{ "best_metric": null, "best_model_checkpoint": null, "epoch": 4.99746835443038, "eval_steps": 100, "global_step": 987, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.005063291139240506, "grad_norm": 520145.53108452284, "learning_rate": 1.4084507042253521e-09, "logits/chosen": -16.270591735839844, "logits/rejected": -16.343984603881836, "logps/chosen": -186.17276000976562, "logps/rejected": -175.8095703125, "loss": 122464.3125, "rewards/accuracies": 0.0, "rewards/chosen": 0.0, "rewards/margins": 0.0, "rewards/rejected": 0.0, "step": 1 }, { "epoch": 0.05063291139240506, "grad_norm": 501181.05233525805, "learning_rate": 1.408450704225352e-08, "logits/chosen": -17.194263458251953, "logits/rejected": -17.04476547241211, "logps/chosen": -220.64031982421875, "logps/rejected": -220.79531860351562, "loss": 124716.2917, "rewards/accuracies": 0.4583333432674408, "rewards/chosen": 1.5937095554363623e-07, "rewards/margins": 1.5358187738456763e-05, "rewards/rejected": -1.5198814253380988e-05, "step": 10 }, { "epoch": 0.10126582278481013, "grad_norm": 537058.8643033113, "learning_rate": 2.816901408450704e-08, "logits/chosen": -16.468345642089844, "logits/rejected": -16.397050857543945, "logps/chosen": -238.16464233398438, "logps/rejected": -234.13320922851562, "loss": 125132.075, "rewards/accuracies": 0.512499988079071, "rewards/chosen": -7.482715773221571e-06, "rewards/margins": 1.4337347238324583e-05, "rewards/rejected": -2.1820069378009066e-05, "step": 20 }, { "epoch": 0.1518987341772152, "grad_norm": 457257.68659374124, "learning_rate": 4.2253521126760564e-08, "logits/chosen": -16.952747344970703, "logits/rejected": -16.70650863647461, "logps/chosen": -242.9259490966797, "logps/rejected": -242.9457244873047, "loss": 124660.25, "rewards/accuracies": 0.4375, "rewards/chosen": -3.21022052958142e-05, "rewards/margins": 3.5706521885003895e-05, "rewards/rejected": -6.780872354283929e-05, "step": 30 }, { "epoch": 0.20253164556962025, "grad_norm": 520967.9129238899, "learning_rate": 5.633802816901408e-08, "logits/chosen": -16.920284271240234, "logits/rejected": -16.8529052734375, "logps/chosen": -243.7992706298828, "logps/rejected": -244.38906860351562, "loss": 124148.0625, "rewards/accuracies": 0.550000011920929, "rewards/chosen": -0.0001235240779351443, "rewards/margins": 8.850651647662744e-05, "rewards/rejected": -0.00021203060168772936, "step": 40 }, { "epoch": 0.25316455696202533, "grad_norm": 722258.4292859514, "learning_rate": 7.042253521126761e-08, "logits/chosen": -16.24307632446289, "logits/rejected": -16.294937133789062, "logps/chosen": -238.68148803710938, "logps/rejected": -240.46337890625, "loss": 125272.85, "rewards/accuracies": 0.44999998807907104, "rewards/chosen": -0.00024854010553099215, "rewards/margins": -7.368279329966754e-05, "rewards/rejected": -0.00017485734133515507, "step": 50 }, { "epoch": 0.3037974683544304, "grad_norm": 521725.51159479923, "learning_rate": 8.450704225352113e-08, "logits/chosen": -16.547048568725586, "logits/rejected": -16.562244415283203, "logps/chosen": -234.24453735351562, "logps/rejected": -236.03823852539062, "loss": 123692.1, "rewards/accuracies": 0.4375, "rewards/chosen": -0.00036723288940265775, "rewards/margins": 6.122588274592999e-06, "rewards/rejected": -0.00037335552042350173, "step": 60 }, { "epoch": 0.35443037974683544, "grad_norm": 446768.20251500694, "learning_rate": 9.859154929577463e-08, "logits/chosen": -16.514156341552734, "logits/rejected": -16.41303062438965, "logps/chosen": -240.8957061767578, "logps/rejected": -235.2915496826172, "loss": 125937.8, "rewards/accuracies": 0.5874999761581421, "rewards/chosen": -0.00028800699510611594, "rewards/margins": 0.00021416530944406986, "rewards/rejected": -0.0005021723336540163, "step": 70 }, { "epoch": 0.4050632911392405, "grad_norm": 463557.5011981856, "learning_rate": 1.1267605633802817e-07, "logits/chosen": -16.711376190185547, "logits/rejected": -16.489612579345703, "logps/chosen": -243.5523681640625, "logps/rejected": -228.8307342529297, "loss": 125818.525, "rewards/accuracies": 0.550000011920929, "rewards/chosen": -0.0005035396316088736, "rewards/margins": 6.90509841660969e-05, "rewards/rejected": -0.0005725906230509281, "step": 80 }, { "epoch": 0.45569620253164556, "grad_norm": 465137.87035599066, "learning_rate": 1.2676056338028167e-07, "logits/chosen": -17.326900482177734, "logits/rejected": -17.396936416625977, "logps/chosen": -240.1623077392578, "logps/rejected": -234.27578735351562, "loss": 123894.4, "rewards/accuracies": 0.44999998807907104, "rewards/chosen": -0.0008088796166703105, "rewards/margins": -0.00010466824460308999, "rewards/rejected": -0.0007042114739306271, "step": 90 }, { "epoch": 0.5063291139240507, "grad_norm": 505006.4054603859, "learning_rate": 1.4084507042253522e-07, "logits/chosen": -16.5346736907959, "logits/rejected": -16.46234893798828, "logps/chosen": -238.9674530029297, "logps/rejected": -235.36239624023438, "loss": 126640.2125, "rewards/accuracies": 0.512499988079071, "rewards/chosen": -0.0006280581001192331, "rewards/margins": 2.4443055735900998e-05, "rewards/rejected": -0.0006525011267513037, "step": 100 }, { "epoch": 0.5569620253164557, "grad_norm": 475489.46555727004, "learning_rate": 1.549295774647887e-07, "logits/chosen": -16.67499351501465, "logits/rejected": -16.584075927734375, "logps/chosen": -240.5388946533203, "logps/rejected": -239.03366088867188, "loss": 122706.3, "rewards/accuracies": 0.5249999761581421, "rewards/chosen": -0.0006213908782228827, "rewards/margins": 0.00010261077841278166, "rewards/rejected": -0.0007240016711875796, "step": 110 }, { "epoch": 0.6075949367088608, "grad_norm": 492764.07090207015, "learning_rate": 1.6901408450704225e-07, "logits/chosen": -16.746532440185547, "logits/rejected": -16.617717742919922, "logps/chosen": -227.05398559570312, "logps/rejected": -225.60214233398438, "loss": 126588.925, "rewards/accuracies": 0.48750001192092896, "rewards/chosen": -0.0008889889577403665, "rewards/margins": 2.5076475139940158e-05, "rewards/rejected": -0.0009140653419308364, "step": 120 }, { "epoch": 0.6582278481012658, "grad_norm": 511084.4558498889, "learning_rate": 1.8309859154929577e-07, "logits/chosen": -16.747934341430664, "logits/rejected": -16.733430862426758, "logps/chosen": -240.7227325439453, "logps/rejected": -240.2967529296875, "loss": 125175.5125, "rewards/accuracies": 0.574999988079071, "rewards/chosen": -0.0010187395382672548, "rewards/margins": 0.002705145161598921, "rewards/rejected": -0.003723885165527463, "step": 130 }, { "epoch": 0.7088607594936709, "grad_norm": 540454.6644647518, "learning_rate": 1.9718309859154927e-07, "logits/chosen": -16.1859073638916, "logits/rejected": -16.264835357666016, "logps/chosen": -231.37173461914062, "logps/rejected": -227.0606689453125, "loss": 126058.6375, "rewards/accuracies": 0.4749999940395355, "rewards/chosen": -0.0008466474828310311, "rewards/margins": -4.7403918870259076e-05, "rewards/rejected": -0.0007992436294443905, "step": 140 }, { "epoch": 0.759493670886076, "grad_norm": 503077.16971538117, "learning_rate": 2.112676056338028e-07, "logits/chosen": -17.280269622802734, "logits/rejected": -17.093780517578125, "logps/chosen": -238.0977325439453, "logps/rejected": -238.93212890625, "loss": 126646.125, "rewards/accuracies": 0.512499988079071, "rewards/chosen": -0.0008783842204138637, "rewards/margins": 0.00026031016022898257, "rewards/rejected": -0.0011386943515390158, "step": 150 }, { "epoch": 0.810126582278481, "grad_norm": 541715.9624559938, "learning_rate": 2.2535211267605633e-07, "logits/chosen": -16.782550811767578, "logits/rejected": -16.79593276977539, "logps/chosen": -250.48593139648438, "logps/rejected": -249.44924926757812, "loss": 124718.425, "rewards/accuracies": 0.5249999761581421, "rewards/chosen": -0.0012445250758901238, "rewards/margins": 5.686017539119348e-05, "rewards/rejected": -0.0013013852294534445, "step": 160 }, { "epoch": 0.8607594936708861, "grad_norm": 548905.0358445289, "learning_rate": 2.394366197183098e-07, "logits/chosen": -17.04167938232422, "logits/rejected": -16.985572814941406, "logps/chosen": -255.06942749023438, "logps/rejected": -260.38128662109375, "loss": 125650.7625, "rewards/accuracies": 0.6499999761581421, "rewards/chosen": -0.0021144188940525055, "rewards/margins": 0.001583110773935914, "rewards/rejected": -0.003697529900819063, "step": 170 }, { "epoch": 0.9113924050632911, "grad_norm": 561549.4959644328, "learning_rate": 2.5352112676056334e-07, "logits/chosen": -16.703407287597656, "logits/rejected": -16.487037658691406, "logps/chosen": -232.529052734375, "logps/rejected": -228.3297576904297, "loss": 127298.1375, "rewards/accuracies": 0.5249999761581421, "rewards/chosen": -0.002397818025201559, "rewards/margins": 0.0011578220874071121, "rewards/rejected": -0.003555640112608671, "step": 180 }, { "epoch": 0.9620253164556962, "grad_norm": 565071.0053763993, "learning_rate": 2.6760563380281686e-07, "logits/chosen": -16.11090660095215, "logits/rejected": -16.053157806396484, "logps/chosen": -239.39205932617188, "logps/rejected": -235.435791015625, "loss": 127009.225, "rewards/accuracies": 0.5874999761581421, "rewards/chosen": -0.0010291519574820995, "rewards/margins": 0.00019036220328416675, "rewards/rejected": -0.0012195140589028597, "step": 190 }, { "epoch": 1.0126582278481013, "grad_norm": 497332.98430491646, "learning_rate": 2.8169014084507043e-07, "logits/chosen": -16.127140045166016, "logits/rejected": -15.988116264343262, "logps/chosen": -225.9070587158203, "logps/rejected": -227.90145874023438, "loss": 126358.875, "rewards/accuracies": 0.6000000238418579, "rewards/chosen": -0.0012379485415294766, "rewards/margins": 0.0006675361073575914, "rewards/rejected": -0.001905484707094729, "step": 200 }, { "epoch": 1.0632911392405062, "grad_norm": 935063.760892245, "learning_rate": 2.957746478873239e-07, "logits/chosen": -16.701793670654297, "logits/rejected": -16.669902801513672, "logps/chosen": -230.3677520751953, "logps/rejected": -229.03921508789062, "loss": 124250.775, "rewards/accuracies": 0.737500011920929, "rewards/chosen": -0.0006614397279918194, "rewards/margins": 0.003238010685890913, "rewards/rejected": -0.003899450646713376, "step": 210 }, { "epoch": 1.1139240506329113, "grad_norm": 517399.2020129059, "learning_rate": 3.098591549295774e-07, "logits/chosen": -16.413972854614258, "logits/rejected": -16.371458053588867, "logps/chosen": -247.8984832763672, "logps/rejected": -249.5322723388672, "loss": 124993.7375, "rewards/accuracies": 0.762499988079071, "rewards/chosen": -0.0012706981506198645, "rewards/margins": 0.003060612827539444, "rewards/rejected": -0.004331310745328665, "step": 220 }, { "epoch": 1.1645569620253164, "grad_norm": 499036.7717944408, "learning_rate": 3.23943661971831e-07, "logits/chosen": -15.908624649047852, "logits/rejected": -15.847338676452637, "logps/chosen": -236.7013397216797, "logps/rejected": -239.3136749267578, "loss": 122842.5, "rewards/accuracies": 0.800000011920929, "rewards/chosen": -0.0006655483739450574, "rewards/margins": 0.0032406128011643887, "rewards/rejected": -0.0039061610586941242, "step": 230 }, { "epoch": 1.2151898734177216, "grad_norm": 540681.7856619481, "learning_rate": 3.380281690140845e-07, "logits/chosen": -16.052249908447266, "logits/rejected": -15.99653148651123, "logps/chosen": -229.74832153320312, "logps/rejected": -230.9803009033203, "loss": 124587.3625, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": -0.0007962372037582099, "rewards/margins": 0.0025483998470008373, "rewards/rejected": -0.003344637108966708, "step": 240 }, { "epoch": 1.2658227848101267, "grad_norm": 1023950.8355601664, "learning_rate": 3.52112676056338e-07, "logits/chosen": -15.299288749694824, "logits/rejected": -15.215815544128418, "logps/chosen": -231.2301788330078, "logps/rejected": -232.03359985351562, "loss": 121822.4, "rewards/accuracies": 0.7250000238418579, "rewards/chosen": -7.512583579227794e-06, "rewards/margins": 0.003883513854816556, "rewards/rejected": -0.0038910270668566227, "step": 250 }, { "epoch": 1.3164556962025316, "grad_norm": 620253.8184950812, "learning_rate": 3.6619718309859155e-07, "logits/chosen": -16.167770385742188, "logits/rejected": -15.915590286254883, "logps/chosen": -238.9904327392578, "logps/rejected": -239.73953247070312, "loss": 123388.8625, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -0.00017356239550281316, "rewards/margins": 0.0050824107602238655, "rewards/rejected": -0.005255972500890493, "step": 260 }, { "epoch": 1.3670886075949367, "grad_norm": 575104.3218096169, "learning_rate": 3.8028169014084507e-07, "logits/chosen": -15.480558395385742, "logits/rejected": -15.386639595031738, "logps/chosen": -241.60879516601562, "logps/rejected": -250.003173828125, "loss": 123555.7, "rewards/accuracies": 0.737500011920929, "rewards/chosen": -0.001139859901741147, "rewards/margins": 0.005077657289803028, "rewards/rejected": -0.0062175169587135315, "step": 270 }, { "epoch": 1.4177215189873418, "grad_norm": 601224.4433091934, "learning_rate": 3.9436619718309853e-07, "logits/chosen": -15.266016960144043, "logits/rejected": -15.313554763793945, "logps/chosen": -230.73397827148438, "logps/rejected": -237.3317108154297, "loss": 125556.675, "rewards/accuracies": 0.762499988079071, "rewards/chosen": 0.0007209269679151475, "rewards/margins": 0.00534270191565156, "rewards/rejected": -0.004621774889528751, "step": 280 }, { "epoch": 1.4683544303797469, "grad_norm": 751936.3077706753, "learning_rate": 4.084507042253521e-07, "logits/chosen": -14.600263595581055, "logits/rejected": -14.538311958312988, "logps/chosen": -224.1177520751953, "logps/rejected": -226.97879028320312, "loss": 123584.675, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": 0.0011863496620208025, "rewards/margins": 0.007649322040379047, "rewards/rejected": -0.006462973542511463, "step": 290 }, { "epoch": 1.518987341772152, "grad_norm": 575660.5828565176, "learning_rate": 4.225352112676056e-07, "logits/chosen": -14.935551643371582, "logits/rejected": -15.062429428100586, "logps/chosen": -235.7123565673828, "logps/rejected": -245.36181640625, "loss": 122562.1375, "rewards/accuracies": 0.75, "rewards/chosen": 0.0014863747637718916, "rewards/margins": 0.0057060932740569115, "rewards/rejected": -0.0042197187431156635, "step": 300 }, { "epoch": 1.5696202531645569, "grad_norm": 619514.1083852616, "learning_rate": 4.366197183098591e-07, "logits/chosen": -14.678690910339355, "logits/rejected": -14.617218017578125, "logps/chosen": -229.6386260986328, "logps/rejected": -234.1474151611328, "loss": 123630.225, "rewards/accuracies": 0.699999988079071, "rewards/chosen": 0.0006864996394142509, "rewards/margins": 0.004933560267090797, "rewards/rejected": -0.004247060976922512, "step": 310 }, { "epoch": 1.620253164556962, "grad_norm": 738538.1512211321, "learning_rate": 4.5070422535211266e-07, "logits/chosen": -14.131611824035645, "logits/rejected": -14.156657218933105, "logps/chosen": -241.20156860351562, "logps/rejected": -248.2321319580078, "loss": 124158.6, "rewards/accuracies": 0.699999988079071, "rewards/chosen": -0.0009155808947980404, "rewards/margins": 0.006913213524967432, "rewards/rejected": -0.007828795351088047, "step": 320 }, { "epoch": 1.6708860759493671, "grad_norm": 688317.7143989427, "learning_rate": 4.647887323943662e-07, "logits/chosen": -13.791796684265137, "logits/rejected": -13.970884323120117, "logps/chosen": -228.53079223632812, "logps/rejected": -235.5008087158203, "loss": 123378.175, "rewards/accuracies": 0.7250000238418579, "rewards/chosen": 0.0017698236042633653, "rewards/margins": 0.006004182621836662, "rewards/rejected": -0.004234359599649906, "step": 330 }, { "epoch": 1.721518987341772, "grad_norm": 693314.5034252935, "learning_rate": 4.788732394366196e-07, "logits/chosen": -13.555567741394043, "logits/rejected": -13.32630729675293, "logps/chosen": -227.0249481201172, "logps/rejected": -232.2772216796875, "loss": 122521.475, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": 0.001143553527072072, "rewards/margins": 0.009070896543562412, "rewards/rejected": -0.00792734231799841, "step": 340 }, { "epoch": 1.7721518987341773, "grad_norm": 758709.6120906892, "learning_rate": 4.929577464788733e-07, "logits/chosen": -13.520563125610352, "logits/rejected": -13.633130073547363, "logps/chosen": -234.7182159423828, "logps/rejected": -248.12890625, "loss": 121557.575, "rewards/accuracies": 0.675000011920929, "rewards/chosen": -0.00047356385039165616, "rewards/margins": 0.00813873577862978, "rewards/rejected": -0.008612299337983131, "step": 350 }, { "epoch": 1.8227848101265822, "grad_norm": 689974.393201542, "learning_rate": 4.992165465371357e-07, "logits/chosen": -12.841153144836426, "logits/rejected": -12.86094856262207, "logps/chosen": -232.314697265625, "logps/rejected": -232.64297485351562, "loss": 121436.65, "rewards/accuracies": 0.7124999761581421, "rewards/chosen": 0.0036194869317114353, "rewards/margins": 0.009506477043032646, "rewards/rejected": -0.005886988714337349, "step": 360 }, { "epoch": 1.8734177215189873, "grad_norm": 883375.543329047, "learning_rate": 4.976496396114071e-07, "logits/chosen": -12.77904224395752, "logits/rejected": -12.76900577545166, "logps/chosen": -239.8730010986328, "logps/rejected": -251.4569549560547, "loss": 122456.925, "rewards/accuracies": 0.75, "rewards/chosen": -0.0006393647054210305, "rewards/margins": 0.008665768429636955, "rewards/rejected": -0.009305133484303951, "step": 370 }, { "epoch": 1.9240506329113924, "grad_norm": 797554.0864386982, "learning_rate": 4.960827326856785e-07, "logits/chosen": -13.028135299682617, "logits/rejected": -13.148831367492676, "logps/chosen": -237.040771484375, "logps/rejected": -244.45181274414062, "loss": 124907.725, "rewards/accuracies": 0.699999988079071, "rewards/chosen": 0.0025544934906065464, "rewards/margins": 0.008132859133183956, "rewards/rejected": -0.005578366108238697, "step": 380 }, { "epoch": 1.9746835443037973, "grad_norm": 793120.1180084129, "learning_rate": 4.945158257599498e-07, "logits/chosen": -12.312803268432617, "logits/rejected": -12.135167121887207, "logps/chosen": -235.60360717773438, "logps/rejected": -242.9219207763672, "loss": 121583.8, "rewards/accuracies": 0.75, "rewards/chosen": 0.003660207614302635, "rewards/margins": 0.011001082137227058, "rewards/rejected": -0.007340874522924423, "step": 390 }, { "epoch": 2.0253164556962027, "grad_norm": 767339.6192091529, "learning_rate": 4.929489188342212e-07, "logits/chosen": -12.052891731262207, "logits/rejected": -11.94625473022461, "logps/chosen": -225.0377197265625, "logps/rejected": -243.81039428710938, "loss": 119737.85, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": 0.006956162396818399, "rewards/margins": 0.01727995090186596, "rewards/rejected": -0.010323788039386272, "step": 400 }, { "epoch": 2.0759493670886076, "grad_norm": 936793.207320047, "learning_rate": 4.913820119084926e-07, "logits/chosen": -11.38767147064209, "logits/rejected": -11.339715957641602, "logps/chosen": -219.8796844482422, "logps/rejected": -252.80581665039062, "loss": 114021.05, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": 0.008199459873139858, "rewards/margins": 0.031510110944509506, "rewards/rejected": -0.023310650140047073, "step": 410 }, { "epoch": 2.1265822784810124, "grad_norm": 1035986.8564166825, "learning_rate": 4.89815104982764e-07, "logits/chosen": -10.819408416748047, "logits/rejected": -10.774351119995117, "logps/chosen": -231.78854370117188, "logps/rejected": -260.20355224609375, "loss": 116051.6, "rewards/accuracies": 0.875, "rewards/chosen": 0.0056950985454022884, "rewards/margins": 0.027868490666151047, "rewards/rejected": -0.02217339165508747, "step": 420 }, { "epoch": 2.1772151898734178, "grad_norm": 1036991.7861177241, "learning_rate": 4.882481980570354e-07, "logits/chosen": -10.84526252746582, "logits/rejected": -10.708145141601562, "logps/chosen": -221.5430908203125, "logps/rejected": -257.36114501953125, "loss": 113501.175, "rewards/accuracies": 0.875, "rewards/chosen": 0.005057200789451599, "rewards/margins": 0.038923002779483795, "rewards/rejected": -0.033865805715322495, "step": 430 }, { "epoch": 2.2278481012658227, "grad_norm": 1227488.243303788, "learning_rate": 4.866812911313068e-07, "logits/chosen": -10.5010986328125, "logits/rejected": -10.63232135772705, "logps/chosen": -233.42373657226562, "logps/rejected": -276.0982666015625, "loss": 112100.4, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 0.004779786802828312, "rewards/margins": 0.040522992610931396, "rewards/rejected": -0.03574320673942566, "step": 440 }, { "epoch": 2.278481012658228, "grad_norm": 1079397.6974786038, "learning_rate": 4.851143842055782e-07, "logits/chosen": -10.104026794433594, "logits/rejected": -10.142271995544434, "logps/chosen": -216.66940307617188, "logps/rejected": -258.98858642578125, "loss": 112483.4, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": 0.0053299954161047935, "rewards/margins": 0.03484385460615158, "rewards/rejected": -0.029513856396079063, "step": 450 }, { "epoch": 2.329113924050633, "grad_norm": 1367054.8438774655, "learning_rate": 4.835474772798496e-07, "logits/chosen": -10.148681640625, "logits/rejected": -10.183786392211914, "logps/chosen": -233.730224609375, "logps/rejected": -278.64349365234375, "loss": 111561.6625, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 0.011530257761478424, "rewards/margins": 0.04578756168484688, "rewards/rejected": -0.034257303923368454, "step": 460 }, { "epoch": 2.379746835443038, "grad_norm": 1298484.9349088285, "learning_rate": 4.819805703541209e-07, "logits/chosen": -10.018949508666992, "logits/rejected": -10.097805976867676, "logps/chosen": -224.6026153564453, "logps/rejected": -270.0591735839844, "loss": 112710.1875, "rewards/accuracies": 0.875, "rewards/chosen": 0.011801879853010178, "rewards/margins": 0.040784891694784164, "rewards/rejected": -0.028983011841773987, "step": 470 }, { "epoch": 2.430379746835443, "grad_norm": 1428524.6930006845, "learning_rate": 4.804136634283923e-07, "logits/chosen": -9.595979690551758, "logits/rejected": -9.634994506835938, "logps/chosen": -265.3009338378906, "logps/rejected": -315.98541259765625, "loss": 110031.3, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": 0.0027348275762051344, "rewards/margins": 0.05151837319135666, "rewards/rejected": -0.048783544450998306, "step": 480 }, { "epoch": 2.481012658227848, "grad_norm": 1467649.8441612076, "learning_rate": 4.788467565026637e-07, "logits/chosen": -8.871723175048828, "logits/rejected": -8.764354705810547, "logps/chosen": -203.2312774658203, "logps/rejected": -241.612548828125, "loss": 110534.325, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 0.01302252896130085, "rewards/margins": 0.03907207027077675, "rewards/rejected": -0.02604953944683075, "step": 490 }, { "epoch": 2.5316455696202533, "grad_norm": 1382959.9591988046, "learning_rate": 4.772798495769351e-07, "logits/chosen": -8.468270301818848, "logits/rejected": -8.384966850280762, "logps/chosen": -226.46237182617188, "logps/rejected": -269.6461181640625, "loss": 110480.175, "rewards/accuracies": 0.824999988079071, "rewards/chosen": 0.015191495418548584, "rewards/margins": 0.0456535741686821, "rewards/rejected": -0.030462080612778664, "step": 500 }, { "epoch": 2.5822784810126582, "grad_norm": 1369494.2190603705, "learning_rate": 4.757129426512065e-07, "logits/chosen": -8.634099006652832, "logits/rejected": -8.640868186950684, "logps/chosen": -232.20022583007812, "logps/rejected": -304.80352783203125, "loss": 109921.975, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 0.013228721916675568, "rewards/margins": 0.07378505170345306, "rewards/rejected": -0.060556329786777496, "step": 510 }, { "epoch": 2.632911392405063, "grad_norm": 1750255.0550240122, "learning_rate": 4.741460357254779e-07, "logits/chosen": -7.8379316329956055, "logits/rejected": -7.4784440994262695, "logps/chosen": -213.3401641845703, "logps/rejected": -258.43743896484375, "loss": 111730.3875, "rewards/accuracies": 0.875, "rewards/chosen": 0.016342563554644585, "rewards/margins": 0.048144370317459106, "rewards/rejected": -0.03180180490016937, "step": 520 }, { "epoch": 2.6835443037974684, "grad_norm": 1447093.2174814222, "learning_rate": 4.7257912879974927e-07, "logits/chosen": -8.354089736938477, "logits/rejected": -7.889782905578613, "logps/chosen": -225.5243682861328, "logps/rejected": -276.7877502441406, "loss": 109226.9625, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 0.013667022809386253, "rewards/margins": 0.05627403408288956, "rewards/rejected": -0.042607005685567856, "step": 530 }, { "epoch": 2.7341772151898733, "grad_norm": 1477083.7533012358, "learning_rate": 4.710122218740207e-07, "logits/chosen": -7.921019077301025, "logits/rejected": -7.979846000671387, "logps/chosen": -237.23715209960938, "logps/rejected": -285.4289855957031, "loss": 109592.125, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 0.014952963218092918, "rewards/margins": 0.055934417992830276, "rewards/rejected": -0.040981464087963104, "step": 540 }, { "epoch": 2.7848101265822782, "grad_norm": 1486366.6324330876, "learning_rate": 4.6944531494829204e-07, "logits/chosen": -7.12634801864624, "logits/rejected": -7.396058082580566, "logps/chosen": -226.1304168701172, "logps/rejected": -276.8672790527344, "loss": 108245.925, "rewards/accuracies": 0.862500011920929, "rewards/chosen": 0.00994019117206335, "rewards/margins": 0.0537477545440197, "rewards/rejected": -0.04380756989121437, "step": 550 }, { "epoch": 2.8354430379746836, "grad_norm": 1560304.698196799, "learning_rate": 4.6787840802256345e-07, "logits/chosen": -7.268878936767578, "logits/rejected": -7.414219856262207, "logps/chosen": -215.24661254882812, "logps/rejected": -276.79437255859375, "loss": 110187.5125, "rewards/accuracies": 0.862500011920929, "rewards/chosen": 0.016926631331443787, "rewards/margins": 0.05572710186243057, "rewards/rejected": -0.03880046680569649, "step": 560 }, { "epoch": 2.8860759493670884, "grad_norm": 1647695.8714812996, "learning_rate": 4.663115010968348e-07, "logits/chosen": -8.584083557128906, "logits/rejected": -8.43793773651123, "logps/chosen": -239.3496856689453, "logps/rejected": -301.948974609375, "loss": 108493.15, "rewards/accuracies": 0.875, "rewards/chosen": 0.007640582975000143, "rewards/margins": 0.06335236132144928, "rewards/rejected": -0.0557117760181427, "step": 570 }, { "epoch": 2.9367088607594938, "grad_norm": 1523200.3846012072, "learning_rate": 4.647445941711062e-07, "logits/chosen": -8.875934600830078, "logits/rejected": -8.860316276550293, "logps/chosen": -234.2982635498047, "logps/rejected": -293.39727783203125, "loss": 107204.65, "rewards/accuracies": 0.9375, "rewards/chosen": 0.0077833631075918674, "rewards/margins": 0.061719853430986404, "rewards/rejected": -0.05393648147583008, "step": 580 }, { "epoch": 2.9873417721518987, "grad_norm": 1605115.356703113, "learning_rate": 4.631776872453776e-07, "logits/chosen": -8.788633346557617, "logits/rejected": -8.637460708618164, "logps/chosen": -257.7025146484375, "logps/rejected": -303.82147216796875, "loss": 108959.225, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": 0.0057020229287445545, "rewards/margins": 0.053022872656583786, "rewards/rejected": -0.04732084274291992, "step": 590 }, { "epoch": 3.037974683544304, "grad_norm": 1435515.2852262415, "learning_rate": 4.61610780319649e-07, "logits/chosen": -7.956998348236084, "logits/rejected": -7.496169090270996, "logps/chosen": -219.92410278320312, "logps/rejected": -310.20123291015625, "loss": 95986.4875, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.016220271587371826, "rewards/margins": 0.09167212247848511, "rewards/rejected": -0.07545184344053268, "step": 600 }, { "epoch": 3.088607594936709, "grad_norm": 1646011.901841717, "learning_rate": 4.6004387339392035e-07, "logits/chosen": -7.747580051422119, "logits/rejected": -7.5227952003479, "logps/chosen": -217.8295440673828, "logps/rejected": -343.4312438964844, "loss": 91538.925, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.02667585015296936, "rewards/margins": 0.12547221779823303, "rewards/rejected": -0.09879636764526367, "step": 610 }, { "epoch": 3.1392405063291138, "grad_norm": 1631989.4144731541, "learning_rate": 4.5847696646819176e-07, "logits/chosen": -6.8127121925354, "logits/rejected": -6.8090972900390625, "logps/chosen": -209.46859741210938, "logps/rejected": -332.0594482421875, "loss": 92242.9, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 0.026208167895674706, "rewards/margins": 0.12268342822790146, "rewards/rejected": -0.0964752584695816, "step": 620 }, { "epoch": 3.189873417721519, "grad_norm": 1627589.9925143481, "learning_rate": 4.569100595424631e-07, "logits/chosen": -6.631221771240234, "logits/rejected": -6.502354621887207, "logps/chosen": -211.57974243164062, "logps/rejected": -333.447265625, "loss": 89921.25, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.022689208388328552, "rewards/margins": 0.12395058572292328, "rewards/rejected": -0.10126137733459473, "step": 630 }, { "epoch": 3.240506329113924, "grad_norm": 1780107.5787213328, "learning_rate": 4.5534315261673453e-07, "logits/chosen": -7.868208885192871, "logits/rejected": -7.755393981933594, "logps/chosen": -209.3970184326172, "logps/rejected": -341.9508056640625, "loss": 89608.1875, "rewards/accuracies": 0.9375, "rewards/chosen": 0.027028566226363182, "rewards/margins": 0.133165642619133, "rewards/rejected": -0.10613708198070526, "step": 640 }, { "epoch": 3.291139240506329, "grad_norm": 1730512.4518714033, "learning_rate": 4.5377624569100595e-07, "logits/chosen": -7.359053134918213, "logits/rejected": -7.324367523193359, "logps/chosen": -193.1954803466797, "logps/rejected": -309.5513610839844, "loss": 93257.225, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 0.028996175155043602, "rewards/margins": 0.11760006099939346, "rewards/rejected": -0.08860386908054352, "step": 650 }, { "epoch": 3.3417721518987342, "grad_norm": 1692816.769511115, "learning_rate": 4.5220933876527736e-07, "logits/chosen": -8.043203353881836, "logits/rejected": -8.003018379211426, "logps/chosen": -211.73648071289062, "logps/rejected": -336.10455322265625, "loss": 88400.4688, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 0.024640550836920738, "rewards/margins": 0.12655004858970642, "rewards/rejected": -0.10190950334072113, "step": 660 }, { "epoch": 3.392405063291139, "grad_norm": 1906377.7496358757, "learning_rate": 4.506424318395487e-07, "logits/chosen": -7.25619649887085, "logits/rejected": -7.37869119644165, "logps/chosen": -197.8258819580078, "logps/rejected": -324.2138671875, "loss": 89983.5688, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 0.026263948529958725, "rewards/margins": 0.12702925503253937, "rewards/rejected": -0.10076530277729034, "step": 670 }, { "epoch": 3.4430379746835444, "grad_norm": 1785643.0594316572, "learning_rate": 4.4907552491382013e-07, "logits/chosen": -6.798577785491943, "logits/rejected": -6.7768073081970215, "logps/chosen": -208.5835723876953, "logps/rejected": -323.3017883300781, "loss": 89767.5, "rewards/accuracies": 0.9375, "rewards/chosen": 0.025741413235664368, "rewards/margins": 0.1167701929807663, "rewards/rejected": -0.09102877229452133, "step": 680 }, { "epoch": 3.4936708860759493, "grad_norm": 2393957.296937455, "learning_rate": 4.475086179880915e-07, "logits/chosen": -6.352355480194092, "logits/rejected": -6.526197910308838, "logps/chosen": -187.56597900390625, "logps/rejected": -306.5972595214844, "loss": 89036.6875, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 0.024061182513833046, "rewards/margins": 0.11990946531295776, "rewards/rejected": -0.09584827721118927, "step": 690 }, { "epoch": 3.5443037974683547, "grad_norm": 1811486.2204670438, "learning_rate": 4.459417110623629e-07, "logits/chosen": -5.7466630935668945, "logits/rejected": -5.797163486480713, "logps/chosen": -212.6585235595703, "logps/rejected": -364.36199951171875, "loss": 88031.3, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.027677077800035477, "rewards/margins": 0.14764061570167542, "rewards/rejected": -0.11996352672576904, "step": 700 }, { "epoch": 3.5949367088607596, "grad_norm": 1724684.5755440604, "learning_rate": 4.4437480413663426e-07, "logits/chosen": -5.412962436676025, "logits/rejected": -5.541121959686279, "logps/chosen": -202.39065551757812, "logps/rejected": -333.0758056640625, "loss": 86956.675, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.0247800350189209, "rewards/margins": 0.12825721502304077, "rewards/rejected": -0.10347716510295868, "step": 710 }, { "epoch": 3.6455696202531644, "grad_norm": 1933271.7611355048, "learning_rate": 4.4280789721090567e-07, "logits/chosen": -5.053005218505859, "logits/rejected": -4.886711597442627, "logps/chosen": -199.10885620117188, "logps/rejected": -317.7257385253906, "loss": 86655.0125, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 0.02152046002447605, "rewards/margins": 0.11774978786706924, "rewards/rejected": -0.09622932970523834, "step": 720 }, { "epoch": 3.6962025316455698, "grad_norm": 2267463.489494214, "learning_rate": 4.4124099028517703e-07, "logits/chosen": -6.616279602050781, "logits/rejected": -6.9615797996521, "logps/chosen": -200.58961486816406, "logps/rejected": -351.6376953125, "loss": 86181.3938, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.032253801822662354, "rewards/margins": 0.14937567710876465, "rewards/rejected": -0.1171218603849411, "step": 730 }, { "epoch": 3.7468354430379747, "grad_norm": 1734288.0953653858, "learning_rate": 4.3967408335944844e-07, "logits/chosen": -5.873335361480713, "logits/rejected": -5.689335823059082, "logps/chosen": -217.43637084960938, "logps/rejected": -350.2752990722656, "loss": 86780.825, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.031159091740846634, "rewards/margins": 0.13692796230316162, "rewards/rejected": -0.10576887428760529, "step": 740 }, { "epoch": 3.7974683544303796, "grad_norm": 1741715.9901586007, "learning_rate": 4.381071764337198e-07, "logits/chosen": -7.123785972595215, "logits/rejected": -7.188807487487793, "logps/chosen": -207.00045776367188, "logps/rejected": -336.5976867675781, "loss": 86139.5625, "rewards/accuracies": 0.9375, "rewards/chosen": 0.03052128478884697, "rewards/margins": 0.13043463230133057, "rewards/rejected": -0.0999133437871933, "step": 750 }, { "epoch": 3.848101265822785, "grad_norm": 1879351.8394690978, "learning_rate": 4.365402695079912e-07, "logits/chosen": -7.820990085601807, "logits/rejected": -7.7128729820251465, "logps/chosen": -213.57388305664062, "logps/rejected": -362.5634460449219, "loss": 87478.3625, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.03660900145769119, "rewards/margins": 0.1480773240327835, "rewards/rejected": -0.11146833002567291, "step": 760 }, { "epoch": 3.8987341772151898, "grad_norm": 1968713.4204386624, "learning_rate": 4.349733625822626e-07, "logits/chosen": -7.314540863037109, "logits/rejected": -7.363668918609619, "logps/chosen": -213.6930694580078, "logps/rejected": -367.44073486328125, "loss": 86825.5813, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 0.026752913370728493, "rewards/margins": 0.15061405301094055, "rewards/rejected": -0.1238611489534378, "step": 770 }, { "epoch": 3.9493670886075947, "grad_norm": 2163439.406665409, "learning_rate": 4.33406455656534e-07, "logits/chosen": -7.67099666595459, "logits/rejected": -7.536408424377441, "logps/chosen": -213.9747772216797, "logps/rejected": -344.7560119628906, "loss": 86913.0375, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 0.029844319447875023, "rewards/margins": 0.12930825352668762, "rewards/rejected": -0.09946390986442566, "step": 780 }, { "epoch": 4.0, "grad_norm": 1866234.1823014135, "learning_rate": 4.3183954873080535e-07, "logits/chosen": -7.922532081604004, "logits/rejected": -7.692726135253906, "logps/chosen": -211.41653442382812, "logps/rejected": -349.7116394042969, "loss": 86592.8938, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.027728911489248276, "rewards/margins": 0.1435452550649643, "rewards/rejected": -0.11581633985042572, "step": 790 }, { "epoch": 4.050632911392405, "grad_norm": 1782853.8797277175, "learning_rate": 4.3027264180507676e-07, "logits/chosen": -8.29829216003418, "logits/rejected": -8.205643653869629, "logps/chosen": -178.8797149658203, "logps/rejected": -378.06121826171875, "loss": 69143.425, "rewards/accuracies": 0.9375, "rewards/chosen": 0.05098045617341995, "rewards/margins": 0.1993386447429657, "rewards/rejected": -0.14835818111896515, "step": 800 }, { "epoch": 4.10126582278481, "grad_norm": 1719472.9461235409, "learning_rate": 4.287057348793481e-07, "logits/chosen": -7.558290958404541, "logits/rejected": -7.646592617034912, "logps/chosen": -186.36911010742188, "logps/rejected": -386.6961975097656, "loss": 67634.3375, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 0.04189852252602577, "rewards/margins": 0.19968575239181519, "rewards/rejected": -0.1577872335910797, "step": 810 }, { "epoch": 4.151898734177215, "grad_norm": 1571399.8942716653, "learning_rate": 4.2713882795361953e-07, "logits/chosen": -7.811161994934082, "logits/rejected": -7.783130645751953, "logps/chosen": -181.81602478027344, "logps/rejected": -402.1683654785156, "loss": 66806.9187, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.049001529812812805, "rewards/margins": 0.21849961578845978, "rewards/rejected": -0.16949808597564697, "step": 820 }, { "epoch": 4.2025316455696204, "grad_norm": 1992030.3917670588, "learning_rate": 4.255719210278909e-07, "logits/chosen": -7.349759101867676, "logits/rejected": -7.380797386169434, "logps/chosen": -175.21702575683594, "logps/rejected": -396.2167053222656, "loss": 67021.875, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.05283821374177933, "rewards/margins": 0.22190704941749573, "rewards/rejected": -0.169068843126297, "step": 830 }, { "epoch": 4.253164556962025, "grad_norm": 1859879.670487208, "learning_rate": 4.2400501410216235e-07, "logits/chosen": -7.482248783111572, "logits/rejected": -7.252910614013672, "logps/chosen": -187.070556640625, "logps/rejected": -401.1556701660156, "loss": 68463.9, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.05697192624211311, "rewards/margins": 0.21645841002464294, "rewards/rejected": -0.15948647260665894, "step": 840 }, { "epoch": 4.30379746835443, "grad_norm": 1688181.1410657803, "learning_rate": 4.224381071764337e-07, "logits/chosen": -5.693742275238037, "logits/rejected": -5.435591697692871, "logps/chosen": -198.21900939941406, "logps/rejected": -398.49981689453125, "loss": 67266.2, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.04546400159597397, "rewards/margins": 0.20465342700481415, "rewards/rejected": -0.15918943285942078, "step": 850 }, { "epoch": 4.3544303797468356, "grad_norm": 1750431.6432656392, "learning_rate": 4.208712002507051e-07, "logits/chosen": -8.664016723632812, "logits/rejected": -8.082508087158203, "logps/chosen": -178.05966186523438, "logps/rejected": -402.77093505859375, "loss": 65760.2625, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.056066203862428665, "rewards/margins": 0.22950176894664764, "rewards/rejected": -0.17343556880950928, "step": 860 }, { "epoch": 4.405063291139241, "grad_norm": 1904336.610304837, "learning_rate": 4.193042933249765e-07, "logits/chosen": -5.778517723083496, "logits/rejected": -5.432709693908691, "logps/chosen": -176.563720703125, "logps/rejected": -379.2276916503906, "loss": 67058.1125, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 0.05091014504432678, "rewards/margins": 0.2058809995651245, "rewards/rejected": -0.15497085452079773, "step": 870 }, { "epoch": 4.455696202531645, "grad_norm": 1779397.1811982268, "learning_rate": 4.177373863992479e-07, "logits/chosen": -6.937778472900391, "logits/rejected": -6.611588954925537, "logps/chosen": -180.23001098632812, "logps/rejected": -400.9800720214844, "loss": 67019.0875, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.05085798352956772, "rewards/margins": 0.2235671728849411, "rewards/rejected": -0.17270918190479279, "step": 880 }, { "epoch": 4.506329113924051, "grad_norm": 1755630.994265544, "learning_rate": 4.1617047947351925e-07, "logits/chosen": -6.663479804992676, "logits/rejected": -6.144991397857666, "logps/chosen": -189.93707275390625, "logps/rejected": -383.9622802734375, "loss": 66060.8813, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.053109876811504364, "rewards/margins": 0.20497091114521027, "rewards/rejected": -0.1518610268831253, "step": 890 }, { "epoch": 4.556962025316456, "grad_norm": 1729683.010514938, "learning_rate": 4.1460357254779067e-07, "logits/chosen": -7.10635232925415, "logits/rejected": -7.227837562561035, "logps/chosen": -184.3021240234375, "logps/rejected": -391.59930419921875, "loss": 67231.6313, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.050502438098192215, "rewards/margins": 0.20674797892570496, "rewards/rejected": -0.15624557435512543, "step": 900 }, { "epoch": 4.6075949367088604, "grad_norm": 1921064.671845176, "learning_rate": 4.13036665622062e-07, "logits/chosen": -7.409733772277832, "logits/rejected": -7.2668256759643555, "logps/chosen": -184.89645385742188, "logps/rejected": -395.2364501953125, "loss": 67370.1875, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.047733135521411896, "rewards/margins": 0.2108074128627777, "rewards/rejected": -0.1630742847919464, "step": 910 }, { "epoch": 4.658227848101266, "grad_norm": 1780170.6356310213, "learning_rate": 4.1146975869633344e-07, "logits/chosen": -8.294339179992676, "logits/rejected": -8.312765121459961, "logps/chosen": -185.74949645996094, "logps/rejected": -405.0606689453125, "loss": 64484.2438, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 0.05801473185420036, "rewards/margins": 0.21365991234779358, "rewards/rejected": -0.15564517676830292, "step": 920 }, { "epoch": 4.708860759493671, "grad_norm": 1755118.627079852, "learning_rate": 4.099028517706048e-07, "logits/chosen": -8.692441940307617, "logits/rejected": -8.729148864746094, "logps/chosen": -177.8703155517578, "logps/rejected": -410.15179443359375, "loss": 65960.6812, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.061922211199998856, "rewards/margins": 0.2333444058895111, "rewards/rejected": -0.17142215371131897, "step": 930 }, { "epoch": 4.759493670886076, "grad_norm": 1801666.0452341542, "learning_rate": 4.083359448448762e-07, "logits/chosen": -8.838138580322266, "logits/rejected": -8.679426193237305, "logps/chosen": -160.35488891601562, "logps/rejected": -387.3427429199219, "loss": 65957.3, "rewards/accuracies": 1.0, "rewards/chosen": 0.061734091490507126, "rewards/margins": 0.2303626835346222, "rewards/rejected": -0.16862855851650238, "step": 940 }, { "epoch": 4.810126582278481, "grad_norm": 1823914.1164093877, "learning_rate": 4.0676903791914757e-07, "logits/chosen": -8.039133071899414, "logits/rejected": -8.235550880432129, "logps/chosen": -181.90818786621094, "logps/rejected": -390.46075439453125, "loss": 65100.0437, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.05453425645828247, "rewards/margins": 0.20622405409812927, "rewards/rejected": -0.1516897976398468, "step": 950 }, { "epoch": 4.860759493670886, "grad_norm": 2552504.752187401, "learning_rate": 4.05202130993419e-07, "logits/chosen": -8.228861808776855, "logits/rejected": -8.044200897216797, "logps/chosen": -175.62306213378906, "logps/rejected": -387.7801818847656, "loss": 65251.5563, "rewards/accuracies": 0.9750000238418579, "rewards/chosen": 0.05643890053033829, "rewards/margins": 0.2162017822265625, "rewards/rejected": -0.15976287424564362, "step": 960 }, { "epoch": 4.911392405063291, "grad_norm": 2112562.829549655, "learning_rate": 4.0363522406769034e-07, "logits/chosen": -8.678482055664062, "logits/rejected": -8.680012702941895, "logps/chosen": -180.9581298828125, "logps/rejected": -402.48944091796875, "loss": 65731.7188, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.05988938361406326, "rewards/margins": 0.22270476818084717, "rewards/rejected": -0.1628153920173645, "step": 970 }, { "epoch": 4.962025316455696, "grad_norm": 1800725.2761679955, "learning_rate": 4.0206831714196175e-07, "logits/chosen": -9.068916320800781, "logits/rejected": -8.908533096313477, "logps/chosen": -191.30018615722656, "logps/rejected": -433.2850036621094, "loss": 64987.5125, "rewards/accuracies": 0.987500011920929, "rewards/chosen": 0.0664498582482338, "rewards/margins": 0.24509286880493164, "rewards/rejected": -0.17864301800727844, "step": 980 } ], "logging_steps": 10, "max_steps": 3546, "num_input_tokens_seen": 0, "num_train_epochs": 18, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }