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": 0.882712113229852, + "accuracy": 0.8829772723348517, "total_bits": 320757760, "q_proj": { "group_size": { @@ -66,7 +66,7 @@ } }, { - "accuracy": 0.8875077649166709, + "accuracy": 0.8871288299560547, "total_bits": 329080832, "q_proj": { "group_size": { @@ -130,7 +130,7 @@ } }, { - "accuracy": 0.9015210239510787, + "accuracy": 0.8969089733926874, "total_bits": 336024576, "q_proj": { "group_size": { @@ -194,7 +194,7 @@ } }, { - "accuracy": 0.9058939971421894, + "accuracy": 0.9019804941980463, "total_bits": 401557504, "q_proj": { "group_size": { @@ -258,7 +258,7 @@ } }, { - "accuracy": 0.9029873421317652, + "accuracy": 0.8988138374529386, "total_bits": 475279360, "q_proj": { "group_size": { @@ -322,7 +322,7 @@ } }, { - "accuracy": 0.9223913895456415, + "accuracy": 0.9226437179665816, "total_bits": 475479040, "q_proj": { "group_size": { @@ -386,7 +386,7 @@ } }, { - "accuracy": 0.9045263654307315, + "accuracy": 0.9003585388785914, "total_bits": 609759232, "q_proj": { "group_size": { @@ -438,7 +438,7 @@ } }, { - "accuracy": 0.9240356370022422, + "accuracy": 0.9241100612439608, "total_bits": 610024448, "q_proj": { "group_size": { @@ -490,7 +490,7 @@ } }, { - "accuracy": 0.9508266166636818, + "accuracy": 0.9507242472548234, "total_bits": 615020544, "q_proj": { "group_size": { @@ -542,7 +542,7 @@ } }, { - "accuracy": 0.9511768943385074, + "accuracy": 0.9510418427617926, "total_bits": 623951872, "q_proj": { "group_size": { @@ -594,7 +594,7 @@ } }, { - "accuracy": 0.9569498269181502, + "accuracy": 0.9563334082302294, "total_bits": 626473984, "q_proj": { "group_size": { @@ -658,7 +658,7 @@ } }, { - "accuracy": 0.9733225860093769, + "accuracy": 0.9733267771570306, "total_bits": 630355968, "q_proj": { "group_size": { @@ -722,7 +722,7 @@ } }, { - "accuracy": 0.9573833816929868, + "accuracy": 0.9568781193933988, "total_bits": 637362176, "q_proj": { "group_size": { @@ -783,7 +783,7 @@ } }, { - "accuracy": 0.9739285785900919, + "accuracy": 0.9739301032141635, "total_bits": 646823936, "q_proj": { "group_size": { @@ -844,7 +844,7 @@ } }, { - "accuracy": 0.9710664278582523, + "accuracy": 0.9708444036935505, "total_bits": 784740352, "q_proj": { "group_size": { @@ -905,7 +905,7 @@ } }, { - "accuracy": 0.9860754452253643, + "accuracy": 0.9860426391425886, "total_bits": 797818880, "q_proj": { "group_size": { @@ -966,7 +966,7 @@ } }, { - "accuracy": 0.971195776211588, + "accuracy": 0.9709927063239249, "total_bits": 911749120, "q_proj": { "group_size": { @@ -1018,7 +1018,7 @@ } }, { - "accuracy": 0.9957216307520866, + "accuracy": 0.9957260799251104, "total_bits": 942718976, "q_proj": { "group_size": { @@ -1070,7 +1070,7 @@ } }, { - "accuracy": 0.9916304906732157, + "accuracy": 0.9917721920891812, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -1124,7 +1124,7 @@ ], "model.layers.0.mlp": [ { - "accuracy": 0.9679982034783614, + "accuracy": 0.9693153218219155, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -1176,7 +1176,7 @@ } }, { - "accuracy": 0.9697063659366808, + "accuracy": 0.9723150886987385, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -1228,7 +1228,7 @@ } }, { - "accuracy": 0.9731203380383944, + "accuracy": 0.9749206288864738, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -1277,7 +1277,7 @@ } }, { - "accuracy": 0.9744064243216264, + "accuracy": 0.9763068054851732, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -1326,7 +1326,7 @@ } }, { - "accuracy": 0.9863242601093493, + "accuracy": 0.9852936754101201, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -1378,7 +1378,7 @@ } }, { - "accuracy": 0.9908170370679152, + "accuracy": 0.990641787648201, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -1430,7 +1430,7 @@ } }, { - "accuracy": 0.9921177296262038, + "accuracy": 0.9919063895940781, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -1479,7 +1479,7 @@ } }, { - "accuracy": 0.9919839736662412, + "accuracy": 0.992201431017173, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -1522,7 +1522,7 @@ } }, { - "accuracy": 0.9942030910598604, + "accuracy": 0.9940490157980668, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -1565,7 +1565,7 @@ } }, { - "accuracy": 0.9929208943718358, + "accuracy": 0.9928872538240332, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -1617,7 +1617,7 @@ } }, { - "accuracy": 0.9952130866678137, + "accuracy": 0.9953875545608369, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -1669,7 +1669,7 @@ } }, { - "accuracy": 0.9952480804763342, + "accuracy": 0.9953974771656489, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -1721,7 +1721,7 @@ } }, { - "accuracy": 0.9970410623048481, + "accuracy": 0.997085116999714, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -1773,7 +1773,7 @@ } }, { - "accuracy": 0.9971312388385597, + "accuracy": 0.9971900323503896, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -1816,7 +1816,7 @@ } }, { - "accuracy": 0.9974316154655657, + "accuracy": 0.9974763660054458, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -1865,7 +1865,7 @@ } }, { - "accuracy": 0.9975916847194496, + "accuracy": 0.9976383014337012, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -1911,7 +1911,7 @@ } }, { - "accuracy": 0.9981952454698714, + "accuracy": 0.9982064409475577, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -1953,7 +1953,7 @@ ], "model.layers.1.self_attn": [ { - "accuracy": 0.9345038250872963, + "accuracy": 0.9353077286168149, "total_bits": 320757760, "q_proj": { "group_size": { @@ -2017,7 +2017,7 @@ } }, { - "accuracy": 0.9402142323945698, + "accuracy": 0.9405258329291093, "total_bits": 329080832, "q_proj": { "group_size": { @@ -2081,7 +2081,7 @@ } }, { - "accuracy": 0.9570221555860419, + "accuracy": 0.9548656846347608, "total_bits": 336024576, "q_proj": { "group_size": { @@ -2145,7 +2145,7 @@ } }, { - "accuracy": 0.9663761697317425, + "accuracy": 0.9660760509340387, "total_bits": 401557504, "q_proj": { "group_size": { @@ -2209,7 +2209,7 @@ } }, { - "accuracy": 0.9674931952827855, + "accuracy": 0.967124669175399, "total_bits": 475279360, "q_proj": { "group_size": { @@ -2273,7 +2273,7 @@ } }, { - "accuracy": 0.9640379359847621, + "accuracy": 0.9659519572007029, "total_bits": 475479040, "q_proj": { "group_size": { @@ -2337,7 +2337,7 @@ } }, { - "accuracy": 0.9710221918005693, + "accuracy": 0.971620848304347, "total_bits": 609759232, "q_proj": { "group_size": { @@ -2389,7 +2389,7 @@ } }, { - "accuracy": 0.9665452179155851, + "accuracy": 0.9697729662845009, "total_bits": 610024448, "q_proj": { "group_size": { @@ -2441,7 +2441,7 @@ } }, { - "accuracy": 0.968627066988694, + "accuracy": 0.9750067136789623, "total_bits": 615020544, "q_proj": { "group_size": { @@ -2493,7 +2493,7 @@ } }, { - "accuracy": 0.9689751549770957, + "accuracy": 0.9754095736302828, "total_bits": 623951872, "q_proj": { "group_size": { @@ -2545,7 +2545,7 @@ } }, { - "accuracy": 0.9760667822862926, + "accuracy": 0.9779488977633024, "total_bits": 626473984, "q_proj": { "group_size": { @@ -2609,7 +2609,7 @@ } }, { - "accuracy": 0.9795027437962984, + "accuracy": 0.978923245480186, "total_bits": 630355968, "q_proj": { "group_size": { @@ -2673,7 +2673,7 @@ } }, { - "accuracy": 0.9764939295618158, + "accuracy": 0.978648442971079, "total_bits": 637362176, "q_proj": { "group_size": { @@ -2734,7 +2734,7 @@ } }, { - "accuracy": 0.9803291621961092, + "accuracy": 0.9797004822053408, "total_bits": 646823936, "q_proj": { "group_size": { @@ -2795,7 +2795,7 @@ } }, { - "accuracy": 0.9835296633996462, + "accuracy": 0.9844391722428171, "total_bits": 784740352, "q_proj": { "group_size": { @@ -2856,7 +2856,7 @@ } }, { - "accuracy": 0.9870956798917369, + "accuracy": 0.9872110125265623, "total_bits": 797818880, "q_proj": { "group_size": { @@ -2917,7 +2917,7 @@ } }, { - "accuracy": 0.9838812617879165, + "accuracy": 0.9849762900879508, "total_bits": 911749120, "q_proj": { "group_size": { @@ -2969,7 +2969,7 @@ } }, { - "accuracy": 0.9957109885780435, + "accuracy": 0.9955952598860389, "total_bits": 942718976, "q_proj": { "group_size": { @@ -3021,7 +3021,7 @@ } }, { - "accuracy": 0.9949270569180187, + "accuracy": 0.9951779026734201, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -3075,7 +3075,7 @@ ], "model.layers.1.mlp": [ { - "accuracy": 0.9636296818130895, + "accuracy": 0.9660471238588032, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -3127,7 +3127,7 @@ } }, { - "accuracy": 0.9648283751387345, + "accuracy": 0.9670314192771912, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -3179,7 +3179,7 @@ } }, { - "accuracy": 0.9707210126676058, + "accuracy": 0.9728078057891444, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -3228,7 +3228,7 @@ } }, { - "accuracy": 0.9726354354306271, + "accuracy": 0.9747436250510969, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -3277,7 +3277,7 @@ } }, { - "accuracy": 0.9822184663069876, + "accuracy": 0.9830711531011682, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -3329,7 +3329,7 @@ } }, { - "accuracy": 0.9844779231046376, + "accuracy": 0.9853199751753556, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -3381,7 +3381,7 @@ } }, { - "accuracy": 0.9871042511965099, + "accuracy": 0.9879218488931656, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -3430,7 +3430,7 @@ } }, { - "accuracy": 0.9908996640067351, + "accuracy": 0.9913273845848284, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -3473,7 +3473,7 @@ } }, { - "accuracy": 0.9919266300766092, + "accuracy": 0.9923539279322875, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -3516,7 +3516,7 @@ } }, { - "accuracy": 0.9909498966053912, + "accuracy": 0.9913923387464724, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -3568,7 +3568,7 @@ } }, { - "accuracy": 0.9925695516561207, + "accuracy": 0.992967564808695, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -3620,7 +3620,7 @@ } }, { - "accuracy": 0.9953060718743425, + "accuracy": 0.9955527441282022, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -3672,7 +3672,7 @@ } }, { - "accuracy": 0.9962322190403938, + "accuracy": 0.9964303464481705, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -3724,7 +3724,7 @@ } }, { - "accuracy": 0.9973503933533242, + "accuracy": 0.9974895219661688, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -3767,7 +3767,7 @@ } }, { - "accuracy": 0.9976117163896561, + "accuracy": 0.9977349644820941, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -3816,7 +3816,7 @@ } }, { - "accuracy": 0.998089674860239, + "accuracy": 0.9982114630310159, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -3862,7 +3862,7 @@ } }, { - "accuracy": 0.9992581057901445, + "accuracy": 0.999285333740868, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -3904,7 +3904,7 @@ ], "model.layers.2.self_attn": [ { - "accuracy": 0.9429669631154913, + "accuracy": 0.9477103415288424, "total_bits": 320757760, "q_proj": { "group_size": { @@ -3968,7 +3968,7 @@ } }, { - "accuracy": 0.9460543017638356, + "accuracy": 0.9506230354309082, "total_bits": 329080832, "q_proj": { "group_size": { @@ -4032,7 +4032,7 @@ } }, { - "accuracy": 0.9596779471949527, + "accuracy": 0.9563302491840563, "total_bits": 336024576, "q_proj": { "group_size": { @@ -4096,7 +4096,7 @@ } }, { - "accuracy": 0.9675747689447904, + "accuracy": 0.9658897801449424, "total_bits": 401557504, "q_proj": { "group_size": { @@ -4160,7 +4160,7 @@ } }, { - "accuracy": 0.9695919783491838, + "accuracy": 0.9678113178202981, "total_bits": 475279360, "q_proj": { "group_size": { @@ -4224,7 +4224,7 @@ } }, { - "accuracy": 0.9679698818608334, + "accuracy": 0.9701801099275288, "total_bits": 475479040, "q_proj": { "group_size": { @@ -4288,7 +4288,7 @@ } }, { - "accuracy": 0.9743048485956693, + "accuracy": 0.9731926949400651, "total_bits": 609759232, "q_proj": { "group_size": { @@ -4340,7 +4340,7 @@ } }, { - "accuracy": 0.9726020536924663, + "accuracy": 0.9758993214682529, "total_bits": 610024448, "q_proj": { "group_size": { @@ -4392,7 +4392,7 @@ } }, { - "accuracy": 0.9775016543112303, + "accuracy": 0.976250559091568, "total_bits": 615020544, "q_proj": { "group_size": { @@ -4444,7 +4444,7 @@ } }, { - "accuracy": 0.9778291692859248, + "accuracy": 0.9766509752524527, "total_bits": 623951872, "q_proj": { "group_size": { @@ -4496,7 +4496,7 @@ } }, { - "accuracy": 0.981070792988727, + "accuracy": 0.9806302127085234, "total_bits": 626473984, "q_proj": { "group_size": { @@ -4560,7 +4560,7 @@ } }, { - "accuracy": 0.9814236477801674, + "accuracy": 0.980520569964459, "total_bits": 630355968, "q_proj": { "group_size": { @@ -4624,7 +4624,7 @@ } }, { - "accuracy": 0.9819275627010747, + "accuracy": 0.9816991184887133, "total_bits": 637362176, "q_proj": { "group_size": { @@ -4685,7 +4685,7 @@ } }, { - "accuracy": 0.9821750756941343, + "accuracy": 0.9817342271930293, "total_bits": 646823936, "q_proj": { "group_size": { @@ -4746,7 +4746,7 @@ } }, { - "accuracy": 0.9846075820295435, + "accuracy": 0.9850820224536093, "total_bits": 784740352, "q_proj": { "group_size": { @@ -4807,7 +4807,7 @@ } }, { - "accuracy": 0.9860277913118664, + "accuracy": 0.9868124409725791, "total_bits": 797818880, "q_proj": { "group_size": { @@ -4868,7 +4868,7 @@ } }, { - "accuracy": 0.9849328414389962, + "accuracy": 0.9855376999629172, "total_bits": 911749120, "q_proj": { "group_size": { @@ -4920,7 +4920,7 @@ } }, { - "accuracy": 0.9950275750536668, + "accuracy": 0.9951489897150743, "total_bits": 942718976, "q_proj": { "group_size": { @@ -4972,7 +4972,7 @@ } }, { - "accuracy": 0.9943904825731328, + "accuracy": 0.9943933004611417, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -5026,7 +5026,7 @@ ], "model.layers.2.mlp": [ { - "accuracy": 0.9546678568187513, + "accuracy": 0.9568619069300199, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -5078,7 +5078,7 @@ } }, { - "accuracy": 0.9558708416788202, + "accuracy": 0.9579651261630812, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -5130,7 +5130,7 @@ } }, { - "accuracy": 0.9630724348519978, + "accuracy": 0.9649813739876998, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -5179,7 +5179,7 @@ } }, { - "accuracy": 0.9651379616636979, + "accuracy": 0.9670103757004989, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -5228,7 +5228,7 @@ } }, { - "accuracy": 0.9767710221441168, + "accuracy": 0.9778620914409035, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -5280,7 +5280,7 @@ } }, { - "accuracy": 0.9800644171865363, + "accuracy": 0.9810298524404827, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -5332,7 +5332,7 @@ } }, { - "accuracy": 0.9829833821246499, + "accuracy": 0.9839037531300595, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -5381,7 +5381,7 @@ } }, { - "accuracy": 0.9884828266344572, + "accuracy": 0.9890311156448565, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -5424,7 +5424,7 @@ } }, { - "accuracy": 0.9899379728656066, + "accuracy": 0.9904321714451438, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -5467,7 +5467,7 @@ } }, { - "accuracy": 0.9879825601452276, + "accuracy": 0.9885472082778027, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -5519,7 +5519,7 @@ } }, { - "accuracy": 0.9904523311476958, + "accuracy": 0.9909106806704873, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -5571,7 +5571,7 @@ } }, { - "accuracy": 0.9937780295547686, + "accuracy": 0.9940608728088831, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -5623,7 +5623,7 @@ } }, { - "accuracy": 0.9952394934861284, + "accuracy": 0.9954741914805613, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -5675,7 +5675,7 @@ } }, { - "accuracy": 0.9966403300824919, + "accuracy": 0.9967868469263378, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -5718,7 +5718,7 @@ } }, { - "accuracy": 0.9967680185248977, + "accuracy": 0.996913617378787, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -5767,7 +5767,7 @@ } }, { - "accuracy": 0.9972429820581487, + "accuracy": 0.9973793529758328, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -5813,7 +5813,7 @@ } }, { - "accuracy": 0.9990788404094545, + "accuracy": 0.9991159573393432, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -5855,7 +5855,7 @@ ], "model.layers.3.self_attn": [ { - "accuracy": 0.9356806403712222, + "accuracy": 0.9323006303686845, "total_bits": 320757760, "q_proj": { "group_size": { @@ -5919,7 +5919,7 @@ } }, { - "accuracy": 0.947466536572105, + "accuracy": 0.9437416540948969, "total_bits": 329080832, "q_proj": { "group_size": { @@ -5983,7 +5983,7 @@ } }, { - "accuracy": 0.9545549624844601, + "accuracy": 0.952155495944776, "total_bits": 336024576, "q_proj": { "group_size": { @@ -6047,7 +6047,7 @@ } }, { - "accuracy": 0.9635144879943446, + "accuracy": 0.9621566188962836, "total_bits": 401557504, "q_proj": { "group_size": { @@ -6111,7 +6111,7 @@ } }, { - "accuracy": 0.9641148604844746, + "accuracy": 0.9627495972733748, "total_bits": 475279360, "q_proj": { "group_size": { @@ -6175,7 +6175,7 @@ } }, { - "accuracy": 0.9665483581392389, + "accuracy": 0.9647820027250993, "total_bits": 475479040, "q_proj": { "group_size": { @@ -6239,7 +6239,7 @@ } }, { - "accuracy": 0.969290996852674, + "accuracy": 0.9688083874551874, "total_bits": 609759232, "q_proj": { "group_size": { @@ -6291,7 +6291,7 @@ } }, { - "accuracy": 0.9717680090352109, + "accuracy": 0.971308099596124, "total_bits": 610024448, "q_proj": { "group_size": { @@ -6343,7 +6343,7 @@ } }, { - "accuracy": 0.9736241792377672, + "accuracy": 0.9731057254891646, "total_bits": 615020544, "q_proj": { "group_size": { @@ -6395,7 +6395,7 @@ } }, { - "accuracy": 0.9739209193932382, + "accuracy": 0.973379345316636, "total_bits": 623951872, "q_proj": { "group_size": { @@ -6447,7 +6447,7 @@ } }, { - "accuracy": 0.9772144383505771, + "accuracy": 0.9768742291550887, "total_bits": 626473984, "q_proj": { "group_size": { @@ -6511,7 +6511,7 @@ } }, { - "accuracy": 0.9791495282399026, + "accuracy": 0.9795826250001004, "total_bits": 630355968, "q_proj": { "group_size": { @@ -6575,7 +6575,7 @@ } }, { - "accuracy": 0.97870293416475, + "accuracy": 0.978502271991027, "total_bits": 637362176, "q_proj": { "group_size": { @@ -6636,7 +6636,7 @@ } }, { - "accuracy": 0.9805299278936888, + "accuracy": 0.9811181943667563, "total_bits": 646823936, "q_proj": { "group_size": { @@ -6697,7 +6697,7 @@ } }, { - "accuracy": 0.9840990430430362, + "accuracy": 0.9828330152913144, "total_bits": 784740352, "q_proj": { "group_size": { @@ -6758,7 +6758,7 @@ } }, { - "accuracy": 0.9880320288633045, + "accuracy": 0.9881509683634105, "total_bits": 797818880, "q_proj": { "group_size": { @@ -6819,7 +6819,7 @@ } }, { - "accuracy": 0.984493120720512, + "accuracy": 0.9832736708615956, "total_bits": 911749120, "q_proj": { "group_size": { @@ -6871,7 +6871,7 @@ } }, { - "accuracy": 0.9963034877651616, + "accuracy": 0.9962223681964373, "total_bits": 942718976, "q_proj": { "group_size": { @@ -6923,7 +6923,7 @@ } }, { - "accuracy": 0.9950510334027441, + "accuracy": 0.994509857344, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -6977,7 +6977,7 @@ ], "model.layers.3.mlp": [ { - "accuracy": 0.8688555767661647, + "accuracy": 0.868526207773309, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -7029,7 +7029,7 @@ } }, { - "accuracy": 0.8729508174093146, + "accuracy": 0.8697502864034552, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -7081,7 +7081,7 @@ } }, { - "accuracy": 0.8726074068169845, + "accuracy": 0.8693351871088931, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -7130,7 +7130,7 @@ } }, { - "accuracy": 0.8727666829761705, + "accuracy": 0.869503234562121, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -7179,7 +7179,7 @@ } }, { - "accuracy": 0.9527197888022975, + "accuracy": 0.9563210732058475, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -7231,7 +7231,7 @@ } }, { - "accuracy": 0.9373646974563599, + "accuracy": 0.9363601584183543, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -7283,7 +7283,7 @@ } }, { - "accuracy": 0.9376337779195685, + "accuracy": 0.9365360799588656, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -7332,7 +7332,7 @@ } }, { - "accuracy": 0.9425607417759142, + "accuracy": 0.9423013423618517, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -7375,7 +7375,7 @@ } }, { - "accuracy": 0.9378391943479839, + "accuracy": 0.9380589723587036, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -7418,7 +7418,7 @@ } }, { - "accuracy": 0.9809844729147459, + "accuracy": 0.9924260652378986, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -7470,7 +7470,7 @@ } }, { - "accuracy": 0.9908502015628313, + "accuracy": 0.9908051255502199, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -7522,7 +7522,7 @@ } }, { - "accuracy": 0.993722823889632, + "accuracy": 0.9671524731736434, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -7574,7 +7574,7 @@ } }, { - "accuracy": 0.9936820417642593, + "accuracy": 0.9937566969739763, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -7626,7 +7626,7 @@ } }, { - "accuracy": 0.9936052427480095, + "accuracy": 0.9677239279997976, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -7669,7 +7669,7 @@ } }, { - "accuracy": 0.9954862288738552, + "accuracy": 0.9953670886002088, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -7718,7 +7718,7 @@ } }, { - "accuracy": 0.9955371888844591, + "accuracy": 0.9954090259577099, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -7764,7 +7764,7 @@ } }, { - "accuracy": 0.9956585462940367, + "accuracy": 0.995540810258765, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -7806,7 +7806,7 @@ ], "model.layers.4.self_attn": [ { - "accuracy": 0.9961339580385309, + "accuracy": 0.9961511382931157, "total_bits": 320757760, "q_proj": { "group_size": { @@ -7870,7 +7870,7 @@ } }, { - "accuracy": 0.9963707559202847, + "accuracy": 0.996427943831996, "total_bits": 329080832, "q_proj": { "group_size": { @@ -7934,7 +7934,7 @@ } }, { - "accuracy": 0.9967529305109852, + "accuracy": 0.9967787587328961, "total_bits": 336024576, "q_proj": { "group_size": { @@ -7998,7 +7998,7 @@ } }, { - "accuracy": 0.9970866112332595, + "accuracy": 0.9970837037422156, "total_bits": 401557504, "q_proj": { "group_size": { @@ -8062,7 +8062,7 @@ } }, { - "accuracy": 0.9971243078379255, + "accuracy": 0.9971108093465629, "total_bits": 475279360, "q_proj": { "group_size": { @@ -8126,7 +8126,7 @@ } }, { - "accuracy": 0.9970655992235008, + "accuracy": 0.9971136129216144, "total_bits": 475479040, "q_proj": { "group_size": { @@ -8190,7 +8190,7 @@ } }, { - "accuracy": 0.9978850891715602, + "accuracy": 0.9979828580429679, "total_bits": 609759232, "q_proj": { "group_size": { @@ -8242,7 +8242,7 @@ } }, { - "accuracy": 0.9980177995013563, + "accuracy": 0.99801814163986, "total_bits": 610024448, "q_proj": { "group_size": { @@ -8294,7 +8294,7 @@ } }, { - "accuracy": 0.9980482343388232, + "accuracy": 0.9981091500897157, "total_bits": 615020544, "q_proj": { "group_size": { @@ -8346,7 +8346,7 @@ } }, { - "accuracy": 0.998059017289626, + "accuracy": 0.9981198126548215, "total_bits": 623951872, "q_proj": { "group_size": { @@ -8398,7 +8398,7 @@ } }, { - "accuracy": 0.9980361679274785, + "accuracy": 0.9980649583433804, "total_bits": 626473984, "q_proj": { "group_size": { @@ -8462,7 +8462,7 @@ } }, { - "accuracy": 0.998129309204064, + "accuracy": 0.9981231265946439, "total_bits": 630355968, "q_proj": { "group_size": { @@ -8526,7 +8526,7 @@ } }, { - "accuracy": 0.9987995422592288, + "accuracy": 0.9987778991068664, "total_bits": 637362176, "q_proj": { "group_size": { @@ -8587,7 +8587,7 @@ } }, { - "accuracy": 0.9988725999076116, + "accuracy": 0.9988637655777367, "total_bits": 646823936, "q_proj": { "group_size": { @@ -8648,7 +8648,7 @@ } }, { - "accuracy": 0.9993180288982234, + "accuracy": 0.9993173989810442, "total_bits": 784740352, "q_proj": { "group_size": { @@ -8709,7 +8709,7 @@ } }, { - "accuracy": 0.9993997273084364, + "accuracy": 0.9994237036199162, "total_bits": 797818880, "q_proj": { "group_size": { @@ -8770,7 +8770,7 @@ } }, { - "accuracy": 0.99934596828136, + "accuracy": 0.9993429669718209, "total_bits": 911749120, "q_proj": { "group_size": { @@ -8822,7 +8822,7 @@ } }, { - "accuracy": 0.9998218625122192, + "accuracy": 0.9998208565980589, "total_bits": 942718976, "q_proj": { "group_size": { @@ -8874,7 +8874,7 @@ } }, { - "accuracy": 0.9998283276432439, + "accuracy": 0.9998294662095999, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -8928,7 +8928,7 @@ ], "model.layers.4.mlp": [ { - "accuracy": 0.9948095651833635, + "accuracy": 0.9948854501310148, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -8980,7 +8980,7 @@ } }, { - "accuracy": 0.9949485036103349, + "accuracy": 0.9950173289368027, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -9032,7 +9032,7 @@ } }, { - "accuracy": 0.9956819144518751, + "accuracy": 0.9957485144075594, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -9081,7 +9081,7 @@ } }, { - "accuracy": 0.9959021924357665, + "accuracy": 0.9959667325019836, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -9130,7 +9130,7 @@ } }, { - "accuracy": 0.9973085289330859, + "accuracy": 0.9973390567067423, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -9182,7 +9182,7 @@ } }, { - "accuracy": 0.9976692891826755, + "accuracy": 0.9977025352418423, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -9234,7 +9234,7 @@ } }, { - "accuracy": 0.9979805171881851, + "accuracy": 0.998012337637575, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -9283,7 +9283,7 @@ } }, { - "accuracy": 0.998644256670224, + "accuracy": 0.9986621163001186, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -9326,7 +9326,7 @@ } }, { - "accuracy": 0.9988181099883819, + "accuracy": 0.9988351751511034, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -9369,7 +9369,7 @@ } }, { - "accuracy": 0.998623604445081, + "accuracy": 0.9986398688664562, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -9421,7 +9421,7 @@ } }, { - "accuracy": 0.998882547511082, + "accuracy": 0.9988985642986862, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -9473,7 +9473,7 @@ } }, { - "accuracy": 0.9992939720028325, + "accuracy": 0.9993015460081791, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -9525,7 +9525,7 @@ } }, { - "accuracy": 0.9994444935827663, + "accuracy": 0.9994513750272361, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -9577,7 +9577,7 @@ } }, { - "accuracy": 0.9996206382240512, + "accuracy": 0.9996236154662543, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -9620,7 +9620,7 @@ } }, { - "accuracy": 0.9996367963521104, + "accuracy": 0.9996409214248783, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -9669,7 +9669,7 @@ } }, { - "accuracy": 0.9996901751544914, + "accuracy": 0.9996939301980954, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -9715,7 +9715,7 @@ } }, { - "accuracy": 0.9998960619865868, + "accuracy": 0.9998944507985327, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -9757,7 +9757,7 @@ ], "model.layers.5.self_attn": [ { - "accuracy": 0.9957247333307016, + "accuracy": 0.9957711873085875, "total_bits": 320757760, "q_proj": { "group_size": { @@ -9821,7 +9821,7 @@ } }, { - "accuracy": 0.9960670141797316, + "accuracy": 0.9960996947790447, "total_bits": 329080832, "q_proj": { "group_size": { @@ -9885,7 +9885,7 @@ } }, { - "accuracy": 0.9966135958307668, + "accuracy": 0.9966530623404604, "total_bits": 336024576, "q_proj": { "group_size": { @@ -9949,7 +9949,7 @@ } }, { - "accuracy": 0.9974373171204015, + "accuracy": 0.9974612832853669, "total_bits": 401557504, "q_proj": { "group_size": { @@ -10013,7 +10013,7 @@ } }, { - "accuracy": 0.9975065667378274, + "accuracy": 0.997531112871672, "total_bits": 475279360, "q_proj": { "group_size": { @@ -10077,7 +10077,7 @@ } }, { - "accuracy": 0.9975468273225584, + "accuracy": 0.9975746558292916, "total_bits": 475479040, "q_proj": { "group_size": { @@ -10141,7 +10141,7 @@ } }, { - "accuracy": 0.9981235473563796, + "accuracy": 0.9981124308940611, "total_bits": 609759232, "q_proj": { "group_size": { @@ -10193,7 +10193,7 @@ } }, { - "accuracy": 0.99817422149997, + "accuracy": 0.9981393427833131, "total_bits": 610024448, "q_proj": { "group_size": { @@ -10245,7 +10245,7 @@ } }, { - "accuracy": 0.9982087094532816, + "accuracy": 0.9982188316552263, "total_bits": 615020544, "q_proj": { "group_size": { @@ -10297,7 +10297,7 @@ } }, { - "accuracy": 0.998215160675739, + "accuracy": 0.9982261761630836, "total_bits": 623951872, "q_proj": { "group_size": { @@ -10349,7 +10349,7 @@ } }, { - "accuracy": 0.9982363395392895, + "accuracy": 0.9982679431375704, "total_bits": 626473984, "q_proj": { "group_size": { @@ -10413,7 +10413,7 @@ } }, { - "accuracy": 0.9982577779967534, + "accuracy": 0.9983036606327483, "total_bits": 630355968, "q_proj": { "group_size": { @@ -10477,7 +10477,7 @@ } }, { - "accuracy": 0.9983370937407017, + "accuracy": 0.9983702411777094, "total_bits": 637362176, "q_proj": { "group_size": { @@ -10538,7 +10538,7 @@ } }, { - "accuracy": 0.9983725559554602, + "accuracy": 0.9983650017529726, "total_bits": 646823936, "q_proj": { "group_size": { @@ -10599,7 +10599,7 @@ } }, { - "accuracy": 0.9988127548835779, + "accuracy": 0.9988060895549623, "total_bits": 784740352, "q_proj": { "group_size": { @@ -10660,7 +10660,7 @@ } }, { - "accuracy": 0.9988320713960811, + "accuracy": 0.9988482339601767, "total_bits": 797818880, "q_proj": { "group_size": { @@ -10721,7 +10721,7 @@ } }, { - "accuracy": 0.998857516305227, + "accuracy": 0.9988487874598879, "total_bits": 911749120, "q_proj": { "group_size": { @@ -10773,7 +10773,7 @@ } }, { - "accuracy": 0.9992073745417752, + "accuracy": 0.9992111180173723, "total_bits": 942718976, "q_proj": { "group_size": { @@ -10825,7 +10825,7 @@ } }, { - "accuracy": 0.9992348922318534, + "accuracy": 0.999242010624393, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -10879,7 +10879,7 @@ ], "model.layers.5.mlp": [ { - "accuracy": 0.9941252226892271, + "accuracy": 0.9942019401412261, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -10931,7 +10931,7 @@ } }, { - "accuracy": 0.9942738025596267, + "accuracy": 0.9943513972194571, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -10983,7 +10983,7 @@ } }, { - "accuracy": 0.9951262744633775, + "accuracy": 0.9951952852700886, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -11032,7 +11032,7 @@ } }, { - "accuracy": 0.9953896603301952, + "accuracy": 0.9954546078255302, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -11081,7 +11081,7 @@ } }, { - "accuracy": 0.9969224920006174, + "accuracy": 0.9969625635758826, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -11133,7 +11133,7 @@ } }, { - "accuracy": 0.9973469282451429, + "accuracy": 0.997387310980182, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -11185,7 +11185,7 @@ } }, { - "accuracy": 0.9977220663506734, + "accuracy": 0.9977569564392692, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -11234,7 +11234,7 @@ } }, { - "accuracy": 0.9982892086631373, + "accuracy": 0.9983108800492788, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -11277,7 +11277,7 @@ } }, { - "accuracy": 0.9985780538500924, + "accuracy": 0.9985986212758642, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -11320,7 +11320,7 @@ } }, { - "accuracy": 0.9983862589456534, + "accuracy": 0.9984243836646017, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -11372,7 +11372,7 @@ } }, { - "accuracy": 0.9987355418699352, + "accuracy": 0.9987534623789159, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -11424,7 +11424,7 @@ } }, { - "accuracy": 0.9990916133515144, + "accuracy": 0.9990982191735193, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -11476,7 +11476,7 @@ } }, { - "accuracy": 0.999336052636959, + "accuracy": 0.9993443131250771, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -11528,7 +11528,7 @@ } }, { - "accuracy": 0.9994621396947064, + "accuracy": 0.9994636802492958, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -11571,7 +11571,7 @@ } }, { - "accuracy": 0.9994835210964084, + "accuracy": 0.9994847207869354, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -11620,7 +11620,7 @@ } }, { - "accuracy": 0.9995349664241076, + "accuracy": 0.999534779326304, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -11666,7 +11666,7 @@ } }, { - "accuracy": 0.9998382587157386, + "accuracy": 0.999829817244685, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -11708,7 +11708,7 @@ ], "model.layers.6.self_attn": [ { - "accuracy": 0.9955444022228843, + "accuracy": 0.9954404352526915, "total_bits": 320757760, "q_proj": { "group_size": { @@ -11772,7 +11772,7 @@ } }, { - "accuracy": 0.9960725730971286, + "accuracy": 0.9960030274171578, "total_bits": 329080832, "q_proj": { "group_size": { @@ -11836,7 +11836,7 @@ } }, { - "accuracy": 0.9968657809260645, + "accuracy": 0.9968925759588417, "total_bits": 336024576, "q_proj": { "group_size": { @@ -11900,7 +11900,7 @@ } }, { - "accuracy": 0.9977938119125994, + "accuracy": 0.9978137502544805, "total_bits": 401557504, "q_proj": { "group_size": { @@ -11964,7 +11964,7 @@ } }, { - "accuracy": 0.997970481452189, + "accuracy": 0.9979901470636067, "total_bits": 475279360, "q_proj": { "group_size": { @@ -12028,7 +12028,7 @@ } }, { - "accuracy": 0.9980321314774061, + "accuracy": 0.9980235658586025, "total_bits": 475479040, "q_proj": { "group_size": { @@ -12092,7 +12092,7 @@ } }, { - "accuracy": 0.9987041152229434, + "accuracy": 0.9987196339980552, "total_bits": 609759232, "q_proj": { "group_size": { @@ -12144,7 +12144,7 @@ } }, { - "accuracy": 0.9987658369109819, + "accuracy": 0.9987481599183459, "total_bits": 610024448, "q_proj": { "group_size": { @@ -12196,7 +12196,7 @@ } }, { - "accuracy": 0.9988313246714441, + "accuracy": 0.9988201237038562, "total_bits": 615020544, "q_proj": { "group_size": { @@ -12248,7 +12248,7 @@ } }, { - "accuracy": 0.9988639126286695, + "accuracy": 0.9988517156360965, "total_bits": 623951872, "q_proj": { "group_size": { @@ -12300,7 +12300,7 @@ } }, { - "accuracy": 0.9988786210551074, + "accuracy": 0.9988896966372666, "total_bits": 626473984, "q_proj": { "group_size": { @@ -12364,7 +12364,7 @@ } }, { - "accuracy": 0.9989327600127772, + "accuracy": 0.9989302174041146, "total_bits": 630355968, "q_proj": { "group_size": { @@ -12428,7 +12428,7 @@ } }, { - "accuracy": 0.9991661738604307, + "accuracy": 0.9991721704994377, "total_bits": 637362176, "q_proj": { "group_size": { @@ -12489,7 +12489,7 @@ } }, { - "accuracy": 0.9992342186895641, + "accuracy": 0.9992313894296163, "total_bits": 646823936, "q_proj": { "group_size": { @@ -12550,7 +12550,7 @@ } }, { - "accuracy": 0.9995447077641362, + "accuracy": 0.9995492891849656, "total_bits": 784740352, "q_proj": { "group_size": { @@ -12611,7 +12611,7 @@ } }, { - "accuracy": 0.9996090260775465, + "accuracy": 0.9996104203164577, "total_bits": 797818880, "q_proj": { "group_size": { @@ -12672,7 +12672,7 @@ } }, { - "accuracy": 0.9996411338154423, + "accuracy": 0.9996454404716036, "total_bits": 911749120, "q_proj": { "group_size": { @@ -12724,7 +12724,7 @@ } }, { - "accuracy": 0.9998188434830425, + "accuracy": 0.9998191342272452, "total_bits": 942718976, "q_proj": { "group_size": { @@ -12776,7 +12776,7 @@ } }, { - "accuracy": 0.9999019349373779, + "accuracy": 0.999900343209064, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -12830,7 +12830,7 @@ ], "model.layers.6.mlp": [ { - "accuracy": 0.9934912445513826, + "accuracy": 0.9935955342493559, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -12882,7 +12882,7 @@ } }, { - "accuracy": 0.9936532103701642, + "accuracy": 0.9937494436376973, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -12934,7 +12934,7 @@ } }, { - "accuracy": 0.9946330483806761, + "accuracy": 0.9947180583288795, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -12983,7 +12983,7 @@ } }, { - "accuracy": 0.9949320864520574, + "accuracy": 0.9950110379018282, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -13032,7 +13032,7 @@ } }, { - "accuracy": 0.9967397616097802, + "accuracy": 0.9968282494105791, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -13084,7 +13084,7 @@ } }, { - "accuracy": 0.9970319337750736, + "accuracy": 0.9970703693596941, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -13136,7 +13136,7 @@ } }, { - "accuracy": 0.9974556377059535, + "accuracy": 0.9974842814630583, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -13185,7 +13185,7 @@ } }, { - "accuracy": 0.9982815063313434, + "accuracy": 0.998300171604282, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -13228,7 +13228,7 @@ } }, { - "accuracy": 0.9984211628570369, + "accuracy": 0.9984373312051359, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -13271,7 +13271,7 @@ } }, { - "accuracy": 0.9983872815378403, + "accuracy": 0.9984212483426458, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -13323,7 +13323,7 @@ } }, { - "accuracy": 0.9985944095410799, + "accuracy": 0.9986181534630688, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -13375,7 +13375,7 @@ } }, { - "accuracy": 0.999181319224207, + "accuracy": 0.9991968783893084, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -13427,7 +13427,7 @@ } }, { - "accuracy": 0.9993016335034841, + "accuracy": 0.9993139457839885, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -13479,7 +13479,7 @@ } }, { - "accuracy": 0.999553513860232, + "accuracy": 0.9995665040455366, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -13522,7 +13522,7 @@ } }, { - "accuracy": 0.9995828847351828, + "accuracy": 0.9995851708870185, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -13571,7 +13571,7 @@ } }, { - "accuracy": 0.9996661209795428, + "accuracy": 0.9996656496813031, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -13617,7 +13617,7 @@ } }, { - "accuracy": 0.9998750927563953, + "accuracy": 0.9998637679841762, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -13659,7 +13659,7 @@ ], "model.layers.7.self_attn": [ { - "accuracy": 0.9957331869947282, + "accuracy": 0.9956815058463498, "total_bits": 320757760, "q_proj": { "group_size": { @@ -13723,7 +13723,7 @@ } }, { - "accuracy": 0.9962008807219958, + "accuracy": 0.9962362298055699, "total_bits": 329080832, "q_proj": { "group_size": { @@ -13787,7 +13787,7 @@ } }, { - "accuracy": 0.9972529174073747, + "accuracy": 0.9972511186803642, "total_bits": 336024576, "q_proj": { "group_size": { @@ -13851,7 +13851,7 @@ } }, { - "accuracy": 0.997928732319882, + "accuracy": 0.9979187016816515, "total_bits": 401557504, "q_proj": { "group_size": { @@ -13915,7 +13915,7 @@ } }, { - "accuracy": 0.9981140782566447, + "accuracy": 0.9981022875167822, "total_bits": 475279360, "q_proj": { "group_size": { @@ -13979,7 +13979,7 @@ } }, { - "accuracy": 0.9981643873217859, + "accuracy": 0.9981404903687929, "total_bits": 475479040, "q_proj": { "group_size": { @@ -14043,7 +14043,7 @@ } }, { - "accuracy": 0.9987523550853917, + "accuracy": 0.9987493056411806, "total_bits": 609759232, "q_proj": { "group_size": { @@ -14095,7 +14095,7 @@ } }, { - "accuracy": 0.9988136368950731, + "accuracy": 0.9987982087033359, "total_bits": 610024448, "q_proj": { "group_size": { @@ -14147,7 +14147,7 @@ } }, { - "accuracy": 0.9988597651061258, + "accuracy": 0.99888949223647, "total_bits": 615020544, "q_proj": { "group_size": { @@ -14199,7 +14199,7 @@ } }, { - "accuracy": 0.9988829455289402, + "accuracy": 0.9989131140081506, "total_bits": 623951872, "q_proj": { "group_size": { @@ -14251,7 +14251,7 @@ } }, { - "accuracy": 0.9989384600990697, + "accuracy": 0.9989387902774309, "total_bits": 626473984, "q_proj": { "group_size": { @@ -14315,7 +14315,7 @@ } }, { - "accuracy": 0.9989787271540416, + "accuracy": 0.99898599931284, "total_bits": 630355968, "q_proj": { "group_size": { @@ -14379,7 +14379,7 @@ } }, { - "accuracy": 0.9992219521909168, + "accuracy": 0.999222860426495, "total_bits": 637362176, "q_proj": { "group_size": { @@ -14440,7 +14440,7 @@ } }, { - "accuracy": 0.9992944628588463, + "accuracy": 0.9992884179381164, "total_bits": 646823936, "q_proj": { "group_size": { @@ -14501,7 +14501,7 @@ } }, { - "accuracy": 0.999572721113892, + "accuracy": 0.9995745191056478, "total_bits": 784740352, "q_proj": { "group_size": { @@ -14562,7 +14562,7 @@ } }, { - "accuracy": 0.9996411371485967, + "accuracy": 0.999639321608763, "total_bits": 797818880, "q_proj": { "group_size": { @@ -14623,7 +14623,7 @@ } }, { - "accuracy": 0.9996374770508785, + "accuracy": 0.9996396662471326, "total_bits": 911749120, "q_proj": { "group_size": { @@ -14675,7 +14675,7 @@ } }, { - "accuracy": 0.9998393113553328, + "accuracy": 0.9998390283435583, "total_bits": 942718976, "q_proj": { "group_size": { @@ -14727,7 +14727,7 @@ } }, { - "accuracy": 0.9999029561570895, + "accuracy": 0.9999008709380991, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -14781,7 +14781,7 @@ ], "model.layers.7.mlp": [ { - "accuracy": 0.9925373455411509, + "accuracy": 0.9926616251468658, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -14833,7 +14833,7 @@ } }, { - "accuracy": 0.9927481169763365, + "accuracy": 0.9928668039409738, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -14885,7 +14885,7 @@ } }, { - "accuracy": 0.9938074169974578, + "accuracy": 0.9939058775964537, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -14934,7 +14934,7 @@ } }, { - "accuracy": 0.9941292495319718, + "accuracy": 0.9942211681290677, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -14983,7 +14983,7 @@ } }, { - "accuracy": 0.9963686113294802, + "accuracy": 0.9964321534100332, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -15035,7 +15035,7 @@ } }, { - "accuracy": 0.9966892658879882, + "accuracy": 0.9967453903273532, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -15087,7 +15087,7 @@ } }, { - "accuracy": 0.9971585659996459, + "accuracy": 0.9972051474216738, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -15136,7 +15136,7 @@ } }, { - "accuracy": 0.9981502500411711, + "accuracy": 0.9981828927993774, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -15179,7 +15179,7 @@ } }, { - "accuracy": 0.9983164942578265, + "accuracy": 0.9983466516592001, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -15222,7 +15222,7 @@ } }, { - "accuracy": 0.998165578042206, + "accuracy": 0.9982010079057593, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -15274,7 +15274,7 @@ } }, { - "accuracy": 0.9984035816435751, + "accuracy": 0.9984306296039569, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -15326,7 +15326,7 @@ } }, { - "accuracy": 0.9990642171745238, + "accuracy": 0.9990826671648967, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -15378,7 +15378,7 @@ } }, { - "accuracy": 0.9992012553113071, + "accuracy": 0.9992101700290253, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -15430,7 +15430,7 @@ } }, { - "accuracy": 0.9994877413111297, + "accuracy": 0.9994991772170914, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -15473,7 +15473,7 @@ } }, { - "accuracy": 0.9995251892036513, + "accuracy": 0.9995327253188742, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -15522,7 +15522,7 @@ } }, { - "accuracy": 0.999616207432394, + "accuracy": 0.9996215938835552, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -15568,7 +15568,7 @@ } }, { - "accuracy": 0.9998622072223378, + "accuracy": 0.9998572975103008, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -15610,7 +15610,7 @@ ], "model.layers.8.self_attn": [ { - "accuracy": 0.9936644540805566, + "accuracy": 0.9935695140769607, "total_bits": 320757760, "q_proj": { "group_size": { @@ -15674,7 +15674,7 @@ } }, { - "accuracy": 0.9940900916331693, + "accuracy": 0.9942599703606806, "total_bits": 329080832, "q_proj": { "group_size": { @@ -15738,7 +15738,7 @@ } }, { - "accuracy": 0.995490870977703, + "accuracy": 0.9955004555614371, "total_bits": 336024576, "q_proj": { "group_size": { @@ -15802,7 +15802,7 @@ } }, { - "accuracy": 0.9963940248677605, + "accuracy": 0.9963807902838054, "total_bits": 401557504, "q_proj": { "group_size": { @@ -15866,7 +15866,7 @@ } }, { - "accuracy": 0.9965187895454859, + "accuracy": 0.9965053443845949, "total_bits": 475279360, "q_proj": { "group_size": { @@ -15930,7 +15930,7 @@ } }, { - "accuracy": 0.9965737411066106, + "accuracy": 0.9965960665753013, "total_bits": 475479040, "q_proj": { "group_size": { @@ -15994,7 +15994,7 @@ } }, { - "accuracy": 0.9974765569755906, + "accuracy": 0.9975225745063079, "total_bits": 609759232, "q_proj": { "group_size": { @@ -16046,7 +16046,7 @@ } }, { - "accuracy": 0.9975893434725309, + "accuracy": 0.9975952029620346, "total_bits": 610024448, "q_proj": { "group_size": { @@ -16098,7 +16098,7 @@ } }, { - "accuracy": 0.9977382078374687, + "accuracy": 0.9976478122959012, "total_bits": 615020544, "q_proj": { "group_size": { @@ -16150,7 +16150,7 @@ } }, { - "accuracy": 0.9977591273031736, + "accuracy": 0.9976685319684053, "total_bits": 623951872, "q_proj": { "group_size": { @@ -16202,7 +16202,7 @@ } }, { - "accuracy": 0.9976815191146574, + "accuracy": 0.99768264768155, "total_bits": 626473984, "q_proj": { "group_size": { @@ -16266,7 +16266,7 @@ } }, { - "accuracy": 0.9977850404224897, + "accuracy": 0.9976957320774856, "total_bits": 630355968, "q_proj": { "group_size": { @@ -16330,7 +16330,7 @@ } }, { - "accuracy": 0.9984539984480331, + "accuracy": 0.9984527883168898, "total_bits": 637362176, "q_proj": { "group_size": { @@ -16391,7 +16391,7 @@ } }, { - "accuracy": 0.9985351376235485, + "accuracy": 0.9985404517501593, "total_bits": 646823936, "q_proj": { "group_size": { @@ -16452,7 +16452,7 @@ } }, { - "accuracy": 0.9991571674809644, + "accuracy": 0.9991547740799817, "total_bits": 784740352, "q_proj": { "group_size": { @@ -16513,7 +16513,7 @@ } }, { - "accuracy": 0.9992716807479921, + "accuracy": 0.9992800235846325, "total_bits": 797818880, "q_proj": { "group_size": { @@ -16574,7 +16574,7 @@ } }, { - "accuracy": 0.9992291019543221, + "accuracy": 0.999227895891588, "total_bits": 911749120, "q_proj": { "group_size": { @@ -16626,7 +16626,7 @@ } }, { - "accuracy": 0.9997504289556098, + "accuracy": 0.9997471468523145, "total_bits": 942718976, "q_proj": { "group_size": { @@ -16678,7 +16678,7 @@ } }, { - "accuracy": 0.9997994363013851, + "accuracy": 0.999794185382167, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -16732,7 +16732,7 @@ ], "model.layers.8.mlp": [ { - "accuracy": 0.9919421963001552, + "accuracy": 0.9920782415490401, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -16784,7 +16784,7 @@ } }, { - "accuracy": 0.9921483962159408, + "accuracy": 0.9922846375327361, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -16836,7 +16836,7 @@ } }, { - "accuracy": 0.9933573431090305, + "accuracy": 0.9934626684377068, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -16885,7 +16885,7 @@ } }, { - "accuracy": 0.9937279275373408, + "accuracy": 0.993823245559868, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -16934,7 +16934,7 @@ } }, { - "accuracy": 0.995588153208557, + "accuracy": 0.9956541480986696, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -16986,7 +16986,7 @@ } }, { - "accuracy": 0.996422888416993, + "accuracy": 0.9964864736324862, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -17038,7 +17038,7 @@ } }, { - "accuracy": 0.9969592541456223, + "accuracy": 0.9970111364596769, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -17087,7 +17087,7 @@ } }, { - "accuracy": 0.9978526589509688, + "accuracy": 0.9978899442051586, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -17130,7 +17130,7 @@ } }, { - "accuracy": 0.9981903545558453, + "accuracy": 0.9982229841775015, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -17173,7 +17173,7 @@ } }, { - "accuracy": 0.9976840915256425, + "accuracy": 0.9977197249076868, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -17225,7 +17225,7 @@ } }, { - "accuracy": 0.9982827472451486, + "accuracy": 0.998313034051343, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -17277,7 +17277,7 @@ } }, { - "accuracy": 0.9987795206865198, + "accuracy": 0.998796857795433, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -17329,7 +17329,7 @@ } }, { - "accuracy": 0.9991447780477373, + "accuracy": 0.9991598001828319, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -17381,7 +17381,7 @@ } }, { - "accuracy": 0.9993357634858081, + "accuracy": 0.9993453850773605, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -17424,7 +17424,7 @@ } }, { - "accuracy": 0.9993567658882392, + "accuracy": 0.9993655886501074, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -17473,7 +17473,7 @@ } }, { - "accuracy": 0.9994370076119116, + "accuracy": 0.9994436081890997, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -17519,7 +17519,7 @@ } }, { - "accuracy": 0.999820676301361, + "accuracy": 0.9998231241356974, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -17561,7 +17561,7 @@ ], "model.layers.9.self_attn": [ { - "accuracy": 0.9924858573235964, + "accuracy": 0.9926480313664988, "total_bits": 320757760, "q_proj": { "group_size": { @@ -17625,7 +17625,7 @@ } }, { - "accuracy": 0.9931424369937495, + "accuracy": 0.9931283271626422, "total_bits": 329080832, "q_proj": { "group_size": { @@ -17689,7 +17689,7 @@ } }, { - "accuracy": 0.9940861636086514, + "accuracy": 0.9941145871814928, "total_bits": 336024576, "q_proj": { "group_size": { @@ -17753,7 +17753,7 @@ } }, { - "accuracy": 0.9951514801696727, + "accuracy": 0.995171661047559, "total_bits": 401557504, "q_proj": { "group_size": { @@ -17817,7 +17817,7 @@ } }, { - "accuracy": 0.995255426356667, + "accuracy": 0.9952712902113011, "total_bits": 475279360, "q_proj": { "group_size": { @@ -17881,7 +17881,7 @@ } }, { - "accuracy": 0.9952889947514785, + "accuracy": 0.995361868880297, "total_bits": 475479040, "q_proj": { "group_size": { @@ -17945,7 +17945,7 @@ } }, { - "accuracy": 0.9963576852491027, + "accuracy": 0.9964091860150036, "total_bits": 609759232, "q_proj": { "group_size": { @@ -17997,7 +17997,7 @@ } }, { - "accuracy": 0.9965018618263697, + "accuracy": 0.9965067890129591, "total_bits": 610024448, "q_proj": { "group_size": { @@ -18049,7 +18049,7 @@ } }, { - "accuracy": 0.9965712055563927, + "accuracy": 0.9966210824878592, "total_bits": 615020544, "q_proj": { "group_size": { @@ -18101,7 +18101,7 @@ } }, { - "accuracy": 0.9965905617726477, + "accuracy": 0.9966483429858559, "total_bits": 623951872, "q_proj": { "group_size": { @@ -18153,7 +18153,7 @@ } }, { - "accuracy": 0.9966421260645515, + "accuracy": 0.9965847169882373, "total_bits": 626473984, "q_proj": { "group_size": { @@ -18217,7 +18217,7 @@ } }, { - "accuracy": 0.9966545367711469, + "accuracy": 0.9967073955033955, "total_bits": 630355968, "q_proj": { "group_size": { @@ -18281,7 +18281,7 @@ } }, { - "accuracy": 0.9976477707295042, + "accuracy": 0.9976071640849113, "total_bits": 637362176, "q_proj": { "group_size": { @@ -18342,7 +18342,7 @@ } }, { - "accuracy": 0.9978620751123679, + "accuracy": 0.9978378248450003, "total_bits": 646823936, "q_proj": { "group_size": { @@ -18403,7 +18403,7 @@ } }, { - "accuracy": 0.9986804054167709, + "accuracy": 0.9986977447804651, "total_bits": 784740352, "q_proj": { "group_size": { @@ -18464,7 +18464,7 @@ } }, { - "accuracy": 0.9988922779693415, + "accuracy": 0.9989216782544789, "total_bits": 797818880, "q_proj": { "group_size": { @@ -18525,7 +18525,7 @@ } }, { - "accuracy": 0.9987556446147593, + "accuracy": 0.9987749096594358, "total_bits": 911749120, "q_proj": { "group_size": { @@ -18577,7 +18577,7 @@ } }, { - "accuracy": 0.9996471527081571, + "accuracy": 0.999649605125581, "total_bits": 942718976, "q_proj": { "group_size": { @@ -18629,7 +18629,7 @@ } }, { - "accuracy": 0.9996841979752246, + "accuracy": 0.999691359527213, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -18683,7 +18683,7 @@ ], "model.layers.9.mlp": [ { - "accuracy": 0.9907768721643248, + "accuracy": 0.9909018027155023, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -18735,7 +18735,7 @@ } }, { - "accuracy": 0.9909801608637759, + "accuracy": 0.9911025731187117, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -18787,7 +18787,7 @@ } }, { - "accuracy": 0.9923760349813261, + "accuracy": 0.9924725694091696, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -18836,7 +18836,7 @@ } }, { - "accuracy": 0.9928061750374342, + "accuracy": 0.9928921421891764, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -18885,7 +18885,7 @@ } }, { - "accuracy": 0.994809023251659, + "accuracy": 0.9948815926909447, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -18937,7 +18937,7 @@ } }, { - "accuracy": 0.9959385587196601, + "accuracy": 0.9960025145035041, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -18989,7 +18989,7 @@ } }, { - "accuracy": 0.9965791165044433, + "accuracy": 0.9966303498337143, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -19038,7 +19038,7 @@ } }, { - "accuracy": 0.9972833055806788, + "accuracy": 0.9973220582071104, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -19081,7 +19081,7 @@ } }, { - "accuracy": 0.9978996123138227, + "accuracy": 0.9979326613247395, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -19124,7 +19124,7 @@ } }, { - "accuracy": 0.9967806074572237, + "accuracy": 0.9968607864881817, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -19176,7 +19176,7 @@ } }, { - "accuracy": 0.9979893524003657, + "accuracy": 0.9980421115301157, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -19228,7 +19228,7 @@ } }, { - "accuracy": 0.9978754551787126, + "accuracy": 0.9978961456370982, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -19280,7 +19280,7 @@ } }, { - "accuracy": 0.9989507515963755, + "accuracy": 0.9989694368682409, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -19332,7 +19332,7 @@ } }, { - "accuracy": 0.9985200171603968, + "accuracy": 0.9985353087908343, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -19375,7 +19375,7 @@ } }, { - "accuracy": 0.9985674269693462, + "accuracy": 0.9985797473866689, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -19424,7 +19424,7 @@ } }, { - "accuracy": 0.9986151128419136, + "accuracy": 0.9986259859839552, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -19470,7 +19470,7 @@ } }, { - "accuracy": 0.9993902202205438, + "accuracy": 0.9993963886150404, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -19512,7 +19512,7 @@ ], "model.layers.10.self_attn": [ { - "accuracy": 0.9909308944877825, + "accuracy": 0.991252338415698, "total_bits": 320757760, "q_proj": { "group_size": { @@ -19576,7 +19576,7 @@ } }, { - "accuracy": 0.9913782485221562, + "accuracy": 0.9914088460959887, "total_bits": 329080832, "q_proj": { "group_size": { @@ -19640,7 +19640,7 @@ } }, { - "accuracy": 0.9926927372028953, + "accuracy": 0.992724492361671, "total_bits": 336024576, "q_proj": { "group_size": { @@ -19704,7 +19704,7 @@ } }, { - "accuracy": 0.9935553920896429, + "accuracy": 0.9935447820707372, "total_bits": 401557504, "q_proj": { "group_size": { @@ -19768,7 +19768,7 @@ } }, { - "accuracy": 0.9937822457991148, + "accuracy": 0.993755642520754, "total_bits": 475279360, "q_proj": { "group_size": { @@ -19832,7 +19832,7 @@ } }, { - "accuracy": 0.9940199879439253, + "accuracy": 0.9940377277763266, "total_bits": 475479040, "q_proj": { "group_size": { @@ -19896,7 +19896,7 @@ } }, { - "accuracy": 0.9952029836805243, + "accuracy": 0.9952342882752419, "total_bits": 609759232, "q_proj": { "group_size": { @@ -19948,7 +19948,7 @@ } }, { - "accuracy": 0.9958504563883731, + "accuracy": 0.9958569917239641, "total_bits": 610024448, "q_proj": { "group_size": { @@ -20000,7 +20000,7 @@ } }, { - "accuracy": 0.9960235467082575, + "accuracy": 0.9959631166175792, "total_bits": 615020544, "q_proj": { "group_size": { @@ -20052,7 +20052,7 @@ } }, { - "accuracy": 0.9960461168697006, + "accuracy": 0.9959931546135953, "total_bits": 623951872, "q_proj": { "group_size": { @@ -20104,7 +20104,7 @@ } }, { - "accuracy": 0.9958318722875494, + "accuracy": 0.9959381222724915, "total_bits": 626473984, "q_proj": { "group_size": { @@ -20168,7 +20168,7 @@ } }, { - "accuracy": 0.9961019809308805, + "accuracy": 0.9961302900000623, "total_bits": 630355968, "q_proj": { "group_size": { @@ -20232,7 +20232,7 @@ } }, { - "accuracy": 0.9975296958888832, + "accuracy": 0.9974831362303934, "total_bits": 637362176, "q_proj": { "group_size": { @@ -20293,7 +20293,7 @@ } }, { - "accuracy": 0.9977128991955205, + "accuracy": 0.9976785277065477, "total_bits": 646823936, "q_proj": { "group_size": { @@ -20354,7 +20354,7 @@ } }, { - "accuracy": 0.9984269189207178, + "accuracy": 0.9984153660112306, "total_bits": 784740352, "q_proj": { "group_size": { @@ -20415,7 +20415,7 @@ } }, { - "accuracy": 0.9988310517449128, + "accuracy": 0.9988531213449804, "total_bits": 797818880, "q_proj": { "group_size": { @@ -20476,7 +20476,7 @@ } }, { - "accuracy": 0.998489229988895, + "accuracy": 0.9984748420354567, "total_bits": 911749120, "q_proj": { "group_size": { @@ -20528,7 +20528,7 @@ } }, { - "accuracy": 0.9996198634371946, + "accuracy": 0.9996221170417572, "total_bits": 942718976, "q_proj": { "group_size": { @@ -20580,7 +20580,7 @@ } }, { - "accuracy": 0.999601783647545, + "accuracy": 0.9996140349999463, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -20634,7 +20634,7 @@ ], "model.layers.10.mlp": [ { - "accuracy": 0.9893906869386372, + "accuracy": 0.9895371308452204, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -20686,7 +20686,7 @@ } }, { - "accuracy": 0.9896614708398518, + "accuracy": 0.9898036992863605, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -20738,7 +20738,7 @@ } }, { - "accuracy": 0.9913950868343052, + "accuracy": 0.9915021038369128, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -20787,7 +20787,7 @@ } }, { - "accuracy": 0.991916127895054, + "accuracy": 0.9920113533735275, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -20836,7 +20836,7 @@ } }, { - "accuracy": 0.9947415402061061, + "accuracy": 0.9948152413493708, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -20888,7 +20888,7 @@ } }, { - "accuracy": 0.9951759153290799, + "accuracy": 0.9952419533541328, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -20940,7 +20940,7 @@ } }, { - "accuracy": 0.9959161277664336, + "accuracy": 0.9959671752233255, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -20989,7 +20989,7 @@ } }, { - "accuracy": 0.9973072256696852, + "accuracy": 0.9973463116115645, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -21032,7 +21032,7 @@ } }, { - "accuracy": 0.997559554874897, + "accuracy": 0.9975938324473406, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -21075,7 +21075,7 @@ } }, { - "accuracy": 0.9973404631018639, + "accuracy": 0.9973773546516895, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -21127,7 +21127,7 @@ } }, { - "accuracy": 0.9976819071330523, + "accuracy": 0.9977138277731443, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -21179,7 +21179,7 @@ } }, { - "accuracy": 0.9986494042371449, + "accuracy": 0.9986682415596748, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -21231,7 +21231,7 @@ } }, { - "accuracy": 0.9988473986128443, + "accuracy": 0.9988632749177908, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -21283,7 +21283,7 @@ } }, { - "accuracy": 0.9992784084262032, + "accuracy": 0.9992884627396339, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -21326,7 +21326,7 @@ } }, { - "accuracy": 0.9993126337955657, + "accuracy": 0.9993219555502659, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -21375,7 +21375,7 @@ } }, { - "accuracy": 0.9994596598767921, + "accuracy": 0.9994656373011438, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -21421,7 +21421,7 @@ } }, { - "accuracy": 0.9998078946465332, + "accuracy": 0.9998105902616915, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -21463,7 +21463,7 @@ ], "model.layers.11.self_attn": [ { - "accuracy": 0.9910436280463871, + "accuracy": 0.9909514620115882, "total_bits": 320757760, "q_proj": { "group_size": { @@ -21527,7 +21527,7 @@ } }, { - "accuracy": 0.9915673709229419, + "accuracy": 0.9915214910318977, "total_bits": 329080832, "q_proj": { "group_size": { @@ -21591,7 +21591,7 @@ } }, { - "accuracy": 0.9931760500920447, + "accuracy": 0.9931138698991976, "total_bits": 336024576, "q_proj": { "group_size": { @@ -21655,7 +21655,7 @@ } }, { - "accuracy": 0.9942498948228987, + "accuracy": 0.9941876475748262, "total_bits": 401557504, "q_proj": { "group_size": { @@ -21719,7 +21719,7 @@ } }, { - "accuracy": 0.9945554513680307, + "accuracy": 0.9944776618166974, "total_bits": 475279360, "q_proj": { "group_size": { @@ -21783,7 +21783,7 @@ } }, { - "accuracy": 0.9945933089444512, + "accuracy": 0.9946575603987041, "total_bits": 475479040, "q_proj": { "group_size": { @@ -21847,7 +21847,7 @@ } }, { - "accuracy": 0.9959382206985825, + "accuracy": 0.9959606704743285, "total_bits": 609759232, "q_proj": { "group_size": { @@ -21899,7 +21899,7 @@ } }, { - "accuracy": 0.9961203517098176, + "accuracy": 0.9961069939952147, "total_bits": 610024448, "q_proj": { "group_size": { @@ -21951,7 +21951,7 @@ } }, { - "accuracy": 0.9962216137271178, + "accuracy": 0.9962361584368505, "total_bits": 615020544, "q_proj": { "group_size": { @@ -22003,7 +22003,7 @@ } }, { - "accuracy": 0.9962381544081789, + "accuracy": 0.9962633091368174, "total_bits": 623951872, "q_proj": { "group_size": { @@ -22055,7 +22055,7 @@ } }, { - "accuracy": 0.9960853355495554, + "accuracy": 0.996189326832169, "total_bits": 626473984, "q_proj": { "group_size": { @@ -22119,7 +22119,7 @@ } }, { - "accuracy": 0.9962003842780465, + "accuracy": 0.9962654368657815, "total_bits": 630355968, "q_proj": { "group_size": { @@ -22183,7 +22183,7 @@ } }, { - "accuracy": 0.9971737453812047, + "accuracy": 0.9971460484360394, "total_bits": 637362176, "q_proj": { "group_size": { @@ -22244,7 +22244,7 @@ } }, { - "accuracy": 0.9972001859231999, + "accuracy": 0.9972606095436373, "total_bits": 646823936, "q_proj": { "group_size": { @@ -22305,7 +22305,7 @@ } }, { - "accuracy": 0.9980616557754969, + "accuracy": 0.9980710964453848, "total_bits": 784740352, "q_proj": { "group_size": { @@ -22366,7 +22366,7 @@ } }, { - "accuracy": 0.9981159571754304, + "accuracy": 0.9981375736625571, "total_bits": 797818880, "q_proj": { "group_size": { @@ -22427,7 +22427,7 @@ } }, { - "accuracy": 0.9981207388796305, + "accuracy": 0.9981342087450781, "total_bits": 911749120, "q_proj": { "group_size": { @@ -22479,7 +22479,7 @@ } }, { - "accuracy": 0.9992105840754352, + "accuracy": 0.9992129936520207, "total_bits": 942718976, "q_proj": { "group_size": { @@ -22531,7 +22531,7 @@ } }, { - "accuracy": 0.9992402338569886, + "accuracy": 0.9992493567007937, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -22585,7 +22585,7 @@ ], "model.layers.11.mlp": [ { - "accuracy": 0.9875197512538809, + "accuracy": 0.9876507520675659, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -22637,7 +22637,7 @@ } }, { - "accuracy": 0.9878317070634741, + "accuracy": 0.9879594272688815, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -22689,7 +22689,7 @@ } }, { - "accuracy": 0.9896890873971739, + "accuracy": 0.9897886883271368, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -22738,7 +22738,7 @@ } }, { - "accuracy": 0.99024708647477, + "accuracy": 0.9903377341596704, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -22787,7 +22787,7 @@ } }, { - "accuracy": 0.9935758156996024, + "accuracy": 0.993638802515833, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -22839,7 +22839,7 @@ } }, { - "accuracy": 0.9943937337712238, + "accuracy": 0.9944544999223006, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -22891,7 +22891,7 @@ } }, { - "accuracy": 0.9952096543029735, + "accuracy": 0.9952580046496893, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -22940,7 +22940,7 @@ } }, { - "accuracy": 0.9968000617074338, + "accuracy": 0.9968334952075231, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -22983,7 +22983,7 @@ } }, { - "accuracy": 0.9971663730317041, + "accuracy": 0.9971976850770021, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -23026,7 +23026,7 @@ } }, { - "accuracy": 0.9967312208916012, + "accuracy": 0.9967603314864008, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -23078,7 +23078,7 @@ } }, { - "accuracy": 0.9973079015157724, + "accuracy": 0.9973367876128146, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -23130,7 +23130,7 @@ } }, { - "accuracy": 0.9983203285618832, + "accuracy": 0.9983356389168062, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -23182,7 +23182,7 @@ } }, { - "accuracy": 0.9986619147423067, + "accuracy": 0.9986765970917124, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -23234,7 +23234,7 @@ } }, { - "accuracy": 0.999103368995221, + "accuracy": 0.9991120307853347, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -23277,7 +23277,7 @@ } }, { - "accuracy": 0.999139403924346, + "accuracy": 0.9991469088157541, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -23326,7 +23326,7 @@ } }, { - "accuracy": 0.9992812698412883, + "accuracy": 0.9992860780636731, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -23372,7 +23372,7 @@ } }, { - "accuracy": 0.999760933244895, + "accuracy": 0.9997630564397887, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -23414,7 +23414,7 @@ ], "model.layers.12.self_attn": [ { - "accuracy": 0.9871261104157096, + "accuracy": 0.9870101503635708, "total_bits": 320757760, "q_proj": { "group_size": { @@ -23478,7 +23478,7 @@ } }, { - "accuracy": 0.9874413107570849, + "accuracy": 0.9876369284956079, "total_bits": 329080832, "q_proj": { "group_size": { @@ -23542,7 +23542,7 @@ } }, { - "accuracy": 0.990122032008673, + "accuracy": 0.9901722497061679, "total_bits": 336024576, "q_proj": { "group_size": { @@ -23606,7 +23606,7 @@ } }, { - "accuracy": 0.9910036599949786, + "accuracy": 0.9910771501691717, "total_bits": 401557504, "q_proj": { "group_size": { @@ -23670,7 +23670,7 @@ } }, { - "accuracy": 0.991171447854293, + "accuracy": 0.991241560170525, "total_bits": 475279360, "q_proj": { "group_size": { @@ -23734,7 +23734,7 @@ } }, { - "accuracy": 0.9913954193654814, + "accuracy": 0.991303781929769, "total_bits": 475479040, "q_proj": { "group_size": { @@ -23798,7 +23798,7 @@ } }, { - "accuracy": 0.9940579227711025, + "accuracy": 0.9941384611945403, "total_bits": 609759232, "q_proj": { "group_size": { @@ -23850,7 +23850,7 @@ } }, { - "accuracy": 0.9945268081991296, + "accuracy": 0.9945388110844713, "total_bits": 610024448, "q_proj": { "group_size": { @@ -23902,7 +23902,7 @@ } }, { - "accuracy": 0.9949111950240637, + "accuracy": 0.9949508191723573, "total_bits": 615020544, "q_proj": { "group_size": { @@ -23954,7 +23954,7 @@ } }, { - "accuracy": 0.9949385508110649, + "accuracy": 0.9949833927185912, "total_bits": 623951872, "q_proj": { "group_size": { @@ -24006,7 +24006,7 @@ } }, { - "accuracy": 0.9946152579627539, + "accuracy": 0.9946003548408809, "total_bits": 626473984, "q_proj": { "group_size": { @@ -24070,7 +24070,7 @@ } }, { - "accuracy": 0.9950193911790848, + "accuracy": 0.9950450925450576, "total_bits": 630355968, "q_proj": { "group_size": { @@ -24134,7 +24134,7 @@ } }, { - "accuracy": 0.9969427077785918, + "accuracy": 0.9970130016537089, "total_bits": 637362176, "q_proj": { "group_size": { @@ -24195,7 +24195,7 @@ } }, { - "accuracy": 0.9972639760296595, + "accuracy": 0.9972280528592435, "total_bits": 646823936, "q_proj": { "group_size": { @@ -24256,7 +24256,7 @@ } }, { - "accuracy": 0.9982951091308343, + "accuracy": 0.9983670686028505, "total_bits": 784740352, "q_proj": { "group_size": { @@ -24317,7 +24317,7 @@ } }, { - "accuracy": 0.9986039905955917, + "accuracy": 0.9986477568745613, "total_bits": 797818880, "q_proj": { "group_size": { @@ -24378,7 +24378,7 @@ } }, { - "accuracy": 0.9983756048114676, + "accuracy": 0.9984464643425063, "total_bits": 911749120, "q_proj": { "group_size": { @@ -24430,7 +24430,7 @@ } }, { - "accuracy": 0.9995518480182478, + "accuracy": 0.9995508000842834, "total_bits": 942718976, "q_proj": { "group_size": { @@ -24482,7 +24482,7 @@ } }, { - "accuracy": 0.9995935640640949, + "accuracy": 0.9996039943622523, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -24536,7 +24536,7 @@ ], "model.layers.12.mlp": [ { - "accuracy": 0.9865677246921941, + "accuracy": 0.986695270789297, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -24588,7 +24588,7 @@ } }, { - "accuracy": 0.986895313388423, + "accuracy": 0.9870199467006483, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -24640,7 +24640,7 @@ } }, { - "accuracy": 0.9890122209724627, + "accuracy": 0.9891081687651182, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -24689,7 +24689,7 @@ } }, { - "accuracy": 0.9896531630503503, + "accuracy": 0.9897396407629314, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -24738,7 +24738,7 @@ } }, { - "accuracy": 0.9934008482255434, + "accuracy": 0.9934639593488291, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -24790,7 +24790,7 @@ } }, { - "accuracy": 0.9939416556766159, + "accuracy": 0.9940013077698255, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -24842,7 +24842,7 @@ } }, { - "accuracy": 0.9948657676577568, + "accuracy": 0.9949125584803129, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -24891,7 +24891,7 @@ } }, { - "accuracy": 0.9966268178663755, + "accuracy": 0.9966604858636856, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -24934,7 +24934,7 @@ } }, { - "accuracy": 0.9969380768506151, + "accuracy": 0.9969685963894191, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -24977,7 +24977,7 @@ } }, { - "accuracy": 0.9966591961289707, + "accuracy": 0.996691306954936, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -25029,7 +25029,7 @@ } }, { - "accuracy": 0.997091862520105, + "accuracy": 0.9971206886203665, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -25081,7 +25081,7 @@ } }, { - "accuracy": 0.998299309101544, + "accuracy": 0.9983154229427639, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -25133,7 +25133,7 @@ } }, { - "accuracy": 0.9985526297241449, + "accuracy": 0.9985665949551683, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -25185,7 +25185,7 @@ } }, { - "accuracy": 0.9990910548520716, + "accuracy": 0.9990996045893744, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -25228,7 +25228,7 @@ } }, { - "accuracy": 0.9991329141745442, + "accuracy": 0.9991408528662041, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -25277,7 +25277,7 @@ } }, { - "accuracy": 0.9993119657431778, + "accuracy": 0.9993168057285642, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -25323,7 +25323,7 @@ } }, { - "accuracy": 0.9997550517467684, + "accuracy": 0.9997572318503731, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -25365,7 +25365,7 @@ ], "model.layers.13.self_attn": [ { - "accuracy": 0.9839275369518682, + "accuracy": 0.9843582808971405, "total_bits": 320757760, "q_proj": { "group_size": { @@ -25429,7 +25429,7 @@ } }, { - "accuracy": 0.9848762044781133, + "accuracy": 0.9848717106016058, "total_bits": 329080832, "q_proj": { "group_size": { @@ -25493,7 +25493,7 @@ } }, { - "accuracy": 0.9864384312378732, + "accuracy": 0.9865383286225168, "total_bits": 336024576, "q_proj": { "group_size": { @@ -25557,7 +25557,7 @@ } }, { - "accuracy": 0.9887322752099288, + "accuracy": 0.9888410497652856, "total_bits": 401557504, "q_proj": { "group_size": { @@ -25621,7 +25621,7 @@ } }, { - "accuracy": 0.9900884847891959, + "accuracy": 0.9901437594702369, "total_bits": 475279360, "q_proj": { "group_size": { @@ -25685,7 +25685,7 @@ } }, { - "accuracy": 0.9902212368814569, + "accuracy": 0.9902054032212809, "total_bits": 475479040, "q_proj": { "group_size": { @@ -25749,7 +25749,7 @@ } }, { - "accuracy": 0.9931259547409258, + "accuracy": 0.9933325483610755, "total_bits": 609759232, "q_proj": { "group_size": { @@ -25801,7 +25801,7 @@ } }, { - "accuracy": 0.993573922859995, + "accuracy": 0.9934128616985521, "total_bits": 610024448, "q_proj": { "group_size": { @@ -25853,7 +25853,7 @@ } }, { - "accuracy": 0.9939442124021681, + "accuracy": 0.9939653936185335, "total_bits": 615020544, "q_proj": { "group_size": { @@ -25905,7 +25905,7 @@ } }, { - "accuracy": 0.993992985079163, + "accuracy": 0.9940492722548937, "total_bits": 623951872, "q_proj": { "group_size": { @@ -25957,7 +25957,7 @@ } }, { - "accuracy": 0.9937816462234447, + "accuracy": 0.9935935912163634, "total_bits": 626473984, "q_proj": { "group_size": { @@ -26021,7 +26021,7 @@ } }, { - "accuracy": 0.9940637110879547, + "accuracy": 0.9940639479379905, "total_bits": 630355968, "q_proj": { "group_size": { @@ -26085,7 +26085,7 @@ } }, { - "accuracy": 0.9961284889202369, + "accuracy": 0.9961333714033428, "total_bits": 637362176, "q_proj": { "group_size": { @@ -26146,7 +26146,7 @@ } }, { - "accuracy": 0.996435731257263, + "accuracy": 0.9964956621590414, "total_bits": 646823936, "q_proj": { "group_size": { @@ -26207,7 +26207,7 @@ } }, { - "accuracy": 0.9979100621452457, + "accuracy": 0.9979398287832737, "total_bits": 784740352, "q_proj": { "group_size": { @@ -26268,7 +26268,7 @@ } }, { - "accuracy": 0.9982152434163972, + "accuracy": 0.9982787329507502, "total_bits": 797818880, "q_proj": { "group_size": { @@ -26329,7 +26329,7 @@ } }, { - "accuracy": 0.9981699676105851, + "accuracy": 0.9982085920086032, "total_bits": 911749120, "q_proj": { "group_size": { @@ -26381,7 +26381,7 @@ } }, { - "accuracy": 0.999332431605772, + "accuracy": 0.9993402297167402, "total_bits": 942718976, "q_proj": { "group_size": { @@ -26433,7 +26433,7 @@ } }, { - "accuracy": 0.9995105332743964, + "accuracy": 0.9995265997652161, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -26487,7 +26487,7 @@ ], "model.layers.13.mlp": [ { - "accuracy": 0.9856217832941758, + "accuracy": 0.9857253526386461, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -26539,7 +26539,7 @@ } }, { - "accuracy": 0.9859292695396825, + "accuracy": 0.9860281897218603, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -26591,7 +26591,7 @@ } }, { - "accuracy": 0.988535244998179, + "accuracy": 0.9886146360322049, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -26640,7 +26640,7 @@ } }, { - "accuracy": 0.9893272864191156, + "accuracy": 0.9893996668489355, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -26689,7 +26689,7 @@ } }, { - "accuracy": 0.992900127643033, + "accuracy": 0.9929526965869101, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -26741,7 +26741,7 @@ } }, { - "accuracy": 0.9934694994437067, + "accuracy": 0.9935186030833345, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -26793,7 +26793,7 @@ } }, { - "accuracy": 0.9946016034013346, + "accuracy": 0.994641392246673, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -26842,7 +26842,7 @@ } }, { - "accuracy": 0.9963686281913205, + "accuracy": 0.9963958118306963, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -26885,7 +26885,7 @@ } }, { - "accuracy": 0.9967179606227499, + "accuracy": 0.9967423257859129, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -26928,7 +26928,7 @@ } }, { - "accuracy": 0.9964085554606036, + "accuracy": 0.9964356014603063, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -26980,7 +26980,7 @@ } }, { - "accuracy": 0.9968671988891927, + "accuracy": 0.9968899631578672, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -27032,7 +27032,7 @@ } }, { - "accuracy": 0.9981744214892387, + "accuracy": 0.9981886885668102, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -27084,7 +27084,7 @@ } }, { - "accuracy": 0.9984414632383146, + "accuracy": 0.9984529572293946, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -27136,7 +27136,7 @@ } }, { - "accuracy": 0.9990289833788809, + "accuracy": 0.9990364588601025, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -27179,7 +27179,7 @@ } }, { - "accuracy": 0.9990714534528946, + "accuracy": 0.9990785945402948, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -27228,7 +27228,7 @@ } }, { - "accuracy": 0.9992947611761721, + "accuracy": 0.9992999734945203, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -27274,7 +27274,7 @@ } }, { - "accuracy": 0.9997391184105685, + "accuracy": 0.999741086613779, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -27316,7 +27316,7 @@ ], "model.layers.14.self_attn": [ { - "accuracy": 0.9814734647148534, + "accuracy": 0.9820133839782915, "total_bits": 320757760, "q_proj": { "group_size": { @@ -27380,7 +27380,7 @@ } }, { - "accuracy": 0.9818269202583715, + "accuracy": 0.9813823448984247, "total_bits": 329080832, "q_proj": { "group_size": { @@ -27444,7 +27444,7 @@ } }, { - "accuracy": 0.983623743057251, + "accuracy": 0.9842198679321691, "total_bits": 336024576, "q_proj": { "group_size": { @@ -27508,7 +27508,7 @@ } }, { - "accuracy": 0.9858457053962507, + "accuracy": 0.9865131644826186, "total_bits": 401557504, "q_proj": { "group_size": { @@ -27572,7 +27572,7 @@ } }, { - "accuracy": 0.9888085989575637, + "accuracy": 0.988668220607858, "total_bits": 475279360, "q_proj": { "group_size": { @@ -27636,7 +27636,7 @@ } }, { - "accuracy": 0.9891440868377686, + "accuracy": 0.9885247067401284, "total_bits": 475479040, "q_proj": { "group_size": { @@ -27700,7 +27700,7 @@ } }, { - "accuracy": 0.9926417085685229, + "accuracy": 0.9926554388121555, "total_bits": 609759232, "q_proj": { "group_size": { @@ -27752,7 +27752,7 @@ } }, { - "accuracy": 0.9930496717754164, + "accuracy": 0.9931092944584394, "total_bits": 610024448, "q_proj": { "group_size": { @@ -27804,7 +27804,7 @@ } }, { - "accuracy": 0.9936817790332594, + "accuracy": 0.9933561925825319, "total_bits": 615020544, "q_proj": { "group_size": { @@ -27856,7 +27856,7 @@ } }, { - "accuracy": 0.9937308481649348, + "accuracy": 0.9934791687287783, "total_bits": 623951872, "q_proj": { "group_size": { @@ -27908,7 +27908,7 @@ } }, { - "accuracy": 0.9930772020628578, + "accuracy": 0.9933117149691832, "total_bits": 626473984, "q_proj": { "group_size": { @@ -27972,7 +27972,7 @@ } }, { - "accuracy": 0.9936293916482675, + "accuracy": 0.9935862508259321, "total_bits": 630355968, "q_proj": { "group_size": { @@ -28036,7 +28036,7 @@ } }, { - "accuracy": 0.995640270804104, + "accuracy": 0.9955351654636232, "total_bits": 637362176, "q_proj": { "group_size": { @@ -28097,7 +28097,7 @@ } }, { - "accuracy": 0.9959704307349104, + "accuracy": 0.9958424548569479, "total_bits": 646823936, "q_proj": { "group_size": { @@ -28158,7 +28158,7 @@ } }, { - "accuracy": 0.997632081963514, + "accuracy": 0.9975629980234724, "total_bits": 784740352, "q_proj": { "group_size": { @@ -28219,7 +28219,7 @@ } }, { - "accuracy": 0.9979567953238362, + "accuracy": 0.9979482214701804, "total_bits": 797818880, "q_proj": { "group_size": { @@ -28280,7 +28280,7 @@ } }, { - "accuracy": 0.9979118644014785, + "accuracy": 0.9978087242496642, "total_bits": 911749120, "q_proj": { "group_size": { @@ -28332,7 +28332,7 @@ } }, { - "accuracy": 0.9992264473418656, + "accuracy": 0.9991968335387739, "total_bits": 942718976, "q_proj": { "group_size": { @@ -28384,7 +28384,7 @@ } }, { - "accuracy": 0.9994615794796693, + "accuracy": 0.9994433216358486, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -28438,7 +28438,7 @@ ], "model.layers.14.mlp": [ { - "accuracy": 0.9827684650295659, + "accuracy": 0.9828639061827409, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -28490,7 +28490,7 @@ } }, { - "accuracy": 0.9831489246142538, + "accuracy": 0.9832409416374407, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -28542,7 +28542,7 @@ } }, { - "accuracy": 0.9862094756804014, + "accuracy": 0.9862834921008662, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -28591,7 +28591,7 @@ } }, { - "accuracy": 0.9871377058719334, + "accuracy": 0.9872073884073057, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -28640,7 +28640,7 @@ } }, { - "accuracy": 0.9914754472280803, + "accuracy": 0.9915214251530798, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -28692,7 +28692,7 @@ } }, { - "accuracy": 0.9921506251159468, + "accuracy": 0.992193529480382, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -28744,7 +28744,7 @@ } }, { - "accuracy": 0.9934787303209305, + "accuracy": 0.9935133566981867, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -28793,7 +28793,7 @@ } }, { - "accuracy": 0.9956304783883848, + "accuracy": 0.9956533755911025, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -28836,7 +28836,7 @@ } }, { - "accuracy": 0.9960436389634484, + "accuracy": 0.9960661428539377, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -28879,7 +28879,7 @@ } }, { - "accuracy": 0.9956834073129454, + "accuracy": 0.995707499745645, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -28931,7 +28931,7 @@ } }, { - "accuracy": 0.9962318229832148, + "accuracy": 0.996252958711825, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -28983,7 +28983,7 @@ } }, { - "accuracy": 0.9978079409583619, + "accuracy": 0.9978218486434535, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -29035,7 +29035,7 @@ } }, { - "accuracy": 0.9981259433062453, + "accuracy": 0.9981356727841654, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -29087,7 +29087,7 @@ } }, { - "accuracy": 0.9988293649726793, + "accuracy": 0.9988386817276478, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -29130,7 +29130,7 @@ } }, { - "accuracy": 0.998884962185433, + "accuracy": 0.9988926025597673, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -29179,7 +29179,7 @@ } }, { - "accuracy": 0.999146368158491, + "accuracy": 0.9991530382907704, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -29225,7 +29225,7 @@ } }, { - "accuracy": 0.9996835034291603, + "accuracy": 0.9996904434244099, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -29267,7 +29267,7 @@ ], "model.layers.15.self_attn": [ { - "accuracy": 0.9790342261916712, + "accuracy": 0.9793184995651245, "total_bits": 320757760, "q_proj": { "group_size": { @@ -29331,7 +29331,7 @@ } }, { - "accuracy": 0.9821612129085943, + "accuracy": 0.9823253405721564, "total_bits": 329080832, "q_proj": { "group_size": { @@ -29395,7 +29395,7 @@ } }, { - "accuracy": 0.9852006200112795, + "accuracy": 0.9853477964275762, "total_bits": 336024576, "q_proj": { "group_size": { @@ -29459,7 +29459,7 @@ } }, { - "accuracy": 0.9876660328162344, + "accuracy": 0.9878452106526023, "total_bits": 401557504, "q_proj": { "group_size": { @@ -29523,7 +29523,7 @@ } }, { - "accuracy": 0.9886201290707839, + "accuracy": 0.988789395282143, "total_bits": 475279360, "q_proj": { "group_size": { @@ -29587,7 +29587,7 @@ } }, { - "accuracy": 0.9891896906651949, + "accuracy": 0.9892573011548895, "total_bits": 475479040, "q_proj": { "group_size": { @@ -29651,7 +29651,7 @@ } }, { - "accuracy": 0.9921950486145521, + "accuracy": 0.9921749116558778, "total_bits": 609759232, "q_proj": { "group_size": { @@ -29703,7 +29703,7 @@ } }, { - "accuracy": 0.992560185099903, + "accuracy": 0.9926794492884686, "total_bits": 610024448, "q_proj": { "group_size": { @@ -29755,7 +29755,7 @@ } }, { - "accuracy": 0.9929025149659106, + "accuracy": 0.9930727442628459, "total_bits": 615020544, "q_proj": { "group_size": { @@ -29807,7 +29807,7 @@ } }, { - "accuracy": 0.993019058516151, + "accuracy": 0.9931681360069075, "total_bits": 623951872, "q_proj": { "group_size": { @@ -29859,7 +29859,7 @@ } }, { - "accuracy": 0.9936291657780346, + "accuracy": 0.9933916706787912, "total_bits": 626473984, "q_proj": { "group_size": { @@ -29923,7 +29923,7 @@ } }, { - "accuracy": 0.9939859101646825, + "accuracy": 0.9938600710348079, "total_bits": 630355968, "q_proj": { "group_size": { @@ -29987,7 +29987,7 @@ } }, { - "accuracy": 0.9956992068573048, + "accuracy": 0.9954636296943614, "total_bits": 637362176, "q_proj": { "group_size": { @@ -30048,7 +30048,7 @@ } }, { - "accuracy": 0.9960234435765367, + "accuracy": 0.9960848987102509, "total_bits": 646823936, "q_proj": { "group_size": { @@ -30109,7 +30109,7 @@ } }, { - "accuracy": 0.9976367846523461, + "accuracy": 0.9976808448371134, "total_bits": 784740352, "q_proj": { "group_size": { @@ -30170,7 +30170,7 @@ } }, { - "accuracy": 0.9979342749636424, + "accuracy": 0.997999407351017, "total_bits": 797818880, "q_proj": { "group_size": { @@ -30231,7 +30231,7 @@ } }, { - "accuracy": 0.997900194243381, + "accuracy": 0.9979457457206751, "total_bits": 911749120, "q_proj": { "group_size": { @@ -30283,7 +30283,7 @@ } }, { - "accuracy": 0.9992511313114512, + "accuracy": 0.999265517843397, "total_bits": 942718976, "q_proj": { "group_size": { @@ -30335,7 +30335,7 @@ } }, { - "accuracy": 0.999450004806644, + "accuracy": 0.9994650766939709, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -30389,7 +30389,7 @@ ], "model.layers.15.mlp": [ { - "accuracy": 0.9800982804674852, + "accuracy": 0.980205320998242, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -30441,7 +30441,7 @@ } }, { - "accuracy": 0.9805547610709542, + "accuracy": 0.9806532593149888, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -30493,7 +30493,7 @@ } }, { - "accuracy": 0.9840819647437647, + "accuracy": 0.9841675648563787, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -30542,7 +30542,7 @@ } }, { - "accuracy": 0.9851276356922952, + "accuracy": 0.9852059436471838, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -30591,7 +30591,7 @@ } }, { - "accuracy": 0.9901798634152663, + "accuracy": 0.9902357317899403, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -30643,7 +30643,7 @@ } }, { - "accuracy": 0.9909477673078838, + "accuracy": 0.9909971136795847, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -30695,7 +30695,7 @@ } }, { - "accuracy": 0.9924567733940325, + "accuracy": 0.9924984174339395, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -30744,7 +30744,7 @@ } }, { - "accuracy": 0.9949847189219374, + "accuracy": 0.9950143887024177, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -30787,7 +30787,7 @@ } }, { - "accuracy": 0.9954444711145601, + "accuracy": 0.9954705430488837, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -30830,7 +30830,7 @@ } }, { - "accuracy": 0.9950298039536727, + "accuracy": 0.9950569954357649, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -30882,7 +30882,7 @@ } }, { - "accuracy": 0.9956484205628696, + "accuracy": 0.9956729439528365, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -30934,7 +30934,7 @@ } }, { - "accuracy": 0.9974746090409002, + "accuracy": 0.997489512554909, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -30986,7 +30986,7 @@ } }, { - "accuracy": 0.9978342985636309, + "accuracy": 0.9978469406303606, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -31038,7 +31038,7 @@ } }, { - "accuracy": 0.9986568450143463, + "accuracy": 0.9986654199463757, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -31081,7 +31081,7 @@ } }, { - "accuracy": 0.9987133937446695, + "accuracy": 0.9987220302420227, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -31130,7 +31130,7 @@ } }, { - "accuracy": 0.9990071982733513, + "accuracy": 0.9990142621099949, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -31176,7 +31176,7 @@ } }, { - "accuracy": 0.9996389649857423, + "accuracy": 0.9996447167114207, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -31218,7 +31218,7 @@ ], "model.layers.16.self_attn": [ { - "accuracy": 0.9774744165571112, + "accuracy": 0.9776952078467921, "total_bits": 320757760, "q_proj": { "group_size": { @@ -31282,7 +31282,7 @@ } }, { - "accuracy": 0.9807607845256203, + "accuracy": 0.9808052684131422, "total_bits": 329080832, "q_proj": { "group_size": { @@ -31346,7 +31346,7 @@ } }, { - "accuracy": 0.9825311161969837, + "accuracy": 0.9839467594498083, "total_bits": 336024576, "q_proj": { "group_size": { @@ -31410,7 +31410,7 @@ } }, { - "accuracy": 0.9857921961106753, + "accuracy": 0.9876000802767905, "total_bits": 401557504, "q_proj": { "group_size": { @@ -31474,7 +31474,7 @@ } }, { - "accuracy": 0.9874563044623325, + "accuracy": 0.9898211901125155, "total_bits": 475279360, "q_proj": { "group_size": { @@ -31538,7 +31538,7 @@ } }, { - "accuracy": 0.9879651148068277, + "accuracy": 0.988416366671261, "total_bits": 475479040, "q_proj": { "group_size": { @@ -31602,7 +31602,7 @@ } }, { - "accuracy": 0.9919261799046868, + "accuracy": 0.9921289078499141, "total_bits": 609759232, "q_proj": { "group_size": { @@ -31654,7 +31654,7 @@ } }, { - "accuracy": 0.9924536049365997, + "accuracy": 0.9924072191903466, "total_bits": 610024448, "q_proj": { "group_size": { @@ -31706,7 +31706,7 @@ } }, { - "accuracy": 0.9930006720517811, + "accuracy": 0.9931799087085222, "total_bits": 615020544, "q_proj": { "group_size": { @@ -31758,7 +31758,7 @@ } }, { - "accuracy": 0.9931202711243379, + "accuracy": 0.993347032289756, "total_bits": 623951872, "q_proj": { "group_size": { @@ -31810,7 +31810,7 @@ } }, { - "accuracy": 0.9932500092606795, + "accuracy": 0.9933705620075527, "total_bits": 626473984, "q_proj": { "group_size": { @@ -31874,7 +31874,7 @@ } }, { - "accuracy": 0.9938105438884935, + "accuracy": 0.9938321835116336, "total_bits": 630355968, "q_proj": { "group_size": { @@ -31938,7 +31938,7 @@ } }, { - "accuracy": 0.9954316843497125, + "accuracy": 0.9952971586271336, "total_bits": 637362176, "q_proj": { "group_size": { @@ -31999,7 +31999,7 @@ } }, { - "accuracy": 0.9958361963692465, + "accuracy": 0.9957568206285176, "total_bits": 646823936, "q_proj": { "group_size": { @@ -32060,7 +32060,7 @@ } }, { - "accuracy": 0.9975923433115608, + "accuracy": 0.9975966944506294, "total_bits": 784740352, "q_proj": { "group_size": { @@ -32121,7 +32121,7 @@ } }, { - "accuracy": 0.9979222944300425, + "accuracy": 0.9979627418674921, "total_bits": 797818880, "q_proj": { "group_size": { @@ -32182,7 +32182,7 @@ } }, { - "accuracy": 0.9980200650660616, + "accuracy": 0.9979964737829409, "total_bits": 911749120, "q_proj": { "group_size": { @@ -32234,7 +32234,7 @@ } }, { - "accuracy": 0.999147955524294, + "accuracy": 0.9991501622686261, "total_bits": 942718976, "q_proj": { "group_size": { @@ -32286,7 +32286,7 @@ } }, { - "accuracy": 0.9994592971511578, + "accuracy": 0.999473822940337, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -32340,7 +32340,7 @@ ], "model.layers.16.mlp": [ { - "accuracy": 0.9759851976444847, + "accuracy": 0.9761039865644354, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -32392,7 +32392,7 @@ } }, { - "accuracy": 0.9765292217856959, + "accuracy": 0.9766463025620109, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -32444,7 +32444,7 @@ } }, { - "accuracy": 0.9807863878576379, + "accuracy": 0.980880828280198, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -32493,7 +32493,7 @@ } }, { - "accuracy": 0.9820449995367151, + "accuracy": 0.9821350135301289, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -32542,7 +32542,7 @@ } }, { - "accuracy": 0.9880792412318682, + "accuracy": 0.988135449196163, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -32594,7 +32594,7 @@ } }, { - "accuracy": 0.9890126303622597, + "accuracy": 0.9890662070951963, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -32646,7 +32646,7 @@ } }, { - "accuracy": 0.9908301681280136, + "accuracy": 0.9908743155630011, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -32695,7 +32695,7 @@ } }, { - "accuracy": 0.9938990732556895, + "accuracy": 0.9939274972206668, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -32738,7 +32738,7 @@ } }, { - "accuracy": 0.9944598976718751, + "accuracy": 0.9944866938810599, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -32781,7 +32781,7 @@ } }, { - "accuracy": 0.9939581881228247, + "accuracy": 0.9939868312917257, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -32833,7 +32833,7 @@ } }, { - "accuracy": 0.9947057695765245, + "accuracy": 0.9947317285757316, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -32885,7 +32885,7 @@ } }, { - "accuracy": 0.9969260043611652, + "accuracy": 0.9969401532097867, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -32937,7 +32937,7 @@ } }, { - "accuracy": 0.9973642163370785, + "accuracy": 0.9973771209387403, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -32989,7 +32989,7 @@ } }, { - "accuracy": 0.9983645933435151, + "accuracy": 0.998372212640549, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -33032,7 +33032,7 @@ } }, { - "accuracy": 0.9984347308544737, + "accuracy": 0.99844167479559, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -33081,7 +33081,7 @@ } }, { - "accuracy": 0.9987973904139117, + "accuracy": 0.998802642092893, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -33127,7 +33127,7 @@ } }, { - "accuracy": 0.9995649672162376, + "accuracy": 0.9995669375516867, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -33169,7 +33169,7 @@ ], "model.layers.17.self_attn": [ { - "accuracy": 0.9715196741254706, + "accuracy": 0.9713898646204095, "total_bits": 320757760, "q_proj": { "group_size": { @@ -33233,7 +33233,7 @@ } }, { - "accuracy": 0.97568058810736, + "accuracy": 0.9760970175266266, "total_bits": 329080832, "q_proj": { "group_size": { @@ -33297,7 +33297,7 @@ } }, { - "accuracy": 0.9790116846561432, + "accuracy": 0.9791428462455147, "total_bits": 336024576, "q_proj": { "group_size": { @@ -33361,7 +33361,7 @@ } }, { - "accuracy": 0.9849396677393663, + "accuracy": 0.9851852184847781, "total_bits": 401557504, "q_proj": { "group_size": { @@ -33425,7 +33425,7 @@ } }, { - "accuracy": 0.9869026259372109, + "accuracy": 0.9872383522359949, "total_bits": 475279360, "q_proj": { "group_size": { @@ -33489,7 +33489,7 @@ } }, { - "accuracy": 0.9870980734887876, + "accuracy": 0.9875977619698173, "total_bits": 475479040, "q_proj": { "group_size": { @@ -33553,7 +33553,7 @@ } }, { - "accuracy": 0.9905216764462622, + "accuracy": 0.9909669976485403, "total_bits": 609759232, "q_proj": { "group_size": { @@ -33605,7 +33605,7 @@ } }, { - "accuracy": 0.9907246608483163, + "accuracy": 0.9914379543379733, "total_bits": 610024448, "q_proj": { "group_size": { @@ -33657,7 +33657,7 @@ } }, { - "accuracy": 0.9915780534869746, + "accuracy": 0.9917164295911789, "total_bits": 615020544, "q_proj": { "group_size": { @@ -33709,7 +33709,7 @@ } }, { - "accuracy": 0.9918630930938219, + "accuracy": 0.9920321820597899, "total_bits": 623951872, "q_proj": { "group_size": { @@ -33761,7 +33761,7 @@ } }, { - "accuracy": 0.9933998271038658, + "accuracy": 0.9934315305007132, "total_bits": 626473984, "q_proj": { "group_size": { @@ -33825,7 +33825,7 @@ } }, { - "accuracy": 0.9938390709851918, + "accuracy": 0.9939096072002461, "total_bits": 630355968, "q_proj": { "group_size": { @@ -33889,7 +33889,7 @@ } }, { - "accuracy": 0.9939844106373034, + "accuracy": 0.9940276938049417, "total_bits": 637362176, "q_proj": { "group_size": { @@ -33950,7 +33950,7 @@ } }, { - "accuracy": 0.9944978054416808, + "accuracy": 0.9945623631540098, "total_bits": 646823936, "q_proj": { "group_size": { @@ -34011,7 +34011,7 @@ } }, { - "accuracy": 0.9967323210286466, + "accuracy": 0.996811139740442, "total_bits": 784740352, "q_proj": { "group_size": { @@ -34072,7 +34072,7 @@ } }, { - "accuracy": 0.9972785481888997, + "accuracy": 0.9972696876839587, "total_bits": 797818880, "q_proj": { "group_size": { @@ -34133,7 +34133,7 @@ } }, { - "accuracy": 0.9973883854323312, + "accuracy": 0.9974801301171905, "total_bits": 911749120, "q_proj": { "group_size": { @@ -34185,7 +34185,7 @@ } }, { - "accuracy": 0.9987902869715503, + "accuracy": 0.9988025772924486, "total_bits": 942718976, "q_proj": { "group_size": { @@ -34237,7 +34237,7 @@ } }, { - "accuracy": 0.9993227730554185, + "accuracy": 0.9993391272758967, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -34291,7 +34291,7 @@ ], "model.layers.17.mlp": [ { - "accuracy": 0.9719695386133695, + "accuracy": 0.9720633218162938, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -34343,7 +34343,7 @@ } }, { - "accuracy": 0.9726032614707947, + "accuracy": 0.9726992186747099, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -34395,7 +34395,7 @@ } }, { - "accuracy": 0.9774761168580306, + "accuracy": 0.9775631600304654, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -34444,7 +34444,7 @@ } }, { - "accuracy": 0.9789059679759177, + "accuracy": 0.9789897768121016, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -34493,7 +34493,7 @@ } }, { - "accuracy": 0.986094782226964, + "accuracy": 0.9861453269657335, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -34545,7 +34545,7 @@ } }, { - "accuracy": 0.9871702554978823, + "accuracy": 0.9872179305867145, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -34597,7 +34597,7 @@ } }, { - "accuracy": 0.9892399130683196, + "accuracy": 0.9892799477828177, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -34646,7 +34646,7 @@ } }, { - "accuracy": 0.9928925060912183, + "accuracy": 0.992917966685797, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -34689,7 +34689,7 @@ } }, { - "accuracy": 0.9935314957248537, + "accuracy": 0.9935543654780639, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -34732,7 +34732,7 @@ } }, { - "accuracy": 0.9929547098122145, + "accuracy": 0.99297953828385, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -34784,7 +34784,7 @@ } }, { - "accuracy": 0.9938160086932936, + "accuracy": 0.9938386098334664, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -34836,7 +34836,7 @@ } }, { - "accuracy": 0.9964171220597468, + "accuracy": 0.9964298566705302, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -34888,7 +34888,7 @@ } }, { - "accuracy": 0.9969223424008018, + "accuracy": 0.9969337596312949, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -34940,7 +34940,7 @@ } }, { - "accuracy": 0.9980961037309546, + "accuracy": 0.9981033374604426, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -34983,7 +34983,7 @@ } }, { - "accuracy": 0.9981768966505402, + "accuracy": 0.9981829802456655, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -35032,7 +35032,7 @@ } }, { - "accuracy": 0.9985866530945426, + "accuracy": 0.9985915269506606, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -35078,7 +35078,7 @@ } }, { - "accuracy": 0.9994965166735806, + "accuracy": 0.9994983225570697, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -35120,7 +35120,7 @@ ], "model.layers.18.self_attn": [ { - "accuracy": 0.9681605163373446, + "accuracy": 0.9680482180495011, "total_bits": 320757760, "q_proj": { "group_size": { @@ -35184,7 +35184,7 @@ } }, { - "accuracy": 0.9730988615437558, + "accuracy": 0.9733026372758966, "total_bits": 329080832, "q_proj": { "group_size": { @@ -35248,7 +35248,7 @@ } }, { - "accuracy": 0.9775737555403459, + "accuracy": 0.9776360235716167, "total_bits": 336024576, "q_proj": { "group_size": { @@ -35312,7 +35312,7 @@ } }, { - "accuracy": 0.9838091809498636, + "accuracy": 0.9838520602176064, "total_bits": 401557504, "q_proj": { "group_size": { @@ -35376,7 +35376,7 @@ } }, { - "accuracy": 0.9859108250392111, + "accuracy": 0.986017318148362, "total_bits": 475279360, "q_proj": { "group_size": { @@ -35440,7 +35440,7 @@ } }, { - "accuracy": 0.9861576666957453, + "accuracy": 0.9863717320718264, "total_bits": 475479040, "q_proj": { "group_size": { @@ -35504,7 +35504,7 @@ } }, { - "accuracy": 0.9900413790815755, + "accuracy": 0.990062855576214, "total_bits": 609759232, "q_proj": { "group_size": { @@ -35556,7 +35556,7 @@ } }, { - "accuracy": 0.9903111959758558, + "accuracy": 0.9905819006656346, "total_bits": 610024448, "q_proj": { "group_size": { @@ -35608,7 +35608,7 @@ } }, { - "accuracy": 0.9910957201531059, + "accuracy": 0.9911310398264935, "total_bits": 615020544, "q_proj": { "group_size": { @@ -35660,7 +35660,7 @@ } }, { - "accuracy": 0.9913321317810762, + "accuracy": 0.991509284628065, "total_bits": 623951872, "q_proj": { "group_size": { @@ -35712,7 +35712,7 @@ } }, { - "accuracy": 0.9930900194142994, + "accuracy": 0.9931538967709792, "total_bits": 626473984, "q_proj": { "group_size": { @@ -35776,7 +35776,7 @@ } }, { - "accuracy": 0.9934194676185909, + "accuracy": 0.9935918740536037, "total_bits": 630355968, "q_proj": { "group_size": { @@ -35840,7 +35840,7 @@ } }, { - "accuracy": 0.9938246576409591, + "accuracy": 0.9938791856953972, "total_bits": 637362176, "q_proj": { "group_size": { @@ -35901,7 +35901,7 @@ } }, { - "accuracy": 0.994211527469911, + "accuracy": 0.9943228203215098, "total_bits": 646823936, "q_proj": { "group_size": { @@ -35962,7 +35962,7 @@ } }, { - "accuracy": 0.9966098215234908, + "accuracy": 0.9966090674463072, "total_bits": 784740352, "q_proj": { "group_size": { @@ -36023,7 +36023,7 @@ } }, { - "accuracy": 0.9971547811047027, + "accuracy": 0.997176323674227, "total_bits": 797818880, "q_proj": { "group_size": { @@ -36084,7 +36084,7 @@ } }, { - "accuracy": 0.997356252450692, + "accuracy": 0.9973441272189743, "total_bits": 911749120, "q_proj": { "group_size": { @@ -36136,7 +36136,7 @@ } }, { - "accuracy": 0.9987107640818546, + "accuracy": 0.9987269806626596, "total_bits": 942718976, "q_proj": { "group_size": { @@ -36188,7 +36188,7 @@ } }, { - "accuracy": 0.999305403791368, + "accuracy": 0.9993074919636312, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -36242,7 +36242,7 @@ ], "model.layers.18.mlp": [ { - "accuracy": 0.9654861280792638, + "accuracy": 0.9655057348703083, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -36294,7 +36294,7 @@ } }, { - "accuracy": 0.9662446442403292, + "accuracy": 0.9662556993333917, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -36346,7 +36346,7 @@ } }, { - "accuracy": 0.9709160076944452, + "accuracy": 0.9708994281919379, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -36395,7 +36395,7 @@ } }, { - "accuracy": 0.9722702911025599, + "accuracy": 0.9722480146508468, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -36444,7 +36444,7 @@ } }, { - "accuracy": 0.9832555588923002, + "accuracy": 0.9832617781664196, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -36496,7 +36496,7 @@ } }, { - "accuracy": 0.985322327990281, + "accuracy": 0.9853509115545374, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -36548,7 +36548,7 @@ } }, { - "accuracy": 0.9874796098784396, + "accuracy": 0.9875085573447379, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -36597,7 +36597,7 @@ } }, { - "accuracy": 0.9918587364648518, + "accuracy": 0.9918761292570516, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -36640,7 +36640,7 @@ } }, { - "accuracy": 0.992540294402524, + "accuracy": 0.9925548136234283, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -36683,7 +36683,7 @@ } }, { - "accuracy": 0.991927727272636, + "accuracy": 0.991952972976785, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -36735,7 +36735,7 @@ } }, { - "accuracy": 0.9928613681542245, + "accuracy": 0.9928739031678752, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -36787,7 +36787,7 @@ } }, { - "accuracy": 0.9958819664622608, + "accuracy": 0.9958894633148846, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -36839,7 +36839,7 @@ } }, { - "accuracy": 0.9964668931145417, + "accuracy": 0.9964724614431983, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -36891,7 +36891,7 @@ } }, { - "accuracy": 0.9977746543131376, + "accuracy": 0.9977779125696734, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -36934,7 +36934,7 @@ } }, { - "accuracy": 0.9978991329277817, + "accuracy": 0.9979063956753204, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -36983,7 +36983,7 @@ } }, { - "accuracy": 0.9983252496703675, + "accuracy": 0.9983331329728428, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -37029,7 +37029,7 @@ } }, { - "accuracy": 0.9993686606421283, + "accuracy": 0.9993839482531736, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -37071,7 +37071,7 @@ ], "model.layers.19.self_attn": [ { - "accuracy": 0.9729026963836268, + "accuracy": 0.9737109322296945, "total_bits": 320757760, "q_proj": { "group_size": { @@ -37135,7 +37135,7 @@ } }, { - "accuracy": 0.9759520856957686, + "accuracy": 0.9762213261503923, "total_bits": 329080832, "q_proj": { "group_size": { @@ -37199,7 +37199,7 @@ } }, { - "accuracy": 0.9790839019574618, + "accuracy": 0.9791505493615803, "total_bits": 336024576, "q_proj": { "group_size": { @@ -37263,7 +37263,7 @@ } }, { - "accuracy": 0.9837972537467354, + "accuracy": 0.9839103582658266, "total_bits": 401557504, "q_proj": { "group_size": { @@ -37327,7 +37327,7 @@ } }, { - "accuracy": 0.985572127919448, + "accuracy": 0.9857514830012071, "total_bits": 475279360, "q_proj": { "group_size": { @@ -37391,7 +37391,7 @@ } }, { - "accuracy": 0.9859857135697415, + "accuracy": 0.9858928667871576, "total_bits": 475479040, "q_proj": { "group_size": { @@ -37455,7 +37455,7 @@ } }, { - "accuracy": 0.9905594485370737, + "accuracy": 0.9905924695102792, "total_bits": 609759232, "q_proj": { "group_size": { @@ -37507,7 +37507,7 @@ } }, { - "accuracy": 0.9911070771907505, + "accuracy": 0.9910375993502768, "total_bits": 610024448, "q_proj": { "group_size": { @@ -37559,7 +37559,7 @@ } }, { - "accuracy": 0.9915314952009603, + "accuracy": 0.9915908731912312, "total_bits": 615020544, "q_proj": { "group_size": { @@ -37611,7 +37611,7 @@ } }, { - "accuracy": 0.9917484419910532, + "accuracy": 0.9918176304352911, "total_bits": 623951872, "q_proj": { "group_size": { @@ -37663,7 +37663,7 @@ } }, { - "accuracy": 0.9923635532981471, + "accuracy": 0.992223880792919, "total_bits": 626473984, "q_proj": { "group_size": { @@ -37727,7 +37727,7 @@ } }, { - "accuracy": 0.9929303809216148, + "accuracy": 0.9927884238330942, "total_bits": 630355968, "q_proj": { "group_size": { @@ -37791,7 +37791,7 @@ } }, { - "accuracy": 0.9942923050961996, + "accuracy": 0.9942889433158072, "total_bits": 637362176, "q_proj": { "group_size": { @@ -37852,7 +37852,7 @@ } }, { - "accuracy": 0.9947933797773562, + "accuracy": 0.9947850614001876, "total_bits": 646823936, "q_proj": { "group_size": { @@ -37913,7 +37913,7 @@ } }, { - "accuracy": 0.9969351044610927, + "accuracy": 0.9970426486903116, "total_bits": 784740352, "q_proj": { "group_size": { @@ -37974,7 +37974,7 @@ } }, { - "accuracy": 0.9973653645107621, + "accuracy": 0.9974029464157004, "total_bits": 797818880, "q_proj": { "group_size": { @@ -38035,7 +38035,7 @@ } }, { - "accuracy": 0.9975087603456095, + "accuracy": 0.9976341604794327, "total_bits": 911749120, "q_proj": { "group_size": { @@ -38087,7 +38087,7 @@ } }, { - "accuracy": 0.9988727033334343, + "accuracy": 0.9988766628268518, "total_bits": 942718976, "q_proj": { "group_size": { @@ -38139,7 +38139,7 @@ } }, { - "accuracy": 0.9993417262540836, + "accuracy": 0.9993570163649949, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -38193,7 +38193,7 @@ ], "model.layers.19.mlp": [ { - "accuracy": 0.9650533952211079, + "accuracy": 0.9650920880468268, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -38245,7 +38245,7 @@ } }, { - "accuracy": 0.9659970716426247, + "accuracy": 0.9660348233423734, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -38297,7 +38297,7 @@ } }, { - "accuracy": 0.9715526292198583, + "accuracy": 0.9715854745162161, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -38346,7 +38346,7 @@ } }, { - "accuracy": 0.9730949495968065, + "accuracy": 0.973126376930036, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -38395,7 +38395,7 @@ } }, { - "accuracy": 0.9828524777763769, + "accuracy": 0.9828653931617737, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -38447,7 +38447,7 @@ } }, { - "accuracy": 0.9841679256213339, + "accuracy": 0.984185350568671, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -38499,7 +38499,7 @@ } }, { - "accuracy": 0.9863829628417367, + "accuracy": 0.9863996913558558, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -38548,7 +38548,7 @@ } }, { - "accuracy": 0.9912811102051484, + "accuracy": 0.9912879388583334, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -38591,7 +38591,7 @@ } }, { - "accuracy": 0.9919791339259398, + "accuracy": 0.9919891890726591, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -38634,7 +38634,7 @@ } }, { - "accuracy": 0.9913126943927062, + "accuracy": 0.9913182423303002, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -38686,7 +38686,7 @@ } }, { - "accuracy": 0.9923596209601352, + "accuracy": 0.9923677577784187, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -38738,7 +38738,7 @@ } }, { - "accuracy": 0.9955839938239047, + "accuracy": 0.9955875481429853, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -38790,7 +38790,7 @@ } }, { - "accuracy": 0.9961992945326003, + "accuracy": 0.9962027343480211, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -38842,7 +38842,7 @@ } }, { - "accuracy": 0.9976443250320459, + "accuracy": 0.9976474803529287, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -38885,7 +38885,7 @@ } }, { - "accuracy": 0.9977479969200335, + "accuracy": 0.997750458748717, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -38934,7 +38934,7 @@ } }, { - "accuracy": 0.9981824632145857, + "accuracy": 0.9981850832700729, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -38980,7 +38980,7 @@ } }, { - "accuracy": 0.9993796464251844, + "accuracy": 0.9993805524549986, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -39022,7 +39022,7 @@ ], "model.layers.20.self_attn": [ { - "accuracy": 0.9767009866865057, + "accuracy": 0.9766052412359338, "total_bits": 320757760, "q_proj": { "group_size": { @@ -39086,7 +39086,7 @@ } }, { - "accuracy": 0.9773092442437222, + "accuracy": 0.977346238337065, "total_bits": 329080832, "q_proj": { "group_size": { @@ -39150,7 +39150,7 @@ } }, { - "accuracy": 0.9821581495435614, + "accuracy": 0.9820998436526248, "total_bits": 336024576, "q_proj": { "group_size": { @@ -39214,7 +39214,7 @@ } }, { - "accuracy": 0.9859638951326671, + "accuracy": 0.985923076930799, "total_bits": 401557504, "q_proj": { "group_size": { @@ -39278,7 +39278,7 @@ } }, { - "accuracy": 0.9872828306336152, + "accuracy": 0.9872548540956095, "total_bits": 475279360, "q_proj": { "group_size": { @@ -39342,7 +39342,7 @@ } }, { - "accuracy": 0.9875408936488, + "accuracy": 0.9877182276625383, "total_bits": 475479040, "q_proj": { "group_size": { @@ -39406,7 +39406,7 @@ } }, { - "accuracy": 0.9896065780991002, + "accuracy": 0.9896546398338518, "total_bits": 609759232, "q_proj": { "group_size": { @@ -39458,7 +39458,7 @@ } }, { - "accuracy": 0.989953706923284, + "accuracy": 0.9900369518681577, "total_bits": 610024448, "q_proj": { "group_size": { @@ -39510,7 +39510,7 @@ } }, { - "accuracy": 0.9905643470977482, + "accuracy": 0.9906222898709146, "total_bits": 615020544, "q_proj": { "group_size": { @@ -39562,7 +39562,7 @@ } }, { - "accuracy": 0.990689818012087, + "accuracy": 0.9907642100986681, "total_bits": 623951872, "q_proj": { "group_size": { @@ -39614,7 +39614,7 @@ } }, { - "accuracy": 0.9931569907226061, + "accuracy": 0.9932596409007123, "total_bits": 626473984, "q_proj": { "group_size": { @@ -39678,7 +39678,7 @@ } }, { - "accuracy": 0.9936121827677676, + "accuracy": 0.99361638999299, "total_bits": 630355968, "q_proj": { "group_size": { @@ -39742,7 +39742,7 @@ } }, { - "accuracy": 0.9936603656724879, + "accuracy": 0.9936650607146715, "total_bits": 637362176, "q_proj": { "group_size": { @@ -39803,7 +39803,7 @@ } }, { - "accuracy": 0.9941092549186004, + "accuracy": 0.9940972045848244, "total_bits": 646823936, "q_proj": { "group_size": { @@ -39864,7 +39864,7 @@ } }, { - "accuracy": 0.9964999685946264, + "accuracy": 0.9965014673377338, "total_bits": 784740352, "q_proj": { "group_size": { @@ -39925,7 +39925,7 @@ } }, { - "accuracy": 0.9969227104202697, + "accuracy": 0.9969358908894815, "total_bits": 797818880, "q_proj": { "group_size": { @@ -39986,7 +39986,7 @@ } }, { - "accuracy": 0.9968381069208446, + "accuracy": 0.9968365503377036, "total_bits": 911749120, "q_proj": { "group_size": { @@ -40038,7 +40038,7 @@ } }, { - "accuracy": 0.998851046946488, + "accuracy": 0.9988400318512791, "total_bits": 942718976, "q_proj": { "group_size": { @@ -40090,7 +40090,7 @@ } }, { - "accuracy": 0.999160528967255, + "accuracy": 0.9991567998536324, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -40144,7 +40144,7 @@ ], "model.layers.20.mlp": [ { - "accuracy": 0.9644091756720292, + "accuracy": 0.964422461233641, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -40196,7 +40196,7 @@ } }, { - "accuracy": 0.9652943485661557, + "accuracy": 0.9653010368347168, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -40248,7 +40248,7 @@ } }, { - "accuracy": 0.9708892640314604, + "accuracy": 0.9708965734431618, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -40297,7 +40297,7 @@ } }, { - "accuracy": 0.9724627388151068, + "accuracy": 0.972467093091262, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -40346,7 +40346,7 @@ } }, { - "accuracy": 0.982363478133553, + "accuracy": 0.9823654560666335, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -40398,7 +40398,7 @@ } }, { - "accuracy": 0.9837264371545691, + "accuracy": 0.9837321764544437, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -40450,7 +40450,7 @@ } }, { - "accuracy": 0.986005016063389, + "accuracy": 0.9860120980363143, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -40499,7 +40499,7 @@ } }, { - "accuracy": 0.9910168945789337, + "accuracy": 0.9910178694285845, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -40542,7 +40542,7 @@ } }, { - "accuracy": 0.9917686965904737, + "accuracy": 0.9917707302068409, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -40585,7 +40585,7 @@ } }, { - "accuracy": 0.991059380926584, + "accuracy": 0.9910599416808078, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -40637,7 +40637,7 @@ } }, { - "accuracy": 0.9921420114605051, + "accuracy": 0.9921441995783856, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -40689,7 +40689,7 @@ } }, { - "accuracy": 0.9954552415170168, + "accuracy": 0.9954544839106108, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -40741,7 +40741,7 @@ } }, { - "accuracy": 0.996090472528809, + "accuracy": 0.9960912073913374, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -40793,7 +40793,7 @@ } }, { - "accuracy": 0.9975828787213877, + "accuracy": 0.9975826069712639, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -40836,7 +40836,7 @@ } }, { - "accuracy": 0.9976841140734521, + "accuracy": 0.9976837868361097, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -40885,7 +40885,7 @@ } }, { - "accuracy": 0.9981321167004736, + "accuracy": 0.9981317163298005, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -40931,7 +40931,7 @@ } }, { - "accuracy": 0.9993634440593029, + "accuracy": 0.9993634718029123, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -40973,7 +40973,7 @@ ], "model.layers.21.self_attn": [ { - "accuracy": 0.9695823286709032, + "accuracy": 0.9700161500980979, "total_bits": 320757760, "q_proj": { "group_size": { @@ -41037,7 +41037,7 @@ } }, { - "accuracy": 0.97170784285194, + "accuracy": 0.9730572449533563, "total_bits": 329080832, "q_proj": { "group_size": { @@ -41101,7 +41101,7 @@ } }, { - "accuracy": 0.9750879108905792, + "accuracy": 0.9754101085035425, "total_bits": 336024576, "q_proj": { "group_size": { @@ -41165,7 +41165,7 @@ } }, { - "accuracy": 0.9799557877214331, + "accuracy": 0.9803256314051779, "total_bits": 401557504, "q_proj": { "group_size": { @@ -41229,7 +41229,7 @@ } }, { - "accuracy": 0.9832003038180502, + "accuracy": 0.9837953777689683, "total_bits": 475279360, "q_proj": { "group_size": { @@ -41293,7 +41293,7 @@ } }, { - "accuracy": 0.9837385714054108, + "accuracy": 0.9840459023651323, "total_bits": 475479040, "q_proj": { "group_size": { @@ -41357,7 +41357,7 @@ } }, { - "accuracy": 0.9897666110804206, + "accuracy": 0.9892111106922752, "total_bits": 609759232, "q_proj": { "group_size": { @@ -41409,7 +41409,7 @@ } }, { - "accuracy": 0.9899154948560815, + "accuracy": 0.9897596757662924, "total_bits": 610024448, "q_proj": { "group_size": { @@ -41461,7 +41461,7 @@ } }, { - "accuracy": 0.9905949109479001, + "accuracy": 0.9909442608293734, "total_bits": 615020544, "q_proj": { "group_size": { @@ -41513,7 +41513,7 @@ } }, { - "accuracy": 0.9907519448744623, + "accuracy": 0.9911437026764217, "total_bits": 623951872, "q_proj": { "group_size": { @@ -41565,7 +41565,7 @@ } }, { - "accuracy": 0.9912045590187374, + "accuracy": 0.9912000353399076, "total_bits": 626473984, "q_proj": { "group_size": { @@ -41629,7 +41629,7 @@ } }, { - "accuracy": 0.991843014171249, + "accuracy": 0.9919030611452303, "total_bits": 630355968, "q_proj": { "group_size": { @@ -41693,7 +41693,7 @@ } }, { - "accuracy": 0.9935931469264784, + "accuracy": 0.9936806504663668, "total_bits": 637362176, "q_proj": { "group_size": { @@ -41754,7 +41754,7 @@ } }, { - "accuracy": 0.9940820493196186, + "accuracy": 0.9942079209967664, "total_bits": 646823936, "q_proj": { "group_size": { @@ -41815,7 +41815,7 @@ } }, { - "accuracy": 0.996579543932488, + "accuracy": 0.996496567208516, "total_bits": 784740352, "q_proj": { "group_size": { @@ -41876,7 +41876,7 @@ } }, { - "accuracy": 0.9971035609119817, + "accuracy": 0.9971058352997428, "total_bits": 797818880, "q_proj": { "group_size": { @@ -41937,7 +41937,7 @@ } }, { - "accuracy": 0.9972943687125256, + "accuracy": 0.9971978733021962, "total_bits": 911749120, "q_proj": { "group_size": { @@ -41989,7 +41989,7 @@ } }, { - "accuracy": 0.998738927178477, + "accuracy": 0.9987386128816166, "total_bits": 942718976, "q_proj": { "group_size": { @@ -42041,7 +42041,7 @@ } }, { - "accuracy": 0.9992956594122868, + "accuracy": 0.9992830469027946, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -42095,7 +42095,7 @@ ], "model.layers.21.mlp": [ { - "accuracy": 0.9620468773339924, + "accuracy": 0.9619991716585661, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -42147,7 +42147,7 @@ } }, { - "accuracy": 0.9630083880926433, + "accuracy": 0.9629664640677602, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -42199,7 +42199,7 @@ } }, { - "accuracy": 0.9688292177099931, + "accuracy": 0.9687966704368591, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -42248,7 +42248,7 @@ } }, { - "accuracy": 0.970436598125257, + "accuracy": 0.9704079784845051, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -42297,7 +42297,7 @@ } }, { - "accuracy": 0.9811940616682956, + "accuracy": 0.9811710432956093, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -42349,7 +42349,7 @@ } }, { - "accuracy": 0.9826522657745763, + "accuracy": 0.982629939129478, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -42401,7 +42401,7 @@ } }, { - "accuracy": 0.9849876337929776, + "accuracy": 0.9849691579216405, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -42450,7 +42450,7 @@ } }, { - "accuracy": 0.9904371609813288, + "accuracy": 0.9904225947041261, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -42493,7 +42493,7 @@ } }, { - "accuracy": 0.991214959244979, + "accuracy": 0.9912022697298151, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -42536,7 +42536,7 @@ } }, { - "accuracy": 0.9904567058149137, + "accuracy": 0.9904456405263198, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -42588,7 +42588,7 @@ } }, { - "accuracy": 0.9916066472467623, + "accuracy": 0.9915948235674908, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -42640,7 +42640,7 @@ } }, { - "accuracy": 0.9951455526446041, + "accuracy": 0.9951382353901863, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -42692,7 +42692,7 @@ } }, { - "accuracy": 0.995820530151066, + "accuracy": 0.9958154425809258, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -42744,7 +42744,7 @@ } }, { - "accuracy": 0.9974156175004808, + "accuracy": 0.9974102630819145, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -42787,7 +42787,7 @@ } }, { - "accuracy": 0.9975236726826743, + "accuracy": 0.9975199599407221, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -42836,7 +42836,7 @@ } }, { - "accuracy": 0.997979996627883, + "accuracy": 0.9979769075779539, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -42882,7 +42882,7 @@ } }, { - "accuracy": 0.9993146075622031, + "accuracy": 0.9993132474881253, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -42924,7 +42924,7 @@ ], "model.layers.22.self_attn": [ { - "accuracy": 0.9745166505637922, + "accuracy": 0.9744256640735426, "total_bits": 320757760, "q_proj": { "group_size": { @@ -42988,7 +42988,7 @@ } }, { - "accuracy": 0.9751788016996885, + "accuracy": 0.975161783005062, "total_bits": 329080832, "q_proj": { "group_size": { @@ -43052,7 +43052,7 @@ } }, { - "accuracy": 0.9795838202300825, + "accuracy": 0.9799705822216837, "total_bits": 336024576, "q_proj": { "group_size": { @@ -43116,7 +43116,7 @@ } }, { - "accuracy": 0.9825019648200587, + "accuracy": 0.982998959327999, "total_bits": 401557504, "q_proj": { "group_size": { @@ -43180,7 +43180,7 @@ } }, { - "accuracy": 0.983738816098163, + "accuracy": 0.9842252872492138, "total_bits": 475279360, "q_proj": { "group_size": { @@ -43244,7 +43244,7 @@ } }, { - "accuracy": 0.9842250017743361, + "accuracy": 0.9854801830492521, "total_bits": 475479040, "q_proj": { "group_size": { @@ -43308,7 +43308,7 @@ } }, { - "accuracy": 0.9857631529632368, + "accuracy": 0.9859433032964405, "total_bits": 609759232, "q_proj": { "group_size": { @@ -43360,7 +43360,7 @@ } }, { - "accuracy": 0.9862003326416016, + "accuracy": 0.9876655779386821, "total_bits": 610024448, "q_proj": { "group_size": { @@ -43412,7 +43412,7 @@ } }, { - "accuracy": 0.9881240145156258, + "accuracy": 0.9884650950369082, "total_bits": 615020544, "q_proj": { "group_size": { @@ -43464,7 +43464,7 @@ } }, { - "accuracy": 0.9882125101591411, + "accuracy": 0.9885191689980658, "total_bits": 623951872, "q_proj": { "group_size": { @@ -43516,7 +43516,7 @@ } }, { - "accuracy": 0.9922403081467277, + "accuracy": 0.992128420032953, "total_bits": 626473984, "q_proj": { "group_size": { @@ -43580,7 +43580,7 @@ } }, { - "accuracy": 0.9929560548380801, + "accuracy": 0.9928544532311591, "total_bits": 630355968, "q_proj": { "group_size": { @@ -43644,7 +43644,7 @@ } }, { - "accuracy": 0.9926713807018179, + "accuracy": 0.992592882953192, "total_bits": 637362176, "q_proj": { "group_size": { @@ -43705,7 +43705,7 @@ } }, { - "accuracy": 0.9934507459402084, + "accuracy": 0.9932609278904764, "total_bits": 646823936, "q_proj": { "group_size": { @@ -43766,7 +43766,7 @@ } }, { - "accuracy": 0.9959760802356821, + "accuracy": 0.9960086937797697, "total_bits": 784740352, "q_proj": { "group_size": { @@ -43827,7 +43827,7 @@ } }, { - "accuracy": 0.9965576262850511, + "accuracy": 0.9965175064770799, "total_bits": 797818880, "q_proj": { "group_size": { @@ -43888,7 +43888,7 @@ } }, { - "accuracy": 0.9962937416214692, + "accuracy": 0.9963155922136808, "total_bits": 911749120, "q_proj": { "group_size": { @@ -43940,7 +43940,7 @@ } }, { - "accuracy": 0.9987868323530021, + "accuracy": 0.9988025383729684, "total_bits": 942718976, "q_proj": { "group_size": { @@ -43992,7 +43992,7 @@ } }, { - "accuracy": 0.9989936473338228, + "accuracy": 0.9990511244457019, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -44046,7 +44046,7 @@ ], "model.layers.22.mlp": [ { - "accuracy": 0.9608007700819718, + "accuracy": 0.9607152123200267, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -44098,7 +44098,7 @@ } }, { - "accuracy": 0.961834989095989, + "accuracy": 0.9617486501994886, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -44150,7 +44150,7 @@ } }, { - "accuracy": 0.9676527192718104, + "accuracy": 0.9675877439348322, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -44199,7 +44199,7 @@ } }, { - "accuracy": 0.9692335818943224, + "accuracy": 0.9691751066007113, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -44248,7 +44248,7 @@ } }, { - "accuracy": 0.9805956827966791, + "accuracy": 0.980549272738005, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -44300,7 +44300,7 @@ } }, { - "accuracy": 0.9821103340701053, + "accuracy": 0.9820693997960341, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -44352,7 +44352,7 @@ } }, { - "accuracy": 0.9844140567277607, + "accuracy": 0.984381062419791, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -44401,7 +44401,7 @@ } }, { - "accuracy": 0.9901505457727533, + "accuracy": 0.9901273242737118, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -44444,7 +44444,7 @@ } }, { - "accuracy": 0.9909463352278659, + "accuracy": 0.9909234768465945, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -44487,7 +44487,7 @@ } }, { - "accuracy": 0.9901514610177592, + "accuracy": 0.9901277391534102, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -44539,7 +44539,7 @@ } }, { - "accuracy": 0.9913530647754669, + "accuracy": 0.991334150496282, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -44591,7 +44591,7 @@ } }, { - "accuracy": 0.9949887763512762, + "accuracy": 0.9949777067491883, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -44643,7 +44643,7 @@ } }, { - "accuracy": 0.995695952914263, + "accuracy": 0.9956871388774169, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -44695,7 +44695,7 @@ } }, { - "accuracy": 0.9973359141302737, + "accuracy": 0.9973292046863782, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -44738,7 +44738,7 @@ } }, { - "accuracy": 0.9974426235023298, + "accuracy": 0.9974376918061784, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -44787,7 +44787,7 @@ } }, { - "accuracy": 0.9978846897812266, + "accuracy": 0.9978808007742229, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -44833,7 +44833,7 @@ } }, { - "accuracy": 0.9993005203279225, + "accuracy": 0.9992990423189966, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -44875,7 +44875,7 @@ ], "model.layers.23.self_attn": [ { - "accuracy": 0.9751913971022556, + "accuracy": 0.9753004845819975, "total_bits": 320757760, "q_proj": { "group_size": { @@ -44939,7 +44939,7 @@ } }, { - "accuracy": 0.9757795224064275, + "accuracy": 0.975831886655406, "total_bits": 329080832, "q_proj": { "group_size": { @@ -45003,7 +45003,7 @@ } }, { - "accuracy": 0.9801134984744223, + "accuracy": 0.9802416641461221, "total_bits": 336024576, "q_proj": { "group_size": { @@ -45067,7 +45067,7 @@ } }, { - "accuracy": 0.9839488958057604, + "accuracy": 0.9841173008868569, "total_bits": 401557504, "q_proj": { "group_size": { @@ -45131,7 +45131,7 @@ } }, { - "accuracy": 0.9858585988220415, + "accuracy": 0.9862215895401804, "total_bits": 475279360, "q_proj": { "group_size": { @@ -45195,7 +45195,7 @@ } }, { - "accuracy": 0.9860215312556216, + "accuracy": 0.9862355480068609, "total_bits": 475479040, "q_proj": { "group_size": { @@ -45259,7 +45259,7 @@ } }, { - "accuracy": 0.9887161184298364, + "accuracy": 0.9889712600331557, "total_bits": 609759232, "q_proj": { "group_size": { @@ -45311,7 +45311,7 @@ } }, { - "accuracy": 0.9887879890830893, + "accuracy": 0.9889935098196331, "total_bits": 610024448, "q_proj": { "group_size": { @@ -45363,7 +45363,7 @@ } }, { - "accuracy": 0.989736546027033, + "accuracy": 0.9899147403867621, "total_bits": 615020544, "q_proj": { "group_size": { @@ -45415,7 +45415,7 @@ } }, { - "accuracy": 0.9899234912897411, + "accuracy": 0.990104458049724, "total_bits": 623951872, "q_proj": { "group_size": { @@ -45467,7 +45467,7 @@ } }, { - "accuracy": 0.9926448597719795, + "accuracy": 0.9927597602731303, "total_bits": 626473984, "q_proj": { "group_size": { @@ -45531,7 +45531,7 @@ } }, { - "accuracy": 0.9932931818460163, + "accuracy": 0.9931924241153818, "total_bits": 630355968, "q_proj": { "group_size": { @@ -45595,7 +45595,7 @@ } }, { - "accuracy": 0.9931735286587163, + "accuracy": 0.993300166569258, "total_bits": 637362176, "q_proj": { "group_size": { @@ -45656,7 +45656,7 @@ } }, { - "accuracy": 0.9937800941498656, + "accuracy": 0.9937387877389005, "total_bits": 646823936, "q_proj": { "group_size": { @@ -45717,7 +45717,7 @@ } }, { - "accuracy": 0.9962995926800527, + "accuracy": 0.9962737026967501, "total_bits": 784740352, "q_proj": { "group_size": { @@ -45778,7 +45778,7 @@ } }, { - "accuracy": 0.9967598387677419, + "accuracy": 0.9967570104881337, "total_bits": 797818880, "q_proj": { "group_size": { @@ -45839,7 +45839,7 @@ } }, { - "accuracy": 0.99675360753348, + "accuracy": 0.9967248859374147, "total_bits": 911749120, "q_proj": { "group_size": { @@ -45891,7 +45891,7 @@ } }, { - "accuracy": 0.9987847340342245, + "accuracy": 0.9987885505941353, "total_bits": 942718976, "q_proj": { "group_size": { @@ -45943,7 +45943,7 @@ } }, { - "accuracy": 0.9991235289918748, + "accuracy": 0.9991154476608101, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -45997,7 +45997,7 @@ ], "model.layers.23.mlp": [ { - "accuracy": 0.9599518964165136, + "accuracy": 0.9598137171644914, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -46049,7 +46049,7 @@ } }, { - "accuracy": 0.9609983563423157, + "accuracy": 0.9608627934204905, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -46101,7 +46101,7 @@ } }, { - "accuracy": 0.9667852928763941, + "accuracy": 0.9666714981982583, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -46150,7 +46150,7 @@ } }, { - "accuracy": 0.9683656912100942, + "accuracy": 0.9682609438896179, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -46199,7 +46199,7 @@ } }, { - "accuracy": 0.9801521285584098, + "accuracy": 0.9800864709051031, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -46251,7 +46251,7 @@ } }, { - "accuracy": 0.9817046821117401, + "accuracy": 0.9816410729759618, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -46303,7 +46303,7 @@ } }, { - "accuracy": 0.9840270171039983, + "accuracy": 0.9839700397692228, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -46352,7 +46352,7 @@ } }, { - "accuracy": 0.9899451395398692, + "accuracy": 0.9899114229177174, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -46395,7 +46395,7 @@ } }, { - "accuracy": 0.990760696561713, + "accuracy": 0.9907288480746118, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -46438,7 +46438,7 @@ } }, { - "accuracy": 0.9899386700830961, + "accuracy": 0.9899101076941741, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -46490,7 +46490,7 @@ } }, { - "accuracy": 0.991173588915875, + "accuracy": 0.9911445441998934, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -46542,7 +46542,7 @@ } }, { - "accuracy": 0.9948889762163162, + "accuracy": 0.9948766451132925, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -46594,7 +46594,7 @@ } }, { - "accuracy": 0.995609241096597, + "accuracy": 0.9955955096765569, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -46646,7 +46646,7 @@ } }, { - "accuracy": 0.997284719230313, + "accuracy": 0.9972787038668206, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -46689,7 +46689,7 @@ } }, { - "accuracy": 0.9973914201714491, + "accuracy": 0.9973846452408716, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -46738,7 +46738,7 @@ } }, { - "accuracy": 0.99783328900996, + "accuracy": 0.997828752782784, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -46784,7 +46784,7 @@ } }, { - "accuracy": 0.9992865890166477, + "accuracy": 0.999285649018068, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -46826,7 +46826,7 @@ ], "model.layers.24.self_attn": [ { - "accuracy": 0.9797939516996083, + "accuracy": 0.9793834121603715, "total_bits": 320757760, "q_proj": { "group_size": { @@ -46890,7 +46890,7 @@ } }, { - "accuracy": 0.980676448658893, + "accuracy": 0.9803548643463537, "total_bits": 329080832, "q_proj": { "group_size": { @@ -46954,7 +46954,7 @@ } }, { - "accuracy": 0.9834664307142559, + "accuracy": 0.9834568390720769, "total_bits": 336024576, "q_proj": { "group_size": { @@ -47018,7 +47018,7 @@ } }, { - "accuracy": 0.9865674784308985, + "accuracy": 0.9865337233794363, "total_bits": 401557504, "q_proj": { "group_size": { @@ -47082,7 +47082,7 @@ } }, { - "accuracy": 0.9879378629358191, + "accuracy": 0.9877654588536212, "total_bits": 475279360, "q_proj": { "group_size": { @@ -47146,7 +47146,7 @@ } }, { - "accuracy": 0.9888774948684793, + "accuracy": 0.9889072415075804, "total_bits": 475479040, "q_proj": { "group_size": { @@ -47210,7 +47210,7 @@ } }, { - "accuracy": 0.9898284563892766, + "accuracy": 0.9895546130443874, "total_bits": 609759232, "q_proj": { "group_size": { @@ -47262,7 +47262,7 @@ } }, { - "accuracy": 0.9909600019454956, + "accuracy": 0.990901212158956, "total_bits": 610024448, "q_proj": { "group_size": { @@ -47314,7 +47314,7 @@ } }, { - "accuracy": 0.9914158527788363, + "accuracy": 0.9917487157018561, "total_bits": 615020544, "q_proj": { "group_size": { @@ -47366,7 +47366,7 @@ } }, { - "accuracy": 0.9915950337522909, + "accuracy": 0.9919557963546953, "total_bits": 623951872, "q_proj": { "group_size": { @@ -47418,7 +47418,7 @@ } }, { - "accuracy": 0.993435404018352, + "accuracy": 0.993690830704413, "total_bits": 626473984, "q_proj": { "group_size": { @@ -47482,7 +47482,7 @@ } }, { - "accuracy": 0.9940588713476533, + "accuracy": 0.9940986225479528, "total_bits": 630355968, "q_proj": { "group_size": { @@ -47546,7 +47546,7 @@ } }, { - "accuracy": 0.9938976678409075, + "accuracy": 0.994076949593268, "total_bits": 637362176, "q_proj": { "group_size": { @@ -47607,7 +47607,7 @@ } }, { - "accuracy": 0.9944490731546753, + "accuracy": 0.994505404249618, "total_bits": 646823936, "q_proj": { "group_size": { @@ -47668,7 +47668,7 @@ } }, { - "accuracy": 0.9967852479925281, + "accuracy": 0.9968567970944079, "total_bits": 784740352, "q_proj": { "group_size": { @@ -47729,7 +47729,7 @@ } }, { - "accuracy": 0.9972363700600047, + "accuracy": 0.9973549570300078, "total_bits": 797818880, "q_proj": { "group_size": { @@ -47790,7 +47790,7 @@ } }, { - "accuracy": 0.9970792420208454, + "accuracy": 0.9971553918562437, "total_bits": 911749120, "q_proj": { "group_size": { @@ -47842,7 +47842,7 @@ } }, { - "accuracy": 0.9989346007963544, + "accuracy": 0.9989398055170712, "total_bits": 942718976, "q_proj": { "group_size": { @@ -47894,7 +47894,7 @@ } }, { - "accuracy": 0.9992452740767285, + "accuracy": 0.9992509412236119, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -47948,7 +47948,7 @@ ], "model.layers.24.mlp": [ { - "accuracy": 0.9591941833496094, + "accuracy": 0.9590147608204892, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -48000,7 +48000,7 @@ } }, { - "accuracy": 0.9602019661351254, + "accuracy": 0.9600452059193662, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -48052,7 +48052,7 @@ } }, { - "accuracy": 0.9659810474044398, + "accuracy": 0.9658340874471163, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -48101,7 +48101,7 @@ } }, { - "accuracy": 0.9675735015618174, + "accuracy": 0.967435011738225, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -48150,7 +48150,7 @@ } }, { - "accuracy": 0.9796106266347986, + "accuracy": 0.9795238469776354, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -48202,7 +48202,7 @@ } }, { - "accuracy": 0.9812336783660086, + "accuracy": 0.9811574252028215, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -48254,7 +48254,7 @@ } }, { - "accuracy": 0.9835864336867082, + "accuracy": 0.9835221014524761, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -48303,7 +48303,7 @@ } }, { - "accuracy": 0.9896470614169773, + "accuracy": 0.9895949857799631, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -48346,7 +48346,7 @@ } }, { - "accuracy": 0.990495251197564, + "accuracy": 0.9904506026129973, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -48389,7 +48389,7 @@ } }, { - "accuracy": 0.9896427361588729, + "accuracy": 0.989595005386754, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -48441,7 +48441,7 @@ } }, { - "accuracy": 0.9909129087862215, + "accuracy": 0.9908719525525445, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -48493,7 +48493,7 @@ } }, { - "accuracy": 0.9947272033283585, + "accuracy": 0.9947001616421499, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -48545,7 +48545,7 @@ } }, { - "accuracy": 0.9954756639505687, + "accuracy": 0.9954543278405541, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -48597,7 +48597,7 @@ } }, { - "accuracy": 0.9971974146993536, + "accuracy": 0.9971842136430112, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -48640,7 +48640,7 @@ } }, { - "accuracy": 0.997309625540909, + "accuracy": 0.9972958852978129, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -48689,7 +48689,7 @@ } }, { - "accuracy": 0.997762279290902, + "accuracy": 0.9977505234511275, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -48735,7 +48735,7 @@ } }, { - "accuracy": 0.9992569040899214, + "accuracy": 0.9992535285847751, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -48777,7 +48777,7 @@ ], "model.layers.25.self_attn": [ { - "accuracy": 0.973833684858523, + "accuracy": 0.9750682821399287, "total_bits": 320757760, "q_proj": { "group_size": { @@ -48841,7 +48841,7 @@ } }, { - "accuracy": 0.9751201397494266, + "accuracy": 0.9756383237085844, "total_bits": 329080832, "q_proj": { "group_size": { @@ -48905,7 +48905,7 @@ } }, { - "accuracy": 0.9769765342536726, + "accuracy": 0.9775036149903348, "total_bits": 336024576, "q_proj": { "group_size": { @@ -48969,7 +48969,7 @@ } }, { - "accuracy": 0.9794703361235166, + "accuracy": 0.9800233276266801, "total_bits": 401557504, "q_proj": { "group_size": { @@ -49033,7 +49033,7 @@ } }, { - "accuracy": 0.982063585206082, + "accuracy": 0.9823900492567765, "total_bits": 475279360, "q_proj": { "group_size": { @@ -49097,7 +49097,7 @@ } }, { - "accuracy": 0.9826588175798717, + "accuracy": 0.9823347486947712, "total_bits": 475479040, "q_proj": { "group_size": { @@ -49161,7 +49161,7 @@ } }, { - "accuracy": 0.9878465846965188, + "accuracy": 0.9882386969892603, "total_bits": 609759232, "q_proj": { "group_size": { @@ -49213,7 +49213,7 @@ } }, { - "accuracy": 0.9883394084478679, + "accuracy": 0.988652196369673, "total_bits": 610024448, "q_proj": { "group_size": { @@ -49265,7 +49265,7 @@ } }, { - "accuracy": 0.9891583895996997, + "accuracy": 0.9895936282057511, "total_bits": 615020544, "q_proj": { "group_size": { @@ -49317,7 +49317,7 @@ } }, { - "accuracy": 0.9892434038613972, + "accuracy": 0.9897250940925196, "total_bits": 623951872, "q_proj": { "group_size": { @@ -49369,7 +49369,7 @@ } }, { - "accuracy": 0.9884542595399054, + "accuracy": 0.989228041548478, "total_bits": 626473984, "q_proj": { "group_size": { @@ -49433,7 +49433,7 @@ } }, { - "accuracy": 0.9896676320778696, + "accuracy": 0.9897393772476598, "total_bits": 630355968, "q_proj": { "group_size": { @@ -49497,7 +49497,7 @@ } }, { - "accuracy": 0.9928039351576253, + "accuracy": 0.9931471637989345, "total_bits": 637362176, "q_proj": { "group_size": { @@ -49558,7 +49558,7 @@ } }, { - "accuracy": 0.9935711245787772, + "accuracy": 0.9936071120594677, "total_bits": 646823936, "q_proj": { "group_size": { @@ -49619,7 +49619,7 @@ } }, { - "accuracy": 0.9959637216831508, + "accuracy": 0.9960861817786568, "total_bits": 784740352, "q_proj": { "group_size": { @@ -49680,7 +49680,7 @@ } }, { - "accuracy": 0.996799597418622, + "accuracy": 0.9968054280861428, "total_bits": 797818880, "q_proj": { "group_size": { @@ -49741,7 +49741,7 @@ } }, { - "accuracy": 0.9963111422563854, + "accuracy": 0.9964329816008869, "total_bits": 911749120, "q_proj": { "group_size": { @@ -49793,7 +49793,7 @@ } }, { - "accuracy": 0.9988186126084704, + "accuracy": 0.998835412491309, "total_bits": 942718976, "q_proj": { "group_size": { @@ -49845,7 +49845,7 @@ } }, { - "accuracy": 0.9990798872160284, + "accuracy": 0.9990991728478357, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -49899,7 +49899,7 @@ ], "model.layers.25.mlp": [ { - "accuracy": 0.9575414343884117, + "accuracy": 0.9573492376427901, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -49951,7 +49951,7 @@ } }, { - "accuracy": 0.958645484949413, + "accuracy": 0.9584630163092362, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -50003,7 +50003,7 @@ } }, { - "accuracy": 0.9645760592661405, + "accuracy": 0.9644158702147634, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -50052,7 +50052,7 @@ } }, { - "accuracy": 0.9662118648227892, + "accuracy": 0.9660590040056329, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -50101,7 +50101,7 @@ } }, { - "accuracy": 0.9788115950007188, + "accuracy": 0.9787109415782126, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -50153,7 +50153,7 @@ } }, { - "accuracy": 0.9805101971877249, + "accuracy": 0.9804178460648185, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -50205,7 +50205,7 @@ } }, { - "accuracy": 0.9829025205812956, + "accuracy": 0.9828245482946697, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -50254,7 +50254,7 @@ } }, { - "accuracy": 0.9892467032921942, + "accuracy": 0.9891913713593232, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -50297,7 +50297,7 @@ } }, { - "accuracy": 0.9901197780119745, + "accuracy": 0.9900673502369931, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -50340,7 +50340,7 @@ } }, { - "accuracy": 0.9892367822559256, + "accuracy": 0.9891837231422725, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -50392,7 +50392,7 @@ } }, { - "accuracy": 0.9905703655983272, + "accuracy": 0.990523091272304, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -50444,7 +50444,7 @@ } }, { - "accuracy": 0.9945201905150163, + "accuracy": 0.9944915073482614, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -50496,7 +50496,7 @@ } }, { - "accuracy": 0.9953027979323739, + "accuracy": 0.995279032148813, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -50548,7 +50548,7 @@ } }, { - "accuracy": 0.9970777813149126, + "accuracy": 0.9970624648818844, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -50591,7 +50591,7 @@ } }, { - "accuracy": 0.9972012152797297, + "accuracy": 0.9971867260572157, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -50640,7 +50640,7 @@ } }, { - "accuracy": 0.9976545152695555, + "accuracy": 0.9976424272907408, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -50686,7 +50686,7 @@ } }, { - "accuracy": 0.9992254464641998, + "accuracy": 0.9992208435739341, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -50728,7 +50728,7 @@ ], "model.layers.26.self_attn": [ { - "accuracy": 0.971840478871998, + "accuracy": 0.9722584015444705, "total_bits": 320757760, "q_proj": { "group_size": { @@ -50792,7 +50792,7 @@ } }, { - "accuracy": 0.9730051881388614, + "accuracy": 0.9733615047053287, "total_bits": 329080832, "q_proj": { "group_size": { @@ -50856,7 +50856,7 @@ } }, { - "accuracy": 0.9768337419158534, + "accuracy": 0.9767298996448517, "total_bits": 336024576, "q_proj": { "group_size": { @@ -50920,7 +50920,7 @@ } }, { - "accuracy": 0.9799825125618985, + "accuracy": 0.9798255500040556, "total_bits": 401557504, "q_proj": { "group_size": { @@ -50984,7 +50984,7 @@ } }, { - "accuracy": 0.9830986684874484, + "accuracy": 0.9826251487982901, "total_bits": 475279360, "q_proj": { "group_size": { @@ -51048,7 +51048,7 @@ } }, { - "accuracy": 0.9840161596473894, + "accuracy": 0.983216006504862, "total_bits": 475479040, "q_proj": { "group_size": { @@ -51112,7 +51112,7 @@ } }, { - "accuracy": 0.9887476210531435, + "accuracy": 0.9881052837560051, "total_bits": 609759232, "q_proj": { "group_size": { @@ -51164,7 +51164,7 @@ } }, { - "accuracy": 0.9894581931202036, + "accuracy": 0.9891905635595322, "total_bits": 610024448, "q_proj": { "group_size": { @@ -51216,7 +51216,7 @@ } }, { - "accuracy": 0.9901097197281686, + "accuracy": 0.9902101370849108, "total_bits": 615020544, "q_proj": { "group_size": { @@ -51268,7 +51268,7 @@ } }, { - "accuracy": 0.9902879054609098, + "accuracy": 0.9903950048120398, "total_bits": 623951872, "q_proj": { "group_size": { @@ -51320,7 +51320,7 @@ } }, { - "accuracy": 0.9899624750802392, + "accuracy": 0.9891203673262345, "total_bits": 626473984, "q_proj": { "group_size": { @@ -51384,7 +51384,7 @@ } }, { - "accuracy": 0.9902578019782117, + "accuracy": 0.9905381445821962, "total_bits": 630355968, "q_proj": { "group_size": { @@ -51448,7 +51448,7 @@ } }, { - "accuracy": 0.9934997315469541, + "accuracy": 0.9932131461407009, "total_bits": 637362176, "q_proj": { "group_size": { @@ -51509,7 +51509,7 @@ } }, { - "accuracy": 0.9937988460848206, + "accuracy": 0.9938407489343694, "total_bits": 646823936, "q_proj": { "group_size": { @@ -51570,7 +51570,7 @@ } }, { - "accuracy": 0.9963542497471759, + "accuracy": 0.9964524123229479, "total_bits": 784740352, "q_proj": { "group_size": { @@ -51631,7 +51631,7 @@ } }, { - "accuracy": 0.996951068702497, + "accuracy": 0.9969074004574826, "total_bits": 797818880, "q_proj": { "group_size": { @@ -51692,7 +51692,7 @@ } }, { - "accuracy": 0.9967846203791467, + "accuracy": 0.9969329653601897, "total_bits": 911749120, "q_proj": { "group_size": { @@ -51744,7 +51744,7 @@ } }, { - "accuracy": 0.9988321958011702, + "accuracy": 0.998837393855578, "total_bits": 942718976, "q_proj": { "group_size": { @@ -51796,7 +51796,7 @@ } }, { - "accuracy": 0.9991387554297322, + "accuracy": 0.999173362984469, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -51850,7 +51850,7 @@ ], "model.layers.26.mlp": [ { - "accuracy": 0.9568879510227003, + "accuracy": 0.9567055545355144, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -51902,7 +51902,7 @@ } }, { - "accuracy": 0.95804552655471, + "accuracy": 0.9578708190667002, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -51954,7 +51954,7 @@ } }, { - "accuracy": 0.9639791877646195, + "accuracy": 0.9638244101875707, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -52003,7 +52003,7 @@ } }, { - "accuracy": 0.9656371097815664, + "accuracy": 0.9654895004473234, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -52052,7 +52052,7 @@ } }, { - "accuracy": 0.978565807405271, + "accuracy": 0.9784697106010035, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -52104,7 +52104,7 @@ } }, { - "accuracy": 0.9802216699248866, + "accuracy": 0.9801376399240995, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -52156,7 +52156,7 @@ } }, { - "accuracy": 0.9826228336284035, + "accuracy": 0.9825480250935805, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -52205,7 +52205,7 @@ } }, { - "accuracy": 0.9890850131448946, + "accuracy": 0.9890343891946893, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -52248,7 +52248,7 @@ } }, { - "accuracy": 0.9899458900878304, + "accuracy": 0.989900814075219, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -52291,7 +52291,7 @@ } }, { - "accuracy": 0.9891297879971956, + "accuracy": 0.9890788495540619, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -52343,7 +52343,7 @@ } }, { - "accuracy": 0.9904278618724722, + "accuracy": 0.9903867558429116, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -52395,7 +52395,7 @@ } }, { - "accuracy": 0.9944740792638377, + "accuracy": 0.994448793561835, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -52447,7 +52447,7 @@ } }, { - "accuracy": 0.9952355348750165, + "accuracy": 0.99521485951386, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -52499,7 +52499,7 @@ } }, { - "accuracy": 0.9970506072828644, + "accuracy": 0.9970367664569303, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -52542,7 +52542,7 @@ } }, { - "accuracy": 0.9971839442457023, + "accuracy": 0.9971706873100055, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -52591,7 +52591,7 @@ } }, { - "accuracy": 0.9976464117828169, + "accuracy": 0.9976354580568639, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -52637,7 +52637,7 @@ } }, { - "accuracy": 0.9992262085311507, + "accuracy": 0.9992230457596873, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -52679,7 +52679,7 @@ ], "model.layers.27.self_attn": [ { - "accuracy": 0.9690439512855128, + "accuracy": 0.9688237591793663, "total_bits": 320757760, "q_proj": { "group_size": { @@ -52743,7 +52743,7 @@ } }, { - "accuracy": 0.9718883570871855, + "accuracy": 0.9725917326776605, "total_bits": 329080832, "q_proj": { "group_size": { @@ -52807,7 +52807,7 @@ } }, { - "accuracy": 0.9761815949490196, + "accuracy": 0.9761837814983568, "total_bits": 336024576, "q_proj": { "group_size": { @@ -52871,7 +52871,7 @@ } }, { - "accuracy": 0.9818104678078702, + "accuracy": 0.9817658098120439, "total_bits": 401557504, "q_proj": { "group_size": { @@ -52935,7 +52935,7 @@ } }, { - "accuracy": 0.9849840355546851, + "accuracy": 0.9848748241600237, "total_bits": 475279360, "q_proj": { "group_size": { @@ -52999,7 +52999,7 @@ } }, { - "accuracy": 0.9853851403060713, + "accuracy": 0.9853376259929255, "total_bits": 475479040, "q_proj": { "group_size": { @@ -53063,7 +53063,7 @@ } }, { - "accuracy": 0.9907040839132509, + "accuracy": 0.9906250387430191, "total_bits": 609759232, "q_proj": { "group_size": { @@ -53115,7 +53115,7 @@ } }, { - "accuracy": 0.9911328263972935, + "accuracy": 0.9910773901562941, "total_bits": 610024448, "q_proj": { "group_size": { @@ -53167,7 +53167,7 @@ } }, { - "accuracy": 0.9917174130678177, + "accuracy": 0.9916931539773941, "total_bits": 615020544, "q_proj": { "group_size": { @@ -53219,7 +53219,7 @@ } }, { - "accuracy": 0.9920583500673896, + "accuracy": 0.99200281814525, "total_bits": 623951872, "q_proj": { "group_size": { @@ -53271,7 +53271,7 @@ } }, { - "accuracy": 0.992110694709577, + "accuracy": 0.9923827726590005, "total_bits": 626473984, "q_proj": { "group_size": { @@ -53335,7 +53335,7 @@ } }, { - "accuracy": 0.9928198605775833, + "accuracy": 0.9927425078655544, "total_bits": 630355968, "q_proj": { "group_size": { @@ -53399,7 +53399,7 @@ } }, { - "accuracy": 0.9937088066025784, + "accuracy": 0.9938014118294967, "total_bits": 637362176, "q_proj": { "group_size": { @@ -53460,7 +53460,7 @@ } }, { - "accuracy": 0.994227225451093, + "accuracy": 0.994293134071325, "total_bits": 646823936, "q_proj": { "group_size": { @@ -53521,7 +53521,7 @@ } }, { - "accuracy": 0.9965697013233837, + "accuracy": 0.9966407559419933, "total_bits": 784740352, "q_proj": { "group_size": { @@ -53582,7 +53582,7 @@ } }, { - "accuracy": 0.9971210756584218, + "accuracy": 0.9971467627114371, "total_bits": 797818880, "q_proj": { "group_size": { @@ -53643,7 +53643,7 @@ } }, { - "accuracy": 0.9974834742514711, + "accuracy": 0.9975773921530497, "total_bits": 911749120, "q_proj": { "group_size": { @@ -53695,7 +53695,7 @@ } }, { - "accuracy": 0.9985932143110978, + "accuracy": 0.9985954443875112, "total_bits": 942718976, "q_proj": { "group_size": { @@ -53747,7 +53747,7 @@ } }, { - "accuracy": 0.9993608279741908, + "accuracy": 0.9993482573251975, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -53801,7 +53801,7 @@ ], "model.layers.27.mlp": [ { - "accuracy": 0.9554918031943471, + "accuracy": 0.9553408654112565, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -53853,7 +53853,7 @@ } }, { - "accuracy": 0.9566749961752641, + "accuracy": 0.9565312485945852, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -53905,7 +53905,7 @@ } }, { - "accuracy": 0.9631398790761044, + "accuracy": 0.9630074061845478, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -53954,7 +53954,7 @@ } }, { - "accuracy": 0.9649206494030199, + "accuracy": 0.9647959784457558, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -54003,7 +54003,7 @@ } }, { - "accuracy": 0.977913814155679, + "accuracy": 0.9778365681045934, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -54055,7 +54055,7 @@ } }, { - "accuracy": 0.979622693438279, + "accuracy": 0.9795510188529366, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -54107,7 +54107,7 @@ } }, { - "accuracy": 0.9822231625255785, + "accuracy": 0.9821623689249942, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -54156,7 +54156,7 @@ } }, { - "accuracy": 0.9887432691298033, + "accuracy": 0.9887051182357889, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -54199,7 +54199,7 @@ } }, { - "accuracy": 0.9896474770809475, + "accuracy": 0.9896135675279718, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -54242,7 +54242,7 @@ } }, { - "accuracy": 0.9887958694445459, + "accuracy": 0.988759770989418, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -54294,7 +54294,7 @@ } }, { - "accuracy": 0.9901427650137952, + "accuracy": 0.990110056180703, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -54346,7 +54346,7 @@ } }, { - "accuracy": 0.9943045514978861, + "accuracy": 0.9942850199969191, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -54398,7 +54398,7 @@ } }, { - "accuracy": 0.9950922707978048, + "accuracy": 0.9950752681807468, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -54450,7 +54450,7 @@ } }, { - "accuracy": 0.9969583575270677, + "accuracy": 0.9969481151354941, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -54493,7 +54493,7 @@ } }, { - "accuracy": 0.9970971043956908, + "accuracy": 0.9970875603980139, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -54542,7 +54542,7 @@ } }, { - "accuracy": 0.9976053669264442, + "accuracy": 0.9975968038565234, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -54588,7 +54588,7 @@ } }, { - "accuracy": 0.9991985839350443, + "accuracy": 0.9991967422691616, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -54630,7 +54630,7 @@ ], "model.layers.28.self_attn": [ { - "accuracy": 0.9707333978853727, + "accuracy": 0.9712290983451041, "total_bits": 320757760, "q_proj": { "group_size": { @@ -54694,7 +54694,7 @@ } }, { - "accuracy": 0.9715769541890997, + "accuracy": 0.9718469287219801, "total_bits": 329080832, "q_proj": { "group_size": { @@ -54758,7 +54758,7 @@ } }, { - "accuracy": 0.976103663444519, + "accuracy": 0.976164427242781, "total_bits": 336024576, "q_proj": { "group_size": { @@ -54822,7 +54822,7 @@ } }, { - "accuracy": 0.9811423907154485, + "accuracy": 0.9813129305839539, "total_bits": 401557504, "q_proj": { "group_size": { @@ -54886,7 +54886,7 @@ } }, { - "accuracy": 0.9847845497884249, + "accuracy": 0.9848552612881911, "total_bits": 475279360, "q_proj": { "group_size": { @@ -54950,7 +54950,7 @@ } }, { - "accuracy": 0.984946702655993, + "accuracy": 0.9850893491192868, "total_bits": 475479040, "q_proj": { "group_size": { @@ -55014,7 +55014,7 @@ } }, { - "accuracy": 0.9887021489833531, + "accuracy": 0.988910544075464, "total_bits": 609759232, "q_proj": { "group_size": { @@ -55066,7 +55066,7 @@ } }, { - "accuracy": 0.9889918079501704, + "accuracy": 0.9891129700761092, "total_bits": 610024448, "q_proj": { "group_size": { @@ -55118,7 +55118,7 @@ } }, { - "accuracy": 0.9898172883610976, + "accuracy": 0.989629339230688, "total_bits": 615020544, "q_proj": { "group_size": { @@ -55170,7 +55170,7 @@ } }, { - "accuracy": 0.9900349480541129, + "accuracy": 0.9898862242698669, "total_bits": 623951872, "q_proj": { "group_size": { @@ -55222,7 +55222,7 @@ } }, { - "accuracy": 0.9920921843302878, + "accuracy": 0.9921314857508007, "total_bits": 626473984, "q_proj": { "group_size": { @@ -55286,7 +55286,7 @@ } }, { - "accuracy": 0.992621837477935, + "accuracy": 0.9926312448162782, "total_bits": 630355968, "q_proj": { "group_size": { @@ -55350,7 +55350,7 @@ } }, { - "accuracy": 0.9926039043225741, + "accuracy": 0.9926711422832388, "total_bits": 637362176, "q_proj": { "group_size": { @@ -55411,7 +55411,7 @@ } }, { - "accuracy": 0.9932573563174197, + "accuracy": 0.9932566308661511, "total_bits": 646823936, "q_proj": { "group_size": { @@ -55472,7 +55472,7 @@ } }, { - "accuracy": 0.9958718277906117, + "accuracy": 0.9959061655559038, "total_bits": 784740352, "q_proj": { "group_size": { @@ -55533,7 +55533,7 @@ } }, { - "accuracy": 0.9965228912861723, + "accuracy": 0.996550758418284, "total_bits": 797818880, "q_proj": { "group_size": { @@ -55594,7 +55594,7 @@ } }, { - "accuracy": 0.9966168176186713, + "accuracy": 0.9966768896893451, "total_bits": 911749120, "q_proj": { "group_size": { @@ -55646,7 +55646,7 @@ } }, { - "accuracy": 0.9984893843923744, + "accuracy": 0.9984707285306955, "total_bits": 942718976, "q_proj": { "group_size": { @@ -55698,7 +55698,7 @@ } }, { - "accuracy": 0.9990549455151746, + "accuracy": 0.9990701816583935, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -55752,7 +55752,7 @@ ], "model.layers.28.mlp": [ { - "accuracy": 0.9539401562590348, + "accuracy": 0.9537898490303441, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -55804,7 +55804,7 @@ } }, { - "accuracy": 0.9551617942358318, + "accuracy": 0.9550186991691589, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -55856,7 +55856,7 @@ } }, { - "accuracy": 0.961907706762615, + "accuracy": 0.9617808649414464, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -55905,7 +55905,7 @@ } }, { - "accuracy": 0.9637738434891951, + "accuracy": 0.9636508351878116, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -55954,7 +55954,7 @@ } }, { - "accuracy": 0.9770849817677548, + "accuracy": 0.9770082501988662, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -56006,7 +56006,7 @@ } }, { - "accuracy": 0.9788798376133567, + "accuracy": 0.9788016535733876, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -56058,7 +56058,7 @@ } }, { - "accuracy": 0.9815997575458727, + "accuracy": 0.9815292217229542, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -56107,7 +56107,7 @@ } }, { - "accuracy": 0.9882910126133969, + "accuracy": 0.9882488046821795, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -56150,7 +56150,7 @@ } }, { - "accuracy": 0.9892301088885257, + "accuracy": 0.9891873731424934, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -56193,7 +56193,7 @@ } }, { - "accuracy": 0.9883437627240231, + "accuracy": 0.9883021422122654, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -56245,7 +56245,7 @@ } }, { - "accuracy": 0.9897430445018568, + "accuracy": 0.9897070801571796, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -56297,7 +56297,7 @@ } }, { - "accuracy": 0.9940563593255846, + "accuracy": 0.9940366768523267, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -56349,7 +56349,7 @@ } }, { - "accuracy": 0.9948868496637595, + "accuracy": 0.9948674538417867, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -56401,7 +56401,7 @@ } }, { - "accuracy": 0.9968196506562986, + "accuracy": 0.9968097133463935, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -56444,7 +56444,7 @@ } }, { - "accuracy": 0.9969656263527117, + "accuracy": 0.9969548175209447, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -56493,7 +56493,7 @@ } }, { - "accuracy": 0.9974974099742738, + "accuracy": 0.9974886998534203, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -56539,7 +56539,7 @@ } }, { - "accuracy": 0.9991560997931581, + "accuracy": 0.9991529548638746, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -56581,7 +56581,7 @@ ], "model.layers.29.self_attn": [ { - "accuracy": 0.9693141830594916, + "accuracy": 0.9698963855442247, "total_bits": 320757760, "q_proj": { "group_size": { @@ -56645,7 +56645,7 @@ } }, { - "accuracy": 0.969690137787869, + "accuracy": 0.9710368576802706, "total_bits": 329080832, "q_proj": { "group_size": { @@ -56709,7 +56709,7 @@ } }, { - "accuracy": 0.9740292633834639, + "accuracy": 0.9743174094902841, "total_bits": 336024576, "q_proj": { "group_size": { @@ -56773,7 +56773,7 @@ } }, { - "accuracy": 0.979405972518419, + "accuracy": 0.9796246650971865, "total_bits": 401557504, "q_proj": { "group_size": { @@ -56837,7 +56837,7 @@ } }, { - "accuracy": 0.9838082711947592, + "accuracy": 0.9840611329204157, "total_bits": 475279360, "q_proj": { "group_size": { @@ -56901,7 +56901,7 @@ } }, { - "accuracy": 0.984044318136416, + "accuracy": 0.9843920939847043, "total_bits": 475479040, "q_proj": { "group_size": { @@ -56965,7 +56965,7 @@ } }, { - "accuracy": 0.9879599182229293, + "accuracy": 0.9881209268381721, "total_bits": 609759232, "q_proj": { "group_size": { @@ -57017,7 +57017,7 @@ } }, { - "accuracy": 0.9881891027877205, + "accuracy": 0.9890243826728118, "total_bits": 610024448, "q_proj": { "group_size": { @@ -57069,7 +57069,7 @@ } }, { - "accuracy": 0.9880705052300504, + "accuracy": 0.9898613205081538, "total_bits": 615020544, "q_proj": { "group_size": { @@ -57121,7 +57121,7 @@ } }, { - "accuracy": 0.9884021823343477, + "accuracy": 0.9901269964481655, "total_bits": 623951872, "q_proj": { "group_size": { @@ -57173,7 +57173,7 @@ } }, { - "accuracy": 0.9917081061162447, + "accuracy": 0.9919500327423999, "total_bits": 626473984, "q_proj": { "group_size": { @@ -57237,7 +57237,7 @@ } }, { - "accuracy": 0.9921486307131616, + "accuracy": 0.9923906922340393, "total_bits": 630355968, "q_proj": { "group_size": { @@ -57301,7 +57301,7 @@ } }, { - "accuracy": 0.9923033518226523, + "accuracy": 0.9925501738723955, "total_bits": 637362176, "q_proj": { "group_size": { @@ -57362,7 +57362,7 @@ } }, { - "accuracy": 0.9927989448371687, + "accuracy": 0.9930783690590608, "total_bits": 646823936, "q_proj": { "group_size": { @@ -57423,7 +57423,7 @@ } }, { - "accuracy": 0.995672627499229, + "accuracy": 0.9957888706734306, "total_bits": 784740352, "q_proj": { "group_size": { @@ -57484,7 +57484,7 @@ } }, { - "accuracy": 0.9963778222077772, + "accuracy": 0.9964651500708178, "total_bits": 797818880, "q_proj": { "group_size": { @@ -57545,7 +57545,7 @@ } }, { - "accuracy": 0.996548562457687, + "accuracy": 0.9966301890580278, "total_bits": 911749120, "q_proj": { "group_size": { @@ -57597,7 +57597,7 @@ } }, { - "accuracy": 0.9983793824518982, + "accuracy": 0.9984079151365318, "total_bits": 942718976, "q_proj": { "group_size": { @@ -57649,7 +57649,7 @@ } }, { - "accuracy": 0.9991246377559084, + "accuracy": 0.9991242695403727, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -57703,7 +57703,7 @@ ], "model.layers.29.mlp": [ { - "accuracy": 0.9530472065273085, + "accuracy": 0.9529313476462113, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -57755,7 +57755,7 @@ } }, { - "accuracy": 0.9543056080215856, + "accuracy": 0.9542128757426613, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -57807,7 +57807,7 @@ } }, { - "accuracy": 0.9611161476687381, + "accuracy": 0.961025783890172, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -57856,7 +57856,7 @@ } }, { - "accuracy": 0.9630119612342433, + "accuracy": 0.9629301899357846, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -57905,7 +57905,7 @@ } }, { - "accuracy": 0.97669896326567, + "accuracy": 0.9766332789471275, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -57957,7 +57957,7 @@ } }, { - "accuracy": 0.9784950237525137, + "accuracy": 0.9784395537878338, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -58009,7 +58009,7 @@ } }, { - "accuracy": 0.9812284394314414, + "accuracy": 0.9811779244949943, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -58058,7 +58058,7 @@ } }, { - "accuracy": 0.9880849397496173, + "accuracy": 0.9880498145755968, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -58101,7 +58101,7 @@ } }, { - "accuracy": 0.9890205993464118, + "accuracy": 0.9889906887945376, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -58144,7 +58144,7 @@ } }, { - "accuracy": 0.9881579648507269, + "accuracy": 0.9881261979278765, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -58196,7 +58196,7 @@ } }, { - "accuracy": 0.9895707266895395, + "accuracy": 0.9895431626784174, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -58248,7 +58248,7 @@ } }, { - "accuracy": 0.9939682762089529, + "accuracy": 0.9939534334759963, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -58300,7 +58300,7 @@ } }, { - "accuracy": 0.994801463265168, + "accuracy": 0.994787364021728, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -58352,7 +58352,7 @@ } }, { - "accuracy": 0.9967632181942463, + "accuracy": 0.9967540084923568, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -58395,7 +58395,7 @@ } }, { - "accuracy": 0.996914646343181, + "accuracy": 0.9969059491628095, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -58444,7 +58444,7 @@ } }, { - "accuracy": 0.9974409735908634, + "accuracy": 0.9974329624521104, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -58490,7 +58490,7 @@ } }, { - "accuracy": 0.9991108630048601, + "accuracy": 0.999108017177174, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -58532,7 +58532,7 @@ ], "model.layers.30.self_attn": [ { - "accuracy": 0.9700600661729512, + "accuracy": 0.9698140244734915, "total_bits": 320757760, "q_proj": { "group_size": { @@ -58596,7 +58596,7 @@ } }, { - "accuracy": 0.9709761958373221, + "accuracy": 0.9708121199356882, "total_bits": 329080832, "q_proj": { "group_size": { @@ -58660,7 +58660,7 @@ } }, { - "accuracy": 0.9749453271690168, + "accuracy": 0.9749842505705985, "total_bits": 336024576, "q_proj": { "group_size": { @@ -58724,7 +58724,7 @@ } }, { - "accuracy": 0.981272523340426, + "accuracy": 0.9813142450232255, "total_bits": 401557504, "q_proj": { "group_size": { @@ -58788,7 +58788,7 @@ } }, { - "accuracy": 0.9842489000998045, + "accuracy": 0.9842463496484255, "total_bits": 475279360, "q_proj": { "group_size": { @@ -58852,7 +58852,7 @@ } }, { - "accuracy": 0.9847229750532853, + "accuracy": 0.9846654659823367, "total_bits": 475479040, "q_proj": { "group_size": { @@ -58916,7 +58916,7 @@ } }, { - "accuracy": 0.9888382757964888, + "accuracy": 0.9889285658535204, "total_bits": 609759232, "q_proj": { "group_size": { @@ -58968,7 +58968,7 @@ } }, { - "accuracy": 0.9891618301993922, + "accuracy": 0.9887766587106805, "total_bits": 610024448, "q_proj": { "group_size": { @@ -59020,7 +59020,7 @@ } }, { - "accuracy": 0.9901792093327171, + "accuracy": 0.9901095401299628, "total_bits": 615020544, "q_proj": { "group_size": { @@ -59072,7 +59072,7 @@ } }, { - "accuracy": 0.9902943388411873, + "accuracy": 0.990447378472278, "total_bits": 623951872, "q_proj": { "group_size": { @@ -59124,7 +59124,7 @@ } }, { - "accuracy": 0.9919377212461672, + "accuracy": 0.9919624548209341, "total_bits": 626473984, "q_proj": { "group_size": { @@ -59188,7 +59188,7 @@ } }, { - "accuracy": 0.9924476154540715, + "accuracy": 0.9923822205317648, "total_bits": 630355968, "q_proj": { "group_size": { @@ -59252,7 +59252,7 @@ } }, { - "accuracy": 0.9926562803356271, + "accuracy": 0.9927032323260057, "total_bits": 637362176, "q_proj": { "group_size": { @@ -59313,7 +59313,7 @@ } }, { - "accuracy": 0.9932450699178796, + "accuracy": 0.9931264637332213, "total_bits": 646823936, "q_proj": { "group_size": { @@ -59374,7 +59374,7 @@ } }, { - "accuracy": 0.9958892754818264, + "accuracy": 0.9959228952464304, "total_bits": 784740352, "q_proj": { "group_size": { @@ -59435,7 +59435,7 @@ } }, { - "accuracy": 0.9965796627496418, + "accuracy": 0.9966222686987174, "total_bits": 797818880, "q_proj": { "group_size": { @@ -59496,7 +59496,7 @@ } }, { - "accuracy": 0.9967494993225524, + "accuracy": 0.9968116128523099, "total_bits": 911749120, "q_proj": { "group_size": { @@ -59548,7 +59548,7 @@ } }, { - "accuracy": 0.9984196089208126, + "accuracy": 0.9984195343169727, "total_bits": 942718976, "q_proj": { "group_size": { @@ -59600,7 +59600,7 @@ } }, { - "accuracy": 0.9991471186084183, + "accuracy": 0.9991410521692351, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -59654,7 +59654,7 @@ ], "model.layers.30.mlp": [ { - "accuracy": 0.9540594878949618, + "accuracy": 0.9539911245044909, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -59706,7 +59706,7 @@ } }, { - "accuracy": 0.9552903834142183, + "accuracy": 0.9552304870203921, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -59758,7 +59758,7 @@ } }, { - "accuracy": 0.9620092981740048, + "accuracy": 0.9619480904779936, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -59807,7 +59807,7 @@ } }, { - "accuracy": 0.9639124117399517, + "accuracy": 0.9638502691921434, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -59856,7 +59856,7 @@ } }, { - "accuracy": 0.9771716484898015, + "accuracy": 0.9771366887970975, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -59908,7 +59908,7 @@ } }, { - "accuracy": 0.9789336260996366, + "accuracy": 0.9788977346922222, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -59960,7 +59960,7 @@ } }, { - "accuracy": 0.9816576948291377, + "accuracy": 0.9816227570960396, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -60009,7 +60009,7 @@ } }, { - "accuracy": 0.9883000905576506, + "accuracy": 0.9882799841855702, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -60052,7 +60052,7 @@ } }, { - "accuracy": 0.989231664883463, + "accuracy": 0.9892141701359498, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -60095,7 +60095,7 @@ } }, { - "accuracy": 0.9883966179270494, + "accuracy": 0.9883756237594705, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -60147,7 +60147,7 @@ } }, { - "accuracy": 0.9897739236292086, + "accuracy": 0.9897538149043134, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -60199,7 +60199,7 @@ } }, { - "accuracy": 0.9940856820658633, + "accuracy": 0.9940737540784635, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -60251,7 +60251,7 @@ } }, { - "accuracy": 0.9949005559871071, + "accuracy": 0.9948922505504206, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -60303,7 +60303,7 @@ } }, { - "accuracy": 0.9968268345845374, + "accuracy": 0.9968220858197463, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -60346,7 +60346,7 @@ } }, { - "accuracy": 0.99697374395634, + "accuracy": 0.9969683665978281, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -60395,7 +60395,7 @@ } }, { - "accuracy": 0.9975162746482774, + "accuracy": 0.9975111727651796, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -60441,7 +60441,7 @@ } }, { - "accuracy": 0.999124084844401, + "accuracy": 0.9991225750234566, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -60483,7 +60483,7 @@ ], "model.layers.31.self_attn": [ { - "accuracy": 0.9562085992411563, + "accuracy": 0.9565619167528654, "total_bits": 320757760, "q_proj": { "group_size": { @@ -60547,7 +60547,7 @@ } }, { - "accuracy": 0.9588543490359658, + "accuracy": 0.9591552521053114, "total_bits": 329080832, "q_proj": { "group_size": { @@ -60611,7 +60611,7 @@ } }, { - "accuracy": 0.96373555848473, + "accuracy": 0.9641510812859786, "total_bits": 336024576, "q_proj": { "group_size": { @@ -60675,7 +60675,7 @@ } }, { - "accuracy": 0.9709683248871251, + "accuracy": 0.9714128155457346, "total_bits": 401557504, "q_proj": { "group_size": { @@ -60739,7 +60739,7 @@ } }, { - "accuracy": 0.9786312674221239, + "accuracy": 0.9786144949887928, "total_bits": 475279360, "q_proj": { "group_size": { @@ -60803,7 +60803,7 @@ } }, { - "accuracy": 0.978660512911646, + "accuracy": 0.9789146577057085, "total_bits": 475479040, "q_proj": { "group_size": { @@ -60867,7 +60867,7 @@ } }, { - "accuracy": 0.9879548290842458, + "accuracy": 0.9880103374782362, "total_bits": 609759232, "q_proj": { "group_size": { @@ -60919,7 +60919,7 @@ } }, { - "accuracy": 0.9882310174013439, + "accuracy": 0.9882392083343706, "total_bits": 610024448, "q_proj": { "group_size": { @@ -60971,7 +60971,7 @@ } }, { - "accuracy": 0.9890933601479781, + "accuracy": 0.9892781659176475, "total_bits": 615020544, "q_proj": { "group_size": { @@ -61023,7 +61023,7 @@ } }, { - "accuracy": 0.989502949149985, + "accuracy": 0.9897166333700481, "total_bits": 623951872, "q_proj": { "group_size": { @@ -61075,7 +61075,7 @@ } }, { - "accuracy": 0.9890668556878441, + "accuracy": 0.9890267778384058, "total_bits": 626473984, "q_proj": { "group_size": { @@ -61139,7 +61139,7 @@ } }, { - "accuracy": 0.9895921451480765, + "accuracy": 0.9899020281277204, "total_bits": 630355968, "q_proj": { "group_size": { @@ -61203,7 +61203,7 @@ } }, { - "accuracy": 0.9912972748279572, + "accuracy": 0.9914378382657704, "total_bits": 637362176, "q_proj": { "group_size": { @@ -61264,7 +61264,7 @@ } }, { - "accuracy": 0.9921817505045941, + "accuracy": 0.9921559314978751, "total_bits": 646823936, "q_proj": { "group_size": { @@ -61325,7 +61325,7 @@ } }, { - "accuracy": 0.9953180190763975, + "accuracy": 0.9953870463528132, "total_bits": 784740352, "q_proj": { "group_size": { @@ -61386,7 +61386,7 @@ } }, { - "accuracy": 0.9960450686906513, + "accuracy": 0.9960647805740959, "total_bits": 797818880, "q_proj": { "group_size": { @@ -61447,7 +61447,7 @@ } }, { - "accuracy": 0.9969052048890215, + "accuracy": 0.9969441678963209, "total_bits": 911749120, "q_proj": { "group_size": { @@ -61499,7 +61499,7 @@ } }, { - "accuracy": 0.9980479494521493, + "accuracy": 0.9980507614581209, "total_bits": 942718976, "q_proj": { "group_size": { @@ -61551,7 +61551,7 @@ } }, { - "accuracy": 0.9991761312672967, + "accuracy": 0.9991810649241272, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -61605,7 +61605,7 @@ ], "model.layers.31.mlp": [ { - "accuracy": 0.9531574562976235, + "accuracy": 0.9531018451640481, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -61657,7 +61657,7 @@ } }, { - "accuracy": 0.9544103678904081, + "accuracy": 0.9543687544370952, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -61709,7 +61709,7 @@ } }, { - "accuracy": 0.9613295511195534, + "accuracy": 0.9612825481515181, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -61758,7 +61758,7 @@ } }, { - "accuracy": 0.9632624952416671, + "accuracy": 0.9632150875894647, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -61807,7 +61807,7 @@ } }, { - "accuracy": 0.9767399210678903, + "accuracy": 0.9767074945725893, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -61859,7 +61859,7 @@ } }, { - "accuracy": 0.9785279129680834, + "accuracy": 0.9785031095931405, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -61911,7 +61911,7 @@ } }, { - "accuracy": 0.9813132380184374, + "accuracy": 0.9812909239216855, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -61960,7 +61960,7 @@ } }, { - "accuracy": 0.9881119728088379, + "accuracy": 0.9880948482375396, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -62003,7 +62003,7 @@ } }, { - "accuracy": 0.9890494754439906, + "accuracy": 0.989033612765764, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -62046,7 +62046,7 @@ } }, { - "accuracy": 0.988177555172067, + "accuracy": 0.9881612031083358, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -62098,7 +62098,7 @@ } }, { - "accuracy": 0.9895799473712319, + "accuracy": 0.9895675284297842, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -62150,7 +62150,7 @@ } }, { - "accuracy": 0.9939796904984274, + "accuracy": 0.99397015218672, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -62202,7 +62202,7 @@ } }, { - "accuracy": 0.9948067320020575, + "accuracy": 0.9947991437817875, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -62254,7 +62254,7 @@ } }, { - "accuracy": 0.9967825289227461, + "accuracy": 0.9967784907080626, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -62297,7 +62297,7 @@ } }, { - "accuracy": 0.996927289978454, + "accuracy": 0.9969206962146258, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -62346,7 +62346,7 @@ } }, { - "accuracy": 0.997468790333522, + "accuracy": 0.9974622657816661, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -62392,7 +62392,7 @@ } }, { - "accuracy": 0.9991494951475608, + "accuracy": 0.9991479683667421, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -62434,7 +62434,7 @@ ], "model.layers.32.self_attn": [ { - "accuracy": 0.9668000967879045, + "accuracy": 0.9673847399259868, "total_bits": 320757760, "q_proj": { "group_size": { @@ -62498,7 +62498,7 @@ } }, { - "accuracy": 0.9678211902317247, + "accuracy": 0.9678471464859812, "total_bits": 329080832, "q_proj": { "group_size": { @@ -62562,7 +62562,7 @@ } }, { - "accuracy": 0.9747328083766135, + "accuracy": 0.9733324646949768, "total_bits": 336024576, "q_proj": { "group_size": { @@ -62626,7 +62626,7 @@ } }, { - "accuracy": 0.9790494551784114, + "accuracy": 0.9772155614275682, "total_bits": 401557504, "q_proj": { "group_size": { @@ -62690,7 +62690,7 @@ } }, { - "accuracy": 0.9813494650941146, + "accuracy": 0.9800048784205788, "total_bits": 475279360, "q_proj": { "group_size": { @@ -62754,7 +62754,7 @@ } }, { - "accuracy": 0.9801986625320033, + "accuracy": 0.9818303004691475, "total_bits": 475479040, "q_proj": { "group_size": { @@ -62818,7 +62818,7 @@ } }, { - "accuracy": 0.989141320711688, + "accuracy": 0.989637199201082, "total_bits": 609759232, "q_proj": { "group_size": { @@ -62870,7 +62870,7 @@ } }, { - "accuracy": 0.9897620066216117, + "accuracy": 0.9896535520490847, "total_bits": 610024448, "q_proj": { "group_size": { @@ -62922,7 +62922,7 @@ } }, { - "accuracy": 0.9905922381501449, + "accuracy": 0.9906184783107356, "total_bits": 615020544, "q_proj": { "group_size": { @@ -62974,7 +62974,7 @@ } }, { - "accuracy": 0.9909453627310301, + "accuracy": 0.9908678163039056, "total_bits": 623951872, "q_proj": { "group_size": { @@ -63026,7 +63026,7 @@ } }, { - "accuracy": 0.9894712786925467, + "accuracy": 0.9899832731799075, "total_bits": 626473984, "q_proj": { "group_size": { @@ -63090,7 +63090,7 @@ } }, { - "accuracy": 0.990896264189168, + "accuracy": 0.990515640691707, "total_bits": 630355968, "q_proj": { "group_size": { @@ -63154,7 +63154,7 @@ } }, { - "accuracy": 0.9933349921515113, + "accuracy": 0.9933732191198751, "total_bits": 637362176, "q_proj": { "group_size": { @@ -63215,7 +63215,7 @@ } }, { - "accuracy": 0.9938801864260122, + "accuracy": 0.9939073665361655, "total_bits": 646823936, "q_proj": { "group_size": { @@ -63276,7 +63276,7 @@ } }, { - "accuracy": 0.9963800287560413, + "accuracy": 0.9964255576855258, "total_bits": 784740352, "q_proj": { "group_size": { @@ -63337,7 +63337,7 @@ } }, { - "accuracy": 0.9969699596496004, + "accuracy": 0.9969285081483816, "total_bits": 797818880, "q_proj": { "group_size": { @@ -63398,7 +63398,7 @@ } }, { - "accuracy": 0.9970364813741884, + "accuracy": 0.9970655558924926, "total_bits": 911749120, "q_proj": { "group_size": { @@ -63450,7 +63450,7 @@ } }, { - "accuracy": 0.9986543675002298, + "accuracy": 0.9986629478241268, "total_bits": 942718976, "q_proj": { "group_size": { @@ -63502,7 +63502,7 @@ } }, { - "accuracy": 0.9991932224580332, + "accuracy": 0.9992052175006584, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -63556,7 +63556,7 @@ ], "model.layers.32.mlp": [ { - "accuracy": 0.9531020867197137, + "accuracy": 0.9530720051966215, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -63608,7 +63608,7 @@ } }, { - "accuracy": 0.9543359154149106, + "accuracy": 0.9543035563669706, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -63660,7 +63660,7 @@ } }, { - "accuracy": 0.9611333545885588, + "accuracy": 0.9611073136329651, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -63709,7 +63709,7 @@ } }, { - "accuracy": 0.9630866176203677, + "accuracy": 0.963060670777371, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -63758,7 +63758,7 @@ } }, { - "accuracy": 0.9765793006671103, + "accuracy": 0.976559756617797, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -63810,7 +63810,7 @@ } }, { - "accuracy": 0.9784038898191953, + "accuracy": 0.9783844132172433, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -63862,7 +63862,7 @@ } }, { - "accuracy": 0.9812154534615969, + "accuracy": 0.9811975093264329, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -63911,7 +63911,7 @@ } }, { - "accuracy": 0.9879822464365708, + "accuracy": 0.9879709521406576, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -63954,7 +63954,7 @@ } }, { - "accuracy": 0.9889682735267439, + "accuracy": 0.9889578317341051, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -63997,7 +63997,7 @@ } }, { - "accuracy": 0.9880814717004174, + "accuracy": 0.9880708918759697, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -64049,7 +64049,7 @@ } }, { - "accuracy": 0.9895141556074745, + "accuracy": 0.9895049615910179, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -64101,7 +64101,7 @@ } }, { - "accuracy": 0.9939185278980356, + "accuracy": 0.9939149026023714, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -64153,7 +64153,7 @@ } }, { - "accuracy": 0.9947691222554759, + "accuracy": 0.9947631986517655, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -64205,7 +64205,7 @@ } }, { - "accuracy": 0.9967396567134481, + "accuracy": 0.9967368849013981, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -64248,7 +64248,7 @@ } }, { - "accuracy": 0.9968942552804947, + "accuracy": 0.9968916148339447, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -64297,7 +64297,7 @@ } }, { - "accuracy": 0.9974491570733095, + "accuracy": 0.9974459907726237, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -64343,7 +64343,7 @@ } }, { - "accuracy": 0.9991157141171003, + "accuracy": 0.9991139255856213, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -64385,7 +64385,7 @@ ], "model.layers.33.self_attn": [ { - "accuracy": 0.9608171676334581, + "accuracy": 0.9611800250254179, "total_bits": 320757760, "q_proj": { "group_size": { @@ -64449,7 +64449,7 @@ } }, { - "accuracy": 0.9631913021991128, + "accuracy": 0.9636226421908328, "total_bits": 329080832, "q_proj": { "group_size": { @@ -64513,7 +64513,7 @@ } }, { - "accuracy": 0.9677893362547222, + "accuracy": 0.9676858029867473, "total_bits": 336024576, "q_proj": { "group_size": { @@ -64577,7 +64577,7 @@ } }, { - "accuracy": 0.974450094135184, + "accuracy": 0.9742776218213534, "total_bits": 401557504, "q_proj": { "group_size": { @@ -64641,7 +64641,7 @@ } }, { - "accuracy": 0.9797495634932267, + "accuracy": 0.9803887009620667, "total_bits": 475279360, "q_proj": { "group_size": { @@ -64705,7 +64705,7 @@ } }, { - "accuracy": 0.9799389870543229, + "accuracy": 0.9804716361196417, "total_bits": 475479040, "q_proj": { "group_size": { @@ -64769,7 +64769,7 @@ } }, { - "accuracy": 0.9885730241474352, + "accuracy": 0.9890069436085852, "total_bits": 609759232, "q_proj": { "group_size": { @@ -64821,7 +64821,7 @@ } }, { - "accuracy": 0.9892245436969557, + "accuracy": 0.9890578514651248, "total_bits": 610024448, "q_proj": { "group_size": { @@ -64873,7 +64873,7 @@ } }, { - "accuracy": 0.9898122298090082, + "accuracy": 0.9896279079349417, "total_bits": 615020544, "q_proj": { "group_size": { @@ -64925,7 +64925,7 @@ } }, { - "accuracy": 0.9903432664118315, + "accuracy": 0.9901924047030901, "total_bits": 623951872, "q_proj": { "group_size": { @@ -64977,7 +64977,7 @@ } }, { - "accuracy": 0.9899086975737622, + "accuracy": 0.9891608859363356, "total_bits": 626473984, "q_proj": { "group_size": { @@ -65041,7 +65041,7 @@ } }, { - "accuracy": 0.9904844368758955, + "accuracy": 0.9906041724117178, "total_bits": 630355968, "q_proj": { "group_size": { @@ -65105,7 +65105,7 @@ } }, { - "accuracy": 0.9920745907645476, + "accuracy": 0.9921298348589948, "total_bits": 637362176, "q_proj": { "group_size": { @@ -65166,7 +65166,7 @@ } }, { - "accuracy": 0.9927323052757665, + "accuracy": 0.9927577344994796, "total_bits": 646823936, "q_proj": { "group_size": { @@ -65227,7 +65227,7 @@ } }, { - "accuracy": 0.9957279488444328, + "accuracy": 0.9957438664216745, "total_bits": 784740352, "q_proj": { "group_size": { @@ -65288,7 +65288,7 @@ } }, { - "accuracy": 0.9963407226298985, + "accuracy": 0.9963670949402609, "total_bits": 797818880, "q_proj": { "group_size": { @@ -65349,7 +65349,7 @@ } }, { - "accuracy": 0.9971870583923239, + "accuracy": 0.9971664104806749, "total_bits": 911749120, "q_proj": { "group_size": { @@ -65401,7 +65401,7 @@ } }, { - "accuracy": 0.9981663095715799, + "accuracy": 0.9981637438269038, "total_bits": 942718976, "q_proj": { "group_size": { @@ -65453,7 +65453,7 @@ } }, { - "accuracy": 0.9992276658549121, + "accuracy": 0.9992381260779343, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -65507,7 +65507,7 @@ ], "model.layers.33.mlp": [ { - "accuracy": 0.9514258598026476, + "accuracy": 0.9513958692550659, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -65559,7 +65559,7 @@ } }, { - "accuracy": 0.9526837342663815, + "accuracy": 0.9526619660226923, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -65611,7 +65611,7 @@ } }, { - "accuracy": 0.9593044613537035, + "accuracy": 0.959282282151674, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -65660,7 +65660,7 @@ } }, { - "accuracy": 0.9611747390345523, + "accuracy": 0.9611514838118302, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -65709,7 +65709,7 @@ } }, { - "accuracy": 0.9716558738758689, + "accuracy": 0.9716518740904959, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -65761,7 +65761,7 @@ } }, { - "accuracy": 0.9769588740248429, + "accuracy": 0.9769532743253206, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -65813,7 +65813,7 @@ } }, { - "accuracy": 0.9795619421883633, + "accuracy": 0.9795588850975037, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -65862,7 +65862,7 @@ } }, { - "accuracy": 0.98369027125208, + "accuracy": 0.9836896093268144, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -65905,7 +65905,7 @@ } }, { - "accuracy": 0.9868637135154322, + "accuracy": 0.9868643307372144, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -65948,7 +65948,7 @@ } }, { - "accuracy": 0.9813115047781091, + "accuracy": 0.9813182542198583, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -66000,7 +66000,7 @@ } }, { - "accuracy": 0.9873405075386951, + "accuracy": 0.9873425756630144, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -66052,7 +66052,7 @@ } }, { - "accuracy": 0.9862524929799532, + "accuracy": 0.9862638100197441, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -66104,7 +66104,7 @@ } }, { - "accuracy": 0.991991827362462, + "accuracy": 0.9919977070469606, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -66156,7 +66156,7 @@ } }, { - "accuracy": 0.9897498104133104, + "accuracy": 0.9897638543655998, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -66199,7 +66199,7 @@ } }, { - "accuracy": 0.9897988744472203, + "accuracy": 0.989812054916432, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -66248,7 +66248,7 @@ } }, { - "accuracy": 0.9899544990376422, + "accuracy": 0.989968406526666, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -66294,7 +66294,7 @@ } }, { - "accuracy": 0.9946170511998629, + "accuracy": 0.9946237951517105, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -66336,7 +66336,7 @@ ], "model.layers.34.self_attn": [ { - "accuracy": 0.9541143743615401, + "accuracy": 0.953845805243442, "total_bits": 320757760, "q_proj": { "group_size": { @@ -66400,7 +66400,7 @@ } }, { - "accuracy": 0.9605768856249357, + "accuracy": 0.9618892418710809, "total_bits": 329080832, "q_proj": { "group_size": { @@ -66464,7 +66464,7 @@ } }, { - "accuracy": 0.96766218699907, + "accuracy": 0.968051276708904, "total_bits": 336024576, "q_proj": { "group_size": { @@ -66528,7 +66528,7 @@ } }, { - "accuracy": 0.9738373709352393, + "accuracy": 0.9743172667528454, "total_bits": 401557504, "q_proj": { "group_size": { @@ -66592,7 +66592,7 @@ } }, { - "accuracy": 0.9800770721937481, + "accuracy": 0.9792975814718949, "total_bits": 475279360, "q_proj": { "group_size": { @@ -66656,7 +66656,7 @@ } }, { - "accuracy": 0.9791887590759679, + "accuracy": 0.9795984280736822, "total_bits": 475479040, "q_proj": { "group_size": { @@ -66720,7 +66720,7 @@ } }, { - "accuracy": 0.988222305711947, + "accuracy": 0.9880754328087756, "total_bits": 609759232, "q_proj": { "group_size": { @@ -66772,7 +66772,7 @@ } }, { - "accuracy": 0.988496123175872, + "accuracy": 0.9886315935536435, "total_bits": 610024448, "q_proj": { "group_size": { @@ -66824,7 +66824,7 @@ } }, { - "accuracy": 0.9883855855778644, + "accuracy": 0.989783304302316, "total_bits": 615020544, "q_proj": { "group_size": { @@ -66876,7 +66876,7 @@ } }, { - "accuracy": 0.9888038243118086, + "accuracy": 0.9903724068089536, "total_bits": 623951872, "q_proj": { "group_size": { @@ -66928,7 +66928,7 @@ } }, { - "accuracy": 0.9895388021280891, + "accuracy": 0.9904633093821374, "total_bits": 626473984, "q_proj": { "group_size": { @@ -66992,7 +66992,7 @@ } }, { - "accuracy": 0.9909956282690952, + "accuracy": 0.9907932501090201, "total_bits": 630355968, "q_proj": { "group_size": { @@ -67056,7 +67056,7 @@ } }, { - "accuracy": 0.9922518855647037, + "accuracy": 0.9922968784445211, "total_bits": 637362176, "q_proj": { "group_size": { @@ -67117,7 +67117,7 @@ } }, { - "accuracy": 0.9927365568123365, + "accuracy": 0.9928401034129294, "total_bits": 646823936, "q_proj": { "group_size": { @@ -67178,7 +67178,7 @@ } }, { - "accuracy": 0.9956777931043976, + "accuracy": 0.9957563018328265, "total_bits": 784740352, "q_proj": { "group_size": { @@ -67239,7 +67239,7 @@ } }, { - "accuracy": 0.9963739655519787, + "accuracy": 0.9964043505881962, "total_bits": 797818880, "q_proj": { "group_size": { @@ -67300,7 +67300,7 @@ } }, { - "accuracy": 0.9970193109229991, + "accuracy": 0.9970912915703497, "total_bits": 911749120, "q_proj": { "group_size": { @@ -67352,7 +67352,7 @@ } }, { - "accuracy": 0.9981610229925105, + "accuracy": 0.9981921756345975, "total_bits": 942718976, "q_proj": { "group_size": { @@ -67404,7 +67404,7 @@ } }, { - "accuracy": 0.999178483886154, + "accuracy": 0.9992004868231321, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -67458,7 +67458,7 @@ ], "model.layers.34.mlp": [ { - "accuracy": 0.9527272770279333, + "accuracy": 0.9527148886730796, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -67510,7 +67510,7 @@ } }, { - "accuracy": 0.9540642625407169, + "accuracy": 0.9540571978217677, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -67562,7 +67562,7 @@ } }, { - "accuracy": 0.9613308404621325, + "accuracy": 0.9613171157084013, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -67611,7 +67611,7 @@ } }, { - "accuracy": 0.9633358842448184, + "accuracy": 0.9633238095986216, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -67660,7 +67660,7 @@ } }, { - "accuracy": 0.9766831884258672, + "accuracy": 0.9766689884035211, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -67712,7 +67712,7 @@ } }, { - "accuracy": 0.9784724633944663, + "accuracy": 0.9784592766510812, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -67764,7 +67764,7 @@ } }, { - "accuracy": 0.9813115534029508, + "accuracy": 0.9813011445497212, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -67813,7 +67813,7 @@ } }, { - "accuracy": 0.9880659572387996, + "accuracy": 0.9880595944429699, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -67856,7 +67856,7 @@ } }, { - "accuracy": 0.9889881093251077, + "accuracy": 0.9889803560156571, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -67899,7 +67899,7 @@ } }, { - "accuracy": 0.9881472415045688, + "accuracy": 0.9881425617556823, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -67951,7 +67951,7 @@ } }, { - "accuracy": 0.9895512022470173, + "accuracy": 0.989547673808901, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -68003,7 +68003,7 @@ } }, { - "accuracy": 0.993960177427844, + "accuracy": 0.9939589696495157, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -68055,7 +68055,7 @@ } }, { - "accuracy": 0.9947900054486174, + "accuracy": 0.9947876369482592, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -68107,7 +68107,7 @@ } }, { - "accuracy": 0.9967600075822127, + "accuracy": 0.9967591674312165, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -68150,7 +68150,7 @@ } }, { - "accuracy": 0.9969049935278139, + "accuracy": 0.9969049301978788, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -68199,7 +68199,7 @@ } }, { - "accuracy": 0.997466776127878, + "accuracy": 0.9974661636117258, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -68245,7 +68245,7 @@ } }, { - "accuracy": 0.9991073220183975, + "accuracy": 0.9991073130973076, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -68287,7 +68287,7 @@ ], "model.layers.35.self_attn": [ { - "accuracy": 0.9590424832544828, + "accuracy": 0.9615837178732219, "total_bits": 320757760, "q_proj": { "group_size": { @@ -68351,7 +68351,7 @@ } }, { - "accuracy": 0.9629373393560711, + "accuracy": 0.9635013812466672, "total_bits": 329080832, "q_proj": { "group_size": { @@ -68415,7 +68415,7 @@ } }, { - "accuracy": 0.9686742770044428, + "accuracy": 0.9689799453082838, "total_bits": 336024576, "q_proj": { "group_size": { @@ -68479,7 +68479,7 @@ } }, { - "accuracy": 0.9752918654366544, + "accuracy": 0.9755679980704659, "total_bits": 401557504, "q_proj": { "group_size": { @@ -68543,7 +68543,7 @@ } }, { - "accuracy": 0.9804421318204779, + "accuracy": 0.9805677846858376, "total_bits": 475279360, "q_proj": { "group_size": { @@ -68607,7 +68607,7 @@ } }, { - "accuracy": 0.9808286475507837, + "accuracy": 0.9801337185658907, "total_bits": 475479040, "q_proj": { "group_size": { @@ -68671,7 +68671,7 @@ } }, { - "accuracy": 0.9887093470284813, + "accuracy": 0.98894853811515, "total_bits": 609759232, "q_proj": { "group_size": { @@ -68723,7 +68723,7 @@ } }, { - "accuracy": 0.9893146204321008, + "accuracy": 0.9890809231682828, "total_bits": 610024448, "q_proj": { "group_size": { @@ -68775,7 +68775,7 @@ } }, { - "accuracy": 0.9898553647493061, + "accuracy": 0.9898843428021983, "total_bits": 615020544, "q_proj": { "group_size": { @@ -68827,7 +68827,7 @@ } }, { - "accuracy": 0.9903425754685151, + "accuracy": 0.9903351664543152, "total_bits": 623951872, "q_proj": { "group_size": { @@ -68879,7 +68879,7 @@ } }, { - "accuracy": 0.989807704561635, + "accuracy": 0.9895566560720143, "total_bits": 626473984, "q_proj": { "group_size": { @@ -68943,7 +68943,7 @@ } }, { - "accuracy": 0.990382825073443, + "accuracy": 0.9904612992939196, "total_bits": 630355968, "q_proj": { "group_size": { @@ -69007,7 +69007,7 @@ } }, { - "accuracy": 0.9920692381105924, + "accuracy": 0.9921052408845801, "total_bits": 637362176, "q_proj": { "group_size": { @@ -69068,7 +69068,7 @@ } }, { - "accuracy": 0.9928046111997805, + "accuracy": 0.9927699495303003, "total_bits": 646823936, "q_proj": { "group_size": { @@ -69129,7 +69129,7 @@ } }, { - "accuracy": 0.9956385167805772, + "accuracy": 0.9956420663940279, "total_bits": 784740352, "q_proj": { "group_size": { @@ -69190,7 +69190,7 @@ } }, { - "accuracy": 0.9963736216488638, + "accuracy": 0.9963480559618849, "total_bits": 797818880, "q_proj": { "group_size": { @@ -69251,7 +69251,7 @@ } }, { - "accuracy": 0.9969365394821292, + "accuracy": 0.9969104397061624, "total_bits": 911749120, "q_proj": { "group_size": { @@ -69303,7 +69303,7 @@ } }, { - "accuracy": 0.9981943249310318, + "accuracy": 0.9981994764193108, "total_bits": 942718976, "q_proj": { "group_size": { @@ -69355,7 +69355,7 @@ } }, { - "accuracy": 0.9991999860656889, + "accuracy": 0.9991907704327452, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -69409,7 +69409,7 @@ ], "model.layers.35.mlp": [ { - "accuracy": 0.9534934539543956, + "accuracy": 0.9534893757418582, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -69461,7 +69461,7 @@ } }, { - "accuracy": 0.9548059827403018, + "accuracy": 0.9548076485332689, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -69513,7 +69513,7 @@ } }, { - "accuracy": 0.9617149233818054, + "accuracy": 0.961725978474868, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -69562,7 +69562,7 @@ } }, { - "accuracy": 0.9636126505701166, + "accuracy": 0.9636240789764806, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -69611,7 +69611,7 @@ } }, { - "accuracy": 0.9770650048004953, + "accuracy": 0.9770617014483401, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -69663,7 +69663,7 @@ } }, { - "accuracy": 0.9788241402099007, + "accuracy": 0.9788179507381037, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -69715,7 +69715,7 @@ } }, { - "accuracy": 0.9815127677039096, + "accuracy": 0.9815116054133365, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -69764,7 +69764,7 @@ } }, { - "accuracy": 0.9882786603350389, + "accuracy": 0.988275472270815, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -69807,7 +69807,7 @@ } }, { - "accuracy": 0.9891888287506605, + "accuracy": 0.9891846736795024, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -69850,7 +69850,7 @@ } }, { - "accuracy": 0.988341115807232, + "accuracy": 0.9883366854567277, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -69902,7 +69902,7 @@ } }, { - "accuracy": 0.9897241874745017, + "accuracy": 0.9897185007208272, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -69954,7 +69954,7 @@ } }, { - "accuracy": 0.9940574306406473, + "accuracy": 0.9940568267514831, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -70006,7 +70006,7 @@ } }, { - "accuracy": 0.9948739695705866, + "accuracy": 0.9948716975356403, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -70058,7 +70058,7 @@ } }, { - "accuracy": 0.996817029032268, + "accuracy": 0.9968149724759554, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -70101,7 +70101,7 @@ } }, { - "accuracy": 0.9969524570593709, + "accuracy": 0.9969496121139902, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -70150,7 +70150,7 @@ } }, { - "accuracy": 0.9974799805173749, + "accuracy": 0.9974775831950339, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -70196,7 +70196,7 @@ } }, { - "accuracy": 0.9991128220173874, + "accuracy": 0.9991114353270907, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -70238,7 +70238,7 @@ ], "model.layers.36.self_attn": [ { - "accuracy": 0.9752436797869833, + "accuracy": 0.9754256464933094, "total_bits": 320757760, "q_proj": { "group_size": { @@ -70302,7 +70302,7 @@ } }, { - "accuracy": 0.9765073704092126, + "accuracy": 0.9766364725012529, "total_bits": 329080832, "q_proj": { "group_size": { @@ -70366,7 +70366,7 @@ } }, { - "accuracy": 0.9799943534951461, + "accuracy": 0.9799172501814993, "total_bits": 336024576, "q_proj": { "group_size": { @@ -70430,7 +70430,7 @@ } }, { - "accuracy": 0.9844045874319578, + "accuracy": 0.9842887865869623, "total_bits": 401557504, "q_proj": { "group_size": { @@ -70494,7 +70494,7 @@ } }, { - "accuracy": 0.9867488977156187, + "accuracy": 0.9866924129034343, "total_bits": 475279360, "q_proj": { "group_size": { @@ -70558,7 +70558,7 @@ } }, { - "accuracy": 0.9871573832474256, + "accuracy": 0.9872963224586687, "total_bits": 475479040, "q_proj": { "group_size": { @@ -70622,7 +70622,7 @@ } }, { - "accuracy": 0.9902573447478445, + "accuracy": 0.9902246837553225, "total_bits": 609759232, "q_proj": { "group_size": { @@ -70674,7 +70674,7 @@ } }, { - "accuracy": 0.9907789614639784, + "accuracy": 0.9909871236274117, "total_bits": 610024448, "q_proj": { "group_size": { @@ -70726,7 +70726,7 @@ } }, { - "accuracy": 0.9914127172608125, + "accuracy": 0.9915269872075633, "total_bits": 615020544, "q_proj": { "group_size": { @@ -70778,7 +70778,7 @@ } }, { - "accuracy": 0.991696589087185, + "accuracy": 0.9917515986844113, "total_bits": 623951872, "q_proj": { "group_size": { @@ -70830,7 +70830,7 @@ } }, { - "accuracy": 0.9930648364518818, + "accuracy": 0.9931377721460242, "total_bits": 626473984, "q_proj": { "group_size": { @@ -70894,7 +70894,7 @@ } }, { - "accuracy": 0.993553894522943, + "accuracy": 0.9935964185156321, "total_bits": 630355968, "q_proj": { "group_size": { @@ -70958,7 +70958,7 @@ } }, { - "accuracy": 0.9937853934733492, + "accuracy": 0.9938667424415287, "total_bits": 637362176, "q_proj": { "group_size": { @@ -71019,7 +71019,7 @@ } }, { - "accuracy": 0.9942909875198415, + "accuracy": 0.9943572906287093, "total_bits": 646823936, "q_proj": { "group_size": { @@ -71080,7 +71080,7 @@ } }, { - "accuracy": 0.9964677165997656, + "accuracy": 0.9965229712818798, "total_bits": 784740352, "q_proj": { "group_size": { @@ -71141,7 +71141,7 @@ } }, { - "accuracy": 0.9970073466630358, + "accuracy": 0.9970489746253741, "total_bits": 797818880, "q_proj": { "group_size": { @@ -71202,7 +71202,7 @@ } }, { - "accuracy": 0.9970194926779521, + "accuracy": 0.9970790606580282, "total_bits": 911749120, "q_proj": { "group_size": { @@ -71254,7 +71254,7 @@ } }, { - "accuracy": 0.9986748862031259, + "accuracy": 0.9986992938149917, "total_bits": 942718976, "q_proj": { "group_size": { @@ -71306,7 +71306,7 @@ } }, { - "accuracy": 0.9992137521897492, + "accuracy": 0.9992215954943707, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -71360,7 +71360,7 @@ ], "model.layers.36.mlp": [ { - "accuracy": 0.9550625117201554, + "accuracy": 0.9550592271905196, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -71412,7 +71412,7 @@ } }, { - "accuracy": 0.9562310545068038, + "accuracy": 0.9562255050006666, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -71464,7 +71464,7 @@ } }, { - "accuracy": 0.9627236692528975, + "accuracy": 0.9627282525363722, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -71513,7 +71513,7 @@ } }, { - "accuracy": 0.9645929179693523, + "accuracy": 0.9646002493406597, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -71562,7 +71562,7 @@ } }, { - "accuracy": 0.9775997870846799, + "accuracy": 0.977595075180656, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -71614,7 +71614,7 @@ } }, { - "accuracy": 0.9793205323972201, + "accuracy": 0.979319343441411, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -71666,7 +71666,7 @@ } }, { - "accuracy": 0.9819824319136771, + "accuracy": 0.9819812947197964, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -71715,7 +71715,7 @@ } }, { - "accuracy": 0.9884894090263467, + "accuracy": 0.988487372272893, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -71758,7 +71758,7 @@ } }, { - "accuracy": 0.9894220946650756, + "accuracy": 0.9894212186336517, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -71801,7 +71801,7 @@ } }, { - "accuracy": 0.9886033864397752, + "accuracy": 0.9885982063255812, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -71853,7 +71853,7 @@ } }, { - "accuracy": 0.9899482209431497, + "accuracy": 0.9899468790543707, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -71905,7 +71905,7 @@ } }, { - "accuracy": 0.9941847959631368, + "accuracy": 0.9941834148607755, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -71957,7 +71957,7 @@ } }, { - "accuracy": 0.9949840001369777, + "accuracy": 0.9949827445180792, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -72009,7 +72009,7 @@ } }, { - "accuracy": 0.9968775708816553, + "accuracy": 0.9968767144570225, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -72052,7 +72052,7 @@ } }, { - "accuracy": 0.9970189285905737, + "accuracy": 0.9970172563273656, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -72101,7 +72101,7 @@ } }, { - "accuracy": 0.9975457138528949, + "accuracy": 0.9975441966794039, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -72147,7 +72147,7 @@ } }, { - "accuracy": 0.9991152036542955, + "accuracy": 0.9991145223183068, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -72189,7 +72189,7 @@ ], "model.layers.37.self_attn": [ { - "accuracy": 0.9729624359231246, + "accuracy": 0.9731402114817971, "total_bits": 320757760, "q_proj": { "group_size": { @@ -72253,7 +72253,7 @@ } }, { - "accuracy": 0.9746807248968827, + "accuracy": 0.974785997679359, "total_bits": 329080832, "q_proj": { "group_size": { @@ -72317,7 +72317,7 @@ } }, { - "accuracy": 0.9777941452829462, + "accuracy": 0.9779112025311119, "total_bits": 336024576, "q_proj": { "group_size": { @@ -72381,7 +72381,7 @@ } }, { - "accuracy": 0.9831059543710006, + "accuracy": 0.9831503708111612, "total_bits": 401557504, "q_proj": { "group_size": { @@ -72445,7 +72445,7 @@ } }, { - "accuracy": 0.9863779638942919, + "accuracy": 0.9864734630835684, "total_bits": 475279360, "q_proj": { "group_size": { @@ -72509,7 +72509,7 @@ } }, { - "accuracy": 0.9866755494945928, + "accuracy": 0.9868635401913994, "total_bits": 475479040, "q_proj": { "group_size": { @@ -72573,7 +72573,7 @@ } }, { - "accuracy": 0.9911819304290571, + "accuracy": 0.9911651321147618, "total_bits": 609759232, "q_proj": { "group_size": { @@ -72625,7 +72625,7 @@ } }, { - "accuracy": 0.9915300403770647, + "accuracy": 0.9916199382982755, "total_bits": 610024448, "q_proj": { "group_size": { @@ -72677,7 +72677,7 @@ } }, { - "accuracy": 0.9920860952452609, + "accuracy": 0.9921998052220595, "total_bits": 615020544, "q_proj": { "group_size": { @@ -72729,7 +72729,7 @@ } }, { - "accuracy": 0.9924097539562928, + "accuracy": 0.9924696754468115, "total_bits": 623951872, "q_proj": { "group_size": { @@ -72781,7 +72781,7 @@ } }, { - "accuracy": 0.992787101551106, + "accuracy": 0.9928990422110808, "total_bits": 626473984, "q_proj": { "group_size": { @@ -72845,7 +72845,7 @@ } }, { - "accuracy": 0.9933146842216191, + "accuracy": 0.9933355811395144, "total_bits": 630355968, "q_proj": { "group_size": { @@ -72909,7 +72909,7 @@ } }, { - "accuracy": 0.9935888581370053, + "accuracy": 0.9937089701232157, "total_bits": 637362176, "q_proj": { "group_size": { @@ -72970,7 +72970,7 @@ } }, { - "accuracy": 0.9941637963056564, + "accuracy": 0.9941742263342205, "total_bits": 646823936, "q_proj": { "group_size": { @@ -73031,7 +73031,7 @@ } }, { - "accuracy": 0.9963858935393786, + "accuracy": 0.9963987289290679, "total_bits": 784740352, "q_proj": { "group_size": { @@ -73092,7 +73092,7 @@ } }, { - "accuracy": 0.9969750681990072, + "accuracy": 0.9969951618266734, "total_bits": 797818880, "q_proj": { "group_size": { @@ -73153,7 +73153,7 @@ } }, { - "accuracy": 0.9971892916058239, + "accuracy": 0.9971955346041604, "total_bits": 911749120, "q_proj": { "group_size": { @@ -73205,7 +73205,7 @@ } }, { - "accuracy": 0.9986069929835043, + "accuracy": 0.9986096399983293, "total_bits": 942718976, "q_proj": { "group_size": { @@ -73257,7 +73257,7 @@ } }, { - "accuracy": 0.9992045642513978, + "accuracy": 0.9992105616746765, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -73311,7 +73311,7 @@ ], "model.layers.37.mlp": [ { - "accuracy": 0.9556482057822377, + "accuracy": 0.9556552516786676, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -73363,7 +73363,7 @@ } }, { - "accuracy": 0.9567890512315851, + "accuracy": 0.9568089842796326, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -73415,7 +73415,7 @@ } }, { - "accuracy": 0.9631767021982294, + "accuracy": 0.9631907877169157, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -73464,7 +73464,7 @@ } }, { - "accuracy": 0.965013115029586, + "accuracy": 0.9650265260746604, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -73513,7 +73513,7 @@ } }, { - "accuracy": 0.9778931391866583, + "accuracy": 0.9779009787659896, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -73565,7 +73565,7 @@ } }, { - "accuracy": 0.9796006083488464, + "accuracy": 0.9796029470468822, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -73617,7 +73617,7 @@ } }, { - "accuracy": 0.9822132603118294, + "accuracy": 0.9822145371060622, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -73666,7 +73666,7 @@ } }, { - "accuracy": 0.9886786247554579, + "accuracy": 0.9886786780859295, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -73709,7 +73709,7 @@ } }, { - "accuracy": 0.9895982993276495, + "accuracy": 0.9895993965236765, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -73752,7 +73752,7 @@ } }, { - "accuracy": 0.988758214210209, + "accuracy": 0.9887637072487881, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -73804,7 +73804,7 @@ } }, { - "accuracy": 0.9901065842101449, + "accuracy": 0.9901102279361925, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -73856,7 +73856,7 @@ } }, { - "accuracy": 0.9942702823563626, + "accuracy": 0.994273530417367, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -73908,7 +73908,7 @@ } }, { - "accuracy": 0.9950691159618529, + "accuracy": 0.995070621763405, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -73960,7 +73960,7 @@ } }, { - "accuracy": 0.99693378355158, + "accuracy": 0.9969352399440188, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -74003,7 +74003,7 @@ } }, { - "accuracy": 0.9970661519389403, + "accuracy": 0.9970672105095888, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -74052,7 +74052,7 @@ } }, { - "accuracy": 0.9975692818039342, + "accuracy": 0.9975703807645723, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -74098,7 +74098,7 @@ } }, { - "accuracy": 0.9991372905083393, + "accuracy": 0.9991372864889471, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -74140,7 +74140,7 @@ ], "model.layers.38.self_attn": [ { - "accuracy": 0.9715856909751892, + "accuracy": 0.9718447515839025, "total_bits": 320757760, "q_proj": { "group_size": { @@ -74204,7 +74204,7 @@ } }, { - "accuracy": 0.9747463496107804, + "accuracy": 0.9745464607288963, "total_bits": 329080832, "q_proj": { "group_size": { @@ -74268,7 +74268,7 @@ } }, { - "accuracy": 0.9776538861425299, + "accuracy": 0.9777676698408628, "total_bits": 336024576, "q_proj": { "group_size": { @@ -74332,7 +74332,7 @@ } }, { - "accuracy": 0.9830130087701898, + "accuracy": 0.9831245604314303, "total_bits": 401557504, "q_proj": { "group_size": { @@ -74396,7 +74396,7 @@ } }, { - "accuracy": 0.9856761348874945, + "accuracy": 0.9858307509045852, "total_bits": 475279360, "q_proj": { "group_size": { @@ -74460,7 +74460,7 @@ } }, { - "accuracy": 0.9861865639686584, + "accuracy": 0.986184476237548, "total_bits": 475479040, "q_proj": { "group_size": { @@ -74524,7 +74524,7 @@ } }, { - "accuracy": 0.9897762223293907, + "accuracy": 0.9899974990832178, "total_bits": 609759232, "q_proj": { "group_size": { @@ -74576,7 +74576,7 @@ } }, { - "accuracy": 0.9904081052855441, + "accuracy": 0.990362284214873, "total_bits": 610024448, "q_proj": { "group_size": { @@ -74628,7 +74628,7 @@ } }, { - "accuracy": 0.9907835478845396, + "accuracy": 0.9909798863687014, "total_bits": 615020544, "q_proj": { "group_size": { @@ -74680,7 +74680,7 @@ } }, { - "accuracy": 0.9910660315501062, + "accuracy": 0.9912428910795011, "total_bits": 623951872, "q_proj": { "group_size": { @@ -74732,7 +74732,7 @@ } }, { - "accuracy": 0.992474335588907, + "accuracy": 0.9924193675580778, "total_bits": 626473984, "q_proj": { "group_size": { @@ -74796,7 +74796,7 @@ } }, { - "accuracy": 0.9929433182666176, + "accuracy": 0.9929072276542061, "total_bits": 630355968, "q_proj": { "group_size": { @@ -74860,7 +74860,7 @@ } }, { - "accuracy": 0.993135384038875, + "accuracy": 0.993086105898807, "total_bits": 637362176, "q_proj": { "group_size": { @@ -74921,7 +74921,7 @@ } }, { - "accuracy": 0.9936725818797162, + "accuracy": 0.9936287285465943, "total_bits": 646823936, "q_proj": { "group_size": { @@ -74982,7 +74982,7 @@ } }, { - "accuracy": 0.9961850258864855, + "accuracy": 0.9961832142189929, "total_bits": 784740352, "q_proj": { "group_size": { @@ -75043,7 +75043,7 @@ } }, { - "accuracy": 0.9968123222260099, + "accuracy": 0.9967413040760317, "total_bits": 797818880, "q_proj": { "group_size": { @@ -75104,7 +75104,7 @@ } }, { - "accuracy": 0.9969043863054953, + "accuracy": 0.996892873197794, "total_bits": 911749120, "q_proj": { "group_size": { @@ -75156,7 +75156,7 @@ } }, { - "accuracy": 0.9985720336829361, + "accuracy": 0.9985820642231327, "total_bits": 942718976, "q_proj": { "group_size": { @@ -75208,7 +75208,7 @@ } }, { - "accuracy": 0.9992022334450954, + "accuracy": 0.9991794954005041, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -75262,7 +75262,7 @@ ], "model.layers.38.mlp": [ { - "accuracy": 0.9560274293548182, + "accuracy": 0.9560815284126684, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -75314,7 +75314,7 @@ } }, { - "accuracy": 0.9571839978820399, + "accuracy": 0.9572354178679616, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -75366,7 +75366,7 @@ } }, { - "accuracy": 0.9634983288614374, + "accuracy": 0.9635435876093412, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -75415,7 +75415,7 @@ } }, { - "accuracy": 0.9652881873281378, + "accuracy": 0.9653327747395164, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -75464,7 +75464,7 @@ } }, { - "accuracy": 0.9781132026722557, + "accuracy": 0.9781347748480345, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -75516,7 +75516,7 @@ } }, { - "accuracy": 0.9798016626583902, + "accuracy": 0.9798228442668915, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -75568,7 +75568,7 @@ } }, { - "accuracy": 0.9823530175183949, + "accuracy": 0.9823685868790275, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -75617,7 +75617,7 @@ } }, { - "accuracy": 0.988787348333158, + "accuracy": 0.9887946122571042, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -75660,7 +75660,7 @@ } }, { - "accuracy": 0.9896664611603084, + "accuracy": 0.9896812729145351, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -75703,7 +75703,7 @@ } }, { - "accuracy": 0.9888751600918017, + "accuracy": 0.9888854881650523, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -75755,7 +75755,7 @@ } }, { - "accuracy": 0.9901876073134573, + "accuracy": 0.9901970562181974, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -75807,7 +75807,7 @@ } }, { - "accuracy": 0.9943314178993827, + "accuracy": 0.9943368434906006, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -75859,7 +75859,7 @@ } }, { - "accuracy": 0.995105194025918, + "accuracy": 0.9951096769226225, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -75911,7 +75911,7 @@ } }, { - "accuracy": 0.9969598139195066, + "accuracy": 0.9969621435984185, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -75954,7 +75954,7 @@ } }, { - "accuracy": 0.9971079834197697, + "accuracy": 0.9971109067923144, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -76003,7 +76003,7 @@ } }, { - "accuracy": 0.9976056539698651, + "accuracy": 0.9976075256341382, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -76049,7 +76049,7 @@ } }, { - "accuracy": 0.9991859996593312, + "accuracy": 0.9991869780382043, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -76091,7 +76091,7 @@ ], "model.layers.39.self_attn": [ { - "accuracy": 0.9727507986520466, + "accuracy": 0.9727698784125479, "total_bits": 320757760, "q_proj": { "group_size": { @@ -76155,7 +76155,7 @@ } }, { - "accuracy": 0.9745930386217017, + "accuracy": 0.9744302787278828, "total_bits": 329080832, "q_proj": { "group_size": { @@ -76219,7 +76219,7 @@ } }, { - "accuracy": 0.9766974511899447, + "accuracy": 0.9764877165618696, "total_bits": 336024576, "q_proj": { "group_size": { @@ -76283,7 +76283,7 @@ } }, { - "accuracy": 0.9818985007311168, + "accuracy": 0.9814491805277372, "total_bits": 401557504, "q_proj": { "group_size": { @@ -76347,7 +76347,7 @@ } }, { - "accuracy": 0.9851567149162292, + "accuracy": 0.98515721685008, "total_bits": 475279360, "q_proj": { "group_size": { @@ -76411,7 +76411,7 @@ } }, { - "accuracy": 0.9852726694784666, + "accuracy": 0.9849978403041237, "total_bits": 475479040, "q_proj": { "group_size": { @@ -76475,7 +76475,7 @@ } }, { - "accuracy": 0.9896653200450697, + "accuracy": 0.9899890085584239, "total_bits": 609759232, "q_proj": { "group_size": { @@ -76527,7 +76527,7 @@ } }, { - "accuracy": 0.9903147134341692, + "accuracy": 0.9904764373051492, "total_bits": 610024448, "q_proj": { "group_size": { @@ -76579,7 +76579,7 @@ } }, { - "accuracy": 0.9911483839938515, + "accuracy": 0.9909533544590599, "total_bits": 615020544, "q_proj": { "group_size": { @@ -76631,7 +76631,7 @@ } }, { - "accuracy": 0.9915320418382946, + "accuracy": 0.9913611843397743, "total_bits": 623951872, "q_proj": { "group_size": { @@ -76683,7 +76683,7 @@ } }, { - "accuracy": 0.991118491480225, + "accuracy": 0.9912619206466173, "total_bits": 626473984, "q_proj": { "group_size": { @@ -76747,7 +76747,7 @@ } }, { - "accuracy": 0.9915628354800375, + "accuracy": 0.9917955704425511, "total_bits": 630355968, "q_proj": { "group_size": { @@ -76811,7 +76811,7 @@ } }, { - "accuracy": 0.992757836454793, + "accuracy": 0.9929927414969394, "total_bits": 637362176, "q_proj": { "group_size": { @@ -76872,7 +76872,7 @@ } }, { - "accuracy": 0.9936416478533494, + "accuracy": 0.9936754374127639, "total_bits": 646823936, "q_proj": { "group_size": { @@ -76933,7 +76933,7 @@ } }, { - "accuracy": 0.9959909315956267, + "accuracy": 0.996027898239462, "total_bits": 784740352, "q_proj": { "group_size": { @@ -76994,7 +76994,7 @@ } }, { - "accuracy": 0.9965674453660062, + "accuracy": 0.9966498558458529, "total_bits": 797818880, "q_proj": { "group_size": { @@ -77055,7 +77055,7 @@ } }, { - "accuracy": 0.9966721464144556, + "accuracy": 0.9966841238109689, "total_bits": 911749120, "q_proj": { "group_size": { @@ -77107,7 +77107,7 @@ } }, { - "accuracy": 0.9985165581303207, + "accuracy": 0.9985108044194547, "total_bits": 942718976, "q_proj": { "group_size": { @@ -77159,7 +77159,7 @@ } }, { - "accuracy": 0.999087089084481, + "accuracy": 0.9990922878251264, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -77213,7 +77213,7 @@ ], "model.layers.39.mlp": [ { - "accuracy": 0.9558397029575548, + "accuracy": 0.9558913299911901, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -77265,7 +77265,7 @@ } }, { - "accuracy": 0.9569758521883112, + "accuracy": 0.9570177385681554, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -77317,7 +77317,7 @@ } }, { - "accuracy": 0.9631552131552445, + "accuracy": 0.9631913618037575, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -77366,7 +77366,7 @@ } }, { - "accuracy": 0.9649330314837004, + "accuracy": 0.9649658893284044, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -77415,7 +77415,7 @@ } }, { - "accuracy": 0.9779258417455774, + "accuracy": 0.9779474029415532, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -77467,7 +77467,7 @@ } }, { - "accuracy": 0.9796351508090371, + "accuracy": 0.979663223028183, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -77519,7 +77519,7 @@ } }, { - "accuracy": 0.9821741408423373, + "accuracy": 0.9821961584844088, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -77568,7 +77568,7 @@ } }, { - "accuracy": 0.9887032908828635, + "accuracy": 0.9887121006062156, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -77611,7 +77611,7 @@ } }, { - "accuracy": 0.9896186268643329, + "accuracy": 0.9896310905092641, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -77654,7 +77654,7 @@ } }, { - "accuracy": 0.9887810357307133, + "accuracy": 0.9887873671556774, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -77706,7 +77706,7 @@ } }, { - "accuracy": 0.9901201128959656, + "accuracy": 0.9901285838139685, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -77758,7 +77758,7 @@ } }, { - "accuracy": 0.9942868493105236, + "accuracy": 0.9942865932458326, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -77810,7 +77810,7 @@ } }, { - "accuracy": 0.9950749266304468, + "accuracy": 0.9950801051760975, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -77862,7 +77862,7 @@ } }, { - "accuracy": 0.996944927659474, + "accuracy": 0.9969455623313, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -77905,7 +77905,7 @@ } }, { - "accuracy": 0.9970731276430582, + "accuracy": 0.9970731986196417, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -77954,7 +77954,7 @@ } }, { - "accuracy": 0.9975561805461582, + "accuracy": 0.9975550145302948, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -78000,7 +78000,7 @@ } }, { - "accuracy": 0.9991489708619682, + "accuracy": 0.9991502485385066, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -78042,7 +78042,7 @@ ], "model.layers.40.self_attn": [ { - "accuracy": 0.9787316181157765, + "accuracy": 0.9790317541674564, "total_bits": 320757760, "q_proj": { "group_size": { @@ -78106,7 +78106,7 @@ } }, { - "accuracy": 0.9810021943167636, + "accuracy": 0.9811096756081832, "total_bits": 329080832, "q_proj": { "group_size": { @@ -78170,7 +78170,7 @@ } }, { - "accuracy": 0.9835346560729178, + "accuracy": 0.9834603149639932, "total_bits": 336024576, "q_proj": { "group_size": { @@ -78234,7 +78234,7 @@ } }, { - "accuracy": 0.986806394238221, + "accuracy": 0.9866805609903837, "total_bits": 401557504, "q_proj": { "group_size": { @@ -78298,7 +78298,7 @@ } }, { - "accuracy": 0.9885824589352858, + "accuracy": 0.9884440843996248, "total_bits": 475279360, "q_proj": { "group_size": { @@ -78362,7 +78362,7 @@ } }, { - "accuracy": 0.9889615899638126, + "accuracy": 0.9889172056787893, "total_bits": 475479040, "q_proj": { "group_size": { @@ -78426,7 +78426,7 @@ } }, { - "accuracy": 0.9910609533912257, + "accuracy": 0.9910143825568651, "total_bits": 609759232, "q_proj": { "group_size": { @@ -78478,7 +78478,7 @@ } }, { - "accuracy": 0.9914473326582658, + "accuracy": 0.9913953244686127, "total_bits": 610024448, "q_proj": { "group_size": { @@ -78530,7 +78530,7 @@ } }, { - "accuracy": 0.9919082287110781, + "accuracy": 0.9920373159019571, "total_bits": 615020544, "q_proj": { "group_size": { @@ -78582,7 +78582,7 @@ } }, { - "accuracy": 0.9919929324012053, + "accuracy": 0.9921322457100216, "total_bits": 623951872, "q_proj": { "group_size": { @@ -78634,7 +78634,7 @@ } }, { - "accuracy": 0.9935811656086069, + "accuracy": 0.993672330520655, "total_bits": 626473984, "q_proj": { "group_size": { @@ -78698,7 +78698,7 @@ } }, { - "accuracy": 0.9938171239275682, + "accuracy": 0.9941347017884254, "total_bits": 630355968, "q_proj": { "group_size": { @@ -78762,7 +78762,7 @@ } }, { - "accuracy": 0.9939935348535839, + "accuracy": 0.99412631086613, "total_bits": 637362176, "q_proj": { "group_size": { @@ -78823,7 +78823,7 @@ } }, { - "accuracy": 0.9943441881945259, + "accuracy": 0.9946203114170777, "total_bits": 646823936, "q_proj": { "group_size": { @@ -78884,7 +78884,7 @@ } }, { - "accuracy": 0.9964963052617876, + "accuracy": 0.9964964625082517, "total_bits": 784740352, "q_proj": { "group_size": { @@ -78945,7 +78945,7 @@ } }, { - "accuracy": 0.9970318177028706, + "accuracy": 0.9969914045773054, "total_bits": 797818880, "q_proj": { "group_size": { @@ -79006,7 +79006,7 @@ } }, { - "accuracy": 0.9968293993488738, + "accuracy": 0.9968402046514185, "total_bits": 911749120, "q_proj": { "group_size": { @@ -79058,7 +79058,7 @@ } }, { - "accuracy": 0.9988946167654112, + "accuracy": 0.9988919597511229, "total_bits": 942718976, "q_proj": { "group_size": { @@ -79110,7 +79110,7 @@ } }, { - "accuracy": 0.9991674414394718, + "accuracy": 0.9991849867725059, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -79164,7 +79164,7 @@ ], "model.layers.40.mlp": [ { - "accuracy": 0.9554162339160317, + "accuracy": 0.9555049695466694, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -79216,7 +79216,7 @@ } }, { - "accuracy": 0.9565403837906687, + "accuracy": 0.9566192689694857, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -79268,7 +79268,7 @@ } }, { - "accuracy": 0.9627267216381273, + "accuracy": 0.9627982660343772, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -79317,7 +79317,7 @@ } }, { - "accuracy": 0.9645298091988814, + "accuracy": 0.9645961083863911, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -79366,7 +79366,7 @@ } }, { - "accuracy": 0.977634365621366, + "accuracy": 0.9776720043859983, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -79418,7 +79418,7 @@ } }, { - "accuracy": 0.9793675635990343, + "accuracy": 0.9794034330468429, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -79470,7 +79470,7 @@ } }, { - "accuracy": 0.9819494046662983, + "accuracy": 0.9819809637571636, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -79519,7 +79519,7 @@ } }, { - "accuracy": 0.9885285771206805, + "accuracy": 0.9885484207617609, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -79562,7 +79562,7 @@ } }, { - "accuracy": 0.9894739044340033, + "accuracy": 0.9894944703892657, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -79605,7 +79605,7 @@ } }, { - "accuracy": 0.9886253734952525, + "accuracy": 0.9886482459934134, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -79657,7 +79657,7 @@ } }, { - "accuracy": 0.9899814081819434, + "accuracy": 0.9900002416811491, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -79709,7 +79709,7 @@ } }, { - "accuracy": 0.9942006519750545, + "accuracy": 0.9942113078738514, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -79761,7 +79761,7 @@ } }, { - "accuracy": 0.9950022387661432, + "accuracy": 0.9950108998700192, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -79813,7 +79813,7 @@ } }, { - "accuracy": 0.9968884889232484, + "accuracy": 0.9968943380211529, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -79856,7 +79856,7 @@ } }, { - "accuracy": 0.9970300934816662, + "accuracy": 0.997037121731984, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -79905,7 +79905,7 @@ } }, { - "accuracy": 0.9975217478839975, + "accuracy": 0.9975283749793705, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -79951,7 +79951,7 @@ } }, { - "accuracy": 0.9991186700369182, + "accuracy": 0.9991211459824914, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -79993,7 +79993,7 @@ ], "model.layers.41.self_attn": [ { - "accuracy": 0.981377008714174, + "accuracy": 0.9817103555327967, "total_bits": 320757760, "q_proj": { "group_size": { @@ -80057,7 +80057,7 @@ } }, { - "accuracy": 0.9823866690460005, + "accuracy": 0.9825625607841894, "total_bits": 329080832, "q_proj": { "group_size": { @@ -80121,7 +80121,7 @@ } }, { - "accuracy": 0.9843117233953977, + "accuracy": 0.9847933979410874, "total_bits": 336024576, "q_proj": { "group_size": { @@ -80185,7 +80185,7 @@ } }, { - "accuracy": 0.9868928578339125, + "accuracy": 0.9874847327408037, "total_bits": 401557504, "q_proj": { "group_size": { @@ -80249,7 +80249,7 @@ } }, { - "accuracy": 0.9894664538534064, + "accuracy": 0.9897471627122477, "total_bits": 475279360, "q_proj": { "group_size": { @@ -80313,7 +80313,7 @@ } }, { - "accuracy": 0.9899954952691731, + "accuracy": 0.9896669599570727, "total_bits": 475479040, "q_proj": { "group_size": { @@ -80377,7 +80377,7 @@ } }, { - "accuracy": 0.9922491233599814, + "accuracy": 0.9920386915144167, "total_bits": 609759232, "q_proj": { "group_size": { @@ -80429,7 +80429,7 @@ } }, { - "accuracy": 0.992517877566187, + "accuracy": 0.992985852454838, "total_bits": 610024448, "q_proj": { "group_size": { @@ -80481,7 +80481,7 @@ } }, { - "accuracy": 0.9933448049582934, + "accuracy": 0.9933604190224096, "total_bits": 615020544, "q_proj": { "group_size": { @@ -80533,7 +80533,7 @@ } }, { - "accuracy": 0.9934868530223244, + "accuracy": 0.9935508044926744, "total_bits": 623951872, "q_proj": { "group_size": { @@ -80585,7 +80585,7 @@ } }, { - "accuracy": 0.9933688005334452, + "accuracy": 0.9933757625128093, "total_bits": 626473984, "q_proj": { "group_size": { @@ -80649,7 +80649,7 @@ } }, { - "accuracy": 0.9937149533315709, + "accuracy": 0.9936821994028593, "total_bits": 630355968, "q_proj": { "group_size": { @@ -80713,7 +80713,7 @@ } }, { - "accuracy": 0.9946313445505343, + "accuracy": 0.9947575679735133, "total_bits": 637362176, "q_proj": { "group_size": { @@ -80774,7 +80774,7 @@ } }, { - "accuracy": 0.995084344948593, + "accuracy": 0.9950749293753975, "total_bits": 646823936, "q_proj": { "group_size": { @@ -80835,7 +80835,7 @@ } }, { - "accuracy": 0.9967530714838129, + "accuracy": 0.9967881835212833, "total_bits": 784740352, "q_proj": { "group_size": { @@ -80896,7 +80896,7 @@ } }, { - "accuracy": 0.9972718334511707, + "accuracy": 0.9972929276133838, "total_bits": 797818880, "q_proj": { "group_size": { @@ -80957,7 +80957,7 @@ } }, { - "accuracy": 0.9970463712356592, + "accuracy": 0.9970766701980641, "total_bits": 911749120, "q_proj": { "group_size": { @@ -81009,7 +81009,7 @@ } }, { - "accuracy": 0.9989881955675388, + "accuracy": 0.999011244722887, "total_bits": 942718976, "q_proj": { "group_size": { @@ -81061,7 +81061,7 @@ } }, { - "accuracy": 0.9992328630760312, + "accuracy": 0.9992383528794897, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -81115,7 +81115,7 @@ ], "model.layers.41.mlp": [ { - "accuracy": 0.9559422135353088, + "accuracy": 0.9560472871127882, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -81167,7 +81167,7 @@ } }, { - "accuracy": 0.9570275262782448, + "accuracy": 0.9571303003712704, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -81219,7 +81219,7 @@ } }, { - "accuracy": 0.9631056471874839, + "accuracy": 0.9631915719885575, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -81268,7 +81268,7 @@ } }, { - "accuracy": 0.9648484901378029, + "accuracy": 0.9649315476417542, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -81317,7 +81317,7 @@ } }, { - "accuracy": 0.97787312457436, + "accuracy": 0.9779269946248907, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -81369,7 +81369,7 @@ } }, { - "accuracy": 0.9795892960146854, + "accuracy": 0.9796352700183266, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -81421,7 +81421,7 @@ } }, { - "accuracy": 0.9821181673752634, + "accuracy": 0.9821589526377226, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -81470,7 +81470,7 @@ } }, { - "accuracy": 0.9886950497564516, + "accuracy": 0.9887240050654662, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -81513,7 +81513,7 @@ } }, { - "accuracy": 0.9896255896279686, + "accuracy": 0.9896520964409176, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -81556,7 +81556,7 @@ } }, { - "accuracy": 0.9887655040151194, + "accuracy": 0.9887936232905639, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -81608,7 +81608,7 @@ } }, { - "accuracy": 0.9901051811481777, + "accuracy": 0.9901290841792759, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -81660,7 +81660,7 @@ } }, { - "accuracy": 0.9942855489881415, + "accuracy": 0.9942981400772145, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -81712,7 +81712,7 @@ } }, { - "accuracy": 0.9950712519256693, + "accuracy": 0.9950835626376303, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -81764,7 +81764,7 @@ } }, { - "accuracy": 0.9969560360830081, + "accuracy": 0.9969624933835707, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -81807,7 +81807,7 @@ } }, { - "accuracy": 0.9970734201763806, + "accuracy": 0.9970806497884425, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -81856,7 +81856,7 @@ } }, { - "accuracy": 0.9975486848699419, + "accuracy": 0.9975549874729231, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -81902,7 +81902,7 @@ } }, { - "accuracy": 0.9991428387400351, + "accuracy": 0.9991434329238377, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -81944,7 +81944,7 @@ ], "model.layers.42.self_attn": [ { - "accuracy": 0.9786916908464933, + "accuracy": 0.980894818117744, "total_bits": 320757760, "q_proj": { "group_size": { @@ -82008,7 +82008,7 @@ } }, { - "accuracy": 0.9811059644347743, + "accuracy": 0.9832934721520072, "total_bits": 329080832, "q_proj": { "group_size": { @@ -82072,7 +82072,7 @@ } }, { - "accuracy": 0.9859621697350552, + "accuracy": 0.9859640410071925, "total_bits": 336024576, "q_proj": { "group_size": { @@ -82136,7 +82136,7 @@ } }, { - "accuracy": 0.9880426463327909, + "accuracy": 0.9879391365929654, "total_bits": 401557504, "q_proj": { "group_size": { @@ -82200,7 +82200,7 @@ } }, { - "accuracy": 0.9892420964805704, + "accuracy": 0.9890787891651455, "total_bits": 475279360, "q_proj": { "group_size": { @@ -82264,7 +82264,7 @@ } }, { - "accuracy": 0.9886894955446845, + "accuracy": 0.9899472249181647, "total_bits": 475479040, "q_proj": { "group_size": { @@ -82328,7 +82328,7 @@ } }, { - "accuracy": 0.9915113503995695, + "accuracy": 0.9913074766334734, "total_bits": 609759232, "q_proj": { "group_size": { @@ -82380,7 +82380,7 @@ } }, { - "accuracy": 0.9924306273460388, + "accuracy": 0.9922790550871899, "total_bits": 610024448, "q_proj": { "group_size": { @@ -82432,7 +82432,7 @@ } }, { - "accuracy": 0.9930313958933479, + "accuracy": 0.9930967994426426, "total_bits": 615020544, "q_proj": { "group_size": { @@ -82484,7 +82484,7 @@ } }, { - "accuracy": 0.9931268966511676, + "accuracy": 0.9931854166482624, "total_bits": 623951872, "q_proj": { "group_size": { @@ -82536,7 +82536,7 @@ } }, { - "accuracy": 0.993076956585834, + "accuracy": 0.9928553825930545, "total_bits": 626473984, "q_proj": { "group_size": { @@ -82600,7 +82600,7 @@ } }, { - "accuracy": 0.99333471922498, + "accuracy": 0.9933174048599444, "total_bits": 630355968, "q_proj": { "group_size": { @@ -82664,7 +82664,7 @@ } }, { - "accuracy": 0.9946826194462023, + "accuracy": 0.9947880396717473, "total_bits": 637362176, "q_proj": { "group_size": { @@ -82725,7 +82725,7 @@ } }, { - "accuracy": 0.9951670310999218, + "accuracy": 0.9952448457479477, "total_bits": 646823936, "q_proj": { "group_size": { @@ -82786,7 +82786,7 @@ } }, { - "accuracy": 0.9967123240624604, + "accuracy": 0.9969708111725355, "total_bits": 784740352, "q_proj": { "group_size": { @@ -82847,7 +82847,7 @@ } }, { - "accuracy": 0.9974709151214675, + "accuracy": 0.997429493422571, "total_bits": 797818880, "q_proj": { "group_size": { @@ -82908,7 +82908,7 @@ } }, { - "accuracy": 0.9968867019603127, + "accuracy": 0.9971518994946229, "total_bits": 911749120, "q_proj": { "group_size": { @@ -82960,7 +82960,7 @@ } }, { - "accuracy": 0.9990777322336247, + "accuracy": 0.9991006518861181, "total_bits": 942718976, "q_proj": { "group_size": { @@ -83012,7 +83012,7 @@ } }, { - "accuracy": 0.999211617941527, + "accuracy": 0.9992156867918215, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -83066,7 +83066,7 @@ ], "model.layers.42.mlp": [ { - "accuracy": 0.9561494588851929, + "accuracy": 0.9562669459142183, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -83118,7 +83118,7 @@ } }, { - "accuracy": 0.9572267908799021, + "accuracy": 0.9573470573676259, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -83170,7 +83170,7 @@ } }, { - "accuracy": 0.9631681222664683, + "accuracy": 0.9632737197374043, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -83219,7 +83219,7 @@ } }, { - "accuracy": 0.9648521542549133, + "accuracy": 0.9649521175183748, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -83268,7 +83268,7 @@ } }, { - "accuracy": 0.9779521995469144, + "accuracy": 0.9780170305779106, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -83320,7 +83320,7 @@ } }, { - "accuracy": 0.9796571355116995, + "accuracy": 0.9797177910804749, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -83372,7 +83372,7 @@ } }, { - "accuracy": 0.982124946619335, + "accuracy": 0.9821780622005463, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -83421,7 +83421,7 @@ } }, { - "accuracy": 0.9887442008445138, + "accuracy": 0.9887767653716238, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -83464,7 +83464,7 @@ } }, { - "accuracy": 0.9896569785318876, + "accuracy": 0.9896865557683142, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -83507,7 +83507,7 @@ } }, { - "accuracy": 0.9888011475926951, + "accuracy": 0.9888323396444321, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -83559,7 +83559,7 @@ } }, { - "accuracy": 0.9901255243702939, + "accuracy": 0.9901557251026756, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -83611,7 +83611,7 @@ } }, { - "accuracy": 0.9942990713997891, + "accuracy": 0.9943152070045471, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -83663,7 +83663,7 @@ } }, { - "accuracy": 0.9950787213287855, + "accuracy": 0.9950926856775033, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -83715,7 +83715,7 @@ } }, { - "accuracy": 0.9969619332175506, + "accuracy": 0.9969711668397251, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -83758,7 +83758,7 @@ } }, { - "accuracy": 0.9970792949591812, + "accuracy": 0.997086314582511, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -83807,7 +83807,7 @@ } }, { - "accuracy": 0.9975409425402942, + "accuracy": 0.9975467739920867, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -83853,7 +83853,7 @@ } }, { - "accuracy": 0.9991413102926392, + "accuracy": 0.9991440702425806, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -83895,7 +83895,7 @@ ], "model.layers.43.self_attn": [ { - "accuracy": 0.9811300039291382, + "accuracy": 0.981926502365815, "total_bits": 320757760, "q_proj": { "group_size": { @@ -83959,7 +83959,7 @@ } }, { - "accuracy": 0.9827878569301806, + "accuracy": 0.9827630519866943, "total_bits": 329080832, "q_proj": { "group_size": { @@ -84023,7 +84023,7 @@ } }, { - "accuracy": 0.9841346144676208, + "accuracy": 0.9843070365880665, "total_bits": 336024576, "q_proj": { "group_size": { @@ -84087,7 +84087,7 @@ } }, { - "accuracy": 0.9864303940220883, + "accuracy": 0.9866969506991538, "total_bits": 401557504, "q_proj": { "group_size": { @@ -84151,7 +84151,7 @@ } }, { - "accuracy": 0.9882989666963878, + "accuracy": 0.9885782677876321, "total_bits": 475279360, "q_proj": { "group_size": { @@ -84215,7 +84215,7 @@ } }, { - "accuracy": 0.9884036199042672, + "accuracy": 0.9884995818138123, "total_bits": 475479040, "q_proj": { "group_size": { @@ -84279,7 +84279,7 @@ } }, { - "accuracy": 0.9911516740133888, + "accuracy": 0.9912183967075849, "total_bits": 609759232, "q_proj": { "group_size": { @@ -84331,7 +84331,7 @@ } }, { - "accuracy": 0.9922710186556766, + "accuracy": 0.9921230116957113, "total_bits": 610024448, "q_proj": { "group_size": { @@ -84383,7 +84383,7 @@ } }, { - "accuracy": 0.9929804794098201, + "accuracy": 0.9929813342659097, "total_bits": 615020544, "q_proj": { "group_size": { @@ -84435,7 +84435,7 @@ } }, { - "accuracy": 0.9931048484225022, + "accuracy": 0.9930923385055441, "total_bits": 623951872, "q_proj": { "group_size": { @@ -84487,7 +84487,7 @@ } }, { - "accuracy": 0.9929938684952887, + "accuracy": 0.992985272093823, "total_bits": 626473984, "q_proj": { "group_size": { @@ -84551,7 +84551,7 @@ } }, { - "accuracy": 0.9933382962879381, + "accuracy": 0.9933626526280454, "total_bits": 630355968, "q_proj": { "group_size": { @@ -84615,7 +84615,7 @@ } }, { - "accuracy": 0.9946055898540899, + "accuracy": 0.9946929706554664, "total_bits": 637362176, "q_proj": { "group_size": { @@ -84676,7 +84676,7 @@ } }, { - "accuracy": 0.9951291127424491, + "accuracy": 0.9952200368831032, "total_bits": 646823936, "q_proj": { "group_size": { @@ -84737,7 +84737,7 @@ } }, { - "accuracy": 0.9968126482869449, + "accuracy": 0.9968842736592418, "total_bits": 784740352, "q_proj": { "group_size": { @@ -84798,7 +84798,7 @@ } }, { - "accuracy": 0.9973898186887565, + "accuracy": 0.997448510245273, "total_bits": 797818880, "q_proj": { "group_size": { @@ -84859,7 +84859,7 @@ } }, { - "accuracy": 0.9970867588723961, + "accuracy": 0.9971634355422697, "total_bits": 911749120, "q_proj": { "group_size": { @@ -84911,7 +84911,7 @@ } }, { - "accuracy": 0.9990212629108053, + "accuracy": 0.9990340371273065, "total_bits": 942718976, "q_proj": { "group_size": { @@ -84963,7 +84963,7 @@ } }, { - "accuracy": 0.9992315131484678, + "accuracy": 0.9992368268339258, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -85017,7 +85017,7 @@ ], "model.layers.43.mlp": [ { - "accuracy": 0.955452749603673, + "accuracy": 0.9555926573903937, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -85069,7 +85069,7 @@ } }, { - "accuracy": 0.9565560002075999, + "accuracy": 0.9566950327471683, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -85121,7 +85121,7 @@ } }, { - "accuracy": 0.962504157894536, + "accuracy": 0.9626272916793823, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -85170,7 +85170,7 @@ } }, { - "accuracy": 0.9642100208684018, + "accuracy": 0.9643297665997556, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -85219,7 +85219,7 @@ } }, { - "accuracy": 0.9775807277152413, + "accuracy": 0.9776488385702434, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -85271,7 +85271,7 @@ } }, { - "accuracy": 0.979320753561823, + "accuracy": 0.9793817055852789, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -85323,7 +85323,7 @@ } }, { - "accuracy": 0.9818057574723896, + "accuracy": 0.9818583648455771, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -85372,7 +85372,7 @@ } }, { - "accuracy": 0.9885394996718356, + "accuracy": 0.9885728665088352, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -85415,7 +85415,7 @@ } }, { - "accuracy": 0.9894793394364809, + "accuracy": 0.9895127603882238, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -85458,7 +85458,7 @@ } }, { - "accuracy": 0.9886156587224257, + "accuracy": 0.988645151257515, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -85510,7 +85510,7 @@ } }, { - "accuracy": 0.9899683163354271, + "accuracy": 0.9899991335053193, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -85562,7 +85562,7 @@ } }, { - "accuracy": 0.9942086578199738, + "accuracy": 0.9942239966047438, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -85614,7 +85614,7 @@ } }, { - "accuracy": 0.9950042755195969, + "accuracy": 0.9950194029431594, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -85666,7 +85666,7 @@ } }, { - "accuracy": 0.9969145175265638, + "accuracy": 0.9969231139280295, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -85709,7 +85709,7 @@ } }, { - "accuracy": 0.997044418791407, + "accuracy": 0.9970504519970793, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -85758,7 +85758,7 @@ } }, { - "accuracy": 0.997513013450723, + "accuracy": 0.997517652613552, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -85804,7 +85804,7 @@ } }, { - "accuracy": 0.9991671236133889, + "accuracy": 0.9991692575184923, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -85846,7 +85846,7 @@ ], "model.layers.44.self_attn": [ { - "accuracy": 0.9680112788551732, + "accuracy": 0.965864084268871, "total_bits": 320757760, "q_proj": { "group_size": { @@ -85910,7 +85910,7 @@ } }, { - "accuracy": 0.9687282097967047, + "accuracy": 0.9699607428751493, "total_bits": 329080832, "q_proj": { "group_size": { @@ -85974,7 +85974,7 @@ } }, { - "accuracy": 0.9785719435465964, + "accuracy": 0.978956473501105, "total_bits": 336024576, "q_proj": { "group_size": { @@ -86038,7 +86038,7 @@ } }, { - "accuracy": 0.9810545711140883, + "accuracy": 0.9813963535584902, "total_bits": 401557504, "q_proj": { "group_size": { @@ -86102,7 +86102,7 @@ } }, { - "accuracy": 0.9829524128060592, + "accuracy": 0.9838796132489255, "total_bits": 475279360, "q_proj": { "group_size": { @@ -86166,7 +86166,7 @@ } }, { - "accuracy": 0.9852537999027654, + "accuracy": 0.984551191329956, "total_bits": 475479040, "q_proj": { "group_size": { @@ -86230,7 +86230,7 @@ } }, { - "accuracy": 0.9898859168353834, + "accuracy": 0.9900776375281183, "total_bits": 609759232, "q_proj": { "group_size": { @@ -86282,7 +86282,7 @@ } }, { - "accuracy": 0.9903467477936494, + "accuracy": 0.9901808147367678, "total_bits": 610024448, "q_proj": { "group_size": { @@ -86334,7 +86334,7 @@ } }, { - "accuracy": 0.9913088161694376, + "accuracy": 0.9918584478528876, "total_bits": 615020544, "q_proj": { "group_size": { @@ -86386,7 +86386,7 @@ } }, { - "accuracy": 0.9914139587628213, + "accuracy": 0.9919776563581667, "total_bits": 623951872, "q_proj": { "group_size": { @@ -86438,7 +86438,7 @@ } }, { - "accuracy": 0.9913108937050167, + "accuracy": 0.9914897680282593, "total_bits": 626473984, "q_proj": { "group_size": { @@ -86502,7 +86502,7 @@ } }, { - "accuracy": 0.9921536484831258, + "accuracy": 0.9921815348298926, "total_bits": 630355968, "q_proj": { "group_size": { @@ -86566,7 +86566,7 @@ } }, { - "accuracy": 0.994212055284726, + "accuracy": 0.9942563082042494, "total_bits": 637362176, "q_proj": { "group_size": { @@ -86627,7 +86627,7 @@ } }, { - "accuracy": 0.9947001181150738, + "accuracy": 0.9948092514747068, "total_bits": 646823936, "q_proj": { "group_size": { @@ -86688,7 +86688,7 @@ } }, { - "accuracy": 0.9966953298763225, + "accuracy": 0.996856768076357, "total_bits": 784740352, "q_proj": { "group_size": { @@ -86749,7 +86749,7 @@ } }, { - "accuracy": 0.9972686649937379, + "accuracy": 0.9973448271814146, "total_bits": 797818880, "q_proj": { "group_size": { @@ -86810,7 +86810,7 @@ } }, { - "accuracy": 0.9970714932209567, + "accuracy": 0.9972651547899372, "total_bits": 911749120, "q_proj": { "group_size": { @@ -86862,7 +86862,7 @@ } }, { - "accuracy": 0.9989672528677865, + "accuracy": 0.9989589220403057, "total_bits": 942718976, "q_proj": { "group_size": { @@ -86914,7 +86914,7 @@ } }, { - "accuracy": 0.999275085565291, + "accuracy": 0.9992830470988625, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -86968,7 +86968,7 @@ ], "model.layers.44.mlp": [ { - "accuracy": 0.9544078927291066, + "accuracy": 0.954548788698096, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -87020,7 +87020,7 @@ } }, { - "accuracy": 0.9555460622436122, + "accuracy": 0.9556845395188582, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -87072,7 +87072,7 @@ } }, { - "accuracy": 0.9616077385450664, + "accuracy": 0.961734087843644, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -87121,7 +87121,7 @@ } }, { - "accuracy": 0.9633574297553614, + "accuracy": 0.9634770876482913, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -87170,7 +87170,7 @@ } }, { - "accuracy": 0.9770243465900421, + "accuracy": 0.9771038607547158, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -87222,7 +87222,7 @@ } }, { - "accuracy": 0.9788118742014232, + "accuracy": 0.9788765609264374, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -87274,7 +87274,7 @@ } }, { - "accuracy": 0.981354147195816, + "accuracy": 0.9814123417201795, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -87323,7 +87323,7 @@ } }, { - "accuracy": 0.9882442135559885, + "accuracy": 0.9882860967987462, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -87366,7 +87366,7 @@ } }, { - "accuracy": 0.9892079838012394, + "accuracy": 0.9892434783672032, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -87409,7 +87409,7 @@ } }, { - "accuracy": 0.9883248860898771, + "accuracy": 0.9883666634559631, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -87461,7 +87461,7 @@ } }, { - "accuracy": 0.989712738677075, + "accuracy": 0.9897472105528179, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -87513,7 +87513,7 @@ } }, { - "accuracy": 0.9940560252258652, + "accuracy": 0.9940802478476575, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -87565,7 +87565,7 @@ } }, { - "accuracy": 0.9948740134897985, + "accuracy": 0.9948915113743982, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -87617,7 +87617,7 @@ } }, { - "accuracy": 0.9968295554189306, + "accuracy": 0.9968416575146349, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -87660,7 +87660,7 @@ } }, { - "accuracy": 0.9969698347543415, + "accuracy": 0.9969812417893034, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -87709,7 +87709,7 @@ } }, { - "accuracy": 0.9974509516828939, + "accuracy": 0.9974608072324803, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -87755,7 +87755,7 @@ } }, { - "accuracy": 0.9991566138832193, + "accuracy": 0.9991598689046345, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -87797,7 +87797,7 @@ ], "model.layers.45.self_attn": [ { - "accuracy": 0.9812730738991186, + "accuracy": 0.9817799392499422, "total_bits": 320757760, "q_proj": { "group_size": { @@ -87861,7 +87861,7 @@ } }, { - "accuracy": 0.9828665711377796, + "accuracy": 0.9828835474817377, "total_bits": 329080832, "q_proj": { "group_size": { @@ -87925,7 +87925,7 @@ } }, { - "accuracy": 0.9852222988480016, + "accuracy": 0.9854413757198736, "total_bits": 336024576, "q_proj": { "group_size": { @@ -87989,7 +87989,7 @@ } }, { - "accuracy": 0.9882442441425825, + "accuracy": 0.9884977113259467, "total_bits": 401557504, "q_proj": { "group_size": { @@ -88053,7 +88053,7 @@ } }, { - "accuracy": 0.9895345192206534, + "accuracy": 0.9898189478798917, "total_bits": 475279360, "q_proj": { "group_size": { @@ -88117,7 +88117,7 @@ } }, { - "accuracy": 0.9901455334926906, + "accuracy": 0.9905303214725695, "total_bits": 475479040, "q_proj": { "group_size": { @@ -88181,7 +88181,7 @@ } }, { - "accuracy": 0.9920123078321156, + "accuracy": 0.991933872825221, "total_bits": 609759232, "q_proj": { "group_size": { @@ -88233,7 +88233,7 @@ } }, { - "accuracy": 0.9928456662516845, + "accuracy": 0.9931696974917462, "total_bits": 610024448, "q_proj": { "group_size": { @@ -88285,7 +88285,7 @@ } }, { - "accuracy": 0.9933516775306902, + "accuracy": 0.9935774665914083, "total_bits": 615020544, "q_proj": { "group_size": { @@ -88337,7 +88337,7 @@ } }, { - "accuracy": 0.9934766657258335, + "accuracy": 0.9937034351261038, "total_bits": 623951872, "q_proj": { "group_size": { @@ -88389,7 +88389,7 @@ } }, { - "accuracy": 0.9933554781110663, + "accuracy": 0.9938189316737024, "total_bits": 626473984, "q_proj": { "group_size": { @@ -88453,7 +88453,7 @@ } }, { - "accuracy": 0.9940603524446487, + "accuracy": 0.9940546327515652, "total_bits": 630355968, "q_proj": { "group_size": { @@ -88517,7 +88517,7 @@ } }, { - "accuracy": 0.9948862461667312, + "accuracy": 0.9949656407299795, "total_bits": 637362176, "q_proj": { "group_size": { @@ -88578,7 +88578,7 @@ } }, { - "accuracy": 0.9952902770356128, + "accuracy": 0.9953373490195525, "total_bits": 646823936, "q_proj": { "group_size": { @@ -88639,7 +88639,7 @@ } }, { - "accuracy": 0.9969850572708406, + "accuracy": 0.9970011617008009, "total_bits": 784740352, "q_proj": { "group_size": { @@ -88700,7 +88700,7 @@ } }, { - "accuracy": 0.9974588514550736, + "accuracy": 0.9974759191666779, "total_bits": 797818880, "q_proj": { "group_size": { @@ -88761,7 +88761,7 @@ } }, { - "accuracy": 0.9972709993782797, + "accuracy": 0.9972833410689705, "total_bits": 911749120, "q_proj": { "group_size": { @@ -88813,7 +88813,7 @@ } }, { - "accuracy": 0.9990287961340264, + "accuracy": 0.9990332597180417, "total_bits": 942718976, "q_proj": { "group_size": { @@ -88865,7 +88865,7 @@ } }, { - "accuracy": 0.9992568209571274, + "accuracy": 0.9992769747776421, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -88919,7 +88919,7 @@ ], "model.layers.45.mlp": [ { - "accuracy": 0.9544578759293807, + "accuracy": 0.9546195801935697, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -88971,7 +88971,7 @@ } }, { - "accuracy": 0.9555876098181072, + "accuracy": 0.9557548228063082, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -89023,7 +89023,7 @@ } }, { - "accuracy": 0.9615707460202669, + "accuracy": 0.9617087778292204, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -89072,7 +89072,7 @@ } }, { - "accuracy": 0.963280439376831, + "accuracy": 0.9634070019972951, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -89121,7 +89121,7 @@ } }, { - "accuracy": 0.9770386203339225, + "accuracy": 0.9771197548038081, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -89173,7 +89173,7 @@ } }, { - "accuracy": 0.9788267957536798, + "accuracy": 0.9789067710700788, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -89225,7 +89225,7 @@ } }, { - "accuracy": 0.9813215403180373, + "accuracy": 0.981391495779941, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -89274,7 +89274,7 @@ } }, { - "accuracy": 0.9882562184020093, + "accuracy": 0.9882945818336386, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -89317,7 +89317,7 @@ } }, { - "accuracy": 0.989216076700311, + "accuracy": 0.9892575881983104, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -89360,7 +89360,7 @@ } }, { - "accuracy": 0.9883247119815726, + "accuracy": 0.9883646172912497, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -89412,7 +89412,7 @@ } }, { - "accuracy": 0.9897141723256362, + "accuracy": 0.9897562045800058, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -89464,7 +89464,7 @@ } }, { - "accuracy": 0.9940549323433324, + "accuracy": 0.9940777817054799, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -89516,7 +89516,7 @@ } }, { - "accuracy": 0.994872017910606, + "accuracy": 0.9948932159888116, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -89568,7 +89568,7 @@ } }, { - "accuracy": 0.9968240682623888, + "accuracy": 0.9968350502221208, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -89611,7 +89611,7 @@ } }, { - "accuracy": 0.9969552124017164, + "accuracy": 0.9969683524809385, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -89660,7 +89660,7 @@ } }, { - "accuracy": 0.9974209638802629, + "accuracy": 0.9974329448059985, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -89706,7 +89706,7 @@ } }, { - "accuracy": 0.9991107241887796, + "accuracy": 0.9991138702944705, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -89748,7 +89748,7 @@ ], "model.layers.46.self_attn": [ { - "accuracy": 0.990240332327391, + "accuracy": 0.9903120978882438, "total_bits": 320757760, "q_proj": { "group_size": { @@ -89812,7 +89812,7 @@ } }, { - "accuracy": 0.990737110376358, + "accuracy": 0.9908270200616435, "total_bits": 329080832, "q_proj": { "group_size": { @@ -89876,7 +89876,7 @@ } }, { - "accuracy": 0.9921460300683975, + "accuracy": 0.9920748291831267, "total_bits": 336024576, "q_proj": { "group_size": { @@ -89940,7 +89940,7 @@ } }, { - "accuracy": 0.9935857641853785, + "accuracy": 0.9934694249379007, "total_bits": 401557504, "q_proj": { "group_size": { @@ -90004,7 +90004,7 @@ } }, { - "accuracy": 0.9940403154806087, + "accuracy": 0.9939531154538456, "total_bits": 475279360, "q_proj": { "group_size": { @@ -90068,7 +90068,7 @@ } }, { - "accuracy": 0.99476161246237, + "accuracy": 0.9946123910577673, "total_bits": 475479040, "q_proj": { "group_size": { @@ -90132,7 +90132,7 @@ } }, { - "accuracy": 0.9951837258903604, + "accuracy": 0.9950732094676871, "total_bits": 609759232, "q_proj": { "group_size": { @@ -90184,7 +90184,7 @@ } }, { - "accuracy": 0.9960623842320944, + "accuracy": 0.9958403784977762, "total_bits": 610024448, "q_proj": { "group_size": { @@ -90236,7 +90236,7 @@ } }, { - "accuracy": 0.996179859889181, + "accuracy": 0.9962053102882285, "total_bits": 615020544, "q_proj": { "group_size": { @@ -90288,7 +90288,7 @@ } }, { - "accuracy": 0.9962137584623537, + "accuracy": 0.9962444583836355, "total_bits": 623951872, "q_proj": { "group_size": { @@ -90340,7 +90340,7 @@ } }, { - "accuracy": 0.996590887245379, + "accuracy": 0.9966770469358093, "total_bits": 626473984, "q_proj": { "group_size": { @@ -90404,7 +90404,7 @@ } }, { - "accuracy": 0.9969367124140263, + "accuracy": 0.9969203789767466, "total_bits": 630355968, "q_proj": { "group_size": { @@ -90468,7 +90468,7 @@ } }, { - "accuracy": 0.996888301678394, + "accuracy": 0.9970059308566546, "total_bits": 637362176, "q_proj": { "group_size": { @@ -90529,7 +90529,7 @@ } }, { - "accuracy": 0.9972432428284695, + "accuracy": 0.9972261110026586, "total_bits": 646823936, "q_proj": { "group_size": { @@ -90590,7 +90590,7 @@ } }, { - "accuracy": 0.9978194234794692, + "accuracy": 0.9978140814131812, "total_bits": 784740352, "q_proj": { "group_size": { @@ -90651,7 +90651,7 @@ } }, { - "accuracy": 0.9983133889342609, + "accuracy": 0.9982939180182783, "total_bits": 797818880, "q_proj": { "group_size": { @@ -90712,7 +90712,7 @@ } }, { - "accuracy": 0.9978983325785712, + "accuracy": 0.9978886332951094, "total_bits": 911749120, "q_proj": { "group_size": { @@ -90764,7 +90764,7 @@ } }, { - "accuracy": 0.9993966417877298, + "accuracy": 0.9994104155584386, "total_bits": 942718976, "q_proj": { "group_size": { @@ -90816,7 +90816,7 @@ } }, { - "accuracy": 0.9993786001577973, + "accuracy": 0.99938335181459, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -90870,7 +90870,7 @@ ], "model.layers.46.mlp": [ { - "accuracy": 0.9545336114732843, + "accuracy": 0.9547016652006852, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -90922,7 +90922,7 @@ } }, { - "accuracy": 0.9556515844244706, + "accuracy": 0.9558164916540447, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -90974,7 +90974,7 @@ } }, { - "accuracy": 0.9616160455502962, + "accuracy": 0.9617605115237989, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -91023,7 +91023,7 @@ } }, { - "accuracy": 0.9632922850157085, + "accuracy": 0.9634271954235277, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -91072,7 +91072,7 @@ } }, { - "accuracy": 0.9770675536833311, + "accuracy": 0.9771536714152286, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -91124,7 +91124,7 @@ } }, { - "accuracy": 0.9788577948745928, + "accuracy": 0.9789396038180903, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -91176,7 +91176,7 @@ } }, { - "accuracy": 0.9813361905123058, + "accuracy": 0.9814090681703467, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -91225,7 +91225,7 @@ } }, { - "accuracy": 0.9882913090680775, + "accuracy": 0.9883333013245934, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -91268,7 +91268,7 @@ } }, { - "accuracy": 0.9892481432149285, + "accuracy": 0.989288395956943, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -91311,7 +91311,7 @@ } }, { - "accuracy": 0.9883350353491934, + "accuracy": 0.9883821583107898, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -91363,7 +91363,7 @@ } }, { - "accuracy": 0.9897306255604091, + "accuracy": 0.9897703614674116, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -91415,7 +91415,7 @@ } }, { - "accuracy": 0.9940647620119547, + "accuracy": 0.9940861832154425, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -91467,7 +91467,7 @@ } }, { - "accuracy": 0.9948862830274984, + "accuracy": 0.9949047236066115, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -91519,7 +91519,7 @@ } }, { - "accuracy": 0.9968435997633558, + "accuracy": 0.9968556016683578, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -91562,7 +91562,7 @@ } }, { - "accuracy": 0.996969203219602, + "accuracy": 0.9969811057181734, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -91611,7 +91611,7 @@ } }, { - "accuracy": 0.9974334239959717, + "accuracy": 0.9974434952202597, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -91657,7 +91657,7 @@ } }, { - "accuracy": 0.999161279907352, + "accuracy": 0.9991642740604124, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -91699,7 +91699,7 @@ ], "model.layers.47.self_attn": [ { - "accuracy": 0.9718637089980277, + "accuracy": 0.9696962864775407, "total_bits": 320757760, "q_proj": { "group_size": { @@ -91763,7 +91763,7 @@ } }, { - "accuracy": 0.9734510528413873, + "accuracy": 0.9714628238427011, "total_bits": 329080832, "q_proj": { "group_size": { @@ -91827,7 +91827,7 @@ } }, { - "accuracy": 0.9827904795345507, + "accuracy": 0.9826132153209887, "total_bits": 336024576, "q_proj": { "group_size": { @@ -91891,7 +91891,7 @@ } }, { - "accuracy": 0.9843323732677259, + "accuracy": 0.9845148858271147, "total_bits": 401557504, "q_proj": { "group_size": { @@ -91955,7 +91955,7 @@ } }, { - "accuracy": 0.9854924537633595, + "accuracy": 0.985698042731536, "total_bits": 475279360, "q_proj": { "group_size": { @@ -92019,7 +92019,7 @@ } }, { - "accuracy": 0.9852719008922577, + "accuracy": 0.9864465186470434, "total_bits": 475479040, "q_proj": { "group_size": { @@ -92083,7 +92083,7 @@ } }, { - "accuracy": 0.9908731666050459, + "accuracy": 0.9909787185882267, "total_bits": 609759232, "q_proj": { "group_size": { @@ -92135,7 +92135,7 @@ } }, { - "accuracy": 0.9914459225378538, + "accuracy": 0.9912735600220529, "total_bits": 610024448, "q_proj": { "group_size": { @@ -92187,7 +92187,7 @@ } }, { - "accuracy": 0.9919471905419701, + "accuracy": 0.9919459835479134, "total_bits": 615020544, "q_proj": { "group_size": { @@ -92239,7 +92239,7 @@ } }, { - "accuracy": 0.9920425407196346, + "accuracy": 0.9920285799001393, "total_bits": 623951872, "q_proj": { "group_size": { @@ -92291,7 +92291,7 @@ } }, { - "accuracy": 0.9918935142065349, + "accuracy": 0.992218764204728, "total_bits": 626473984, "q_proj": { "group_size": { @@ -92355,7 +92355,7 @@ } }, { - "accuracy": 0.9927697385612287, + "accuracy": 0.992703168015731, "total_bits": 630355968, "q_proj": { "group_size": { @@ -92419,7 +92419,7 @@ } }, { - "accuracy": 0.9948119560354635, + "accuracy": 0.9948346403084303, "total_bits": 637362176, "q_proj": { "group_size": { @@ -92480,7 +92480,7 @@ } }, { - "accuracy": 0.9952459056910715, + "accuracy": 0.9952432693619477, "total_bits": 646823936, "q_proj": { "group_size": { @@ -92541,7 +92541,7 @@ } }, { - "accuracy": 0.9971814398702822, + "accuracy": 0.9971037799198377, "total_bits": 784740352, "q_proj": { "group_size": { @@ -92602,7 +92602,7 @@ } }, { - "accuracy": 0.9976436382061556, + "accuracy": 0.9976299071782514, "total_bits": 797818880, "q_proj": { "group_size": { @@ -92663,7 +92663,7 @@ } }, { - "accuracy": 0.9973358070771945, + "accuracy": 0.9972570342452902, "total_bits": 911749120, "q_proj": { "group_size": { @@ -92715,7 +92715,7 @@ } }, { - "accuracy": 0.9990714648248333, + "accuracy": 0.9990967536639226, "total_bits": 942718976, "q_proj": { "group_size": { @@ -92767,7 +92767,7 @@ } }, { - "accuracy": 0.999295882488552, + "accuracy": 0.9992793574929237, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -92821,7 +92821,7 @@ ], "model.layers.47.mlp": [ { - "accuracy": 0.9535999015757912, + "accuracy": 0.953785501028362, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -92873,7 +92873,7 @@ } }, { - "accuracy": 0.9547503621954667, + "accuracy": 0.9549355663751301, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -92925,7 +92925,7 @@ } }, { - "accuracy": 0.9607919015382466, + "accuracy": 0.9609503589178386, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -92974,7 +92974,7 @@ } }, { - "accuracy": 0.9624981660591929, + "accuracy": 0.962649709299991, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -93023,7 +93023,7 @@ } }, { - "accuracy": 0.9765963962203578, + "accuracy": 0.9766870517479745, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -93075,7 +93075,7 @@ } }, { - "accuracy": 0.9784212253595653, + "accuracy": 0.9785057761167225, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -93127,7 +93127,7 @@ } }, { - "accuracy": 0.9809415293367285, + "accuracy": 0.9810172240985068, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -93176,7 +93176,7 @@ } }, { - "accuracy": 0.9880472907894536, + "accuracy": 0.9880919417268351, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -93219,7 +93219,7 @@ } }, { - "accuracy": 0.9890272178147969, + "accuracy": 0.9890702029592112, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -93262,7 +93262,7 @@ } }, { - "accuracy": 0.9881063189945722, + "accuracy": 0.9881541964254881, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -93314,7 +93314,7 @@ } }, { - "accuracy": 0.9895285634618056, + "accuracy": 0.9895665143665514, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -93366,7 +93366,7 @@ } }, { - "accuracy": 0.9939479314182934, + "accuracy": 0.9939703376669633, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -93418,7 +93418,7 @@ } }, { - "accuracy": 0.9947802157778489, + "accuracy": 0.9948004433198979, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -93470,7 +93470,7 @@ } }, { - "accuracy": 0.9967759200616887, + "accuracy": 0.9967878927525721, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -93513,7 +93513,7 @@ } }, { - "accuracy": 0.9968973531534797, + "accuracy": 0.9969084878501139, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -93562,7 +93562,7 @@ } }, { - "accuracy": 0.9973662777950889, + "accuracy": 0.9973760543293074, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -93608,7 +93608,7 @@ } }, { - "accuracy": 0.9990876035666779, + "accuracy": 0.9990916582510659, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -93650,7 +93650,7 @@ ], "model.layers.48.self_attn": [ { - "accuracy": 0.9707853856839632, + "accuracy": 0.973650618603355, "total_bits": 320757760, "q_proj": { "group_size": { @@ -93714,7 +93714,7 @@ } }, { - "accuracy": 0.9753189730016809, + "accuracy": 0.9738533104720869, "total_bits": 329080832, "q_proj": { "group_size": { @@ -93778,7 +93778,7 @@ } }, { - "accuracy": 0.9830662372865175, + "accuracy": 0.9840000303168046, "total_bits": 336024576, "q_proj": { "group_size": { @@ -93842,7 +93842,7 @@ } }, { - "accuracy": 0.984518046441831, + "accuracy": 0.9853309675266868, "total_bits": 401557504, "q_proj": { "group_size": { @@ -93906,7 +93906,7 @@ } }, { - "accuracy": 0.9852773719712308, + "accuracy": 0.9861489989255604, "total_bits": 475279360, "q_proj": { "group_size": { @@ -93970,7 +93970,7 @@ } }, { - "accuracy": 0.9870115377401051, + "accuracy": 0.9858287949311105, "total_bits": 475479040, "q_proj": { "group_size": { @@ -94034,7 +94034,7 @@ } }, { - "accuracy": 0.9906641535068813, + "accuracy": 0.9897834164531607, "total_bits": 609759232, "q_proj": { "group_size": { @@ -94086,7 +94086,7 @@ } }, { - "accuracy": 0.9910827922193628, + "accuracy": 0.9905844785665211, "total_bits": 610024448, "q_proj": { "group_size": { @@ -94138,7 +94138,7 @@ } }, { - "accuracy": 0.991862311174995, + "accuracy": 0.9917805971283662, "total_bits": 615020544, "q_proj": { "group_size": { @@ -94190,7 +94190,7 @@ } }, { - "accuracy": 0.9919349574729016, + "accuracy": 0.9918425420397207, "total_bits": 623951872, "q_proj": { "group_size": { @@ -94242,7 +94242,7 @@ } }, { - "accuracy": 0.9915537042053122, + "accuracy": 0.9917887943355661, "total_bits": 626473984, "q_proj": { "group_size": { @@ -94306,7 +94306,7 @@ } }, { - "accuracy": 0.9923101671432194, + "accuracy": 0.9922218675676145, "total_bits": 630355968, "q_proj": { "group_size": { @@ -94370,7 +94370,7 @@ } }, { - "accuracy": 0.9945805092391214, + "accuracy": 0.9946161547773763, "total_bits": 637362176, "q_proj": { "group_size": { @@ -94431,7 +94431,7 @@ } }, { - "accuracy": 0.9951862637933931, + "accuracy": 0.9952415518070522, "total_bits": 646823936, "q_proj": { "group_size": { @@ -94492,7 +94492,7 @@ } }, { - "accuracy": 0.9969263594401511, + "accuracy": 0.9969497272058537, "total_bits": 784740352, "q_proj": { "group_size": { @@ -94553,7 +94553,7 @@ } }, { - "accuracy": 0.997509024449085, + "accuracy": 0.9974691965862325, "total_bits": 797818880, "q_proj": { "group_size": { @@ -94614,7 +94614,7 @@ } }, { - "accuracy": 0.9970690084523276, + "accuracy": 0.9970825683129462, "total_bits": 911749120, "q_proj": { "group_size": { @@ -94666,7 +94666,7 @@ } }, { - "accuracy": 0.9991375347109217, + "accuracy": 0.9991406175847116, "total_bits": 942718976, "q_proj": { "group_size": { @@ -94718,7 +94718,7 @@ } }, { - "accuracy": 0.99920739895223, + "accuracy": 0.9992407968659934, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -94772,7 +94772,7 @@ ], "model.layers.48.mlp": [ { - "accuracy": 0.9523789568951255, + "accuracy": 0.95257084934335, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -94824,7 +94824,7 @@ } }, { - "accuracy": 0.9535833377587168, + "accuracy": 0.9537530039486132, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -94876,7 +94876,7 @@ } }, { - "accuracy": 0.9597355566526714, + "accuracy": 0.9598884645261263, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -94925,7 +94925,7 @@ } }, { - "accuracy": 0.9614972001627872, + "accuracy": 0.9616413963468451, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -94974,7 +94974,7 @@ } }, { - "accuracy": 0.9759538989318045, + "accuracy": 0.976046842964072, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -95026,7 +95026,7 @@ } }, { - "accuracy": 0.9778455291923723, + "accuracy": 0.9779337268126639, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -95078,7 +95078,7 @@ } }, { - "accuracy": 0.9804322092156661, + "accuracy": 0.9805085690397966, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -95127,7 +95127,7 @@ } }, { - "accuracy": 0.9877064047675383, + "accuracy": 0.9877465728082155, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -95170,7 +95170,7 @@ } }, { - "accuracy": 0.9887156274757887, + "accuracy": 0.988757439349827, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -95213,7 +95213,7 @@ } }, { - "accuracy": 0.9877732905902361, + "accuracy": 0.9878147283667013, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -95265,7 +95265,7 @@ } }, { - "accuracy": 0.9892407773356688, + "accuracy": 0.989278961953364, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -95317,7 +95317,7 @@ } }, { - "accuracy": 0.9937712534477836, + "accuracy": 0.9937894552161819, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -95369,7 +95369,7 @@ } }, { - "accuracy": 0.9946335789404417, + "accuracy": 0.9946530767177281, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -95421,7 +95421,7 @@ } }, { - "accuracy": 0.9966674243149004, + "accuracy": 0.9966767383249182, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -95464,7 +95464,7 @@ } }, { - "accuracy": 0.99680898083668, + "accuracy": 0.9968175886100844, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -95513,7 +95513,7 @@ } }, { - "accuracy": 0.9972923370568376, + "accuracy": 0.9972982073300763, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -95559,7 +95559,7 @@ } }, { - "accuracy": 0.9990622544367063, + "accuracy": 0.9990652780979872, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -95601,7 +95601,7 @@ ], "model.layers.49.self_attn": [ { - "accuracy": 0.9728309010204516, + "accuracy": 0.9736765748576114, "total_bits": 320757760, "q_proj": { "group_size": { @@ -95665,7 +95665,7 @@ } }, { - "accuracy": 0.975760701455568, + "accuracy": 0.976259752323753, "total_bits": 329080832, "q_proj": { "group_size": { @@ -95729,7 +95729,7 @@ } }, { - "accuracy": 0.9838828068030508, + "accuracy": 0.9838466612916243, "total_bits": 336024576, "q_proj": { "group_size": { @@ -95793,7 +95793,7 @@ } }, { - "accuracy": 0.9863561878078863, + "accuracy": 0.9862394081918817, "total_bits": 401557504, "q_proj": { "group_size": { @@ -95857,7 +95857,7 @@ } }, { - "accuracy": 0.9879493634951743, + "accuracy": 0.9878114916776356, "total_bits": 475279360, "q_proj": { "group_size": { @@ -95921,7 +95921,7 @@ } }, { - "accuracy": 0.9866690933704376, + "accuracy": 0.9871745478165778, "total_bits": 475479040, "q_proj": { "group_size": { @@ -95985,7 +95985,7 @@ } }, { - "accuracy": 0.9913150511289898, + "accuracy": 0.9914371943787524, "total_bits": 609759232, "q_proj": { "group_size": { @@ -96037,7 +96037,7 @@ } }, { - "accuracy": 0.991711479268576, + "accuracy": 0.9922066142684535, "total_bits": 610024448, "q_proj": { "group_size": { @@ -96089,7 +96089,7 @@ } }, { - "accuracy": 0.9930442093234313, + "accuracy": 0.9929761690528769, "total_bits": 615020544, "q_proj": { "group_size": { @@ -96141,7 +96141,7 @@ } }, { - "accuracy": 0.9932263085716649, + "accuracy": 0.9931250402801916, "total_bits": 623951872, "q_proj": { "group_size": { @@ -96193,7 +96193,7 @@ } }, { - "accuracy": 0.9931547531956121, + "accuracy": 0.9934837775010812, "total_bits": 626473984, "q_proj": { "group_size": { @@ -96257,7 +96257,7 @@ } }, { - "accuracy": 0.9936989781103636, + "accuracy": 0.9934867463613811, "total_bits": 630355968, "q_proj": { "group_size": { @@ -96321,7 +96321,7 @@ } }, { - "accuracy": 0.995125266282182, + "accuracy": 0.9951163330360463, "total_bits": 637362176, "q_proj": { "group_size": { @@ -96382,7 +96382,7 @@ } }, { - "accuracy": 0.9955175577809936, + "accuracy": 0.9954110650639785, "total_bits": 646823936, "q_proj": { "group_size": { @@ -96443,7 +96443,7 @@ } }, { - "accuracy": 0.9970778669965895, + "accuracy": 0.9970911492250467, "total_bits": 784740352, "q_proj": { "group_size": { @@ -96504,7 +96504,7 @@ } }, { - "accuracy": 0.997561834164356, + "accuracy": 0.9975869839912966, "total_bits": 797818880, "q_proj": { "group_size": { @@ -96565,7 +96565,7 @@ } }, { - "accuracy": 0.9973693288078433, + "accuracy": 0.9973806719246664, "total_bits": 911749120, "q_proj": { "group_size": { @@ -96617,7 +96617,7 @@ } }, { - "accuracy": 0.9990587572714216, + "accuracy": 0.9990796069369504, "total_bits": 942718976, "q_proj": { "group_size": { @@ -96669,7 +96669,7 @@ } }, { - "accuracy": 0.9993275819150241, + "accuracy": 0.9993107529651177, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -96723,7 +96723,7 @@ ], "model.layers.49.mlp": [ { - "accuracy": 0.9522395667276884, + "accuracy": 0.9524390164174532, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -96775,7 +96775,7 @@ } }, { - "accuracy": 0.9534494218073393, + "accuracy": 0.953640740168722, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -96827,7 +96827,7 @@ } }, { - "accuracy": 0.959526611001868, + "accuracy": 0.9596935447893644, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -96876,7 +96876,7 @@ } }, { - "accuracy": 0.9612475759104678, + "accuracy": 0.9614119812061912, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -96925,7 +96925,7 @@ } }, { - "accuracy": 0.9758811953820681, + "accuracy": 0.9759855788005026, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -96977,7 +96977,7 @@ } }, { - "accuracy": 0.9777612246965107, + "accuracy": 0.9778511210491783, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -97029,7 +97029,7 @@ } }, { - "accuracy": 0.9803049062427721, + "accuracy": 0.9803861003173026, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -97078,7 +97078,7 @@ } }, { - "accuracy": 0.987662712210103, + "accuracy": 0.9877126067876816, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -97121,7 +97121,7 @@ } }, { - "accuracy": 0.9886705577373505, + "accuracy": 0.9887164509610126, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -97164,7 +97164,7 @@ } }, { - "accuracy": 0.9877381434566096, + "accuracy": 0.987784857028409, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -97216,7 +97216,7 @@ } }, { - "accuracy": 0.9891960832633471, + "accuracy": 0.9892393993703943, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -97268,7 +97268,7 @@ } }, { - "accuracy": 0.993755920545051, + "accuracy": 0.9937830210516327, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -97320,7 +97320,7 @@ } }, { - "accuracy": 0.9946144795731494, + "accuracy": 0.9946390590385387, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -97372,7 +97372,7 @@ } }, { - "accuracy": 0.9966667067063483, + "accuracy": 0.99668064909546, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -97415,7 +97415,7 @@ } }, { - "accuracy": 0.9967990523498309, + "accuracy": 0.9968127943575382, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -97464,7 +97464,7 @@ } }, { - "accuracy": 0.997273470422155, + "accuracy": 0.9972852260658616, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -97510,7 +97510,7 @@ } }, { - "accuracy": 0.9990534672611638, + "accuracy": 0.9990568863914201, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -97552,7 +97552,7 @@ ], "model.layers.50.self_attn": [ { - "accuracy": 0.9782357513904572, + "accuracy": 0.9789437886915708, "total_bits": 320757760, "q_proj": { "group_size": { @@ -97616,7 +97616,7 @@ } }, { - "accuracy": 0.9801090171462611, + "accuracy": 0.981423756009654, "total_bits": 329080832, "q_proj": { "group_size": { @@ -97680,7 +97680,7 @@ } }, { - "accuracy": 0.986800691014842, + "accuracy": 0.987398018178187, "total_bits": 336024576, "q_proj": { "group_size": { @@ -97744,7 +97744,7 @@ } }, { - "accuracy": 0.98805638441914, + "accuracy": 0.9886414777291449, "total_bits": 401557504, "q_proj": { "group_size": { @@ -97808,7 +97808,7 @@ } }, { - "accuracy": 0.9889158959451475, + "accuracy": 0.9894950366333911, "total_bits": 475279360, "q_proj": { "group_size": { @@ -97872,7 +97872,7 @@ } }, { - "accuracy": 0.9899598650242153, + "accuracy": 0.9908012355628767, "total_bits": 475479040, "q_proj": { "group_size": { @@ -97936,7 +97936,7 @@ } }, { - "accuracy": 0.9924314265188418, + "accuracy": 0.9925088176601812, "total_bits": 609759232, "q_proj": { "group_size": { @@ -97988,7 +97988,7 @@ } }, { - "accuracy": 0.9930633565312937, + "accuracy": 0.9931973399300325, "total_bits": 610024448, "q_proj": { "group_size": { @@ -98040,7 +98040,7 @@ } }, { - "accuracy": 0.9940313724310774, + "accuracy": 0.9936682425047222, "total_bits": 615020544, "q_proj": { "group_size": { @@ -98092,7 +98092,7 @@ } }, { - "accuracy": 0.9941065240847436, + "accuracy": 0.9937166877483067, "total_bits": 623951872, "q_proj": { "group_size": { @@ -98144,7 +98144,7 @@ } }, { - "accuracy": 0.9938154969560472, + "accuracy": 0.9935260815055746, "total_bits": 626473984, "q_proj": { "group_size": { @@ -98208,7 +98208,7 @@ } }, { - "accuracy": 0.9942026016743559, + "accuracy": 0.9940698091921053, "total_bits": 630355968, "q_proj": { "group_size": { @@ -98272,7 +98272,7 @@ } }, { - "accuracy": 0.9956283087008878, + "accuracy": 0.9955692181461736, "total_bits": 637362176, "q_proj": { "group_size": { @@ -98333,7 +98333,7 @@ } }, { - "accuracy": 0.995943325130563, + "accuracy": 0.9960455153333513, "total_bits": 646823936, "q_proj": { "group_size": { @@ -98394,7 +98394,7 @@ } }, { - "accuracy": 0.9974359434686209, + "accuracy": 0.9974517457579312, "total_bits": 784740352, "q_proj": { "group_size": { @@ -98455,7 +98455,7 @@ } }, { - "accuracy": 0.9978666097710007, + "accuracy": 0.9979515612909668, "total_bits": 797818880, "q_proj": { "group_size": { @@ -98516,7 +98516,7 @@ } }, { - "accuracy": 0.9975634029037074, + "accuracy": 0.9975724075185625, "total_bits": 911749120, "q_proj": { "group_size": { @@ -98568,7 +98568,7 @@ } }, { - "accuracy": 0.9992759509600306, + "accuracy": 0.9992771730022995, "total_bits": 942718976, "q_proj": { "group_size": { @@ -98620,7 +98620,7 @@ } }, { - "accuracy": 0.9993463368890317, + "accuracy": 0.9993545348314863, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -98674,7 +98674,7 @@ ], "model.layers.50.mlp": [ { - "accuracy": 0.9522501606690257, + "accuracy": 0.9524460215317576, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -98726,7 +98726,7 @@ } }, { - "accuracy": 0.9534431068520797, + "accuracy": 0.9536399621712535, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -98778,7 +98778,7 @@ } }, { - "accuracy": 0.9595448061039573, + "accuracy": 0.9597147577687314, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -98827,7 +98827,7 @@ } }, { - "accuracy": 0.9612456152313634, + "accuracy": 0.9614091044978091, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -98876,7 +98876,7 @@ } }, { - "accuracy": 0.9758785539551785, + "accuracy": 0.9759870438199294, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -98928,7 +98928,7 @@ } }, { - "accuracy": 0.9777600420148749, + "accuracy": 0.977851927280426, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -98980,7 +98980,7 @@ } }, { - "accuracy": 0.9803020091433274, + "accuracy": 0.9803829977386876, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -99029,7 +99029,7 @@ } }, { - "accuracy": 0.9876872120719207, + "accuracy": 0.9877382650187141, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -99072,7 +99072,7 @@ } }, { - "accuracy": 0.9886843820935801, + "accuracy": 0.9887354828809437, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -99115,7 +99115,7 @@ } }, { - "accuracy": 0.9877362627732126, + "accuracy": 0.9877857228643016, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -99167,7 +99167,7 @@ } }, { - "accuracy": 0.9891910184370843, + "accuracy": 0.9892377131863644, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -99219,7 +99219,7 @@ } }, { - "accuracy": 0.9937606810739166, + "accuracy": 0.9937846260635477, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -99271,7 +99271,7 @@ } }, { - "accuracy": 0.9946162943777285, + "accuracy": 0.9946380626214179, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -99323,7 +99323,7 @@ } }, { - "accuracy": 0.9966823050850316, + "accuracy": 0.9966957176986494, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -99366,7 +99366,7 @@ } }, { - "accuracy": 0.9968126600510195, + "accuracy": 0.9968242747218985, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -99415,7 +99415,7 @@ } }, { - "accuracy": 0.9972866408919033, + "accuracy": 0.9972958837292696, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -99461,7 +99461,7 @@ } }, { - "accuracy": 0.9991182530004727, + "accuracy": 0.9991223654268604, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -99503,7 +99503,7 @@ ], "model.layers.51.self_attn": [ { - "accuracy": 0.9717070491690385, + "accuracy": 0.9756422152644709, "total_bits": 320757760, "q_proj": { "group_size": { @@ -99567,7 +99567,7 @@ } }, { - "accuracy": 0.9768564669709456, + "accuracy": 0.9782204769159618, "total_bits": 329080832, "q_proj": { "group_size": { @@ -99631,7 +99631,7 @@ } }, { - "accuracy": 0.9827174826672203, + "accuracy": 0.9828002060714521, "total_bits": 336024576, "q_proj": { "group_size": { @@ -99695,7 +99695,7 @@ } }, { - "accuracy": 0.9842949650789562, + "accuracy": 0.9844595397773542, "total_bits": 401557504, "q_proj": { "group_size": { @@ -99759,7 +99759,7 @@ } }, { - "accuracy": 0.9856502947054411, + "accuracy": 0.9858658000042564, "total_bits": 475279360, "q_proj": { "group_size": { @@ -99823,7 +99823,7 @@ } }, { - "accuracy": 0.9867192193081504, + "accuracy": 0.9874649432144667, "total_bits": 475479040, "q_proj": { "group_size": { @@ -99887,7 +99887,7 @@ } }, { - "accuracy": 0.9900217056274414, + "accuracy": 0.9909046927565023, "total_bits": 609759232, "q_proj": { "group_size": { @@ -99939,7 +99939,7 @@ } }, { - "accuracy": 0.9912679360101098, + "accuracy": 0.991140952235774, "total_bits": 610024448, "q_proj": { "group_size": { @@ -99991,7 +99991,7 @@ } }, { - "accuracy": 0.992065512820294, + "accuracy": 0.9922617509176856, "total_bits": 615020544, "q_proj": { "group_size": { @@ -100043,7 +100043,7 @@ } }, { - "accuracy": 0.9921534610422034, + "accuracy": 0.9923441661031622, "total_bits": 623951872, "q_proj": { "group_size": { @@ -100095,7 +100095,7 @@ } }, { - "accuracy": 0.9918215690474761, + "accuracy": 0.9923028365561837, "total_bits": 626473984, "q_proj": { "group_size": { @@ -100159,7 +100159,7 @@ } }, { - "accuracy": 0.9925132182083631, + "accuracy": 0.9922885863404525, "total_bits": 630355968, "q_proj": { "group_size": { @@ -100223,7 +100223,7 @@ } }, { - "accuracy": 0.9944439569586202, + "accuracy": 0.9943990683869312, "total_bits": 637362176, "q_proj": { "group_size": { @@ -100284,7 +100284,7 @@ } }, { - "accuracy": 0.9947702359212073, + "accuracy": 0.9949361274116918, "total_bits": 646823936, "q_proj": { "group_size": { @@ -100345,7 +100345,7 @@ } }, { - "accuracy": 0.996866859103504, + "accuracy": 0.9968350237529529, "total_bits": 784740352, "q_proj": { "group_size": { @@ -100406,7 +100406,7 @@ } }, { - "accuracy": 0.9973240322188327, + "accuracy": 0.9973576404154301, "total_bits": 797818880, "q_proj": { "group_size": { @@ -100467,7 +100467,7 @@ } }, { - "accuracy": 0.9970542247358122, + "accuracy": 0.997004064486215, "total_bits": 911749120, "q_proj": { "group_size": { @@ -100519,7 +100519,7 @@ } }, { - "accuracy": 0.9990589809849074, + "accuracy": 0.9990761095755979, "total_bits": 942718976, "q_proj": { "group_size": { @@ -100571,7 +100571,7 @@ } }, { - "accuracy": 0.9992085459985232, + "accuracy": 0.9992227229828897, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -100625,7 +100625,7 @@ ], "model.layers.51.mlp": [ { - "accuracy": 0.9515406458001388, + "accuracy": 0.951741237389414, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -100677,7 +100677,7 @@ } }, { - "accuracy": 0.9527595890195746, + "accuracy": 0.9529588849920976, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -100729,7 +100729,7 @@ } }, { - "accuracy": 0.9589149951934814, + "accuracy": 0.959081768989563, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -100778,7 +100778,7 @@ } }, { - "accuracy": 0.9606449666776156, + "accuracy": 0.9608025362617091, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -100827,7 +100827,7 @@ } }, { - "accuracy": 0.9755138833271829, + "accuracy": 0.975617484042519, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -100879,7 +100879,7 @@ } }, { - "accuracy": 0.977427507701673, + "accuracy": 0.9775224986829256, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -100931,7 +100931,7 @@ } }, { - "accuracy": 0.9799997273244356, + "accuracy": 0.9800845008147391, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -100980,7 +100980,7 @@ } }, { - "accuracy": 0.9874737121556935, + "accuracy": 0.9875276661232898, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -101023,7 +101023,7 @@ } }, { - "accuracy": 0.9884968823508212, + "accuracy": 0.9885453346528505, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -101066,7 +101066,7 @@ } }, { - "accuracy": 0.987537892241227, + "accuracy": 0.98759148152251, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -101118,7 +101118,7 @@ } }, { - "accuracy": 0.9890196001843402, + "accuracy": 0.9890686916677576, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -101170,7 +101170,7 @@ } }, { - "accuracy": 0.9936509708824911, + "accuracy": 0.9936767075406877, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -101222,7 +101222,7 @@ } }, { - "accuracy": 0.9945234958278505, + "accuracy": 0.9945478878523174, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -101274,7 +101274,7 @@ } }, { - "accuracy": 0.9966063895507863, + "accuracy": 0.9966229216048592, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -101317,7 +101317,7 @@ } }, { - "accuracy": 0.9967570918563166, + "accuracy": 0.9967714740257514, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -101366,7 +101366,7 @@ } }, { - "accuracy": 0.9972379768365308, + "accuracy": 0.9972504245999613, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -101412,7 +101412,7 @@ } }, { - "accuracy": 0.9990856668078586, + "accuracy": 0.9990904725303775, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -101454,7 +101454,7 @@ ], "model.layers.52.self_attn": [ { - "accuracy": 0.9616607584451374, + "accuracy": 0.9632019337854887, "total_bits": 320757760, "q_proj": { "group_size": { @@ -101518,7 +101518,7 @@ } }, { - "accuracy": 0.9654575900027627, + "accuracy": 0.9652874595240543, "total_bits": 329080832, "q_proj": { "group_size": { @@ -101582,7 +101582,7 @@ } }, { - "accuracy": 0.9760926381537789, + "accuracy": 0.9764390509379538, "total_bits": 336024576, "q_proj": { "group_size": { @@ -101646,7 +101646,7 @@ } }, { - "accuracy": 0.9795675763958379, + "accuracy": 0.979779651290492, "total_bits": 401557504, "q_proj": { "group_size": { @@ -101710,7 +101710,7 @@ } }, { - "accuracy": 0.9827596639332018, + "accuracy": 0.9828834314095346, "total_bits": 475279360, "q_proj": { "group_size": { @@ -101774,7 +101774,7 @@ } }, { - "accuracy": 0.9828241106710935, + "accuracy": 0.9834098486523879, "total_bits": 475479040, "q_proj": { "group_size": { @@ -101838,7 +101838,7 @@ } }, { - "accuracy": 0.9893668207683062, + "accuracy": 0.9898819829288282, "total_bits": 609759232, "q_proj": { "group_size": { @@ -101890,7 +101890,7 @@ } }, { - "accuracy": 0.9903941468188637, + "accuracy": 0.9900930116051122, "total_bits": 610024448, "q_proj": { "group_size": { @@ -101942,7 +101942,7 @@ } }, { - "accuracy": 0.9909701112069582, + "accuracy": 0.9910328333315096, "total_bits": 615020544, "q_proj": { "group_size": { @@ -101994,7 +101994,7 @@ } }, { - "accuracy": 0.9912358051852176, + "accuracy": 0.9912514043481726, "total_bits": 623951872, "q_proj": { "group_size": { @@ -102046,7 +102046,7 @@ } }, { - "accuracy": 0.990916274880108, + "accuracy": 0.9910939476991955, "total_bits": 626473984, "q_proj": { "group_size": { @@ -102110,7 +102110,7 @@ } }, { - "accuracy": 0.9917350215347189, + "accuracy": 0.9915816525095388, "total_bits": 630355968, "q_proj": { "group_size": { @@ -102174,7 +102174,7 @@ } }, { - "accuracy": 0.9938088365291294, + "accuracy": 0.9939057223106685, "total_bits": 637362176, "q_proj": { "group_size": { @@ -102235,7 +102235,7 @@ } }, { - "accuracy": 0.994367789281042, + "accuracy": 0.9943429012047617, "total_bits": 646823936, "q_proj": { "group_size": { @@ -102296,7 +102296,7 @@ } }, { - "accuracy": 0.9966513490990588, + "accuracy": 0.996669132850672, "total_bits": 784740352, "q_proj": { "group_size": { @@ -102357,7 +102357,7 @@ } }, { - "accuracy": 0.9970847958404767, + "accuracy": 0.997193737837829, "total_bits": 797818880, "q_proj": { "group_size": { @@ -102418,7 +102418,7 @@ } }, { - "accuracy": 0.9971992344056305, + "accuracy": 0.9972291496631346, "total_bits": 911749120, "q_proj": { "group_size": { @@ -102470,7 +102470,7 @@ } }, { - "accuracy": 0.9986063950744114, + "accuracy": 0.9986110029644087, "total_bits": 942718976, "q_proj": { "group_size": { @@ -102522,7 +102522,7 @@ } }, { - "accuracy": 0.9992447471932361, + "accuracy": 0.9992610874438757, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -102576,7 +102576,7 @@ ], "model.layers.52.mlp": [ { - "accuracy": 0.9510125768812079, + "accuracy": 0.9512310749606082, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -102628,7 +102628,7 @@ } }, { - "accuracy": 0.9522602118943867, + "accuracy": 0.9524676611548976, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -102680,7 +102680,7 @@ } }, { - "accuracy": 0.9584509761709916, + "accuracy": 0.9586491898486489, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -102729,7 +102729,7 @@ } }, { - "accuracy": 0.9602161049842834, + "accuracy": 0.9604067331866214, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -102778,7 +102778,7 @@ } }, { - "accuracy": 0.9752470945057116, + "accuracy": 0.9753568486163491, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -102830,7 +102830,7 @@ } }, { - "accuracy": 0.9771770301618075, + "accuracy": 0.9772829585953763, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -102882,7 +102882,7 @@ } }, { - "accuracy": 0.9797813170834592, + "accuracy": 0.9798761841497923, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -102931,7 +102931,7 @@ } }, { - "accuracy": 0.9873286266075937, + "accuracy": 0.9873861443055304, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -102974,7 +102974,7 @@ } }, { - "accuracy": 0.9883735509295213, + "accuracy": 0.988423324729267, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -103017,7 +103017,7 @@ } }, { - "accuracy": 0.987412804835721, + "accuracy": 0.9874666607693622, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -103069,7 +103069,7 @@ } }, { - "accuracy": 0.9889202345358697, + "accuracy": 0.9889684617519379, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -103121,7 +103121,7 @@ } }, { - "accuracy": 0.9935879730864575, + "accuracy": 0.9936172958267363, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -103173,7 +103173,7 @@ } }, { - "accuracy": 0.9944735420377631, + "accuracy": 0.9944977336808255, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -103225,7 +103225,7 @@ } }, { - "accuracy": 0.9965662513124315, + "accuracy": 0.9965833856871253, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -103268,7 +103268,7 @@ } }, { - "accuracy": 0.9967106153306208, + "accuracy": 0.9967253594414184, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -103317,7 +103317,7 @@ } }, { - "accuracy": 0.9971978521268619, + "accuracy": 0.9972105549746438, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -103363,7 +103363,7 @@ } }, { - "accuracy": 0.9990172112654698, + "accuracy": 0.9990209317521045, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -103405,7 +103405,7 @@ ], "model.layers.53.self_attn": [ { - "accuracy": 0.9708366362672103, + "accuracy": 0.9714708390988802, "total_bits": 320757760, "q_proj": { "group_size": { @@ -103469,7 +103469,7 @@ } }, { - "accuracy": 0.9729972669952794, + "accuracy": 0.9747991499147917, "total_bits": 329080832, "q_proj": { "group_size": { @@ -103533,7 +103533,7 @@ } }, { - "accuracy": 0.9828286202330339, + "accuracy": 0.9824532380229548, "total_bits": 336024576, "q_proj": { "group_size": { @@ -103597,7 +103597,7 @@ } }, { - "accuracy": 0.9846750388019964, + "accuracy": 0.9841656245683369, "total_bits": 401557504, "q_proj": { "group_size": { @@ -103661,7 +103661,7 @@ } }, { - "accuracy": 0.986217410940873, + "accuracy": 0.9856206664913579, "total_bits": 475279360, "q_proj": { "group_size": { @@ -103725,7 +103725,7 @@ } }, { - "accuracy": 0.986575972092779, + "accuracy": 0.9871051350706502, "total_bits": 475479040, "q_proj": { "group_size": { @@ -103789,7 +103789,7 @@ } }, { - "accuracy": 0.9908226516685987, + "accuracy": 0.9909222376974005, "total_bits": 609759232, "q_proj": { "group_size": { @@ -103841,7 +103841,7 @@ } }, { - "accuracy": 0.9919285452679584, + "accuracy": 0.9919240827623167, "total_bits": 610024448, "q_proj": { "group_size": { @@ -103893,7 +103893,7 @@ } }, { - "accuracy": 0.9922851386823153, + "accuracy": 0.9926015860156009, "total_bits": 615020544, "q_proj": { "group_size": { @@ -103945,7 +103945,7 @@ } }, { - "accuracy": 0.9924029449098989, + "accuracy": 0.992701107734128, "total_bits": 623951872, "q_proj": { "group_size": { @@ -103997,7 +103997,7 @@ } }, { - "accuracy": 0.9924263758094687, + "accuracy": 0.9927321578326979, "total_bits": 626473984, "q_proj": { "group_size": { @@ -104061,7 +104061,7 @@ } }, { - "accuracy": 0.9932256348823246, + "accuracy": 0.9930105656385422, "total_bits": 630355968, "q_proj": { "group_size": { @@ -104125,7 +104125,7 @@ } }, { - "accuracy": 0.9948894910906491, + "accuracy": 0.994939240185838, "total_bits": 637362176, "q_proj": { "group_size": { @@ -104186,7 +104186,7 @@ } }, { - "accuracy": 0.9952417439536044, + "accuracy": 0.9953013756557515, "total_bits": 646823936, "q_proj": { "group_size": { @@ -104247,7 +104247,7 @@ } }, { - "accuracy": 0.9969604976083103, + "accuracy": 0.9970141212014776, "total_bits": 784740352, "q_proj": { "group_size": { @@ -104308,7 +104308,7 @@ } }, { - "accuracy": 0.9974792633009585, + "accuracy": 0.9974847581041487, "total_bits": 797818880, "q_proj": { "group_size": { @@ -104369,7 +104369,7 @@ } }, { - "accuracy": 0.9972146310304341, + "accuracy": 0.9972618888867529, "total_bits": 911749120, "q_proj": { "group_size": { @@ -104421,7 +104421,7 @@ } }, { - "accuracy": 0.9990739128307292, + "accuracy": 0.9990700541162177, "total_bits": 942718976, "q_proj": { "group_size": { @@ -104473,7 +104473,7 @@ } }, { - "accuracy": 0.9992740701785997, + "accuracy": 0.9992741605168894, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -104527,7 +104527,7 @@ ], "model.layers.53.mlp": [ { - "accuracy": 0.951221836240668, + "accuracy": 0.9514437788411191, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -104579,7 +104579,7 @@ } }, { - "accuracy": 0.9524522674711127, + "accuracy": 0.9526577748750386, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -104631,7 +104631,7 @@ } }, { - "accuracy": 0.9586023625574613, + "accuracy": 0.9587897909315009, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -104680,7 +104680,7 @@ } }, { - "accuracy": 0.9603345143167596, + "accuracy": 0.9605153673573544, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -104729,7 +104729,7 @@ } }, { - "accuracy": 0.975344036754809, + "accuracy": 0.9754566785536314, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -104781,7 +104781,7 @@ } }, { - "accuracy": 0.977261737773293, + "accuracy": 0.9773615284969932, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -104833,7 +104833,7 @@ } }, { - "accuracy": 0.9798405625318226, + "accuracy": 0.9799303710460663, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -104882,7 +104882,7 @@ } }, { - "accuracy": 0.987392470240593, + "accuracy": 0.9874544190733057, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -104925,7 +104925,7 @@ } }, { - "accuracy": 0.9884179469786192, + "accuracy": 0.9884738082948484, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -104968,7 +104968,7 @@ } }, { - "accuracy": 0.9874616877028817, + "accuracy": 0.9875220883833734, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -105020,7 +105020,7 @@ } }, { - "accuracy": 0.9889469287897411, + "accuracy": 0.9890011156860151, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -105072,7 +105072,7 @@ } }, { - "accuracy": 0.9936191761179974, + "accuracy": 0.9936493964571702, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -105124,7 +105124,7 @@ } }, { - "accuracy": 0.9944883600661629, + "accuracy": 0.9945142453438357, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -105176,7 +105176,7 @@ } }, { - "accuracy": 0.9965928067502222, + "accuracy": 0.9966100085722772, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -105219,7 +105219,7 @@ } }, { - "accuracy": 0.996721377890361, + "accuracy": 0.9967355028186974, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -105268,7 +105268,7 @@ } }, { - "accuracy": 0.9972012260634648, + "accuracy": 0.9972132752208334, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -105314,7 +105314,7 @@ } }, { - "accuracy": 0.9990112124697158, + "accuracy": 0.9990130021776024, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -105356,7 +105356,7 @@ ], "model.layers.54.self_attn": [ { - "accuracy": 0.9799219683596962, + "accuracy": 0.9807969080774408, "total_bits": 320757760, "q_proj": { "group_size": { @@ -105420,7 +105420,7 @@ } }, { - "accuracy": 0.9827559543283362, + "accuracy": 0.9843302416174036, "total_bits": 329080832, "q_proj": { "group_size": { @@ -105484,7 +105484,7 @@ } }, { - "accuracy": 0.9858648824064355, + "accuracy": 0.9864619311533476, "total_bits": 336024576, "q_proj": { "group_size": { @@ -105548,7 +105548,7 @@ } }, { - "accuracy": 0.9882254796592813, + "accuracy": 0.9887527854819047, "total_bits": 401557504, "q_proj": { "group_size": { @@ -105612,7 +105612,7 @@ } }, { - "accuracy": 0.98966373424781, + "accuracy": 0.9902207569072121, "total_bits": 475279360, "q_proj": { "group_size": { @@ -105676,7 +105676,7 @@ } }, { - "accuracy": 0.9900302620310533, + "accuracy": 0.9913410065989745, "total_bits": 475479040, "q_proj": { "group_size": { @@ -105740,7 +105740,7 @@ } }, { - "accuracy": 0.9929633642497816, + "accuracy": 0.9930600312195326, "total_bits": 609759232, "q_proj": { "group_size": { @@ -105792,7 +105792,7 @@ } }, { - "accuracy": 0.9935926402869978, + "accuracy": 0.993552033446337, "total_bits": 610024448, "q_proj": { "group_size": { @@ -105844,7 +105844,7 @@ } }, { - "accuracy": 0.9939405937728129, + "accuracy": 0.9941471740603447, "total_bits": 615020544, "q_proj": { "group_size": { @@ -105896,7 +105896,7 @@ } }, { - "accuracy": 0.9940407793772849, + "accuracy": 0.9942821374064997, "total_bits": 623951872, "q_proj": { "group_size": { @@ -105948,7 +105948,7 @@ } }, { - "accuracy": 0.9942950786728608, + "accuracy": 0.9946720384453472, "total_bits": 626473984, "q_proj": { "group_size": { @@ -106012,7 +106012,7 @@ } }, { - "accuracy": 0.9947249920744645, + "accuracy": 0.9947298263248644, "total_bits": 630355968, "q_proj": { "group_size": { @@ -106076,7 +106076,7 @@ } }, { - "accuracy": 0.9955062352513012, + "accuracy": 0.9955019954788057, "total_bits": 637362176, "q_proj": { "group_size": { @@ -106137,7 +106137,7 @@ } }, { - "accuracy": 0.9958743629486937, + "accuracy": 0.9958185102594527, "total_bits": 646823936, "q_proj": { "group_size": { @@ -106198,7 +106198,7 @@ } }, { - "accuracy": 0.9972753254206557, + "accuracy": 0.9972704276442528, "total_bits": 784740352, "q_proj": { "group_size": { @@ -106259,7 +106259,7 @@ } }, { - "accuracy": 0.9977015343151594, + "accuracy": 0.9977572032887685, "total_bits": 797818880, "q_proj": { "group_size": { @@ -106320,7 +106320,7 @@ } }, { - "accuracy": 0.9975112611918073, + "accuracy": 0.9974996884794611, "total_bits": 911749120, "q_proj": { "group_size": { @@ -106372,7 +106372,7 @@ } }, { - "accuracy": 0.9990533099166656, + "accuracy": 0.9990648461603805, "total_bits": 942718976, "q_proj": { "group_size": { @@ -106424,7 +106424,7 @@ } }, { - "accuracy": 0.999312882850829, + "accuracy": 0.9993193394651538, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -106478,7 +106478,7 @@ ], "model.layers.54.mlp": [ { - "accuracy": 0.9511978218429967, + "accuracy": 0.9514284102540267, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -106530,7 +106530,7 @@ } }, { - "accuracy": 0.9524217561671608, + "accuracy": 0.9526470554502386, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -106582,7 +106582,7 @@ } }, { - "accuracy": 0.9585994513411271, + "accuracy": 0.9587910614515606, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -106631,7 +106631,7 @@ } }, { - "accuracy": 0.9603085486512435, + "accuracy": 0.9604954248980472, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -106680,7 +106680,7 @@ } }, { - "accuracy": 0.9753205744843734, + "accuracy": 0.9754466492878763, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -106732,7 +106732,7 @@ } }, { - "accuracy": 0.9772532707766483, + "accuracy": 0.9773634844704678, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -106784,7 +106784,7 @@ } }, { - "accuracy": 0.9798384355871301, + "accuracy": 0.9799385086486214, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -106833,7 +106833,7 @@ } }, { - "accuracy": 0.9874173685124046, + "accuracy": 0.9874812176353053, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -106876,7 +106876,7 @@ } }, { - "accuracy": 0.9884515436072099, + "accuracy": 0.9885105584797106, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -106919,7 +106919,7 @@ } }, { - "accuracy": 0.9874624304081264, + "accuracy": 0.987525758774657, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -106971,7 +106971,7 @@ } }, { - "accuracy": 0.9889650250736036, + "accuracy": 0.9890201954465163, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -107023,7 +107023,7 @@ } }, { - "accuracy": 0.9936212622805646, + "accuracy": 0.9936559808097387, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -107075,7 +107075,7 @@ } }, { - "accuracy": 0.9945019534543941, + "accuracy": 0.9945289002437341, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -107127,7 +107127,7 @@ } }, { - "accuracy": 0.9966081584754743, + "accuracy": 0.9966260590835622, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -107170,7 +107170,7 @@ } }, { - "accuracy": 0.9967306942531937, + "accuracy": 0.9967504528008009, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -107219,7 +107219,7 @@ } }, { - "accuracy": 0.9972106692822356, + "accuracy": 0.9972288894810175, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -107265,7 +107265,7 @@ } }, { - "accuracy": 0.9990528244525194, + "accuracy": 0.9990585088533791, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -107307,7 +107307,7 @@ ], "model.layers.55.self_attn": [ { - "accuracy": 0.9789862428840838, + "accuracy": 0.980118414289073, "total_bits": 320757760, "q_proj": { "group_size": { @@ -107371,7 +107371,7 @@ } }, { - "accuracy": 0.9804508905661734, + "accuracy": 0.980436561923278, "total_bits": 329080832, "q_proj": { "group_size": { @@ -107435,7 +107435,7 @@ } }, { - "accuracy": 0.9842272745935541, + "accuracy": 0.9846044088664808, "total_bits": 336024576, "q_proj": { "group_size": { @@ -107499,7 +107499,7 @@ } }, { - "accuracy": 0.9878974266742405, + "accuracy": 0.9883365027214351, "total_bits": 401557504, "q_proj": { "group_size": { @@ -107563,7 +107563,7 @@ } }, { - "accuracy": 0.9891126540146375, + "accuracy": 0.989549823497471, "total_bits": 475279360, "q_proj": { "group_size": { @@ -107627,7 +107627,7 @@ } }, { - "accuracy": 0.9895536640757009, + "accuracy": 0.988960253564935, "total_bits": 475479040, "q_proj": { "group_size": { @@ -107691,7 +107691,7 @@ } }, { - "accuracy": 0.9923767369044455, + "accuracy": 0.9926875751269492, "total_bits": 609759232, "q_proj": { "group_size": { @@ -107743,7 +107743,7 @@ } }, { - "accuracy": 0.9928120586432909, + "accuracy": 0.9929375530857789, "total_bits": 610024448, "q_proj": { "group_size": { @@ -107795,7 +107795,7 @@ } }, { - "accuracy": 0.9933218446217085, + "accuracy": 0.9936534044773955, "total_bits": 615020544, "q_proj": { "group_size": { @@ -107847,7 +107847,7 @@ } }, { - "accuracy": 0.9934699151076769, + "accuracy": 0.9937895591321745, "total_bits": 623951872, "q_proj": { "group_size": { @@ -107899,7 +107899,7 @@ } }, { - "accuracy": 0.9935824882827307, + "accuracy": 0.9936909189349726, "total_bits": 626473984, "q_proj": { "group_size": { @@ -107963,7 +107963,7 @@ } }, { - "accuracy": 0.993922868449437, + "accuracy": 0.994028774531264, "total_bits": 630355968, "q_proj": { "group_size": { @@ -108027,7 +108027,7 @@ } }, { - "accuracy": 0.9951866959270678, + "accuracy": 0.9951857281358618, "total_bits": 637362176, "q_proj": { "group_size": { @@ -108088,7 +108088,7 @@ } }, { - "accuracy": 0.9957305451757029, + "accuracy": 0.9956477456971219, "total_bits": 646823936, "q_proj": { "group_size": { @@ -108149,7 +108149,7 @@ } }, { - "accuracy": 0.9971681017624704, + "accuracy": 0.9972445872661314, "total_bits": 784740352, "q_proj": { "group_size": { @@ -108210,7 +108210,7 @@ } }, { - "accuracy": 0.9977009184658527, + "accuracy": 0.9976520291284511, "total_bits": 797818880, "q_proj": { "group_size": { @@ -108271,7 +108271,7 @@ } }, { - "accuracy": 0.9975690355426386, + "accuracy": 0.997652460477854, "total_bits": 911749120, "q_proj": { "group_size": { @@ -108323,7 +108323,7 @@ } }, { - "accuracy": 0.9989169223332092, + "accuracy": 0.9989340744520488, "total_bits": 942718976, "q_proj": { "group_size": { @@ -108375,7 +108375,7 @@ } }, { - "accuracy": 0.9992208738664263, + "accuracy": 0.999218310621616, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -108429,7 +108429,7 @@ ], "model.layers.55.mlp": [ { - "accuracy": 0.9506603887206629, + "accuracy": 0.9509090185165405, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -108481,7 +108481,7 @@ } }, { - "accuracy": 0.9519139967466655, + "accuracy": 0.9521476532283583, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -108533,7 +108533,7 @@ } }, { - "accuracy": 0.9581057868505779, + "accuracy": 0.9583042640435069, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -108582,7 +108582,7 @@ } }, { - "accuracy": 0.9598335498257687, + "accuracy": 0.9600202252990321, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -108631,7 +108631,7 @@ } }, { - "accuracy": 0.9750490329767528, + "accuracy": 0.9751751485623812, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -108683,7 +108683,7 @@ } }, { - "accuracy": 0.9769971668720245, + "accuracy": 0.9771133206392589, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -108735,7 +108735,7 @@ } }, { - "accuracy": 0.9795962368187151, + "accuracy": 0.9796978156817587, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -108784,7 +108784,7 @@ } }, { - "accuracy": 0.9872632552134363, + "accuracy": 0.9873284893600565, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -108827,7 +108827,7 @@ } }, { - "accuracy": 0.9883105833279459, + "accuracy": 0.98837122870119, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -108870,7 +108870,7 @@ } }, { - "accuracy": 0.9873227257477609, + "accuracy": 0.9873874187469482, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -108922,7 +108922,7 @@ } }, { - "accuracy": 0.9888414701348857, + "accuracy": 0.988898040432679, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -108974,7 +108974,7 @@ } }, { - "accuracy": 0.993550673911446, + "accuracy": 0.9935834674458754, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -109026,7 +109026,7 @@ } }, { - "accuracy": 0.9944373047665546, + "accuracy": 0.9944651440570229, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -109078,7 +109078,7 @@ } }, { - "accuracy": 0.9965581266503585, + "accuracy": 0.9965761472520075, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -109121,7 +109121,7 @@ } }, { - "accuracy": 0.9966876557783076, + "accuracy": 0.9967055261919373, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -109170,7 +109170,7 @@ } }, { - "accuracy": 0.9971699787205771, + "accuracy": 0.9971857410120336, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -109216,7 +109216,7 @@ } }, { - "accuracy": 0.9990027316502834, + "accuracy": 0.9990074588476043, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -109258,7 +109258,7 @@ ], "model.layers.56.self_attn": [ { - "accuracy": 0.9634280079289487, + "accuracy": 0.9631827379527845, "total_bits": 320757760, "q_proj": { "group_size": { @@ -109322,7 +109322,7 @@ } }, { - "accuracy": 0.9665420243614599, + "accuracy": 0.9671248919085452, "total_bits": 329080832, "q_proj": { "group_size": { @@ -109386,7 +109386,7 @@ } }, { - "accuracy": 0.976246095017383, + "accuracy": 0.9774175992137507, "total_bits": 336024576, "q_proj": { "group_size": { @@ -109450,7 +109450,7 @@ } }, { - "accuracy": 0.9795936722504465, + "accuracy": 0.9804748500648298, "total_bits": 401557504, "q_proj": { "group_size": { @@ -109514,7 +109514,7 @@ } }, { - "accuracy": 0.9832922251600968, + "accuracy": 0.9839301423022622, "total_bits": 475279360, "q_proj": { "group_size": { @@ -109578,7 +109578,7 @@ } }, { - "accuracy": 0.9831936218236622, + "accuracy": 0.9845556930491799, "total_bits": 475479040, "q_proj": { "group_size": { @@ -109642,7 +109642,7 @@ } }, { - "accuracy": 0.9910126610806114, + "accuracy": 0.9909385144710541, "total_bits": 609759232, "q_proj": { "group_size": { @@ -109694,7 +109694,7 @@ } }, { - "accuracy": 0.991245301930528, + "accuracy": 0.9911981107372987, "total_bits": 610024448, "q_proj": { "group_size": { @@ -109746,7 +109746,7 @@ } }, { - "accuracy": 0.9918757520223919, + "accuracy": 0.9918422424479535, "total_bits": 615020544, "q_proj": { "group_size": { @@ -109798,7 +109798,7 @@ } }, { - "accuracy": 0.9920777051072371, + "accuracy": 0.9920470989064166, "total_bits": 623951872, "q_proj": { "group_size": { @@ -109850,7 +109850,7 @@ } }, { - "accuracy": 0.991915477733863, + "accuracy": 0.9917567936997664, "total_bits": 626473984, "q_proj": { "group_size": { @@ -109914,7 +109914,7 @@ } }, { - "accuracy": 0.9924442023038864, + "accuracy": 0.9923739997964156, "total_bits": 630355968, "q_proj": { "group_size": { @@ -109978,7 +109978,7 @@ } }, { - "accuracy": 0.994210505171826, + "accuracy": 0.9944297914442263, "total_bits": 637362176, "q_proj": { "group_size": { @@ -110039,7 +110039,7 @@ } }, { - "accuracy": 0.9948194552409021, + "accuracy": 0.9949034672034415, "total_bits": 646823936, "q_proj": { "group_size": { @@ -110100,7 +110100,7 @@ } }, { - "accuracy": 0.9969312080035084, + "accuracy": 0.9968523767433668, "total_bits": 784740352, "q_proj": { "group_size": { @@ -110161,7 +110161,7 @@ } }, { - "accuracy": 0.9973623529076576, + "accuracy": 0.9974034020775243, "total_bits": 797818880, "q_proj": { "group_size": { @@ -110222,7 +110222,7 @@ } }, { - "accuracy": 0.9975241871648713, + "accuracy": 0.9974353619311985, "total_bits": 911749120, "q_proj": { "group_size": { @@ -110274,7 +110274,7 @@ } }, { - "accuracy": 0.9988003293738553, + "accuracy": 0.9988003995661673, "total_bits": 942718976, "q_proj": { "group_size": { @@ -110326,7 +110326,7 @@ } }, { - "accuracy": 0.999330161188386, + "accuracy": 0.9993325462074656, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -110380,7 +110380,7 @@ ], "model.layers.56.mlp": [ { - "accuracy": 0.9495103704301935, + "accuracy": 0.9497662437589545, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -110432,7 +110432,7 @@ } }, { - "accuracy": 0.9507956598934374, + "accuracy": 0.951040478129136, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -110484,7 +110484,7 @@ } }, { - "accuracy": 0.9571198632842616, + "accuracy": 0.9573314378136083, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -110533,7 +110533,7 @@ } }, { - "accuracy": 0.958906352519989, + "accuracy": 0.959112032463676, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -110582,7 +110582,7 @@ } }, { - "accuracy": 0.9744714137754942, + "accuracy": 0.9746022381280598, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -110634,7 +110634,7 @@ } }, { - "accuracy": 0.9764562437408849, + "accuracy": 0.9765803406113073, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -110686,7 +110686,7 @@ } }, { - "accuracy": 0.9791206921401777, + "accuracy": 0.9792312132684808, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -110735,7 +110735,7 @@ } }, { - "accuracy": 0.9869518303557446, + "accuracy": 0.9870156645774841, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -110778,7 +110778,7 @@ } }, { - "accuracy": 0.9880132902609674, + "accuracy": 0.9880749536188025, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -110821,7 +110821,7 @@ } }, { - "accuracy": 0.9870239182522422, + "accuracy": 0.9870883814598385, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -110873,7 +110873,7 @@ } }, { - "accuracy": 0.9885697404020711, + "accuracy": 0.9886274722061659, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -110925,7 +110925,7 @@ } }, { - "accuracy": 0.9933961159304568, + "accuracy": 0.9934294580628997, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -110977,7 +110977,7 @@ } }, { - "accuracy": 0.9943023351462263, + "accuracy": 0.9943319206175051, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -111029,7 +111029,7 @@ } }, { - "accuracy": 0.9964769823770774, + "accuracy": 0.9964950021944547, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -111072,7 +111072,7 @@ } }, { - "accuracy": 0.9966103285551071, + "accuracy": 0.9966277013483801, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -111121,7 +111121,7 @@ } }, { - "accuracy": 0.9971059803899965, + "accuracy": 0.9971208172409158, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -111167,7 +111167,7 @@ } }, { - "accuracy": 0.9989948474655026, + "accuracy": 0.9990003831488522, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -111209,7 +111209,7 @@ ], "model.layers.57.self_attn": [ { - "accuracy": 0.9790824510549244, + "accuracy": 0.9798179629601931, "total_bits": 320757760, "q_proj": { "group_size": { @@ -111273,7 +111273,7 @@ } }, { - "accuracy": 0.9809796841521012, + "accuracy": 0.9818102921310224, "total_bits": 329080832, "q_proj": { "group_size": { @@ -111337,7 +111337,7 @@ } }, { - "accuracy": 0.9839377920878561, + "accuracy": 0.9844887774241599, "total_bits": 336024576, "q_proj": { "group_size": { @@ -111401,7 +111401,7 @@ } }, { - "accuracy": 0.9867779314517975, + "accuracy": 0.9873375853425578, "total_bits": 401557504, "q_proj": { "group_size": { @@ -111465,7 +111465,7 @@ } }, { - "accuracy": 0.9881865303767355, + "accuracy": 0.988739644226275, "total_bits": 475279360, "q_proj": { "group_size": { @@ -111529,7 +111529,7 @@ } }, { - "accuracy": 0.9884520024061203, + "accuracy": 0.9892209674182691, "total_bits": 475479040, "q_proj": { "group_size": { @@ -111593,7 +111593,7 @@ } }, { - "accuracy": 0.9912155490172537, + "accuracy": 0.9914976123132204, "total_bits": 609759232, "q_proj": { "group_size": { @@ -111645,7 +111645,7 @@ } }, { - "accuracy": 0.9922054645262266, + "accuracy": 0.9920270678244139, "total_bits": 610024448, "q_proj": { "group_size": { @@ -111697,7 +111697,7 @@ } }, { - "accuracy": 0.992621226530326, + "accuracy": 0.9929344716824984, "total_bits": 615020544, "q_proj": { "group_size": { @@ -111749,7 +111749,7 @@ } }, { - "accuracy": 0.9927386704244112, + "accuracy": 0.9930637039636311, "total_bits": 623951872, "q_proj": { "group_size": { @@ -111801,7 +111801,7 @@ } }, { - "accuracy": 0.9927314080690083, + "accuracy": 0.9931192696094513, "total_bits": 626473984, "q_proj": { "group_size": { @@ -111865,7 +111865,7 @@ } }, { - "accuracy": 0.993340496954165, + "accuracy": 0.9932935575121328, "total_bits": 630355968, "q_proj": { "group_size": { @@ -111929,7 +111929,7 @@ } }, { - "accuracy": 0.994636674068476, + "accuracy": 0.9946954536594843, "total_bits": 637362176, "q_proj": { "group_size": { @@ -111990,7 +111990,7 @@ } }, { - "accuracy": 0.995002779913576, + "accuracy": 0.9951429614110997, "total_bits": 646823936, "q_proj": { "group_size": { @@ -112051,7 +112051,7 @@ } }, { - "accuracy": 0.996961023658514, + "accuracy": 0.996921792430313, "total_bits": 784740352, "q_proj": { "group_size": { @@ -112112,7 +112112,7 @@ } }, { - "accuracy": 0.9973941057136184, + "accuracy": 0.9974401922602403, "total_bits": 797818880, "q_proj": { "group_size": { @@ -112173,7 +112173,7 @@ } }, { - "accuracy": 0.9972537305009993, + "accuracy": 0.9972118313767409, "total_bits": 911749120, "q_proj": { "group_size": { @@ -112225,7 +112225,7 @@ } }, { - "accuracy": 0.9989990814539947, + "accuracy": 0.9989868494632997, "total_bits": 942718976, "q_proj": { "group_size": { @@ -112277,7 +112277,7 @@ } }, { - "accuracy": 0.9992643443279361, + "accuracy": 0.9992321309094366, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -112331,7 +112331,7 @@ ], "model.layers.57.mlp": [ { - "accuracy": 0.9491821903931468, + "accuracy": 0.9494552486821225, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -112383,7 +112383,7 @@ } }, { - "accuracy": 0.9504819826075905, + "accuracy": 0.9507334608780711, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -112435,7 +112435,7 @@ } }, { - "accuracy": 0.9568110955388922, + "accuracy": 0.9570276329391881, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -112484,7 +112484,7 @@ } }, { - "accuracy": 0.9585840890282079, + "accuracy": 0.9587931570253874, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -112533,7 +112533,7 @@ } }, { - "accuracy": 0.9742984897211978, + "accuracy": 0.9744346251613215, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -112585,7 +112585,7 @@ } }, { - "accuracy": 0.9762962115438361, + "accuracy": 0.9764170411385988, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -112637,7 +112637,7 @@ } }, { - "accuracy": 0.978955256311517, + "accuracy": 0.9790648159227873, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -112686,7 +112686,7 @@ } }, { - "accuracy": 0.9868731875168649, + "accuracy": 0.9869428175060373, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -112729,7 +112729,7 @@ } }, { - "accuracy": 0.9879485972617802, + "accuracy": 0.9880149591910211, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -112772,7 +112772,7 @@ } }, { - "accuracy": 0.9869412011221835, + "accuracy": 0.9870122075080872, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -112824,7 +112824,7 @@ } }, { - "accuracy": 0.9885021354022779, + "accuracy": 0.9885647806682085, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -112876,7 +112876,7 @@ } }, { - "accuracy": 0.9933619577633707, + "accuracy": 0.9933982844415464, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -112928,7 +112928,7 @@ } }, { - "accuracy": 0.9942734814003894, + "accuracy": 0.9943049902978697, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -112980,7 +112980,7 @@ } }, { - "accuracy": 0.9964668178244641, + "accuracy": 0.9964869147852847, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -113023,7 +113023,7 @@ } }, { - "accuracy": 0.9966126339215982, + "accuracy": 0.9966287695263562, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -113072,7 +113072,7 @@ } }, { - "accuracy": 0.9971125221958286, + "accuracy": 0.9971256528637911, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -113118,7 +113118,7 @@ } }, { - "accuracy": 0.9990566214056391, + "accuracy": 0.9990614170306608, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -113160,7 +113160,7 @@ ], "model.layers.58.self_attn": [ { - "accuracy": 0.9829953124648646, + "accuracy": 0.9832098609522769, "total_bits": 320757760, "q_proj": { "group_size": { @@ -113224,7 +113224,7 @@ } }, { - "accuracy": 0.9840947232748333, + "accuracy": 0.9840953726517526, "total_bits": 329080832, "q_proj": { "group_size": { @@ -113288,7 +113288,7 @@ } }, { - "accuracy": 0.9866198724821994, + "accuracy": 0.9864305947956286, "total_bits": 336024576, "q_proj": { "group_size": { @@ -113352,7 +113352,7 @@ } }, { - "accuracy": 0.9891383938099209, + "accuracy": 0.9888289876674351, "total_bits": 401557504, "q_proj": { "group_size": { @@ -113416,7 +113416,7 @@ } }, { - "accuracy": 0.9903772332166371, + "accuracy": 0.9899887552386836, "total_bits": 475279360, "q_proj": { "group_size": { @@ -113480,7 +113480,7 @@ } }, { - "accuracy": 0.9905720353126526, + "accuracy": 0.9906402371431652, "total_bits": 475479040, "q_proj": { "group_size": { @@ -113544,7 +113544,7 @@ } }, { - "accuracy": 0.9928626990632007, + "accuracy": 0.9928043939565357, "total_bits": 609759232, "q_proj": { "group_size": { @@ -113596,7 +113596,7 @@ } }, { - "accuracy": 0.9933237386377234, + "accuracy": 0.9934644577534575, "total_bits": 610024448, "q_proj": { "group_size": { @@ -113648,7 +113648,7 @@ } }, { - "accuracy": 0.9938536396152094, + "accuracy": 0.9940411479849565, "total_bits": 615020544, "q_proj": { "group_size": { @@ -113700,7 +113700,7 @@ } }, { - "accuracy": 0.9939847765000243, + "accuracy": 0.994147473259976, "total_bits": 623951872, "q_proj": { "group_size": { @@ -113752,7 +113752,7 @@ } }, { - "accuracy": 0.9939296371058414, + "accuracy": 0.9940437678443758, "total_bits": 626473984, "q_proj": { "group_size": { @@ -113816,7 +113816,7 @@ } }, { - "accuracy": 0.9943331821184409, + "accuracy": 0.9943619280269271, "total_bits": 630355968, "q_proj": { "group_size": { @@ -113880,7 +113880,7 @@ } }, { - "accuracy": 0.9954104690175307, + "accuracy": 0.9952621930523923, "total_bits": 637362176, "q_proj": { "group_size": { @@ -113941,7 +113941,7 @@ } }, { - "accuracy": 0.9956836567113274, + "accuracy": 0.9957024364879257, "total_bits": 646823936, "q_proj": { "group_size": { @@ -114002,7 +114002,7 @@ } }, { - "accuracy": 0.997183732688427, + "accuracy": 0.9972457823000456, "total_bits": 784740352, "q_proj": { "group_size": { @@ -114063,7 +114063,7 @@ } }, { - "accuracy": 0.9976733654345337, + "accuracy": 0.9976774381571695, "total_bits": 797818880, "q_proj": { "group_size": { @@ -114124,7 +114124,7 @@ } }, { - "accuracy": 0.9974185810669473, + "accuracy": 0.9974815502370659, "total_bits": 911749120, "q_proj": { "group_size": { @@ -114176,7 +114176,7 @@ } }, { - "accuracy": 0.9991080876635877, + "accuracy": 0.9991099321724553, "total_bits": 942718976, "q_proj": { "group_size": { @@ -114228,7 +114228,7 @@ } }, { - "accuracy": 0.9993254018359279, + "accuracy": 0.9993150261671919, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -114282,7 +114282,7 @@ ], "model.layers.58.mlp": [ { - "accuracy": 0.9490075017276564, + "accuracy": 0.9492900026471991, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -114334,7 +114334,7 @@ } }, { - "accuracy": 0.9502944977659928, + "accuracy": 0.9505696014354104, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -114386,7 +114386,7 @@ } }, { - "accuracy": 0.9566593232907747, + "accuracy": 0.9568976979506643, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -114435,7 +114435,7 @@ } }, { - "accuracy": 0.9584141781455592, + "accuracy": 0.9586438944465235, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -114484,7 +114484,7 @@ } }, { - "accuracy": 0.9741703256180412, + "accuracy": 0.9743154629280693, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -114536,7 +114536,7 @@ } }, { - "accuracy": 0.9761996002573716, + "accuracy": 0.9763255040896567, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -114588,7 +114588,7 @@ } }, { - "accuracy": 0.9788700075525987, + "accuracy": 0.9789826885650033, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -114637,7 +114637,7 @@ } }, { - "accuracy": 0.986824642670782, + "accuracy": 0.9868973132811094, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -114680,7 +114680,7 @@ } }, { - "accuracy": 0.9879198615488253, + "accuracy": 0.9879831381534275, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -114723,7 +114723,7 @@ } }, { - "accuracy": 0.9868744109806261, + "accuracy": 0.9869484015201268, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -114775,7 +114775,7 @@ } }, { - "accuracy": 0.9884518839811024, + "accuracy": 0.9885147637442538, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -114827,7 +114827,7 @@ } }, { - "accuracy": 0.9933249636700279, + "accuracy": 0.9933643066569379, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -114879,7 +114879,7 @@ } }, { - "accuracy": 0.9942489533047927, + "accuracy": 0.9942809398237028, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -114931,7 +114931,7 @@ } }, { - "accuracy": 0.9964532381609866, + "accuracy": 0.9964733010059909, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -114974,7 +114974,7 @@ } }, { - "accuracy": 0.9965904700128656, + "accuracy": 0.9966097881919459, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -115023,7 +115023,7 @@ } }, { - "accuracy": 0.9970889158154789, + "accuracy": 0.99710595097981, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -115069,7 +115069,7 @@ } }, { - "accuracy": 0.9990534108916396, + "accuracy": 0.9990588424629286, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -115111,7 +115111,7 @@ ], "model.layers.59.self_attn": [ { - "accuracy": 0.9833300490128366, + "accuracy": 0.9842811744464072, "total_bits": 320757760, "q_proj": { "group_size": { @@ -115175,7 +115175,7 @@ } }, { - "accuracy": 0.9838761969616538, + "accuracy": 0.9853646300340954, "total_bits": 329080832, "q_proj": { "group_size": { @@ -115239,7 +115239,7 @@ } }, { - "accuracy": 0.9887080404319262, + "accuracy": 0.9887697704528507, "total_bits": 336024576, "q_proj": { "group_size": { @@ -115303,7 +115303,7 @@ } }, { - "accuracy": 0.9904703239077016, + "accuracy": 0.9905944113668642, "total_bits": 401557504, "q_proj": { "group_size": { @@ -115367,7 +115367,7 @@ } }, { - "accuracy": 0.9910813201414911, + "accuracy": 0.9912205479646984, "total_bits": 475279360, "q_proj": { "group_size": { @@ -115431,7 +115431,7 @@ } }, { - "accuracy": 0.9913861955467024, + "accuracy": 0.9907317192930924, "total_bits": 475479040, "q_proj": { "group_size": { @@ -115495,7 +115495,7 @@ } }, { - "accuracy": 0.993184960202167, + "accuracy": 0.9931072694690604, "total_bits": 609759232, "q_proj": { "group_size": { @@ -115547,7 +115547,7 @@ } }, { - "accuracy": 0.9939505685316888, + "accuracy": 0.9938294252282694, "total_bits": 610024448, "q_proj": { "group_size": { @@ -115599,7 +115599,7 @@ } }, { - "accuracy": 0.994353996295678, + "accuracy": 0.994358532522854, "total_bits": 615020544, "q_proj": { "group_size": { @@ -115651,7 +115651,7 @@ } }, { - "accuracy": 0.9944278542932711, + "accuracy": 0.9944223945862368, "total_bits": 623951872, "q_proj": { "group_size": { @@ -115703,7 +115703,7 @@ } }, { - "accuracy": 0.9942180679032677, + "accuracy": 0.9944595898452558, "total_bits": 626473984, "q_proj": { "group_size": { @@ -115767,7 +115767,7 @@ } }, { - "accuracy": 0.9944803804943436, + "accuracy": 0.9947014368678394, "total_bits": 630355968, "q_proj": { "group_size": { @@ -115831,7 +115831,7 @@ } }, { - "accuracy": 0.9958677338926416, + "accuracy": 0.9958038322235409, "total_bits": 637362176, "q_proj": { "group_size": { @@ -115892,7 +115892,7 @@ } }, { - "accuracy": 0.9962392284681922, + "accuracy": 0.9963004142045975, "total_bits": 646823936, "q_proj": { "group_size": { @@ -115953,7 +115953,7 @@ } }, { - "accuracy": 0.9975154256742251, + "accuracy": 0.9974967958895784, "total_bits": 784740352, "q_proj": { "group_size": { @@ -116014,7 +116014,7 @@ } }, { - "accuracy": 0.9979438085697199, + "accuracy": 0.9980194254925376, "total_bits": 797818880, "q_proj": { "group_size": { @@ -116075,7 +116075,7 @@ } }, { - "accuracy": 0.9976660168092502, + "accuracy": 0.9976522085305891, "total_bits": 911749120, "q_proj": { "group_size": { @@ -116127,7 +116127,7 @@ } }, { - "accuracy": 0.9992655914668974, + "accuracy": 0.9992616087884495, "total_bits": 942718976, "q_proj": { "group_size": { @@ -116179,7 +116179,7 @@ } }, { - "accuracy": 0.9993602508482965, + "accuracy": 0.9993683818825766, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -116233,7 +116233,7 @@ ], "model.layers.59.mlp": [ { - "accuracy": 0.9478302001953125, + "accuracy": 0.9481125066154882, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -116285,7 +116285,7 @@ } }, { - "accuracy": 0.9491698365462453, + "accuracy": 0.9494428509160092, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -116337,7 +116337,7 @@ } }, { - "accuracy": 0.9556074111085189, + "accuracy": 0.9558422879168862, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -116386,7 +116386,7 @@ } }, { - "accuracy": 0.9573921373015956, + "accuracy": 0.9576230582438017, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -116435,7 +116435,7 @@ } }, { - "accuracy": 0.9736095760997973, + "accuracy": 0.9737527856701299, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -116487,7 +116487,7 @@ } }, { - "accuracy": 0.9756593390514976, + "accuracy": 0.9757893571728155, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -116539,7 +116539,7 @@ } }, { - "accuracy": 0.9783662949737749, + "accuracy": 0.9784829930255288, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -116588,7 +116588,7 @@ } }, { - "accuracy": 0.9865264735723797, + "accuracy": 0.9866003441183191, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -116631,7 +116631,7 @@ } }, { - "accuracy": 0.9876171985739156, + "accuracy": 0.9876818594179655, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -116674,7 +116674,7 @@ } }, { - "accuracy": 0.9865897689995012, + "accuracy": 0.9866610781142586, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -116726,7 +116726,7 @@ } }, { - "accuracy": 0.9881767073744222, + "accuracy": 0.9882409509859587, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -116778,7 +116778,7 @@ } }, { - "accuracy": 0.9931817352771759, + "accuracy": 0.9932162989127008, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -116830,7 +116830,7 @@ } }, { - "accuracy": 0.9941078526409048, + "accuracy": 0.994140141104397, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -116882,7 +116882,7 @@ } }, { - "accuracy": 0.9963688587671832, + "accuracy": 0.9963874232612158, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -116925,7 +116925,7 @@ } }, { - "accuracy": 0.9965046620682666, + "accuracy": 0.996522410527656, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -116974,7 +116974,7 @@ } }, { - "accuracy": 0.9970086765916724, + "accuracy": 0.9970235303044319, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -117020,7 +117020,7 @@ } }, { - "accuracy": 0.9989822238291565, + "accuracy": 0.9989862374373173, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -117062,7 +117062,7 @@ ], "model.layers.60.self_attn": [ { - "accuracy": 0.9578190226303904, + "accuracy": 0.9582568061979193, "total_bits": 320757760, "q_proj": { "group_size": { @@ -117126,7 +117126,7 @@ } }, { - "accuracy": 0.9594920497191579, + "accuracy": 0.9602970104468496, "total_bits": 329080832, "q_proj": { "group_size": { @@ -117190,7 +117190,7 @@ } }, { - "accuracy": 0.9756089119534743, + "accuracy": 0.9756001030143938, "total_bits": 336024576, "q_proj": { "group_size": { @@ -117254,7 +117254,7 @@ } }, { - "accuracy": 0.9785418353582683, + "accuracy": 0.978312948816701, "total_bits": 401557504, "q_proj": { "group_size": { @@ -117318,7 +117318,7 @@ } }, { - "accuracy": 0.9814661976538206, + "accuracy": 0.9813581579609921, "total_bits": 475279360, "q_proj": { "group_size": { @@ -117382,7 +117382,7 @@ } }, { - "accuracy": 0.9814753281442743, + "accuracy": 0.9817807094046944, "total_bits": 475479040, "q_proj": { "group_size": { @@ -117446,7 +117446,7 @@ } }, { - "accuracy": 0.9893943816423416, + "accuracy": 0.9895560137535396, "total_bits": 609759232, "q_proj": { "group_size": { @@ -117498,7 +117498,7 @@ } }, { - "accuracy": 0.9899491173656363, + "accuracy": 0.9901725069472664, "total_bits": 610024448, "q_proj": { "group_size": { @@ -117550,7 +117550,7 @@ } }, { - "accuracy": 0.9906094646767566, + "accuracy": 0.9909621500655225, "total_bits": 615020544, "q_proj": { "group_size": { @@ -117602,7 +117602,7 @@ } }, { - "accuracy": 0.9907553933168712, + "accuracy": 0.9911495423630664, "total_bits": 623951872, "q_proj": { "group_size": { @@ -117654,7 +117654,7 @@ } }, { - "accuracy": 0.9908654274124848, + "accuracy": 0.9909107151784395, "total_bits": 626473984, "q_proj": { "group_size": { @@ -117718,7 +117718,7 @@ } }, { - "accuracy": 0.9911731983485975, + "accuracy": 0.9911761103492034, "total_bits": 630355968, "q_proj": { "group_size": { @@ -117782,7 +117782,7 @@ } }, { - "accuracy": 0.9937042825316128, + "accuracy": 0.9939513198639217, "total_bits": 637362176, "q_proj": { "group_size": { @@ -117843,7 +117843,7 @@ } }, { - "accuracy": 0.9943244786638963, + "accuracy": 0.9944392070174217, "total_bits": 646823936, "q_proj": { "group_size": { @@ -117904,7 +117904,7 @@ } }, { - "accuracy": 0.996607265974346, + "accuracy": 0.9967199528687879, "total_bits": 784740352, "q_proj": { "group_size": { @@ -117965,7 +117965,7 @@ } }, { - "accuracy": 0.99708755137889, + "accuracy": 0.9971940109604284, "total_bits": 797818880, "q_proj": { "group_size": { @@ -118026,7 +118026,7 @@ } }, { - "accuracy": 0.9971153679254808, + "accuracy": 0.9972327729981196, "total_bits": 911749120, "q_proj": { "group_size": { @@ -118078,7 +118078,7 @@ } }, { - "accuracy": 0.9988228905161745, + "accuracy": 0.9988272733202106, "total_bits": 942718976, "q_proj": { "group_size": { @@ -118130,7 +118130,7 @@ } }, { - "accuracy": 0.9992392588602869, + "accuracy": 0.9992560075203839, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -118184,7 +118184,7 @@ ], "model.layers.60.mlp": [ { - "accuracy": 0.9467590608094868, + "accuracy": 0.9470314414877641, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -118236,7 +118236,7 @@ } }, { - "accuracy": 0.9481437049413982, + "accuracy": 0.9484032172905772, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -118288,7 +118288,7 @@ } }, { - "accuracy": 0.9546760697113841, + "accuracy": 0.9549135565757751, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -118337,7 +118337,7 @@ } }, { - "accuracy": 0.9565189135702032, + "accuracy": 0.9567477828577945, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -118386,7 +118386,7 @@ } }, { - "accuracy": 0.9730447468004728, + "accuracy": 0.9731941160402799, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -118438,7 +118438,7 @@ } }, { - "accuracy": 0.9751514761071456, + "accuracy": 0.9752823914352217, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -118490,7 +118490,7 @@ } }, { - "accuracy": 0.9779013975670463, + "accuracy": 0.9780205111754569, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -118539,7 +118539,7 @@ } }, { - "accuracy": 0.9862160353284133, + "accuracy": 0.986293154327493, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -118582,7 +118582,7 @@ } }, { - "accuracy": 0.9873286634683609, + "accuracy": 0.9874009584125719, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -118625,7 +118625,7 @@ } }, { - "accuracy": 0.986288120872096, + "accuracy": 0.9863648445982682, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -118677,7 +118677,7 @@ } }, { - "accuracy": 0.9879277544586282, + "accuracy": 0.987996021383687, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -118729,7 +118729,7 @@ } }, { - "accuracy": 0.9930261459789778, + "accuracy": 0.9930622405127475, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -118781,7 +118781,7 @@ } }, { - "accuracy": 0.9939801892951915, + "accuracy": 0.994013733769718, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -118833,7 +118833,7 @@ } }, { - "accuracy": 0.9962720772937724, + "accuracy": 0.9962924412990871, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -118876,7 +118876,7 @@ } }, { - "accuracy": 0.9964163040644244, + "accuracy": 0.99643753312136, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -118925,7 +118925,7 @@ } }, { - "accuracy": 0.9969279179839712, + "accuracy": 0.9969474439950365, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -118971,7 +118971,7 @@ } }, { - "accuracy": 0.9989222045987844, + "accuracy": 0.9989305419945403, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -119013,7 +119013,7 @@ ], "model.layers.61.self_attn": [ { - "accuracy": 0.9845995118743495, + "accuracy": 0.9847357429956135, "total_bits": 320757760, "q_proj": { "group_size": { @@ -119077,7 +119077,7 @@ } }, { - "accuracy": 0.9857342384363476, + "accuracy": 0.9859363107304824, "total_bits": 329080832, "q_proj": { "group_size": { @@ -119141,7 +119141,7 @@ } }, { - "accuracy": 0.9887276778095647, + "accuracy": 0.988815869155683, "total_bits": 336024576, "q_proj": { "group_size": { @@ -119205,7 +119205,7 @@ } }, { - "accuracy": 0.9906210711127833, + "accuracy": 0.9906943369852869, "total_bits": 401557504, "q_proj": { "group_size": { @@ -119269,7 +119269,7 @@ } }, { - "accuracy": 0.9918210490753776, + "accuracy": 0.991860148153807, "total_bits": 475279360, "q_proj": { "group_size": { @@ -119333,7 +119333,7 @@ } }, { - "accuracy": 0.9919786343449041, + "accuracy": 0.9920872810639834, "total_bits": 475479040, "q_proj": { "group_size": { @@ -119397,7 +119397,7 @@ } }, { - "accuracy": 0.9935639735899473, + "accuracy": 0.9936265910142347, "total_bits": 609759232, "q_proj": { "group_size": { @@ -119449,7 +119449,7 @@ } }, { - "accuracy": 0.9937015015043711, + "accuracy": 0.9938337363694844, "total_bits": 610024448, "q_proj": { "group_size": { @@ -119501,7 +119501,7 @@ } }, { - "accuracy": 0.9943795596298418, + "accuracy": 0.994372255708042, "total_bits": 615020544, "q_proj": { "group_size": { @@ -119553,7 +119553,7 @@ } }, { - "accuracy": 0.9944548132388216, + "accuracy": 0.9944548987244305, "total_bits": 623951872, "q_proj": { "group_size": { @@ -119605,7 +119605,7 @@ } }, { - "accuracy": 0.9952905483936009, + "accuracy": 0.9952154904603958, "total_bits": 626473984, "q_proj": { "group_size": { @@ -119669,7 +119669,7 @@ } }, { - "accuracy": 0.9956011121210299, + "accuracy": 0.9955816323819914, "total_bits": 630355968, "q_proj": { "group_size": { @@ -119733,7 +119733,7 @@ } }, { - "accuracy": 0.9956828736160931, + "accuracy": 0.9955973609497673, "total_bits": 637362176, "q_proj": { "group_size": { @@ -119794,7 +119794,7 @@ } }, { - "accuracy": 0.9960351488307903, + "accuracy": 0.9959951439186147, "total_bits": 646823936, "q_proj": { "group_size": { @@ -119855,7 +119855,7 @@ } }, { - "accuracy": 0.9972815646937019, + "accuracy": 0.9973123838242731, "total_bits": 784740352, "q_proj": { "group_size": { @@ -119916,7 +119916,7 @@ } }, { - "accuracy": 0.9977832826736727, + "accuracy": 0.9977227453338472, "total_bits": 797818880, "q_proj": { "group_size": { @@ -119977,7 +119977,7 @@ } }, { - "accuracy": 0.9974293214710135, + "accuracy": 0.9974606627304303, "total_bits": 911749120, "q_proj": { "group_size": { @@ -120029,7 +120029,7 @@ } }, { - "accuracy": 0.9992015193657655, + "accuracy": 0.9992084942365947, "total_bits": 942718976, "q_proj": { "group_size": { @@ -120081,7 +120081,7 @@ } }, { - "accuracy": 0.9993260185185232, + "accuracy": 0.9993183374110806, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -120135,7 +120135,7 @@ ], "model.layers.61.mlp": [ { - "accuracy": 0.9467084031356008, + "accuracy": 0.9470049582029644, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -120187,7 +120187,7 @@ } }, { - "accuracy": 0.9480967490296615, + "accuracy": 0.9483971062459444, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -120239,7 +120239,7 @@ } }, { - "accuracy": 0.9546087823416057, + "accuracy": 0.9548575282096863, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -120288,7 +120288,7 @@ } }, { - "accuracy": 0.9564303504793268, + "accuracy": 0.9566706858183208, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -120337,7 +120337,7 @@ } }, { - "accuracy": 0.9730409132806879, + "accuracy": 0.9731926353354203, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -120389,7 +120389,7 @@ } }, { - "accuracy": 0.9751395896861428, + "accuracy": 0.9752752420149351, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -120441,7 +120441,7 @@ } }, { - "accuracy": 0.9778776576644496, + "accuracy": 0.9779988699837735, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -120490,7 +120490,7 @@ } }, { - "accuracy": 0.986225132879458, + "accuracy": 0.9863048509547585, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -120533,7 +120533,7 @@ } }, { - "accuracy": 0.9873419827536533, + "accuracy": 0.9874109178781509, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -120576,7 +120576,7 @@ } }, { - "accuracy": 0.9863044760729137, + "accuracy": 0.9863856058371695, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -120628,7 +120628,7 @@ } }, { - "accuracy": 0.9879322585306669, + "accuracy": 0.9880001152816572, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -120680,7 +120680,7 @@ } }, { - "accuracy": 0.9930352168647867, + "accuracy": 0.9930781118179622, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -120732,7 +120732,7 @@ } }, { - "accuracy": 0.993987257543363, + "accuracy": 0.9940218172575298, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -120784,7 +120784,7 @@ } }, { - "accuracy": 0.9962864118187051, + "accuracy": 0.9963099831028989, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -120827,7 +120827,7 @@ } }, { - "accuracy": 0.9964421791465659, + "accuracy": 0.9964613569410223, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -120876,7 +120876,7 @@ } }, { - "accuracy": 0.9969553282778514, + "accuracy": 0.9969715099585684, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -120922,7 +120922,7 @@ } }, { - "accuracy": 0.9990017922889245, + "accuracy": 0.9990080888138005, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -120964,7 +120964,7 @@ ], "model.layers.62.self_attn": [ { - "accuracy": 0.987798129257403, + "accuracy": 0.9879055187890404, "total_bits": 320757760, "q_proj": { "group_size": { @@ -121028,7 +121028,7 @@ } }, { - "accuracy": 0.9883374712969127, + "accuracy": 0.9884194425846401, "total_bits": 329080832, "q_proj": { "group_size": { @@ -121092,7 +121092,7 @@ } }, { - "accuracy": 0.9905390872767097, + "accuracy": 0.9904907157546595, "total_bits": 336024576, "q_proj": { "group_size": { @@ -121156,7 +121156,7 @@ } }, { - "accuracy": 0.9922894584505182, + "accuracy": 0.9922887431947809, "total_bits": 401557504, "q_proj": { "group_size": { @@ -121220,7 +121220,7 @@ } }, { - "accuracy": 0.9926662280371314, + "accuracy": 0.9926462691081198, "total_bits": 475279360, "q_proj": { "group_size": { @@ -121284,7 +121284,7 @@ } }, { - "accuracy": 0.9934994531305212, + "accuracy": 0.9935622129001116, "total_bits": 475479040, "q_proj": { "group_size": { @@ -121348,7 +121348,7 @@ } }, { - "accuracy": 0.9939317750303369, + "accuracy": 0.9938962185069135, "total_bits": 609759232, "q_proj": { "group_size": { @@ -121400,7 +121400,7 @@ } }, { - "accuracy": 0.9948685616254807, + "accuracy": 0.9949211176288756, "total_bits": 610024448, "q_proj": { "group_size": { @@ -121452,7 +121452,7 @@ } }, { - "accuracy": 0.9953876349486803, + "accuracy": 0.9954725131392479, "total_bits": 615020544, "q_proj": { "group_size": { @@ -121504,7 +121504,7 @@ } }, { - "accuracy": 0.9954220774142366, + "accuracy": 0.995503737738258, "total_bits": 623951872, "q_proj": { "group_size": { @@ -121556,7 +121556,7 @@ } }, { - "accuracy": 0.9963315937079882, + "accuracy": 0.9964008433254141, "total_bits": 626473984, "q_proj": { "group_size": { @@ -121620,7 +121620,7 @@ } }, { - "accuracy": 0.9965766468330434, + "accuracy": 0.9965987891743058, "total_bits": 630355968, "q_proj": { "group_size": { @@ -121684,7 +121684,7 @@ } }, { - "accuracy": 0.9967070704227999, + "accuracy": 0.9967527085621106, "total_bits": 637362176, "q_proj": { "group_size": { @@ -121745,7 +121745,7 @@ } }, { - "accuracy": 0.9969257337874488, + "accuracy": 0.9969489903826463, "total_bits": 646823936, "q_proj": { "group_size": { @@ -121806,7 +121806,7 @@ } }, { - "accuracy": 0.9975505353588807, + "accuracy": 0.9975564291602687, "total_bits": 784740352, "q_proj": { "group_size": { @@ -121867,7 +121867,7 @@ } }, { - "accuracy": 0.9984063416915504, + "accuracy": 0.9984150878888997, "total_bits": 797818880, "q_proj": { "group_size": { @@ -121928,7 +121928,7 @@ } }, { - "accuracy": 0.997657255318604, + "accuracy": 0.9976633618536749, "total_bits": 911749120, "q_proj": { "group_size": { @@ -121980,7 +121980,7 @@ } }, { - "accuracy": 0.9993905098128476, + "accuracy": 0.9993882206219592, "total_bits": 942718976, "q_proj": { "group_size": { @@ -122032,7 +122032,7 @@ } }, { - "accuracy": 0.9993752637191823, + "accuracy": 0.999385707717585, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -122086,7 +122086,7 @@ ], "model.layers.62.mlp": [ { - "accuracy": 0.9467521780415585, + "accuracy": 0.9470622978712383, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -122138,7 +122138,7 @@ } }, { - "accuracy": 0.9481369853019714, + "accuracy": 0.9484273446233649, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -122190,7 +122190,7 @@ } }, { - "accuracy": 0.9546577365774858, + "accuracy": 0.9549165901384855, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -122239,7 +122239,7 @@ } }, { - "accuracy": 0.9564441097410101, + "accuracy": 0.9566907788577833, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -122288,7 +122288,7 @@ } }, { - "accuracy": 0.9730446558249625, + "accuracy": 0.9731941442740591, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -122340,7 +122340,7 @@ } }, { - "accuracy": 0.9751368463039398, + "accuracy": 0.9752855912635201, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -122392,7 +122392,7 @@ } }, { - "accuracy": 0.9778767714374944, + "accuracy": 0.9780097854764838, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -122441,7 +122441,7 @@ } }, { - "accuracy": 0.9862547908958635, + "accuracy": 0.9863386436512596, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -122484,7 +122484,7 @@ } }, { - "accuracy": 0.9873705467111186, + "accuracy": 0.9874456924827475, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -122527,7 +122527,7 @@ } }, { - "accuracy": 0.9863053089693973, + "accuracy": 0.986383529085862, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -122579,7 +122579,7 @@ } }, { - "accuracy": 0.9879336615926341, + "accuracy": 0.9880072568592272, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -122631,7 +122631,7 @@ } }, { - "accuracy": 0.9930359234935359, + "accuracy": 0.993079358025601, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -122683,7 +122683,7 @@ } }, { - "accuracy": 0.9939866164012959, + "accuracy": 0.9940234257986671, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -122735,7 +122735,7 @@ } }, { - "accuracy": 0.996295484665193, + "accuracy": 0.9963178928745421, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -122778,7 +122778,7 @@ } }, { - "accuracy": 0.9964285583088273, + "accuracy": 0.9964501014665553, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -122827,7 +122827,7 @@ } }, { - "accuracy": 0.9969372502283046, + "accuracy": 0.9969561648996252, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -122873,7 +122873,7 @@ } }, { - "accuracy": 0.9989498121369826, + "accuracy": 0.998955713683053, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -122915,7 +122915,7 @@ ], "model.layers.63.self_attn": [ { - "accuracy": 0.9848332624686392, + "accuracy": 0.9849690230269181, "total_bits": 320757760, "q_proj": { "group_size": { @@ -122979,7 +122979,7 @@ } }, { - "accuracy": 0.9856680192445454, + "accuracy": 0.985709537016718, "total_bits": 329080832, "q_proj": { "group_size": { @@ -123043,7 +123043,7 @@ } }, { - "accuracy": 0.9882192423469142, + "accuracy": 0.9883561738227543, "total_bits": 336024576, "q_proj": { "group_size": { @@ -123107,7 +123107,7 @@ } }, { - "accuracy": 0.9904338419437408, + "accuracy": 0.9905238653484144, "total_bits": 401557504, "q_proj": { "group_size": { @@ -123171,7 +123171,7 @@ } }, { - "accuracy": 0.9915329531619423, + "accuracy": 0.991599399008249, "total_bits": 475279360, "q_proj": { "group_size": { @@ -123235,7 +123235,7 @@ } }, { - "accuracy": 0.9920317381620407, + "accuracy": 0.992143778424514, "total_bits": 475479040, "q_proj": { "group_size": { @@ -123299,7 +123299,7 @@ } }, { - "accuracy": 0.9932122826576233, + "accuracy": 0.9931903928518295, "total_bits": 609759232, "q_proj": { "group_size": { @@ -123351,7 +123351,7 @@ } }, { - "accuracy": 0.9936989616406592, + "accuracy": 0.9937756814454731, "total_bits": 610024448, "q_proj": { "group_size": { @@ -123403,7 +123403,7 @@ } }, { - "accuracy": 0.994277793325876, + "accuracy": 0.9943413365828363, "total_bits": 615020544, "q_proj": { "group_size": { @@ -123455,7 +123455,7 @@ } }, { - "accuracy": 0.9943792843504956, + "accuracy": 0.9944381898171023, "total_bits": 623951872, "q_proj": { "group_size": { @@ -123507,7 +123507,7 @@ } }, { - "accuracy": 0.9955674719653631, + "accuracy": 0.9955105852139624, "total_bits": 626473984, "q_proj": { "group_size": { @@ -123571,7 +123571,7 @@ } }, { - "accuracy": 0.9959607077272314, + "accuracy": 0.9959723553375194, "total_bits": 630355968, "q_proj": { "group_size": { @@ -123635,7 +123635,7 @@ } }, { - "accuracy": 0.9958901311221876, + "accuracy": 0.9958406161320837, "total_bits": 637362176, "q_proj": { "group_size": { @@ -123696,7 +123696,7 @@ } }, { - "accuracy": 0.9963408963460672, + "accuracy": 0.9963343221890298, "total_bits": 646823936, "q_proj": { "group_size": { @@ -123757,7 +123757,7 @@ } }, { - "accuracy": 0.9976364948639744, + "accuracy": 0.9977170981858906, "total_bits": 784740352, "q_proj": { "group_size": { @@ -123818,7 +123818,7 @@ } }, { - "accuracy": 0.9980832497148138, + "accuracy": 0.998073079672299, "total_bits": 797818880, "q_proj": { "group_size": { @@ -123879,7 +123879,7 @@ } }, { - "accuracy": 0.9978396180821093, + "accuracy": 0.9979274523885626, "total_bits": 911749120, "q_proj": { "group_size": { @@ -123931,7 +123931,7 @@ } }, { - "accuracy": 0.9992215063815054, + "accuracy": 0.9992315925069546, "total_bits": 942718976, "q_proj": { "group_size": { @@ -123983,7 +123983,7 @@ } }, { - "accuracy": 0.9994191440116418, + "accuracy": 0.9993993507109975, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -124037,7 +124037,7 @@ ], "model.layers.63.mlp": [ { - "accuracy": 0.9453399683299818, + "accuracy": 0.9456410721728676, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -124089,7 +124089,7 @@ } }, { - "accuracy": 0.9467709064483643, + "accuracy": 0.9470702974419845, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -124141,7 +124141,7 @@ } }, { - "accuracy": 0.9534110771982294, + "accuracy": 0.9536710318766142, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -124190,7 +124190,7 @@ } }, { - "accuracy": 0.9552448391914368, + "accuracy": 0.955493754462192, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -124239,7 +124239,7 @@ } }, { - "accuracy": 0.9723404551807203, + "accuracy": 0.9724983761185094, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -124291,7 +124291,7 @@ } }, { - "accuracy": 0.9744852514643418, + "accuracy": 0.9746329141290564, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -124343,7 +124343,7 @@ } }, { - "accuracy": 0.9772710847227197, + "accuracy": 0.9773999389849211, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -124392,7 +124392,7 @@ } }, { - "accuracy": 0.9858801992315995, + "accuracy": 0.9859619564131686, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -124435,7 +124435,7 @@ } }, { - "accuracy": 0.9870127964960901, + "accuracy": 0.9870905248742354, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -124478,7 +124478,7 @@ } }, { - "accuracy": 0.9859519177361539, + "accuracy": 0.9860346348662126, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -124530,7 +124530,7 @@ } }, { - "accuracy": 0.9876092688033455, + "accuracy": 0.9876815072799984, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -124582,7 +124582,7 @@ } }, { - "accuracy": 0.9928600341081619, + "accuracy": 0.992900657810663, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -124634,7 +124634,7 @@ } }, { - "accuracy": 0.9938271622124472, + "accuracy": 0.9938623430697542, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -124686,7 +124686,7 @@ } }, { - "accuracy": 0.9961958225620421, + "accuracy": 0.9962187201568955, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -124729,7 +124729,7 @@ } }, { - "accuracy": 0.9963524314133745, + "accuracy": 0.9963739565328548, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -124778,7 +124778,7 @@ } }, { - "accuracy": 0.9968741383207472, + "accuracy": 0.9968930951466686, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -124824,7 +124824,7 @@ } }, { - "accuracy": 0.9989813342690468, + "accuracy": 0.998987332672665, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -124866,7 +124866,7 @@ ], "model.layers.64.self_attn": [ { - "accuracy": 0.9612342715263367, + "accuracy": 0.959673451749902, "total_bits": 320757760, "q_proj": { "group_size": { @@ -124930,7 +124930,7 @@ } }, { - "accuracy": 0.9639760381297061, + "accuracy": 0.962479359225223, "total_bits": 329080832, "q_proj": { "group_size": { @@ -124994,7 +124994,7 @@ } }, { - "accuracy": 0.976591307865946, + "accuracy": 0.9766104566423517, "total_bits": 336024576, "q_proj": { "group_size": { @@ -125058,7 +125058,7 @@ } }, { - "accuracy": 0.9794718874128241, + "accuracy": 0.9794882332023821, "total_bits": 401557504, "q_proj": { "group_size": { @@ -125122,7 +125122,7 @@ } }, { - "accuracy": 0.9817855107156854, + "accuracy": 0.9818383863097743, "total_bits": 475279360, "q_proj": { "group_size": { @@ -125186,7 +125186,7 @@ } }, { - "accuracy": 0.9821480873383974, + "accuracy": 0.9821225012603559, "total_bits": 475479040, "q_proj": { "group_size": { @@ -125250,7 +125250,7 @@ } }, { - "accuracy": 0.9891321172839717, + "accuracy": 0.9886361737000314, "total_bits": 609759232, "q_proj": { "group_size": { @@ -125302,7 +125302,7 @@ } }, { - "accuracy": 0.9895083363118925, + "accuracy": 0.9893183700348201, "total_bits": 610024448, "q_proj": { "group_size": { @@ -125354,7 +125354,7 @@ } }, { - "accuracy": 0.9904286037934454, + "accuracy": 0.9900589797057604, "total_bits": 615020544, "q_proj": { "group_size": { @@ -125406,7 +125406,7 @@ } }, { - "accuracy": 0.9905912295768136, + "accuracy": 0.9902222172210091, "total_bits": 623951872, "q_proj": { "group_size": { @@ -125458,7 +125458,7 @@ } }, { - "accuracy": 0.9902796886469188, + "accuracy": 0.99008961414036, "total_bits": 626473984, "q_proj": { "group_size": { @@ -125522,7 +125522,7 @@ } }, { - "accuracy": 0.9910481760376378, + "accuracy": 0.9907773709610889, "total_bits": 630355968, "q_proj": { "group_size": { @@ -125586,7 +125586,7 @@ } }, { - "accuracy": 0.9934319147938176, + "accuracy": 0.993672389733164, "total_bits": 637362176, "q_proj": { "group_size": { @@ -125647,7 +125647,7 @@ } }, { - "accuracy": 0.9941026501749691, + "accuracy": 0.9942229076435691, "total_bits": 646823936, "q_proj": { "group_size": { @@ -125708,7 +125708,7 @@ } }, { - "accuracy": 0.9963697595031638, + "accuracy": 0.9963652079826907, "total_bits": 784740352, "q_proj": { "group_size": { @@ -125769,7 +125769,7 @@ } }, { - "accuracy": 0.9970310781347124, + "accuracy": 0.9970274495058938, "total_bits": 797818880, "q_proj": { "group_size": { @@ -125830,7 +125830,7 @@ } }, { - "accuracy": 0.9967682624333783, + "accuracy": 0.9967514543156875, "total_bits": 911749120, "q_proj": { "group_size": { @@ -125882,7 +125882,7 @@ } }, { - "accuracy": 0.9987691050689471, + "accuracy": 0.9987614562636927, "total_bits": 942718976, "q_proj": { "group_size": { @@ -125934,7 +125934,7 @@ } }, { - "accuracy": 0.9991455320268869, + "accuracy": 0.9991754234621399, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -125988,7 +125988,7 @@ ], "model.layers.64.mlp": [ { - "accuracy": 0.9438887709065488, + "accuracy": 0.944195314457542, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -126040,7 +126040,7 @@ } }, { - "accuracy": 0.9453709501969187, + "accuracy": 0.9456706360766762, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -126092,7 +126092,7 @@ } }, { - "accuracy": 0.9521120692554274, + "accuracy": 0.952382859430815, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -126141,7 +126141,7 @@ } }, { - "accuracy": 0.9540201456923234, + "accuracy": 0.9542783059571919, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -126190,7 +126190,7 @@ } }, { - "accuracy": 0.9715995349382099, + "accuracy": 0.9717675416093123, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -126242,7 +126242,7 @@ } }, { - "accuracy": 0.9738027343624517, + "accuracy": 0.9739492943412379, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -126294,7 +126294,7 @@ } }, { - "accuracy": 0.9766444673663691, + "accuracy": 0.9767746046969765, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -126343,7 +126343,7 @@ } }, { - "accuracy": 0.9854651767956583, + "accuracy": 0.9855425326447738, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -126386,7 +126386,7 @@ } }, { - "accuracy": 0.986632293776462, + "accuracy": 0.9867079697157207, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -126429,7 +126429,7 @@ } }, { - "accuracy": 0.9855724228055853, + "accuracy": 0.9856552528707605, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -126481,7 +126481,7 @@ } }, { - "accuracy": 0.9872758019911615, + "accuracy": 0.9873484184867457, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -126533,7 +126533,7 @@ } }, { - "accuracy": 0.9926622502113643, + "accuracy": 0.9927021429726952, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -126585,7 +126585,7 @@ } }, { - "accuracy": 0.9936535742722059, + "accuracy": 0.9936903515144399, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -126637,7 +126637,7 @@ } }, { - "accuracy": 0.9960683070515332, + "accuracy": 0.9960901262728792, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -126680,7 +126680,7 @@ } }, { - "accuracy": 0.9962361654952953, + "accuracy": 0.9962572200517905, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -126729,7 +126729,7 @@ } }, { - "accuracy": 0.9967678038305358, + "accuracy": 0.9967855671910864, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -126775,7 +126775,7 @@ } }, { - "accuracy": 0.9988851171771163, + "accuracy": 0.9988910321538386, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -126817,7 +126817,7 @@ ], "model.layers.65.self_attn": [ { - "accuracy": 0.9857259800559596, + "accuracy": 0.9859549842382732, "total_bits": 320757760, "q_proj": { "group_size": { @@ -126881,7 +126881,7 @@ } }, { - "accuracy": 0.9865271260863856, + "accuracy": 0.9866217798308322, "total_bits": 329080832, "q_proj": { "group_size": { @@ -126945,7 +126945,7 @@ } }, { - "accuracy": 0.9883712420338079, + "accuracy": 0.988494400131075, "total_bits": 336024576, "q_proj": { "group_size": { @@ -127009,7 +127009,7 @@ } }, { - "accuracy": 0.9903496394031926, + "accuracy": 0.9906278417298668, "total_bits": 401557504, "q_proj": { "group_size": { @@ -127073,7 +127073,7 @@ } }, { - "accuracy": 0.9912963682099393, + "accuracy": 0.9914681927153939, "total_bits": 475279360, "q_proj": { "group_size": { @@ -127137,7 +127137,7 @@ } }, { - "accuracy": 0.9920228743239453, + "accuracy": 0.9921597312939795, "total_bits": 475479040, "q_proj": { "group_size": { @@ -127201,7 +127201,7 @@ } }, { - "accuracy": 0.9927175680273458, + "accuracy": 0.9929823679359335, "total_bits": 609759232, "q_proj": { "group_size": { @@ -127253,7 +127253,7 @@ } }, { - "accuracy": 0.9935887440254814, + "accuracy": 0.993740340988887, "total_bits": 610024448, "q_proj": { "group_size": { @@ -127305,7 +127305,7 @@ } }, { - "accuracy": 0.9951740522917948, + "accuracy": 0.9951921615161394, "total_bits": 615020544, "q_proj": { "group_size": { @@ -127357,7 +127357,7 @@ } }, { - "accuracy": 0.9952824154966756, + "accuracy": 0.9952981864151201, "total_bits": 623951872, "q_proj": { "group_size": { @@ -127409,7 +127409,7 @@ } }, { - "accuracy": 0.9954923144296596, + "accuracy": 0.9954871507851701, "total_bits": 626473984, "q_proj": { "group_size": { @@ -127473,7 +127473,7 @@ } }, { - "accuracy": 0.996081708685348, + "accuracy": 0.9961230841122175, "total_bits": 630355968, "q_proj": { "group_size": { @@ -127537,7 +127537,7 @@ } }, { - "accuracy": 0.9958282097389823, + "accuracy": 0.9958308452838346, "total_bits": 637362176, "q_proj": { "group_size": { @@ -127598,7 +127598,7 @@ } }, { - "accuracy": 0.9964959417518816, + "accuracy": 0.9965333887620976, "total_bits": 646823936, "q_proj": { "group_size": { @@ -127659,7 +127659,7 @@ } }, { - "accuracy": 0.9973654970526695, + "accuracy": 0.9974093684240392, "total_bits": 784740352, "q_proj": { "group_size": { @@ -127720,7 +127720,7 @@ } }, { - "accuracy": 0.9980278383744391, + "accuracy": 0.9980371549333397, "total_bits": 797818880, "q_proj": { "group_size": { @@ -127781,7 +127781,7 @@ } }, { - "accuracy": 0.9975048411441477, + "accuracy": 0.9975490793585777, "total_bits": 911749120, "q_proj": { "group_size": { @@ -127833,7 +127833,7 @@ } }, { - "accuracy": 0.9992067012935877, + "accuracy": 0.999210161990241, "total_bits": 942718976, "q_proj": { "group_size": { @@ -127885,7 +127885,7 @@ } }, { - "accuracy": 0.9992332331051952, + "accuracy": 0.9992343408398723, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -127939,7 +127939,7 @@ ], "model.layers.65.mlp": [ { - "accuracy": 0.943446347587987, + "accuracy": 0.9437939305054515, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -127991,7 +127991,7 @@ } }, { - "accuracy": 0.9449616043191207, + "accuracy": 0.9452927677254928, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -128043,7 +128043,7 @@ } }, { - "accuracy": 0.9516874081210086, + "accuracy": 0.9519863818821154, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -128092,7 +128092,7 @@ } }, { - "accuracy": 0.9535779984373796, + "accuracy": 0.9538649540198476, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -128141,7 +128141,7 @@ } }, { - "accuracy": 0.9713825332491022, + "accuracy": 0.971560471936276, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -128193,7 +128193,7 @@ } }, { - "accuracy": 0.9735991421498751, + "accuracy": 0.9737621200712103, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -128245,7 +128245,7 @@ } }, { - "accuracy": 0.9764294843924674, + "accuracy": 0.9765777776115819, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -128294,7 +128294,7 @@ } }, { - "accuracy": 0.985366835405952, + "accuracy": 0.9854584916641838, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -128337,7 +128337,7 @@ } }, { - "accuracy": 0.9865373231862721, + "accuracy": 0.9866189141022531, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -128380,7 +128380,7 @@ } }, { - "accuracy": 0.9854681209514016, + "accuracy": 0.9855533555934304, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -128432,7 +128432,7 @@ } }, { - "accuracy": 0.9871789687558225, + "accuracy": 0.9872549521295648, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -128484,7 +128484,7 @@ } }, { - "accuracy": 0.9926147562892813, + "accuracy": 0.992660746762627, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -128536,7 +128536,7 @@ } }, { - "accuracy": 0.9936122878601676, + "accuracy": 0.9936525229560701, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -128588,7 +128588,7 @@ } }, { - "accuracy": 0.9960594079212138, + "accuracy": 0.9960847477379599, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -128631,7 +128631,7 @@ } }, { - "accuracy": 0.996224838259973, + "accuracy": 0.9962489679455757, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -128680,7 +128680,7 @@ } }, { - "accuracy": 0.9967555842901531, + "accuracy": 0.9967771176444856, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -128726,7 +128726,7 @@ } }, { - "accuracy": 0.9989404039163339, + "accuracy": 0.9989473282506591, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -128768,7 +128768,7 @@ ], "model.layers.66.self_attn": [ { - "accuracy": 0.9821379592544154, + "accuracy": 0.9824391493671819, "total_bits": 320757760, "q_proj": { "group_size": { @@ -128832,7 +128832,7 @@ } }, { - "accuracy": 0.9832240892084021, + "accuracy": 0.983380565517827, "total_bits": 329080832, "q_proj": { "group_size": { @@ -128896,7 +128896,7 @@ } }, { - "accuracy": 0.9861854157949749, + "accuracy": 0.986210076432479, "total_bits": 336024576, "q_proj": { "group_size": { @@ -128960,7 +128960,7 @@ } }, { - "accuracy": 0.9889904001825734, + "accuracy": 0.9890080110022896, "total_bits": 401557504, "q_proj": { "group_size": { @@ -129024,7 +129024,7 @@ } }, { - "accuracy": 0.9900476516861665, + "accuracy": 0.990054765814229, "total_bits": 475279360, "q_proj": { "group_size": { @@ -129088,7 +129088,7 @@ } }, { - "accuracy": 0.990634422553213, + "accuracy": 0.990631239194619, "total_bits": 475479040, "q_proj": { "group_size": { @@ -129152,7 +129152,7 @@ } }, { - "accuracy": 0.992154996646078, + "accuracy": 0.9921125189254159, "total_bits": 609759232, "q_proj": { "group_size": { @@ -129204,7 +129204,7 @@ } }, { - "accuracy": 0.9928280609218698, + "accuracy": 0.9928176308933058, "total_bits": 610024448, "q_proj": { "group_size": { @@ -129256,7 +129256,7 @@ } }, { - "accuracy": 0.99369116127491, + "accuracy": 0.993629125388045, "total_bits": 615020544, "q_proj": { "group_size": { @@ -129308,7 +129308,7 @@ } }, { - "accuracy": 0.9938450239990887, + "accuracy": 0.9937984315972579, "total_bits": 623951872, "q_proj": { "group_size": { @@ -129360,7 +129360,7 @@ } }, { - "accuracy": 0.9946816708696516, + "accuracy": 0.994719611578866, "total_bits": 626473984, "q_proj": { "group_size": { @@ -129424,7 +129424,7 @@ } }, { - "accuracy": 0.995229955174421, + "accuracy": 0.9952413816201059, "total_bits": 630355968, "q_proj": { "group_size": { @@ -129488,7 +129488,7 @@ } }, { - "accuracy": 0.9952023335193333, + "accuracy": 0.9952630922198296, "total_bits": 637362176, "q_proj": { "group_size": { @@ -129549,7 +129549,7 @@ } }, { - "accuracy": 0.9957494747481848, + "accuracy": 0.9958029930528841, "total_bits": 646823936, "q_proj": { "group_size": { @@ -129610,7 +129610,7 @@ } }, { - "accuracy": 0.9972571393376902, + "accuracy": 0.9972702990237036, "total_bits": 784740352, "q_proj": { "group_size": { @@ -129671,7 +129671,7 @@ } }, { - "accuracy": 0.9978441437216181, + "accuracy": 0.9978582078689023, "total_bits": 797818880, "q_proj": { "group_size": { @@ -129732,7 +129732,7 @@ } }, { - "accuracy": 0.9974968588273776, + "accuracy": 0.9975086123143372, "total_bits": 911749120, "q_proj": { "group_size": { @@ -129784,7 +129784,7 @@ } }, { - "accuracy": 0.999062563733835, + "accuracy": 0.9990663124542487, "total_bits": 942718976, "q_proj": { "group_size": { @@ -129836,7 +129836,7 @@ } }, { - "accuracy": 0.9993167545548395, + "accuracy": 0.9993162390432859, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -129890,7 +129890,7 @@ ], "model.layers.66.mlp": [ { - "accuracy": 0.9438636804881849, + "accuracy": 0.9441965379213032, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -129942,7 +129942,7 @@ } }, { - "accuracy": 0.9453435948020533, + "accuracy": 0.9456821554585507, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -129994,7 +129994,7 @@ } }, { - "accuracy": 0.9520565428231892, + "accuracy": 0.9523621044660869, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -130043,7 +130043,7 @@ } }, { - "accuracy": 0.9538899911077399, + "accuracy": 0.9541843878595453, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -130092,7 +130092,7 @@ } }, { - "accuracy": 0.9715566572390104, + "accuracy": 0.9717403430687753, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -130144,7 +130144,7 @@ } }, { - "accuracy": 0.9737720991435804, + "accuracy": 0.9739360825011605, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -130196,7 +130196,7 @@ } }, { - "accuracy": 0.9765964824902383, + "accuracy": 0.9767420197788038, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -130245,7 +130245,7 @@ } }, { - "accuracy": 0.985482539001264, + "accuracy": 0.9855814042844271, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -130288,7 +130288,7 @@ } }, { - "accuracy": 0.9866547129656139, + "accuracy": 0.9867413734134874, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -130331,7 +130331,7 @@ } }, { - "accuracy": 0.985542324028517, + "accuracy": 0.9856371848206771, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -130383,7 +130383,7 @@ } }, { - "accuracy": 0.9872544556856155, + "accuracy": 0.987334776081537, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -130435,7 +130435,7 @@ } }, { - "accuracy": 0.9926447899718034, + "accuracy": 0.9926943104518088, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -130487,7 +130487,7 @@ } }, { - "accuracy": 0.9936461142803493, + "accuracy": 0.9936881488875339, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -130539,7 +130539,7 @@ } }, { - "accuracy": 0.9960818039743524, + "accuracy": 0.996108762527767, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -130582,7 +130582,7 @@ } }, { - "accuracy": 0.9962216898014671, + "accuracy": 0.9962480017229131, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -130631,7 +130631,7 @@ } }, { - "accuracy": 0.9967447866342569, + "accuracy": 0.9967683081172014, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -130677,7 +130677,7 @@ } }, { - "accuracy": 0.998880942597201, + "accuracy": 0.998889844570505, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -130719,7 +130719,7 @@ ], "model.layers.67.self_attn": [ { - "accuracy": 0.9609534960044057, + "accuracy": 0.9596483017268934, "total_bits": 320757760, "q_proj": { "group_size": { @@ -130783,7 +130783,7 @@ } }, { - "accuracy": 0.9609217079062211, + "accuracy": 0.9624556961812472, "total_bits": 329080832, "q_proj": { "group_size": { @@ -130847,7 +130847,7 @@ } }, { - "accuracy": 0.9766820982882851, + "accuracy": 0.9773485033135665, "total_bits": 336024576, "q_proj": { "group_size": { @@ -130911,7 +130911,7 @@ } }, { - "accuracy": 0.9797116800358421, + "accuracy": 0.9803220739490107, "total_bits": 401557504, "q_proj": { "group_size": { @@ -130975,7 +130975,7 @@ } }, { - "accuracy": 0.9819115635595823, + "accuracy": 0.9825714669729534, "total_bits": 475279360, "q_proj": { "group_size": { @@ -131039,7 +131039,7 @@ } }, { - "accuracy": 0.9829012861377314, + "accuracy": 0.9816983373541581, "total_bits": 475479040, "q_proj": { "group_size": { @@ -131103,7 +131103,7 @@ } }, { - "accuracy": 0.9897047304793408, + "accuracy": 0.9894436323329022, "total_bits": 609759232, "q_proj": { "group_size": { @@ -131155,7 +131155,7 @@ } }, { - "accuracy": 0.989957908266469, + "accuracy": 0.9900826105945989, "total_bits": 610024448, "q_proj": { "group_size": { @@ -131207,7 +131207,7 @@ } }, { - "accuracy": 0.990889236330986, + "accuracy": 0.9909424083797556, "total_bits": 615020544, "q_proj": { "group_size": { @@ -131259,7 +131259,7 @@ } }, { - "accuracy": 0.9910206975121247, + "accuracy": 0.9910853779629657, "total_bits": 623951872, "q_proj": { "group_size": { @@ -131311,7 +131311,7 @@ } }, { - "accuracy": 0.9906162407837416, + "accuracy": 0.9904378456504721, "total_bits": 626473984, "q_proj": { "group_size": { @@ -131375,7 +131375,7 @@ } }, { - "accuracy": 0.9912718236446381, + "accuracy": 0.9912401978906832, "total_bits": 630355968, "q_proj": { "group_size": { @@ -131439,7 +131439,7 @@ } }, { - "accuracy": 0.9940109139210299, + "accuracy": 0.993947413406874, "total_bits": 637362176, "q_proj": { "group_size": { @@ -131500,7 +131500,7 @@ } }, { - "accuracy": 0.9945488274097443, + "accuracy": 0.994540000432416, "total_bits": 646823936, "q_proj": { "group_size": { @@ -131561,7 +131561,7 @@ } }, { - "accuracy": 0.996563504793142, + "accuracy": 0.9966456329351977, "total_bits": 784740352, "q_proj": { "group_size": { @@ -131622,7 +131622,7 @@ } }, { - "accuracy": 0.9971607315697169, + "accuracy": 0.9971794889945733, "total_bits": 797818880, "q_proj": { "group_size": { @@ -131683,7 +131683,7 @@ } }, { - "accuracy": 0.997003506084806, + "accuracy": 0.9971032717118138, "total_bits": 911749120, "q_proj": { "group_size": { @@ -131735,7 +131735,7 @@ } }, { - "accuracy": 0.9988423028078518, + "accuracy": 0.998859591782093, "total_bits": 942718976, "q_proj": { "group_size": { @@ -131787,7 +131787,7 @@ } }, { - "accuracy": 0.9992397651076317, + "accuracy": 0.9992253047561175, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -131841,7 +131841,7 @@ ], "model.layers.67.mlp": [ { - "accuracy": 0.942307848679392, + "accuracy": 0.9426512843684146, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -131893,7 +131893,7 @@ } }, { - "accuracy": 0.94382895921406, + "accuracy": 0.9441774769833213, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -131945,7 +131945,7 @@ } }, { - "accuracy": 0.9506990846834684, + "accuracy": 0.9510059701768976, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -131994,7 +131994,7 @@ } }, { - "accuracy": 0.9525863214542991, + "accuracy": 0.9528830584726835, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -132043,7 +132043,7 @@ } }, { - "accuracy": 0.9707813106085125, + "accuracy": 0.9709694730608087, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -132095,7 +132095,7 @@ } }, { - "accuracy": 0.9730525769685444, + "accuracy": 0.9732202228746916, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -132147,7 +132147,7 @@ } }, { - "accuracy": 0.9759402353512613, + "accuracy": 0.9760893348016237, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -132196,7 +132196,7 @@ } }, { - "accuracy": 0.9850847109367973, + "accuracy": 0.9851793882093931, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -132239,7 +132239,7 @@ } }, { - "accuracy": 0.9862796413271051, + "accuracy": 0.9863681306964472, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -132282,7 +132282,7 @@ } }, { - "accuracy": 0.9851538601674532, + "accuracy": 0.9852459556178043, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -132334,7 +132334,7 @@ } }, { - "accuracy": 0.9869086256152705, + "accuracy": 0.9869917984071531, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -132386,7 +132386,7 @@ } }, { - "accuracy": 0.9924544174420206, + "accuracy": 0.9925030211084768, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -132438,7 +132438,7 @@ } }, { - "accuracy": 0.9934799369228514, + "accuracy": 0.9935220119200254, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -132490,7 +132490,7 @@ } }, { - "accuracy": 0.9959850464212266, + "accuracy": 0.996011029340719, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -132533,7 +132533,7 @@ } }, { - "accuracy": 0.9961448641199815, + "accuracy": 0.9961692122252364, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -132582,7 +132582,7 @@ } }, { - "accuracy": 0.9966846488808331, + "accuracy": 0.9967056238337567, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -132628,7 +132628,7 @@ } }, { - "accuracy": 0.9989304168051795, + "accuracy": 0.9989371250726675, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -132670,7 +132670,7 @@ ], "model.layers.68.self_attn": [ { - "accuracy": 0.9419417318544889, + "accuracy": 0.9452797048970273, "total_bits": 320757760, "q_proj": { "group_size": { @@ -132734,7 +132734,7 @@ } }, { - "accuracy": 0.9465973314486051, + "accuracy": 0.9476478413531655, "total_bits": 329080832, "q_proj": { "group_size": { @@ -132798,7 +132798,7 @@ } }, { - "accuracy": 0.9689077672205473, + "accuracy": 0.968694762179726, "total_bits": 336024576, "q_proj": { "group_size": { @@ -132862,7 +132862,7 @@ } }, { - "accuracy": 0.9718637340947202, + "accuracy": 0.9720780598489862, "total_bits": 401557504, "q_proj": { "group_size": { @@ -132926,7 +132926,7 @@ } }, { - "accuracy": 0.9753499815338537, + "accuracy": 0.9756318236652174, "total_bits": 475279360, "q_proj": { "group_size": { @@ -132990,7 +132990,7 @@ } }, { - "accuracy": 0.9753626145814595, + "accuracy": 0.9753532503780565, "total_bits": 475479040, "q_proj": { "group_size": { @@ -133054,7 +133054,7 @@ } }, { - "accuracy": 0.9862606556792008, + "accuracy": 0.9865289393224215, "total_bits": 609759232, "q_proj": { "group_size": { @@ -133106,7 +133106,7 @@ } }, { - "accuracy": 0.9875748416310862, + "accuracy": 0.9872832706100062, "total_bits": 610024448, "q_proj": { "group_size": { @@ -133158,7 +133158,7 @@ } }, { - "accuracy": 0.9884564892241829, + "accuracy": 0.9882933630755073, "total_bits": 615020544, "q_proj": { "group_size": { @@ -133210,7 +133210,7 @@ } }, { - "accuracy": 0.9886217383961928, + "accuracy": 0.9884843418472692, "total_bits": 623951872, "q_proj": { "group_size": { @@ -133262,7 +133262,7 @@ } }, { - "accuracy": 0.9882743531151822, + "accuracy": 0.9881194414276826, "total_bits": 626473984, "q_proj": { "group_size": { @@ -133326,7 +133326,7 @@ } }, { - "accuracy": 0.9887473606749585, + "accuracy": 0.989339250483011, "total_bits": 630355968, "q_proj": { "group_size": { @@ -133390,7 +133390,7 @@ } }, { - "accuracy": 0.9925556927919388, + "accuracy": 0.9925261328094884, "total_bits": 637362176, "q_proj": { "group_size": { @@ -133451,7 +133451,7 @@ } }, { - "accuracy": 0.9931462101246181, + "accuracy": 0.9933177750361594, "total_bits": 646823936, "q_proj": { "group_size": { @@ -133512,7 +133512,7 @@ } }, { - "accuracy": 0.995919789530729, + "accuracy": 0.9959385795028586, "total_bits": 784740352, "q_proj": { "group_size": { @@ -133573,7 +133573,7 @@ } }, { - "accuracy": 0.9966436259840664, + "accuracy": 0.9966347107761785, "total_bits": 797818880, "q_proj": { "group_size": { @@ -133634,7 +133634,7 @@ } }, { - "accuracy": 0.9964060026564097, + "accuracy": 0.99640761747172, "total_bits": 911749120, "q_proj": { "group_size": { @@ -133686,7 +133686,7 @@ } }, { - "accuracy": 0.9986179994517251, + "accuracy": 0.9986340065339678, "total_bits": 942718976, "q_proj": { "group_size": { @@ -133738,7 +133738,7 @@ } }, { - "accuracy": 0.9990228285130701, + "accuracy": 0.9990116577399405, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -133792,7 +133792,7 @@ ], "model.layers.68.mlp": [ { - "accuracy": 0.9397669716885215, + "accuracy": 0.940133345754523, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -133844,7 +133844,7 @@ } }, { - "accuracy": 0.9414037591532657, + "accuracy": 0.9417683074348852, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -133896,7 +133896,7 @@ } }, { - "accuracy": 0.9484167475449412, + "accuracy": 0.9487407395714208, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -133945,7 +133945,7 @@ } }, { - "accuracy": 0.950395405292511, + "accuracy": 0.9507107891534504, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -133994,7 +133994,7 @@ } }, { - "accuracy": 0.9695258046451368, + "accuracy": 0.9697231744465075, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -134046,7 +134046,7 @@ } }, { - "accuracy": 0.9718884731593885, + "accuracy": 0.9720653766079953, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -134098,7 +134098,7 @@ } }, { - "accuracy": 0.9748411900118777, + "accuracy": 0.9750009210486161, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -134147,7 +134147,7 @@ } }, { - "accuracy": 0.9843972027301788, + "accuracy": 0.9844968067972284, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -134190,7 +134190,7 @@ } }, { - "accuracy": 0.985634949646498, + "accuracy": 0.9857231817747417, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -134233,7 +134233,7 @@ } }, { - "accuracy": 0.9845208274690729, + "accuracy": 0.9846170356399134, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -134285,7 +134285,7 @@ } }, { - "accuracy": 0.9863450731101789, + "accuracy": 0.986429994043551, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -134337,7 +134337,7 @@ } }, { - "accuracy": 0.9921295587953768, + "accuracy": 0.9921802956806985, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -134389,7 +134389,7 @@ } }, { - "accuracy": 0.993193509547334, + "accuracy": 0.9932350492791125, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -134441,7 +134441,7 @@ } }, { - "accuracy": 0.9957867037308844, + "accuracy": 0.9958148175164273, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -134484,7 +134484,7 @@ } }, { - "accuracy": 0.9959590760500807, + "accuracy": 0.9959853722860939, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -134533,7 +134533,7 @@ } }, { - "accuracy": 0.9965116216948158, + "accuracy": 0.9965340546087215, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -134579,7 +134579,7 @@ } }, { - "accuracy": 0.9987977243175632, + "accuracy": 0.9988052900880575, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -134621,7 +134621,7 @@ ], "model.layers.69.self_attn": [ { - "accuracy": 0.9577348483236212, + "accuracy": 0.9578068664199427, "total_bits": 320757760, "q_proj": { "group_size": { @@ -134685,7 +134685,7 @@ } }, { - "accuracy": 0.9590638217173124, + "accuracy": 0.9623721806626571, "total_bits": 329080832, "q_proj": { "group_size": { @@ -134749,7 +134749,7 @@ } }, { - "accuracy": 0.975088706142024, + "accuracy": 0.9744345733993932, "total_bits": 336024576, "q_proj": { "group_size": { @@ -134813,7 +134813,7 @@ } }, { - "accuracy": 0.9777729260294061, + "accuracy": 0.977204314972225, "total_bits": 401557504, "q_proj": { "group_size": { @@ -134877,7 +134877,7 @@ } }, { - "accuracy": 0.9792740752822474, + "accuracy": 0.9786755301450428, "total_bits": 475279360, "q_proj": { "group_size": { @@ -134941,7 +134941,7 @@ } }, { - "accuracy": 0.9797721128714713, + "accuracy": 0.9795108610077908, "total_bits": 475479040, "q_proj": { "group_size": { @@ -135005,7 +135005,7 @@ } }, { - "accuracy": 0.9871726561533777, + "accuracy": 0.9870510422869733, "total_bits": 609759232, "q_proj": { "group_size": { @@ -135057,7 +135057,7 @@ } }, { - "accuracy": 0.9875928963485517, + "accuracy": 0.98801116096346, "total_bits": 610024448, "q_proj": { "group_size": { @@ -135109,7 +135109,7 @@ } }, { - "accuracy": 0.9889761052633587, + "accuracy": 0.9891380659843746, "total_bits": 615020544, "q_proj": { "group_size": { @@ -135161,7 +135161,7 @@ } }, { - "accuracy": 0.989124806303727, + "accuracy": 0.9893103367403934, "total_bits": 623951872, "q_proj": { "group_size": { @@ -135213,7 +135213,7 @@ } }, { - "accuracy": 0.9886743504750101, + "accuracy": 0.9889919985281793, "total_bits": 626473984, "q_proj": { "group_size": { @@ -135277,7 +135277,7 @@ } }, { - "accuracy": 0.9899733921414927, + "accuracy": 0.9900609019555544, "total_bits": 630355968, "q_proj": { "group_size": { @@ -135341,7 +135341,7 @@ } }, { - "accuracy": 0.9928061256283208, + "accuracy": 0.9929361562979849, "total_bits": 637362176, "q_proj": { "group_size": { @@ -135402,7 +135402,7 @@ } }, { - "accuracy": 0.993536182532185, + "accuracy": 0.9935311082946626, "total_bits": 646823936, "q_proj": { "group_size": { @@ -135463,7 +135463,7 @@ } }, { - "accuracy": 0.9958907624608592, + "accuracy": 0.9960133009835294, "total_bits": 784740352, "q_proj": { "group_size": { @@ -135524,7 +135524,7 @@ } }, { - "accuracy": 0.9966939299514419, + "accuracy": 0.9967410805586138, "total_bits": 797818880, "q_proj": { "group_size": { @@ -135585,7 +135585,7 @@ } }, { - "accuracy": 0.9962244955332655, + "accuracy": 0.9963545387512759, "total_bits": 911749120, "q_proj": { "group_size": { @@ -135637,7 +135637,7 @@ } }, { - "accuracy": 0.9987089622177576, + "accuracy": 0.9987127190749896, "total_bits": 942718976, "q_proj": { "group_size": { @@ -135689,7 +135689,7 @@ } }, { - "accuracy": 0.9990010306631264, + "accuracy": 0.9990083233110214, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -135743,7 +135743,7 @@ ], "model.layers.69.mlp": [ { - "accuracy": 0.9392343194861161, + "accuracy": 0.9396113220014071, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -135795,7 +135795,7 @@ } }, { - "accuracy": 0.9408864159333079, + "accuracy": 0.9412573011297929, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -135847,7 +135847,7 @@ } }, { - "accuracy": 0.947979145928433, + "accuracy": 0.9483077588834261, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -135896,7 +135896,7 @@ } }, { - "accuracy": 0.9499631116264745, + "accuracy": 0.9502791009451214, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -135945,7 +135945,7 @@ } }, { - "accuracy": 0.969258753876937, + "accuracy": 0.9694518603776631, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -135997,7 +135997,7 @@ } }, { - "accuracy": 0.9716319880987468, + "accuracy": 0.971817063657861, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -136049,7 +136049,7 @@ } }, { - "accuracy": 0.974617107918388, + "accuracy": 0.9747853545766128, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -136098,7 +136098,7 @@ } }, { - "accuracy": 0.9842761158943176, + "accuracy": 0.9843724103350389, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -136141,7 +136141,7 @@ } }, { - "accuracy": 0.985524888101377, + "accuracy": 0.985620086130343, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -136184,7 +136184,7 @@ } }, { - "accuracy": 0.9843835250327462, + "accuracy": 0.984482981656727, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -136236,7 +136236,7 @@ } }, { - "accuracy": 0.986223186317243, + "accuracy": 0.9863124709380301, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -136288,7 +136288,7 @@ } }, { - "accuracy": 0.9920629325665926, + "accuracy": 0.9921151101589203, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -136340,7 +136340,7 @@ } }, { - "accuracy": 0.9931394732312152, + "accuracy": 0.9931829399184177, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -136392,7 +136392,7 @@ } }, { - "accuracy": 0.9957637179054712, + "accuracy": 0.9957908066479784, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -136435,7 +136435,7 @@ } }, { - "accuracy": 0.9959463943776331, + "accuracy": 0.9959717887012582, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -136484,7 +136484,7 @@ } }, { - "accuracy": 0.9965069674347576, + "accuracy": 0.9965289191980111, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -136530,7 +136530,7 @@ } }, { - "accuracy": 0.9988704455134115, + "accuracy": 0.9988776304219898, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -136572,7 +136572,7 @@ ], "model.layers.70.self_attn": [ { - "accuracy": 0.9711955095592298, + "accuracy": 0.9713438121896041, "total_bits": 320757760, "q_proj": { "group_size": { @@ -136636,7 +136636,7 @@ } }, { - "accuracy": 0.9732321971341183, + "accuracy": 0.973576329256359, "total_bits": 329080832, "q_proj": { "group_size": { @@ -136700,7 +136700,7 @@ } }, { - "accuracy": 0.9760182876335947, + "accuracy": 0.9762139304688102, "total_bits": 336024576, "q_proj": { "group_size": { @@ -136764,7 +136764,7 @@ } }, { - "accuracy": 0.9797634200045937, + "accuracy": 0.9799719954791822, "total_bits": 401557504, "q_proj": { "group_size": { @@ -136828,7 +136828,7 @@ } }, { - "accuracy": 0.9831656280316805, + "accuracy": 0.9833419056315171, "total_bits": 475279360, "q_proj": { "group_size": { @@ -136892,7 +136892,7 @@ } }, { - "accuracy": 0.9836665203696803, + "accuracy": 0.9840130962823567, "total_bits": 475479040, "q_proj": { "group_size": { @@ -136956,7 +136956,7 @@ } }, { - "accuracy": 0.9893202162102649, + "accuracy": 0.989356498184957, "total_bits": 609759232, "q_proj": { "group_size": { @@ -137008,7 +137008,7 @@ } }, { - "accuracy": 0.9898630749238165, + "accuracy": 0.989799166196271, "total_bits": 610024448, "q_proj": { "group_size": { @@ -137060,7 +137060,7 @@ } }, { - "accuracy": 0.990433259229911, + "accuracy": 0.9906799675602662, "total_bits": 615020544, "q_proj": { "group_size": { @@ -137112,7 +137112,7 @@ } }, { - "accuracy": 0.990695794946269, + "accuracy": 0.9908786023917951, "total_bits": 623951872, "q_proj": { "group_size": { @@ -137164,7 +137164,7 @@ } }, { - "accuracy": 0.9904983326008445, + "accuracy": 0.9906234654941057, "total_bits": 626473984, "q_proj": { "group_size": { @@ -137228,7 +137228,7 @@ } }, { - "accuracy": 0.9910258093946859, + "accuracy": 0.991198405623436, "total_bits": 630355968, "q_proj": { "group_size": { @@ -137292,7 +137292,7 @@ } }, { - "accuracy": 0.993239363557414, + "accuracy": 0.993208782453286, "total_bits": 637362176, "q_proj": { "group_size": { @@ -137353,7 +137353,7 @@ } }, { - "accuracy": 0.9937599462113882, + "accuracy": 0.993845180853417, "total_bits": 646823936, "q_proj": { "group_size": { @@ -137414,7 +137414,7 @@ } }, { - "accuracy": 0.9962397311863146, + "accuracy": 0.9963346033504135, "total_bits": 784740352, "q_proj": { "group_size": { @@ -137475,7 +137475,7 @@ } }, { - "accuracy": 0.9967872529829803, + "accuracy": 0.9968793164742621, "total_bits": 797818880, "q_proj": { "group_size": { @@ -137536,7 +137536,7 @@ } }, { - "accuracy": 0.9968453482969811, + "accuracy": 0.996942656997003, "total_bits": 911749120, "q_proj": { "group_size": { @@ -137588,7 +137588,7 @@ } }, { - "accuracy": 0.998606634865466, + "accuracy": 0.9986275568800537, "total_bits": 942718976, "q_proj": { "group_size": { @@ -137640,7 +137640,7 @@ } }, { - "accuracy": 0.9991581978178338, + "accuracy": 0.9991588415087838, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -137694,7 +137694,7 @@ ], "model.layers.70.mlp": [ { - "accuracy": 0.9381677665208515, + "accuracy": 0.9385518588517842, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -137746,7 +137746,7 @@ } }, { - "accuracy": 0.9398394383882221, + "accuracy": 0.9402131971560026, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -137798,7 +137798,7 @@ } }, { - "accuracy": 0.947066815275895, + "accuracy": 0.9474110258252997, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -137847,7 +137847,7 @@ } }, { - "accuracy": 0.9490433398045992, + "accuracy": 0.9493765203576339, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -137896,7 +137896,7 @@ } }, { - "accuracy": 0.9686733233301263, + "accuracy": 0.968872186384703, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -137948,7 +137948,7 @@ } }, { - "accuracy": 0.971104875991219, + "accuracy": 0.9712900331145838, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -138000,7 +138000,7 @@ } }, { - "accuracy": 0.9741481919037668, + "accuracy": 0.9743145923865469, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -138049,7 +138049,7 @@ } }, { - "accuracy": 0.9839913954860285, + "accuracy": 0.9840980234899019, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -138092,7 +138092,7 @@ } }, { - "accuracy": 0.9852674038786637, + "accuracy": 0.9853638818389491, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -138135,7 +138135,7 @@ } }, { - "accuracy": 0.9840732781510604, + "accuracy": 0.984175798140074, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -138187,7 +138187,7 @@ } }, { - "accuracy": 0.985956039867903, + "accuracy": 0.9860457762291557, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -138239,7 +138239,7 @@ } }, { - "accuracy": 0.9918960058375409, + "accuracy": 0.9919529902307611, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -138291,7 +138291,7 @@ } }, { - "accuracy": 0.9929940967183364, + "accuracy": 0.9930398981822165, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -138343,7 +138343,7 @@ } }, { - "accuracy": 0.9956643942155337, + "accuracy": 0.9956934691259736, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -138386,7 +138386,7 @@ } }, { - "accuracy": 0.9958534774027372, + "accuracy": 0.9958811857198414, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -138435,7 +138435,7 @@ } }, { - "accuracy": 0.996421570448499, + "accuracy": 0.9964457926781554, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -138481,7 +138481,7 @@ } }, { - "accuracy": 0.9988104310867033, + "accuracy": 0.9988176377588197, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -138523,7 +138523,7 @@ ], "model.layers.71.self_attn": [ { - "accuracy": 0.9407630154961034, + "accuracy": 0.9439701281095806, "total_bits": 320757760, "q_proj": { "group_size": { @@ -138587,7 +138587,7 @@ } }, { - "accuracy": 0.9481912600366693, + "accuracy": 0.9479871956925643, "total_bits": 329080832, "q_proj": { "group_size": { @@ -138651,7 +138651,7 @@ } }, { - "accuracy": 0.9654816859646848, + "accuracy": 0.9656026833935788, "total_bits": 336024576, "q_proj": { "group_size": { @@ -138715,7 +138715,7 @@ } }, { - "accuracy": 0.9704953118374473, + "accuracy": 0.9704507746194538, "total_bits": 401557504, "q_proj": { "group_size": { @@ -138779,7 +138779,7 @@ } }, { - "accuracy": 0.9739467407527723, + "accuracy": 0.9738372156494542, "total_bits": 475279360, "q_proj": { "group_size": { @@ -138843,7 +138843,7 @@ } }, { - "accuracy": 0.9757474033456099, + "accuracy": 0.9745329665510278, "total_bits": 475479040, "q_proj": { "group_size": { @@ -138907,7 +138907,7 @@ } }, { - "accuracy": 0.9855372121459559, + "accuracy": 0.9853218103709974, "total_bits": 609759232, "q_proj": { "group_size": { @@ -138959,7 +138959,7 @@ } }, { - "accuracy": 0.9862602635433799, + "accuracy": 0.9861307206906771, "total_bits": 610024448, "q_proj": { "group_size": { @@ -139011,7 +139011,7 @@ } }, { - "accuracy": 0.9874742556559412, + "accuracy": 0.9872631501210364, "total_bits": 615020544, "q_proj": { "group_size": { @@ -139063,7 +139063,7 @@ } }, { - "accuracy": 0.9876996427774429, + "accuracy": 0.9875349002449136, "total_bits": 623951872, "q_proj": { "group_size": { @@ -139115,7 +139115,7 @@ } }, { - "accuracy": 0.9874448737031535, + "accuracy": 0.9873394542618802, "total_bits": 626473984, "q_proj": { "group_size": { @@ -139179,7 +139179,7 @@ } }, { - "accuracy": 0.9884187908549058, + "accuracy": 0.988410508162097, "total_bits": 630355968, "q_proj": { "group_size": { @@ -139243,7 +139243,7 @@ } }, { - "accuracy": 0.9914206862449646, + "accuracy": 0.9915678069779748, "total_bits": 637362176, "q_proj": { "group_size": { @@ -139304,7 +139304,7 @@ } }, { - "accuracy": 0.9922665467387751, + "accuracy": 0.9924182468339017, "total_bits": 646823936, "q_proj": { "group_size": { @@ -139365,7 +139365,7 @@ } }, { - "accuracy": 0.9953128519026857, + "accuracy": 0.9954021753449189, "total_bits": 784740352, "q_proj": { "group_size": { @@ -139426,7 +139426,7 @@ } }, { - "accuracy": 0.9961062959934536, + "accuracy": 0.9961357395115652, "total_bits": 797818880, "q_proj": { "group_size": { @@ -139487,7 +139487,7 @@ } }, { - "accuracy": 0.9960838187682001, + "accuracy": 0.9961911518322794, "total_bits": 911749120, "q_proj": { "group_size": { @@ -139539,7 +139539,7 @@ } }, { - "accuracy": 0.9982225747877046, + "accuracy": 0.9982506648490304, "total_bits": 942718976, "q_proj": { "group_size": { @@ -139591,7 +139591,7 @@ } }, { - "accuracy": 0.9990023861786252, + "accuracy": 0.9989911058035336, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -139645,7 +139645,7 @@ ], "model.layers.71.mlp": [ { - "accuracy": 0.935518176932084, + "accuracy": 0.9359502227682817, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -139697,7 +139697,7 @@ } }, { - "accuracy": 0.9372616818076686, + "accuracy": 0.9376871774071142, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -139749,7 +139749,7 @@ } }, { - "accuracy": 0.9447692444449977, + "accuracy": 0.9451361957349276, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -139798,7 +139798,7 @@ } }, { - "accuracy": 0.9468450797231573, + "accuracy": 0.947200335954365, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -139847,7 +139847,7 @@ } }, { - "accuracy": 0.9673541627432171, + "accuracy": 0.9675758575138292, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -139899,7 +139899,7 @@ } }, { - "accuracy": 0.9698815000684637, + "accuracy": 0.970090577476903, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -139951,7 +139951,7 @@ } }, { - "accuracy": 0.9730451546217266, + "accuracy": 0.9732323728109661, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -140000,7 +140000,7 @@ } }, { - "accuracy": 0.9832980444556788, + "accuracy": 0.9834123096967998, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -140043,7 +140043,7 @@ } }, { - "accuracy": 0.9846315305483969, + "accuracy": 0.9847364268804851, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -140086,7 +140086,7 @@ } }, { - "accuracy": 0.9834011981361791, + "accuracy": 0.9835116392687747, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -140138,7 +140138,7 @@ } }, { - "accuracy": 0.9853638081174148, + "accuracy": 0.985458116782339, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -140190,7 +140190,7 @@ } }, { - "accuracy": 0.9915555692032764, + "accuracy": 0.9916136437340787, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -140242,7 +140242,7 @@ } }, { - "accuracy": 0.9926997619239908, + "accuracy": 0.9927484040197573, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -140294,7 +140294,7 @@ } }, { - "accuracy": 0.9954867551201269, + "accuracy": 0.9955156469031384, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -140337,7 +140337,7 @@ } }, { - "accuracy": 0.995680296891614, + "accuracy": 0.9957098168762106, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -140386,7 +140386,7 @@ } }, { - "accuracy": 0.9962723462989456, + "accuracy": 0.9962984468591841, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -140432,7 +140432,7 @@ } }, { - "accuracy": 0.998776415460988, + "accuracy": 0.9987837120302414, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -140474,7 +140474,7 @@ ], "model.layers.72.self_attn": [ { - "accuracy": 0.9457032555028012, + "accuracy": 0.9485957246077688, "total_bits": 320757760, "q_proj": { "group_size": { @@ -140538,7 +140538,7 @@ } }, { - "accuracy": 0.9473653843528346, + "accuracy": 0.9467097458086515, "total_bits": 329080832, "q_proj": { "group_size": { @@ -140602,7 +140602,7 @@ } }, { - "accuracy": 0.9650907045916507, + "accuracy": 0.9661189945120561, "total_bits": 336024576, "q_proj": { "group_size": { @@ -140666,7 +140666,7 @@ } }, { - "accuracy": 0.9697697413595099, + "accuracy": 0.9710033661440799, "total_bits": 401557504, "q_proj": { "group_size": { @@ -140730,7 +140730,7 @@ } }, { - "accuracy": 0.9747143387794495, + "accuracy": 0.9754261092135781, "total_bits": 475279360, "q_proj": { "group_size": { @@ -140794,7 +140794,7 @@ } }, { - "accuracy": 0.9754952051137623, + "accuracy": 0.9760242935858274, "total_bits": 475479040, "q_proj": { "group_size": { @@ -140858,7 +140858,7 @@ } }, { - "accuracy": 0.98638432590585, + "accuracy": 0.9861753708437869, "total_bits": 609759232, "q_proj": { "group_size": { @@ -140910,7 +140910,7 @@ } }, { - "accuracy": 0.9868126260606866, + "accuracy": 0.986395454720447, "total_bits": 610024448, "q_proj": { "group_size": { @@ -140962,7 +140962,7 @@ } }, { - "accuracy": 0.9880781636426323, + "accuracy": 0.9879814347154215, "total_bits": 615020544, "q_proj": { "group_size": { @@ -141014,7 +141014,7 @@ } }, { - "accuracy": 0.9882987816082803, + "accuracy": 0.9881844395085385, "total_bits": 623951872, "q_proj": { "group_size": { @@ -141066,7 +141066,7 @@ } }, { - "accuracy": 0.9878992038337808, + "accuracy": 0.988135081372763, "total_bits": 626473984, "q_proj": { "group_size": { @@ -141130,7 +141130,7 @@ } }, { - "accuracy": 0.9888089267831099, + "accuracy": 0.9889352752974159, "total_bits": 630355968, "q_proj": { "group_size": { @@ -141194,7 +141194,7 @@ } }, { - "accuracy": 0.9915042503883964, + "accuracy": 0.9918131396958703, "total_bits": 637362176, "q_proj": { "group_size": { @@ -141255,7 +141255,7 @@ } }, { - "accuracy": 0.9924118973706898, + "accuracy": 0.9923152170683208, "total_bits": 646823936, "q_proj": { "group_size": { @@ -141316,7 +141316,7 @@ } }, { - "accuracy": 0.9954736687635121, + "accuracy": 0.9955231296388727, "total_bits": 784740352, "q_proj": { "group_size": { @@ -141377,7 +141377,7 @@ } }, { - "accuracy": 0.9961340274465712, + "accuracy": 0.9962273859664014, "total_bits": 797818880, "q_proj": { "group_size": { @@ -141438,7 +141438,7 @@ } }, { - "accuracy": 0.9963498746878222, + "accuracy": 0.9963740361364264, "total_bits": 911749120, "q_proj": { "group_size": { @@ -141490,7 +141490,7 @@ } }, { - "accuracy": 0.9982680737187988, + "accuracy": 0.998292795529491, "total_bits": 942718976, "q_proj": { "group_size": { @@ -141542,7 +141542,7 @@ } }, { - "accuracy": 0.999003409260982, + "accuracy": 0.9990036376800976, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -141596,7 +141596,7 @@ ], "model.layers.72.mlp": [ { - "accuracy": 0.9337879419326782, + "accuracy": 0.934178214324148, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -141648,7 +141648,7 @@ } }, { - "accuracy": 0.935604779343856, + "accuracy": 0.9359802321383828, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -141700,7 +141700,7 @@ } }, { - "accuracy": 0.9432544708251953, + "accuracy": 0.9435891477685225, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -141749,7 +141749,7 @@ } }, { - "accuracy": 0.9454545598281057, + "accuracy": 0.9457825045836599, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -141798,7 +141798,7 @@ } }, { - "accuracy": 0.9665264800975197, + "accuracy": 0.9667216884462457, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -141850,7 +141850,7 @@ } }, { - "accuracy": 0.9691304501734281, + "accuracy": 0.9693053082415932, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -141902,7 +141902,7 @@ } }, { - "accuracy": 0.9723373180941531, + "accuracy": 0.9724994991955004, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -141951,7 +141951,7 @@ } }, { - "accuracy": 0.9828026859383834, + "accuracy": 0.9829002085484957, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -141994,7 +141994,7 @@ } }, { - "accuracy": 0.984175045239298, + "accuracy": 0.9842645965124431, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -142037,7 +142037,7 @@ } }, { - "accuracy": 0.9829912938569721, + "accuracy": 0.9830900995354903, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -142089,7 +142089,7 @@ } }, { - "accuracy": 0.9850000143051147, + "accuracy": 0.985087270799436, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -142141,7 +142141,7 @@ } }, { - "accuracy": 0.991352946350449, + "accuracy": 0.991402804851532, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -142193,7 +142193,7 @@ } }, { - "accuracy": 0.9925220216575422, + "accuracy": 0.9925636751087088, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -142245,7 +142245,7 @@ } }, { - "accuracy": 0.9953518435359001, + "accuracy": 0.9953773802048281, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -142288,7 +142288,7 @@ } }, { - "accuracy": 0.9955703373018064, + "accuracy": 0.9955951336183047, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -142337,7 +142337,7 @@ } }, { - "accuracy": 0.996174126079208, + "accuracy": 0.9961956245334525, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -142383,7 +142383,7 @@ } }, { - "accuracy": 0.9986938511658656, + "accuracy": 0.9987016437869323, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -142425,7 +142425,7 @@ ], "model.layers.73.self_attn": [ { - "accuracy": 0.9641388466483668, + "accuracy": 0.9648380436395344, "total_bits": 320757760, "q_proj": { "group_size": { @@ -142489,7 +142489,7 @@ } }, { - "accuracy": 0.9655978303206594, + "accuracy": 0.9657484606692666, "total_bits": 329080832, "q_proj": { "group_size": { @@ -142553,7 +142553,7 @@ } }, { - "accuracy": 0.9693322652264645, + "accuracy": 0.969718525284215, "total_bits": 336024576, "q_proj": { "group_size": { @@ -142617,7 +142617,7 @@ } }, { - "accuracy": 0.9746346003130862, + "accuracy": 0.975013770555195, "total_bits": 401557504, "q_proj": { "group_size": { @@ -142681,7 +142681,7 @@ } }, { - "accuracy": 0.9802468136737221, + "accuracy": 0.9805005851544832, "total_bits": 475279360, "q_proj": { "group_size": { @@ -142745,7 +142745,7 @@ } }, { - "accuracy": 0.9804162194854334, + "accuracy": 0.9808616246047773, "total_bits": 475479040, "q_proj": { "group_size": { @@ -142809,7 +142809,7 @@ } }, { - "accuracy": 0.9868680426948949, + "accuracy": 0.987425306125691, "total_bits": 609759232, "q_proj": { "group_size": { @@ -142861,7 +142861,7 @@ } }, { - "accuracy": 0.987165378112542, + "accuracy": 0.9875823918141817, "total_bits": 610024448, "q_proj": { "group_size": { @@ -142913,7 +142913,7 @@ } }, { - "accuracy": 0.9885438578693491, + "accuracy": 0.9886332844433031, "total_bits": 615020544, "q_proj": { "group_size": { @@ -142965,7 +142965,7 @@ } }, { - "accuracy": 0.9888820256057539, + "accuracy": 0.9889473813144785, "total_bits": 623951872, "q_proj": { "group_size": { @@ -143017,7 +143017,7 @@ } }, { - "accuracy": 0.9890319077592147, + "accuracy": 0.9887825266311043, "total_bits": 626473984, "q_proj": { "group_size": { @@ -143081,7 +143081,7 @@ } }, { - "accuracy": 0.9894924822606539, + "accuracy": 0.9898217822376051, "total_bits": 630355968, "q_proj": { "group_size": { @@ -143145,7 +143145,7 @@ } }, { - "accuracy": 0.9915448019379064, + "accuracy": 0.9917343925488623, "total_bits": 637362176, "q_proj": { "group_size": { @@ -143206,7 +143206,7 @@ } }, { - "accuracy": 0.9922681623383572, + "accuracy": 0.9922923916264584, "total_bits": 646823936, "q_proj": { "group_size": { @@ -143267,7 +143267,7 @@ } }, { - "accuracy": 0.9952292191354852, + "accuracy": 0.995334271929766, "total_bits": 784740352, "q_proj": { "group_size": { @@ -143328,7 +143328,7 @@ } }, { - "accuracy": 0.9960674306279734, + "accuracy": 0.9960319933138395, "total_bits": 797818880, "q_proj": { "group_size": { @@ -143389,7 +143389,7 @@ } }, { - "accuracy": 0.9961954465037898, + "accuracy": 0.9963119155482242, "total_bits": 911749120, "q_proj": { "group_size": { @@ -143441,7 +143441,7 @@ } }, { - "accuracy": 0.9981912156861079, + "accuracy": 0.9981783973543268, "total_bits": 942718976, "q_proj": { "group_size": { @@ -143493,7 +143493,7 @@ } }, { - "accuracy": 0.9989889446449908, + "accuracy": 0.999003971779817, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -143547,7 +143547,7 @@ ], "model.layers.73.mlp": [ { - "accuracy": 0.9331154133144178, + "accuracy": 0.9335026301835713, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -143599,7 +143599,7 @@ } }, { - "accuracy": 0.9350030422210693, + "accuracy": 0.9354074816954763, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -143651,7 +143651,7 @@ } }, { - "accuracy": 0.9427665660255834, + "accuracy": 0.9431361399198833, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -143700,7 +143700,7 @@ } }, { - "accuracy": 0.9449635304902729, + "accuracy": 0.9453236178347939, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -143749,7 +143749,7 @@ } }, { - "accuracy": 0.9660358083875555, + "accuracy": 0.9662442615157679, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -143801,7 +143801,7 @@ } }, { - "accuracy": 0.9687754323607997, + "accuracy": 0.9689593064157587, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -143853,7 +143853,7 @@ } }, { - "accuracy": 0.9719971073301215, + "accuracy": 0.9721629086293673, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -143902,7 +143902,7 @@ } }, { - "accuracy": 0.9824772979083815, + "accuracy": 0.9825934422643561, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -143945,7 +143945,7 @@ } }, { - "accuracy": 0.9838315750423231, + "accuracy": 0.9839427894667575, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -143988,7 +143988,7 @@ } }, { - "accuracy": 0.9825791434237832, + "accuracy": 0.9826967794644205, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -144040,7 +144040,7 @@ } }, { - "accuracy": 0.9846870961942171, + "accuracy": 0.9847893510994158, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -144092,7 +144092,7 @@ } }, { - "accuracy": 0.9910650747387033, + "accuracy": 0.9911187471527803, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -144144,7 +144144,7 @@ } }, { - "accuracy": 0.9923356300906131, + "accuracy": 0.9923868477344513, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -144196,7 +144196,7 @@ } }, { - "accuracy": 0.9951563034402696, + "accuracy": 0.995194077883896, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -144239,7 +144239,7 @@ } }, { - "accuracy": 0.9953978242058503, + "accuracy": 0.9954321729509454, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -144288,7 +144288,7 @@ } }, { - "accuracy": 0.9959851722968253, + "accuracy": 0.9960172368507636, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -144334,7 +144334,7 @@ } }, { - "accuracy": 0.9986519849810161, + "accuracy": 0.9986623157992175, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -144376,7 +144376,7 @@ ], "model.layers.74.self_attn": [ { - "accuracy": 0.9355527601743999, + "accuracy": 0.9350374937057495, "total_bits": 320757760, "q_proj": { "group_size": { @@ -144440,7 +144440,7 @@ } }, { - "accuracy": 0.9415584237951982, + "accuracy": 0.9373297879570409, "total_bits": 329080832, "q_proj": { "group_size": { @@ -144504,7 +144504,7 @@ } }, { - "accuracy": 0.9594148083737022, + "accuracy": 0.9605710161359686, "total_bits": 336024576, "q_proj": { "group_size": { @@ -144568,7 +144568,7 @@ } }, { - "accuracy": 0.964950326241945, + "accuracy": 0.9660376906394958, "total_bits": 401557504, "q_proj": { "group_size": { @@ -144632,7 +144632,7 @@ } }, { - "accuracy": 0.9695519303020678, + "accuracy": 0.9700005713262057, "total_bits": 475279360, "q_proj": { "group_size": { @@ -144696,7 +144696,7 @@ } }, { - "accuracy": 0.9714412438242059, + "accuracy": 0.9711148519265024, "total_bits": 475479040, "q_proj": { "group_size": { @@ -144760,7 +144760,7 @@ } }, { - "accuracy": 0.9832599649303838, + "accuracy": 0.9828267066102279, "total_bits": 609759232, "q_proj": { "group_size": { @@ -144812,7 +144812,7 @@ } }, { - "accuracy": 0.984428628494865, + "accuracy": 0.9840543991640994, "total_bits": 610024448, "q_proj": { "group_size": { @@ -144864,7 +144864,7 @@ } }, { - "accuracy": 0.9856422245502472, + "accuracy": 0.9859218330759751, "total_bits": 615020544, "q_proj": { "group_size": { @@ -144916,7 +144916,7 @@ } }, { - "accuracy": 0.9859842987436997, + "accuracy": 0.9862542497484308, "total_bits": 623951872, "q_proj": { "group_size": { @@ -144968,7 +144968,7 @@ } }, { - "accuracy": 0.985555953101108, + "accuracy": 0.9846131911403254, "total_bits": 626473984, "q_proj": { "group_size": { @@ -145032,7 +145032,7 @@ } }, { - "accuracy": 0.9862603043255053, + "accuracy": 0.9867275545471593, "total_bits": 630355968, "q_proj": { "group_size": { @@ -145096,7 +145096,7 @@ } }, { - "accuracy": 0.9904399043635318, + "accuracy": 0.9902359843254089, "total_bits": 637362176, "q_proj": { "group_size": { @@ -145157,7 +145157,7 @@ } }, { - "accuracy": 0.9910461612437901, + "accuracy": 0.9912664749120411, "total_bits": 646823936, "q_proj": { "group_size": { @@ -145218,7 +145218,7 @@ } }, { - "accuracy": 0.9947020682065111, + "accuracy": 0.9947653483403357, "total_bits": 784740352, "q_proj": { "group_size": { @@ -145279,7 +145279,7 @@ } }, { - "accuracy": 0.9955620550011334, + "accuracy": 0.9956081819377447, "total_bits": 797818880, "q_proj": { "group_size": { @@ -145340,7 +145340,7 @@ } }, { - "accuracy": 0.9957687548900905, + "accuracy": 0.9957344979047775, "total_bits": 911749120, "q_proj": { "group_size": { @@ -145392,7 +145392,7 @@ } }, { - "accuracy": 0.9979437344560498, + "accuracy": 0.9979277700185776, "total_bits": 942718976, "q_proj": { "group_size": { @@ -145444,7 +145444,7 @@ } }, { - "accuracy": 0.9988271120543543, + "accuracy": 0.9987860436698324, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -145498,7 +145498,7 @@ ], "model.layers.74.mlp": [ { - "accuracy": 0.9301935371599699, + "accuracy": 0.9305787337453741, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -145550,7 +145550,7 @@ } }, { - "accuracy": 0.9321970500444111, + "accuracy": 0.9324950607199418, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -145602,7 +145602,7 @@ } }, { - "accuracy": 0.940350381951583, + "accuracy": 0.9406311888443797, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -145651,7 +145651,7 @@ } }, { - "accuracy": 0.942656253513537, + "accuracy": 0.9429237403367695, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -145700,7 +145700,7 @@ } }, { - "accuracy": 0.9647265358975059, + "accuracy": 0.9649315194079751, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -145752,7 +145752,7 @@ } }, { - "accuracy": 0.9674996294473347, + "accuracy": 0.967693219059392, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -145804,7 +145804,7 @@ } }, { - "accuracy": 0.9708791281047621, + "accuracy": 0.9710519658891779, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -145853,7 +145853,7 @@ } }, { - "accuracy": 0.9818425131471533, + "accuracy": 0.9819557760891161, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -145896,7 +145896,7 @@ } }, { - "accuracy": 0.9832795074111537, + "accuracy": 0.9833854907437375, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -145939,7 +145939,7 @@ } }, { - "accuracy": 0.9819929019400948, + "accuracy": 0.9820959944474069, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -145991,7 +145991,7 @@ } }, { - "accuracy": 0.9841553600210893, + "accuracy": 0.984253255944503, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -146043,7 +146043,7 @@ } }, { - "accuracy": 0.9908085261520586, + "accuracy": 0.990857678808664, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -146095,7 +146095,7 @@ } }, { - "accuracy": 0.9920865822779504, + "accuracy": 0.9921333774140007, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -146147,7 +146147,7 @@ } }, { - "accuracy": 0.9950419091864636, + "accuracy": 0.9950704747124722, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -146190,7 +146190,7 @@ } }, { - "accuracy": 0.9952828354741398, + "accuracy": 0.9953109037719274, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -146239,7 +146239,7 @@ } }, { - "accuracy": 0.9959118283892933, + "accuracy": 0.9959361094393229, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -146285,7 +146285,7 @@ } }, { - "accuracy": 0.9986207640092624, + "accuracy": 0.9986285747666108, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -146327,7 +146327,7 @@ ], "model.layers.75.self_attn": [ { - "accuracy": 0.9415766063489412, + "accuracy": 0.9419503839392411, "total_bits": 320757760, "q_proj": { "group_size": { @@ -146391,7 +146391,7 @@ } }, { - "accuracy": 0.9456013252860621, + "accuracy": 0.9451723977139121, "total_bits": 329080832, "q_proj": { "group_size": { @@ -146455,7 +146455,7 @@ } }, { - "accuracy": 0.955046744723069, + "accuracy": 0.9547628917192158, "total_bits": 336024576, "q_proj": { "group_size": { @@ -146519,7 +146519,7 @@ } }, { - "accuracy": 0.9637423973334462, + "accuracy": 0.9634079180265728, "total_bits": 401557504, "q_proj": { "group_size": { @@ -146583,7 +146583,7 @@ } }, { - "accuracy": 0.9701066330859536, + "accuracy": 0.9700148042879606, "total_bits": 475279360, "q_proj": { "group_size": { @@ -146647,7 +146647,7 @@ } }, { - "accuracy": 0.9708288939375627, + "accuracy": 0.971065734562121, "total_bits": 475479040, "q_proj": { "group_size": { @@ -146711,7 +146711,7 @@ } }, { - "accuracy": 0.9827741133539301, + "accuracy": 0.9823036021307895, "total_bits": 609759232, "q_proj": { "group_size": { @@ -146763,7 +146763,7 @@ } }, { - "accuracy": 0.9833453172131589, + "accuracy": 0.9833877604258688, "total_bits": 610024448, "q_proj": { "group_size": { @@ -146815,7 +146815,7 @@ } }, { - "accuracy": 0.9846391505316684, + "accuracy": 0.9846199060741224, "total_bits": 615020544, "q_proj": { "group_size": { @@ -146867,7 +146867,7 @@ } }, { - "accuracy": 0.9851064195758418, + "accuracy": 0.9850691415761647, "total_bits": 623951872, "q_proj": { "group_size": { @@ -146919,7 +146919,7 @@ } }, { - "accuracy": 0.9845221921017295, + "accuracy": 0.9846572107390353, "total_bits": 626473984, "q_proj": { "group_size": { @@ -146983,7 +146983,7 @@ } }, { - "accuracy": 0.9855536928302363, + "accuracy": 0.9857006574931898, "total_bits": 630355968, "q_proj": { "group_size": { @@ -147047,7 +147047,7 @@ } }, { - "accuracy": 0.9885978604617872, + "accuracy": 0.988655664418873, "total_bits": 637362176, "q_proj": { "group_size": { @@ -147108,7 +147108,7 @@ } }, { - "accuracy": 0.9895637294179515, + "accuracy": 0.9894590103312543, "total_bits": 646823936, "q_proj": { "group_size": { @@ -147169,7 +147169,7 @@ } }, { - "accuracy": 0.9937499690996973, + "accuracy": 0.993778330715079, "total_bits": 784740352, "q_proj": { "group_size": { @@ -147230,7 +147230,7 @@ } }, { - "accuracy": 0.9947519204334209, + "accuracy": 0.9947749540994042, "total_bits": 797818880, "q_proj": { "group_size": { @@ -147291,7 +147291,7 @@ } }, { - "accuracy": 0.9953722142075238, + "accuracy": 0.9954042007264338, "total_bits": 911749120, "q_proj": { "group_size": { @@ -147343,7 +147343,7 @@ } }, { - "accuracy": 0.9974178589488331, + "accuracy": 0.9974468624905536, "total_bits": 942718976, "q_proj": { "group_size": { @@ -147395,7 +147395,7 @@ } }, { - "accuracy": 0.9987789710101328, + "accuracy": 0.998786607169007, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -147449,7 +147449,7 @@ ], "model.layers.75.mlp": [ { - "accuracy": 0.9281256700816908, + "accuracy": 0.9285151895723844, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -147501,7 +147501,7 @@ } }, { - "accuracy": 0.9301207191065738, + "accuracy": 0.9304846525192261, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -147553,7 +147553,7 @@ } }, { - "accuracy": 0.9385271009645964, + "accuracy": 0.9388513941513865, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -147602,7 +147602,7 @@ } }, { - "accuracy": 0.9410342668232164, + "accuracy": 0.9413422597082037, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -147651,7 +147651,7 @@ } }, { - "accuracy": 0.9635903301991915, + "accuracy": 0.9637917719389263, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -147703,7 +147703,7 @@ } }, { - "accuracy": 0.96645524627284, + "accuracy": 0.9666336900309512, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -147755,7 +147755,7 @@ } }, { - "accuracy": 0.9700135494533338, + "accuracy": 0.9701740145683289, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -147804,7 +147804,7 @@ } }, { - "accuracy": 0.9811526003636812, + "accuracy": 0.9812607027982411, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -147847,7 +147847,7 @@ } }, { - "accuracy": 0.9826849824503848, + "accuracy": 0.9827814619792136, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -147890,7 +147890,7 @@ } }, { - "accuracy": 0.9814143525926691, + "accuracy": 0.9815204849368647, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -147942,7 +147942,7 @@ } }, { - "accuracy": 0.9836429184988925, + "accuracy": 0.9837301075458527, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -147994,7 +147994,7 @@ } }, { - "accuracy": 0.9905113326875787, + "accuracy": 0.9905663540488795, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -148046,7 +148046,7 @@ } }, { - "accuracy": 0.9918279561557268, + "accuracy": 0.9918694464783919, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -148098,7 +148098,7 @@ } }, { - "accuracy": 0.9948589993934882, + "accuracy": 0.9948895024625879, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -148141,7 +148141,7 @@ } }, { - "accuracy": 0.9951201547917566, + "accuracy": 0.9951484771935564, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -148190,7 +148190,7 @@ } }, { - "accuracy": 0.9957844595375814, + "accuracy": 0.9958111463408721, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -148236,7 +148236,7 @@ } }, { - "accuracy": 0.998512778725279, + "accuracy": 0.9985225276139221, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -148278,7 +148278,7 @@ ], "model.layers.76.self_attn": [ { - "accuracy": 0.9465230515128688, + "accuracy": 0.9466392868443539, "total_bits": 320757760, "q_proj": { "group_size": { @@ -148342,7 +148342,7 @@ } }, { - "accuracy": 0.947117617255763, + "accuracy": 0.9478922235338312, "total_bits": 329080832, "q_proj": { "group_size": { @@ -148406,7 +148406,7 @@ } }, { - "accuracy": 0.9629774501449183, + "accuracy": 0.9592910534457156, "total_bits": 336024576, "q_proj": { "group_size": { @@ -148470,7 +148470,7 @@ } }, { - "accuracy": 0.9676497735475239, + "accuracy": 0.9635316604062131, "total_bits": 401557504, "q_proj": { "group_size": { @@ -148534,7 +148534,7 @@ } }, { - "accuracy": 0.9739012718200684, + "accuracy": 0.9698867358659443, "total_bits": 475279360, "q_proj": { "group_size": { @@ -148598,7 +148598,7 @@ } }, { - "accuracy": 0.974125800948394, + "accuracy": 0.9749371550585094, "total_bits": 475479040, "q_proj": { "group_size": { @@ -148662,7 +148662,7 @@ } }, { - "accuracy": 0.9847747056107772, + "accuracy": 0.9848634459470448, "total_bits": 609759232, "q_proj": { "group_size": { @@ -148714,7 +148714,7 @@ } }, { - "accuracy": 0.9849110135906621, + "accuracy": 0.9858982751243993, "total_bits": 610024448, "q_proj": { "group_size": { @@ -148766,7 +148766,7 @@ } }, { - "accuracy": 0.9867063619588551, + "accuracy": 0.9870155767390603, "total_bits": 615020544, "q_proj": { "group_size": { @@ -148818,7 +148818,7 @@ } }, { - "accuracy": 0.9871315799261394, + "accuracy": 0.9872767219417974, "total_bits": 623951872, "q_proj": { "group_size": { @@ -148870,7 +148870,7 @@ } }, { - "accuracy": 0.9862926649419885, + "accuracy": 0.986915282512966, "total_bits": 626473984, "q_proj": { "group_size": { @@ -148934,7 +148934,7 @@ } }, { - "accuracy": 0.9878160514329609, + "accuracy": 0.9878154365639937, "total_bits": 630355968, "q_proj": { "group_size": { @@ -148998,7 +148998,7 @@ } }, { - "accuracy": 0.990951869832842, + "accuracy": 0.9911975688055942, "total_bits": 637362176, "q_proj": { "group_size": { @@ -149059,7 +149059,7 @@ } }, { - "accuracy": 0.9916702681466153, + "accuracy": 0.9916812840260958, "total_bits": 646823936, "q_proj": { "group_size": { @@ -149120,7 +149120,7 @@ } }, { - "accuracy": 0.9950284569671279, + "accuracy": 0.9949615574196765, "total_bits": 784740352, "q_proj": { "group_size": { @@ -149181,7 +149181,7 @@ } }, { - "accuracy": 0.9958347482116598, + "accuracy": 0.9958723712908594, "total_bits": 797818880, "q_proj": { "group_size": { @@ -149242,7 +149242,7 @@ } }, { - "accuracy": 0.9959516466448182, + "accuracy": 0.9959185739096842, "total_bits": 911749120, "q_proj": { "group_size": { @@ -149294,7 +149294,7 @@ } }, { - "accuracy": 0.9980640064336752, + "accuracy": 0.998046513646841, "total_bits": 942718976, "q_proj": { "group_size": { @@ -149346,7 +149346,7 @@ } }, { - "accuracy": 0.9989599681606418, + "accuracy": 0.9989433320925424, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -149400,7 +149400,7 @@ ], "model.layers.76.mlp": [ { - "accuracy": 0.9217840997796309, + "accuracy": 0.9224104504836232, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -149452,7 +149452,7 @@ } }, { - "accuracy": 0.923889768751044, + "accuracy": 0.9245201725708811, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -149504,7 +149504,7 @@ } }, { - "accuracy": 0.9328631539093821, + "accuracy": 0.933307221061305, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -149553,7 +149553,7 @@ } }, { - "accuracy": 0.9358106286902177, + "accuracy": 0.9362416518361945, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -149602,7 +149602,7 @@ } }, { - "accuracy": 0.9601799500615973, + "accuracy": 0.9602458947583249, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -149654,7 +149654,7 @@ } }, { - "accuracy": 0.963267561636473, + "accuracy": 0.9635057104261298, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -149706,7 +149706,7 @@ } }, { - "accuracy": 0.9671702730028253, + "accuracy": 0.9674004033992165, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -149755,7 +149755,7 @@ } }, { - "accuracy": 0.9789612105018214, + "accuracy": 0.9790288743219877, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -149798,7 +149798,7 @@ } }, { - "accuracy": 0.9807390178504743, + "accuracy": 0.9808048198097631, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -149841,7 +149841,7 @@ } }, { - "accuracy": 0.9795043390048178, + "accuracy": 0.9795721392882498, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -149893,7 +149893,7 @@ } }, { - "accuracy": 0.9819698333740234, + "accuracy": 0.9820494604738135, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -149945,7 +149945,7 @@ } }, { - "accuracy": 0.9894600847834035, + "accuracy": 0.9894944288228688, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -149997,7 +149997,7 @@ } }, { - "accuracy": 0.9909542148050509, + "accuracy": 0.9909933758409399, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -150049,7 +150049,7 @@ } }, { - "accuracy": 0.9941997798649889, + "accuracy": 0.9942325012464273, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -150092,7 +150092,7 @@ } }, { - "accuracy": 0.994594511232878, + "accuracy": 0.9946107629098391, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -150141,7 +150141,7 @@ } }, { - "accuracy": 0.9953547559286419, + "accuracy": 0.9953672960400581, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -150187,7 +150187,7 @@ } }, { - "accuracy": 0.9984304063806408, + "accuracy": 0.9984301671777901, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -150229,7 +150229,7 @@ ], "model.layers.77.self_attn": [ { - "accuracy": 0.9404934205506977, + "accuracy": 0.9433507417377672, "total_bits": 320757760, "q_proj": { "group_size": { @@ -150293,7 +150293,7 @@ } }, { - "accuracy": 0.9429259174748471, + "accuracy": 0.9443995764381007, "total_bits": 329080832, "q_proj": { "group_size": { @@ -150357,7 +150357,7 @@ } }, { - "accuracy": 0.9562844163493106, + "accuracy": 0.9545540119472303, "total_bits": 336024576, "q_proj": { "group_size": { @@ -150421,7 +150421,7 @@ } }, { - "accuracy": 0.9627126612161335, + "accuracy": 0.9604054626665617, "total_bits": 401557504, "q_proj": { "group_size": { @@ -150485,7 +150485,7 @@ } }, { - "accuracy": 0.9707808525938737, + "accuracy": 0.96844251218595, "total_bits": 475279360, "q_proj": { "group_size": { @@ -150549,7 +150549,7 @@ } }, { - "accuracy": 0.9691147365068135, + "accuracy": 0.9704476532183195, "total_bits": 475479040, "q_proj": { "group_size": { @@ -150613,7 +150613,7 @@ } }, { - "accuracy": 0.9844424803006021, + "accuracy": 0.9836739725188205, "total_bits": 609759232, "q_proj": { "group_size": { @@ -150665,7 +150665,7 @@ } }, { - "accuracy": 0.9850527597101111, + "accuracy": 0.9849120394179696, "total_bits": 610024448, "q_proj": { "group_size": { @@ -150717,7 +150717,7 @@ } }, { - "accuracy": 0.9862690442486813, + "accuracy": 0.986130394433674, "total_bits": 615020544, "q_proj": { "group_size": { @@ -150769,7 +150769,7 @@ } }, { - "accuracy": 0.9867599496715947, + "accuracy": 0.9864964563595621, "total_bits": 623951872, "q_proj": { "group_size": { @@ -150821,7 +150821,7 @@ } }, { - "accuracy": 0.9855489526924334, + "accuracy": 0.9850449844410545, "total_bits": 626473984, "q_proj": { "group_size": { @@ -150885,7 +150885,7 @@ } }, { - "accuracy": 0.9863408960794148, + "accuracy": 0.9869591240820131, "total_bits": 630355968, "q_proj": { "group_size": { @@ -150949,7 +150949,7 @@ } }, { - "accuracy": 0.9900789613786497, + "accuracy": 0.9900276762874503, "total_bits": 637362176, "q_proj": { "group_size": { @@ -151010,7 +151010,7 @@ } }, { - "accuracy": 0.9906859970406482, + "accuracy": 0.9909521066828778, "total_bits": 646823936, "q_proj": { "group_size": { @@ -151071,7 +151071,7 @@ } }, { - "accuracy": 0.9946052953600883, + "accuracy": 0.9946969633823947, "total_bits": 784740352, "q_proj": { "group_size": { @@ -151132,7 +151132,7 @@ } }, { - "accuracy": 0.9954711230177629, + "accuracy": 0.9955139622876519, "total_bits": 797818880, "q_proj": { "group_size": { @@ -151193,7 +151193,7 @@ } }, { - "accuracy": 0.9958753856389146, + "accuracy": 0.9959747599143731, "total_bits": 911749120, "q_proj": { "group_size": { @@ -151245,7 +151245,7 @@ } }, { - "accuracy": 0.9978627148819598, + "accuracy": 0.9978704595644223, "total_bits": 942718976, "q_proj": { "group_size": { @@ -151297,7 +151297,7 @@ } }, { - "accuracy": 0.9989234181611162, + "accuracy": 0.9989212251415378, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -151351,7 +151351,7 @@ ], "model.layers.77.mlp": [ { - "accuracy": 0.9187319906134355, + "accuracy": 0.9183139801025391, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -151403,7 +151403,7 @@ } }, { - "accuracy": 0.9210905213105052, + "accuracy": 0.9208045507732191, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -151455,7 +151455,7 @@ } }, { - "accuracy": 0.9298149974722612, + "accuracy": 0.9292756883721602, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -151504,7 +151504,7 @@ } }, { - "accuracy": 0.9325851076527646, + "accuracy": 0.9320116858733327, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -151553,7 +151553,7 @@ } }, { - "accuracy": 0.9604029655456543, + "accuracy": 0.9606047053086131, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -151605,7 +151605,7 @@ } }, { - "accuracy": 0.9634295984318382, + "accuracy": 0.9638074918797142, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -151657,7 +151657,7 @@ } }, { - "accuracy": 0.9673381667388112, + "accuracy": 0.9677240221123946, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -151706,7 +151706,7 @@ } }, { - "accuracy": 0.9792220529757047, + "accuracy": 0.9793694348711717, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -151749,7 +151749,7 @@ } }, { - "accuracy": 0.9809435809913435, + "accuracy": 0.9810052012142382, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -151792,7 +151792,7 @@ } }, { - "accuracy": 0.9797642136874952, + "accuracy": 0.9797676770310653, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -151844,7 +151844,7 @@ } }, { - "accuracy": 0.9822441229694768, + "accuracy": 0.9822212990961576, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -151896,7 +151896,7 @@ } }, { - "accuracy": 0.989603158674742, + "accuracy": 0.9896385332471446, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -151948,7 +151948,7 @@ } }, { - "accuracy": 0.9910800676596793, + "accuracy": 0.9910989215499476, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -152000,7 +152000,7 @@ } }, { - "accuracy": 0.9942138473454275, + "accuracy": 0.9942727037950566, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -152043,7 +152043,7 @@ } }, { - "accuracy": 0.9946485385298729, + "accuracy": 0.994648120120952, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -152092,7 +152092,7 @@ } }, { - "accuracy": 0.995416701624268, + "accuracy": 0.9954083848156428, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -152138,7 +152138,7 @@ } }, { - "accuracy": 0.9983493244569552, + "accuracy": 0.9983445849073561, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -152180,7 +152180,7 @@ ], "model.layers.78.self_attn": [ { - "accuracy": 0.9409458574495817, + "accuracy": 0.9435614284716154, "total_bits": 320757760, "q_proj": { "group_size": { @@ -152244,7 +152244,7 @@ } }, { - "accuracy": 0.9479577698205647, + "accuracy": 0.9458739255603991, "total_bits": 329080832, "q_proj": { "group_size": { @@ -152308,7 +152308,7 @@ } }, { - "accuracy": 0.9646490529963845, + "accuracy": 0.964287845711959, "total_bits": 336024576, "q_proj": { "group_size": { @@ -152372,7 +152372,7 @@ } }, { - "accuracy": 0.9703242590552882, + "accuracy": 0.9701100713328311, "total_bits": 401557504, "q_proj": { "group_size": { @@ -152436,7 +152436,7 @@ } }, { - "accuracy": 0.9735132800905328, + "accuracy": 0.9739637327821631, "total_bits": 475279360, "q_proj": { "group_size": { @@ -152500,7 +152500,7 @@ } }, { - "accuracy": 0.9739738985111839, + "accuracy": 0.9745071420544072, "total_bits": 475479040, "q_proj": { "group_size": { @@ -152564,7 +152564,7 @@ } }, { - "accuracy": 0.9856243431568146, + "accuracy": 0.9854652897307747, "total_bits": 609759232, "q_proj": { "group_size": { @@ -152616,7 +152616,7 @@ } }, { - "accuracy": 0.9858504282800775, + "accuracy": 0.9859525153511449, "total_bits": 610024448, "q_proj": { "group_size": { @@ -152668,7 +152668,7 @@ } }, { - "accuracy": 0.9870884489071997, + "accuracy": 0.9872107529326489, "total_bits": 615020544, "q_proj": { "group_size": { @@ -152720,7 +152720,7 @@ } }, { - "accuracy": 0.9874161128935061, + "accuracy": 0.987496366626338, "total_bits": 623951872, "q_proj": { "group_size": { @@ -152772,7 +152772,7 @@ } }, { - "accuracy": 0.9871393638221841, + "accuracy": 0.9868261829802865, "total_bits": 626473984, "q_proj": { "group_size": { @@ -152836,7 +152836,7 @@ } }, { - "accuracy": 0.9880114903575495, + "accuracy": 0.9877449015253469, "total_bits": 630355968, "q_proj": { "group_size": { @@ -152900,7 +152900,7 @@ } }, { - "accuracy": 0.9912842825839394, + "accuracy": 0.991370763433607, "total_bits": 637362176, "q_proj": { "group_size": { @@ -152961,7 +152961,7 @@ } }, { - "accuracy": 0.9920164072199872, + "accuracy": 0.9919306306462539, "total_bits": 646823936, "q_proj": { "group_size": { @@ -153022,7 +153022,7 @@ } }, { - "accuracy": 0.9952015872848662, + "accuracy": 0.9952077763645273, "total_bits": 784740352, "q_proj": { "group_size": { @@ -153083,7 +153083,7 @@ } }, { - "accuracy": 0.995973930154976, + "accuracy": 0.9960647417526496, "total_bits": 797818880, "q_proj": { "group_size": { @@ -153144,7 +153144,7 @@ } }, { - "accuracy": 0.9962569671241861, + "accuracy": 0.9961731331913095, "total_bits": 911749120, "q_proj": { "group_size": { @@ -153196,7 +153196,7 @@ } }, { - "accuracy": 0.9980999162714732, + "accuracy": 0.9981092520450291, "total_bits": 942718976, "q_proj": { "group_size": { @@ -153248,7 +153248,7 @@ } }, { - "accuracy": 0.9990158794741881, + "accuracy": 0.998986813876974, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -153302,7 +153302,7 @@ ], "model.layers.78.mlp": [ { - "accuracy": 0.905286475231773, + "accuracy": 0.9056085348129272, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -153354,7 +153354,7 @@ } }, { - "accuracy": 0.9087460229271337, + "accuracy": 0.9076831905465377, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -153406,7 +153406,7 @@ } }, { - "accuracy": 0.9169142058021144, + "accuracy": 0.9179282062932065, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -153455,7 +153455,7 @@ } }, { - "accuracy": 0.921290535675852, + "accuracy": 0.922288455461201, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -153504,7 +153504,7 @@ } }, { - "accuracy": 0.955506262026335, + "accuracy": 0.9552601701334903, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -153556,7 +153556,7 @@ } }, { - "accuracy": 0.9596404652846486, + "accuracy": 0.9584519643532603, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -153608,7 +153608,7 @@ } }, { - "accuracy": 0.964368738626179, + "accuracy": 0.9631062746047974, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -153657,7 +153657,7 @@ } }, { - "accuracy": 0.9758242949059135, + "accuracy": 0.9757449673978906, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -153700,7 +153700,7 @@ } }, { - "accuracy": 0.9780167027523643, + "accuracy": 0.9779634616876903, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -153743,7 +153743,7 @@ } }, { - "accuracy": 0.9771964863726967, + "accuracy": 0.9772762907178778, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -153795,7 +153795,7 @@ } }, { - "accuracy": 0.9802003016597346, + "accuracy": 0.9802230251462836, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -153847,7 +153847,7 @@ } }, { - "accuracy": 0.9882458777804124, + "accuracy": 0.9882395541981647, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -153899,7 +153899,7 @@ } }, { - "accuracy": 0.9899806113619554, + "accuracy": 0.9899796059257105, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -153951,7 +153951,7 @@ } }, { - "accuracy": 0.9930825060919711, + "accuracy": 0.9930541468294043, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -153994,7 +153994,7 @@ } }, { - "accuracy": 0.9938182128887427, + "accuracy": 0.9938322533118097, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -154043,7 +154043,7 @@ } }, { - "accuracy": 0.9948064155484501, + "accuracy": 0.9948169604728097, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -154089,7 +154089,7 @@ } }, { - "accuracy": 0.9975997936960898, + "accuracy": 0.997605842587195, "total_bits": 5660623360, "gate_proj": { "group_size": { @@ -154131,7 +154131,7 @@ ], "model.layers.79.self_attn": [ { - "accuracy": 0.9692286943134508, + "accuracy": 0.9692164690870988, "total_bits": 320757760, "q_proj": { "group_size": { @@ -154195,7 +154195,7 @@ } }, { - "accuracy": 0.9703308312516463, + "accuracy": 0.9697453818823162, "total_bits": 329080832, "q_proj": { "group_size": { @@ -154259,7 +154259,7 @@ } }, { - "accuracy": 0.9787688239624626, + "accuracy": 0.9787844231254176, "total_bits": 336024576, "q_proj": { "group_size": { @@ -154323,7 +154323,7 @@ } }, { - "accuracy": 0.9836779974008861, + "accuracy": 0.9835086245285837, "total_bits": 401557504, "q_proj": { "group_size": { @@ -154387,7 +154387,7 @@ } }, { - "accuracy": 0.9848514011031703, + "accuracy": 0.9846274664527491, "total_bits": 475279360, "q_proj": { "group_size": { @@ -154451,7 +154451,7 @@ } }, { - "accuracy": 0.9849773817940762, + "accuracy": 0.9852589604101682, "total_bits": 475479040, "q_proj": { "group_size": { @@ -154515,7 +154515,7 @@ } }, { - "accuracy": 0.991734336081304, + "accuracy": 0.9916843905260688, "total_bits": 609759232, "q_proj": { "group_size": { @@ -154567,7 +154567,7 @@ } }, { - "accuracy": 0.9919739420476713, + "accuracy": 0.9920145790827902, "total_bits": 610024448, "q_proj": { "group_size": { @@ -154619,7 +154619,7 @@ } }, { - "accuracy": 0.9925556488727268, + "accuracy": 0.9926326651322214, "total_bits": 615020544, "q_proj": { "group_size": { @@ -154671,7 +154671,7 @@ } }, { - "accuracy": 0.9927907374344374, + "accuracy": 0.9928497970104218, "total_bits": 623951872, "q_proj": { "group_size": { @@ -154723,7 +154723,7 @@ } }, { - "accuracy": 0.992471769452095, + "accuracy": 0.9924301128638419, "total_bits": 626473984, "q_proj": { "group_size": { @@ -154787,7 +154787,7 @@ } }, { - "accuracy": 0.9928735753423289, + "accuracy": 0.9930194749643928, "total_bits": 630355968, "q_proj": { "group_size": { @@ -154851,7 +154851,7 @@ } }, { - "accuracy": 0.994721231884078, + "accuracy": 0.9947665439624536, "total_bits": 637362176, "q_proj": { "group_size": { @@ -154912,7 +154912,7 @@ } }, { - "accuracy": 0.995207794402775, + "accuracy": 0.9951960444450378, "total_bits": 646823936, "q_proj": { "group_size": { @@ -154973,7 +154973,7 @@ } }, { - "accuracy": 0.9970686914105165, + "accuracy": 0.9971092623707495, "total_bits": 784740352, "q_proj": { "group_size": { @@ -155034,7 +155034,7 @@ } }, { - "accuracy": 0.997573757250058, + "accuracy": 0.9976081777560083, "total_bits": 797818880, "q_proj": { "group_size": { @@ -155095,7 +155095,7 @@ } }, { - "accuracy": 0.9977605325218878, + "accuracy": 0.9977982338321837, "total_bits": 911749120, "q_proj": { "group_size": { @@ -155147,7 +155147,7 @@ } }, { - "accuracy": 0.9988547173377714, + "accuracy": 0.9988587939817655, "total_bits": 942718976, "q_proj": { "group_size": { @@ -155199,7 +155199,7 @@ } }, { - "accuracy": 0.999392632591097, + "accuracy": 0.9993969189297212, "total_bits": 1213739008, "q_proj": { "group_size": { @@ -155253,7 +155253,7 @@ ], "model.layers.79.mlp": [ { - "accuracy": 0.9239560930352462, + "accuracy": 0.9239540162839388, "total_bits": 1575618640, "gate_proj": { "group_size": { @@ -155305,7 +155305,7 @@ } }, { - "accuracy": 0.9270516508503964, + "accuracy": 0.927004757680391, "total_bits": 1632503888, "gate_proj": { "group_size": { @@ -155357,7 +155357,7 @@ } }, { - "accuracy": 0.9334851064180073, + "accuracy": 0.9334283690703542, "total_bits": 1823397888, "gate_proj": { "group_size": { @@ -155406,7 +155406,7 @@ } }, { - "accuracy": 0.9357873389595434, + "accuracy": 0.9357349935330843, "total_bits": 2046482432, "gate_proj": { "group_size": { @@ -155455,7 +155455,7 @@ } }, { - "accuracy": 0.9623585374731767, + "accuracy": 0.9623635787712901, "total_bits": 2306033760, "gate_proj": { "group_size": { @@ -155507,7 +155507,7 @@ } }, { - "accuracy": 0.9660178956232572, + "accuracy": 0.9659812983713651, "total_bits": 2368669696, "gate_proj": { "group_size": { @@ -155559,7 +155559,7 @@ } }, { - "accuracy": 0.9686695086328607, + "accuracy": 0.9686264991760254, "total_bits": 2547224672, "gate_proj": { "group_size": { @@ -155608,7 +155608,7 @@ } }, { - "accuracy": 0.9804582799735823, + "accuracy": 0.9804466743218271, "total_bits": 2912382048, "gate_proj": { "group_size": { @@ -155651,7 +155651,7 @@ } }, { - "accuracy": 0.9818845579498693, + "accuracy": 0.9818733899216903, "total_bits": 2955347968, "gate_proj": { "group_size": { @@ -155694,7 +155694,7 @@ } }, { - "accuracy": 0.9806664695865229, + "accuracy": 0.9806787999052751, "total_bits": 2998880352, "gate_proj": { "group_size": { @@ -155746,7 +155746,7 @@ } }, { - "accuracy": 0.9833919186341135, + "accuracy": 0.9833984171089373, "total_bits": 3061516288, "gate_proj": { "group_size": { @@ -155798,7 +155798,7 @@ } }, { - "accuracy": 0.9900450957448859, + "accuracy": 0.9900529894389605, "total_bits": 3691726944, "gate_proj": { "group_size": { @@ -155850,7 +155850,7 @@ } }, { - "accuracy": 0.9916454616345858, + "accuracy": 0.9916460318000693, "total_bits": 3754362880, "gate_proj": { "group_size": { @@ -155902,7 +155902,7 @@ } }, { - "accuracy": 0.9943741065891165, + "accuracy": 0.9943661211352599, "total_bits": 4276045920, "gate_proj": { "group_size": { @@ -155945,7 +155945,7 @@ } }, { - "accuracy": 0.994851262553742, + "accuracy": 0.9948561348413166, "total_bits": 4425400832, "gate_proj": { "group_size": { @@ -155994,7 +155994,7 @@ } }, { - "accuracy": 0.995405664961589, + "accuracy": 0.995407624464286, "total_bits": 4823859712, "gate_proj": { "group_size": { @@ -156040,7 +156040,7 @@ } }, { - "accuracy": 0.9983516002171918, + "accuracy": 0.9983551927695149, "total_bits": 5660623360, "gate_proj": { "group_size": {