }, { - "accuracy": 0.9850014071715506, + "accuracy": 0.9850585539090005, "total_bits": 401557504, "q_proj": { "group_size": { @@ -70494,7 +70494,7 @@ } }, { - "accuracy": 0.9873851827884975, + "accuracy": 0.9875227981492093, "total_bits": 475279360, "q_proj": { "group_size": { @@ -70558,7 +70558,7 @@ } }, { - "accuracy": 0.987860055346238, + "accuracy": 0.987871882162596, "total_bits": 475479040, "q_proj": { "group_size": { @@ -70622,7 +70622,7 @@ } }, { - "accuracy": 0.9911970903998927, + "accuracy": 0.9912476288644891, "total_bits": 609759232, "q_proj": { "group_size": { @@ -70674,7 +70674,7 @@ } }, { - "accuracy": 0.9917356348351428, + "accuracy": 0.9916545262462214, "total_bits": 610024448, "q_proj": { "group_size": { @@ -70726,7 +70726,7 @@ } }, { - "accuracy": 0.9922425519478949, + "accuracy": 0.9923341870307922, "total_bits": 615020544, "q_proj": { "group_size": { @@ -70778,7 +70778,7 @@ } }, { - "accuracy": 0.9924730956554413, + "accuracy": 0.9925917237997055, "total_bits": 623951872, "q_proj": { "group_size": { @@ -70830,7 +70830,7 @@ } }, { - "accuracy": 0.9935210186399912, + "accuracy": 0.9935919234627172, "total_bits": 626473984, "q_proj": { "group_size": { @@ -70894,7 +70894,7 @@ } }, { - "accuracy": 0.9939764569464483, + "accuracy": 0.9940329970497834, "total_bits": 630355968, "q_proj": { "group_size": { @@ -70958,7 +70958,7 @@ } }, { - "accuracy": 0.9942762330174446, + "accuracy": 0.9943072964486322, "total_bits": 637362176, "q_proj": { "group_size": { @@ -71019,7 +71019,7 @@ } }, { - "accuracy": 0.9947153769041363, + "accuracy": 0.9947851739431682, "total_bits": 646823936, "q_proj": { "group_size": { @@ -71080,7 +71080,7 @@ } }, { - "accuracy": 0.9968579082112563, + "accuracy": 0.9968664336361384, "total_bits": 784740352, "q_proj": { "group_size": { @@ -71141,7 +71141,7 @@ } }, { - "accuracy": 0.9973393027719698, + "accuracy": 0.9973529528238272, "total_bits": 797818880, "q_proj": { "group_size": { @@ -71202,7 +71202,7 @@ } }, { - "accuracy": 0.99748773715998, + "accuracy": 0.9974901142873263, "total_bits": 911749120, "q_proj": { "group_size": { @@ -71254,7 +71254,7 @@ } }, { - "accuracy": 0.9987603112270957, + "accuracy": 0.9987713543600157, "total_bits": 942718976, "q_proj": { "group_size": { @@ -71306,7 +71306,7 @@ } }, { - "accuracy": 0.9993458556893625, + "accuracy": 0.999346732505058, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -71360,7 +71360,7 @@ ], "model.layers.36.mlp": [ { - "accuracy": 0.9536179241381193, + "accuracy": 0.9538569764087075, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -71412,7 +71412,7 @@ } }, { - "accuracy": 0.9548384202154059, + "accuracy": 0.9550655327345196, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -71464,7 +71464,7 @@ } }, { - "accuracy": 0.961607321312553, + "accuracy": 0.9617913522218403, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -71513,7 +71513,7 @@ } }, { - "accuracy": 0.9635515150270964, + "accuracy": 0.963728810611524, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -71562,7 +71562,7 @@ } }, { - "accuracy": 0.9769380422014939, + "accuracy": 0.9770551810139104, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -71614,7 +71614,7 @@ } }, { - "accuracy": 0.9787183780419199, + "accuracy": 0.9788247754699305, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -71666,7 +71666,7 @@ } }, { - "accuracy": 0.9814680422607222, + "accuracy": 0.9815609596277538, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -71715,7 +71715,7 @@ } }, { - "accuracy": 0.9881506428906792, + "accuracy": 0.9882144284875769, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -71758,7 +71758,7 @@ } }, { - "accuracy": 0.9891077899619153, + "accuracy": 0.9891658833152369, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -71801,7 +71801,7 @@ } }, { - "accuracy": 0.9882712128915285, + "accuracy": 0.9883307061697307, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -71853,7 +71853,7 @@ } }, { - "accuracy": 0.989659887395407, + "accuracy": 0.9897117528476214, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -71905,7 +71905,7 @@ } }, { - "accuracy": 0.9940158222851, + "accuracy": 0.9940470998224459, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -71957,7 +71957,7 @@ } }, { - "accuracy": 0.994839286725772, + "accuracy": 0.9948646469335807, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -72009,7 +72009,7 @@ } }, { - "accuracy": 0.9967837306229692, + "accuracy": 0.996800114057566, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -72052,7 +72052,7 @@ } }, { - "accuracy": 0.9969312807447032, + "accuracy": 0.9969481069006418, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -72101,7 +72101,7 @@ } }, { - "accuracy": 0.9974758507389772, + "accuracy": 0.9974890710099747, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -72147,7 +72147,7 @@ } }, { - "accuracy": 0.9990848202846552, + "accuracy": 0.9990905952688894, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -72189,7 +72189,7 @@ ], "model.layers.37.self_attn": [ { - "accuracy": 0.9734072402903908, + "accuracy": 0.9736684623517489, "total_bits": 320757760, "q_proj": { "group_size": { @@ -72253,7 +72253,7 @@ } }, { - "accuracy": 0.9747191745983926, + "accuracy": 0.974877351208737, "total_bits": 329080832, "q_proj": { "group_size": { @@ -72317,7 +72317,7 @@ } }, { - "accuracy": 0.9770153039380124, + "accuracy": 0.9771743573640522, "total_bits": 336024576, "q_proj": { "group_size": { @@ -72381,7 +72381,7 @@ } }, { - "accuracy": 0.98238762899449, + "accuracy": 0.9825231150576943, "total_bits": 401557504, "q_proj": { "group_size": { @@ -72445,7 +72445,7 @@ } }, { - "accuracy": 0.9857920580788663, + "accuracy": 0.9859860115929654, "total_bits": 475279360, "q_proj": { "group_size": { @@ -72509,7 +72509,7 @@ } }, { - "accuracy": 0.9861583646975065, + "accuracy": 0.986345543673164, "total_bits": 475479040, "q_proj": { "group_size": { @@ -72573,7 +72573,7 @@ } }, { - "accuracy": 0.9913545039139295, + "accuracy": 0.9914663786950865, "total_bits": 609759232, "q_proj": { "group_size": { @@ -72625,7 +72625,7 @@ } }, { - "accuracy": 0.9918378928774282, + "accuracy": 0.9919301569461823, "total_bits": 610024448, "q_proj": { "group_size": { @@ -72677,7 +72677,7 @@ } }, { - "accuracy": 0.9923400973018847, + "accuracy": 0.9922733573537124, "total_bits": 615020544, "q_proj": { "group_size": { @@ -72729,7 +72729,7 @@ } }, { - "accuracy": 0.9926131626493052, + "accuracy": 0.9926130826535978, "total_bits": 623951872, "q_proj": { "group_size": { @@ -72781,7 +72781,7 @@ } }, { - "accuracy": 0.992264164905799, + "accuracy": 0.9922303345642591, "total_bits": 626473984, "q_proj": { "group_size": { @@ -72845,7 +72845,7 @@ } }, { - "accuracy": 0.9926986717864087, + "accuracy": 0.9926969087437579, "total_bits": 630355968, "q_proj": { "group_size": { @@ -72909,7 +72909,7 @@ } }, { - "accuracy": 0.993861220384899, + "accuracy": 0.9939636646916992, "total_bits": 637362176, "q_proj": { "group_size": { @@ -72970,7 +72970,7 @@ } }, { - "accuracy": 0.9943830178756463, + "accuracy": 0.9943796392334135, "total_bits": 646823936, "q_proj": { "group_size": { @@ -73031,7 +73031,7 @@ } }, { - "accuracy": 0.9966215063866816, + "accuracy": 0.9966130593889638, "total_bits": 784740352, "q_proj": { "group_size": { @@ -73092,7 +73092,7 @@ } }, { - "accuracy": 0.9971451696596647, + "accuracy": 0.997181199491024, "total_bits": 797818880, "q_proj": { "group_size": { @@ -73153,7 +73153,7 @@ } }, { - "accuracy": 0.997537368026219, + "accuracy": 0.9975239685491512, "total_bits": 911749120, "q_proj": { "group_size": { @@ -73205,7 +73205,7 @@ } }, { - "accuracy": 0.9986399111190909, + "accuracy": 0.9986461765672031, "total_bits": 942718976, "q_proj": { "group_size": { @@ -73257,7 +73257,7 @@ } }, { - "accuracy": 0.9993058462676249, + "accuracy": 0.9993142507676231, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -73311,7 +73311,7 @@ ], "model.layers.37.mlp": [ { - "accuracy": 0.9545840150431583, + "accuracy": 0.9548400797341999, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -73363,7 +73363,7 @@ } }, { - "accuracy": 0.9557558925528276, + "accuracy": 0.9560113988424602, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -73415,7 +73415,7 @@ } }, { - "accuracy": 0.9623626533307528, + "accuracy": 0.9625703410098427, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -73464,7 +73464,7 @@ } }, { - "accuracy": 0.9642636807341325, + "accuracy": 0.9644562068738436, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -73513,7 +73513,7 @@ } }, { - "accuracy": 0.977420155939303, + "accuracy": 0.9775516014350089, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -73565,7 +73565,7 @@ } }, { - "accuracy": 0.979162700866398, + "accuracy": 0.9792865561811548, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -73617,7 +73617,7 @@ } }, { - "accuracy": 0.9818508374063593, + "accuracy": 0.9819533479841132, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -73666,7 +73666,7 @@ } }, { - "accuracy": 0.9884332457655355, + "accuracy": 0.9885043305786032, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -73709,7 +73709,7 @@ } }, { - "accuracy": 0.9893741733149478, + "accuracy": 0.9894366593737352, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -73752,7 +73752,7 @@ } }, { - "accuracy": 0.9885255278725373, + "accuracy": 0.9885927438735962, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -73804,7 +73804,7 @@ } }, { - "accuracy": 0.9899032970792369, + "accuracy": 0.9899595019064451, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -73856,7 +73856,7 @@ } }, { - "accuracy": 0.9941524314252954, + "accuracy": 0.9941861041282353, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -73908,7 +73908,7 @@ } }, { - "accuracy": 0.9949675418044391, + "accuracy": 0.9949968516042358, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -73960,7 +73960,7 @@ } }, { - "accuracy": 0.9968677198416308, + "accuracy": 0.9968855796675933, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -74003,7 +74003,7 @@ } }, { - "accuracy": 0.997004408193262, + "accuracy": 0.9970214449261364, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -74052,7 +74052,7 @@ } }, { - "accuracy": 0.9975218843472632, + "accuracy": 0.9975359353579973, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -74098,7 +74098,7 @@ } }, { - "accuracy": 0.9991112136723179, + "accuracy": 0.9991180783039645, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -74140,7 +74140,7 @@ ], "model.layers.38.self_attn": [ { - "accuracy": 0.9732631790010553, + "accuracy": 0.9731590465495461, "total_bits": 320757760, "q_proj": { "group_size": { @@ -74204,7 +74204,7 @@ } }, { - "accuracy": 0.9747288038856105, + "accuracy": 0.9748598148948268, "total_bits": 329080832, "q_proj": { "group_size": { @@ -74268,7 +74268,7 @@ } }, { - "accuracy": 0.977204851414028, + "accuracy": 0.9773606407014948, "total_bits": 336024576, "q_proj": { "group_size": { @@ -74332,7 +74332,7 @@ } }, { - "accuracy": 0.9827529568421213, + "accuracy": 0.9828555238874335, "total_bits": 401557504, "q_proj": { "group_size": { @@ -74396,7 +74396,7 @@ } }, { - "accuracy": 0.9856219605395669, + "accuracy": 0.9855974865587134, "total_bits": 475279360, "q_proj": { "group_size": { @@ -74460,7 +74460,7 @@ } }, { - "accuracy": 0.9858519921177312, + "accuracy": 0.9858034786425138, "total_bits": 475479040, "q_proj": { "group_size": { @@ -74524,7 +74524,7 @@ } }, { - "accuracy": 0.9908913342576278, + "accuracy": 0.990911057905147, "total_bits": 609759232, "q_proj": { "group_size": { @@ -74576,7 +74576,7 @@ } }, { - "accuracy": 0.9912266950858267, + "accuracy": 0.9912193088155044, "total_bits": 610024448, "q_proj": { "group_size": { @@ -74628,7 +74628,7 @@ } }, { - "accuracy": 0.9917801422508139, + "accuracy": 0.9919070099529467, "total_bits": 615020544, "q_proj": { "group_size": { @@ -74680,7 +74680,7 @@ } }, { - "accuracy": 0.9920531660318375, + "accuracy": 0.9921656196054659, "total_bits": 623951872, "q_proj": { "group_size": { @@ -74732,7 +74732,7 @@ } }, { - "accuracy": 0.9916981168483433, + "accuracy": 0.9917008218012358, "total_bits": 626473984, "q_proj": { "group_size": { @@ -74796,7 +74796,7 @@ } }, { - "accuracy": 0.9921769829172837, + "accuracy": 0.9923484458735115, "total_bits": 630355968, "q_proj": { "group_size": { @@ -74860,7 +74860,7 @@ } }, { - "accuracy": 0.9935987736833723, + "accuracy": 0.9936834471790414, "total_bits": 637362176, "q_proj": { "group_size": { @@ -74921,7 +74921,7 @@ } }, { - "accuracy": 0.994104546543799, + "accuracy": 0.9942606554219597, "total_bits": 646823936, "q_proj": { "group_size": { @@ -74982,7 +74982,7 @@ } }, { - "accuracy": 0.996488140209725, + "accuracy": 0.9965041279792786, "total_bits": 784740352, "q_proj": { "group_size": { @@ -75043,7 +75043,7 @@ } }, { - "accuracy": 0.9970581398198479, + "accuracy": 0.997077508780517, "total_bits": 797818880, "q_proj": { "group_size": { @@ -75104,7 +75104,7 @@ } }, { - "accuracy": 0.9973150844636717, + "accuracy": 0.9973022582891741, "total_bits": 911749120, "q_proj": { "group_size": { @@ -75156,7 +75156,7 @@ } }, { - "accuracy": 0.9986201291413683, + "accuracy": 0.9986288448501575, "total_bits": 942718976, "q_proj": { "group_size": { @@ -75208,7 +75208,7 @@ } }, { - "accuracy": 0.9993016353661293, + "accuracy": 0.9992993620567416, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -75262,7 +75262,7 @@ ], "model.layers.38.mlp": [ { - "accuracy": 0.9550319470857319, + "accuracy": 0.9553007295257167, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -75314,7 +75314,7 @@ } }, { - "accuracy": 0.9562157141534906, + "accuracy": 0.9564734760083651, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -75366,7 +75366,7 @@ } }, { - "accuracy": 0.9627514073723241, + "accuracy": 0.9629718912275214, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -75415,7 +75415,7 @@ } }, { - "accuracy": 0.9646004407029403, + "accuracy": 0.9648088781457198, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -75464,7 +75464,7 @@ } }, { - "accuracy": 0.9776902481129295, + "accuracy": 0.9778151998394414, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -75516,7 +75516,7 @@ } }, { - "accuracy": 0.9794065183714816, + "accuracy": 0.9795242140167638, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -75568,7 +75568,7 @@ } }, { - "accuracy": 0.982025345689372, + "accuracy": 0.9821275739293349, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -75617,7 +75617,7 @@ } }, { - "accuracy": 0.9885694047338084, + "accuracy": 0.988637379125545, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -75660,7 +75660,7 @@ } }, { - "accuracy": 0.9894597624477587, + "accuracy": 0.9895219151910982, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -75703,7 +75703,7 @@ } }, { - "accuracy": 0.9886643282676998, + "accuracy": 0.9887267131554452, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -75755,7 +75755,7 @@ } }, { - "accuracy": 0.9900016580757341, + "accuracy": 0.9900606988291991, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -75807,7 +75807,7 @@ } }, { - "accuracy": 0.9942201003432274, + "accuracy": 0.9942541153807389, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -75859,7 +75859,7 @@ } }, { - "accuracy": 0.9950096113117117, + "accuracy": 0.9950334394448682, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -75911,7 +75911,7 @@ } }, { - "accuracy": 0.9968955248202148, + "accuracy": 0.9969114590632288, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -75954,7 +75954,7 @@ } }, { - "accuracy": 0.99705044631111, + "accuracy": 0.9970662989898732, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -76003,7 +76003,7 @@ } }, { - "accuracy": 0.9975603485577985, + "accuracy": 0.9975738905762371, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -76049,7 +76049,7 @@ } }, { - "accuracy": 0.9991625364085561, + "accuracy": 0.9991682192408725, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -76091,7 +76091,7 @@ ], "model.layers.39.self_attn": [ { - "accuracy": 0.9742820043312875, + "accuracy": 0.9745715574214333, "total_bits": 320757760, "q_proj": { "group_size": { @@ -76155,7 +76155,7 @@ } }, { - "accuracy": 0.9755710739838449, + "accuracy": 0.9756983502915031, "total_bits": 329080832, "q_proj": { "group_size": { @@ -76219,7 +76219,7 @@ } }, { - "accuracy": 0.9775258835993315, + "accuracy": 0.9778741190308019, "total_bits": 336024576, "q_proj": { "group_size": { @@ -76283,7 +76283,7 @@ } }, { - "accuracy": 0.9827893548890164, + "accuracy": 0.983182574573316, "total_bits": 401557504, "q_proj": { "group_size": { @@ -76347,7 +76347,7 @@ } }, { - "accuracy": 0.9859862625598907, + "accuracy": 0.9861479903522291, "total_bits": 475279360, "q_proj": { "group_size": { @@ -76411,7 +76411,7 @@ } }, { - "accuracy": 0.9858777209332115, + "accuracy": 0.986138757906462, "total_bits": 475479040, "q_proj": { "group_size": { @@ -76475,7 +76475,7 @@ } }, { - "accuracy": 0.9909130044673619, + "accuracy": 0.9910844878146523, "total_bits": 609759232, "q_proj": { "group_size": { @@ -76527,7 +76527,7 @@ } }, { - "accuracy": 0.9913590511209086, + "accuracy": 0.9913240459404493, "total_bits": 610024448, "q_proj": { "group_size": { @@ -76579,7 +76579,7 @@ } }, { - "accuracy": 0.9919900831423307, + "accuracy": 0.9920112380855962, "total_bits": 615020544, "q_proj": { "group_size": { @@ -76631,7 +76631,7 @@ } }, { - "accuracy": 0.9923613353779441, + "accuracy": 0.9922812784972944, "total_bits": 623951872, "q_proj": { "group_size": { @@ -76683,7 +76683,7 @@ } }, { - "accuracy": 0.9919920391158054, + "accuracy": 0.9918967132505617, "total_bits": 626473984, "q_proj": { "group_size": { @@ -76747,7 +76747,7 @@ } }, { - "accuracy": 0.992446885297173, + "accuracy": 0.9925229737633153, "total_bits": 630355968, "q_proj": { "group_size": { @@ -76811,7 +76811,7 @@ } }, { - "accuracy": 0.9936587053694224, + "accuracy": 0.9936834342385593, "total_bits": 637362176, "q_proj": { "group_size": { @@ -76872,7 +76872,7 @@ } }, { - "accuracy": 0.9941496237328178, + "accuracy": 0.9942439190651241, "total_bits": 646823936, "q_proj": { "group_size": { @@ -76933,7 +76933,7 @@ } }, { - "accuracy": 0.996439864760951, + "accuracy": 0.9965091402593412, "total_bits": 784740352, "q_proj": { "group_size": { @@ -76994,7 +76994,7 @@ } }, { - "accuracy": 0.9970563963839882, + "accuracy": 0.9970410768138734, "total_bits": 797818880, "q_proj": { "group_size": { @@ -77055,7 +77055,7 @@ } }, { - "accuracy": 0.9971718982254204, + "accuracy": 0.9972635133093909, "total_bits": 911749120, "q_proj": { "group_size": { @@ -77107,7 +77107,7 @@ } }, { - "accuracy": 0.9986238863907362, + "accuracy": 0.9986196827927702, "total_bits": 942718976, "q_proj": { "group_size": { @@ -77159,7 +77159,7 @@ } }, { - "accuracy": 0.9992456891524949, + "accuracy": 0.9992434080494078, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -77213,7 +77213,7 @@ ], "model.layers.39.mlp": [ { - "accuracy": 0.9554032432405573, + "accuracy": 0.9556911085781298, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -77265,7 +77265,7 @@ } }, { - "accuracy": 0.9565606776036715, + "accuracy": 0.9568321422526711, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -77317,7 +77317,7 @@ } }, { - "accuracy": 0.962859360795272, + "accuracy": 0.9630867211442244, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -77366,7 +77366,7 @@ } }, { - "accuracy": 0.9646706298777932, + "accuracy": 0.964882420866113, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -77415,7 +77415,7 @@ } }, { - "accuracy": 0.9777787484620747, + "accuracy": 0.9779269852136311, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -77467,7 +77467,7 @@ } }, { - "accuracy": 0.979503145343379, + "accuracy": 0.9796300232410431, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -77519,7 +77519,7 @@ } }, { - "accuracy": 0.982074459916667, + "accuracy": 0.9821811741904208, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -77568,7 +77568,7 @@ } }, { - "accuracy": 0.9886320703908017, + "accuracy": 0.9887039755520067, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -77611,7 +77611,7 @@ } }, { - "accuracy": 0.9895499364325875, + "accuracy": 0.9896187303881896, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -77654,7 +77654,7 @@ } }, { - "accuracy": 0.9887079400451559, + "accuracy": 0.9887785927245492, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -77706,7 +77706,7 @@ } }, { - "accuracy": 0.9900601482705066, + "accuracy": 0.990122308072291, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -77758,7 +77758,7 @@ } }, { - "accuracy": 0.99424930073713, + "accuracy": 0.9942836028180624, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -77810,7 +77810,7 @@ } }, { - "accuracy": 0.9950447706015486, + "accuracy": 0.9950765359558558, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -77862,7 +77862,7 @@ } }, { - "accuracy": 0.9969233574443742, + "accuracy": 0.9969421083989897, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -77905,7 +77905,7 @@ } }, { - "accuracy": 0.9970519774054226, + "accuracy": 0.9970707893371582, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -77954,7 +77954,7 @@ } }, { - "accuracy": 0.9975422142367614, + "accuracy": 0.9975570971636396, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -78000,7 +78000,7 @@ } }, { - "accuracy": 0.9991344356615293, + "accuracy": 0.9991404248499557, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -78042,7 +78042,7 @@ ], "model.layers.40.self_attn": [ { - "accuracy": 0.9797092754589883, + "accuracy": 0.9793066084384918, "total_bits": 320757760, "q_proj": { "group_size": { @@ -78106,7 +78106,7 @@ } }, { - "accuracy": 0.9804392833458749, + "accuracy": 0.980920390078896, "total_bits": 329080832, "q_proj": { "group_size": { @@ -78170,7 +78170,7 @@ } }, { - "accuracy": 0.9824234560916298, + "accuracy": 0.9831697015385878, "total_bits": 336024576, "q_proj": { "group_size": { @@ -78234,7 +78234,7 @@ } }, { - "accuracy": 0.985767104123768, + "accuracy": 0.9866717802850824, "total_bits": 401557504, "q_proj": { "group_size": { @@ -78298,7 +78298,7 @@ } }, { - "accuracy": 0.9881533258839658, + "accuracy": 0.989391291612073, "total_bits": 475279360, "q_proj": { "group_size": { @@ -78362,7 +78362,7 @@ } }, { - "accuracy": 0.9886579286111029, + "accuracy": 0.9888025506546623, "total_bits": 475479040, "q_proj": { "group_size": { @@ -78426,7 +78426,7 @@ } }, { - "accuracy": 0.9920373362930197, + "accuracy": 0.9920980757788608, "total_bits": 609759232, "q_proj": { "group_size": { @@ -78478,7 +78478,7 @@ } }, { - "accuracy": 0.9918922632932663, + "accuracy": 0.9922085075001967, "total_bits": 610024448, "q_proj": { "group_size": { @@ -78530,7 +78530,7 @@ } }, { - "accuracy": 0.9927962179246702, + "accuracy": 0.9928913994839317, "total_bits": 615020544, "q_proj": { "group_size": { @@ -78582,7 +78582,7 @@ } }, { - "accuracy": 0.9929505108218444, + "accuracy": 0.9930733214867743, "total_bits": 623951872, "q_proj": { "group_size": { @@ -78634,7 +78634,7 @@ } }, { - "accuracy": 0.9926058563746905, + "accuracy": 0.9928425668101561, "total_bits": 626473984, "q_proj": { "group_size": { @@ -78698,7 +78698,7 @@ } }, { - "accuracy": 0.9931644813010567, + "accuracy": 0.9929139355295583, "total_bits": 630355968, "q_proj": { "group_size": { @@ -78762,7 +78762,7 @@ } }, { - "accuracy": 0.9945383601282772, + "accuracy": 0.994438886642456, "total_bits": 637362176, "q_proj": { "group_size": { @@ -78823,7 +78823,7 @@ } }, { - "accuracy": 0.9950028708890865, + "accuracy": 0.9949209572453248, "total_bits": 646823936, "q_proj": { "group_size": { @@ -78884,7 +78884,7 @@ } }, { - "accuracy": 0.996914617913334, + "accuracy": 0.9969786023230929, "total_bits": 784740352, "q_proj": { "group_size": { @@ -78945,7 +78945,7 @@ } }, { - "accuracy": 0.9974455035438663, + "accuracy": 0.9974690630639854, "total_bits": 797818880, "q_proj": { "group_size": { @@ -79006,7 +79006,7 @@ } }, { - "accuracy": 0.9973337273848685, + "accuracy": 0.9973948233221707, "total_bits": 911749120, "q_proj": { "group_size": { @@ -79058,7 +79058,7 @@ } }, { - "accuracy": 0.9989414832701808, + "accuracy": 0.9989473215843502, "total_bits": 942718976, "q_proj": { "group_size": { @@ -79110,7 +79110,7 @@ } }, { - "accuracy": 0.9993197986562001, + "accuracy": 0.9993007831569565, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -79164,7 +79164,7 @@ ], "model.layers.40.mlp": [ { - "accuracy": 0.9552281969471982, + "accuracy": 0.9555281432051408, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -79216,7 +79216,7 @@ } }, { - "accuracy": 0.9563342772032085, + "accuracy": 0.9566291853001243, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -79268,7 +79268,7 @@ } }, { - "accuracy": 0.9626029855326602, + "accuracy": 0.9628483276618154, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -79317,7 +79317,7 @@ } }, { - "accuracy": 0.9644410672940706, + "accuracy": 0.9646693719060797, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -79366,7 +79366,7 @@ } }, { - "accuracy": 0.9775899899633307, + "accuracy": 0.9777202731684634, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -79418,7 +79418,7 @@ } }, { - "accuracy": 0.9793292817316557, + "accuracy": 0.9794589641847109, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -79470,7 +79470,7 @@ } }, { - "accuracy": 0.9819396389158148, + "accuracy": 0.9820507811872583, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -79519,7 +79519,7 @@ } }, { - "accuracy": 0.9885055203186838, + "accuracy": 0.9885740789927935, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -79562,7 +79562,7 @@ } }, { - "accuracy": 0.9894494108463588, + "accuracy": 0.9895144685318595, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -79605,7 +79605,7 @@ } }, { - "accuracy": 0.9886044593233811, + "accuracy": 0.9886689680187326, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -79657,7 +79657,7 @@ } }, { - "accuracy": 0.9899650757250033, + "accuracy": 0.9900291295428025, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -79709,7 +79709,7 @@ } }, { - "accuracy": 0.9941902815511352, + "accuracy": 0.9942181420169378, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -79761,7 +79761,7 @@ } }, { - "accuracy": 0.9949920173538359, + "accuracy": 0.9950224921891564, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -79813,7 +79813,7 @@ } }, { - "accuracy": 0.996880266619356, + "accuracy": 0.9968949140686738, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -79856,7 +79856,7 @@ } }, { - "accuracy": 0.9970214825711752, + "accuracy": 0.9970344887360146, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -79905,7 +79905,7 @@ } }, { - "accuracy": 0.9975181218040617, + "accuracy": 0.9975260347127914, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -79951,7 +79951,7 @@ } }, { - "accuracy": 0.9991066198992101, + "accuracy": 0.9991103456796784, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -79993,7 +79993,7 @@ ], "model.layers.41.self_attn": [ { - "accuracy": 0.9817931997148615, + "accuracy": 0.9820245739660765, "total_bits": 320757760, "q_proj": { "group_size": { @@ -80057,7 +80057,7 @@ } }, { - "accuracy": 0.9826164276976335, + "accuracy": 0.9828215915905801, "total_bits": 329080832, "q_proj": { "group_size": { @@ -80121,7 +80121,7 @@ } }, { - "accuracy": 0.9844932618894076, + "accuracy": 0.9847946308161083, "total_bits": 336024576, "q_proj": { "group_size": { @@ -80185,7 +80185,7 @@ } }, { - "accuracy": 0.9875000895638215, + "accuracy": 0.9878091906246386, "total_bits": 401557504, "q_proj": { "group_size": { @@ -80249,7 +80249,7 @@ } }, { - "accuracy": 0.9898595566812315, + "accuracy": 0.9900020807981491, "total_bits": 475279360, "q_proj": { "group_size": { @@ -80313,7 +80313,7 @@ } }, { - "accuracy": 0.9900228600752982, + "accuracy": 0.9902586521286714, "total_bits": 475479040, "q_proj": { "group_size": { @@ -80377,7 +80377,7 @@ } }, { - "accuracy": 0.9931456344692331, + "accuracy": 0.9930134313671213, "total_bits": 609759232, "q_proj": { "group_size": { @@ -80429,7 +80429,7 @@ } }, { - "accuracy": 0.9932646108301062, + "accuracy": 0.9933127870685176, "total_bits": 610024448, "q_proj": { "group_size": { @@ -80481,7 +80481,7 @@ } }, { - "accuracy": 0.9938144746579622, + "accuracy": 0.9937073619742143, "total_bits": 615020544, "q_proj": { "group_size": { @@ -80533,7 +80533,7 @@ } }, { - "accuracy": 0.993970381978311, + "accuracy": 0.99388750368043, "total_bits": 623951872, "q_proj": { "group_size": { @@ -80585,7 +80585,7 @@ } }, { - "accuracy": 0.9936641435089865, + "accuracy": 0.9937051017033426, "total_bits": 626473984, "q_proj": { "group_size": { @@ -80649,7 +80649,7 @@ } }, { - "accuracy": 0.9941253756221972, + "accuracy": 0.9942337411798929, "total_bits": 630355968, "q_proj": { "group_size": { @@ -80713,7 +80713,7 @@ } }, { - "accuracy": 0.9951528400966996, + "accuracy": 0.9951638595054024, "total_bits": 637362176, "q_proj": { "group_size": { @@ -80774,7 +80774,7 @@ } }, { - "accuracy": 0.9954839184095985, + "accuracy": 0.9955563419743588, "total_bits": 646823936, "q_proj": { "group_size": { @@ -80835,7 +80835,7 @@ } }, { - "accuracy": 0.9971672412204114, + "accuracy": 0.997185392207221, "total_bits": 784740352, "q_proj": { "group_size": { @@ -80896,7 +80896,7 @@ } }, { - "accuracy": 0.9976280808056656, + "accuracy": 0.9976375326514244, "total_bits": 797818880, "q_proj": { "group_size": { @@ -80957,7 +80957,7 @@ } }, { - "accuracy": 0.9975425850011801, + "accuracy": 0.9975553480418105, "total_bits": 911749120, "q_proj": { "group_size": { @@ -81009,7 +81009,7 @@ } }, { - "accuracy": 0.9990448224309244, + "accuracy": 0.9990515108955534, "total_bits": 942718976, "q_proj": { "group_size": { @@ -81061,7 +81061,7 @@ } }, { - "accuracy": 0.9993565899172896, + "accuracy": 0.9993750588282159, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -81115,7 +81115,7 @@ ], "model.layers.41.mlp": [ { - "accuracy": 0.9562880459584688, + "accuracy": 0.9565535846509432, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -81167,7 +81167,7 @@ } }, { - "accuracy": 0.9573628714210108, + "accuracy": 0.9576228166881361, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -81219,7 +81219,7 @@ } }, { - "accuracy": 0.9634424165675515, + "accuracy": 0.9636591014109159, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -81268,7 +81268,7 @@ } }, { - "accuracy": 0.9651949437041032, + "accuracy": 0.9653998550615812, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -81317,7 +81317,7 @@ } }, { - "accuracy": 0.9781073927879333, + "accuracy": 0.9782398750907496, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -81369,7 +81369,7 @@ } }, { - "accuracy": 0.9798022398823186, + "accuracy": 0.9799308933709797, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -81421,7 +81421,7 @@ } }, { - "accuracy": 0.9823228089432967, + "accuracy": 0.982431731725994, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -81470,7 +81470,7 @@ } }, { - "accuracy": 0.9888174533843994, + "accuracy": 0.9888902651636224, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -81513,7 +81513,7 @@ } }, { - "accuracy": 0.9897387325763702, + "accuracy": 0.9898021032935694, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -81556,7 +81556,7 @@ } }, { - "accuracy": 0.9888859014762076, + "accuracy": 0.9889562647593649, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -81608,7 +81608,7 @@ } }, { - "accuracy": 0.9902160167694092, + "accuracy": 0.9902767554709786, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -81660,7 +81660,7 @@ } }, { - "accuracy": 0.9943471727402586, + "accuracy": 0.9943814724683762, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -81712,7 +81712,7 @@ } }, { - "accuracy": 0.9951270579507476, + "accuracy": 0.9951572539774995, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -81764,7 +81764,7 @@ } }, { - "accuracy": 0.996986735612154, + "accuracy": 0.9970051158023508, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -81807,7 +81807,7 @@ } }, { - "accuracy": 0.9971056562897406, + "accuracy": 0.9971234782745964, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -81856,7 +81856,7 @@ } }, { - "accuracy": 0.9975798835879878, + "accuracy": 0.9975941473324048, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -81902,7 +81902,7 @@ } }, { - "accuracy": 0.9991477336734533, + "accuracy": 0.999152126182851, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -81944,7 +81944,7 @@ ], "model.layers.42.self_attn": [ { - "accuracy": 0.9830969430898365, + "accuracy": 0.9824334069302207, "total_bits": 320757760, "q_proj": { "group_size": { @@ -82008,7 +82008,7 @@ } }, { - "accuracy": 0.9843516255679884, + "accuracy": 0.9837725193876969, "total_bits": 329080832, "q_proj": { "group_size": { @@ -82072,7 +82072,7 @@ } }, { - "accuracy": 0.9869298589857001, + "accuracy": 0.9872976102327046, "total_bits": 336024576, "q_proj": { "group_size": { @@ -82136,7 +82136,7 @@ } }, { - "accuracy": 0.9892101578022304, + "accuracy": 0.9895960641534705, "total_bits": 401557504, "q_proj": { "group_size": { @@ -82200,7 +82200,7 @@ } }, { - "accuracy": 0.990269956620116, + "accuracy": 0.9907013240613436, "total_bits": 475279360, "q_proj": { "group_size": { @@ -82264,7 +82264,7 @@ } }, { - "accuracy": 0.9911504921160246, + "accuracy": 0.9906391854348936, "total_bits": 475479040, "q_proj": { "group_size": { @@ -82328,7 +82328,7 @@ } }, { - "accuracy": 0.9929095130217703, + "accuracy": 0.9929247827906358, "total_bits": 609759232, "q_proj": { "group_size": { @@ -82380,7 +82380,7 @@ } }, { - "accuracy": 0.9936545565724373, + "accuracy": 0.9935074468192301, "total_bits": 610024448, "q_proj": { "group_size": { @@ -82432,7 +82432,7 @@ } }, { - "accuracy": 0.9940914480309737, + "accuracy": 0.9941695736427056, "total_bits": 615020544, "q_proj": { "group_size": { @@ -82484,7 +82484,7 @@ } }, { - "accuracy": 0.9942251483076497, + "accuracy": 0.9942970271957549, "total_bits": 623951872, "q_proj": { "group_size": { @@ -82536,7 +82536,7 @@ } }, { - "accuracy": 0.9940528065750474, + "accuracy": 0.9940173861227537, "total_bits": 626473984, "q_proj": { "group_size": { @@ -82600,7 +82600,7 @@ } }, { - "accuracy": 0.9944141691452578, + "accuracy": 0.9942979228339697, "total_bits": 630355968, "q_proj": { "group_size": { @@ -82664,7 +82664,7 @@ } }, { - "accuracy": 0.9956185029525506, + "accuracy": 0.9955870893440748, "total_bits": 637362176, "q_proj": { "group_size": { @@ -82725,7 +82725,7 @@ } }, { - "accuracy": 0.9959789796879417, + "accuracy": 0.9960328383665336, "total_bits": 646823936, "q_proj": { "group_size": { @@ -82786,7 +82786,7 @@ } }, { - "accuracy": 0.9974598276771998, + "accuracy": 0.997448889048476, "total_bits": 784740352, "q_proj": { "group_size": { @@ -82847,7 +82847,7 @@ } }, { - "accuracy": 0.9978488509200121, + "accuracy": 0.9979035813165339, "total_bits": 797818880, "q_proj": { "group_size": { @@ -82908,7 +82908,7 @@ } }, { - "accuracy": 0.9976865890386858, + "accuracy": 0.9976659342646599, "total_bits": 911749120, "q_proj": { "group_size": { @@ -82960,7 +82960,7 @@ } }, { - "accuracy": 0.9991700480643072, + "accuracy": 0.9991910574271491, "total_bits": 942718976, "q_proj": { "group_size": { @@ -83012,7 +83012,7 @@ } }, { - "accuracy": 0.9993796258380538, + "accuracy": 0.9993833908321041, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -83066,7 +83066,7 @@ ], "model.layers.42.mlp": [ { - "accuracy": 0.9569482615119532, + "accuracy": 0.9572096498388993, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -83118,7 +83118,7 @@ } }, { - "accuracy": 0.958022183493564, + "accuracy": 0.9582767768910057, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -83170,7 +83170,7 @@ } }, { - "accuracy": 0.9639153198191994, + "accuracy": 0.9641302698536923, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -83219,7 +83219,7 @@ } }, { - "accuracy": 0.9655893131306297, + "accuracy": 0.9657911438690989, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -83268,7 +83268,7 @@ } }, { - "accuracy": 0.9784307746510756, + "accuracy": 0.9785720564817127, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -83320,7 +83320,7 @@ } }, { - "accuracy": 0.9800992215934553, + "accuracy": 0.9802211350516269, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -83372,7 +83372,7 @@ } }, { - "accuracy": 0.9825308448389957, + "accuracy": 0.9826358682230899, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -83421,7 +83421,7 @@ } }, { - "accuracy": 0.9889967402345256, + "accuracy": 0.9890637750688352, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -83464,7 +83464,7 @@ } }, { - "accuracy": 0.9898840275249983, + "accuracy": 0.989946142623299, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -83507,7 +83507,7 @@ } }, { - "accuracy": 0.9890469320510563, + "accuracy": 0.9891143723538047, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -83559,7 +83559,7 @@ } }, { - "accuracy": 0.9903457651012822, + "accuracy": 0.9904065477220636, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -83611,7 +83611,7 @@ } }, { - "accuracy": 0.9944236827524084, + "accuracy": 0.9944593149580454, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -83663,7 +83663,7 @@ } }, { - "accuracy": 0.9951882095713365, + "accuracy": 0.9952182738404525, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -83715,7 +83715,7 @@ } }, { - "accuracy": 0.9970237979371297, + "accuracy": 0.9970426343773541, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -83758,7 +83758,7 @@ } }, { - "accuracy": 0.9971382841467857, + "accuracy": 0.9971554696952042, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -83807,7 +83807,7 @@ } }, { - "accuracy": 0.997592465461869, + "accuracy": 0.9976057929820136, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -83853,7 +83853,7 @@ } }, { - "accuracy": 0.9991397337105713, + "accuracy": 0.999144579528978, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -83895,7 +83895,7 @@ ], "model.layers.43.self_attn": [ { - "accuracy": 0.9823797156936244, + "accuracy": 0.982325698200025, "total_bits": 320757760, "q_proj": { "group_size": { @@ -83959,7 +83959,7 @@ } }, { - "accuracy": 0.9843285177883349, + "accuracy": 0.98441271405471, "total_bits": 329080832, "q_proj": { "group_size": { @@ -84023,7 +84023,7 @@ } }, { - "accuracy": 0.986210109371888, + "accuracy": 0.9864120420656706, "total_bits": 336024576, "q_proj": { "group_size": { @@ -84087,7 +84087,7 @@ } }, { - "accuracy": 0.9891883134841919, + "accuracy": 0.9893656082843479, "total_bits": 401557504, "q_proj": { "group_size": { @@ -84151,7 +84151,7 @@ } }, { - "accuracy": 0.9906747207829827, + "accuracy": 0.9908248154740584, "total_bits": 475279360, "q_proj": { "group_size": { @@ -84215,7 +84215,7 @@ } }, { - "accuracy": 0.9909894027208027, + "accuracy": 0.9911323574028517, "total_bits": 475479040, "q_proj": { "group_size": { @@ -84279,7 +84279,7 @@ } }, { - "accuracy": 0.992772616838154, + "accuracy": 0.99295663755191, "total_bits": 609759232, "q_proj": { "group_size": { @@ -84331,7 +84331,7 @@ } }, { - "accuracy": 0.9931059793422097, + "accuracy": 0.9932821471440164, "total_bits": 610024448, "q_proj": { "group_size": { @@ -84383,7 +84383,7 @@ } }, { - "accuracy": 0.9937656823741762, + "accuracy": 0.993888983601018, "total_bits": 615020544, "q_proj": { "group_size": { @@ -84435,7 +84435,7 @@ } }, { - "accuracy": 0.9939158707857132, + "accuracy": 0.9940254225542671, "total_bits": 623951872, "q_proj": { "group_size": { @@ -84487,7 +84487,7 @@ } }, { - "accuracy": 0.99495162775642, + "accuracy": 0.9949555412719124, "total_bits": 626473984, "q_proj": { "group_size": { @@ -84551,7 +84551,7 @@ } }, { - "accuracy": 0.9953230321407318, + "accuracy": 0.9953332445339153, "total_bits": 630355968, "q_proj": { "group_size": { @@ -84615,7 +84615,7 @@ } }, { - "accuracy": 0.9953588466895255, + "accuracy": 0.9953706252731775, "total_bits": 637362176, "q_proj": { "group_size": { @@ -84676,7 +84676,7 @@ } }, { - "accuracy": 0.9957619897629085, + "accuracy": 0.9957713069100129, "total_bits": 646823936, "q_proj": { "group_size": { @@ -84737,7 +84737,7 @@ } }, { - "accuracy": 0.9973575594393831, + "accuracy": 0.9973924197256565, "total_bits": 784740352, "q_proj": { "group_size": { @@ -84798,7 +84798,7 @@ } }, { - "accuracy": 0.9978296480288631, + "accuracy": 0.9978575330031546, "total_bits": 797818880, "q_proj": { "group_size": { @@ -84859,7 +84859,7 @@ } }, { - "accuracy": 0.997661608026216, + "accuracy": 0.997695264847655, "total_bits": 911749120, "q_proj": { "group_size": { @@ -84911,7 +84911,7 @@ } }, { - "accuracy": 0.9990955208869357, + "accuracy": 0.999102646583005, "total_bits": 942718976, "q_proj": { "group_size": { @@ -84963,7 +84963,7 @@ } }, { - "accuracy": 0.999374607430869, + "accuracy": 0.9993687937232224, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -85017,7 +85017,7 @@ ], "model.layers.43.mlp": [ { - "accuracy": 0.9564006548178823, + "accuracy": 0.9566827573274311, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -85069,7 +85069,7 @@ } }, { - "accuracy": 0.9574721298719707, + "accuracy": 0.9577432243447555, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -85121,7 +85121,7 @@ } }, { - "accuracy": 0.9633637258881017, + "accuracy": 0.9635925136114422, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -85170,7 +85170,7 @@ } }, { - "accuracy": 0.9650578938032451, + "accuracy": 0.9652739073100843, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -85219,7 +85219,7 @@ } }, { - "accuracy": 0.9781335984405718, + "accuracy": 0.9782734560339075, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -85271,7 +85271,7 @@ } }, { - "accuracy": 0.9798272879500138, + "accuracy": 0.9799567523755526, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -85323,7 +85323,7 @@ } }, { - "accuracy": 0.9822711395589929, + "accuracy": 0.9823813469786393, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -85372,7 +85372,7 @@ } }, { - "accuracy": 0.9888274850029695, + "accuracy": 0.9889002724697715, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -85415,7 +85415,7 @@ } }, { - "accuracy": 0.9897398838871404, + "accuracy": 0.9898054670346411, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -85458,7 +85458,7 @@ } }, { - "accuracy": 0.9888983510042492, + "accuracy": 0.9889678923707259, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -85510,7 +85510,7 @@ } }, { - "accuracy": 0.9902221834973285, + "accuracy": 0.9902845966188532, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -85562,7 +85562,7 @@ } }, { - "accuracy": 0.9943556565987436, + "accuracy": 0.994390411988685, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -85614,7 +85614,7 @@ } }, { - "accuracy": 0.9951306797171894, + "accuracy": 0.9951621160695427, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -85666,7 +85666,7 @@ } }, { - "accuracy": 0.99699026503061, + "accuracy": 0.9970091710749426, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -85709,7 +85709,7 @@ } }, { - "accuracy": 0.9971191063523293, + "accuracy": 0.9971363277811753, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -85758,7 +85758,7 @@ } }, { - "accuracy": 0.9975798386884364, + "accuracy": 0.997593748334207, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -85804,7 +85804,7 @@ } }, { - "accuracy": 0.9991810268379355, + "accuracy": 0.9991863772371098, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -85846,7 +85846,7 @@ ], "model.layers.44.self_attn": [ { - "accuracy": 0.9730414654079237, + "accuracy": 0.9732459312991092, "total_bits": 320757760, "q_proj": { "group_size": { @@ -85910,7 +85910,7 @@ } }, { - "accuracy": 0.9756704066929064, + "accuracy": 0.9731757703580355, "total_bits": 329080832, "q_proj": { "group_size": { @@ -85974,7 +85974,7 @@ } }, { - "accuracy": 0.9815379930169958, + "accuracy": 0.9813058439053988, "total_bits": 336024576, "q_proj": { "group_size": { @@ -86038,7 +86038,7 @@ } }, { - "accuracy": 0.9841942520518052, + "accuracy": 0.983951167056435, "total_bits": 401557504, "q_proj": { "group_size": { @@ -86102,7 +86102,7 @@ } }, { - "accuracy": 0.9868487166730981, + "accuracy": 0.9864479146505657, "total_bits": 475279360, "q_proj": { "group_size": { @@ -86166,7 +86166,7 @@ } }, { - "accuracy": 0.9869143288386496, + "accuracy": 0.987035447829648, "total_bits": 475479040, "q_proj": { "group_size": { @@ -86230,7 +86230,7 @@ } }, { - "accuracy": 0.9918188789957448, + "accuracy": 0.9916887832315344, "total_bits": 609759232, "q_proj": { "group_size": { @@ -86282,7 +86282,7 @@ } }, { - "accuracy": 0.9922087757210982, + "accuracy": 0.992288202047348, "total_bits": 610024448, "q_proj": { "group_size": { @@ -86334,7 +86334,7 @@ } }, { - "accuracy": 0.9929692157005009, + "accuracy": 0.9926982161245848, "total_bits": 615020544, "q_proj": { "group_size": { @@ -86386,7 +86386,7 @@ } }, { - "accuracy": 0.9930917706928755, + "accuracy": 0.9927970382728075, "total_bits": 623951872, "q_proj": { "group_size": { @@ -86438,7 +86438,7 @@ } }, { - "accuracy": 0.9928022113285566, + "accuracy": 0.992726060120683, "total_bits": 626473984, "q_proj": { "group_size": { @@ -86502,7 +86502,7 @@ } }, { - "accuracy": 0.9929060065432599, + "accuracy": 0.9931099555994335, "total_bits": 630355968, "q_proj": { "group_size": { @@ -86566,7 +86566,7 @@ } }, { - "accuracy": 0.995056896617538, + "accuracy": 0.995014460071137, "total_bits": 637362176, "q_proj": { "group_size": { @@ -86627,7 +86627,7 @@ } }, { - "accuracy": 0.9954406721027274, + "accuracy": 0.9954403780008617, "total_bits": 646823936, "q_proj": { "group_size": { @@ -86688,7 +86688,7 @@ } }, { - "accuracy": 0.9972772778649079, + "accuracy": 0.9972987025976181, "total_bits": 784740352, "q_proj": { "group_size": { @@ -86749,7 +86749,7 @@ } }, { - "accuracy": 0.9977538820944334, + "accuracy": 0.9977082988541377, "total_bits": 797818880, "q_proj": { "group_size": { @@ -86810,7 +86810,7 @@ } }, { - "accuracy": 0.997709425852487, + "accuracy": 0.9977288036362121, "total_bits": 911749120, "q_proj": { "group_size": { @@ -86862,7 +86862,7 @@ } }, { - "accuracy": 0.999053031010063, + "accuracy": 0.999058907459441, "total_bits": 942718976, "q_proj": { "group_size": { @@ -86914,7 +86914,7 @@ } }, { - "accuracy": 0.9993736075825597, + "accuracy": 0.999398975191932, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -86968,7 +86968,7 @@ ], "model.layers.44.mlp": [ { - "accuracy": 0.955465831254658, + "accuracy": 0.9557425630720038, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -87020,7 +87020,7 @@ } }, { - "accuracy": 0.9565646805261311, + "accuracy": 0.9568239293600383, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -87072,7 +87072,7 @@ } }, { - "accuracy": 0.9625380760744998, + "accuracy": 0.9627644011848852, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -87121,7 +87121,7 @@ } }, { - "accuracy": 0.9642741303694875, + "accuracy": 0.9644875714653417, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -87170,7 +87170,7 @@ } }, { - "accuracy": 0.97762121181739, + "accuracy": 0.9777541882113406, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -87222,7 +87222,7 @@ } }, { - "accuracy": 0.9793603279088673, + "accuracy": 0.9794821613713315, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -87274,7 +87274,7 @@ } }, { - "accuracy": 0.9818637151467172, + "accuracy": 0.9819672829226443, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -87323,7 +87323,7 @@ } }, { - "accuracy": 0.9885484137033161, + "accuracy": 0.9886170123752794, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -87366,7 +87366,7 @@ } }, { - "accuracy": 0.9894832255024659, + "accuracy": 0.9895496391936353, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -87409,7 +87409,7 @@ } }, { - "accuracy": 0.9886260722812853, + "accuracy": 0.9886969453410098, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -87461,7 +87461,7 @@ } }, { - "accuracy": 0.9899833633711463, + "accuracy": 0.990044534990662, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -87513,7 +87513,7 @@ } }, { - "accuracy": 0.9942096040437096, + "accuracy": 0.9942440990554658, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -87565,7 +87565,7 @@ } }, { - "accuracy": 0.9950077819981074, + "accuracy": 0.9950383403583577, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -87617,7 +87617,7 @@ } }, { - "accuracy": 0.9969052195941147, + "accuracy": 0.9969239334918951, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -87660,7 +87660,7 @@ } }, { - "accuracy": 0.9970472547176638, + "accuracy": 0.9970653061019746, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -87709,7 +87709,7 @@ } }, { - "accuracy": 0.9975197985768318, + "accuracy": 0.9975342711335734, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -87755,7 +87755,7 @@ } }, { - "accuracy": 0.9991705229407862, + "accuracy": 0.9991751367128209, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -87797,7 +87797,7 @@ ], "model.layers.45.self_attn": [ { - "accuracy": 0.982853418902347, + "accuracy": 0.9824475975413072, "total_bits": 320757760, "q_proj": { "group_size": { @@ -87861,7 +87861,7 @@ } }, { - "accuracy": 0.9837191324484976, + "accuracy": 0.9838520555119765, "total_bits": 329080832, "q_proj": { "group_size": { @@ -87925,7 +87925,7 @@ } }, { - "accuracy": 0.9858417636469791, + "accuracy": 0.9858118970143167, "total_bits": 336024576, "q_proj": { "group_size": { @@ -87989,7 +87989,7 @@ } }, { - "accuracy": 0.9893110857198113, + "accuracy": 0.9892793925184953, "total_bits": 401557504, "q_proj": { "group_size": { @@ -88053,7 +88053,7 @@ } }, { - "accuracy": 0.9906263555351057, + "accuracy": 0.9906207072107416, "total_bits": 475279360, "q_proj": { "group_size": { @@ -88117,7 +88117,7 @@ } }, { - "accuracy": 0.9909175555956992, + "accuracy": 0.9909735008289939, "total_bits": 475479040, "q_proj": { "group_size": { @@ -88181,7 +88181,7 @@ } }, { - "accuracy": 0.9931228945129796, + "accuracy": 0.993268790997957, "total_bits": 609759232, "q_proj": { "group_size": { @@ -88233,7 +88233,7 @@ } }, { - "accuracy": 0.9937789773470477, + "accuracy": 0.993914075980061, "total_bits": 610024448, "q_proj": { "group_size": { @@ -88285,7 +88285,7 @@ } }, { - "accuracy": 0.994074456001583, + "accuracy": 0.9940957776025722, "total_bits": 615020544, "q_proj": { "group_size": { @@ -88337,7 +88337,7 @@ } }, { - "accuracy": 0.9942134795220274, + "accuracy": 0.9942527793739971, "total_bits": 623951872, "q_proj": { "group_size": { @@ -88389,7 +88389,7 @@ } }, { - "accuracy": 0.9944044273150595, + "accuracy": 0.9944206319357219, "total_bits": 626473984, "q_proj": { "group_size": { @@ -88453,7 +88453,7 @@ } }, { - "accuracy": 0.9946309453562686, + "accuracy": 0.9946443387552312, "total_bits": 630355968, "q_proj": { "group_size": { @@ -88517,7 +88517,7 @@ } }, { - "accuracy": 0.9955286003257099, + "accuracy": 0.9954936312217462, "total_bits": 637362176, "q_proj": { "group_size": { @@ -88578,7 +88578,7 @@ } }, { - "accuracy": 0.9959037080407143, + "accuracy": 0.9958983761699576, "total_bits": 646823936, "q_proj": { "group_size": { @@ -88639,7 +88639,7 @@ } }, { - "accuracy": 0.9973872184361282, + "accuracy": 0.9973826935808909, "total_bits": 784740352, "q_proj": { "group_size": { @@ -88700,7 +88700,7 @@ } }, { - "accuracy": 0.9978876023700363, + "accuracy": 0.9978909966977019, "total_bits": 797818880, "q_proj": { "group_size": { @@ -88761,7 +88761,7 @@ } }, { - "accuracy": 0.9977436946018746, + "accuracy": 0.9977342676567403, "total_bits": 911749120, "q_proj": { "group_size": { @@ -88813,7 +88813,7 @@ } }, { - "accuracy": 0.9990780104539896, + "accuracy": 0.9990940974339059, "total_bits": 942718976, "q_proj": { "group_size": { @@ -88865,7 +88865,7 @@ } }, { - "accuracy": 0.9993914923091468, + "accuracy": 0.9994043405903014, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -88919,7 +88919,7 @@ ], "model.layers.45.mlp": [ { - "accuracy": 0.955856298145495, + "accuracy": 0.956102728843689, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -88971,7 +88971,7 @@ } }, { - "accuracy": 0.9569508182375055, + "accuracy": 0.957200564836201, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -89023,7 +89023,7 @@ } }, { - "accuracy": 0.9627949281742698, + "accuracy": 0.9630050125874972, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -89072,7 +89072,7 @@ } }, { - "accuracy": 0.9644691912751449, + "accuracy": 0.9646707271274767, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -89121,7 +89121,7 @@ } }, { - "accuracy": 0.9778132187692743, + "accuracy": 0.9779460586999592, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -89173,7 +89173,7 @@ } }, { - "accuracy": 0.9795462912634799, + "accuracy": 0.9796592969643442, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -89225,7 +89225,7 @@ } }, { - "accuracy": 0.9819757954070443, + "accuracy": 0.9820750371405953, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -89274,7 +89274,7 @@ } }, { - "accuracy": 0.9886552613032492, + "accuracy": 0.9887196327510633, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -89317,7 +89317,7 @@ } }, { - "accuracy": 0.9895810226076528, + "accuracy": 0.9896382077744132, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -89360,7 +89360,7 @@ } }, { - "accuracy": 0.9887195543238991, + "accuracy": 0.988784377512179, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -89412,7 +89412,7 @@ } }, { - "accuracy": 0.9900700732281333, + "accuracy": 0.9901257180853894, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -89464,7 +89464,7 @@ } }, { - "accuracy": 0.9942552792398553, + "accuracy": 0.9942903134383654, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -89516,7 +89516,7 @@ } }, { - "accuracy": 0.9950489197906695, + "accuracy": 0.9950772653284826, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -89568,7 +89568,7 @@ } }, { - "accuracy": 0.9969288644037748, + "accuracy": 0.9969461772002672, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -89611,7 +89611,7 @@ } }, { - "accuracy": 0.9970582151099255, + "accuracy": 0.9970748949992029, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -89660,7 +89660,7 @@ } }, { - "accuracy": 0.9975122085919506, + "accuracy": 0.9975263182269899, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -89706,7 +89706,7 @@ } }, { - "accuracy": 0.9991333953251964, + "accuracy": 0.9991382930035654, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -89748,7 +89748,7 @@ ], "model.layers.46.self_attn": [ { - "accuracy": 0.9909227662964871, + "accuracy": 0.9909246422742543, "total_bits": 320757760, "q_proj": { "group_size": { @@ -89812,7 +89812,7 @@ } }, { - "accuracy": 0.9914209803468302, + "accuracy": 0.9914267078826302, "total_bits": 329080832, "q_proj": { "group_size": { @@ -89876,7 +89876,7 @@ } }, { - "accuracy": 0.9927722301922346, + "accuracy": 0.9928660855481499, "total_bits": 336024576, "q_proj": { "group_size": { @@ -89940,7 +89940,7 @@ } }, { - "accuracy": 0.9941842532471606, + "accuracy": 0.9942549149456777, "total_bits": 401557504, "q_proj": { "group_size": { @@ -90004,7 +90004,7 @@ } }, { - "accuracy": 0.9945745107374693, + "accuracy": 0.9946460194493595, "total_bits": 475279360, "q_proj": { "group_size": { @@ -90068,7 +90068,7 @@ } }, { - "accuracy": 0.9949821390603718, + "accuracy": 0.9950308689945623, "total_bits": 475479040, "q_proj": { "group_size": { @@ -90132,7 +90132,7 @@ } }, { - "accuracy": 0.9956122040748596, + "accuracy": 0.995691850389305, "total_bits": 609759232, "q_proj": { "group_size": { @@ -90184,7 +90184,7 @@ } }, { - "accuracy": 0.996096390250482, + "accuracy": 0.9961762816498154, "total_bits": 610024448, "q_proj": { "group_size": { @@ -90236,7 +90236,7 @@ } }, { - "accuracy": 0.9967141676890222, + "accuracy": 0.9967031906309881, "total_bits": 615020544, "q_proj": { "group_size": { @@ -90288,7 +90288,7 @@ } }, { - "accuracy": 0.9967640114850119, + "accuracy": 0.9967470818052167, "total_bits": 623951872, "q_proj": { "group_size": { @@ -90340,7 +90340,7 @@ } }, { - "accuracy": 0.9968840689643433, + "accuracy": 0.9969318473809644, "total_bits": 626473984, "q_proj": { "group_size": { @@ -90404,7 +90404,7 @@ } }, { - "accuracy": 0.997186500186983, + "accuracy": 0.9971998796651238, "total_bits": 630355968, "q_proj": { "group_size": { @@ -90468,7 +90468,7 @@ } }, { - "accuracy": 0.9971811920404434, + "accuracy": 0.9972249126355899, "total_bits": 637362176, "q_proj": { "group_size": { @@ -90529,7 +90529,7 @@ } }, { - "accuracy": 0.9974680954688474, + "accuracy": 0.9974709921761563, "total_bits": 646823936, "q_proj": { "group_size": { @@ -90590,7 +90590,7 @@ } }, { - "accuracy": 0.9981132690843783, + "accuracy": 0.9981212935557491, "total_bits": 784740352, "q_proj": { "group_size": { @@ -90651,7 +90651,7 @@ } }, { - "accuracy": 0.998570566996932, + "accuracy": 0.9985853832607207, "total_bits": 797818880, "q_proj": { "group_size": { @@ -90712,7 +90712,7 @@ } }, { - "accuracy": 0.9981938639753744, + "accuracy": 0.9981995820999146, "total_bits": 911749120, "q_proj": { "group_size": { @@ -90764,7 +90764,7 @@ } }, { - "accuracy": 0.999483040190841, + "accuracy": 0.99948624746972, "total_bits": 942718976, "q_proj": { "group_size": { @@ -90816,7 +90816,7 @@ } }, { - "accuracy": 0.9995181019368925, + "accuracy": 0.9995203483359594, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -90870,7 +90870,7 @@ ], "model.layers.46.mlp": [ { - "accuracy": 0.9564528277045802, + "accuracy": 0.9566980914065712, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -90922,7 +90922,7 @@ } }, { - "accuracy": 0.9575276751267283, + "accuracy": 0.9577666301476329, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -90974,7 +90974,7 @@ } }, { - "accuracy": 0.9633109600920426, + "accuracy": 0.9635028839111328, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -91023,7 +91023,7 @@ } }, { - "accuracy": 0.9649377120168585, + "accuracy": 0.9651191705151608, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -91072,7 +91072,7 @@ } }, { - "accuracy": 0.9781249918435749, + "accuracy": 0.9782475923237047, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -91124,7 +91124,7 @@ } }, { - "accuracy": 0.9798254668712616, + "accuracy": 0.9799431985930392, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -91176,7 +91176,7 @@ } }, { - "accuracy": 0.9822103804663608, + "accuracy": 0.9823121193208193, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -91225,7 +91225,7 @@ } }, { - "accuracy": 0.9888431139682469, + "accuracy": 0.9889051914215088, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -91268,7 +91268,7 @@ } }, { - "accuracy": 0.9897488096826955, + "accuracy": 0.9898071006724709, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -91311,7 +91311,7 @@ } }, { - "accuracy": 0.9888875664849031, + "accuracy": 0.9889486463446366, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -91363,7 +91363,7 @@ } }, { - "accuracy": 0.9902130247730958, + "accuracy": 0.9902690554919996, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -91415,7 +91415,7 @@ } }, { - "accuracy": 0.9943476578122691, + "accuracy": 0.99437838596733, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -91467,7 +91467,7 @@ } }, { - "accuracy": 0.9951255572469611, + "accuracy": 0.9951528738203802, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -91519,7 +91519,7 @@ } }, { - "accuracy": 0.9969930197847517, + "accuracy": 0.9970100839671335, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -91562,7 +91562,7 @@ } }, { - "accuracy": 0.9971152996938479, + "accuracy": 0.9971317739078873, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -91611,7 +91611,7 @@ } }, { - "accuracy": 0.9975629576334828, + "accuracy": 0.9975768896309953, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -91657,7 +91657,7 @@ } }, { - "accuracy": 0.9991951579424111, + "accuracy": 0.9991993467862669, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -91699,7 +91699,7 @@ ], "model.layers.47.self_attn": [ { - "accuracy": 0.9771184701668588, + "accuracy": 0.9763606504390114, "total_bits": 320757760, "q_proj": { "group_size": { @@ -91763,7 +91763,7 @@ } }, { - "accuracy": 0.9767287295115622, + "accuracy": 0.9789709354701795, "total_bits": 329080832, "q_proj": { "group_size": { @@ -91827,7 +91827,7 @@ } }, { - "accuracy": 0.9852774190275293, + "accuracy": 0.9854976534843445, "total_bits": 336024576, "q_proj": { "group_size": { @@ -91891,7 +91891,7 @@ } }, { - "accuracy": 0.9870990648081428, + "accuracy": 0.9871449596003482, "total_bits": 401557504, "q_proj": { "group_size": { @@ -91955,7 +91955,7 @@ } }, { - "accuracy": 0.9882681989356091, + "accuracy": 0.9883203271188234, "total_bits": 475279360, "q_proj": { "group_size": { @@ -92019,7 +92019,7 @@ } }, { - "accuracy": 0.9886436878066314, + "accuracy": 0.989045732899716, "total_bits": 475479040, "q_proj": { "group_size": { @@ -92083,7 +92083,7 @@ } }, { - "accuracy": 0.9925911026565652, + "accuracy": 0.9924910970424351, "total_bits": 609759232, "q_proj": { "group_size": { @@ -92135,7 +92135,7 @@ } }, { - "accuracy": 0.993072479179031, + "accuracy": 0.9931109806424693, "total_bits": 610024448, "q_proj": { "group_size": { @@ -92187,7 +92187,7 @@ } }, { - "accuracy": 0.993752544647769, + "accuracy": 0.9937556919298673, "total_bits": 615020544, "q_proj": { "group_size": { @@ -92239,7 +92239,7 @@ } }, { - "accuracy": 0.9938463509867066, + "accuracy": 0.993848379505308, "total_bits": 623951872, "q_proj": { "group_size": { @@ -92291,7 +92291,7 @@ } }, { - "accuracy": 0.993597149848938, + "accuracy": 0.9938085439958071, "total_bits": 626473984, "q_proj": { "group_size": { @@ -92355,7 +92355,7 @@ } }, { - "accuracy": 0.9939589288673902, + "accuracy": 0.9941532294216909, "total_bits": 630355968, "q_proj": { "group_size": { @@ -92419,7 +92419,7 @@ } }, { - "accuracy": 0.9957792284457307, + "accuracy": 0.9958264188546884, "total_bits": 637362176, "q_proj": { "group_size": { @@ -92480,7 +92480,7 @@ } }, { - "accuracy": 0.9962445148511937, + "accuracy": 0.9961864948272705, "total_bits": 646823936, "q_proj": { "group_size": { @@ -92541,7 +92541,7 @@ } }, { - "accuracy": 0.9976301287349901, + "accuracy": 0.9976854706673246, "total_bits": 784740352, "q_proj": { "group_size": { @@ -92602,7 +92602,7 @@ } }, { - "accuracy": 0.9980449643182127, + "accuracy": 0.9980345936982256, "total_bits": 797818880, "q_proj": { "group_size": { @@ -92663,7 +92663,7 @@ } }, { - "accuracy": 0.9978328370734265, + "accuracy": 0.99788370610852, "total_bits": 911749120, "q_proj": { "group_size": { @@ -92715,7 +92715,7 @@ } }, { - "accuracy": 0.9992013331992846, + "accuracy": 0.9992037309627784, "total_bits": 942718976, "q_proj": { "group_size": { @@ -92767,7 +92767,7 @@ } }, { - "accuracy": 0.999454221639194, + "accuracy": 0.9994290053942486, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -92821,7 +92821,7 @@ ], "model.layers.47.mlp": [ { - "accuracy": 0.9557193298088877, + "accuracy": 0.9559574158568132, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -92873,7 +92873,7 @@ } }, { - "accuracy": 0.9568147816156086, + "accuracy": 0.9570435442422566, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -92925,7 +92925,7 @@ } }, { - "accuracy": 0.9626382746194538, + "accuracy": 0.9628379470423648, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -92974,7 +92974,7 @@ } }, { - "accuracy": 0.96429254193055, + "accuracy": 0.9644837253972104, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -93023,7 +93023,7 @@ } }, { - "accuracy": 0.977759262448863, + "accuracy": 0.9778788889709272, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -93075,7 +93075,7 @@ } }, { - "accuracy": 0.9794835275725314, + "accuracy": 0.9795970508926793, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -93127,7 +93127,7 @@ } }, { - "accuracy": 0.9819005492486452, + "accuracy": 0.9819990223959872, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -93176,7 +93176,7 @@ } }, { - "accuracy": 0.9886470186082941, + "accuracy": 0.9887120213947798, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -93219,7 +93219,7 @@ } }, { - "accuracy": 0.9895717219302529, + "accuracy": 0.989628645150285, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -93262,7 +93262,7 @@ } }, { - "accuracy": 0.9887023411299053, + "accuracy": 0.9887647079794031, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -93314,7 +93314,7 @@ } }, { - "accuracy": 0.9900501213575664, + "accuracy": 0.9901054752500433, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -93366,7 +93366,7 @@ } }, { - "accuracy": 0.9942513151388419, + "accuracy": 0.9942834538064504, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -93418,7 +93418,7 @@ } }, { - "accuracy": 0.9950420209451726, + "accuracy": 0.9950689932233409, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -93470,7 +93470,7 @@ } }, { - "accuracy": 0.9969353669960248, + "accuracy": 0.9969522688341769, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -93513,7 +93513,7 @@ } }, { - "accuracy": 0.9970500326078189, + "accuracy": 0.9970662638937172, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -93562,7 +93562,7 @@ } }, { - "accuracy": 0.9975004874561962, + "accuracy": 0.9975138316421133, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -93608,7 +93608,7 @@ } }, { - "accuracy": 0.9991161361532775, + "accuracy": 0.9991221218124816, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -93650,7 +93650,7 @@ ], "model.layers.48.self_attn": [ { - "accuracy": 0.9778017291897222, + "accuracy": 0.9803269630984256, "total_bits": 320757760, "q_proj": { "group_size": { @@ -93714,7 +93714,7 @@ } }, { - "accuracy": 0.97982057929039, + "accuracy": 0.9811478523831618, "total_bits": 329080832, "q_proj": { "group_size": { @@ -93778,7 +93778,7 @@ } }, { - "accuracy": 0.9866230754475844, + "accuracy": 0.9865748396045283, "total_bits": 336024576, "q_proj": { "group_size": { @@ -93842,7 +93842,7 @@ } }, { - "accuracy": 0.9882355034351349, + "accuracy": 0.9881792695898759, "total_bits": 401557504, "q_proj": { "group_size": { @@ -93906,7 +93906,7 @@ } }, { - "accuracy": 0.989134923407906, + "accuracy": 0.9890995849119989, "total_bits": 475279360, "q_proj": { "group_size": { @@ -93970,7 +93970,7 @@ } }, { - "accuracy": 0.9896195162283746, + "accuracy": 0.9893111492458143, "total_bits": 475479040, "q_proj": { "group_size": { @@ -94034,7 +94034,7 @@ } }, { - "accuracy": 0.9925148557675513, + "accuracy": 0.9921609798544332, "total_bits": 609759232, "q_proj": { "group_size": { @@ -94086,7 +94086,7 @@ } }, { - "accuracy": 0.9926538851700331, + "accuracy": 0.9929094534171256, "total_bits": 610024448, "q_proj": { "group_size": { @@ -94138,7 +94138,7 @@ } }, { - "accuracy": 0.9934964630948869, + "accuracy": 0.9935594299121907, "total_bits": 615020544, "q_proj": { "group_size": { @@ -94190,7 +94190,7 @@ } }, { - "accuracy": 0.9935890836151022, + "accuracy": 0.9936424093811136, "total_bits": 623951872, "q_proj": { "group_size": { @@ -94242,7 +94242,7 @@ } }, { - "accuracy": 0.993412476621176, + "accuracy": 0.9932677620335629, "total_bits": 626473984, "q_proj": { "group_size": { @@ -94306,7 +94306,7 @@ } }, { - "accuracy": 0.99385357295212, + "accuracy": 0.9936460801645329, "total_bits": 630355968, "q_proj": { "group_size": { @@ -94370,7 +94370,7 @@ } }, { - "accuracy": 0.9956947690562198, + "accuracy": 0.995630959539037, "total_bits": 637362176, "q_proj": { "group_size": { @@ -94431,7 +94431,7 @@ } }, { - "accuracy": 0.9959376266128138, + "accuracy": 0.9959845397817461, "total_bits": 646823936, "q_proj": { "group_size": { @@ -94492,7 +94492,7 @@ } }, { - "accuracy": 0.9973975645476266, + "accuracy": 0.9974982693399254, "total_bits": 784740352, "q_proj": { "group_size": { @@ -94553,7 +94553,7 @@ } }, { - "accuracy": 0.9979184775760299, + "accuracy": 0.9979106117235986, "total_bits": 797818880, "q_proj": { "group_size": { @@ -94614,7 +94614,7 @@ } }, { - "accuracy": 0.9975654102469745, + "accuracy": 0.9976757829518694, "total_bits": 911749120, "q_proj": { "group_size": { @@ -94666,7 +94666,7 @@ } }, { - "accuracy": 0.9992347068496441, + "accuracy": 0.9992339374301465, "total_bits": 942718976, "q_proj": { "group_size": { @@ -94718,7 +94718,7 @@ } }, { - "accuracy": 0.9993538344769102, + "accuracy": 0.9993756399735024, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -94772,7 +94772,7 @@ ], "model.layers.48.mlp": [ { - "accuracy": 0.954665083634226, + "accuracy": 0.9549049295877156, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -94824,7 +94824,7 @@ } }, { - "accuracy": 0.9557959029549047, + "accuracy": 0.9560289728014093, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -94876,7 +94876,7 @@ } }, { - "accuracy": 0.9617160872409218, + "accuracy": 0.9619138805489791, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -94925,7 +94925,7 @@ } }, { - "accuracy": 0.9634244598840412, + "accuracy": 0.9636140214769464, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -94974,7 +94974,7 @@ } }, { - "accuracy": 0.9772009896604639, + "accuracy": 0.9773109461131849, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -95026,7 +95026,7 @@ } }, { - "accuracy": 0.9789809019942033, + "accuracy": 0.9790893557824587, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -95078,7 +95078,7 @@ } }, { - "accuracy": 0.9814607940222088, + "accuracy": 0.9815537992276644, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -95127,7 +95127,7 @@ } }, { - "accuracy": 0.9883461084805036, + "accuracy": 0.9883982311738165, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -95170,7 +95170,7 @@ } }, { - "accuracy": 0.9892965798315249, + "accuracy": 0.9893504608618585, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -95213,7 +95213,7 @@ } }, { - "accuracy": 0.9884168231173566, + "accuracy": 0.9884674917710455, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -95265,7 +95265,7 @@ } }, { - "accuracy": 0.9898004014241067, + "accuracy": 0.9898517657267419, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -95317,7 +95317,7 @@ } }, { - "accuracy": 0.9940976516196602, + "accuracy": 0.9941205911730465, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -95369,7 +95369,7 @@ } }, { - "accuracy": 0.9949105691752935, + "accuracy": 0.9949368489416022, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -95421,7 +95421,7 @@ } }, { - "accuracy": 0.9968364195604074, + "accuracy": 0.996849293771543, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -95464,7 +95464,7 @@ } }, { - "accuracy": 0.9969747113554102, + "accuracy": 0.9969855123444608, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -95513,7 +95513,7 @@ } }, { - "accuracy": 0.9974387250840664, + "accuracy": 0.9974460572396454, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -95559,7 +95559,7 @@ } }, { - "accuracy": 0.9990936500069342, + "accuracy": 0.9990976247936487, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -95601,7 +95601,7 @@ ], "model.layers.49.self_attn": [ { - "accuracy": 0.9771806676136819, + "accuracy": 0.9795065098687222, "total_bits": 320757760, "q_proj": { "group_size": { @@ -95665,7 +95665,7 @@ } }, { - "accuracy": 0.9797037824204093, + "accuracy": 0.9808972430856604, "total_bits": 329080832, "q_proj": { "group_size": { @@ -95729,7 +95729,7 @@ } }, { - "accuracy": 0.9856564857457814, + "accuracy": 0.9856021827773044, "total_bits": 336024576, "q_proj": { "group_size": { @@ -95793,7 +95793,7 @@ } }, { - "accuracy": 0.9881385823613719, + "accuracy": 0.9880856722593307, "total_bits": 401557504, "q_proj": { "group_size": { @@ -95857,7 +95857,7 @@ } }, { - "accuracy": 0.9896699174454338, + "accuracy": 0.9895824962540677, "total_bits": 475279360, "q_proj": { "group_size": { @@ -95921,7 +95921,7 @@ } }, { - "accuracy": 0.990174620559341, + "accuracy": 0.9903511914767718, "total_bits": 475479040, "q_proj": { "group_size": { @@ -95985,7 +95985,7 @@ } }, { - "accuracy": 0.9928828626871109, + "accuracy": 0.9928882239680541, "total_bits": 609759232, "q_proj": { "group_size": { @@ -96037,7 +96037,7 @@ } }, { - "accuracy": 0.9938027745014743, + "accuracy": 0.9934174912540537, "total_bits": 610024448, "q_proj": { "group_size": { @@ -96089,7 +96089,7 @@ } }, { - "accuracy": 0.9940999126748035, + "accuracy": 0.9941077499013198, "total_bits": 615020544, "q_proj": { "group_size": { @@ -96141,7 +96141,7 @@ } }, { - "accuracy": 0.9942782489876998, + "accuracy": 0.9942726010554715, "total_bits": 623951872, "q_proj": { "group_size": { @@ -96193,7 +96193,7 @@ } }, { - "accuracy": 0.99421489709302, + "accuracy": 0.9943881030929717, "total_bits": 626473984, "q_proj": { "group_size": { @@ -96257,7 +96257,7 @@ } }, { - "accuracy": 0.9945666570412485, + "accuracy": 0.9946412432350611, "total_bits": 630355968, "q_proj": { "group_size": { @@ -96321,7 +96321,7 @@ } }, { - "accuracy": 0.9957767093652173, + "accuracy": 0.9957740138235845, "total_bits": 637362176, "q_proj": { "group_size": { @@ -96382,7 +96382,7 @@ } }, { - "accuracy": 0.9960975548938701, + "accuracy": 0.9961561764541426, "total_bits": 646823936, "q_proj": { "group_size": { @@ -96443,7 +96443,7 @@ } }, { - "accuracy": 0.9976410271697923, + "accuracy": 0.9976760052928799, "total_bits": 784740352, "q_proj": { "group_size": { @@ -96504,7 +96504,7 @@ } }, { - "accuracy": 0.9979071775941473, + "accuracy": 0.9979555091183436, "total_bits": 797818880, "q_proj": { "group_size": { @@ -96565,7 +96565,7 @@ } }, { - "accuracy": 0.9979646084339995, + "accuracy": 0.9979952787490267, "total_bits": 911749120, "q_proj": { "group_size": { @@ -96617,7 +96617,7 @@ } }, { - "accuracy": 0.9991523704834675, + "accuracy": 0.9991562793913641, "total_bits": 942718976, "q_proj": { "group_size": { @@ -96669,7 +96669,7 @@ } }, { - "accuracy": 0.999430173566859, + "accuracy": 0.9994309222031581, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -96723,7 +96723,7 @@ ], "model.layers.49.mlp": [ { - "accuracy": 0.9549214024292796, + "accuracy": 0.955123992342698, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -96775,7 +96775,7 @@ } }, { - "accuracy": 0.9560443100176359, + "accuracy": 0.9562490802062185, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -96827,7 +96827,7 @@ } }, { - "accuracy": 0.96182907568781, + "accuracy": 0.9620112337564167, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -96876,7 +96876,7 @@ } }, { - "accuracy": 0.9634750203082436, + "accuracy": 0.9636515567177221, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -96925,7 +96925,7 @@ } }, { - "accuracy": 0.9773181080818176, + "accuracy": 0.9774342540063357, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -96977,7 +96977,7 @@ } }, { - "accuracy": 0.979092604235599, + "accuracy": 0.9791921753632394, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -97029,7 +97029,7 @@ } }, { - "accuracy": 0.9815000154470143, + "accuracy": 0.9815866555038252, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -97078,7 +97078,7 @@ } }, { - "accuracy": 0.9884053751042015, + "accuracy": 0.9884624402774008, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -97121,7 +97121,7 @@ } }, { - "accuracy": 0.9893499840247003, + "accuracy": 0.9893997374333834, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -97164,7 +97164,7 @@ } }, { - "accuracy": 0.9884731534280276, + "accuracy": 0.988530075863788, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -97216,7 +97216,7 @@ } }, { - "accuracy": 0.989851229284939, + "accuracy": 0.9899011544491115, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -97268,7 +97268,7 @@ } }, { - "accuracy": 0.9941323101520538, + "accuracy": 0.9941611964451639, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -97320,7 +97320,7 @@ } }, { - "accuracy": 0.9949404867856126, + "accuracy": 0.9949655423038885, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -97372,7 +97372,7 @@ } }, { - "accuracy": 0.996862114652207, + "accuracy": 0.9968779104712763, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -97415,7 +97415,7 @@ } }, { - "accuracy": 0.9969879696635824, + "accuracy": 0.9970021624314157, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -97464,7 +97464,7 @@ } }, { - "accuracy": 0.9974364040321425, + "accuracy": 0.9974485065199827, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -97510,7 +97510,7 @@ } }, { - "accuracy": 0.9990882270626331, + "accuracy": 0.9990918166739376, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -97552,7 +97552,7 @@ ], "model.layers.50.self_attn": [ { - "accuracy": 0.9831345489150599, + "accuracy": 0.9848041863817918, "total_bits": 320757760, "q_proj": { "group_size": { @@ -97616,7 +97616,7 @@ } }, { - "accuracy": 0.9836885458544681, + "accuracy": 0.98433034984689, "total_bits": 329080832, "q_proj": { "group_size": { @@ -97680,7 +97680,7 @@ } }, { - "accuracy": 0.989333199043023, + "accuracy": 0.9892950693243429, "total_bits": 336024576, "q_proj": { "group_size": { @@ -97744,7 +97744,7 @@ } }, { - "accuracy": 0.9908435101571836, + "accuracy": 0.9906208060289684, "total_bits": 401557504, "q_proj": { "group_size": { @@ -97808,7 +97808,7 @@ } }, { - "accuracy": 0.9916029682284907, + "accuracy": 0.9913997540348455, "total_bits": 475279360, "q_proj": { "group_size": { @@ -97872,7 +97872,7 @@ } }, { - "accuracy": 0.9922801295393392, + "accuracy": 0.9922376612299367, "total_bits": 475479040, "q_proj": { "group_size": { @@ -97936,7 +97936,7 @@ } }, { - "accuracy": 0.9940976641680065, + "accuracy": 0.9936363908805346, "total_bits": 609759232, "q_proj": { "group_size": { @@ -97988,7 +97988,7 @@ } }, { - "accuracy": 0.9942952892497966, + "accuracy": 0.9945472322012249, "total_bits": 610024448, "q_proj": { "group_size": { @@ -98040,7 +98040,7 @@ } }, { - "accuracy": 0.9948650037771777, + "accuracy": 0.9950141075410341, "total_bits": 615020544, "q_proj": { "group_size": { @@ -98092,7 +98092,7 @@ } }, { - "accuracy": 0.9949100092053413, + "accuracy": 0.9950730212424931, "total_bits": 623951872, "q_proj": { "group_size": { @@ -98144,7 +98144,7 @@ } }, { - "accuracy": 0.9949714855143899, + "accuracy": 0.9948739985886373, "total_bits": 626473984, "q_proj": { "group_size": { @@ -98208,7 +98208,7 @@ } }, { - "accuracy": 0.9950751403444692, + "accuracy": 0.995385638977352, "total_bits": 630355968, "q_proj": { "group_size": { @@ -98272,7 +98272,7 @@ } }, { - "accuracy": 0.9962984303894796, + "accuracy": 0.996289860261114, "total_bits": 637362176, "q_proj": { "group_size": { @@ -98333,7 +98333,7 @@ } }, { - "accuracy": 0.9967387597027578, + "accuracy": 0.9966834760025928, "total_bits": 646823936, "q_proj": { "group_size": { @@ -98394,7 +98394,7 @@ } }, { - "accuracy": 0.9979673255431024, + "accuracy": 0.9980179877265504, "total_bits": 784740352, "q_proj": { "group_size": { @@ -98455,7 +98455,7 @@ } }, { - "accuracy": 0.998331021909651, + "accuracy": 0.9983252894721533, "total_bits": 797818880, "q_proj": { "group_size": { @@ -98516,7 +98516,7 @@ } }, { - "accuracy": 0.9981099716142604, + "accuracy": 0.998160220682621, "total_bits": 911749120, "q_proj": { "group_size": { @@ -98568,7 +98568,7 @@ } }, { - "accuracy": 0.9993757746721569, + "accuracy": 0.9993665535983286, "total_bits": 942718976, "q_proj": { "group_size": { @@ -98620,7 +98620,7 @@ } }, { - "accuracy": 0.9994893842621854, + "accuracy": 0.9994877917005828, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -98674,7 +98674,7 @@ ], "model.layers.50.mlp": [ { - "accuracy": 0.9553957769745275, + "accuracy": 0.9555993550702145, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -98726,7 +98726,7 @@ } }, { - "accuracy": 0.9564977162762692, + "accuracy": 0.9567010904613295, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -98778,7 +98778,7 @@ } }, { - "accuracy": 0.962271859771327, + "accuracy": 0.9624442928715756, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -98827,7 +98827,7 @@ } }, { - "accuracy": 0.9638827819573252, + "accuracy": 0.9640467166900635, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -98876,7 +98876,7 @@ } }, { - "accuracy": 0.9775698890811518, + "accuracy": 0.9776747744334372, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -98928,7 +98928,7 @@ } }, { - "accuracy": 0.9793196336219185, + "accuracy": 0.9794128490121741, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -98980,7 +98980,7 @@ } }, { - "accuracy": 0.9817033112049103, + "accuracy": 0.9817852613173033, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -99029,7 +99029,7 @@ } }, { - "accuracy": 0.988556729335534, + "accuracy": 0.9886085359673751, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -99072,7 +99072,7 @@ } }, { - "accuracy": 0.9894829549287495, + "accuracy": 0.9895326369687131, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -99115,7 +99115,7 @@ } }, { - "accuracy": 0.9885949437555513, + "accuracy": 0.9886450438123, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -99167,7 +99167,7 @@ } }, { - "accuracy": 0.989955197823675, + "accuracy": 0.9900018502222864, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -99219,7 +99219,7 @@ } }, { - "accuracy": 0.9941960800635187, + "accuracy": 0.9942203818967468, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -99271,7 +99271,7 @@ } }, { - "accuracy": 0.9949958277376074, + "accuracy": 0.9950199613445684, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -99323,7 +99323,7 @@ } }, { - "accuracy": 0.9969136916885251, + "accuracy": 0.9969268913723921, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -99366,7 +99366,7 @@ } }, { - "accuracy": 0.997035488290222, + "accuracy": 0.9970477462599152, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -99415,7 +99415,7 @@ } }, { - "accuracy": 0.9974796524957606, + "accuracy": 0.9974897017604426, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -99461,7 +99461,7 @@ } }, { - "accuracy": 0.9991711442799944, + "accuracy": 0.999175162691819, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -99503,7 +99503,7 @@ ], "model.layers.51.self_attn": [ { - "accuracy": 0.9837285076317034, + "accuracy": 0.9843988747973191, "total_bits": 320757760, "q_proj": { "group_size": { @@ -99567,7 +99567,7 @@ } }, { - "accuracy": 0.985254770831058, + "accuracy": 0.9856640947492499, "total_bits": 329080832, "q_proj": { "group_size": { @@ -99631,7 +99631,7 @@ } }, { - "accuracy": 0.9867942333221436, + "accuracy": 0.9870009704640037, "total_bits": 336024576, "q_proj": { "group_size": { @@ -99695,7 +99695,7 @@ } }, { - "accuracy": 0.9887226051405856, + "accuracy": 0.9888954703745089, "total_bits": 401557504, "q_proj": { "group_size": { @@ -99759,7 +99759,7 @@ } }, { - "accuracy": 0.9902999173653754, + "accuracy": 0.9902932181170112, "total_bits": 475279360, "q_proj": { "group_size": { @@ -99823,7 +99823,7 @@ } }, { - "accuracy": 0.9902898473174948, + "accuracy": 0.9906027928779, "total_bits": 475479040, "q_proj": { "group_size": { @@ -99887,7 +99887,7 @@ } }, { - "accuracy": 0.992943383361164, + "accuracy": 0.9928822070360184, "total_bits": 609759232, "q_proj": { "group_size": { @@ -99939,7 +99939,7 @@ } }, { - "accuracy": 0.9933192463297593, + "accuracy": 0.9935516103317863, "total_bits": 610024448, "q_proj": { "group_size": { @@ -99991,7 +99991,7 @@ } }, { - "accuracy": 0.9938842905195135, + "accuracy": 0.993758440409836, "total_bits": 615020544, "q_proj": { "group_size": { @@ -100043,7 +100043,7 @@ } }, { - "accuracy": 0.9939678342718827, + "accuracy": 0.9938343684924277, "total_bits": 623951872, "q_proj": { "group_size": { @@ -100095,7 +100095,7 @@ } }, { - "accuracy": 0.9940933608695081, + "accuracy": 0.9940696370444799, "total_bits": 626473984, "q_proj": { "group_size": { @@ -100159,7 +100159,7 @@ } }, { - "accuracy": 0.9942922501971847, + "accuracy": 0.9943441085909542, "total_bits": 630355968, "q_proj": { "group_size": { @@ -100223,7 +100223,7 @@ } }, { - "accuracy": 0.9957079950131869, + "accuracy": 0.995740425821982, "total_bits": 637362176, "q_proj": { "group_size": { @@ -100284,7 +100284,7 @@ } }, { - "accuracy": 0.9960582566104437, + "accuracy": 0.9960718582335272, "total_bits": 646823936, "q_proj": { "group_size": { @@ -100345,7 +100345,7 @@ } }, { - "accuracy": 0.9975685785083395, + "accuracy": 0.9976055627982867, "total_bits": 784740352, "q_proj": { "group_size": { @@ -100406,7 +100406,7 @@ } }, { - "accuracy": 0.9979770752160173, + "accuracy": 0.9980227055126115, "total_bits": 797818880, "q_proj": { "group_size": { @@ -100467,7 +100467,7 @@ } }, { - "accuracy": 0.9977616665786818, + "accuracy": 0.997799910212818, "total_bits": 911749120, "q_proj": { "group_size": { @@ -100519,7 +100519,7 @@ } }, { - "accuracy": 0.9992050155507106, + "accuracy": 0.9992067832499743, "total_bits": 942718976, "q_proj": { "group_size": { @@ -100571,7 +100571,7 @@ } }, { - "accuracy": 0.9993939664410917, + "accuracy": 0.9994087753033167, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -100625,7 +100625,7 @@ ], "model.layers.51.mlp": [ { - "accuracy": 0.9548031970074302, + "accuracy": 0.9550098306254337, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -100677,7 +100677,7 @@ } }, { - "accuracy": 0.9559393117302343, + "accuracy": 0.9561306645995692, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -100729,7 +100729,7 @@ } }, { - "accuracy": 0.9617459742646468, + "accuracy": 0.9619123747474269, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -100778,7 +100778,7 @@ } }, { - "accuracy": 0.9633846565296775, + "accuracy": 0.9635423327747145, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -100827,7 +100827,7 @@ } }, { - "accuracy": 0.9772648795654899, + "accuracy": 0.9773615128115604, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -100879,7 +100879,7 @@ } }, { - "accuracy": 0.979033721120734, + "accuracy": 0.9791261287111985, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -100931,7 +100931,7 @@ } }, { - "accuracy": 0.9814436608239224, + "accuracy": 0.9815250431236467, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -100980,7 +100980,7 @@ } }, { - "accuracy": 0.9883787004571212, + "accuracy": 0.9884330708729593, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -101023,7 +101023,7 @@ } }, { - "accuracy": 0.989325546904614, + "accuracy": 0.9893712801368613, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -101066,7 +101066,7 @@ } }, { - "accuracy": 0.9884342410062489, + "accuracy": 0.9884860868516722, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -101118,7 +101118,7 @@ } }, { - "accuracy": 0.9898153817967364, + "accuracy": 0.9898602813482285, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -101170,7 +101170,7 @@ } }, { - "accuracy": 0.9941082424239108, + "accuracy": 0.9941351272557911, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -101222,7 +101222,7 @@ } }, { - "accuracy": 0.9949190291134935, + "accuracy": 0.9949421682640126, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -101274,7 +101274,7 @@ } }, { - "accuracy": 0.9968493210249826, + "accuracy": 0.9968627952039242, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -101317,7 +101317,7 @@ } }, { - "accuracy": 0.9969899709287443, + "accuracy": 0.9970049352238053, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -101366,7 +101366,7 @@ } }, { - "accuracy": 0.9974413557272208, + "accuracy": 0.997454121512802, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -101412,7 +101412,7 @@ } }, { - "accuracy": 0.9991385486761206, + "accuracy": 0.9991422003429187, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -101454,7 +101454,7 @@ ], "model.layers.52.self_attn": [ { - "accuracy": 0.9679202594255146, + "accuracy": 0.9674007327933061, "total_bits": 320757760, "q_proj": { "group_size": { @@ -101518,7 +101518,7 @@ } }, { - "accuracy": 0.9714520918695551, + "accuracy": 0.9713884435201946, "total_bits": 329080832, "q_proj": { "group_size": { @@ -101582,7 +101582,7 @@ } }, { - "accuracy": 0.9788006261775368, + "accuracy": 0.9791521586869892, "total_bits": 336024576, "q_proj": { "group_size": { @@ -101646,7 +101646,7 @@ } }, { - "accuracy": 0.9822152696157757, + "accuracy": 0.9826102601854425, "total_bits": 401557504, "q_proj": { "group_size": { @@ -101710,7 +101710,7 @@ } }, { - "accuracy": 0.9854190914254439, + "accuracy": 0.9856055974960327, "total_bits": 475279360, "q_proj": { "group_size": { @@ -101774,7 +101774,7 @@ } }, { - "accuracy": 0.9859513671774613, + "accuracy": 0.985610864664379, "total_bits": 475479040, "q_proj": { "group_size": { @@ -101838,7 +101838,7 @@ } }, { - "accuracy": 0.9910903871059418, + "accuracy": 0.9913786453636069, "total_bits": 609759232, "q_proj": { "group_size": { @@ -101890,7 +101890,7 @@ } }, { - "accuracy": 0.9916917752278479, + "accuracy": 0.9917869112993541, "total_bits": 610024448, "q_proj": { "group_size": { @@ -101942,7 +101942,7 @@ } }, { - "accuracy": 0.9922006679208655, + "accuracy": 0.9921901798562, "total_bits": 615020544, "q_proj": { "group_size": { @@ -101994,7 +101994,7 @@ } }, { - "accuracy": 0.9924397296027133, + "accuracy": 0.9924446720825998, "total_bits": 623951872, "q_proj": { "group_size": { @@ -102046,7 +102046,7 @@ } }, { - "accuracy": 0.9925545995172701, + "accuracy": 0.9926101996710426, "total_bits": 626473984, "q_proj": { "group_size": { @@ -102110,7 +102110,7 @@ } }, { - "accuracy": 0.9931570338575464, + "accuracy": 0.993257180640572, "total_bits": 630355968, "q_proj": { "group_size": { @@ -102174,7 +102174,7 @@ } }, { - "accuracy": 0.9946699636547189, + "accuracy": 0.99472697236036, "total_bits": 637362176, "q_proj": { "group_size": { @@ -102235,7 +102235,7 @@ } }, { - "accuracy": 0.9950944800910196, + "accuracy": 0.9950658098647469, "total_bits": 646823936, "q_proj": { "group_size": { @@ -102296,7 +102296,7 @@ } }, { - "accuracy": 0.9970547201994219, + "accuracy": 0.9970999462039847, "total_bits": 784740352, "q_proj": { "group_size": { @@ -102357,7 +102357,7 @@ } }, { - "accuracy": 0.9975504443833703, + "accuracy": 0.9975253672976243, "total_bits": 797818880, "q_proj": { "group_size": { @@ -102418,7 +102418,7 @@ } }, { - "accuracy": 0.9975999046705271, + "accuracy": 0.9976515203322235, "total_bits": 911749120, "q_proj": { "group_size": { @@ -102470,7 +102470,7 @@ } }, { - "accuracy": 0.998747704256522, + "accuracy": 0.9987561321376186, "total_bits": 942718976, "q_proj": { "group_size": { @@ -102522,7 +102522,7 @@ } }, { - "accuracy": 0.9993591631615633, + "accuracy": 0.9993684087438803, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -102576,7 +102576,7 @@ ], "model.layers.52.mlp": [ { - "accuracy": 0.9541540208615755, + "accuracy": 0.9543598388370714, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -102628,7 +102628,7 @@ } }, { - "accuracy": 0.955306162959651, + "accuracy": 0.9555064690740485, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -102680,7 +102680,7 @@ } }, { - "accuracy": 0.9611707329750061, + "accuracy": 0.9613294381844371, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -102729,7 +102729,7 @@ } }, { - "accuracy": 0.9628563774259467, + "accuracy": 0.9630061921320463, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -102778,7 +102778,7 @@ } }, { - "accuracy": 0.9769216054364255, + "accuracy": 0.9770250116523943, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -102830,7 +102830,7 @@ } }, { - "accuracy": 0.9787252827694541, + "accuracy": 0.978819246354856, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -102882,7 +102882,7 @@ } }, { - "accuracy": 0.9811784107434122, + "accuracy": 0.9812597553980978, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -102931,7 +102931,7 @@ } }, { - "accuracy": 0.9881916783357921, + "accuracy": 0.9882422983646393, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -102974,7 +102974,7 @@ } }, { - "accuracy": 0.9891607902551952, + "accuracy": 0.989210351517326, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -103017,7 +103017,7 @@ } }, { - "accuracy": 0.9882620384818629, + "accuracy": 0.988316358704316, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -103069,7 +103069,7 @@ } }, { - "accuracy": 0.9896765806173023, + "accuracy": 0.9897231828225287, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -103121,7 +103121,7 @@ } }, { - "accuracy": 0.994015423875106, + "accuracy": 0.9940443999673191, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -103173,7 +103173,7 @@ } }, { - "accuracy": 0.9948441264660735, + "accuracy": 0.9948670456283971, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -103225,7 +103225,7 @@ } }, { - "accuracy": 0.9967828283184453, + "accuracy": 0.996798570414907, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -103268,7 +103268,7 @@ } }, { - "accuracy": 0.9969238927097697, + "accuracy": 0.9969373157149867, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -103317,7 +103317,7 @@ } }, { - "accuracy": 0.9973813571820134, + "accuracy": 0.9973924481555035, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -103363,7 +103363,7 @@ } }, { - "accuracy": 0.9990573236228604, + "accuracy": 0.9990608032400671, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -103405,7 +103405,7 @@ ], "model.layers.53.self_attn": [ { - "accuracy": 0.9772020452900937, + "accuracy": 0.9790887832641602, "total_bits": 320757760, "q_proj": { "group_size": { @@ -103469,7 +103469,7 @@ } }, { - "accuracy": 0.980516603118495, + "accuracy": 0.9796983458493885, "total_bits": 329080832, "q_proj": { "group_size": { @@ -103533,7 +103533,7 @@ } }, { - "accuracy": 0.9853213554934451, + "accuracy": 0.9852609399117922, "total_bits": 336024576, "q_proj": { "group_size": { @@ -103597,7 +103597,7 @@ } }, { - "accuracy": 0.9873996925981421, + "accuracy": 0.9873411365245518, "total_bits": 401557504, "q_proj": { "group_size": { @@ -103661,7 +103661,7 @@ } }, { - "accuracy": 0.9889684523406782, + "accuracy": 0.9889142560331445, "total_bits": 475279360, "q_proj": { "group_size": { @@ -103725,7 +103725,7 @@ } }, { - "accuracy": 0.9896289518004969, + "accuracy": 0.9890722404969366, "total_bits": 475479040, "q_proj": { "group_size": { @@ -103789,7 +103789,7 @@ } }, { - "accuracy": 0.993024251962963, + "accuracy": 0.9931079392370424, "total_bits": 609759232, "q_proj": { "group_size": { @@ -103841,7 +103841,7 @@ } }, { - "accuracy": 0.9933966837431255, + "accuracy": 0.9936127439141273, "total_bits": 610024448, "q_proj": { "group_size": { @@ -103893,7 +103893,7 @@ } }, { - "accuracy": 0.9939692702732588, + "accuracy": 0.9939858007587885, "total_bits": 615020544, "q_proj": { "group_size": { @@ -103945,7 +103945,7 @@ } }, { - "accuracy": 0.9941108050315004, + "accuracy": 0.9941161870956421, "total_bits": 623951872, "q_proj": { "group_size": { @@ -103997,7 +103997,7 @@ } }, { - "accuracy": 0.9939401212491488, + "accuracy": 0.9938848716648001, "total_bits": 626473984, "q_proj": { "group_size": { @@ -104061,7 +104061,7 @@ } }, { - "accuracy": 0.9943710706735912, + "accuracy": 0.9946469048920431, "total_bits": 630355968, "q_proj": { "group_size": { @@ -104125,7 +104125,7 @@ } }, { - "accuracy": 0.9957380941823909, + "accuracy": 0.9958325122532091, "total_bits": 637362176, "q_proj": { "group_size": { @@ -104186,7 +104186,7 @@ } }, { - "accuracy": 0.9960697532484406, + "accuracy": 0.9961259176856593, "total_bits": 646823936, "q_proj": { "group_size": { @@ -104247,7 +104247,7 @@ } }, { - "accuracy": 0.9976370999295461, + "accuracy": 0.9975916500154295, "total_bits": 784740352, "q_proj": { "group_size": { @@ -104308,7 +104308,7 @@ } }, { - "accuracy": 0.9979898813915881, + "accuracy": 0.9979977327349939, "total_bits": 797818880, "q_proj": { "group_size": { @@ -104369,7 +104369,7 @@ } }, { - "accuracy": 0.9979332622728849, + "accuracy": 0.9978728910025797, "total_bits": 911749120, "q_proj": { "group_size": { @@ -104421,7 +104421,7 @@ } }, { - "accuracy": 0.9991882159129569, + "accuracy": 0.9991773055180123, "total_bits": 942718976, "q_proj": { "group_size": { @@ -104473,7 +104473,7 @@ } }, { - "accuracy": 0.9994185047322198, + "accuracy": 0.999419099749311, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -104527,7 +104527,7 @@ ], "model.layers.53.mlp": [ { - "accuracy": 0.9547754588880037, + "accuracy": 0.9549691206530521, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -104579,7 +104579,7 @@ } }, { - "accuracy": 0.9559084490725869, + "accuracy": 0.956084038081922, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -104631,7 +104631,7 @@ } }, { - "accuracy": 0.9616870880126953, + "accuracy": 0.9618426875064248, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -104680,7 +104680,7 @@ } }, { - "accuracy": 0.9633228841580843, + "accuracy": 0.963468284983384, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -104729,7 +104729,7 @@ } }, { - "accuracy": 0.9772479502778304, + "accuracy": 0.9773458211045516, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -104781,7 +104781,7 @@ } }, { - "accuracy": 0.979005178338603, + "accuracy": 0.9790968847902197, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -104833,7 +104833,7 @@ } }, { - "accuracy": 0.9814082588020124, + "accuracy": 0.9814878482567636, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -104882,7 +104882,7 @@ } }, { - "accuracy": 0.9883726599969362, + "accuracy": 0.9884265535756162, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -104925,7 +104925,7 @@ } }, { - "accuracy": 0.9893131177676352, + "accuracy": 0.9893600336815181, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -104968,7 +104968,7 @@ } }, { - "accuracy": 0.988435644068216, + "accuracy": 0.9884879840047736, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -105020,7 +105020,7 @@ } }, { - "accuracy": 0.9898058113298918, + "accuracy": 0.9898481690569928, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -105072,7 +105072,7 @@ } }, { - "accuracy": 0.9941091808049303, + "accuracy": 0.9941365475717344, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -105124,7 +105124,7 @@ } }, { - "accuracy": 0.9949111797307667, + "accuracy": 0.9949318178390202, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -105176,7 +105176,7 @@ } }, { - "accuracy": 0.9968481101095676, + "accuracy": 0.9968612090145287, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -105219,7 +105219,7 @@ } }, { - "accuracy": 0.99697390159494, + "accuracy": 0.9969855662631361, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -105268,7 +105268,7 @@ } }, { - "accuracy": 0.997420103142136, + "accuracy": 0.9974302292654389, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -105314,7 +105314,7 @@ } }, { - "accuracy": 0.9990737793084822, + "accuracy": 0.9990774264657184, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -105356,7 +105356,7 @@ ], "model.layers.54.self_attn": [ { - "accuracy": 0.9852072376953928, + "accuracy": 0.9845947293858779, "total_bits": 320757760, "q_proj": { "group_size": { @@ -105420,7 +105420,7 @@ } }, { - "accuracy": 0.9865362957904213, + "accuracy": 0.9861050075606296, "total_bits": 329080832, "q_proj": { "group_size": { @@ -105484,7 +105484,7 @@ } }, { - "accuracy": 0.9883846311192763, + "accuracy": 0.9883411942343963, "total_bits": 336024576, "q_proj": { "group_size": { @@ -105548,7 +105548,7 @@ } }, { - "accuracy": 0.9908759288097683, + "accuracy": 0.9908029837043661, "total_bits": 401557504, "q_proj": { "group_size": { @@ -105612,7 +105612,7 @@ } }, { - "accuracy": 0.992276132106781, + "accuracy": 0.9921458936051318, "total_bits": 475279360, "q_proj": { "group_size": { @@ -105676,7 +105676,7 @@ } }, { - "accuracy": 0.9926291813975886, + "accuracy": 0.9925032861922917, "total_bits": 475479040, "q_proj": { "group_size": { @@ -105740,7 +105740,7 @@ } }, { - "accuracy": 0.9938453086896947, + "accuracy": 0.9942021773833978, "total_bits": 609759232, "q_proj": { "group_size": { @@ -105792,7 +105792,7 @@ } }, { - "accuracy": 0.9945389675466638, + "accuracy": 0.994654359786134, "total_bits": 610024448, "q_proj": { "group_size": { @@ -105844,7 +105844,7 @@ } }, { - "accuracy": 0.9950207381656295, + "accuracy": 0.995129391158882, "total_bits": 615020544, "q_proj": { "group_size": { @@ -105896,7 +105896,7 @@ } }, { - "accuracy": 0.9951851709108603, + "accuracy": 0.9952726415113399, "total_bits": 623951872, "q_proj": { "group_size": { @@ -105948,7 +105948,7 @@ } }, { - "accuracy": 0.9952885500694576, + "accuracy": 0.995423333033135, "total_bits": 626473984, "q_proj": { "group_size": { @@ -106012,7 +106012,7 @@ } }, { - "accuracy": 0.9956268581904864, + "accuracy": 0.9956734294169828, "total_bits": 630355968, "q_proj": { "group_size": { @@ -106076,7 +106076,7 @@ } }, { - "accuracy": 0.9962920499475378, + "accuracy": 0.9963827544921323, "total_bits": 637362176, "q_proj": { "group_size": { @@ -106137,7 +106137,7 @@ } }, { - "accuracy": 0.9965613507910779, + "accuracy": 0.9966287671735412, "total_bits": 646823936, "q_proj": { "group_size": { @@ -106198,7 +106198,7 @@ } }, { - "accuracy": 0.997813479876832, + "accuracy": 0.9978458928434473, "total_bits": 784740352, "q_proj": { "group_size": { @@ -106259,7 +106259,7 @@ } }, { - "accuracy": 0.9981937337862817, + "accuracy": 0.9982172911496539, "total_bits": 797818880, "q_proj": { "group_size": { @@ -106320,7 +106320,7 @@ } }, { - "accuracy": 0.9980638884007931, + "accuracy": 0.9980932403551904, "total_bits": 911749120, "q_proj": { "group_size": { @@ -106372,7 +106372,7 @@ } }, { - "accuracy": 0.9991730873129869, + "accuracy": 0.9991833168130956, "total_bits": 942718976, "q_proj": { "group_size": { @@ -106424,7 +106424,7 @@ } }, { - "accuracy": 0.9994705375283957, + "accuracy": 0.9994833386552177, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -106478,7 +106478,7 @@ ], "model.layers.54.mlp": [ { - "accuracy": 0.9552012054543746, + "accuracy": 0.9553918587534052, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -106530,7 +106530,7 @@ } }, { - "accuracy": 0.9563162985600924, + "accuracy": 0.9565090003766512, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -106582,7 +106582,7 @@ } }, { - "accuracy": 0.9620732916028876, + "accuracy": 0.962232276013023, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -106631,7 +106631,7 @@ } }, { - "accuracy": 0.9636660343722293, + "accuracy": 0.9638158428041559, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -106680,7 +106680,7 @@ } }, { - "accuracy": 0.9774715241632963, + "accuracy": 0.9775658955699519, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -106732,7 +106732,7 @@ } }, { - "accuracy": 0.9792262613773346, + "accuracy": 0.9793148950526589, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -106784,7 +106784,7 @@ } }, { - "accuracy": 0.9816080818050786, + "accuracy": 0.9816853168763613, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -106833,7 +106833,7 @@ } }, { - "accuracy": 0.9885208928271344, + "accuracy": 0.9885703199788144, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -106876,7 +106876,7 @@ } }, { - "accuracy": 0.989464346515505, + "accuracy": 0.9895095864408895, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -106919,7 +106919,7 @@ } }, { - "accuracy": 0.9885608334290353, + "accuracy": 0.9886092778883482, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -106971,7 +106971,7 @@ } }, { - "accuracy": 0.9899310767650604, + "accuracy": 0.9899744618880121, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -107023,7 +107023,7 @@ } }, { - "accuracy": 0.9941827850906473, + "accuracy": 0.9942076378747037, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -107075,7 +107075,7 @@ } }, { - "accuracy": 0.9949828778442583, + "accuracy": 0.9950037159417805, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -107127,7 +107127,7 @@ } }, { - "accuracy": 0.9969016989987147, + "accuracy": 0.996915195725466, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -107170,7 +107170,7 @@ } }, { - "accuracy": 0.9970176061125178, + "accuracy": 0.9970303754273214, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -107219,7 +107219,7 @@ } }, { - "accuracy": 0.9974606807686781, + "accuracy": 0.9974714143103675, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -107265,7 +107265,7 @@ } }, { - "accuracy": 0.9991186066089492, + "accuracy": 0.9991222595501887, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -107307,7 +107307,7 @@ ], "model.layers.55.self_attn": [ { - "accuracy": 0.982862180785129, + "accuracy": 0.9825350501035389, "total_bits": 320757760, "q_proj": { "group_size": { @@ -107371,7 +107371,7 @@ } }, { - "accuracy": 0.9834734201431274, + "accuracy": 0.9832451939582825, "total_bits": 329080832, "q_proj": { "group_size": { @@ -107435,7 +107435,7 @@ } }, { - "accuracy": 0.9859893525901594, + "accuracy": 0.986233441453231, "total_bits": 336024576, "q_proj": { "group_size": { @@ -107499,7 +107499,7 @@ } }, { - "accuracy": 0.9896264719335657, + "accuracy": 0.9898849584554371, "total_bits": 401557504, "q_proj": { "group_size": { @@ -107563,7 +107563,7 @@ } }, { - "accuracy": 0.9907047262317256, + "accuracy": 0.9909536038574419, "total_bits": 475279360, "q_proj": { "group_size": { @@ -107627,7 +107627,7 @@ } }, { - "accuracy": 0.9909745478316357, + "accuracy": 0.9909136914893201, "total_bits": 475479040, "q_proj": { "group_size": { @@ -107691,7 +107691,7 @@ } }, { - "accuracy": 0.9938508366283617, + "accuracy": 0.9938871644829449, "total_bits": 609759232, "q_proj": { "group_size": { @@ -107743,7 +107743,7 @@ } }, { - "accuracy": 0.9942040831634873, + "accuracy": 0.9943071015571293, "total_bits": 610024448, "q_proj": { "group_size": { @@ -107795,7 +107795,7 @@ } }, { - "accuracy": 0.9946965167396947, + "accuracy": 0.9947122233478647, "total_bits": 615020544, "q_proj": { "group_size": { @@ -107847,7 +107847,7 @@ } }, { - "accuracy": 0.9948540682855406, + "accuracy": 0.9948673840416106, "total_bits": 623951872, "q_proj": { "group_size": { @@ -107899,7 +107899,7 @@ } }, { - "accuracy": 0.9947331394804152, + "accuracy": 0.9946535355166385, "total_bits": 626473984, "q_proj": { "group_size": { @@ -107963,7 +107963,7 @@ } }, { - "accuracy": 0.994943239971211, + "accuracy": 0.994991127989794, "total_bits": 630355968, "q_proj": { "group_size": { @@ -108027,7 +108027,7 @@ } }, { - "accuracy": 0.996045686696705, + "accuracy": 0.996066685177778, "total_bits": 637362176, "q_proj": { "group_size": { @@ -108088,7 +108088,7 @@ } }, { - "accuracy": 0.9963499409587759, + "accuracy": 0.9964086315349528, "total_bits": 646823936, "q_proj": { "group_size": { @@ -108149,7 +108149,7 @@ } }, { - "accuracy": 0.9977005700531759, + "accuracy": 0.9977198980356518, "total_bits": 784740352, "q_proj": { "group_size": { @@ -108210,7 +108210,7 @@ } }, { - "accuracy": 0.9981149587976305, + "accuracy": 0.9981073162665492, "total_bits": 797818880, "q_proj": { "group_size": { @@ -108271,7 +108271,7 @@ } }, { - "accuracy": 0.9981531849817226, + "accuracy": 0.9981631730732164, "total_bits": 911749120, "q_proj": { "group_size": { @@ -108323,7 +108323,7 @@ } }, { - "accuracy": 0.9990748936604512, + "accuracy": 0.9990759279186788, "total_bits": 942718976, "q_proj": { "group_size": { @@ -108375,7 +108375,7 @@ } }, { - "accuracy": 0.9993276446077385, + "accuracy": 0.9993379451334476, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -108429,7 +108429,7 @@ ], "model.layers.55.mlp": [ { - "accuracy": 0.9549555935357746, + "accuracy": 0.9551384543117724, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -108481,7 +108481,7 @@ } }, { - "accuracy": 0.9560795551852176, + "accuracy": 0.956251577327126, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -108533,7 +108533,7 @@ } }, { - "accuracy": 0.9617903452170522, + "accuracy": 0.961945725114722, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -108582,7 +108582,7 @@ } }, { - "accuracy": 0.9633887974839461, + "accuracy": 0.9635377369428936, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -108631,7 +108631,7 @@ } }, { - "accuracy": 0.977338969707489, + "accuracy": 0.9774343732156252, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -108683,7 +108683,7 @@ } }, { - "accuracy": 0.9791024939010018, + "accuracy": 0.9791854776834187, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -108735,7 +108735,7 @@ } }, { - "accuracy": 0.98148176544591, + "accuracy": 0.9815538211872703, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -108784,7 +108784,7 @@ } }, { - "accuracy": 0.9884484614196577, + "accuracy": 0.9884931547077078, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -108827,7 +108827,7 @@ } }, { - "accuracy": 0.9893891081998223, + "accuracy": 0.9894301491348367, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -108870,7 +108870,7 @@ } }, { - "accuracy": 0.9884918355628064, + "accuracy": 0.9885368417752417, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -108922,7 +108922,7 @@ } }, { - "accuracy": 0.9898713372255626, + "accuracy": 0.9899108394196159, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -108974,7 +108974,7 @@ } }, { - "accuracy": 0.994148099500882, + "accuracy": 0.9941706033913713, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -109026,7 +109026,7 @@ } }, { - "accuracy": 0.9949499619634528, + "accuracy": 0.9949692326940989, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -109078,7 +109078,7 @@ } }, { - "accuracy": 0.9968728872114107, + "accuracy": 0.9968861431667679, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -109121,7 +109121,7 @@ } }, { - "accuracy": 0.9969885664943018, + "accuracy": 0.9970001686168345, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -109170,7 +109170,7 @@ } }, { - "accuracy": 0.9974300316289851, + "accuracy": 0.9974395526867164, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -109216,7 +109216,7 @@ } }, { - "accuracy": 0.9990677004189867, + "accuracy": 0.999071564231264, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -109258,7 +109258,7 @@ ], "model.layers.56.self_attn": [ { - "accuracy": 0.9710961893985146, + "accuracy": 0.9696619196941978, "total_bits": 320757760, "q_proj": { "group_size": { @@ -109322,7 +109322,7 @@ } }, { - "accuracy": 0.973085880279541, + "accuracy": 0.9726328504712958, "total_bits": 329080832, "q_proj": { "group_size": { @@ -109386,7 +109386,7 @@ } }, { - "accuracy": 0.9801304857981833, + "accuracy": 0.9798208318258587, "total_bits": 336024576, "q_proj": { "group_size": { @@ -109450,7 +109450,7 @@ } }, { - "accuracy": 0.9833614888944124, + "accuracy": 0.9829887465426796, "total_bits": 401557504, "q_proj": { "group_size": { @@ -109514,7 +109514,7 @@ } }, { - "accuracy": 0.9865957467179549, + "accuracy": 0.9863494509144833, "total_bits": 475279360, "q_proj": { "group_size": { @@ -109578,7 +109578,7 @@ } }, { - "accuracy": 0.9868490029322473, + "accuracy": 0.9868397289200833, "total_bits": 475479040, "q_proj": { "group_size": { @@ -109642,7 +109642,7 @@ } }, { - "accuracy": 0.9926328063011169, + "accuracy": 0.9925593318123567, "total_bits": 609759232, "q_proj": { "group_size": { @@ -109694,7 +109694,7 @@ } }, { - "accuracy": 0.9927956940312135, + "accuracy": 0.9930566800268072, "total_bits": 610024448, "q_proj": { "group_size": { @@ -109746,7 +109746,7 @@ } }, { - "accuracy": 0.9934755516679663, + "accuracy": 0.993447880995901, "total_bits": 615020544, "q_proj": { "group_size": { @@ -109798,7 +109798,7 @@ } }, { - "accuracy": 0.9936303119910391, + "accuracy": 0.9936016382355439, "total_bits": 623951872, "q_proj": { "group_size": { @@ -109850,7 +109850,7 @@ } }, { - "accuracy": 0.9932129336030859, + "accuracy": 0.9934567154238099, "total_bits": 626473984, "q_proj": { "group_size": { @@ -109914,7 +109914,7 @@ } }, { - "accuracy": 0.9937804470721044, + "accuracy": 0.9937952729432207, "total_bits": 630355968, "q_proj": { "group_size": { @@ -109978,7 +109978,7 @@ } }, { - "accuracy": 0.9952146014884898, + "accuracy": 0.9952551071581087, "total_bits": 637362176, "q_proj": { "group_size": { @@ -110039,7 +110039,7 @@ } }, { - "accuracy": 0.9956191005675417, + "accuracy": 0.9957154652005747, "total_bits": 646823936, "q_proj": { "group_size": { @@ -110100,7 +110100,7 @@ } }, { - "accuracy": 0.9974246021164092, + "accuracy": 0.9974495986182439, "total_bits": 784740352, "q_proj": { "group_size": { @@ -110161,7 +110161,7 @@ } }, { - "accuracy": 0.9978159709196341, + "accuracy": 0.9978339140744585, "total_bits": 797818880, "q_proj": { "group_size": { @@ -110222,7 +110222,7 @@ } }, { - "accuracy": 0.9980006221877901, + "accuracy": 0.9980273556552435, "total_bits": 911749120, "q_proj": { "group_size": { @@ -110274,7 +110274,7 @@ } }, { - "accuracy": 0.9989508513949419, + "accuracy": 0.9989531586240781, "total_bits": 942718976, "q_proj": { "group_size": { @@ -110326,7 +110326,7 @@ } }, { - "accuracy": 0.999459263672562, + "accuracy": 0.9994602518548307, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -110380,7 +110380,7 @@ ], "model.layers.56.mlp": [ { - "accuracy": 0.9540310596164904, + "accuracy": 0.954205594564739, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -110432,7 +110432,7 @@ } }, { - "accuracy": 0.9551830260377181, + "accuracy": 0.955358740530516, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -110484,7 +110484,7 @@ } }, { - "accuracy": 0.9610013303003813, + "accuracy": 0.9611466401501706, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -110533,7 +110533,7 @@ } }, { - "accuracy": 0.962656181109579, + "accuracy": 0.962796782192431, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -110582,7 +110582,7 @@ } }, { - "accuracy": 0.9768621513718053, + "accuracy": 0.9769497027522639, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -110634,7 +110634,7 @@ } }, { - "accuracy": 0.9786587859454908, + "accuracy": 0.9787395251424689, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -110686,7 +110686,7 @@ } }, { - "accuracy": 0.9810948418943506, + "accuracy": 0.9811664600121347, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -110735,7 +110735,7 @@ } }, { - "accuracy": 0.988176490915449, + "accuracy": 0.9882250914448186, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -110778,7 +110778,7 @@ } }, { - "accuracy": 0.989137202501297, + "accuracy": 0.9891794308235771, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -110821,7 +110821,7 @@ } }, { - "accuracy": 0.9882421705283617, + "accuracy": 0.988289978943373, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -110873,7 +110873,7 @@ } }, { - "accuracy": 0.9896478817651146, + "accuracy": 0.9896879949067768, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -110925,7 +110925,7 @@ } }, { - "accuracy": 0.9940155403394448, + "accuracy": 0.9940396762992206, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -110977,7 +110977,7 @@ } }, { - "accuracy": 0.9948396243547138, + "accuracy": 0.9948594374092001, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -111029,7 +111029,7 @@ } }, { - "accuracy": 0.9968013706568041, + "accuracy": 0.9968159441885195, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -111072,7 +111072,7 @@ } }, { - "accuracy": 0.9969241973993025, + "accuracy": 0.9969363357675703, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -111121,7 +111121,7 @@ } }, { - "accuracy": 0.9973770480014776, + "accuracy": 0.997387124911735, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -111167,7 +111167,7 @@ } }, { - "accuracy": 0.9990655232809091, + "accuracy": 0.9990695991406315, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -111209,7 +111209,7 @@ ], "model.layers.57.self_attn": [ { - "accuracy": 0.9833979653684717, + "accuracy": 0.9833751728660182, "total_bits": 320757760, "q_proj": { "group_size": { @@ -111273,7 +111273,7 @@ } }, { - "accuracy": 0.9842442446633389, + "accuracy": 0.9843801636444894, "total_bits": 329080832, "q_proj": { "group_size": { @@ -111337,7 +111337,7 @@ } }, { - "accuracy": 0.9863859587594083, + "accuracy": 0.9864853291135085, "total_bits": 336024576, "q_proj": { "group_size": { @@ -111401,7 +111401,7 @@ } }, { - "accuracy": 0.9893027465594443, + "accuracy": 0.9893891544718492, "total_bits": 401557504, "q_proj": { "group_size": { @@ -111465,7 +111465,7 @@ } }, { - "accuracy": 0.9906628257349918, + "accuracy": 0.9907292551115939, "total_bits": 475279360, "q_proj": { "group_size": { @@ -111529,7 +111529,7 @@ } }, { - "accuracy": 0.990861617420849, + "accuracy": 0.9910626176156496, "total_bits": 475479040, "q_proj": { "group_size": { @@ -111593,7 +111593,7 @@ } }, { - "accuracy": 0.9934507306469115, + "accuracy": 0.9933420811828814, "total_bits": 609759232, "q_proj": { "group_size": { @@ -111645,7 +111645,7 @@ } }, { - "accuracy": 0.9938584228879527, + "accuracy": 0.9938500849039931, "total_bits": 610024448, "q_proj": { "group_size": { @@ -111697,7 +111697,7 @@ } }, { - "accuracy": 0.99421484454682, + "accuracy": 0.9943515654457243, "total_bits": 615020544, "q_proj": { "group_size": { @@ -111749,7 +111749,7 @@ } }, { - "accuracy": 0.994341435793199, + "accuracy": 0.9944773457552257, "total_bits": 623951872, "q_proj": { "group_size": { @@ -111801,7 +111801,7 @@ } }, { - "accuracy": 0.9944766646153048, + "accuracy": 0.9943860455563194, "total_bits": 626473984, "q_proj": { "group_size": { @@ -111865,7 +111865,7 @@ } }, { - "accuracy": 0.9947260488805018, + "accuracy": 0.9947546857752299, "total_bits": 630355968, "q_proj": { "group_size": { @@ -111929,7 +111929,7 @@ } }, { - "accuracy": 0.9958024660223409, + "accuracy": 0.9958807080984116, "total_bits": 637362176, "q_proj": { "group_size": { @@ -111990,7 +111990,7 @@ } }, { - "accuracy": 0.9960411081188604, + "accuracy": 0.9961550463187067, "total_bits": 646823936, "q_proj": { "group_size": { @@ -112051,7 +112051,7 @@ } }, { - "accuracy": 0.9976178962541254, + "accuracy": 0.9975846160791422, "total_bits": 784740352, "q_proj": { "group_size": { @@ -112112,7 +112112,7 @@ } }, { - "accuracy": 0.9979873552526298, + "accuracy": 0.9979969096419058, "total_bits": 797818880, "q_proj": { "group_size": { @@ -112173,7 +112173,7 @@ } }, { - "accuracy": 0.997938258279311, + "accuracy": 0.9978946602266086, "total_bits": 911749120, "q_proj": { "group_size": { @@ -112225,7 +112225,7 @@ } }, { - "accuracy": 0.9991168804270657, + "accuracy": 0.9991309445743498, "total_bits": 942718976, "q_proj": { "group_size": { @@ -112277,7 +112277,7 @@ } }, { - "accuracy": 0.9994251730018541, + "accuracy": 0.9994316796625131, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -112331,7 +112331,7 @@ ], "model.layers.57.mlp": [ { - "accuracy": 0.9542011649985063, + "accuracy": 0.954383068963101, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -112383,7 +112383,7 @@ } }, { - "accuracy": 0.955358740530516, + "accuracy": 0.955534018968281, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -112435,7 +112435,7 @@ } }, { - "accuracy": 0.9611416302229229, + "accuracy": 0.9612957960680911, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -112484,7 +112484,7 @@ } }, { - "accuracy": 0.9627676668920016, + "accuracy": 0.9629161891184355, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -112533,7 +112533,7 @@ } }, { - "accuracy": 0.9769567313947176, + "accuracy": 0.9770450309703225, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -112585,7 +112585,7 @@ } }, { - "accuracy": 0.9787444221346002, + "accuracy": 0.9788221183576082, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -112637,7 +112637,7 @@ } }, { - "accuracy": 0.981152068627508, + "accuracy": 0.9812211535478893, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -112686,7 +112686,7 @@ } }, { - "accuracy": 0.9882402318088632, + "accuracy": 0.9882848803934298, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -112729,7 +112729,7 @@ } }, { - "accuracy": 0.9892013167080126, + "accuracy": 0.989242498027651, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -112772,7 +112772,7 @@ } }, { - "accuracy": 0.9883010473690534, + "accuracy": 0.9883473358656231, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -112824,7 +112824,7 @@ } }, { - "accuracy": 0.9897006318757409, + "accuracy": 0.9897383561259822, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -112876,7 +112876,7 @@ } }, { - "accuracy": 0.9940543433553294, + "accuracy": 0.9940777922931471, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -112928,7 +112928,7 @@ } }, { - "accuracy": 0.9948689663096478, + "accuracy": 0.9948895083446252, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -112980,7 +112980,7 @@ } }, { - "accuracy": 0.9968327874023664, + "accuracy": 0.9968463080494028, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -113023,7 +113023,7 @@ } }, { - "accuracy": 0.9969651999050065, + "accuracy": 0.9969758142374063, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -113072,7 +113072,7 @@ } }, { - "accuracy": 0.9974176783702875, + "accuracy": 0.9974263708450293, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -113118,7 +113118,7 @@ } }, { - "accuracy": 0.9991427275695299, + "accuracy": 0.9991463744326642, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -113160,7 +113160,7 @@ ], "model.layers.58.self_attn": [ { - "accuracy": 0.98523810505867, + "accuracy": 0.9852778644938218, "total_bits": 320757760, "q_proj": { "group_size": { @@ -113224,7 +113224,7 @@ } }, { - "accuracy": 0.9859842297277952, + "accuracy": 0.986109219099346, "total_bits": 329080832, "q_proj": { "group_size": { @@ -113288,7 +113288,7 @@ } }, { - "accuracy": 0.9884573307476545, + "accuracy": 0.9886723796003744, "total_bits": 336024576, "q_proj": { "group_size": { @@ -113352,7 +113352,7 @@ } }, { - "accuracy": 0.9909849872714594, + "accuracy": 0.9911558345744484, "total_bits": 401557504, "q_proj": { "group_size": { @@ -113416,7 +113416,7 @@ } }, { - "accuracy": 0.9921554295640242, + "accuracy": 0.9922415025924382, "total_bits": 475279360, "q_proj": { "group_size": { @@ -113480,7 +113480,7 @@ } }, { - "accuracy": 0.9924782734168204, + "accuracy": 0.9925786782252161, "total_bits": 475479040, "q_proj": { "group_size": { @@ -113544,7 +113544,7 @@ } }, { - "accuracy": 0.9941879538329024, + "accuracy": 0.9941972850968963, "total_bits": 609759232, "q_proj": { "group_size": { @@ -113596,7 +113596,7 @@ } }, { - "accuracy": 0.9945060297062522, + "accuracy": 0.9946118106967524, "total_bits": 610024448, "q_proj": { "group_size": { @@ -113648,7 +113648,7 @@ } }, { - "accuracy": 0.9949631361584914, + "accuracy": 0.995075740704411, "total_bits": 615020544, "q_proj": { "group_size": { @@ -113700,7 +113700,7 @@ } }, { - "accuracy": 0.9950628672775469, + "accuracy": 0.9951731370467889, "total_bits": 623951872, "q_proj": { "group_size": { @@ -113752,7 +113752,7 @@ } }, { - "accuracy": 0.9957421970994849, + "accuracy": 0.9957711500556845, "total_bits": 626473984, "q_proj": { "group_size": { @@ -113816,7 +113816,7 @@ } }, { - "accuracy": 0.9960934562902701, + "accuracy": 0.9960872546622628, "total_bits": 630355968, "q_proj": { "group_size": { @@ -113880,7 +113880,7 @@ } }, { - "accuracy": 0.996176904361499, + "accuracy": 0.9962108103852523, "total_bits": 637362176, "q_proj": { "group_size": { @@ -113941,7 +113941,7 @@ } }, { - "accuracy": 0.9965251793986872, + "accuracy": 0.9965184503480008, "total_bits": 646823936, "q_proj": { "group_size": { @@ -114002,7 +114002,7 @@ } }, { - "accuracy": 0.9978361651301384, + "accuracy": 0.9978606702857896, "total_bits": 784740352, "q_proj": { "group_size": { @@ -114063,7 +114063,7 @@ } }, { - "accuracy": 0.9982101619243622, + "accuracy": 0.9982196661202531, "total_bits": 797818880, "q_proj": { "group_size": { @@ -114124,7 +114124,7 @@ } }, { - "accuracy": 0.9980808355306324, + "accuracy": 0.9981002199806666, "total_bits": 911749120, "q_proj": { "group_size": { @@ -114176,7 +114176,7 @@ } }, { - "accuracy": 0.9992366172863465, + "accuracy": 0.9992433184463727, "total_bits": 942718976, "q_proj": { "group_size": { @@ -114228,7 +114228,7 @@ } }, { - "accuracy": 0.9994779643847754, + "accuracy": 0.9994772182483422, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -114282,7 +114282,7 @@ ], "model.layers.58.mlp": [ { - "accuracy": 0.9545369524704783, + "accuracy": 0.954705724590703, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -114334,7 +114334,7 @@ } }, { - "accuracy": 0.9556693748423928, + "accuracy": 0.9558377736493161, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -114386,7 +114386,7 @@ } }, { - "accuracy": 0.9614286014908239, + "accuracy": 0.9615764021873474, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -114435,7 +114435,7 @@ } }, { - "accuracy": 0.9630126733528941, + "accuracy": 0.9631525874137878, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -114484,7 +114484,7 @@ } }, { - "accuracy": 0.9771041305441606, + "accuracy": 0.9771881636820341, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -114536,7 +114536,7 @@ } }, { - "accuracy": 0.9788894621949447, + "accuracy": 0.9789621014344064, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -114588,7 +114588,7 @@ } }, { - "accuracy": 0.9812810781754946, + "accuracy": 0.9813453884501206, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -114637,7 +114637,7 @@ } }, { - "accuracy": 0.9883390390559247, + "accuracy": 0.9883799223523391, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -114680,7 +114680,7 @@ } }, { - "accuracy": 0.9892968386411667, + "accuracy": 0.9893356891054856, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -114723,7 +114723,7 @@ } }, { - "accuracy": 0.9883744434306496, + "accuracy": 0.9884188551651804, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -114775,7 +114775,7 @@ } }, { - "accuracy": 0.9897693238760296, + "accuracy": 0.9898046231583545, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -114827,7 +114827,7 @@ } }, { - "accuracy": 0.9940917248788633, + "accuracy": 0.9941142840605033, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -114879,7 +114879,7 @@ } }, { - "accuracy": 0.994904326373025, + "accuracy": 0.9949225453953994, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -114931,7 +114931,7 @@ } }, { - "accuracy": 0.9968593491143302, + "accuracy": 0.996871001430248, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -114974,7 +114974,7 @@ } }, { - "accuracy": 0.9969814506016279, + "accuracy": 0.9969924537366942, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -115023,7 +115023,7 @@ } }, { - "accuracy": 0.9974286618985628, + "accuracy": 0.9974375449513134, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -115069,7 +115069,7 @@ } }, { - "accuracy": 0.9991506934165955, + "accuracy": 0.999153899715135, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -115111,7 +115111,7 @@ ], "model.layers.59.self_attn": [ { - "accuracy": 0.9882408678531647, + "accuracy": 0.9884738796635678, "total_bits": 320757760, "q_proj": { "group_size": { @@ -115175,7 +115175,7 @@ } }, { - "accuracy": 0.9889689480003557, + "accuracy": 0.9890111198550776, "total_bits": 329080832, "q_proj": { "group_size": { @@ -115239,7 +115239,7 @@ } }, { - "accuracy": 0.9910667836666107, + "accuracy": 0.9911598257328335, "total_bits": 336024576, "q_proj": { "group_size": { @@ -115303,7 +115303,7 @@ } }, { - "accuracy": 0.9927340306733784, + "accuracy": 0.9927990303227776, "total_bits": 401557504, "q_proj": { "group_size": { @@ -115367,7 +115367,7 @@ } }, { - "accuracy": 0.9933291422693353, + "accuracy": 0.9933927859130659, "total_bits": 475279360, "q_proj": { "group_size": { @@ -115431,7 +115431,7 @@ } }, { - "accuracy": 0.9936061658357319, + "accuracy": 0.9936711631323162, "total_bits": 475479040, "q_proj": { "group_size": { @@ -115495,7 +115495,7 @@ } }, { - "accuracy": 0.9951092130259463, + "accuracy": 0.995201318279693, "total_bits": 609759232, "q_proj": { "group_size": { @@ -115547,7 +115547,7 @@ } }, { - "accuracy": 0.9953971540457324, + "accuracy": 0.9954357911881647, "total_bits": 610024448, "q_proj": { "group_size": { @@ -115599,7 +115599,7 @@ } }, { - "accuracy": 0.9957605702312369, + "accuracy": 0.9957483414756624, "total_bits": 615020544, "q_proj": { "group_size": { @@ -115651,7 +115651,7 @@ } }, { - "accuracy": 0.995830452363742, + "accuracy": 0.9958316585735271, "total_bits": 623951872, "q_proj": { "group_size": { @@ -115703,7 +115703,7 @@ } }, { - "accuracy": 0.9958065163932348, + "accuracy": 0.9958241268208152, "total_bits": 626473984, "q_proj": { "group_size": { @@ -115767,7 +115767,7 @@ } }, { - "accuracy": 0.9960235380812695, + "accuracy": 0.9960570449107572, "total_bits": 630355968, "q_proj": { "group_size": { @@ -115831,7 +115831,7 @@ } }, { - "accuracy": 0.9969271133212667, + "accuracy": 0.9968634039947861, "total_bits": 637362176, "q_proj": { "group_size": { @@ -115892,7 +115892,7 @@ } }, { - "accuracy": 0.9971408000902126, + "accuracy": 0.997145444939011, "total_bits": 646823936, "q_proj": { "group_size": { @@ -115953,7 +115953,7 @@ } }, { - "accuracy": 0.9981886513139072, + "accuracy": 0.9981977194547653, "total_bits": 784740352, "q_proj": { "group_size": { @@ -116014,7 +116014,7 @@ } }, { - "accuracy": 0.9985304531690321, + "accuracy": 0.9985570533103064, "total_bits": 797818880, "q_proj": { "group_size": { @@ -116075,7 +116075,7 @@ } }, { - "accuracy": 0.9983531103322381, + "accuracy": 0.9983579064474294, "total_bits": 911749120, "q_proj": { "group_size": { @@ -116127,7 +116127,7 @@ } }, { - "accuracy": 0.9994003335994325, + "accuracy": 0.9994072075933218, "total_bits": 942718976, "q_proj": { "group_size": { @@ -116179,7 +116179,7 @@ } }, { - "accuracy": 0.999530895319032, + "accuracy": 0.999542680863095, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -116233,7 +116233,7 @@ ], "model.layers.59.mlp": [ { - "accuracy": 0.9534630399001272, + "accuracy": 0.9536421706801966, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -116285,7 +116285,7 @@ } }, { - "accuracy": 0.9546411382524591, + "accuracy": 0.9548060078369944, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -116337,7 +116337,7 @@ } }, { - "accuracy": 0.9604682859621549, + "accuracy": 0.9606017846810191, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -116386,7 +116386,7 @@ } }, { - "accuracy": 0.9620847764768099, + "accuracy": 0.9622087353154233, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -116435,7 +116435,7 @@ } }, { - "accuracy": 0.9765801445433968, + "accuracy": 0.9766598971266496, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -116487,7 +116487,7 @@ } }, { - "accuracy": 0.9784006617571178, + "accuracy": 0.9784740617400721, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -116539,7 +116539,7 @@ } }, { - "accuracy": 0.9808209710999539, + "accuracy": 0.9808864687618456, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -116588,7 +116588,7 @@ } }, { - "accuracy": 0.9880539618040386, + "accuracy": 0.9880993679950112, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -116631,7 +116631,7 @@ } }, { - "accuracy": 0.9890213349932119, + "accuracy": 0.9890584239834234, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -116674,7 +116674,7 @@ } }, { - "accuracy": 0.9881005545980052, + "accuracy": 0.9881433193620882, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -116726,7 +116726,7 @@ } }, { - "accuracy": 0.9895192918024565, + "accuracy": 0.9895549416542053, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -116778,7 +116778,7 @@ } }, { - "accuracy": 0.9939489368545381, + "accuracy": 0.9939695204559126, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -116830,7 +116830,7 @@ } }, { - "accuracy": 0.9947767845894161, + "accuracy": 0.9947924221816816, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -116882,7 +116882,7 @@ } }, { - "accuracy": 0.9967731215844029, + "accuracy": 0.9967854517070871, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -116925,7 +116925,7 @@ } }, { - "accuracy": 0.9968936902127767, + "accuracy": 0.9969046259004819, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -116974,7 +116974,7 @@ } }, { - "accuracy": 0.9973437327303385, + "accuracy": 0.9973528716517123, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -117020,7 +117020,7 @@ } }, { - "accuracy": 0.9990744487823624, + "accuracy": 0.999077346371977, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -117062,7 +117062,7 @@ ], "model.layers.60.self_attn": [ { - "accuracy": 0.9683017322891637, + "accuracy": 0.9676071436781633, "total_bits": 320757760, "q_proj": { "group_size": { @@ -117126,7 +117126,7 @@ } }, { - "accuracy": 0.9671071297244022, + "accuracy": 0.9709349996165225, "total_bits": 329080832, "q_proj": { "group_size": { @@ -117190,7 +117190,7 @@ } }, { - "accuracy": 0.9797070183252033, + "accuracy": 0.9794246946510515, "total_bits": 336024576, "q_proj": { "group_size": { @@ -117254,7 +117254,7 @@ } }, { - "accuracy": 0.9824073189183286, + "accuracy": 0.9821221875516992, "total_bits": 401557504, "q_proj": { "group_size": { @@ -117318,7 +117318,7 @@ } }, { - "accuracy": 0.9855164712981174, + "accuracy": 0.9850607326156214, "total_bits": 475279360, "q_proj": { "group_size": { @@ -117382,7 +117382,7 @@ } }, { - "accuracy": 0.9853995771784532, + "accuracy": 0.9852273840653268, "total_bits": 475479040, "q_proj": { "group_size": { @@ -117446,7 +117446,7 @@ } }, { - "accuracy": 0.9919655095589789, + "accuracy": 0.9915210432127902, "total_bits": 609759232, "q_proj": { "group_size": { @@ -117498,7 +117498,7 @@ } }, { - "accuracy": 0.9917988957543122, + "accuracy": 0.9921969889025939, "total_bits": 610024448, "q_proj": { "group_size": { @@ -117550,7 +117550,7 @@ } }, { - "accuracy": 0.9926195411305678, + "accuracy": 0.9928622481070066, "total_bits": 615020544, "q_proj": { "group_size": { @@ -117602,7 +117602,7 @@ } }, { - "accuracy": 0.9927670453724108, + "accuracy": 0.9930044875333184, "total_bits": 623951872, "q_proj": { "group_size": { @@ -117654,7 +117654,7 @@ } }, { - "accuracy": 0.9923084605681268, + "accuracy": 0.9926339889827528, "total_bits": 626473984, "q_proj": { "group_size": { @@ -117718,7 +117718,7 @@ } }, { - "accuracy": 0.9929973804637006, + "accuracy": 0.9931154753032484, "total_bits": 630355968, "q_proj": { "group_size": { @@ -117782,7 +117782,7 @@ } }, { - "accuracy": 0.9949899770711598, + "accuracy": 0.9948994470270056, "total_bits": 637362176, "q_proj": { "group_size": { @@ -117843,7 +117843,7 @@ } }, { - "accuracy": 0.9954269289186126, + "accuracy": 0.9954268712746469, "total_bits": 646823936, "q_proj": { "group_size": { @@ -117904,7 +117904,7 @@ } }, { - "accuracy": 0.9972538365737388, + "accuracy": 0.9972561497829462, "total_bits": 784740352, "q_proj": { "group_size": { @@ -117965,7 +117965,7 @@ } }, { - "accuracy": 0.9976385506360155, + "accuracy": 0.997677927150538, "total_bits": 797818880, "q_proj": { "group_size": { @@ -118026,7 +118026,7 @@ } }, { - "accuracy": 0.9977654461797915, + "accuracy": 0.9977525745175386, "total_bits": 911749120, "q_proj": { "group_size": { @@ -118078,7 +118078,7 @@ } }, { - "accuracy": 0.9989691419821036, + "accuracy": 0.9989611223143967, "total_bits": 942718976, "q_proj": { "group_size": { @@ -118130,7 +118130,7 @@ } }, { - "accuracy": 0.9993849555520635, + "accuracy": 0.9993964707184779, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -118184,7 +118184,7 @@ ], "model.layers.60.mlp": [ { - "accuracy": 0.9526093570809615, + "accuracy": 0.9527643134719447, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -118236,7 +118236,7 @@ } }, { - "accuracy": 0.9538243601196691, + "accuracy": 0.9539704762007061, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -118288,7 +118288,7 @@ } }, { - "accuracy": 0.9596875498169347, + "accuracy": 0.9598257196576971, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -118337,7 +118337,7 @@ } }, { - "accuracy": 0.9613552250360188, + "accuracy": 0.9614852980563515, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -118386,7 +118386,7 @@ } }, { - "accuracy": 0.9761155137890264, + "accuracy": 0.9761953197027508, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -118438,7 +118438,7 @@ } }, { - "accuracy": 0.9779863279116782, + "accuracy": 0.9780601953205309, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -118490,7 +118490,7 @@ } }, { - "accuracy": 0.9804449002993735, + "accuracy": 0.980511011261689, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -118539,7 +118539,7 @@ } }, { - "accuracy": 0.9877993056648656, + "accuracy": 0.9878378941824562, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -118582,7 +118582,7 @@ } }, { - "accuracy": 0.988778658603367, + "accuracy": 0.9888142057155308, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -118625,7 +118625,7 @@ } }, { - "accuracy": 0.9878573221595663, + "accuracy": 0.9878977654795897, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -118677,7 +118677,7 @@ } }, { - "accuracy": 0.9893161858382978, + "accuracy": 0.9893513753225929, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -118729,7 +118729,7 @@ } }, { - "accuracy": 0.9938219024946815, + "accuracy": 0.993841500658738, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -118781,7 +118781,7 @@ } }, { - "accuracy": 0.9946700307099443, + "accuracy": 0.9946875042821232, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -118833,7 +118833,7 @@ } }, { - "accuracy": 0.9966906748319927, + "accuracy": 0.9967014397445478, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -118876,7 +118876,7 @@ } }, { - "accuracy": 0.9968228540138194, + "accuracy": 0.9968342630094603, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -118925,7 +118925,7 @@ } }, { - "accuracy": 0.9972795626442683, + "accuracy": 0.9972895668133309, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -118971,7 +118971,7 @@ } }, { - "accuracy": 0.999021132231543, + "accuracy": 0.9990250046708082, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -119013,7 +119013,7 @@ ], "model.layers.61.self_attn": [ { - "accuracy": 0.9866692737529152, + "accuracy": 0.9868555068969727, "total_bits": 320757760, "q_proj": { "group_size": { @@ -119077,7 +119077,7 @@ } }, { - "accuracy": 0.9876672484372792, + "accuracy": 0.98776870181686, "total_bits": 329080832, "q_proj": { "group_size": { @@ -119141,7 +119141,7 @@ } }, { - "accuracy": 0.990165388113574, + "accuracy": 0.9902620997868086, "total_bits": 336024576, "q_proj": { "group_size": { @@ -119205,7 +119205,7 @@ } }, { - "accuracy": 0.9919962796725725, + "accuracy": 0.9920590770872015, "total_bits": 401557504, "q_proj": { "group_size": { @@ -119269,7 +119269,7 @@ } }, { - "accuracy": 0.9929962166045841, + "accuracy": 0.9930817124090696, "total_bits": 475279360, "q_proj": { "group_size": { @@ -119333,7 +119333,7 @@ } }, { - "accuracy": 0.9932357315954409, + "accuracy": 0.9932901271079716, "total_bits": 475479040, "q_proj": { "group_size": { @@ -119397,7 +119397,7 @@ } }, { - "accuracy": 0.9946102096061957, + "accuracy": 0.9946656725124309, "total_bits": 609759232, "q_proj": { "group_size": { @@ -119449,7 +119449,7 @@ } }, { - "accuracy": 0.9948291704058647, + "accuracy": 0.994857035969433, "total_bits": 610024448, "q_proj": { "group_size": { @@ -119501,7 +119501,7 @@ } }, { - "accuracy": 0.995382801482552, + "accuracy": 0.9953889188013578, "total_bits": 615020544, "q_proj": { "group_size": { @@ -119553,7 +119553,7 @@ } }, { - "accuracy": 0.9954536796400422, + "accuracy": 0.9954786657502777, "total_bits": 623951872, "q_proj": { "group_size": { @@ -119605,7 +119605,7 @@ } }, { - "accuracy": 0.9961052874201223, + "accuracy": 0.996190261683966, "total_bits": 626473984, "q_proj": { "group_size": { @@ -119669,7 +119669,7 @@ } }, { - "accuracy": 0.9964408949017525, + "accuracy": 0.9964199815141527, "total_bits": 630355968, "q_proj": { "group_size": { @@ -119733,7 +119733,7 @@ } }, { - "accuracy": 0.996454936501227, + "accuracy": 0.996509453967998, "total_bits": 637362176, "q_proj": { "group_size": { @@ -119794,7 +119794,7 @@ } }, { - "accuracy": 0.9967977096767802, + "accuracy": 0.9967538118362427, "total_bits": 646823936, "q_proj": { "group_size": { @@ -119855,7 +119855,7 @@ } }, { - "accuracy": 0.997962794805828, + "accuracy": 0.9979831348908575, "total_bits": 784740352, "q_proj": { "group_size": { @@ -119916,7 +119916,7 @@ } }, { - "accuracy": 0.9982981887695036, + "accuracy": 0.9982766150251815, "total_bits": 797818880, "q_proj": { "group_size": { @@ -119977,7 +119977,7 @@ } }, { - "accuracy": 0.9981280729958886, + "accuracy": 0.9981469898239562, "total_bits": 911749120, "q_proj": { "group_size": { @@ -120029,7 +120029,7 @@ } }, { - "accuracy": 0.9993261769904118, + "accuracy": 0.999338778569118, "total_bits": 942718976, "q_proj": { "group_size": { @@ -120081,7 +120081,7 @@ } }, { - "accuracy": 0.9994834871766599, + "accuracy": 0.9994933480200799, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -120135,7 +120135,7 @@ ], "model.layers.61.mlp": [ { - "accuracy": 0.9533023959711978, + "accuracy": 0.9534526843773692, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -120187,7 +120187,7 @@ } }, { - "accuracy": 0.9545012712478638, + "accuracy": 0.9546403100616053, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -120239,7 +120239,7 @@ } }, { - "accuracy": 0.9602776358002111, + "accuracy": 0.9603994771053916, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -120288,7 +120288,7 @@ } }, { - "accuracy": 0.9618996977806091, + "accuracy": 0.9620168020850733, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -120337,7 +120337,7 @@ } }, { - "accuracy": 0.976500043743535, + "accuracy": 0.9765821397304535, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -120389,7 +120389,7 @@ } }, { - "accuracy": 0.9783278421351784, + "accuracy": 0.978392514743303, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -120441,7 +120441,7 @@ } }, { - "accuracy": 0.9807340251772028, + "accuracy": 0.9807899374710886, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -120490,7 +120490,7 @@ } }, { - "accuracy": 0.9880057549790332, + "accuracy": 0.9880424957526358, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -120533,7 +120533,7 @@ } }, { - "accuracy": 0.9889708694658781, + "accuracy": 0.989007620435012, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -120576,7 +120576,7 @@ } }, { - "accuracy": 0.9880727294244265, + "accuracy": 0.9881076875485872, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -120628,7 +120628,7 @@ } }, { - "accuracy": 0.9894889193145853, + "accuracy": 0.989524350354546, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -120680,7 +120680,7 @@ } }, { - "accuracy": 0.993934949369807, + "accuracy": 0.993955371803359, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -120732,7 +120732,7 @@ } }, { - "accuracy": 0.9947626033895894, + "accuracy": 0.994779488758037, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -120784,7 +120784,7 @@ } }, { - "accuracy": 0.9967616659245992, + "accuracy": 0.996772475540638, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -120827,7 +120827,7 @@ } }, { - "accuracy": 0.9969003604430902, + "accuracy": 0.996910434020193, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -120876,7 +120876,7 @@ } }, { - "accuracy": 0.9973522163927555, + "accuracy": 0.9973599159795987, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -120922,7 +120922,7 @@ } }, { - "accuracy": 0.9991146074117798, + "accuracy": 0.9991170870826432, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -120964,7 +120964,7 @@ ], "model.layers.62.self_attn": [ { - "accuracy": 0.989991904873597, + "accuracy": 0.9900583107220499, "total_bits": 320757760, "q_proj": { "group_size": { @@ -121028,7 +121028,7 @@ } }, { - "accuracy": 0.9904067414371591, + "accuracy": 0.9904956794098804, "total_bits": 329080832, "q_proj": { "group_size": { @@ -121092,7 +121092,7 @@ } }, { - "accuracy": 0.9924340075568149, + "accuracy": 0.9924851114812651, "total_bits": 336024576, "q_proj": { "group_size": { @@ -121156,7 +121156,7 @@ } }, { - "accuracy": 0.9938977352882686, + "accuracy": 0.9939134677774027, "total_bits": 401557504, "q_proj": { "group_size": { @@ -121220,7 +121220,7 @@ } }, { - "accuracy": 0.9941973792094934, + "accuracy": 0.994201881320853, "total_bits": 475279360, "q_proj": { "group_size": { @@ -121284,7 +121284,7 @@ } }, { - "accuracy": 0.9946700205144129, + "accuracy": 0.9947545199017775, "total_bits": 475479040, "q_proj": { "group_size": { @@ -121348,7 +121348,7 @@ } }, { - "accuracy": 0.995233269898515, + "accuracy": 0.9952662849896833, "total_bits": 609759232, "q_proj": { "group_size": { @@ -121400,7 +121400,7 @@ } }, { - "accuracy": 0.995805659968602, + "accuracy": 0.9958687863851848, "total_bits": 610024448, "q_proj": { "group_size": { @@ -121452,7 +121452,7 @@ } }, { - "accuracy": 0.9961536589421724, + "accuracy": 0.996189506430375, "total_bits": 615020544, "q_proj": { "group_size": { @@ -121504,7 +121504,7 @@ } }, { - "accuracy": 0.996188065723369, + "accuracy": 0.9962224458393297, "total_bits": 623951872, "q_proj": { "group_size": { @@ -121556,7 +121556,7 @@ } }, { - "accuracy": 0.9970767001964544, + "accuracy": 0.9970806776300857, "total_bits": 626473984, "q_proj": { "group_size": { @@ -121620,7 +121620,7 @@ } }, { - "accuracy": 0.9972768325946832, + "accuracy": 0.9972786891617273, "total_bits": 630355968, "q_proj": { "group_size": { @@ -121684,7 +121684,7 @@ } }, { - "accuracy": 0.9973702066038784, + "accuracy": 0.9973858610579842, "total_bits": 637362176, "q_proj": { "group_size": { @@ -121745,7 +121745,7 @@ } }, { - "accuracy": 0.9975458740403778, + "accuracy": 0.9975426585266465, "total_bits": 646823936, "q_proj": { "group_size": { @@ -121806,7 +121806,7 @@ } }, { - "accuracy": 0.9981863296737796, + "accuracy": 0.9981983768704691, "total_bits": 784740352, "q_proj": { "group_size": { @@ -121867,7 +121867,7 @@ } }, { - "accuracy": 0.998758642297042, + "accuracy": 0.9987690466407099, "total_bits": 797818880, "q_proj": { "group_size": { @@ -121928,7 +121928,7 @@ } }, { - "accuracy": 0.9982870838751918, + "accuracy": 0.9982992645941282, "total_bits": 911749120, "q_proj": { "group_size": { @@ -121980,7 +121980,7 @@ } }, { - "accuracy": 0.9994856801962382, + "accuracy": 0.9994810135839017, "total_bits": 942718976, "q_proj": { "group_size": { @@ -122032,7 +122032,7 @@ } }, { - "accuracy": 0.9995287012210802, + "accuracy": 0.9995289700792024, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -122086,7 +122086,7 @@ ], "model.layers.62.mlp": [ { - "accuracy": 0.95394787349199, + "accuracy": 0.9540912822673195, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -122138,7 +122138,7 @@ } }, { - "accuracy": 0.9551180977570384, + "accuracy": 0.9552668270311857, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -122190,7 +122190,7 @@ } }, { - "accuracy": 0.9608404510899594, + "accuracy": 0.9609683720689071, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -122239,7 +122239,7 @@ } }, { - "accuracy": 0.9623982812228956, + "accuracy": 0.9625209162109777, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -122288,7 +122288,7 @@ } }, { - "accuracy": 0.9768083252404866, + "accuracy": 0.9768875084425274, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -122340,7 +122340,7 @@ } }, { - "accuracy": 0.978611094386954, + "accuracy": 0.9786788413399145, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -122392,7 +122392,7 @@ } }, { - "accuracy": 0.9809871692406503, + "accuracy": 0.9810468836834556, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -122441,7 +122441,7 @@ } }, { - "accuracy": 0.9881951134455832, + "accuracy": 0.9882331216021588, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -122484,7 +122484,7 @@ } }, { - "accuracy": 0.9891475243003744, + "accuracy": 0.989185727740589, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -122527,7 +122527,7 @@ } }, { - "accuracy": 0.988225740821738, + "accuracy": 0.9882640070036838, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -122579,7 +122579,7 @@ } }, { - "accuracy": 0.9896280436139357, + "accuracy": 0.9896614653499503, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -122631,7 +122631,7 @@ } }, { - "accuracy": 0.9940136275009105, + "accuracy": 0.9940318198580491, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -122683,7 +122683,7 @@ } }, { - "accuracy": 0.9948295919518721, + "accuracy": 0.9948466439780436, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -122735,7 +122735,7 @@ } }, { - "accuracy": 0.9968105080096346, + "accuracy": 0.9968211721432837, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -122778,7 +122778,7 @@ } }, { - "accuracy": 0.9969216936120862, + "accuracy": 0.9969321526587009, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -122827,7 +122827,7 @@ } }, { - "accuracy": 0.9973618358765778, + "accuracy": 0.9973707303012672, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -122873,7 +122873,7 @@ } }, { - "accuracy": 0.9990637937658712, + "accuracy": 0.9990666875321614, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -122915,7 +122915,7 @@ ], "model.layers.63.self_attn": [ { - "accuracy": 0.9873794278031901, + "accuracy": 0.9873796944555483, "total_bits": 320757760, "q_proj": { "group_size": { @@ -122979,7 +122979,7 @@ } }, { - "accuracy": 0.9879948238008901, + "accuracy": 0.9879534370020816, "total_bits": 329080832, "q_proj": { "group_size": { @@ -123043,7 +123043,7 @@ } }, { - "accuracy": 0.9901303076430371, + "accuracy": 0.9901590221806577, "total_bits": 336024576, "q_proj": { "group_size": { @@ -123107,7 +123107,7 @@ } }, { - "accuracy": 0.9922922386934883, + "accuracy": 0.9922506017120261, "total_bits": 401557504, "q_proj": { "group_size": { @@ -123171,7 +123171,7 @@ } }, { - "accuracy": 0.99331229140884, + "accuracy": 0.9932505715834467, "total_bits": 475279360, "q_proj": { "group_size": { @@ -123235,7 +123235,7 @@ } }, { - "accuracy": 0.9934504769350353, + "accuracy": 0.9934662670681351, "total_bits": 475479040, "q_proj": { "group_size": { @@ -123299,7 +123299,7 @@ } }, { - "accuracy": 0.994724990898057, + "accuracy": 0.994718613201066, "total_bits": 609759232, "q_proj": { "group_size": { @@ -123351,7 +123351,7 @@ } }, { - "accuracy": 0.9949122177142846, + "accuracy": 0.9949236425914263, "total_bits": 610024448, "q_proj": { "group_size": { @@ -123403,7 +123403,7 @@ } }, { - "accuracy": 0.9955287516901368, + "accuracy": 0.9955466946488932, "total_bits": 615020544, "q_proj": { "group_size": { @@ -123455,7 +123455,7 @@ } }, { - "accuracy": 0.9956315202932609, + "accuracy": 0.9956376650615743, "total_bits": 623951872, "q_proj": { "group_size": { @@ -123507,7 +123507,7 @@ } }, { - "accuracy": 0.9964708458436163, + "accuracy": 0.9965052083134651, "total_bits": 626473984, "q_proj": { "group_size": { @@ -123571,7 +123571,7 @@ } }, { - "accuracy": 0.9967888676022229, + "accuracy": 0.9967978788833869, "total_bits": 630355968, "q_proj": { "group_size": { @@ -123635,7 +123635,7 @@ } }, { - "accuracy": 0.9967883229255676, + "accuracy": 0.9968196224225195, "total_bits": 637362176, "q_proj": { "group_size": { @@ -123696,7 +123696,7 @@ } }, { - "accuracy": 0.9971301100755993, + "accuracy": 0.9971159771084785, "total_bits": 646823936, "q_proj": { "group_size": { @@ -123757,7 +123757,7 @@ } }, { - "accuracy": 0.9981787114551193, + "accuracy": 0.9982056166780623, "total_bits": 784740352, "q_proj": { "group_size": { @@ -123818,7 +123818,7 @@ } }, { - "accuracy": 0.9985434733527271, + "accuracy": 0.9985685405370436, "total_bits": 797818880, "q_proj": { "group_size": { @@ -123879,7 +123879,7 @@ } }, { - "accuracy": 0.9983874614301481, + "accuracy": 0.9984134777792191, "total_bits": 911749120, "q_proj": { "group_size": { @@ -123931,7 +123931,7 @@ } }, { - "accuracy": 0.9993511811388951, + "accuracy": 0.9993538010473314, "total_bits": 942718976, "q_proj": { "group_size": { @@ -123983,7 +123983,7 @@ } }, { - "accuracy": 0.9995477287294833, + "accuracy": 0.9995547236972734, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -124037,7 +124037,7 @@ ], "model.layers.63.mlp": [ { - "accuracy": 0.9526161457362928, + "accuracy": 0.9527552943480642, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -124089,7 +124089,7 @@ } }, { - "accuracy": 0.9538318671678242, + "accuracy": 0.9539701373953569, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -124141,7 +124141,7 @@ } }, { - "accuracy": 0.9596643792955499, + "accuracy": 0.9597854049582231, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -124190,7 +124190,7 @@ } }, { - "accuracy": 0.9612686100759005, + "accuracy": 0.9613843560218811, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -124239,7 +124239,7 @@ } }, { - "accuracy": 0.9761513252007333, + "accuracy": 0.9762265258713773, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -124291,7 +124291,7 @@ } }, { - "accuracy": 0.977993485174681, + "accuracy": 0.9780668381013369, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -124343,7 +124343,7 @@ } }, { - "accuracy": 0.9804111750502336, + "accuracy": 0.9804752970996656, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -124392,7 +124392,7 @@ } }, { - "accuracy": 0.9878375577299219, + "accuracy": 0.9878783084844288, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -124435,7 +124435,7 @@ } }, { - "accuracy": 0.9888122630746741, + "accuracy": 0.9888485858314916, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -124478,7 +124478,7 @@ } }, { - "accuracy": 0.9878952173810256, + "accuracy": 0.9879320389346072, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -124530,7 +124530,7 @@ } }, { - "accuracy": 0.98932562141042, + "accuracy": 0.9893598423192376, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -124582,7 +124582,7 @@ } }, { - "accuracy": 0.9938440569921544, + "accuracy": 0.9938632316495243, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -124634,7 +124634,7 @@ } }, { - "accuracy": 0.9946781024336815, + "accuracy": 0.9946955344394633, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -124686,7 +124686,7 @@ } }, { - "accuracy": 0.9967163628653476, + "accuracy": 0.9967261974356676, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -124729,7 +124729,7 @@ } }, { - "accuracy": 0.9968513344463549, + "accuracy": 0.9968615842885092, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -124778,7 +124778,7 @@ } }, { - "accuracy": 0.9973034066589255, + "accuracy": 0.9973117368001687, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -124824,7 +124824,7 @@ } }, { - "accuracy": 0.9990917024643797, + "accuracy": 0.9990939312663517, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -124866,7 +124866,7 @@ ], "model.layers.64.self_attn": [ { - "accuracy": 0.9724600189610532, + "accuracy": 0.9732836139829535, "total_bits": 320757760, "q_proj": { "group_size": { @@ -124930,7 +124930,7 @@ } }, { - "accuracy": 0.9750219301173562, + "accuracy": 0.9764141032570287, "total_bits": 329080832, "q_proj": { "group_size": { @@ -124994,7 +124994,7 @@ } }, { - "accuracy": 0.9818364366104728, + "accuracy": 0.9820090297021364, "total_bits": 336024576, "q_proj": { "group_size": { @@ -125058,7 +125058,7 @@ } }, { - "accuracy": 0.9847958589854994, + "accuracy": 0.9849194115714023, "total_bits": 401557504, "q_proj": { "group_size": { @@ -125122,7 +125122,7 @@ } }, { - "accuracy": 0.9871879070997238, + "accuracy": 0.9872207672972428, "total_bits": 475279360, "q_proj": { "group_size": { @@ -125186,7 +125186,7 @@ } }, { - "accuracy": 0.9874718071598756, + "accuracy": 0.9880188750593286, "total_bits": 475479040, "q_proj": { "group_size": { @@ -125250,7 +125250,7 @@ } }, { - "accuracy": 0.9917073594896417, + "accuracy": 0.991942971944809, "total_bits": 609759232, "q_proj": { "group_size": { @@ -125302,7 +125302,7 @@ } }, { - "accuracy": 0.9923483078417025, + "accuracy": 0.9923678605180037, "total_bits": 610024448, "q_proj": { "group_size": { @@ -125354,7 +125354,7 @@ } }, { - "accuracy": 0.9930230112452256, + "accuracy": 0.9931385172040839, "total_bits": 615020544, "q_proj": { "group_size": { @@ -125406,7 +125406,7 @@ } }, { - "accuracy": 0.9932016596982354, + "accuracy": 0.993283665493915, "total_bits": 623951872, "q_proj": { "group_size": { @@ -125458,7 +125458,7 @@ } }, { - "accuracy": 0.9930914389459711, + "accuracy": 0.993052993950091, "total_bits": 626473984, "q_proj": { "group_size": { @@ -125522,7 +125522,7 @@ } }, { - "accuracy": 0.9936126705847288, + "accuracy": 0.9934566699360546, "total_bits": 630355968, "q_proj": { "group_size": { @@ -125586,7 +125586,7 @@ } }, { - "accuracy": 0.9951119203316537, + "accuracy": 0.9950340421576249, "total_bits": 637362176, "q_proj": { "group_size": { @@ -125647,7 +125647,7 @@ } }, { - "accuracy": 0.9954698077942196, + "accuracy": 0.9954988513337938, "total_bits": 646823936, "q_proj": { "group_size": { @@ -125708,7 +125708,7 @@ } }, { - "accuracy": 0.997290227758257, + "accuracy": 0.997308327375274, "total_bits": 784740352, "q_proj": { "group_size": { @@ -125769,7 +125769,7 @@ } }, { - "accuracy": 0.9977303327698457, + "accuracy": 0.9977406475104784, "total_bits": 797818880, "q_proj": { "group_size": { @@ -125830,7 +125830,7 @@ } }, { - "accuracy": 0.9977145879284331, + "accuracy": 0.997729940045821, "total_bits": 911749120, "q_proj": { "group_size": { @@ -125882,7 +125882,7 @@ } }, { - "accuracy": 0.9989682776167205, + "accuracy": 0.9989781037561203, "total_bits": 942718976, "q_proj": { "group_size": { @@ -125934,7 +125934,7 @@ } }, { - "accuracy": 0.9993537964397355, + "accuracy": 0.9993762413137838, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -125988,7 +125988,7 @@ ], "model.layers.64.mlp": [ { - "accuracy": 0.9512565481035333, + "accuracy": 0.9514001388298838, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -126040,7 +126040,7 @@ } }, { - "accuracy": 0.9525316075274819, + "accuracy": 0.9526638608229787, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -126092,7 +126092,7 @@ } }, { - "accuracy": 0.9584410316065738, + "accuracy": 0.9585528436459994, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -126141,7 +126141,7 @@ } }, { - "accuracy": 0.9601289222114965, + "accuracy": 0.9602373869795549, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -126190,7 +126190,7 @@ } }, { - "accuracy": 0.9754608100966403, + "accuracy": 0.9755296738524186, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -126242,7 +126242,7 @@ } }, { - "accuracy": 0.9773552041304739, + "accuracy": 0.9774283186385506, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -126294,7 +126294,7 @@ } }, { - "accuracy": 0.9798251939447302, + "accuracy": 0.979891239028228, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -126343,7 +126343,7 @@ } }, { - "accuracy": 0.987441897392273, + "accuracy": 0.9874787824718576, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -126386,7 +126386,7 @@ } }, { - "accuracy": 0.9884478794900995, + "accuracy": 0.9884835928678513, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -126429,7 +126429,7 @@ } }, { - "accuracy": 0.9875318698192898, + "accuracy": 0.9875681988502804, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -126481,7 +126481,7 @@ } }, { - "accuracy": 0.9890117237442418, + "accuracy": 0.9890459250462683, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -126533,7 +126533,7 @@ } }, { - "accuracy": 0.9936553416283507, + "accuracy": 0.9936732320409072, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -126585,7 +126585,7 @@ } }, { - "accuracy": 0.9945140112387506, + "accuracy": 0.994531054637934, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -126637,7 +126637,7 @@ } }, { - "accuracy": 0.9965891453780626, + "accuracy": 0.9966002898780923, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -126680,7 +126680,7 @@ } }, { - "accuracy": 0.9967410213461048, + "accuracy": 0.9967495597114688, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -126729,7 +126729,7 @@ } }, { - "accuracy": 0.9972025663837006, + "accuracy": 0.9972093503334021, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -126775,7 +126775,7 @@ } }, { - "accuracy": 0.9990021005076798, + "accuracy": 0.9990057335480264, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -126817,7 +126817,7 @@ ], "model.layers.65.self_attn": [ { - "accuracy": 0.9886775236380728, + "accuracy": 0.9886473480023836, "total_bits": 320757760, "q_proj": { "group_size": { @@ -126881,7 +126881,7 @@ } }, { - "accuracy": 0.9892002399030485, + "accuracy": 0.9893025897051159, "total_bits": 329080832, "q_proj": { "group_size": { @@ -126945,7 +126945,7 @@ } }, { - "accuracy": 0.990904793927544, + "accuracy": 0.990879384310622, "total_bits": 336024576, "q_proj": { "group_size": { @@ -127009,7 +127009,7 @@ } }, { - "accuracy": 0.9926663629318538, + "accuracy": 0.9926816774042029, "total_bits": 401557504, "q_proj": { "group_size": { @@ -127073,7 +127073,7 @@ } }, { - "accuracy": 0.9931739647137491, + "accuracy": 0.9932138143401397, "total_bits": 475279360, "q_proj": { "group_size": { @@ -127137,7 +127137,7 @@ } }, { - "accuracy": 0.9938776614634615, + "accuracy": 0.9938137343055323, "total_bits": 475479040, "q_proj": { "group_size": { @@ -127201,7 +127201,7 @@ } }, { - "accuracy": 0.9943064490431234, + "accuracy": 0.9943868737471732, "total_bits": 609759232, "q_proj": { "group_size": { @@ -127253,7 +127253,7 @@ } }, { - "accuracy": 0.9950956690468287, + "accuracy": 0.9950657141836066, "total_bits": 610024448, "q_proj": { "group_size": { @@ -127305,7 +127305,7 @@ } }, { - "accuracy": 0.996090321556518, + "accuracy": 0.9960794174357465, "total_bits": 615020544, "q_proj": { "group_size": { @@ -127357,7 +127357,7 @@ } }, { - "accuracy": 0.9961823283841735, + "accuracy": 0.9961794167757034, "total_bits": 623951872, "q_proj": { "group_size": { @@ -127409,7 +127409,7 @@ } }, { - "accuracy": 0.9963892580647218, + "accuracy": 0.9964048799715544, "total_bits": 626473984, "q_proj": { "group_size": { @@ -127473,7 +127473,7 @@ } }, { - "accuracy": 0.9968599294753451, + "accuracy": 0.996848207947455, "total_bits": 630355968, "q_proj": { "group_size": { @@ -127537,7 +127537,7 @@ } }, { - "accuracy": 0.9966929374556792, + "accuracy": 0.9967230723092431, "total_bits": 637362176, "q_proj": { "group_size": { @@ -127598,7 +127598,7 @@ } }, { - "accuracy": 0.9971812177253397, + "accuracy": 0.9971968892373537, "total_bits": 646823936, "q_proj": { "group_size": { @@ -127659,7 +127659,7 @@ } }, { - "accuracy": 0.9979950797400976, + "accuracy": 0.9980208146336832, "total_bits": 784740352, "q_proj": { "group_size": { @@ -127720,7 +127720,7 @@ } }, { - "accuracy": 0.9984924297191595, + "accuracy": 0.998508428468516, "total_bits": 797818880, "q_proj": { "group_size": { @@ -127781,7 +127781,7 @@ } }, { - "accuracy": 0.998115999722167, + "accuracy": 0.9981431204237436, "total_bits": 911749120, "q_proj": { "group_size": { @@ -127833,7 +127833,7 @@ } }, { - "accuracy": 0.9993736863038257, + "accuracy": 0.9993740619209251, "total_bits": 942718976, "q_proj": { "group_size": { @@ -127885,7 +127885,7 @@ } }, { - "accuracy": 0.9994294468901659, + "accuracy": 0.9994410100442014, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -127939,7 +127939,7 @@ ], "model.layers.65.mlp": [ { - "accuracy": 0.9514733364707545, + "accuracy": 0.9516053733072782, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -127991,7 +127991,7 @@ } }, { - "accuracy": 0.9527467301017359, + "accuracy": 0.9528721947419017, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -128043,7 +128043,7 @@ } }, { - "accuracy": 0.9585831416280646, + "accuracy": 0.9586871674186305, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -128092,7 +128092,7 @@ } }, { - "accuracy": 0.9602278972926893, + "accuracy": 0.9603239297866821, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -128141,7 +128141,7 @@ } }, { - "accuracy": 0.975577646180203, + "accuracy": 0.9756513771257902, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -128193,7 +128193,7 @@ } }, { - "accuracy": 0.9774690709616008, + "accuracy": 0.977534460394006, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -128245,7 +128245,7 @@ } }, { - "accuracy": 0.9798987947012249, + "accuracy": 0.9799553250011644, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -128294,7 +128294,7 @@ } }, { - "accuracy": 0.9875220350529018, + "accuracy": 0.9875584989786148, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -128337,7 +128337,7 @@ } }, { - "accuracy": 0.9885159487787046, + "accuracy": 0.9885492889504683, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -128380,7 +128380,7 @@ } }, { - "accuracy": 0.9876041200600172, + "accuracy": 0.9876407204489959, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -128432,7 +128432,7 @@ } }, { - "accuracy": 0.9890672180213427, + "accuracy": 0.9891013346220318, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -128484,7 +128484,7 @@ } }, { - "accuracy": 0.9937008462454143, + "accuracy": 0.9937207761563753, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -128536,7 +128536,7 @@ } }, { - "accuracy": 0.9945515696155397, + "accuracy": 0.9945686228181186, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -128588,7 +128588,7 @@ } }, { - "accuracy": 0.9966345621567023, + "accuracy": 0.996644699651944, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -128631,7 +128631,7 @@ } }, { - "accuracy": 0.9967785359997499, + "accuracy": 0.9967881725414803, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -128680,7 +128680,7 @@ } }, { - "accuracy": 0.9972341611589256, + "accuracy": 0.997242234843342, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -128726,7 +128726,7 @@ } }, { - "accuracy": 0.9990753609883157, + "accuracy": 0.9990777172344295, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -128768,7 +128768,7 @@ ], "model.layers.66.self_attn": [ { - "accuracy": 0.9847671875828191, + "accuracy": 0.9849560637223093, "total_bits": 320757760, "q_proj": { "group_size": { @@ -128832,7 +128832,7 @@ } }, { - "accuracy": 0.985621828781931, + "accuracy": 0.9858163908908242, "total_bits": 329080832, "q_proj": { "group_size": { @@ -128896,7 +128896,7 @@ } }, { - "accuracy": 0.988165189561091, + "accuracy": 0.9882930650522834, "total_bits": 336024576, "q_proj": { "group_size": { @@ -128960,7 +128960,7 @@ } }, { - "accuracy": 0.990691674383063, + "accuracy": 0.9908143462319123, "total_bits": 401557504, "q_proj": { "group_size": { @@ -129024,7 +129024,7 @@ } }, { - "accuracy": 0.9915279306863484, + "accuracy": 0.9916313957226904, "total_bits": 475279360, "q_proj": { "group_size": { @@ -129088,7 +129088,7 @@ } }, { - "accuracy": 0.9921874858831105, + "accuracy": 0.9921768166516957, "total_bits": 475479040, "q_proj": { "group_size": { @@ -129152,7 +129152,7 @@ } }, { - "accuracy": 0.993317731901219, + "accuracy": 0.993363043195323, "total_bits": 609759232, "q_proj": { "group_size": { @@ -129204,7 +129204,7 @@ } }, { - "accuracy": 0.9940647580905965, + "accuracy": 0.9940550605717459, "total_bits": 610024448, "q_proj": { "group_size": { @@ -129256,7 +129256,7 @@ } }, { - "accuracy": 0.9946503552951311, + "accuracy": 0.9946951199519006, "total_bits": 615020544, "q_proj": { "group_size": { @@ -129308,7 +129308,7 @@ } }, { - "accuracy": 0.9948027882921068, + "accuracy": 0.9948435359095272, "total_bits": 623951872, "q_proj": { "group_size": { @@ -129360,7 +129360,7 @@ } }, { - "accuracy": 0.9957686462684682, + "accuracy": 0.9957915619015694, "total_bits": 626473984, "q_proj": { "group_size": { @@ -129424,7 +129424,7 @@ } }, { - "accuracy": 0.996218850345988, + "accuracy": 0.996213363973718, "total_bits": 630355968, "q_proj": { "group_size": { @@ -129488,7 +129488,7 @@ } }, { - "accuracy": 0.996201118748439, + "accuracy": 0.9962339028716087, "total_bits": 637362176, "q_proj": { "group_size": { @@ -129549,7 +129549,7 @@ } }, { - "accuracy": 0.9967117091934932, + "accuracy": 0.9967158181886924, "total_bits": 646823936, "q_proj": { "group_size": { @@ -129610,7 +129610,7 @@ } }, { - "accuracy": 0.9978320547624638, + "accuracy": 0.9978375619179324, "total_bits": 784740352, "q_proj": { "group_size": { @@ -129671,7 +129671,7 @@ } }, { - "accuracy": 0.9983256239640085, + "accuracy": 0.9983352308994845, "total_bits": 797818880, "q_proj": { "group_size": { @@ -129732,7 +129732,7 @@ } }, { - "accuracy": 0.9980450052964059, + "accuracy": 0.998049666418841, "total_bits": 911749120, "q_proj": { "group_size": { @@ -129784,7 +129784,7 @@ } }, { - "accuracy": 0.9991774412950403, + "accuracy": 0.9991903258487582, "total_bits": 942718976, "q_proj": { "group_size": { @@ -129836,7 +129836,7 @@ } }, { - "accuracy": 0.9994528928869649, + "accuracy": 0.9994576519943381, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -129890,7 +129890,7 @@ ], "model.layers.66.mlp": [ { - "accuracy": 0.9522774564592462, + "accuracy": 0.9524128656638295, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -129942,7 +129942,7 @@ } }, { - "accuracy": 0.9535019491848192, + "accuracy": 0.9536432749346683, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -129994,7 +129994,7 @@ } }, { - "accuracy": 0.9592978232785275, + "accuracy": 0.9594006695245442, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -130043,7 +130043,7 @@ } }, { - "accuracy": 0.9608661657885501, + "accuracy": 0.9609654200704474, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -130092,7 +130092,7 @@ } }, { - "accuracy": 0.975963377638867, + "accuracy": 0.9760275890952662, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -130144,7 +130144,7 @@ } }, { - "accuracy": 0.9778263317911249, + "accuracy": 0.9778872414639121, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -130196,7 +130196,7 @@ } }, { - "accuracy": 0.9802261998778895, + "accuracy": 0.9802811318322232, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -130245,7 +130245,7 @@ } }, { - "accuracy": 0.9877519325206154, + "accuracy": 0.9877886003569553, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -130288,7 +130288,7 @@ } }, { - "accuracy": 0.9887346758654243, + "accuracy": 0.9887675478270179, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -130331,7 +130331,7 @@ } }, { - "accuracy": 0.9877897642160717, + "accuracy": 0.9878222228665101, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -130383,7 +130383,7 @@ } }, { - "accuracy": 0.9892358528940302, + "accuracy": 0.9892630506502954, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -130435,7 +130435,7 @@ } }, { - "accuracy": 0.9937873717985655, + "accuracy": 0.9938061088323593, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -130487,7 +130487,7 @@ } }, { - "accuracy": 0.9946309516304418, + "accuracy": 0.9946458696534759, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -130539,7 +130539,7 @@ } }, { - "accuracy": 0.9966874197125435, + "accuracy": 0.9966969478287195, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -130582,7 +130582,7 @@ } }, { - "accuracy": 0.9968002066016197, + "accuracy": 0.9968092341564203, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -130631,7 +130631,7 @@ } }, { - "accuracy": 0.9972436677076315, + "accuracy": 0.997251497875703, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -130677,7 +130677,7 @@ } }, { - "accuracy": 0.9990231825136825, + "accuracy": 0.9990257264948205, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -130719,7 +130719,7 @@ ], "model.layers.67.self_attn": [ { - "accuracy": 0.9680956476613095, + "accuracy": 0.967251109449487, "total_bits": 320757760, "q_proj": { "group_size": { @@ -130783,7 +130783,7 @@ } }, { - "accuracy": 0.9702999372231332, + "accuracy": 0.9691291231858102, "total_bits": 329080832, "q_proj": { "group_size": { @@ -130847,7 +130847,7 @@ } }, { - "accuracy": 0.9802610215387846, + "accuracy": 0.9803621878749446, "total_bits": 336024576, "q_proj": { "group_size": { @@ -130911,7 +130911,7 @@ } }, { - "accuracy": 0.983278657260694, + "accuracy": 0.9832138685803664, "total_bits": 401557504, "q_proj": { "group_size": { @@ -130975,7 +130975,7 @@ } }, { - "accuracy": 0.9853858602674384, + "accuracy": 0.9850692560798243, "total_bits": 475279360, "q_proj": { "group_size": { @@ -131039,7 +131039,7 @@ } }, { - "accuracy": 0.9857442222143474, + "accuracy": 0.9857016692036077, "total_bits": 475479040, "q_proj": { "group_size": { @@ -131103,7 +131103,7 @@ } }, { - "accuracy": 0.9917238715447878, + "accuracy": 0.9917802849882528, "total_bits": 609759232, "q_proj": { "group_size": { @@ -131155,7 +131155,7 @@ } }, { - "accuracy": 0.9920969840727354, + "accuracy": 0.9922733565694407, "total_bits": 610024448, "q_proj": { "group_size": { @@ -131207,7 +131207,7 @@ } }, { - "accuracy": 0.9928652306920603, + "accuracy": 0.9931130817061976, "total_bits": 615020544, "q_proj": { "group_size": { @@ -131259,7 +131259,7 @@ } }, { - "accuracy": 0.99301179223939, + "accuracy": 0.9932503441446706, "total_bits": 623951872, "q_proj": { "group_size": { @@ -131311,7 +131311,7 @@ } }, { - "accuracy": 0.9925184963565123, + "accuracy": 0.9927378688987932, "total_bits": 626473984, "q_proj": { "group_size": { @@ -131375,7 +131375,7 @@ } }, { - "accuracy": 0.9931629958905672, + "accuracy": 0.9931812717726356, "total_bits": 630355968, "q_proj": { "group_size": { @@ -131439,7 +131439,7 @@ } }, { - "accuracy": 0.9950290592877489, + "accuracy": 0.9950359687209129, "total_bits": 637362176, "q_proj": { "group_size": { @@ -131500,7 +131500,7 @@ } }, { - "accuracy": 0.9954348986870364, + "accuracy": 0.9955107565773161, "total_bits": 646823936, "q_proj": { "group_size": { @@ -131561,7 +131561,7 @@ } }, { - "accuracy": 0.9972943316556906, + "accuracy": 0.9972717426717281, "total_bits": 784740352, "q_proj": { "group_size": { @@ -131622,7 +131622,7 @@ } }, { - "accuracy": 0.9977306943190726, + "accuracy": 0.9977828734799435, "total_bits": 797818880, "q_proj": { "group_size": { @@ -131683,7 +131683,7 @@ } }, { - "accuracy": 0.9977564529368752, + "accuracy": 0.9977205964295488, "total_bits": 911749120, "q_proj": { "group_size": { @@ -131735,7 +131735,7 @@ } }, { - "accuracy": 0.9990140038885569, + "accuracy": 0.9990226247004772, "total_bits": 942718976, "q_proj": { "group_size": { @@ -131787,7 +131787,7 @@ } }, { - "accuracy": 0.9993936405762246, + "accuracy": 0.9993906185815209, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -131841,7 +131841,7 @@ ], "model.layers.67.mlp": [ { - "accuracy": 0.9506734170411762, + "accuracy": 0.9508171238397297, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -131893,7 +131893,7 @@ } }, { - "accuracy": 0.9519626435480619, + "accuracy": 0.952096910853135, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -131945,7 +131945,7 @@ } }, { - "accuracy": 0.9579086209598341, + "accuracy": 0.9580228924751282, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -131994,7 +131994,7 @@ } }, { - "accuracy": 0.9595306609806261, + "accuracy": 0.9596395492553711, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -132043,7 +132043,7 @@ } }, { - "accuracy": 0.9751536344227038, + "accuracy": 0.9752307581274133, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -132095,7 +132095,7 @@ } }, { - "accuracy": 0.9770921327565846, + "accuracy": 0.9771575566969419, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -132147,7 +132147,7 @@ } }, { - "accuracy": 0.9795581886642858, + "accuracy": 0.9796152130553597, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -132196,7 +132196,7 @@ } }, { - "accuracy": 0.9873376229875966, + "accuracy": 0.9873744163073992, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -132239,7 +132239,7 @@ } }, { - "accuracy": 0.9883554358231393, + "accuracy": 0.9883849440436614, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -132282,7 +132282,7 @@ } }, { - "accuracy": 0.9873815986670946, + "accuracy": 0.9874194742817628, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -132334,7 +132334,7 @@ } }, { - "accuracy": 0.9888839643252524, + "accuracy": 0.9889135721482729, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -132386,7 +132386,7 @@ } }, { - "accuracy": 0.9935872048923844, + "accuracy": 0.9936039416413558, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -132438,7 +132438,7 @@ } }, { - "accuracy": 0.9944620191266662, + "accuracy": 0.9944774688858735, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -132490,7 +132490,7 @@ } }, { - "accuracy": 0.9965848412952925, + "accuracy": 0.9965937855212312, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -132533,7 +132533,7 @@ } }, { - "accuracy": 0.9967208167440013, + "accuracy": 0.9967295190221385, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -132582,7 +132582,7 @@ } }, { - "accuracy": 0.9971810218534971, + "accuracy": 0.9971882940122956, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -132628,7 +132628,7 @@ } }, { - "accuracy": 0.9990757447912505, + "accuracy": 0.9990786474786306, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -132670,7 +132670,7 @@ ], "model.layers.68.self_attn": [ { - "accuracy": 0.9616809424601103, + "accuracy": 0.963059519466601, "total_bits": 320757760, "q_proj": { "group_size": { @@ -132734,7 +132734,7 @@ } }, { - "accuracy": 0.965724345884825, + "accuracy": 0.9629727507892408, "total_bits": 329080832, "q_proj": { "group_size": { @@ -132798,7 +132798,7 @@ } }, { - "accuracy": 0.9759116643353513, + "accuracy": 0.975863549270128, "total_bits": 336024576, "q_proj": { "group_size": { @@ -132862,7 +132862,7 @@ } }, { - "accuracy": 0.9787850991675728, + "accuracy": 0.978840454628593, "total_bits": 401557504, "q_proj": { "group_size": { @@ -132926,7 +132926,7 @@ } }, { - "accuracy": 0.9824284864099402, + "accuracy": 0.9825243071505898, "total_bits": 475279360, "q_proj": { "group_size": { @@ -132990,7 +132990,7 @@ } }, { - "accuracy": 0.9830899191530127, + "accuracy": 0.9829700024504411, "total_bits": 475479040, "q_proj": { "group_size": { @@ -133054,7 +133054,7 @@ } }, { - "accuracy": 0.9897449981225165, + "accuracy": 0.9901765302607888, "total_bits": 609759232, "q_proj": { "group_size": { @@ -133106,7 +133106,7 @@ } }, { - "accuracy": 0.9904360543740424, + "accuracy": 0.990544426598047, "total_bits": 610024448, "q_proj": { "group_size": { @@ -133158,7 +133158,7 @@ } }, { - "accuracy": 0.9914890292443728, + "accuracy": 0.9917062897431222, "total_bits": 615020544, "q_proj": { "group_size": { @@ -133210,7 +133210,7 @@ } }, { - "accuracy": 0.9917257326213937, + "accuracy": 0.9919187975557227, "total_bits": 623951872, "q_proj": { "group_size": { @@ -133262,7 +133262,7 @@ } }, { - "accuracy": 0.9913913748766247, + "accuracy": 0.99102651837625, "total_bits": 626473984, "q_proj": { "group_size": { @@ -133326,7 +133326,7 @@ } }, { - "accuracy": 0.9920386052445361, + "accuracy": 0.9919882902973577, "total_bits": 630355968, "q_proj": { "group_size": { @@ -133390,7 +133390,7 @@ } }, { - "accuracy": 0.994333280544532, + "accuracy": 0.99435403550926, "total_bits": 637362176, "q_proj": { "group_size": { @@ -133451,7 +133451,7 @@ } }, { - "accuracy": 0.9948401133480825, + "accuracy": 0.994795764355283, "total_bits": 646823936, "q_proj": { "group_size": { @@ -133512,7 +133512,7 @@ } }, { - "accuracy": 0.9968623385617608, + "accuracy": 0.9968650011639846, "total_bits": 784740352, "q_proj": { "group_size": { @@ -133573,7 +133573,7 @@ } }, { - "accuracy": 0.9973834449131238, + "accuracy": 0.9973680502489993, "total_bits": 797818880, "q_proj": { "group_size": { @@ -133634,7 +133634,7 @@ } }, { - "accuracy": 0.9973323394201303, + "accuracy": 0.9973344569535632, "total_bits": 911749120, "q_proj": { "group_size": { @@ -133686,7 +133686,7 @@ } }, { - "accuracy": 0.9988364800810814, + "accuracy": 0.9988629552290628, "total_bits": 942718976, "q_proj": { "group_size": { @@ -133738,7 +133738,7 @@ } }, { - "accuracy": 0.999282259935219, + "accuracy": 0.9992848528843177, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -133792,7 +133792,7 @@ ], "model.layers.68.mlp": [ { - "accuracy": 0.9481778740882874, + "accuracy": 0.9483006565194381, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -133844,7 +133844,7 @@ } }, { - "accuracy": 0.9495776170178464, + "accuracy": 0.9497035428097373, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -133896,7 +133896,7 @@ } }, { - "accuracy": 0.9556498841235512, + "accuracy": 0.9557671452823439, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -133945,7 +133945,7 @@ } }, { - "accuracy": 0.9573746041247719, + "accuracy": 0.9574876709988243, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -133994,7 +133994,7 @@ } }, { - "accuracy": 0.9739435660211664, + "accuracy": 0.9740105537991774, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -134046,7 +134046,7 @@ } }, { - "accuracy": 0.9759573779608074, + "accuracy": 0.9760175802205738, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -134098,7 +134098,7 @@ } }, { - "accuracy": 0.9784879904044302, + "accuracy": 0.9785389962949251, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -134147,7 +134147,7 @@ } }, { - "accuracy": 0.9866601228713989, + "accuracy": 0.986695016685285, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -134190,7 +134190,7 @@ } }, { - "accuracy": 0.9877148427461323, + "accuracy": 0.9877472033626155, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -134233,7 +134233,7 @@ } }, { - "accuracy": 0.9867700605016005, + "accuracy": 0.9868029356002808, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -134285,7 +134285,7 @@ } }, { - "accuracy": 0.9883308018508711, + "accuracy": 0.9883617554840288, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -134337,7 +134337,7 @@ } }, { - "accuracy": 0.9932736785788285, + "accuracy": 0.9932894440073716, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -134389,7 +134389,7 @@ } }, { - "accuracy": 0.9941803765924353, + "accuracy": 0.9941944295638486, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -134441,7 +134441,7 @@ } }, { - "accuracy": 0.996387164451574, + "accuracy": 0.9963956467415157, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -134484,7 +134484,7 @@ } }, { - "accuracy": 0.9965413553934348, + "accuracy": 0.9965485793195272, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -134533,7 +134533,7 @@ } }, { - "accuracy": 0.9970150699740962, + "accuracy": 0.9970203449851588, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -134579,7 +134579,7 @@ } }, { - "accuracy": 0.9989331606775522, + "accuracy": 0.9989349884226134, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -134621,7 +134621,7 @@ ], "model.layers.69.self_attn": [ { - "accuracy": 0.9719282984733582, + "accuracy": 0.9731421188304299, "total_bits": 320757760, "q_proj": { "group_size": { @@ -134685,7 +134685,7 @@ } }, { - "accuracy": 0.9734755346649572, + "accuracy": 0.9742942515172457, "total_bits": 329080832, "q_proj": { "group_size": { @@ -134749,7 +134749,7 @@ } }, { - "accuracy": 0.9811996033317164, + "accuracy": 0.9812786767357274, "total_bits": 336024576, "q_proj": { "group_size": { @@ -134813,7 +134813,7 @@ } }, { - "accuracy": 0.9840275943279266, + "accuracy": 0.9841285598905463, "total_bits": 401557504, "q_proj": { "group_size": { @@ -134877,7 +134877,7 @@ } }, { - "accuracy": 0.9857176761878165, + "accuracy": 0.9857202078166761, "total_bits": 475279360, "q_proj": { "group_size": { @@ -134941,7 +134941,7 @@ } }, { - "accuracy": 0.9861648584667005, + "accuracy": 0.9861245986662412, "total_bits": 475479040, "q_proj": { "group_size": { @@ -135005,7 +135005,7 @@ } }, { - "accuracy": 0.9911245491943861, + "accuracy": 0.9912796279317454, "total_bits": 609759232, "q_proj": { "group_size": { @@ -135057,7 +135057,7 @@ } }, { - "accuracy": 0.9915365757126557, + "accuracy": 0.9916220919082039, "total_bits": 610024448, "q_proj": { "group_size": { @@ -135109,7 +135109,7 @@ } }, { - "accuracy": 0.9921094877155203, + "accuracy": 0.9922799193545392, "total_bits": 615020544, "q_proj": { "group_size": { @@ -135161,7 +135161,7 @@ } }, { - "accuracy": 0.9922796895629481, + "accuracy": 0.992445729280773, "total_bits": 623951872, "q_proj": { "group_size": { @@ -135213,7 +135213,7 @@ } }, { - "accuracy": 0.9918579498403951, + "accuracy": 0.9923241373739744, "total_bits": 626473984, "q_proj": { "group_size": { @@ -135277,7 +135277,7 @@ } }, { - "accuracy": 0.9928624198624962, + "accuracy": 0.9930078897037005, "total_bits": 630355968, "q_proj": { "group_size": { @@ -135341,7 +135341,7 @@ } }, { - "accuracy": 0.9946375246110716, + "accuracy": 0.9948777587790238, "total_bits": 637362176, "q_proj": { "group_size": { @@ -135402,7 +135402,7 @@ } }, { - "accuracy": 0.9951408011348624, + "accuracy": 0.9952517308686909, "total_bits": 646823936, "q_proj": { "group_size": { @@ -135463,7 +135463,7 @@ } }, { - "accuracy": 0.9970644641863672, + "accuracy": 0.9971392693880358, "total_bits": 784740352, "q_proj": { "group_size": { @@ -135524,7 +135524,7 @@ } }, { - "accuracy": 0.9975674364127611, + "accuracy": 0.9976008603055226, "total_bits": 797818880, "q_proj": { "group_size": { @@ -135585,7 +135585,7 @@ } }, { - "accuracy": 0.9974270749248957, + "accuracy": 0.9975023565715865, "total_bits": 911749120, "q_proj": { "group_size": { @@ -135637,7 +135637,7 @@ } }, { - "accuracy": 0.9989273563811654, + "accuracy": 0.9989356443678078, "total_bits": 942718976, "q_proj": { "group_size": { @@ -135689,7 +135689,7 @@ } }, { - "accuracy": 0.9992990270747166, + "accuracy": 0.999307446720961, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -135743,7 +135743,7 @@ ], "model.layers.69.mlp": [ { - "accuracy": 0.9479836915668688, + "accuracy": 0.9481430743869982, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -135795,7 +135795,7 @@ } }, { - "accuracy": 0.9493896051457054, + "accuracy": 0.9495157881786949, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -135847,7 +135847,7 @@ } }, { - "accuracy": 0.9555206738020244, + "accuracy": 0.9556330756137246, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -135896,7 +135896,7 @@ } }, { - "accuracy": 0.9572368922986483, + "accuracy": 0.9573479169293454, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -135945,7 +135945,7 @@ } }, { - "accuracy": 0.9738567863640032, + "accuracy": 0.9739245317484203, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -135997,7 +135997,7 @@ } }, { - "accuracy": 0.9758700869585338, + "accuracy": 0.9759377523472434, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -136049,7 +136049,7 @@ } }, { - "accuracy": 0.9784177698587117, + "accuracy": 0.9784819922949138, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -136098,7 +136098,7 @@ } }, { - "accuracy": 0.9866395137811962, + "accuracy": 0.9866739495804435, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -136141,7 +136141,7 @@ } }, { - "accuracy": 0.9876962751150131, + "accuracy": 0.987731399504762, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -136184,7 +136184,7 @@ } }, { - "accuracy": 0.9867274823941683, + "accuracy": 0.9867617017344424, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -136236,7 +136236,7 @@ } }, { - "accuracy": 0.9882925984106565, + "accuracy": 0.9883242531826621, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -136288,7 +136288,7 @@ } }, { - "accuracy": 0.9932580558877242, + "accuracy": 0.9932741601216165, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -136340,7 +136340,7 @@ } }, { - "accuracy": 0.994167587082637, + "accuracy": 0.994182365113183, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -136392,7 +136392,7 @@ } }, { - "accuracy": 0.9963955965481306, + "accuracy": 0.9964054309223828, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -136435,7 +136435,7 @@ } }, { - "accuracy": 0.9965555981585854, + "accuracy": 0.9965614543149346, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -136484,7 +136484,7 @@ } }, { - "accuracy": 0.9970353290830788, + "accuracy": 0.997039689829475, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -136530,7 +136530,7 @@ } }, { - "accuracy": 0.9990238449291179, + "accuracy": 0.9990264846894302, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -136572,7 +136572,7 @@ ], "model.layers.70.self_attn": [ { - "accuracy": 0.9753127458848452, + "accuracy": 0.9756929874420166, "total_bits": 320757760, "q_proj": { "group_size": { @@ -136636,7 +136636,7 @@ } }, { - "accuracy": 0.978001029867875, + "accuracy": 0.9782119534517589, "total_bits": 329080832, "q_proj": { "group_size": { @@ -136700,7 +136700,7 @@ } }, { - "accuracy": 0.9814547598361969, + "accuracy": 0.9815600263445001, "total_bits": 336024576, "q_proj": { "group_size": { @@ -136764,7 +136764,7 @@ } }, { - "accuracy": 0.9855752069699136, + "accuracy": 0.985666982437435, "total_bits": 401557504, "q_proj": { "group_size": { @@ -136828,7 +136828,7 @@ } }, { - "accuracy": 0.9883284647213785, + "accuracy": 0.9883667011010019, "total_bits": 475279360, "q_proj": { "group_size": { @@ -136892,7 +136892,7 @@ } }, { - "accuracy": 0.9884691850135201, + "accuracy": 0.9884860539122632, "total_bits": 475479040, "q_proj": { "group_size": { @@ -136956,7 +136956,7 @@ } }, { - "accuracy": 0.9918487330800608, + "accuracy": 0.9918131020508314, "total_bits": 609759232, "q_proj": { "group_size": { @@ -137008,7 +137008,7 @@ } }, { - "accuracy": 0.9919357895851135, + "accuracy": 0.9919553775536386, "total_bits": 610024448, "q_proj": { "group_size": { @@ -137060,7 +137060,7 @@ } }, { - "accuracy": 0.9926539753612719, + "accuracy": 0.9925335167269957, "total_bits": 615020544, "q_proj": { "group_size": { @@ -137112,7 +137112,7 @@ } }, { - "accuracy": 0.9928939915017078, + "accuracy": 0.9927793827496076, "total_bits": 623951872, "q_proj": { "group_size": { @@ -137164,7 +137164,7 @@ } }, { - "accuracy": 0.993934397634707, + "accuracy": 0.9939424870045561, "total_bits": 626473984, "q_proj": { "group_size": { @@ -137228,7 +137228,7 @@ } }, { - "accuracy": 0.9944100728944728, + "accuracy": 0.9944373047665546, "total_bits": 630355968, "q_proj": { "group_size": { @@ -137292,7 +137292,7 @@ } }, { - "accuracy": 0.9945862681457871, + "accuracy": 0.9946412808800998, "total_bits": 637362176, "q_proj": { "group_size": { @@ -137353,7 +137353,7 @@ } }, { - "accuracy": 0.9950996978502524, + "accuracy": 0.9951387761454833, "total_bits": 646823936, "q_proj": { "group_size": { @@ -137414,7 +137414,7 @@ } }, { - "accuracy": 0.997060423030665, + "accuracy": 0.9970702854426283, "total_bits": 784740352, "q_proj": { "group_size": { @@ -137475,7 +137475,7 @@ } }, { - "accuracy": 0.9975251939735914, + "accuracy": 0.9975201048349079, "total_bits": 797818880, "q_proj": { "group_size": { @@ -137536,7 +137536,7 @@ } }, { - "accuracy": 0.9976680178783441, + "accuracy": 0.9976738777599836, "total_bits": 911749120, "q_proj": { "group_size": { @@ -137588,7 +137588,7 @@ } }, { - "accuracy": 0.9988106539178836, + "accuracy": 0.9988159632408306, "total_bits": 942718976, "q_proj": { "group_size": { @@ -137640,7 +137640,7 @@ } }, { - "accuracy": 0.9993458891189412, + "accuracy": 0.9993457395191255, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -137694,7 +137694,7 @@ ], "model.layers.70.mlp": [ { - "accuracy": 0.9470372388237401, + "accuracy": 0.9471649307953685, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -137746,7 +137746,7 @@ } }, { - "accuracy": 0.9484451475896334, + "accuracy": 0.9485684037208557, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -137798,7 +137798,7 @@ } }, { - "accuracy": 0.9547245471101058, + "accuracy": 0.9548365191409462, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -137847,7 +137847,7 @@ } }, { - "accuracy": 0.9564230316563657, + "accuracy": 0.956532873605427, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -137896,7 +137896,7 @@ } }, { - "accuracy": 0.97333950745432, + "accuracy": 0.9733910090044925, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -137948,7 +137948,7 @@ } }, { - "accuracy": 0.9753995224049217, + "accuracy": 0.9754564997396971, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -138000,7 +138000,7 @@ } }, { - "accuracy": 0.9780077683298212, + "accuracy": 0.9780551226515519, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -138049,7 +138049,7 @@ } }, { - "accuracy": 0.9864019422154677, + "accuracy": 0.986433921675933, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -138092,7 +138092,7 @@ } }, { - "accuracy": 0.9874805180650008, + "accuracy": 0.9875103141132154, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -138135,7 +138135,7 @@ } }, { - "accuracy": 0.9864424859222612, + "accuracy": 0.9864756543385355, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -138187,7 +138187,7 @@ } }, { - "accuracy": 0.9880545335380655, + "accuracy": 0.988085505209471, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -138239,7 +138239,7 @@ } }, { - "accuracy": 0.993100139655565, + "accuracy": 0.9931157254859021, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -138291,7 +138291,7 @@ } }, { - "accuracy": 0.9940385908672684, + "accuracy": 0.9940516619305861, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -138343,7 +138343,7 @@ } }, { - "accuracy": 0.9963042704682601, + "accuracy": 0.9963123508189854, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -138386,7 +138386,7 @@ } }, { - "accuracy": 0.9964646289223119, + "accuracy": 0.9964712019029417, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -138435,7 +138435,7 @@ } }, { - "accuracy": 0.9969491919404582, + "accuracy": 0.9969539922711096, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -138481,7 +138481,7 @@ } }, { - "accuracy": 0.9989598289524254, + "accuracy": 0.9989618696272373, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -138523,7 +138523,7 @@ ], "model.layers.71.self_attn": [ { - "accuracy": 0.9593416264182643, + "accuracy": 0.959749394341519, "total_bits": 320757760, "q_proj": { "group_size": { @@ -138587,7 +138587,7 @@ } }, { - "accuracy": 0.9638629053768358, + "accuracy": 0.962921280609934, "total_bits": 329080832, "q_proj": { "group_size": { @@ -138651,7 +138651,7 @@ } }, { - "accuracy": 0.9728467433076156, + "accuracy": 0.973560443049983, "total_bits": 336024576, "q_proj": { "group_size": { @@ -138715,7 +138715,7 @@ } }, { - "accuracy": 0.9780299569431105, + "accuracy": 0.9786072891009482, "total_bits": 401557504, "q_proj": { "group_size": { @@ -138779,7 +138779,7 @@ } }, { - "accuracy": 0.9813431517073983, + "accuracy": 0.9815290240865004, "total_bits": 475279360, "q_proj": { "group_size": { @@ -138843,7 +138843,7 @@ } }, { - "accuracy": 0.9818414449691772, + "accuracy": 0.9824165858720478, "total_bits": 475479040, "q_proj": { "group_size": { @@ -138907,7 +138907,7 @@ } }, { - "accuracy": 0.9890080659013045, + "accuracy": 0.9892412988763106, "total_bits": 609759232, "q_proj": { "group_size": { @@ -138959,7 +138959,7 @@ } }, { - "accuracy": 0.9897611439228058, + "accuracy": 0.9900229094844115, "total_bits": 610024448, "q_proj": { "group_size": { @@ -139011,7 +139011,7 @@ } }, { - "accuracy": 0.9905684143304825, + "accuracy": 0.9906133883877805, "total_bits": 615020544, "q_proj": { "group_size": { @@ -139063,7 +139063,7 @@ } }, { - "accuracy": 0.9908161437825153, + "accuracy": 0.99089657083938, "total_bits": 623951872, "q_proj": { "group_size": { @@ -139115,7 +139115,7 @@ } }, { - "accuracy": 0.9905224215043219, + "accuracy": 0.9905767456481331, "total_bits": 626473984, "q_proj": { "group_size": { @@ -139179,7 +139179,7 @@ } }, { - "accuracy": 0.9910085060094532, + "accuracy": 0.9912648436270262, "total_bits": 630355968, "q_proj": { "group_size": { @@ -139243,7 +139243,7 @@ } }, { - "accuracy": 0.9933376218143263, + "accuracy": 0.9932230452173635, "total_bits": 637362176, "q_proj": { "group_size": { @@ -139304,7 +139304,7 @@ } }, { - "accuracy": 0.9938937496197852, + "accuracy": 0.9939599903790575, "total_bits": 646823936, "q_proj": { "group_size": { @@ -139365,7 +139365,7 @@ } }, { - "accuracy": 0.9963714049050683, + "accuracy": 0.9964095016843394, "total_bits": 784740352, "q_proj": { "group_size": { @@ -139426,7 +139426,7 @@ } }, { - "accuracy": 0.9968532633624578, + "accuracy": 0.9969101808965206, "total_bits": 797818880, "q_proj": { "group_size": { @@ -139487,7 +139487,7 @@ } }, { - "accuracy": 0.9971767803163905, + "accuracy": 0.9972101865630401, "total_bits": 911749120, "q_proj": { "group_size": { @@ -139539,7 +139539,7 @@ } }, { - "accuracy": 0.9985135463311484, + "accuracy": 0.9985055710728231, "total_bits": 942718976, "q_proj": { "group_size": { @@ -139591,7 +139591,7 @@ } }, { - "accuracy": 0.9992206295167929, + "accuracy": 0.9992433968735369, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -139645,7 +139645,7 @@ ], "model.layers.71.mlp": [ { - "accuracy": 0.9443847442928114, + "accuracy": 0.9445230835362485, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -139697,7 +139697,7 @@ } }, { - "accuracy": 0.9458719178249961, + "accuracy": 0.9460194424579018, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -139749,7 +139749,7 @@ } }, { - "accuracy": 0.9524345429320085, + "accuracy": 0.9525534792950279, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -139798,7 +139798,7 @@ } }, { - "accuracy": 0.9542398327275327, + "accuracy": 0.954355377899973, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -139847,7 +139847,7 @@ } }, { - "accuracy": 0.9720322709334525, + "accuracy": 0.9721018264168187, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -139899,7 +139899,7 @@ } }, { - "accuracy": 0.9741882430879694, + "accuracy": 0.9742551053825178, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -139951,7 +139951,7 @@ } }, { - "accuracy": 0.9769138270302823, + "accuracy": 0.9769744324056726, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -140000,7 +140000,7 @@ } }, { - "accuracy": 0.9857053537117807, + "accuracy": 0.9857411745347475, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -140043,7 +140043,7 @@ } }, { - "accuracy": 0.9868424675966564, + "accuracy": 0.9868772414169813, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -140086,7 +140086,7 @@ } }, { - "accuracy": 0.9857852349155828, + "accuracy": 0.9858238822535464, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -140138,7 +140138,7 @@ } }, { - "accuracy": 0.987464933803207, + "accuracy": 0.987499013543129, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -140190,7 +140190,7 @@ } }, { - "accuracy": 0.9927704247989153, + "accuracy": 0.9927884348128971, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -140242,7 +140242,7 @@ } }, { - "accuracy": 0.9937475237407183, + "accuracy": 0.9937643181336554, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -140294,7 +140294,7 @@ } }, { - "accuracy": 0.9961268101867876, + "accuracy": 0.9961365567226159, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -140337,7 +140337,7 @@ } }, { - "accuracy": 0.9962998393334841, + "accuracy": 0.9963105132705287, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -140386,7 +140386,7 @@ } }, { - "accuracy": 0.9968105629086494, + "accuracy": 0.9968195894831106, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -140432,7 +140432,7 @@ } }, { - "accuracy": 0.9989309446199944, + "accuracy": 0.9989336747676134, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -140474,7 +140474,7 @@ ], "model.layers.72.self_attn": [ { - "accuracy": 0.958266954672964, + "accuracy": 0.9590053809316534, "total_bits": 320757760, "q_proj": { "group_size": { @@ -140538,7 +140538,7 @@ } }, { - "accuracy": 0.9605250640919334, + "accuracy": 0.9598574261916311, "total_bits": 329080832, "q_proj": { "group_size": { @@ -140602,7 +140602,7 @@ } }, { - "accuracy": 0.9705838247349388, + "accuracy": 0.9714048034266421, "total_bits": 336024576, "q_proj": { "group_size": { @@ -140666,7 +140666,7 @@ } }, { - "accuracy": 0.975100112588782, + "accuracy": 0.9758997622289156, "total_bits": 401557504, "q_proj": { "group_size": { @@ -140730,7 +140730,7 @@ } }, { - "accuracy": 0.9800523692055753, + "accuracy": 0.9805179222633964, "total_bits": 475279360, "q_proj": { "group_size": { @@ -140794,7 +140794,7 @@ } }, { - "accuracy": 0.9807554404986533, + "accuracy": 0.9806656696294483, "total_bits": 475479040, "q_proj": { "group_size": { @@ -140858,7 +140858,7 @@ } }, { - "accuracy": 0.9894296401425412, + "accuracy": 0.989623600715085, "total_bits": 609759232, "q_proj": { "group_size": { @@ -140910,7 +140910,7 @@ } }, { - "accuracy": 0.9897790143364354, + "accuracy": 0.9900902031283629, "total_bits": 610024448, "q_proj": { "group_size": { @@ -140962,7 +140962,7 @@ } }, { - "accuracy": 0.9907498116555967, + "accuracy": 0.9906908548191974, "total_bits": 615020544, "q_proj": { "group_size": { @@ -141014,7 +141014,7 @@ } }, { - "accuracy": 0.9910355837721574, + "accuracy": 0.9910058653668353, "total_bits": 623951872, "q_proj": { "group_size": { @@ -141066,7 +141066,7 @@ } }, { - "accuracy": 0.9906069416748849, + "accuracy": 0.9904858313108745, "total_bits": 626473984, "q_proj": { "group_size": { @@ -141130,7 +141130,7 @@ } }, { - "accuracy": 0.9911165206055892, + "accuracy": 0.9913358978535, "total_bits": 630355968, "q_proj": { "group_size": { @@ -141194,7 +141194,7 @@ } }, { - "accuracy": 0.9933118004547922, + "accuracy": 0.993250089256387, "total_bits": 637362176, "q_proj": { "group_size": { @@ -141255,7 +141255,7 @@ } }, { - "accuracy": 0.993826161873968, + "accuracy": 0.9939011982397029, "total_bits": 646823936, "q_proj": { "group_size": { @@ -141316,7 +141316,7 @@ } }, { - "accuracy": 0.9963796574034189, + "accuracy": 0.9964499704931912, "total_bits": 784740352, "q_proj": { "group_size": { @@ -141377,7 +141377,7 @@ } }, { - "accuracy": 0.9968893982862171, + "accuracy": 0.9969506087271791, "total_bits": 797818880, "q_proj": { "group_size": { @@ -141438,7 +141438,7 @@ } }, { - "accuracy": 0.9972879867020407, + "accuracy": 0.9973593511079487, "total_bits": 911749120, "q_proj": { "group_size": { @@ -141490,7 +141490,7 @@ } }, { - "accuracy": 0.9985086961992478, + "accuracy": 0.9985217176573841, "total_bits": 942718976, "q_proj": { "group_size": { @@ -141542,7 +141542,7 @@ } }, { - "accuracy": 0.9992099818528483, + "accuracy": 0.9992093515925502, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -141596,7 +141596,7 @@ ], "model.layers.72.mlp": [ { - "accuracy": 0.9409767452039217, + "accuracy": 0.9411135096299021, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -141648,7 +141648,7 @@ } }, { - "accuracy": 0.942588115993299, + "accuracy": 0.9427082036670885, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -141700,7 +141700,7 @@ } }, { - "accuracy": 0.9494850259078176, + "accuracy": 0.9495874360988015, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -141749,7 +141749,7 @@ } }, { - "accuracy": 0.9514875976662887, + "accuracy": 0.9515890196750039, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -141798,7 +141798,7 @@ } }, { - "accuracy": 0.9703451646001715, + "accuracy": 0.9704094184072394, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -141850,7 +141850,7 @@ } }, { - "accuracy": 0.9726317744506033, + "accuracy": 0.9726912787086085, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -141902,7 +141902,7 @@ } }, { - "accuracy": 0.97549769282341, + "accuracy": 0.9755489010559885, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -141951,7 +141951,7 @@ } }, { - "accuracy": 0.9847571598856073, + "accuracy": 0.9847905494664845, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -141994,7 +141994,7 @@ } }, { - "accuracy": 0.9859671012351388, + "accuracy": 0.985996823561819, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -142037,7 +142037,7 @@ } }, { - "accuracy": 0.984940441031205, + "accuracy": 0.9849710323308644, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -142089,7 +142089,7 @@ } }, { - "accuracy": 0.9867189134422102, + "accuracy": 0.9867459817936546, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -142141,7 +142141,7 @@ } }, { - "accuracy": 0.9923403270934757, + "accuracy": 0.9923578712501024, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -142193,7 +142193,7 @@ } }, { - "accuracy": 0.9933749900052422, + "accuracy": 0.9933894229562659, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -142245,7 +142245,7 @@ } }, { - "accuracy": 0.9958736994548848, + "accuracy": 0.9958823095811041, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -142288,7 +142288,7 @@ } }, { - "accuracy": 0.9960688446697435, + "accuracy": 0.9960776900774554, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -142337,7 +142337,7 @@ } }, { - "accuracy": 0.9966083604254221, + "accuracy": 0.9966157133641996, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -142383,7 +142383,7 @@ } }, { - "accuracy": 0.998797341691036, + "accuracy": 0.9987997261709288, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -142425,7 +142425,7 @@ ], "model.layers.73.self_attn": [ { - "accuracy": 0.9687358423283225, + "accuracy": 0.9688164874127037, "total_bits": 320757760, "q_proj": { "group_size": { @@ -142489,7 +142489,7 @@ } }, { - "accuracy": 0.9698699336302908, + "accuracy": 0.9702643250164232, "total_bits": 329080832, "q_proj": { "group_size": { @@ -142553,7 +142553,7 @@ } }, { - "accuracy": 0.973290923394655, + "accuracy": 0.9734862980089689, "total_bits": 336024576, "q_proj": { "group_size": { @@ -142617,7 +142617,7 @@ } }, { - "accuracy": 0.9790390165228593, + "accuracy": 0.9791515610719982, "total_bits": 401557504, "q_proj": { "group_size": { @@ -142681,7 +142681,7 @@ } }, { - "accuracy": 0.9836690598412564, + "accuracy": 0.9836268879865345, "total_bits": 475279360, "q_proj": { "group_size": { @@ -142745,7 +142745,7 @@ } }, { - "accuracy": 0.9838814845210627, + "accuracy": 0.9839303524870622, "total_bits": 475479040, "q_proj": { "group_size": { @@ -142809,7 +142809,7 @@ } }, { - "accuracy": 0.9893126472046501, + "accuracy": 0.9893185590442858, "total_bits": 609759232, "q_proj": { "group_size": { @@ -142861,7 +142861,7 @@ } }, { - "accuracy": 0.9896222313767985, + "accuracy": 0.98968547896335, "total_bits": 610024448, "q_proj": { "group_size": { @@ -142913,7 +142913,7 @@ } }, { - "accuracy": 0.9904547702325018, + "accuracy": 0.9905350663160023, "total_bits": 615020544, "q_proj": { "group_size": { @@ -142965,7 +142965,7 @@ } }, { - "accuracy": 0.9907562505257758, + "accuracy": 0.9909018623201471, "total_bits": 623951872, "q_proj": { "group_size": { @@ -143017,7 +143017,7 @@ } }, { - "accuracy": 0.9912498303149876, + "accuracy": 0.9913553423003146, "total_bits": 626473984, "q_proj": { "group_size": { @@ -143081,7 +143081,7 @@ } }, { - "accuracy": 0.9916952072005523, + "accuracy": 0.9918488703275982, "total_bits": 630355968, "q_proj": { "group_size": { @@ -143145,7 +143145,7 @@ } }, { - "accuracy": 0.9928448027686069, + "accuracy": 0.9928811161141646, "total_bits": 637362176, "q_proj": { "group_size": { @@ -143206,7 +143206,7 @@ } }, { - "accuracy": 0.9933922628038808, + "accuracy": 0.9934719306857962, "total_bits": 646823936, "q_proj": { "group_size": { @@ -143267,7 +143267,7 @@ } }, { - "accuracy": 0.9960441012915812, + "accuracy": 0.9960852630044285, "total_bits": 784740352, "q_proj": { "group_size": { @@ -143328,7 +143328,7 @@ } }, { - "accuracy": 0.9966483155363485, + "accuracy": 0.9966549191035723, "total_bits": 797818880, "q_proj": { "group_size": { @@ -143389,7 +143389,7 @@ } }, { - "accuracy": 0.9970357559229198, + "accuracy": 0.9970747148127932, "total_bits": 911749120, "q_proj": { "group_size": { @@ -143441,7 +143441,7 @@ } }, { - "accuracy": 0.9983462169766426, + "accuracy": 0.9983641394463024, "total_bits": 942718976, "q_proj": { "group_size": { @@ -143493,7 +143493,7 @@ } }, { - "accuracy": 0.9991984448248619, + "accuracy": 0.9991999816051439, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -143547,7 +143547,7 @@ ], "model.layers.73.mlp": [ { - "accuracy": 0.9387619307166651, + "accuracy": 0.9388958843130815, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -143599,7 +143599,7 @@ } }, { - "accuracy": 0.940447029314543, + "accuracy": 0.9405795712220042, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -143651,7 +143651,7 @@ } }, { - "accuracy": 0.9476411060283059, + "accuracy": 0.9477591765554327, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -143700,7 +143700,7 @@ } }, { - "accuracy": 0.9496879609007585, + "accuracy": 0.9498017524418078, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -143749,7 +143749,7 @@ } }, { - "accuracy": 0.9690853859248915, + "accuracy": 0.9691559452759592, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -143801,7 +143801,7 @@ } }, { - "accuracy": 0.971556516070115, + "accuracy": 0.9716347424607528, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -143853,7 +143853,7 @@ } }, { - "accuracy": 0.9745109112639176, + "accuracy": 0.9745798785435525, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -143902,7 +143902,7 @@ } }, { - "accuracy": 0.9840573292029532, + "accuracy": 0.9840870076104214, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -143945,7 +143945,7 @@ } }, { - "accuracy": 0.9852917868840066, + "accuracy": 0.9853190497348183, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -143988,7 +143988,7 @@ } }, { - "accuracy": 0.9841610334421459, + "accuracy": 0.9842022171145991, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -144040,7 +144040,7 @@ } }, { - "accuracy": 0.9860794701074299, + "accuracy": 0.9861072395977221, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -144092,7 +144092,7 @@ } }, { - "accuracy": 0.99187402662478, + "accuracy": 0.991896675605523, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -144144,7 +144144,7 @@ } }, { - "accuracy": 0.9930302234072435, + "accuracy": 0.9930433827011209, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -144196,7 +144196,7 @@ } }, { - "accuracy": 0.9955854015915018, + "accuracy": 0.9955962727728643, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -144239,7 +144239,7 @@ } }, { - "accuracy": 0.9958063654209438, + "accuracy": 0.995822286135272, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -144288,7 +144288,7 @@ } }, { - "accuracy": 0.9963451106297342, + "accuracy": 0.9963608486087698, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -144334,7 +144334,7 @@ } }, { - "accuracy": 0.9987503977394417, + "accuracy": 0.9987528853510556, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -144376,7 +144376,7 @@ ], "model.layers.74.self_attn": [ { - "accuracy": 0.9532695349894071, + "accuracy": 0.9520598399011713, "total_bits": 320757760, "q_proj": { "group_size": { @@ -144440,7 +144440,7 @@ } }, { - "accuracy": 0.9549891164428309, + "accuracy": 0.9521704974927401, "total_bits": 329080832, "q_proj": { "group_size": { @@ -144504,7 +144504,7 @@ } }, { - "accuracy": 0.967025615667042, + "accuracy": 0.9671432156311839, "total_bits": 336024576, "q_proj": { "group_size": { @@ -144568,7 +144568,7 @@ } }, { - "accuracy": 0.9729490248780501, + "accuracy": 0.9732759877255088, "total_bits": 401557504, "q_proj": { "group_size": { @@ -144632,7 +144632,7 @@ } }, { - "accuracy": 0.9771636520561419, + "accuracy": 0.9775018237139049, "total_bits": 475279360, "q_proj": { "group_size": { @@ -144696,7 +144696,7 @@ } }, { - "accuracy": 0.9784928434773495, + "accuracy": 0.9784553521557858, "total_bits": 475479040, "q_proj": { "group_size": { @@ -144760,7 +144760,7 @@ } }, { - "accuracy": 0.9869983658978814, + "accuracy": 0.9873311754904295, "total_bits": 609759232, "q_proj": { "group_size": { @@ -144812,7 +144812,7 @@ } }, { - "accuracy": 0.9880109554842899, + "accuracy": 0.9880616837426236, "total_bits": 610024448, "q_proj": { "group_size": { @@ -144864,7 +144864,7 @@ } }, { - "accuracy": 0.9889527520066813, + "accuracy": 0.9891287480529986, "total_bits": 615020544, "q_proj": { "group_size": { @@ -144916,7 +144916,7 @@ } }, { - "accuracy": 0.9892723999525371, + "accuracy": 0.9895200172537252, "total_bits": 623951872, "q_proj": { "group_size": { @@ -144968,7 +144968,7 @@ } }, { - "accuracy": 0.9895799356071573, + "accuracy": 0.9887377188393944, "total_bits": 626473984, "q_proj": { "group_size": { @@ -145032,7 +145032,7 @@ } }, { - "accuracy": 0.9895975174088227, + "accuracy": 0.9895497607557397, "total_bits": 630355968, "q_proj": { "group_size": { @@ -145096,7 +145096,7 @@ } }, { - "accuracy": 0.9919582872014296, + "accuracy": 0.9921470927564722, "total_bits": 637362176, "q_proj": { "group_size": { @@ -145157,7 +145157,7 @@ } }, { - "accuracy": 0.9926978741821489, + "accuracy": 0.9927732160216883, "total_bits": 646823936, "q_proj": { "group_size": { @@ -145218,7 +145218,7 @@ } }, { - "accuracy": 0.9956604430550023, + "accuracy": 0.99561316206267, "total_bits": 784740352, "q_proj": { "group_size": { @@ -145279,7 +145279,7 @@ } }, { - "accuracy": 0.9962868404231573, + "accuracy": 0.9963185657796106, "total_bits": 797818880, "q_proj": { "group_size": { @@ -145340,7 +145340,7 @@ } }, { - "accuracy": 0.9967410950676391, + "accuracy": 0.996674113759869, "total_bits": 911749120, "q_proj": { "group_size": { @@ -145392,7 +145392,7 @@ } }, { - "accuracy": 0.9981542631591621, + "accuracy": 0.9981626087897703, "total_bits": 942718976, "q_proj": { "group_size": { @@ -145444,7 +145444,7 @@ } }, { - "accuracy": 0.9990794752773485, + "accuracy": 0.9990782974974105, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -145498,7 +145498,7 @@ ], "model.layers.74.mlp": [ { - "accuracy": 0.9357352758708753, + "accuracy": 0.9358943638048673, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -145550,7 +145550,7 @@ } }, { - "accuracy": 0.9375437058900532, + "accuracy": 0.9376749490436754, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -145602,7 +145602,7 @@ } }, { - "accuracy": 0.9451386677591425, + "accuracy": 0.9452935582713077, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -145651,7 +145651,7 @@ } }, { - "accuracy": 0.9472972091875578, + "accuracy": 0.9474468231201172, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -145700,7 +145700,7 @@ } }, { - "accuracy": 0.9677153010117381, + "accuracy": 0.9677961468696594, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -145752,7 +145752,7 @@ } }, { - "accuracy": 0.9702465816547996, + "accuracy": 0.9703230073577479, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -145804,7 +145804,7 @@ } }, { - "accuracy": 0.9733567394708332, + "accuracy": 0.9734304892389398, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -145853,7 +145853,7 @@ } }, { - "accuracy": 0.9833957615651583, + "accuracy": 0.9834315243520235, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -145896,7 +145896,7 @@ } }, { - "accuracy": 0.984704924257178, + "accuracy": 0.9847326325742822, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -145939,7 +145939,7 @@ } }, { - "accuracy": 0.9835391201471028, + "accuracy": 0.9835718525083441, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -145991,7 +145991,7 @@ } }, { - "accuracy": 0.9855146031630667, + "accuracy": 0.9855485150688573, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -146043,7 +146043,7 @@ } }, { - "accuracy": 0.9915904983093864, + "accuracy": 0.9916136382441771, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -146095,7 +146095,7 @@ } }, { - "accuracy": 0.9927605672886497, + "accuracy": 0.9927773914839092, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -146147,7 +146147,7 @@ } }, { - "accuracy": 0.9954525271528646, + "accuracy": 0.9954650115809942, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -146190,7 +146190,7 @@ } }, { - "accuracy": 0.9956788946139185, + "accuracy": 0.995691679418087, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -146239,7 +146239,7 @@ } }, { - "accuracy": 0.9962581725496995, + "accuracy": 0.9962696005639277, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -146285,7 +146285,7 @@ } }, { - "accuracy": 0.9987028423500689, + "accuracy": 0.998706318536087, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -146327,7 +146327,7 @@ ], "model.layers.75.self_attn": [ { - "accuracy": 0.9504855557491905, + "accuracy": 0.9511424177571347, "total_bits": 320757760, "q_proj": { "group_size": { @@ -146391,7 +146391,7 @@ } }, { - "accuracy": 0.9527984104658428, + "accuracy": 0.9536239944006267, "total_bits": 329080832, "q_proj": { "group_size": { @@ -146455,7 +146455,7 @@ } }, { - "accuracy": 0.9605616017391807, + "accuracy": 0.9607455699067367, "total_bits": 336024576, "q_proj": { "group_size": { @@ -146519,7 +146519,7 @@ } }, { - "accuracy": 0.969519621447513, + "accuracy": 0.9696345235172071, "total_bits": 401557504, "q_proj": { "group_size": { @@ -146583,7 +146583,7 @@ } }, { - "accuracy": 0.9749437978393153, + "accuracy": 0.9751238415115758, "total_bits": 475279360, "q_proj": { "group_size": { @@ -146647,7 +146647,7 @@ } }, { - "accuracy": 0.9753653485524026, + "accuracy": 0.9757265448570251, "total_bits": 475479040, "q_proj": { "group_size": { @@ -146711,7 +146711,7 @@ } }, { - "accuracy": 0.9859468027165061, + "accuracy": 0.9858428835868835, "total_bits": 609759232, "q_proj": { "group_size": { @@ -146763,7 +146763,7 @@ } }, { - "accuracy": 0.9862361283678758, + "accuracy": 0.9863356916527999, "total_bits": 610024448, "q_proj": { "group_size": { @@ -146815,7 +146815,7 @@ } }, { - "accuracy": 0.9871686987186733, + "accuracy": 0.9871953890511864, "total_bits": 615020544, "q_proj": { "group_size": { @@ -146867,7 +146867,7 @@ } }, { - "accuracy": 0.9876733539920104, + "accuracy": 0.9876802328385805, "total_bits": 623951872, "q_proj": { "group_size": { @@ -146919,7 +146919,7 @@ } }, { - "accuracy": 0.9873188639941969, + "accuracy": 0.9870878324696892, "total_bits": 626473984, "q_proj": { "group_size": { @@ -146983,7 +146983,7 @@ } }, { - "accuracy": 0.9882075355241173, + "accuracy": 0.9882550396417317, "total_bits": 630355968, "q_proj": { "group_size": { @@ -147047,7 +147047,7 @@ } }, { - "accuracy": 0.990222617983818, + "accuracy": 0.9902215545114718, "total_bits": 637362176, "q_proj": { "group_size": { @@ -147108,7 +147108,7 @@ } }, { - "accuracy": 0.9909159792096991, + "accuracy": 0.9909958831573787, "total_bits": 646823936, "q_proj": { "group_size": { @@ -147169,7 +147169,7 @@ } }, { - "accuracy": 0.9946545076213384, + "accuracy": 0.9946737736463547, "total_bits": 784740352, "q_proj": { "group_size": { @@ -147230,7 +147230,7 @@ } }, { - "accuracy": 0.9954398654793438, + "accuracy": 0.9954672491079882, "total_bits": 797818880, "q_proj": { "group_size": { @@ -147291,7 +147291,7 @@ } }, { - "accuracy": 0.9962535547582727, + "accuracy": 0.996261819020698, "total_bits": 911749120, "q_proj": { "group_size": { @@ -147343,7 +147343,7 @@ } }, { - "accuracy": 0.99769839369937, + "accuracy": 0.997713955609422, "total_bits": 942718976, "q_proj": { "group_size": { @@ -147395,7 +147395,7 @@ } }, { - "accuracy": 0.9989935860626007, + "accuracy": 0.9989922830933019, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -147449,7 +147449,7 @@ ], "model.layers.75.mlp": [ { - "accuracy": 0.9325390865928248, + "accuracy": 0.9326740064119038, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -147501,7 +147501,7 @@ } }, { - "accuracy": 0.9343912162278828, + "accuracy": 0.9345267885609677, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -147553,7 +147553,7 @@ } }, { - "accuracy": 0.9423857299905074, + "accuracy": 0.9425097703933716, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -147602,7 +147602,7 @@ } }, { - "accuracy": 0.9447897609911466, + "accuracy": 0.94491151759499, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -147651,7 +147651,7 @@ } }, { - "accuracy": 0.9660165529502065, + "accuracy": 0.966078868037776, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -147703,7 +147703,7 @@ } }, { - "accuracy": 0.9686701454614338, + "accuracy": 0.9687256373857197, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -147755,7 +147755,7 @@ } }, { - "accuracy": 0.9720105215122825, + "accuracy": 0.9720590271447834, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -147804,7 +147804,7 @@ } }, { - "accuracy": 0.9823982699921257, + "accuracy": 0.9824323936512596, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -147847,7 +147847,7 @@ } }, { - "accuracy": 0.9838235880199232, + "accuracy": 0.9838601821347287, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -147890,7 +147890,7 @@ } }, { - "accuracy": 0.9826624550317463, + "accuracy": 0.9826893382950833, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -147942,7 +147942,7 @@ } }, { - "accuracy": 0.9847391875166642, + "accuracy": 0.9847717253785384, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -147994,7 +147994,7 @@ } }, { - "accuracy": 0.9911509273867858, + "accuracy": 0.9911652019149378, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -148046,7 +148046,7 @@ } }, { - "accuracy": 0.9923728077035201, + "accuracy": 0.9923925807601527, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -148098,7 +148098,7 @@ } }, { - "accuracy": 0.9951968871449169, + "accuracy": 0.9952049851417542, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -148141,7 +148141,7 @@ } }, { - "accuracy": 0.9954509495904571, + "accuracy": 0.9954607282814226, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -148190,7 +148190,7 @@ } }, { - "accuracy": 0.9960785637560644, + "accuracy": 0.9960861751123479, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -148236,7 +148236,7 @@ } }, { - "accuracy": 0.9986008677239481, + "accuracy": 0.9986032779867712, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -148278,7 +148278,7 @@ ], "model.layers.76.self_attn": [ { - "accuracy": 0.9576328302684584, + "accuracy": 0.9598539032434162, "total_bits": 320757760, "q_proj": { "group_size": { @@ -148342,7 +148342,7 @@ } }, { - "accuracy": 0.9603988904702037, + "accuracy": 0.9612067780996624, "total_bits": 329080832, "q_proj": { "group_size": { @@ -148406,7 +148406,7 @@ } }, { - "accuracy": 0.9692210868785256, + "accuracy": 0.9696876500782213, "total_bits": 336024576, "q_proj": { "group_size": { @@ -148470,7 +148470,7 @@ } }, { - "accuracy": 0.9741584878218802, + "accuracy": 0.9747306218272761, "total_bits": 401557504, "q_proj": { "group_size": { @@ -148534,7 +148534,7 @@ } }, { - "accuracy": 0.9804136658969679, + "accuracy": 0.9800062414846922, "total_bits": 475279360, "q_proj": { "group_size": { @@ -148598,7 +148598,7 @@ } }, { - "accuracy": 0.9791977389862663, + "accuracy": 0.9804688770520059, "total_bits": 475479040, "q_proj": { "group_size": { @@ -148662,7 +148662,7 @@ } }, { - "accuracy": 0.988084730349089, + "accuracy": 0.9882607169841465, "total_bits": 609759232, "q_proj": { "group_size": { @@ -148714,7 +148714,7 @@ } }, { - "accuracy": 0.9886295552316465, + "accuracy": 0.9889381433788099, "total_bits": 610024448, "q_proj": { "group_size": { @@ -148766,7 +148766,7 @@ } }, { - "accuracy": 0.9898896138919028, + "accuracy": 0.989904268791801, "total_bits": 615020544, "q_proj": { "group_size": { @@ -148818,7 +148818,7 @@ } }, { - "accuracy": 0.9903094125421423, + "accuracy": 0.9902000772325616, "total_bits": 623951872, "q_proj": { "group_size": { @@ -148870,7 +148870,7 @@ } }, { - "accuracy": 0.9897474419129523, + "accuracy": 0.9901906777369348, "total_bits": 626473984, "q_proj": { "group_size": { @@ -148934,7 +148934,7 @@ } }, { - "accuracy": 0.9905012924420206, + "accuracy": 0.9904350897199229, "total_bits": 630355968, "q_proj": { "group_size": { @@ -148998,7 +148998,7 @@ } }, { - "accuracy": 0.9925915559655741, + "accuracy": 0.9927126773093876, "total_bits": 637362176, "q_proj": { "group_size": { @@ -149059,7 +149059,7 @@ } }, { - "accuracy": 0.9931174053957588, + "accuracy": 0.9932632477659928, "total_bits": 646823936, "q_proj": { "group_size": { @@ -149120,7 +149120,7 @@ } }, { - "accuracy": 0.9959662211568732, + "accuracy": 0.9960171772461188, "total_bits": 784740352, "q_proj": { "group_size": { @@ -149181,7 +149181,7 @@ } }, { - "accuracy": 0.996578015779194, + "accuracy": 0.9966133805482011, "total_bits": 797818880, "q_proj": { "group_size": { @@ -149242,7 +149242,7 @@ } }, { - "accuracy": 0.9969893260613868, + "accuracy": 0.9970465012286839, "total_bits": 911749120, "q_proj": { "group_size": { @@ -149294,7 +149294,7 @@ } }, { - "accuracy": 0.9983010743009416, + "accuracy": 0.9983243238376943, "total_bits": 942718976, "q_proj": { "group_size": { @@ -149346,7 +149346,7 @@ } }, { - "accuracy": 0.9991908261160317, + "accuracy": 0.9991836522362734, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -149400,7 +149400,7 @@ ], "model.layers.76.mlp": [ { - "accuracy": 0.9165186003634804, + "accuracy": 0.9154162595146581, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -149452,7 +149452,7 @@ } }, { - "accuracy": 0.9214056353819997, + "accuracy": 0.9174676757109792, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -149504,7 +149504,7 @@ } }, { - "accuracy": 0.9253076252184416, + "accuracy": 0.9271793490961978, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -149553,7 +149553,7 @@ } }, { - "accuracy": 0.9281382623471712, + "accuracy": 0.9300364444130346, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -149602,7 +149602,7 @@ } }, { - "accuracy": 0.9627170500002409, + "accuracy": 0.9623225111710398, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -149654,7 +149654,7 @@ } }, { - "accuracy": 0.9643168449401855, + "accuracy": 0.9650987010253103, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -149706,7 +149706,7 @@ } }, { - "accuracy": 0.9680661151283666, + "accuracy": 0.9688891360634252, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -149755,7 +149755,7 @@ } }, { - "accuracy": 0.9801758355215976, + "accuracy": 0.9802883800707365, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -149798,7 +149798,7 @@ } }, { - "accuracy": 0.9819759506928293, + "accuracy": 0.9820773428992221, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -149841,7 +149841,7 @@ } }, { - "accuracy": 0.9809661570348238, + "accuracy": 0.9805129390013846, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -149893,7 +149893,7 @@ } }, { - "accuracy": 0.9830994402107439, + "accuracy": 0.9831525761830179, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -149945,7 +149945,7 @@ } }, { - "accuracy": 0.9902128475277048, + "accuracy": 0.990229562709206, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -149997,7 +149997,7 @@ } }, { - "accuracy": 0.9915821834614402, + "accuracy": 0.9915412413446527, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -150049,7 +150049,7 @@ } }, { - "accuracy": 0.9945259309912983, + "accuracy": 0.9945399584738832, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -150092,7 +150092,7 @@ } }, { - "accuracy": 0.9949488808449946, + "accuracy": 0.994964884299981, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -150141,7 +150141,7 @@ } }, { - "accuracy": 0.9957180395722389, + "accuracy": 0.9957343928123775, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -150187,7 +150187,7 @@ } }, { - "accuracy": 0.9983322473340913, + "accuracy": 0.9983408251091054, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -150229,7 +150229,7 @@ ], "model.layers.77.self_attn": [ { - "accuracy": 0.9555310512843885, + "accuracy": 0.9546028250142148, "total_bits": 320757760, "q_proj": { "group_size": { @@ -150293,7 +150293,7 @@ } }, { - "accuracy": 0.9562044300531086, + "accuracy": 0.956999493272681, "total_bits": 329080832, "q_proj": { "group_size": { @@ -150357,7 +150357,7 @@ } }, { - "accuracy": 0.9649047569224709, + "accuracy": 0.965076678677609, "total_bits": 336024576, "q_proj": { "group_size": { @@ -150421,7 +150421,7 @@ } }, { - "accuracy": 0.9714704814710116, + "accuracy": 0.9718132364122491, "total_bits": 401557504, "q_proj": { "group_size": { @@ -150485,7 +150485,7 @@ } }, { - "accuracy": 0.9775286254129911, + "accuracy": 0.9775706388448414, "total_bits": 475279360, "q_proj": { "group_size": { @@ -150549,7 +150549,7 @@ } }, { - "accuracy": 0.9776585400104523, + "accuracy": 0.9774357425539117, "total_bits": 475479040, "q_proj": { "group_size": { @@ -150613,7 +150613,7 @@ } }, { - "accuracy": 0.9877475680489289, + "accuracy": 0.9876045631734949, "total_bits": 609759232, "q_proj": { "group_size": { @@ -150665,7 +150665,7 @@ } }, { - "accuracy": 0.9879247381498939, + "accuracy": 0.9882702145137285, "total_bits": 610024448, "q_proj": { "group_size": { @@ -150717,7 +150717,7 @@ } }, { - "accuracy": 0.988276503588024, + "accuracy": 0.9886605880762401, "total_bits": 615020544, "q_proj": { "group_size": { @@ -150769,7 +150769,7 @@ } }, { - "accuracy": 0.9887568950653076, + "accuracy": 0.9891019683135184, "total_bits": 623951872, "q_proj": { "group_size": { @@ -150821,7 +150821,7 @@ } }, { - "accuracy": 0.9889212768328818, + "accuracy": 0.9887682246534448, "total_bits": 626473984, "q_proj": { "group_size": { @@ -150885,7 +150885,7 @@ } }, { - "accuracy": 0.9898713003647955, + "accuracy": 0.9897155989157526, "total_bits": 630355968, "q_proj": { "group_size": { @@ -150949,7 +150949,7 @@ } }, { - "accuracy": 0.9919237965031674, + "accuracy": 0.9919874934773696, "total_bits": 637362176, "q_proj": { "group_size": { @@ -151010,7 +151010,7 @@ } }, { - "accuracy": 0.9925362412866793, + "accuracy": 0.9925542489478463, "total_bits": 646823936, "q_proj": { "group_size": { @@ -151071,7 +151071,7 @@ } }, { - "accuracy": 0.9956372835134205, + "accuracy": 0.9956450105497712, "total_bits": 784740352, "q_proj": { "group_size": { @@ -151132,7 +151132,7 @@ } }, { - "accuracy": 0.9962711942039038, + "accuracy": 0.9963011286760631, "total_bits": 797818880, "q_proj": { "group_size": { @@ -151193,7 +151193,7 @@ } }, { - "accuracy": 0.9968896794476008, + "accuracy": 0.9968933447411186, "total_bits": 911749120, "q_proj": { "group_size": { @@ -151245,7 +151245,7 @@ } }, { - "accuracy": 0.998160973779465, + "accuracy": 0.9981745203074656, "total_bits": 942718976, "q_proj": { "group_size": { @@ -151297,7 +151297,7 @@ } }, { - "accuracy": 0.9991594641224334, + "accuracy": 0.9991582902638536, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -151351,7 +151351,7 @@ ], "model.layers.77.mlp": [ { - "accuracy": 0.9265327077162893, + "accuracy": 0.9246674650593808, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -151403,7 +151403,7 @@ } }, { - "accuracy": 0.9284265919735557, + "accuracy": 0.926391413337306, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -151455,7 +151455,7 @@ } }, { - "accuracy": 0.9369054091604132, + "accuracy": 0.9334992672267713, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -151504,7 +151504,7 @@ } }, { - "accuracy": 0.939652587238111, + "accuracy": 0.9361524268200523, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -151553,7 +151553,7 @@ } }, { - "accuracy": 0.9629123493244773, + "accuracy": 0.9630641184355083, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -151605,7 +151605,7 @@ } }, { - "accuracy": 0.966239057089153, + "accuracy": 0.9664167291239688, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -151657,7 +151657,7 @@ } }, { - "accuracy": 0.9699445649197227, + "accuracy": 0.9701298036073384, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -151706,7 +151706,7 @@ } }, { - "accuracy": 0.9806229597643802, + "accuracy": 0.9806497614634665, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -151749,7 +151749,7 @@ } }, { - "accuracy": 0.9822149402216861, + "accuracy": 0.9822695616044497, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -151792,7 +151792,7 @@ } }, { - "accuracy": 0.9811981445864627, + "accuracy": 0.9811966889782956, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -151844,7 +151844,7 @@ } }, { - "accuracy": 0.9834548219254142, + "accuracy": 0.9834777132460946, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -151896,7 +151896,7 @@ } }, { - "accuracy": 0.990325634416781, + "accuracy": 0.9903464097725717, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -151948,7 +151948,7 @@ } }, { - "accuracy": 0.9917018915477552, + "accuracy": 0.9916966620244478, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -152000,7 +152000,7 @@ } }, { - "accuracy": 0.9946297187554208, + "accuracy": 0.9946000713266825, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -152043,7 +152043,7 @@ } }, { - "accuracy": 0.9950180077239087, + "accuracy": 0.9950356950101099, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -152092,7 +152092,7 @@ } }, { - "accuracy": 0.9957410520628879, + "accuracy": 0.9957596761615652, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -152138,7 +152138,7 @@ } }, { - "accuracy": 0.9984598571532651, + "accuracy": 0.9984643107378169, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -152180,7 +152180,7 @@ ], "model.layers.78.self_attn": [ { - "accuracy": 0.9608115177405507, + "accuracy": 0.9609328288781016, "total_bits": 320757760, "q_proj": { "group_size": { @@ -152244,7 +152244,7 @@ } }, { - "accuracy": 0.9621012587296336, + "accuracy": 0.9635737161887319, "total_bits": 329080832, "q_proj": { "group_size": { @@ -152308,7 +152308,7 @@ } }, { - "accuracy": 0.971352712104195, + "accuracy": 0.9717470313373365, "total_bits": 336024576, "q_proj": { "group_size": { @@ -152372,7 +152372,7 @@ } }, { - "accuracy": 0.9768468094499487, + "accuracy": 0.9772938834993463, "total_bits": 401557504, "q_proj": { "group_size": { @@ -152436,7 +152436,7 @@ } }, { - "accuracy": 0.9804564165441614, + "accuracy": 0.9803166593375959, "total_bits": 475279360, "q_proj": { "group_size": { @@ -152500,7 +152500,7 @@ } }, { - "accuracy": 0.981044065011175, + "accuracy": 0.9810376010443035, "total_bits": 475479040, "q_proj": { "group_size": { @@ -152564,7 +152564,7 @@ } }, { - "accuracy": 0.9892457456965196, + "accuracy": 0.9893760241960224, "total_bits": 609759232, "q_proj": { "group_size": { @@ -152616,7 +152616,7 @@ } }, { - "accuracy": 0.9896798800480994, + "accuracy": 0.989600943107354, "total_bits": 610024448, "q_proj": { "group_size": { @@ -152668,7 +152668,7 @@ } }, { - "accuracy": 0.9904145982704664, + "accuracy": 0.990477600379994, "total_bits": 615020544, "q_proj": { "group_size": { @@ -152720,7 +152720,7 @@ } }, { - "accuracy": 0.990713568110215, + "accuracy": 0.9908007375503841, "total_bits": 623951872, "q_proj": { "group_size": { @@ -152772,7 +152772,7 @@ } }, { - "accuracy": 0.9904221061028933, + "accuracy": 0.9905406715054261, "total_bits": 626473984, "q_proj": { "group_size": { @@ -152836,7 +152836,7 @@ } }, { - "accuracy": 0.9910882719253239, + "accuracy": 0.9912443200224325, "total_bits": 630355968, "q_proj": { "group_size": { @@ -152900,7 +152900,7 @@ } }, { - "accuracy": 0.9931860785735281, + "accuracy": 0.9932827761298731, "total_bits": 637362176, "q_proj": { "group_size": { @@ -152961,7 +152961,7 @@ } }, { - "accuracy": 0.993721146724726, + "accuracy": 0.9938445930418215, "total_bits": 646823936, "q_proj": { "group_size": { @@ -153022,7 +153022,7 @@ } }, { - "accuracy": 0.9962413758039474, + "accuracy": 0.9962926557973811, "total_bits": 784740352, "q_proj": { "group_size": { @@ -153083,7 +153083,7 @@ } }, { - "accuracy": 0.9968583738725436, + "accuracy": 0.9969092354570565, "total_bits": 797818880, "q_proj": { "group_size": { @@ -153144,7 +153144,7 @@ } }, { - "accuracy": 0.9972342140972614, + "accuracy": 0.9972586637656939, "total_bits": 911749120, "q_proj": { "group_size": { @@ -153196,7 +153196,7 @@ } }, { - "accuracy": 0.9984208420899353, + "accuracy": 0.9984186587757186, "total_bits": 942718976, "q_proj": { "group_size": { @@ -153248,7 +153248,7 @@ } }, { - "accuracy": 0.9992402038095813, + "accuracy": 0.9992476062064892, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -153302,7 +153302,7 @@ ], "model.layers.78.mlp": [ { - "accuracy": 0.9016723946521157, + "accuracy": 0.9017260702032792, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -153354,7 +153354,7 @@ } }, { - "accuracy": 0.9068721658305118, + "accuracy": 0.9067678137829429, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -153406,7 +153406,7 @@ } }, { - "accuracy": 0.9143687172939903, + "accuracy": 0.9147176868037173, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -153455,7 +153455,7 @@ } }, { - "accuracy": 0.9185661642174971, + "accuracy": 0.9189069271087646, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -153504,7 +153504,7 @@ } }, { - "accuracy": 0.9555696688200298, + "accuracy": 0.9556146671897486, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -153556,7 +153556,7 @@ } }, { - "accuracy": 0.9592325091362, + "accuracy": 0.9596629425099021, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -153608,7 +153608,7 @@ } }, { - "accuracy": 0.9638316035270691, + "accuracy": 0.9642732959044606, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -153657,7 +153657,7 @@ } }, { - "accuracy": 0.9762036204338074, + "accuracy": 0.9761616054334139, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -153700,7 +153700,7 @@ } }, { - "accuracy": 0.978296232850928, + "accuracy": 0.9782806713330118, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -153743,7 +153743,7 @@ } }, { - "accuracy": 0.9771567332117181, + "accuracy": 0.977502095071893, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -153795,7 +153795,7 @@ } }, { - "accuracy": 0.9804113773923171, + "accuracy": 0.9804548072187524, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -153847,7 +153847,7 @@ } }, { - "accuracy": 0.9881863335245534, + "accuracy": 0.9882350791441766, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -153899,7 +153899,7 @@ } }, { - "accuracy": 0.9898369069162168, + "accuracy": 0.9899588768419466, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -153951,7 +153951,7 @@ } }, { - "accuracy": 0.9929252698233253, + "accuracy": 0.9929529734347996, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -153994,7 +153994,7 @@ } }, { - "accuracy": 0.9936153708319915, + "accuracy": 0.9936350948716465, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -154043,7 +154043,7 @@ } }, { - "accuracy": 0.9945418646461085, + "accuracy": 0.9945626227479232, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -154089,7 +154089,7 @@ } }, { - "accuracy": 0.9970294190080542, + "accuracy": 0.9970518580000651, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -154131,7 +154131,7 @@ ], "model.layers.79.self_attn": [ { - "accuracy": 0.9742901811474248, + "accuracy": 0.9744882536561865, "total_bits": 320757760, "q_proj": { "group_size": { @@ -154195,7 +154195,7 @@ } }, { - "accuracy": 0.9754879506010758, + "accuracy": 0.9758317047043851, "total_bits": 329080832, "q_proj": { "group_size": { @@ -154259,7 +154259,7 @@ } }, { - "accuracy": 0.981796572082921, + "accuracy": 0.981840591681631, "total_bits": 336024576, "q_proj": { "group_size": { @@ -154323,7 +154323,7 @@ } }, { - "accuracy": 0.9865622755728269, + "accuracy": 0.9865898254670595, "total_bits": 401557504, "q_proj": { "group_size": { @@ -154387,7 +154387,7 @@ } }, { - "accuracy": 0.9875059598370602, + "accuracy": 0.9875182227084511, "total_bits": 475279360, "q_proj": { "group_size": { @@ -154451,7 +154451,7 @@ } }, { - "accuracy": 0.9877419354099977, + "accuracy": 0.9879516700380727, "total_bits": 475479040, "q_proj": { "group_size": { @@ -154515,7 +154515,7 @@ } }, { - "accuracy": 0.9932194916825545, + "accuracy": 0.993208322870104, "total_bits": 609759232, "q_proj": { "group_size": { @@ -154567,7 +154567,7 @@ } }, { - "accuracy": 0.9934876455288184, + "accuracy": 0.9935055637830182, "total_bits": 610024448, "q_proj": { "group_size": { @@ -154619,7 +154619,7 @@ } }, { - "accuracy": 0.9939487043179964, + "accuracy": 0.9939195890175668, "total_bits": 615020544, "q_proj": { "group_size": { @@ -154671,7 +154671,7 @@ } }, { - "accuracy": 0.9941644958759609, + "accuracy": 0.994135538998403, "total_bits": 623951872, "q_proj": { "group_size": { @@ -154723,7 +154723,7 @@ } }, { - "accuracy": 0.9937432671063825, + "accuracy": 0.9938993948070627, "total_bits": 626473984, "q_proj": { "group_size": { @@ -154787,7 +154787,7 @@ } }, { - "accuracy": 0.9942606973804926, + "accuracy": 0.9942472686893061, "total_bits": 630355968, "q_proj": { "group_size": { @@ -154851,7 +154851,7 @@ } }, { - "accuracy": 0.995505802333355, + "accuracy": 0.9955289618749368, "total_bits": 637362176, "q_proj": { "group_size": { @@ -154912,7 +154912,7 @@ } }, { - "accuracy": 0.9958669507974073, + "accuracy": 0.9959371901656452, "total_bits": 646823936, "q_proj": { "group_size": { @@ -154973,7 +154973,7 @@ } }, { - "accuracy": 0.9975266823250997, + "accuracy": 0.9975318393032802, "total_bits": 784740352, "q_proj": { "group_size": { @@ -155034,7 +155034,7 @@ } }, { - "accuracy": 0.9979193189034337, + "accuracy": 0.9979452655503624, "total_bits": 797818880, "q_proj": { "group_size": { @@ -155095,7 +155095,7 @@ } }, { - "accuracy": 0.9982093911814062, + "accuracy": 0.9982059917559749, "total_bits": 911749120, "q_proj": { "group_size": { @@ -155147,7 +155147,7 @@ } }, { - "accuracy": 0.9989787662695897, + "accuracy": 0.9989853911101818, "total_bits": 942718976, "q_proj": { "group_size": { @@ -155199,7 +155199,7 @@ } }, { - "accuracy": 0.9994885818542618, + "accuracy": 0.9994978151823345, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -155253,7 +155253,7 @@ ], "model.layers.79.mlp": [ { - "accuracy": 0.9259147895009894, + "accuracy": 0.9259069718812641, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -155305,7 +155305,7 @@ } }, { - "accuracy": 0.9288122026543868, + "accuracy": 0.9288101133547331, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -155357,7 +155357,7 @@ } }, { - "accuracy": 0.9352708051079198, + "accuracy": 0.9352696694825825, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -155406,7 +155406,7 @@ } }, { - "accuracy": 0.937591132364775, + "accuracy": 0.9375925879729421, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -155455,7 +155455,7 @@ } }, { - "accuracy": 0.9632547654603657, + "accuracy": 0.9632977059012965, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -155507,7 +155507,7 @@ } }, { - "accuracy": 0.9668143862172177, + "accuracy": 0.9668587979517484, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -155559,7 +155559,7 @@ } }, { - "accuracy": 0.9694928244540566, + "accuracy": 0.9695284397978532, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -155608,7 +155608,7 @@ } }, { - "accuracy": 0.9809373130923823, + "accuracy": 0.9809528008887642, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -155651,7 +155651,7 @@ } }, { - "accuracy": 0.9823573576776605, + "accuracy": 0.9823656960537559, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -155694,7 +155694,7 @@ } }, { - "accuracy": 0.9811331676809412, + "accuracy": 0.9811395218497828, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -155746,7 +155746,7 @@ } }, { - "accuracy": 0.9838030291231055, + "accuracy": 0.9838232460774874, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -155798,7 +155798,7 @@ } }, { - "accuracy": 0.9902922526786202, + "accuracy": 0.9903029673977902, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -155850,7 +155850,7 @@ } }, { - "accuracy": 0.991854019855198, + "accuracy": 0.9918624931260159, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -155902,7 +155902,7 @@ } }, { - "accuracy": 0.9945175843803506, + "accuracy": 0.9945159213323342, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -155945,7 +155945,7 @@ } }, { - "accuracy": 0.9949754703986017, + "accuracy": 0.9949883669614792, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -155994,7 +155994,7 @@ } }, { - "accuracy": 0.9955358544462606, + "accuracy": 0.9955492839217186, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -156040,7 +156040,7 @@ } }, { - "accuracy": 0.9983863326671877, + "accuracy": 0.9983892271197156, "total_bits": 5660623360, "gate_proj": { "group_size": {