{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.9886363636363638, "eval_steps": 50.0, "global_step": 210, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.00946969696969697, "grad_norm": 29.0056470375542, "learning_rate": 4.761904761904761e-09, "logits/chosen": -1.2287073135375977, "logits/rejected": -1.226149559020996, "logps/chosen": -298.41802978515625, "logps/rejected": -125.05058288574219, "loss": 0.6931, "rewards/accuracies": 0.0, "rewards/chosen": 0.0, "rewards/margins": 0.0, "rewards/rejected": 0.0, "step": 1 }, { "epoch": 0.01893939393939394, "grad_norm": 29.598707271309717, "learning_rate": 9.523809523809522e-09, "logits/chosen": -1.1785930395126343, "logits/rejected": -1.2503385543823242, "logps/chosen": -327.542236328125, "logps/rejected": -152.739990234375, "loss": 0.6931, "rewards/accuracies": 0.0, "rewards/chosen": 0.0, "rewards/margins": 0.0, "rewards/rejected": 0.0, "step": 2 }, { "epoch": 0.028409090909090908, "grad_norm": 30.815773047267687, "learning_rate": 1.4285714285714284e-08, "logits/chosen": -1.2728688716888428, "logits/rejected": -1.2958910465240479, "logps/chosen": -162.8279266357422, "logps/rejected": -92.91324615478516, "loss": 0.6939, "rewards/accuracies": 0.46666663885116577, "rewards/chosen": 0.0023424469400197268, "rewards/margins": 0.005515547469258308, "rewards/rejected": -0.003173100296407938, "step": 3 }, { "epoch": 0.03787878787878788, "grad_norm": 30.410758788816498, "learning_rate": 1.9047619047619045e-08, "logits/chosen": -1.3308671712875366, "logits/rejected": -1.4291285276412964, "logps/chosen": -361.55377197265625, "logps/rejected": -185.06399536132812, "loss": 0.6944, "rewards/accuracies": 0.2666666805744171, "rewards/chosen": -0.0036378882359713316, "rewards/margins": -0.005622108932584524, "rewards/rejected": 0.001984220463782549, "step": 4 }, { "epoch": 0.04734848484848485, "grad_norm": 33.72650012517962, "learning_rate": 2.3809523809523807e-08, "logits/chosen": -1.2832980155944824, "logits/rejected": -1.2578494548797607, "logps/chosen": -311.66046142578125, "logps/rejected": -113.93507385253906, "loss": 0.6944, "rewards/accuracies": 0.5333333611488342, "rewards/chosen": 0.0004738619609270245, "rewards/margins": 0.000339136749971658, "rewards/rejected": 0.0001347253128187731, "step": 5 }, { "epoch": 0.056818181818181816, "grad_norm": 33.54755632608565, "learning_rate": 2.857142857142857e-08, "logits/chosen": -1.2438254356384277, "logits/rejected": -1.3154207468032837, "logps/chosen": -237.88046264648438, "logps/rejected": -144.86170959472656, "loss": 0.6927, "rewards/accuracies": 0.6000000238418579, "rewards/chosen": 0.0015382947167381644, "rewards/margins": 0.0072678267024457455, "rewards/rejected": -0.005729531869292259, "step": 6 }, { "epoch": 0.06628787878787878, "grad_norm": 32.506024543201846, "learning_rate": 3.333333333333333e-08, "logits/chosen": -1.1219017505645752, "logits/rejected": -1.3356969356536865, "logps/chosen": -377.674072265625, "logps/rejected": -208.0950164794922, "loss": 0.6946, "rewards/accuracies": 0.46666663885116577, "rewards/chosen": 0.003615258727222681, "rewards/margins": 0.004607487935572863, "rewards/rejected": -0.0009922297904267907, "step": 7 }, { "epoch": 0.07575757575757576, "grad_norm": 26.869719248200415, "learning_rate": 3.809523809523809e-08, "logits/chosen": -1.2021812200546265, "logits/rejected": -1.2994741201400757, "logps/chosen": -223.2887420654297, "logps/rejected": -210.76904296875, "loss": 0.6941, "rewards/accuracies": 0.46666663885116577, "rewards/chosen": -0.0010813308181241155, "rewards/margins": -0.0014500105753540993, "rewards/rejected": 0.00036868025199510157, "step": 8 }, { "epoch": 0.08522727272727272, "grad_norm": 28.851260723123424, "learning_rate": 4.285714285714285e-08, "logits/chosen": -1.2686998844146729, "logits/rejected": -1.2887214422225952, "logps/chosen": -414.627685546875, "logps/rejected": -157.76890563964844, "loss": 0.6925, "rewards/accuracies": 0.3333333432674408, "rewards/chosen": -0.008543035015463829, "rewards/margins": -0.008067367598414421, "rewards/rejected": -0.00047566817374899983, "step": 9 }, { "epoch": 0.0946969696969697, "grad_norm": 29.908833040399635, "learning_rate": 4.7619047619047613e-08, "logits/chosen": -1.2252384424209595, "logits/rejected": -1.4116103649139404, "logps/chosen": -379.18719482421875, "logps/rejected": -134.59957885742188, "loss": 0.6913, "rewards/accuracies": 0.5333333611488342, "rewards/chosen": 0.003142343368381262, "rewards/margins": 0.004382309503853321, "rewards/rejected": -0.0012399652041494846, "step": 10 }, { "epoch": 0.10416666666666667, "grad_norm": 34.95508468311557, "learning_rate": 5.238095238095238e-08, "logits/chosen": -1.2304970026016235, "logits/rejected": -1.4173061847686768, "logps/chosen": -274.32415771484375, "logps/rejected": -143.79139709472656, "loss": 0.6934, "rewards/accuracies": 0.5333333015441895, "rewards/chosen": 0.001248523243702948, "rewards/margins": 0.00037327370955608785, "rewards/rejected": 0.0008752490393817425, "step": 11 }, { "epoch": 0.11363636363636363, "grad_norm": 33.36383008667387, "learning_rate": 5.714285714285714e-08, "logits/chosen": -1.2315393686294556, "logits/rejected": -1.3125958442687988, "logps/chosen": -414.57220458984375, "logps/rejected": -175.58706665039062, "loss": 0.6913, "rewards/accuracies": 0.6000000238418579, "rewards/chosen": 0.009653424844145775, "rewards/margins": 0.005845812149345875, "rewards/rejected": 0.0038076122291386127, "step": 12 }, { "epoch": 0.12310606060606061, "grad_norm": 33.31390804979405, "learning_rate": 6.19047619047619e-08, "logits/chosen": -1.226459264755249, "logits/rejected": -1.333916425704956, "logps/chosen": -220.2379913330078, "logps/rejected": -109.96634674072266, "loss": 0.6914, "rewards/accuracies": 0.5333333611488342, "rewards/chosen": 0.003358738962560892, "rewards/margins": 0.0006390871712937951, "rewards/rejected": 0.002719651209190488, "step": 13 }, { "epoch": 0.13257575757575757, "grad_norm": 30.861328252470425, "learning_rate": 6.666666666666665e-08, "logits/chosen": -1.2634985446929932, "logits/rejected": -1.4346576929092407, "logps/chosen": -280.9840087890625, "logps/rejected": -140.47079467773438, "loss": 0.6921, "rewards/accuracies": 0.5333333611488342, "rewards/chosen": -0.0012296470813453197, "rewards/margins": -0.0024472051300108433, "rewards/rejected": 0.0012175582814961672, "step": 14 }, { "epoch": 0.14204545454545456, "grad_norm": 29.301187205022977, "learning_rate": 7.142857142857142e-08, "logits/chosen": -1.2319475412368774, "logits/rejected": -1.3931156396865845, "logps/chosen": -279.740234375, "logps/rejected": -63.6292724609375, "loss": 0.6916, "rewards/accuracies": 0.6666666865348816, "rewards/chosen": 0.005358916707336903, "rewards/margins": 0.00306467292830348, "rewards/rejected": 0.002294243313372135, "step": 15 }, { "epoch": 0.15151515151515152, "grad_norm": 33.40165234179044, "learning_rate": 7.619047619047618e-08, "logits/chosen": -1.2616360187530518, "logits/rejected": -1.2523891925811768, "logps/chosen": -248.74118041992188, "logps/rejected": -139.3803253173828, "loss": 0.6899, "rewards/accuracies": 0.6000000238418579, "rewards/chosen": 0.00993367563933134, "rewards/margins": 0.0073961131274700165, "rewards/rejected": 0.0025375632103532553, "step": 16 }, { "epoch": 0.16098484848484848, "grad_norm": 31.850407532445676, "learning_rate": 8.095238095238095e-08, "logits/chosen": -1.300764560699463, "logits/rejected": -1.4897656440734863, "logps/chosen": -199.72067260742188, "logps/rejected": -117.5413818359375, "loss": 0.6876, "rewards/accuracies": 0.5333333611488342, "rewards/chosen": 0.005169999785721302, "rewards/margins": 0.004817936569452286, "rewards/rejected": 0.0003520629834383726, "step": 17 }, { "epoch": 0.17045454545454544, "grad_norm": 31.778736512108093, "learning_rate": 8.57142857142857e-08, "logits/chosen": -1.2460119724273682, "logits/rejected": -1.3352010250091553, "logps/chosen": -430.80096435546875, "logps/rejected": -209.36929321289062, "loss": 0.6886, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.022602636367082596, "rewards/margins": 0.02024615928530693, "rewards/rejected": 0.002356476616114378, "step": 18 }, { "epoch": 0.17992424242424243, "grad_norm": 31.888287525403957, "learning_rate": 9.047619047619047e-08, "logits/chosen": -1.2491956949234009, "logits/rejected": -1.2977077960968018, "logps/chosen": -226.297607421875, "logps/rejected": -110.85136413574219, "loss": 0.6849, "rewards/accuracies": 0.5333333015441895, "rewards/chosen": 0.0118248350918293, "rewards/margins": 0.008896713145077229, "rewards/rejected": 0.0029281210154294968, "step": 19 }, { "epoch": 0.1893939393939394, "grad_norm": 30.890781750471852, "learning_rate": 9.523809523809523e-08, "logits/chosen": -1.2209726572036743, "logits/rejected": -1.4927175045013428, "logps/chosen": -337.1795959472656, "logps/rejected": -123.1536865234375, "loss": 0.6865, "rewards/accuracies": 0.6666666865348816, "rewards/chosen": 0.019501259550452232, "rewards/margins": 0.021095363423228264, "rewards/rejected": -0.001594100845977664, "step": 20 }, { "epoch": 0.19886363636363635, "grad_norm": 29.49005689794317, "learning_rate": 1e-07, "logits/chosen": -1.2870718240737915, "logits/rejected": -1.3624413013458252, "logps/chosen": -226.2629852294922, "logps/rejected": -94.34611511230469, "loss": 0.6852, "rewards/accuracies": 0.6666666269302368, "rewards/chosen": 0.011245525442063808, "rewards/margins": 0.0074931420385837555, "rewards/rejected": 0.0037523836363106966, "step": 21 }, { "epoch": 0.20833333333333334, "grad_norm": 35.528523903361425, "learning_rate": 9.947089947089947e-08, "logits/chosen": -1.149263858795166, "logits/rejected": -1.1778504848480225, "logps/chosen": -522.42626953125, "logps/rejected": -187.16226196289062, "loss": 0.6755, "rewards/accuracies": 1.0, "rewards/chosen": 0.05816841125488281, "rewards/margins": 0.0636506900191307, "rewards/rejected": -0.005482281092554331, "step": 22 }, { "epoch": 0.2178030303030303, "grad_norm": 30.463057051689077, "learning_rate": 9.894179894179894e-08, "logits/chosen": -1.2491704225540161, "logits/rejected": -1.4007856845855713, "logps/chosen": -221.8249053955078, "logps/rejected": -94.48908233642578, "loss": 0.6733, "rewards/accuracies": 0.6666666865348816, "rewards/chosen": 0.03221073001623154, "rewards/margins": 0.02945198491215706, "rewards/rejected": 0.0027587474323809147, "step": 23 }, { "epoch": 0.22727272727272727, "grad_norm": 32.66406919328301, "learning_rate": 9.84126984126984e-08, "logits/chosen": -1.213392734527588, "logits/rejected": -1.2315367460250854, "logps/chosen": -484.7833557128906, "logps/rejected": -179.44985961914062, "loss": 0.6688, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.07118237763643265, "rewards/margins": 0.07119101285934448, "rewards/rejected": -8.63661989569664e-06, "step": 24 }, { "epoch": 0.23674242424242425, "grad_norm": 27.949898051091324, "learning_rate": 9.788359788359788e-08, "logits/chosen": -1.2136789560317993, "logits/rejected": -1.3596720695495605, "logps/chosen": -315.8499755859375, "logps/rejected": -146.63519287109375, "loss": 0.6695, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.05525461584329605, "rewards/margins": 0.05192388966679573, "rewards/rejected": 0.003330723149701953, "step": 25 }, { "epoch": 0.24621212121212122, "grad_norm": 31.30737207115018, "learning_rate": 9.735449735449735e-08, "logits/chosen": -1.326267957687378, "logits/rejected": -1.3485438823699951, "logps/chosen": -318.2695007324219, "logps/rejected": -93.18646240234375, "loss": 0.6664, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.06859429180622101, "rewards/margins": 0.06752636283636093, "rewards/rejected": 0.0010679200058802962, "step": 26 }, { "epoch": 0.2556818181818182, "grad_norm": 27.163805337499927, "learning_rate": 9.682539682539682e-08, "logits/chosen": -1.3347467184066772, "logits/rejected": -1.39474618434906, "logps/chosen": -252.51211547851562, "logps/rejected": -130.60934448242188, "loss": 0.6685, "rewards/accuracies": 0.6666666269302368, "rewards/chosen": 0.04859574884176254, "rewards/margins": 0.04069235175848007, "rewards/rejected": 0.007903401739895344, "step": 27 }, { "epoch": 0.26515151515151514, "grad_norm": 26.747663753533093, "learning_rate": 9.629629629629629e-08, "logits/chosen": -1.1969085931777954, "logits/rejected": -1.3106380701065063, "logps/chosen": -282.17694091796875, "logps/rejected": -131.32461547851562, "loss": 0.6674, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.053814828395843506, "rewards/margins": 0.05295072868466377, "rewards/rejected": 0.0008641040185466409, "step": 28 }, { "epoch": 0.2746212121212121, "grad_norm": 28.110282404080507, "learning_rate": 9.576719576719576e-08, "logits/chosen": -1.205107569694519, "logits/rejected": -1.3407138586044312, "logps/chosen": -262.7644958496094, "logps/rejected": -134.28187561035156, "loss": 0.6658, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.05242891237139702, "rewards/margins": 0.05327294394373894, "rewards/rejected": -0.0008440311066806316, "step": 29 }, { "epoch": 0.2840909090909091, "grad_norm": 24.744190890187145, "learning_rate": 9.523809523809523e-08, "logits/chosen": -1.2627979516983032, "logits/rejected": -1.3917124271392822, "logps/chosen": -238.56619262695312, "logps/rejected": -120.55537414550781, "loss": 0.6677, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.04753679037094116, "rewards/margins": 0.04130668193101883, "rewards/rejected": 0.006230102851986885, "step": 30 }, { "epoch": 0.2935606060606061, "grad_norm": 27.968644068426777, "learning_rate": 9.470899470899471e-08, "logits/chosen": -1.3139160871505737, "logits/rejected": -1.4469630718231201, "logps/chosen": -368.27374267578125, "logps/rejected": -130.02462768554688, "loss": 0.6626, "rewards/accuracies": 0.8666666150093079, "rewards/chosen": 0.07182395458221436, "rewards/margins": 0.07256630808115005, "rewards/rejected": -0.000742347736377269, "step": 31 }, { "epoch": 0.30303030303030304, "grad_norm": 26.902166854256873, "learning_rate": 9.417989417989417e-08, "logits/chosen": -1.2979252338409424, "logits/rejected": -1.2534756660461426, "logps/chosen": -361.4090576171875, "logps/rejected": -205.6992645263672, "loss": 0.6487, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.10240572690963745, "rewards/margins": 0.09547950327396393, "rewards/rejected": 0.006926238536834717, "step": 32 }, { "epoch": 0.3125, "grad_norm": 24.41388186700755, "learning_rate": 9.365079365079365e-08, "logits/chosen": -1.2682493925094604, "logits/rejected": -1.3996570110321045, "logps/chosen": -301.144775390625, "logps/rejected": -178.94155883789062, "loss": 0.6366, "rewards/accuracies": 0.6666666865348816, "rewards/chosen": 0.1342059224843979, "rewards/margins": 0.09688643366098404, "rewards/rejected": 0.03731948137283325, "step": 33 }, { "epoch": 0.32196969696969696, "grad_norm": 23.025986498306636, "learning_rate": 9.312169312169311e-08, "logits/chosen": -1.2429295778274536, "logits/rejected": -1.3105340003967285, "logps/chosen": -333.86077880859375, "logps/rejected": -116.7296371459961, "loss": 0.6333, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.1280738264322281, "rewards/margins": 0.13231658935546875, "rewards/rejected": -0.004242760129272938, "step": 34 }, { "epoch": 0.3314393939393939, "grad_norm": 24.498334455805622, "learning_rate": 9.259259259259259e-08, "logits/chosen": -1.3265695571899414, "logits/rejected": -1.3860602378845215, "logps/chosen": -299.00750732421875, "logps/rejected": -140.71592712402344, "loss": 0.6286, "rewards/accuracies": 0.6666666269302368, "rewards/chosen": 0.11404500156641006, "rewards/margins": 0.11178859323263168, "rewards/rejected": 0.002256409265100956, "step": 35 }, { "epoch": 0.3409090909090909, "grad_norm": 24.884921186077833, "learning_rate": 9.206349206349205e-08, "logits/chosen": -1.2984504699707031, "logits/rejected": -1.3744697570800781, "logps/chosen": -285.7705078125, "logps/rejected": -122.30088806152344, "loss": 0.6168, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.15358424186706543, "rewards/margins": 0.15173228085041046, "rewards/rejected": 0.001851982669904828, "step": 36 }, { "epoch": 0.3503787878787879, "grad_norm": 23.82555156631703, "learning_rate": 9.153439153439153e-08, "logits/chosen": -1.2476222515106201, "logits/rejected": -1.3255198001861572, "logps/chosen": -366.74029541015625, "logps/rejected": -130.18682861328125, "loss": 0.6238, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.1664523035287857, "rewards/margins": 0.15406028926372528, "rewards/rejected": 0.012392010539770126, "step": 37 }, { "epoch": 0.35984848484848486, "grad_norm": 24.07477880888288, "learning_rate": 9.1005291005291e-08, "logits/chosen": -1.335256814956665, "logits/rejected": -1.4410061836242676, "logps/chosen": -219.09616088867188, "logps/rejected": -99.69376373291016, "loss": 0.6137, "rewards/accuracies": 0.7333332896232605, "rewards/chosen": 0.1590159386396408, "rewards/margins": 0.1581113636493683, "rewards/rejected": 0.0009045897168107331, "step": 38 }, { "epoch": 0.3693181818181818, "grad_norm": 23.16275958406235, "learning_rate": 9.047619047619047e-08, "logits/chosen": -1.2550255060195923, "logits/rejected": -1.3113772869110107, "logps/chosen": -370.3097839355469, "logps/rejected": -170.9385986328125, "loss": 0.6112, "rewards/accuracies": 1.0, "rewards/chosen": 0.23174059391021729, "rewards/margins": 0.23885612189769745, "rewards/rejected": -0.00711551308631897, "step": 39 }, { "epoch": 0.3787878787878788, "grad_norm": 21.543309745960354, "learning_rate": 8.994708994708994e-08, "logits/chosen": -1.3221899271011353, "logits/rejected": -1.4326074123382568, "logps/chosen": -290.75933837890625, "logps/rejected": -167.4901123046875, "loss": 0.6159, "rewards/accuracies": 0.6666666865348816, "rewards/chosen": 0.13931059837341309, "rewards/margins": 0.13773180544376373, "rewards/rejected": 0.0015788152813911438, "step": 40 }, { "epoch": 0.38825757575757575, "grad_norm": 22.418812006239037, "learning_rate": 8.941798941798941e-08, "logits/chosen": -1.2849162817001343, "logits/rejected": -1.3343697786331177, "logps/chosen": -369.89764404296875, "logps/rejected": -133.05355834960938, "loss": 0.6093, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.19583266973495483, "rewards/margins": 0.19024832546710968, "rewards/rejected": 0.005584346130490303, "step": 41 }, { "epoch": 0.3977272727272727, "grad_norm": 25.71519469395891, "learning_rate": 8.888888888888888e-08, "logits/chosen": -1.286842703819275, "logits/rejected": -1.3505761623382568, "logps/chosen": -447.53778076171875, "logps/rejected": -138.85006713867188, "loss": 0.5864, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.31389278173446655, "rewards/margins": 0.3263290822505951, "rewards/rejected": -0.01243629027158022, "step": 42 }, { "epoch": 0.4071969696969697, "grad_norm": 19.24790527585849, "learning_rate": 8.835978835978835e-08, "logits/chosen": -1.3817297220230103, "logits/rejected": -1.4520673751831055, "logps/chosen": -279.92059326171875, "logps/rejected": -121.86234283447266, "loss": 0.6187, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.15583908557891846, "rewards/margins": 0.15511645376682281, "rewards/rejected": 0.0007226344314403832, "step": 43 }, { "epoch": 0.4166666666666667, "grad_norm": 22.13389198587296, "learning_rate": 8.783068783068782e-08, "logits/chosen": -1.3588807582855225, "logits/rejected": -1.3591158390045166, "logps/chosen": -430.5394592285156, "logps/rejected": -167.34222412109375, "loss": 0.5983, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.22864654660224915, "rewards/margins": 0.19965213537216187, "rewards/rejected": 0.028994422405958176, "step": 44 }, { "epoch": 0.42613636363636365, "grad_norm": 19.709066122231608, "learning_rate": 8.730158730158729e-08, "logits/chosen": -1.319620132446289, "logits/rejected": -1.3754303455352783, "logps/chosen": -301.44891357421875, "logps/rejected": -139.29481506347656, "loss": 0.6036, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.22895975410938263, "rewards/margins": 0.213044673204422, "rewards/rejected": 0.015915077179670334, "step": 45 }, { "epoch": 0.4356060606060606, "grad_norm": 20.432654588017233, "learning_rate": 8.677248677248676e-08, "logits/chosen": -1.314937710762024, "logits/rejected": -1.3968394994735718, "logps/chosen": -240.7213897705078, "logps/rejected": -84.15922546386719, "loss": 0.6013, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.15969710052013397, "rewards/margins": 0.16978248953819275, "rewards/rejected": -0.010085375979542732, "step": 46 }, { "epoch": 0.44507575757575757, "grad_norm": 17.302724108599822, "learning_rate": 8.624338624338625e-08, "logits/chosen": -1.3449150323867798, "logits/rejected": -1.4744312763214111, "logps/chosen": -345.28387451171875, "logps/rejected": -119.1164321899414, "loss": 0.6147, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.1876756250858307, "rewards/margins": 0.18070700764656067, "rewards/rejected": 0.006968595087528229, "step": 47 }, { "epoch": 0.45454545454545453, "grad_norm": 18.038050767324112, "learning_rate": 8.57142857142857e-08, "logits/chosen": -1.3185852766036987, "logits/rejected": -1.4122602939605713, "logps/chosen": -212.75405883789062, "logps/rejected": -140.37655639648438, "loss": 0.6088, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.1278039515018463, "rewards/margins": 0.04179360717535019, "rewards/rejected": 0.08601033687591553, "step": 48 }, { "epoch": 0.4640151515151515, "grad_norm": 22.986861694117593, "learning_rate": 8.518518518518519e-08, "logits/chosen": -1.2588977813720703, "logits/rejected": -1.4479458332061768, "logps/chosen": -338.35858154296875, "logps/rejected": -192.6273956298828, "loss": 0.5795, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.22073189914226532, "rewards/margins": 0.24155020713806152, "rewards/rejected": -0.02081829309463501, "step": 49 }, { "epoch": 0.4734848484848485, "grad_norm": 19.82460905470704, "learning_rate": 8.465608465608464e-08, "logits/chosen": -1.2549892663955688, "logits/rejected": -1.4687315225601196, "logps/chosen": -310.2621154785156, "logps/rejected": -137.7465057373047, "loss": 0.5936, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.2339044064283371, "rewards/margins": 0.22999358177185059, "rewards/rejected": 0.003910848870873451, "step": 50 }, { "epoch": 0.48295454545454547, "grad_norm": 19.652707095228752, "learning_rate": 8.412698412698413e-08, "logits/chosen": -1.397512674331665, "logits/rejected": -1.4295556545257568, "logps/chosen": -350.259033203125, "logps/rejected": -154.9229278564453, "loss": 0.5959, "rewards/accuracies": 0.73333340883255, "rewards/chosen": 0.2114483118057251, "rewards/margins": 0.1936836689710617, "rewards/rejected": 0.017764627933502197, "step": 51 }, { "epoch": 0.49242424242424243, "grad_norm": 18.932470131303603, "learning_rate": 8.359788359788358e-08, "logits/chosen": -1.3295793533325195, "logits/rejected": -1.4178295135498047, "logps/chosen": -228.3561248779297, "logps/rejected": -152.6548309326172, "loss": 0.5968, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.17612284421920776, "rewards/margins": 0.1884087324142456, "rewards/rejected": -0.012285866774618626, "step": 52 }, { "epoch": 0.5018939393939394, "grad_norm": 22.56023512380409, "learning_rate": 8.306878306878307e-08, "logits/chosen": -1.329941749572754, "logits/rejected": -1.437350869178772, "logps/chosen": -272.7892150878906, "logps/rejected": -130.10751342773438, "loss": 0.562, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.2890825569629669, "rewards/margins": 0.3040001392364502, "rewards/rejected": -0.014917601831257343, "step": 53 }, { "epoch": 0.5113636363636364, "grad_norm": 16.32960137441355, "learning_rate": 8.253968253968254e-08, "logits/chosen": -1.4222466945648193, "logits/rejected": -1.6047979593276978, "logps/chosen": -327.20196533203125, "logps/rejected": -94.08998107910156, "loss": 0.5619, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.3207431137561798, "rewards/margins": 0.362914502620697, "rewards/rejected": -0.0421714186668396, "step": 54 }, { "epoch": 0.5208333333333334, "grad_norm": 16.970648675586034, "learning_rate": 8.201058201058201e-08, "logits/chosen": -1.4205853939056396, "logits/rejected": -1.4251158237457275, "logps/chosen": -283.05108642578125, "logps/rejected": -105.49992370605469, "loss": 0.547, "rewards/accuracies": 0.6000000238418579, "rewards/chosen": 0.31010201573371887, "rewards/margins": 0.3413071930408478, "rewards/rejected": -0.031205186620354652, "step": 55 }, { "epoch": 0.5303030303030303, "grad_norm": 13.112771276876163, "learning_rate": 8.148148148148148e-08, "logits/chosen": -1.3618199825286865, "logits/rejected": -1.506471037864685, "logps/chosen": -345.26434326171875, "logps/rejected": -164.83676147460938, "loss": 0.5754, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.41369715332984924, "rewards/margins": 0.42479389905929565, "rewards/rejected": -0.011096751317381859, "step": 56 }, { "epoch": 0.5397727272727273, "grad_norm": 15.00178283487482, "learning_rate": 8.095238095238095e-08, "logits/chosen": -1.3916102647781372, "logits/rejected": -1.4226547479629517, "logps/chosen": -266.93389892578125, "logps/rejected": -123.35113525390625, "loss": 0.5448, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.2940821647644043, "rewards/margins": 0.32441002130508423, "rewards/rejected": -0.030327826738357544, "step": 57 }, { "epoch": 0.5492424242424242, "grad_norm": 13.894167304668269, "learning_rate": 8.042328042328042e-08, "logits/chosen": -1.4024507999420166, "logits/rejected": -1.499112844467163, "logps/chosen": -299.7988586425781, "logps/rejected": -212.5677032470703, "loss": 0.5271, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.4066668450832367, "rewards/margins": 0.4445282816886902, "rewards/rejected": -0.03786151856184006, "step": 58 }, { "epoch": 0.5587121212121212, "grad_norm": 13.686152249765941, "learning_rate": 7.989417989417989e-08, "logits/chosen": -1.4456459283828735, "logits/rejected": -1.4960581064224243, "logps/chosen": -218.0140380859375, "logps/rejected": -91.45022583007812, "loss": 0.5359, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.23951879143714905, "rewards/margins": 0.2641497552394867, "rewards/rejected": -0.02463097684085369, "step": 59 }, { "epoch": 0.5681818181818182, "grad_norm": 13.294977223020641, "learning_rate": 7.936507936507936e-08, "logits/chosen": -1.446040391921997, "logits/rejected": -1.4068689346313477, "logps/chosen": -271.784912109375, "logps/rejected": -104.23944091796875, "loss": 0.5248, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.33832794427871704, "rewards/margins": 0.3952670991420746, "rewards/rejected": -0.05693921446800232, "step": 60 }, { "epoch": 0.5776515151515151, "grad_norm": 14.213258268574368, "learning_rate": 7.883597883597883e-08, "logits/chosen": -1.525424599647522, "logits/rejected": -1.5090603828430176, "logps/chosen": -307.25115966796875, "logps/rejected": -123.07181549072266, "loss": 0.5072, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.2689701318740845, "rewards/margins": 0.3598218560218811, "rewards/rejected": -0.09085171669721603, "step": 61 }, { "epoch": 0.5871212121212122, "grad_norm": 11.942720090111948, "learning_rate": 7.83068783068783e-08, "logits/chosen": -1.341205358505249, "logits/rejected": -1.4334052801132202, "logps/chosen": -235.2104034423828, "logps/rejected": -95.51054382324219, "loss": 0.512, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.2976926863193512, "rewards/margins": 0.3563759922981262, "rewards/rejected": -0.05868334323167801, "step": 62 }, { "epoch": 0.5965909090909091, "grad_norm": 14.098400355821441, "learning_rate": 7.777777777777778e-08, "logits/chosen": -1.5624217987060547, "logits/rejected": -1.550048828125, "logps/chosen": -395.7930603027344, "logps/rejected": -119.40779113769531, "loss": 0.487, "rewards/accuracies": 1.0, "rewards/chosen": 0.40899619460105896, "rewards/margins": 0.47930893301963806, "rewards/rejected": -0.07031276077032089, "step": 63 }, { "epoch": 0.6060606060606061, "grad_norm": 11.1041564672598, "learning_rate": 7.724867724867724e-08, "logits/chosen": -1.4298704862594604, "logits/rejected": -1.354206919670105, "logps/chosen": -312.83697509765625, "logps/rejected": -110.90681457519531, "loss": 0.5224, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.32718759775161743, "rewards/margins": 0.3963702321052551, "rewards/rejected": -0.06918264925479889, "step": 64 }, { "epoch": 0.615530303030303, "grad_norm": 13.357200204611026, "learning_rate": 7.671957671957672e-08, "logits/chosen": -1.423771619796753, "logits/rejected": -1.495290756225586, "logps/chosen": -341.0536193847656, "logps/rejected": -195.5428466796875, "loss": 0.4798, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.47605428099632263, "rewards/margins": 0.5310575366020203, "rewards/rejected": -0.055003322660923004, "step": 65 }, { "epoch": 0.625, "grad_norm": 13.546486727992278, "learning_rate": 7.619047619047618e-08, "logits/chosen": -1.5028284788131714, "logits/rejected": -1.4564129114151, "logps/chosen": -285.58917236328125, "logps/rejected": -131.5936279296875, "loss": 0.4836, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.36600998044013977, "rewards/margins": 0.4425935745239258, "rewards/rejected": -0.0765836089849472, "step": 66 }, { "epoch": 0.634469696969697, "grad_norm": 12.596961700455271, "learning_rate": 7.566137566137566e-08, "logits/chosen": -1.4141170978546143, "logits/rejected": -1.4340078830718994, "logps/chosen": -261.4342041015625, "logps/rejected": -95.14923858642578, "loss": 0.4883, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.44839000701904297, "rewards/margins": 0.5747641324996948, "rewards/rejected": -0.12637418508529663, "step": 67 }, { "epoch": 0.6439393939393939, "grad_norm": 12.384595380406164, "learning_rate": 7.513227513227512e-08, "logits/chosen": -1.4482860565185547, "logits/rejected": -1.5673211812973022, "logps/chosen": -237.4990234375, "logps/rejected": -84.68649291992188, "loss": 0.5, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.37629473209381104, "rewards/margins": 0.42733898758888245, "rewards/rejected": -0.05104423686861992, "step": 68 }, { "epoch": 0.6534090909090909, "grad_norm": 11.573197865338908, "learning_rate": 7.46031746031746e-08, "logits/chosen": -1.4828598499298096, "logits/rejected": -1.5483115911483765, "logps/chosen": -294.4015197753906, "logps/rejected": -105.67552185058594, "loss": 0.4914, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.42700010538101196, "rewards/margins": 0.5032913684844971, "rewards/rejected": -0.07629130035638809, "step": 69 }, { "epoch": 0.6628787878787878, "grad_norm": 11.526501769285554, "learning_rate": 7.407407407407407e-08, "logits/chosen": -1.3778324127197266, "logits/rejected": -1.3970375061035156, "logps/chosen": -356.0700988769531, "logps/rejected": -174.7995147705078, "loss": 0.4855, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.5114725232124329, "rewards/margins": 0.5073350071907043, "rewards/rejected": 0.004137503914535046, "step": 70 }, { "epoch": 0.6723484848484849, "grad_norm": 11.786777529592367, "learning_rate": 7.354497354497354e-08, "logits/chosen": -1.4470304250717163, "logits/rejected": -1.5260601043701172, "logps/chosen": -257.9070739746094, "logps/rejected": -82.23036193847656, "loss": 0.5138, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.3575395941734314, "rewards/margins": 0.4714414179325104, "rewards/rejected": -0.11390187591314316, "step": 71 }, { "epoch": 0.6818181818181818, "grad_norm": 13.911478948655224, "learning_rate": 7.301587301587301e-08, "logits/chosen": -1.4967644214630127, "logits/rejected": -1.379651427268982, "logps/chosen": -456.3885803222656, "logps/rejected": -181.92654418945312, "loss": 0.4686, "rewards/accuracies": 1.0, "rewards/chosen": 0.6715552806854248, "rewards/margins": 0.8583688735961914, "rewards/rejected": -0.18681350350379944, "step": 72 }, { "epoch": 0.6912878787878788, "grad_norm": 11.631616954008805, "learning_rate": 7.248677248677248e-08, "logits/chosen": -1.4026285409927368, "logits/rejected": -1.4543156623840332, "logps/chosen": -337.62078857421875, "logps/rejected": -139.98934936523438, "loss": 0.5001, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.5019295811653137, "rewards/margins": 0.5479447245597839, "rewards/rejected": -0.04601515084505081, "step": 73 }, { "epoch": 0.7007575757575758, "grad_norm": 9.936203845833878, "learning_rate": 7.195767195767195e-08, "logits/chosen": -1.5050134658813477, "logits/rejected": -1.5326780080795288, "logps/chosen": -387.0226135253906, "logps/rejected": -181.61231994628906, "loss": 0.5285, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.47982874512672424, "rewards/margins": 0.5457210540771484, "rewards/rejected": -0.0658923089504242, "step": 74 }, { "epoch": 0.7102272727272727, "grad_norm": 12.184941282416535, "learning_rate": 7.142857142857142e-08, "logits/chosen": -1.4602888822555542, "logits/rejected": -1.4935940504074097, "logps/chosen": -471.65643310546875, "logps/rejected": -181.0192108154297, "loss": 0.4831, "rewards/accuracies": 1.0, "rewards/chosen": 0.696118175983429, "rewards/margins": 0.8435744047164917, "rewards/rejected": -0.1474563181400299, "step": 75 }, { "epoch": 0.7196969696969697, "grad_norm": 11.235163839695879, "learning_rate": 7.08994708994709e-08, "logits/chosen": -1.5082679986953735, "logits/rejected": -1.5974117517471313, "logps/chosen": -399.4970703125, "logps/rejected": -201.58975219726562, "loss": 0.4768, "rewards/accuracies": 1.0, "rewards/chosen": 0.6427886486053467, "rewards/margins": 0.7292692065238953, "rewards/rejected": -0.08648059517145157, "step": 76 }, { "epoch": 0.7291666666666666, "grad_norm": 12.9574754516344, "learning_rate": 7.037037037037036e-08, "logits/chosen": -1.4920450448989868, "logits/rejected": -1.4171216487884521, "logps/chosen": -332.895263671875, "logps/rejected": -176.00601196289062, "loss": 0.4783, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.4683101773262024, "rewards/margins": 0.5503407716751099, "rewards/rejected": -0.08203061670064926, "step": 77 }, { "epoch": 0.7386363636363636, "grad_norm": 11.558597299124692, "learning_rate": 6.984126984126983e-08, "logits/chosen": -1.4621433019638062, "logits/rejected": -1.5110585689544678, "logps/chosen": -361.80889892578125, "logps/rejected": -157.5341339111328, "loss": 0.4715, "rewards/accuracies": 1.0, "rewards/chosen": 0.555467963218689, "rewards/margins": 0.682567834854126, "rewards/rejected": -0.12709984183311462, "step": 78 }, { "epoch": 0.7481060606060606, "grad_norm": 13.548458950264191, "learning_rate": 6.931216931216932e-08, "logits/chosen": -1.523284673690796, "logits/rejected": -1.4904415607452393, "logps/chosen": -309.17816162109375, "logps/rejected": -115.497802734375, "loss": 0.493, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.4258233904838562, "rewards/margins": 0.4799923896789551, "rewards/rejected": -0.05416899174451828, "step": 79 }, { "epoch": 0.7575757575757576, "grad_norm": 10.624056172874939, "learning_rate": 6.878306878306877e-08, "logits/chosen": -1.4721736907958984, "logits/rejected": -1.556168794631958, "logps/chosen": -349.3233337402344, "logps/rejected": -119.07454681396484, "loss": 0.4971, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4824630320072174, "rewards/margins": 0.6270266771316528, "rewards/rejected": -0.1445636749267578, "step": 80 }, { "epoch": 0.7670454545454546, "grad_norm": 10.877663952011163, "learning_rate": 6.825396825396826e-08, "logits/chosen": -1.448540210723877, "logits/rejected": -1.5003968477249146, "logps/chosen": -341.19549560546875, "logps/rejected": -161.4232635498047, "loss": 0.4997, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.512890636920929, "rewards/margins": 0.5583428144454956, "rewards/rejected": -0.04545217379927635, "step": 81 }, { "epoch": 0.7765151515151515, "grad_norm": 13.655792524642862, "learning_rate": 6.772486772486771e-08, "logits/chosen": -1.4932104349136353, "logits/rejected": -1.4824309349060059, "logps/chosen": -377.6290588378906, "logps/rejected": -145.4966583251953, "loss": 0.4506, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5544673204421997, "rewards/margins": 0.7561152577400208, "rewards/rejected": -0.20164790749549866, "step": 82 }, { "epoch": 0.7859848484848485, "grad_norm": 13.094215069002113, "learning_rate": 6.71957671957672e-08, "logits/chosen": -1.5782554149627686, "logits/rejected": -1.660269021987915, "logps/chosen": -349.7228088378906, "logps/rejected": -182.11880493164062, "loss": 0.4818, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.6241387128829956, "rewards/margins": 0.7656179666519165, "rewards/rejected": -0.1414792239665985, "step": 83 }, { "epoch": 0.7954545454545454, "grad_norm": 12.779977757002404, "learning_rate": 6.666666666666665e-08, "logits/chosen": -1.5646089315414429, "logits/rejected": -1.5739343166351318, "logps/chosen": -311.35760498046875, "logps/rejected": -131.74862670898438, "loss": 0.4655, "rewards/accuracies": 1.0, "rewards/chosen": 0.5058180093765259, "rewards/margins": 0.6643168926239014, "rewards/rejected": -0.1584988683462143, "step": 84 }, { "epoch": 0.8049242424242424, "grad_norm": 10.185073249389749, "learning_rate": 6.613756613756614e-08, "logits/chosen": -1.4740874767303467, "logits/rejected": -1.4708284139633179, "logps/chosen": -319.00286865234375, "logps/rejected": -152.86892700195312, "loss": 0.4822, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5287100076675415, "rewards/margins": 0.6535229682922363, "rewards/rejected": -0.1248130053281784, "step": 85 }, { "epoch": 0.8143939393939394, "grad_norm": 11.203293253753769, "learning_rate": 6.560846560846561e-08, "logits/chosen": -1.5367662906646729, "logits/rejected": -1.4478954076766968, "logps/chosen": -201.14149475097656, "logps/rejected": -86.86661529541016, "loss": 0.4671, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.37096017599105835, "rewards/margins": 0.45440879464149475, "rewards/rejected": -0.0834486186504364, "step": 86 }, { "epoch": 0.8238636363636364, "grad_norm": 10.604116644121472, "learning_rate": 6.507936507936508e-08, "logits/chosen": -1.4273946285247803, "logits/rejected": -1.4703240394592285, "logps/chosen": -301.97882080078125, "logps/rejected": -157.4232635498047, "loss": 0.4514, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.624559223651886, "rewards/margins": 0.7758625745773315, "rewards/rejected": -0.15130344033241272, "step": 87 }, { "epoch": 0.8333333333333334, "grad_norm": 9.972392236357932, "learning_rate": 6.455026455026455e-08, "logits/chosen": -1.482578158378601, "logits/rejected": -1.4413843154907227, "logps/chosen": -254.91702270507812, "logps/rejected": -123.97944641113281, "loss": 0.4721, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.48685789108276367, "rewards/margins": 0.6410495638847351, "rewards/rejected": -0.1541917622089386, "step": 88 }, { "epoch": 0.8428030303030303, "grad_norm": 10.523325958099775, "learning_rate": 6.402116402116402e-08, "logits/chosen": -1.480163335800171, "logits/rejected": -1.5355768203735352, "logps/chosen": -304.09197998046875, "logps/rejected": -142.2599639892578, "loss": 0.4744, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.5564759969711304, "rewards/margins": 0.6676918864250183, "rewards/rejected": -0.11121580749750137, "step": 89 }, { "epoch": 0.8522727272727273, "grad_norm": 9.519106187436845, "learning_rate": 6.349206349206349e-08, "logits/chosen": -1.5268163681030273, "logits/rejected": -1.5646402835845947, "logps/chosen": -286.10540771484375, "logps/rejected": -157.7232208251953, "loss": 0.4714, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.4824487268924713, "rewards/margins": 0.5735751986503601, "rewards/rejected": -0.09112647920846939, "step": 90 }, { "epoch": 0.8617424242424242, "grad_norm": 9.51590024478788, "learning_rate": 6.296296296296296e-08, "logits/chosen": -1.5810682773590088, "logits/rejected": -1.561959981918335, "logps/chosen": -292.758056640625, "logps/rejected": -114.890625, "loss": 0.5022, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.47544798254966736, "rewards/margins": 0.5433242321014404, "rewards/rejected": -0.06787623465061188, "step": 91 }, { "epoch": 0.8712121212121212, "grad_norm": 9.702200423985945, "learning_rate": 6.243386243386243e-08, "logits/chosen": -1.4159653186798096, "logits/rejected": -1.5273025035858154, "logps/chosen": -310.64093017578125, "logps/rejected": -160.01473999023438, "loss": 0.4457, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.5551848411560059, "rewards/margins": 0.6525019407272339, "rewards/rejected": -0.097317174077034, "step": 92 }, { "epoch": 0.8806818181818182, "grad_norm": 12.532628428602317, "learning_rate": 6.19047619047619e-08, "logits/chosen": -1.4812260866165161, "logits/rejected": -1.5698903799057007, "logps/chosen": -313.69659423828125, "logps/rejected": -125.74027252197266, "loss": 0.4221, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.512529730796814, "rewards/margins": 0.7795059084892273, "rewards/rejected": -0.26697611808776855, "step": 93 }, { "epoch": 0.8901515151515151, "grad_norm": 9.937234934609606, "learning_rate": 6.137566137566137e-08, "logits/chosen": -1.4872493743896484, "logits/rejected": -1.4849263429641724, "logps/chosen": -306.2376708984375, "logps/rejected": -151.22828674316406, "loss": 0.4843, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.49157848954200745, "rewards/margins": 0.6916278004646301, "rewards/rejected": -0.2000492364168167, "step": 94 }, { "epoch": 0.8996212121212122, "grad_norm": 10.966223203578922, "learning_rate": 6.084656084656084e-08, "logits/chosen": -1.5097806453704834, "logits/rejected": -1.548575520515442, "logps/chosen": -248.2743682861328, "logps/rejected": -143.56008911132812, "loss": 0.4619, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.29616254568099976, "rewards/margins": 0.44729262590408325, "rewards/rejected": -0.1511300802230835, "step": 95 }, { "epoch": 0.9090909090909091, "grad_norm": 10.290146618878584, "learning_rate": 6.031746031746031e-08, "logits/chosen": -1.3948135375976562, "logits/rejected": -1.4724781513214111, "logps/chosen": -200.97560119628906, "logps/rejected": -59.250823974609375, "loss": 0.4559, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.31829187273979187, "rewards/margins": 0.4826883375644684, "rewards/rejected": -0.1643964797258377, "step": 96 }, { "epoch": 0.9185606060606061, "grad_norm": 10.023113393783985, "learning_rate": 5.978835978835979e-08, "logits/chosen": -1.4924052953720093, "logits/rejected": -1.481069803237915, "logps/chosen": -210.6460418701172, "logps/rejected": -101.57307434082031, "loss": 0.4809, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.3884206712245941, "rewards/margins": 0.5861561298370361, "rewards/rejected": -0.19773544371128082, "step": 97 }, { "epoch": 0.928030303030303, "grad_norm": 9.340197405706663, "learning_rate": 5.925925925925925e-08, "logits/chosen": -1.5180368423461914, "logits/rejected": -1.4653794765472412, "logps/chosen": -321.68426513671875, "logps/rejected": -144.31082153320312, "loss": 0.4381, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.4283772110939026, "rewards/margins": 0.8070729970932007, "rewards/rejected": -0.3786957859992981, "step": 98 }, { "epoch": 0.9375, "grad_norm": 9.85547123624337, "learning_rate": 5.873015873015873e-08, "logits/chosen": -1.5044419765472412, "logits/rejected": -1.5305202007293701, "logps/chosen": -289.38946533203125, "logps/rejected": -109.14179992675781, "loss": 0.4682, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.4926362931728363, "rewards/margins": 0.6987640261650085, "rewards/rejected": -0.20612768828868866, "step": 99 }, { "epoch": 0.946969696969697, "grad_norm": 10.445428542153785, "learning_rate": 5.82010582010582e-08, "logits/chosen": -1.4615473747253418, "logits/rejected": -1.507441520690918, "logps/chosen": -233.7719268798828, "logps/rejected": -87.18647766113281, "loss": 0.4687, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.3351677358150482, "rewards/margins": 0.5392307043075562, "rewards/rejected": -0.20406293869018555, "step": 100 }, { "epoch": 0.9564393939393939, "grad_norm": 10.068872305516752, "learning_rate": 5.7671957671957674e-08, "logits/chosen": -1.4435585737228394, "logits/rejected": -1.592414379119873, "logps/chosen": -288.0964660644531, "logps/rejected": -245.06857299804688, "loss": 0.4671, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.519100546836853, "rewards/margins": 0.7276883721351624, "rewards/rejected": -0.20858784019947052, "step": 101 }, { "epoch": 0.9659090909090909, "grad_norm": 9.94558673339714, "learning_rate": 5.714285714285714e-08, "logits/chosen": -1.5683315992355347, "logits/rejected": -1.5737329721450806, "logps/chosen": -380.98223876953125, "logps/rejected": -148.6276092529297, "loss": 0.4173, "rewards/accuracies": 1.0, "rewards/chosen": 0.6376638412475586, "rewards/margins": 0.9165285229682922, "rewards/rejected": -0.2788645625114441, "step": 102 }, { "epoch": 0.9753787878787878, "grad_norm": 9.63450569739456, "learning_rate": 5.6613756613756614e-08, "logits/chosen": -1.5396722555160522, "logits/rejected": -1.5394847393035889, "logps/chosen": -313.8475341796875, "logps/rejected": -103.21199035644531, "loss": 0.4177, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6744638681411743, "rewards/margins": 0.8535541296005249, "rewards/rejected": -0.17909026145935059, "step": 103 }, { "epoch": 0.9848484848484849, "grad_norm": 10.940866953801127, "learning_rate": 5.608465608465608e-08, "logits/chosen": -1.5241594314575195, "logits/rejected": -1.637131690979004, "logps/chosen": -340.30072021484375, "logps/rejected": -126.37461853027344, "loss": 0.4387, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5279501080513, "rewards/margins": 0.7279192805290222, "rewards/rejected": -0.1999691277742386, "step": 104 }, { "epoch": 0.9943181818181818, "grad_norm": 9.396057853660293, "learning_rate": 5.5555555555555555e-08, "logits/chosen": -1.4265496730804443, "logits/rejected": -1.5132142305374146, "logps/chosen": -269.0768127441406, "logps/rejected": -128.25860595703125, "loss": 0.4426, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.511640727519989, "rewards/margins": 0.7144440412521362, "rewards/rejected": -0.20280325412750244, "step": 105 }, { "epoch": 1.003787878787879, "grad_norm": 7.831685341832891, "learning_rate": 5.502645502645502e-08, "logits/chosen": -1.5368983745574951, "logits/rejected": -1.4743095636367798, "logps/chosen": -280.42279052734375, "logps/rejected": -121.66993713378906, "loss": 0.4581, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.5917797684669495, "rewards/margins": 0.7320090532302856, "rewards/rejected": -0.14022931456565857, "step": 106 }, { "epoch": 1.0132575757575757, "grad_norm": 8.940391987752884, "learning_rate": 5.4497354497354495e-08, "logits/chosen": -1.491202712059021, "logits/rejected": -1.5914320945739746, "logps/chosen": -321.1029052734375, "logps/rejected": -147.7200164794922, "loss": 0.4276, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.5120298862457275, "rewards/margins": 0.6171775460243225, "rewards/rejected": -0.10514764487743378, "step": 107 }, { "epoch": 1.0227272727272727, "grad_norm": 7.697241759390277, "learning_rate": 5.3968253968253965e-08, "logits/chosen": -1.5066335201263428, "logits/rejected": -1.528194546699524, "logps/chosen": -292.2257385253906, "logps/rejected": -174.4923553466797, "loss": 0.4566, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5518472790718079, "rewards/margins": 0.8220493197441101, "rewards/rejected": -0.270202100276947, "step": 108 }, { "epoch": 1.0321969696969697, "grad_norm": 7.9963617328841226, "learning_rate": 5.343915343915344e-08, "logits/chosen": -1.5413943529129028, "logits/rejected": -1.6188589334487915, "logps/chosen": -312.98065185546875, "logps/rejected": -115.0594253540039, "loss": 0.4259, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.5731326341629028, "rewards/margins": 0.9009225964546204, "rewards/rejected": -0.32778996229171753, "step": 109 }, { "epoch": 1.0416666666666667, "grad_norm": 9.995827902761533, "learning_rate": 5.2910052910052905e-08, "logits/chosen": -1.6093279123306274, "logits/rejected": -1.713158369064331, "logps/chosen": -373.02288818359375, "logps/rejected": -202.49237060546875, "loss": 0.4242, "rewards/accuracies": 1.0, "rewards/chosen": 0.4924096167087555, "rewards/margins": 0.7550744414329529, "rewards/rejected": -0.2626648545265198, "step": 110 }, { "epoch": 1.0511363636363635, "grad_norm": 6.7346694181058, "learning_rate": 5.238095238095238e-08, "logits/chosen": -1.571112871170044, "logits/rejected": -1.5796878337860107, "logps/chosen": -312.8564453125, "logps/rejected": -136.47171020507812, "loss": 0.4814, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5543137788772583, "rewards/margins": 0.8626272082328796, "rewards/rejected": -0.30831339955329895, "step": 111 }, { "epoch": 1.0606060606060606, "grad_norm": 6.9001053561660015, "learning_rate": 5.1851851851851846e-08, "logits/chosen": -1.5522701740264893, "logits/rejected": -1.6721560955047607, "logps/chosen": -323.85064697265625, "logps/rejected": -118.36842346191406, "loss": 0.4539, "rewards/accuracies": 1.0, "rewards/chosen": 0.6884499192237854, "rewards/margins": 0.9783046841621399, "rewards/rejected": -0.28985467553138733, "step": 112 }, { "epoch": 1.0700757575757576, "grad_norm": 8.261082045277393, "learning_rate": 5.132275132275132e-08, "logits/chosen": -1.5419814586639404, "logits/rejected": -1.6027164459228516, "logps/chosen": -262.14154052734375, "logps/rejected": -85.59474182128906, "loss": 0.42, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4974077641963959, "rewards/margins": 0.7110608220100403, "rewards/rejected": -0.21365304291248322, "step": 113 }, { "epoch": 1.0795454545454546, "grad_norm": 7.731297247151736, "learning_rate": 5.0793650793650786e-08, "logits/chosen": -1.5612366199493408, "logits/rejected": -1.6464793682098389, "logps/chosen": -212.521484375, "logps/rejected": -63.86298751831055, "loss": 0.408, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.43088382482528687, "rewards/margins": 0.6654319167137146, "rewards/rejected": -0.23454804718494415, "step": 114 }, { "epoch": 1.0890151515151516, "grad_norm": 8.268687229033572, "learning_rate": 5.026455026455026e-08, "logits/chosen": -1.4982279539108276, "logits/rejected": -1.5205714702606201, "logps/chosen": -284.63787841796875, "logps/rejected": -165.73988342285156, "loss": 0.4353, "rewards/accuracies": 1.0, "rewards/chosen": 0.5018287301063538, "rewards/margins": 0.7900735139846802, "rewards/rejected": -0.2882448136806488, "step": 115 }, { "epoch": 1.0984848484848484, "grad_norm": 7.675182494137689, "learning_rate": 4.973544973544973e-08, "logits/chosen": -1.3966763019561768, "logits/rejected": -1.5771713256835938, "logps/chosen": -306.68499755859375, "logps/rejected": -174.55003356933594, "loss": 0.3851, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.6541851162910461, "rewards/margins": 0.9976884722709656, "rewards/rejected": -0.34350335597991943, "step": 116 }, { "epoch": 1.1079545454545454, "grad_norm": 7.288879950855222, "learning_rate": 4.92063492063492e-08, "logits/chosen": -1.6342532634735107, "logits/rejected": -1.7281395196914673, "logps/chosen": -312.36895751953125, "logps/rejected": -198.22055053710938, "loss": 0.4289, "rewards/accuracies": 1.0, "rewards/chosen": 0.6851889491081238, "rewards/margins": 0.873565673828125, "rewards/rejected": -0.18837669491767883, "step": 117 }, { "epoch": 1.1174242424242424, "grad_norm": 7.471188637782996, "learning_rate": 4.867724867724867e-08, "logits/chosen": -1.5435848236083984, "logits/rejected": -1.5624885559082031, "logps/chosen": -276.4817810058594, "logps/rejected": -140.15501403808594, "loss": 0.4277, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4910324513912201, "rewards/margins": 0.6622015237808228, "rewards/rejected": -0.17116901278495789, "step": 118 }, { "epoch": 1.1268939393939394, "grad_norm": 6.073468496072038, "learning_rate": 4.814814814814814e-08, "logits/chosen": -1.549534559249878, "logits/rejected": -1.7240833044052124, "logps/chosen": -237.54714965820312, "logps/rejected": -110.4820327758789, "loss": 0.4246, "rewards/accuracies": 1.0, "rewards/chosen": 0.43428367376327515, "rewards/margins": 0.675010085105896, "rewards/rejected": -0.2407263219356537, "step": 119 }, { "epoch": 1.1363636363636362, "grad_norm": 6.229810704784674, "learning_rate": 4.7619047619047613e-08, "logits/chosen": -1.5288211107254028, "logits/rejected": -1.6095609664916992, "logps/chosen": -300.1777038574219, "logps/rejected": -178.78512573242188, "loss": 0.4332, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.45809292793273926, "rewards/margins": 0.775213360786438, "rewards/rejected": -0.31712037324905396, "step": 120 }, { "epoch": 1.1458333333333333, "grad_norm": 6.153778206401315, "learning_rate": 4.7089947089947084e-08, "logits/chosen": -1.446118950843811, "logits/rejected": -1.6403175592422485, "logps/chosen": -261.62823486328125, "logps/rejected": -110.68792724609375, "loss": 0.4185, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5491897463798523, "rewards/margins": 0.9256091117858887, "rewards/rejected": -0.3764193654060364, "step": 121 }, { "epoch": 1.1553030303030303, "grad_norm": 8.460733071100409, "learning_rate": 4.6560846560846554e-08, "logits/chosen": -1.5496113300323486, "logits/rejected": -1.6899559497833252, "logps/chosen": -339.3661193847656, "logps/rejected": -121.66202545166016, "loss": 0.3873, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.7354398369789124, "rewards/margins": 1.1846985816955566, "rewards/rejected": -0.44925880432128906, "step": 122 }, { "epoch": 1.1647727272727273, "grad_norm": 7.315349675684202, "learning_rate": 4.6031746031746024e-08, "logits/chosen": -1.6141818761825562, "logits/rejected": -1.6296923160552979, "logps/chosen": -238.61257934570312, "logps/rejected": -122.18446350097656, "loss": 0.3669, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.37908780574798584, "rewards/margins": 0.6342587471008301, "rewards/rejected": -0.2551709711551666, "step": 123 }, { "epoch": 1.1742424242424243, "grad_norm": 7.280988098748099, "learning_rate": 4.55026455026455e-08, "logits/chosen": -1.5299112796783447, "logits/rejected": -1.5701675415039062, "logps/chosen": -316.45538330078125, "logps/rejected": -162.38204956054688, "loss": 0.3653, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.628962516784668, "rewards/margins": 1.195112943649292, "rewards/rejected": -0.566150426864624, "step": 124 }, { "epoch": 1.183712121212121, "grad_norm": 6.336179594301882, "learning_rate": 4.497354497354497e-08, "logits/chosen": -1.557908058166504, "logits/rejected": -1.6704943180084229, "logps/chosen": -325.92266845703125, "logps/rejected": -167.68130493164062, "loss": 0.3747, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.6076933741569519, "rewards/margins": 0.9305548667907715, "rewards/rejected": -0.3228614330291748, "step": 125 }, { "epoch": 1.1931818181818181, "grad_norm": 6.805130421455224, "learning_rate": 4.444444444444444e-08, "logits/chosen": -1.4893940687179565, "logits/rejected": -1.5416975021362305, "logps/chosen": -164.71548461914062, "logps/rejected": -102.4400634765625, "loss": 0.3971, "rewards/accuracies": 1.0, "rewards/chosen": 0.2525160014629364, "rewards/margins": 0.4948664605617523, "rewards/rejected": -0.2423505038022995, "step": 126 }, { "epoch": 1.2026515151515151, "grad_norm": 6.005160247149641, "learning_rate": 4.391534391534391e-08, "logits/chosen": -1.6034702062606812, "logits/rejected": -1.7156803607940674, "logps/chosen": -327.7132568359375, "logps/rejected": -122.96253967285156, "loss": 0.4048, "rewards/accuracies": 1.0, "rewards/chosen": 0.611197292804718, "rewards/margins": 0.990843653678894, "rewards/rejected": -0.3796464204788208, "step": 127 }, { "epoch": 1.2121212121212122, "grad_norm": 6.064066327086945, "learning_rate": 4.338624338624338e-08, "logits/chosen": -1.4845657348632812, "logits/rejected": -1.6025197505950928, "logps/chosen": -266.589599609375, "logps/rejected": -185.93374633789062, "loss": 0.3821, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.46795588731765747, "rewards/margins": 0.7385295033454895, "rewards/rejected": -0.27057361602783203, "step": 128 }, { "epoch": 1.2215909090909092, "grad_norm": 7.118625992094876, "learning_rate": 4.285714285714285e-08, "logits/chosen": -1.5417770147323608, "logits/rejected": -1.6401869058609009, "logps/chosen": -364.9320983886719, "logps/rejected": -157.976806640625, "loss": 0.3608, "rewards/accuracies": 1.0, "rewards/chosen": 0.7918987274169922, "rewards/margins": 1.300087571144104, "rewards/rejected": -0.5081888437271118, "step": 129 }, { "epoch": 1.231060606060606, "grad_norm": 5.890825166861846, "learning_rate": 4.232804232804232e-08, "logits/chosen": -1.6265586614608765, "logits/rejected": -1.7117183208465576, "logps/chosen": -360.13519287109375, "logps/rejected": -177.2240753173828, "loss": 0.4508, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.7431641221046448, "rewards/margins": 1.2213959693908691, "rewards/rejected": -0.47823190689086914, "step": 130 }, { "epoch": 1.240530303030303, "grad_norm": 6.0694144195552315, "learning_rate": 4.179894179894179e-08, "logits/chosen": -1.5653178691864014, "logits/rejected": -1.6441690921783447, "logps/chosen": -263.4643859863281, "logps/rejected": -122.57698059082031, "loss": 0.4199, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.47927695512771606, "rewards/margins": 0.9024354219436646, "rewards/rejected": -0.4231584668159485, "step": 131 }, { "epoch": 1.25, "grad_norm": 5.640397757089755, "learning_rate": 4.126984126984127e-08, "logits/chosen": -1.5248239040374756, "logits/rejected": -1.5807774066925049, "logps/chosen": -232.2427520751953, "logps/rejected": -182.7039794921875, "loss": 0.4312, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.33463865518569946, "rewards/margins": 0.5401297807693481, "rewards/rejected": -0.20549115538597107, "step": 132 }, { "epoch": 1.259469696969697, "grad_norm": 6.188835951236042, "learning_rate": 4.074074074074074e-08, "logits/chosen": -1.5644571781158447, "logits/rejected": -1.5982444286346436, "logps/chosen": -316.8279724121094, "logps/rejected": -114.7803726196289, "loss": 0.3815, "rewards/accuracies": 1.0, "rewards/chosen": 0.6954594254493713, "rewards/margins": 1.1970874071121216, "rewards/rejected": -0.5016279816627502, "step": 133 }, { "epoch": 1.268939393939394, "grad_norm": 5.913640581542438, "learning_rate": 4.021164021164021e-08, "logits/chosen": -1.5394070148468018, "logits/rejected": -1.6787744760513306, "logps/chosen": -210.8607940673828, "logps/rejected": -104.78348541259766, "loss": 0.4341, "rewards/accuracies": 1.0, "rewards/chosen": 0.422776460647583, "rewards/margins": 0.6558708548545837, "rewards/rejected": -0.2330944538116455, "step": 134 }, { "epoch": 1.2784090909090908, "grad_norm": 5.979621420732666, "learning_rate": 3.968253968253968e-08, "logits/chosen": -1.5477142333984375, "logits/rejected": -1.659967064857483, "logps/chosen": -297.19622802734375, "logps/rejected": -153.1532745361328, "loss": 0.3554, "rewards/accuracies": 1.0, "rewards/chosen": 0.6835674047470093, "rewards/margins": 1.379019021987915, "rewards/rejected": -0.695451557636261, "step": 135 }, { "epoch": 1.2878787878787878, "grad_norm": 5.932788102313539, "learning_rate": 3.915343915343915e-08, "logits/chosen": -1.4924932718276978, "logits/rejected": -1.6393858194351196, "logps/chosen": -246.09750366210938, "logps/rejected": -120.7838363647461, "loss": 0.3555, "rewards/accuracies": 1.0, "rewards/chosen": 0.5313190221786499, "rewards/margins": 0.9934673309326172, "rewards/rejected": -0.4621482789516449, "step": 136 }, { "epoch": 1.2973484848484849, "grad_norm": 14.930220486136447, "learning_rate": 3.862433862433862e-08, "logits/chosen": -1.5214557647705078, "logits/rejected": -1.6087700128555298, "logps/chosen": -384.80938720703125, "logps/rejected": -210.44235229492188, "loss": 0.3889, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.7282799482345581, "rewards/margins": 1.31002676486969, "rewards/rejected": -0.5817468762397766, "step": 137 }, { "epoch": 1.3068181818181819, "grad_norm": 5.971012894528325, "learning_rate": 3.809523809523809e-08, "logits/chosen": -1.5881704092025757, "logits/rejected": -1.569000005722046, "logps/chosen": -297.71490478515625, "logps/rejected": -158.21453857421875, "loss": 0.3827, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.487607479095459, "rewards/margins": 0.8952295184135437, "rewards/rejected": -0.40762215852737427, "step": 138 }, { "epoch": 1.316287878787879, "grad_norm": 5.470550316559649, "learning_rate": 3.756613756613756e-08, "logits/chosen": -1.5592622756958008, "logits/rejected": -1.6217191219329834, "logps/chosen": -208.8889923095703, "logps/rejected": -99.66307067871094, "loss": 0.3742, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.47726184129714966, "rewards/margins": 0.8086546063423157, "rewards/rejected": -0.3313927948474884, "step": 139 }, { "epoch": 1.3257575757575757, "grad_norm": 5.53149417210968, "learning_rate": 3.7037037037037036e-08, "logits/chosen": -1.5759552717208862, "logits/rejected": -1.6656872034072876, "logps/chosen": -282.05584716796875, "logps/rejected": -137.61387634277344, "loss": 0.3673, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5351212024688721, "rewards/margins": 0.9544070363044739, "rewards/rejected": -0.41928577423095703, "step": 140 }, { "epoch": 1.3352272727272727, "grad_norm": 5.589581006217635, "learning_rate": 3.6507936507936506e-08, "logits/chosen": -1.6098802089691162, "logits/rejected": -1.6987054347991943, "logps/chosen": -248.8202362060547, "logps/rejected": -105.59635925292969, "loss": 0.4226, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5064533352851868, "rewards/margins": 0.982274055480957, "rewards/rejected": -0.47582077980041504, "step": 141 }, { "epoch": 1.3446969696969697, "grad_norm": 6.277665119781322, "learning_rate": 3.5978835978835977e-08, "logits/chosen": -1.5494630336761475, "logits/rejected": -1.6693788766860962, "logps/chosen": -351.03125, "logps/rejected": -129.0759735107422, "loss": 0.3673, "rewards/accuracies": 1.0, "rewards/chosen": 0.7771791219711304, "rewards/margins": 1.4887819290161133, "rewards/rejected": -0.7116026878356934, "step": 142 }, { "epoch": 1.3541666666666667, "grad_norm": 5.8355671091327075, "learning_rate": 3.544973544973545e-08, "logits/chosen": -1.559160828590393, "logits/rejected": -1.5491154193878174, "logps/chosen": -335.740234375, "logps/rejected": -170.25552368164062, "loss": 0.4013, "rewards/accuracies": 1.0, "rewards/chosen": 0.7102228999137878, "rewards/margins": 1.2653253078460693, "rewards/rejected": -0.555102527141571, "step": 143 }, { "epoch": 1.3636363636363638, "grad_norm": 5.768525584457291, "learning_rate": 3.492063492063492e-08, "logits/chosen": -1.6074180603027344, "logits/rejected": -1.7357871532440186, "logps/chosen": -264.52679443359375, "logps/rejected": -100.31982421875, "loss": 0.408, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4864141345024109, "rewards/margins": 0.7962800860404968, "rewards/rejected": -0.30986595153808594, "step": 144 }, { "epoch": 1.3731060606060606, "grad_norm": 5.592283665576824, "learning_rate": 3.439153439153439e-08, "logits/chosen": -1.5794652700424194, "logits/rejected": -1.6425234079360962, "logps/chosen": -305.49005126953125, "logps/rejected": -128.72915649414062, "loss": 0.3636, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6161352396011353, "rewards/margins": 1.2325931787490845, "rewards/rejected": -0.6164579391479492, "step": 145 }, { "epoch": 1.3825757575757576, "grad_norm": 5.2043920846765515, "learning_rate": 3.386243386243386e-08, "logits/chosen": -1.6179447174072266, "logits/rejected": -1.7832599878311157, "logps/chosen": -317.20703125, "logps/rejected": -164.9311065673828, "loss": 0.3351, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6158957481384277, "rewards/margins": 1.1830036640167236, "rewards/rejected": -0.5671079158782959, "step": 146 }, { "epoch": 1.3920454545454546, "grad_norm": 5.389376859152257, "learning_rate": 3.333333333333333e-08, "logits/chosen": -1.5436028242111206, "logits/rejected": -1.7525603771209717, "logps/chosen": -217.86569213867188, "logps/rejected": -93.09127044677734, "loss": 0.3912, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4181972146034241, "rewards/margins": 0.8372113108634949, "rewards/rejected": -0.4190141260623932, "step": 147 }, { "epoch": 1.4015151515151514, "grad_norm": 5.429011665107999, "learning_rate": 3.2804232804232804e-08, "logits/chosen": -1.5702577829360962, "logits/rejected": -1.6486690044403076, "logps/chosen": -319.3020324707031, "logps/rejected": -182.95730590820312, "loss": 0.3457, "rewards/accuracies": 0.7333333492279053, "rewards/chosen": 0.5197921395301819, "rewards/margins": 0.8637401461601257, "rewards/rejected": -0.34394803643226624, "step": 148 }, { "epoch": 1.4109848484848486, "grad_norm": 5.368022221479501, "learning_rate": 3.2275132275132274e-08, "logits/chosen": -1.4937713146209717, "logits/rejected": -1.5651482343673706, "logps/chosen": -234.36328125, "logps/rejected": -148.08285522460938, "loss": 0.3444, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.23662462830543518, "rewards/margins": 0.6261420845985413, "rewards/rejected": -0.3895173966884613, "step": 149 }, { "epoch": 1.4204545454545454, "grad_norm": 5.493890942440773, "learning_rate": 3.1746031746031744e-08, "logits/chosen": -1.6106767654418945, "logits/rejected": -1.5224909782409668, "logps/chosen": -302.6184997558594, "logps/rejected": -146.08517456054688, "loss": 0.3663, "rewards/accuracies": 1.0, "rewards/chosen": 0.48458927869796753, "rewards/margins": 0.9824308156967163, "rewards/rejected": -0.49784159660339355, "step": 150 }, { "epoch": 1.4299242424242424, "grad_norm": 5.480765334687315, "learning_rate": 3.1216931216931215e-08, "logits/chosen": -1.5799682140350342, "logits/rejected": -1.7092950344085693, "logps/chosen": -245.9119110107422, "logps/rejected": -150.98989868164062, "loss": 0.4027, "rewards/accuracies": 1.0, "rewards/chosen": 0.4121614396572113, "rewards/margins": 0.8830633163452148, "rewards/rejected": -0.47090187668800354, "step": 151 }, { "epoch": 1.4393939393939394, "grad_norm": 6.044424074410309, "learning_rate": 3.0687830687830685e-08, "logits/chosen": -1.5871868133544922, "logits/rejected": -1.6946001052856445, "logps/chosen": -346.11822509765625, "logps/rejected": -244.7041778564453, "loss": 0.4391, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.3353393077850342, "rewards/margins": 0.7316913604736328, "rewards/rejected": -0.39635199308395386, "step": 152 }, { "epoch": 1.4488636363636362, "grad_norm": 4.974140552150539, "learning_rate": 3.0158730158730155e-08, "logits/chosen": -1.5874955654144287, "logits/rejected": -1.64119553565979, "logps/chosen": -270.36883544921875, "logps/rejected": -129.67535400390625, "loss": 0.3404, "rewards/accuracies": 1.0, "rewards/chosen": 0.7120252251625061, "rewards/margins": 1.3487727642059326, "rewards/rejected": -0.6367474794387817, "step": 153 }, { "epoch": 1.4583333333333333, "grad_norm": 5.350141771926167, "learning_rate": 2.9629629629629625e-08, "logits/chosen": -1.5650367736816406, "logits/rejected": -1.6837724447250366, "logps/chosen": -271.3443298339844, "logps/rejected": -136.472412109375, "loss": 0.3361, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.6383367776870728, "rewards/margins": 1.251798391342163, "rewards/rejected": -0.6134616136550903, "step": 154 }, { "epoch": 1.4678030303030303, "grad_norm": 5.429121594474915, "learning_rate": 2.91005291005291e-08, "logits/chosen": -1.568342924118042, "logits/rejected": -1.6543916463851929, "logps/chosen": -284.5910339355469, "logps/rejected": -152.6836395263672, "loss": 0.3293, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5576558113098145, "rewards/margins": 0.9742603302001953, "rewards/rejected": -0.41660451889038086, "step": 155 }, { "epoch": 1.4772727272727273, "grad_norm": 5.748560761291252, "learning_rate": 2.857142857142857e-08, "logits/chosen": -1.4832626581192017, "logits/rejected": -1.566612958908081, "logps/chosen": -323.40142822265625, "logps/rejected": -178.55923461914062, "loss": 0.421, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.6060717105865479, "rewards/margins": 1.2732772827148438, "rewards/rejected": -0.6672054529190063, "step": 156 }, { "epoch": 1.4867424242424243, "grad_norm": 5.310060511796978, "learning_rate": 2.804232804232804e-08, "logits/chosen": -1.5736675262451172, "logits/rejected": -1.6186530590057373, "logps/chosen": -316.97698974609375, "logps/rejected": -155.43638610839844, "loss": 0.3521, "rewards/accuracies": 1.0, "rewards/chosen": 0.6889209747314453, "rewards/margins": 1.1352983713150024, "rewards/rejected": -0.44637736678123474, "step": 157 }, { "epoch": 1.496212121212121, "grad_norm": 5.118308720530541, "learning_rate": 2.751322751322751e-08, "logits/chosen": -1.560807704925537, "logits/rejected": -1.7643381357192993, "logps/chosen": -402.16632080078125, "logps/rejected": -132.56011962890625, "loss": 0.3576, "rewards/accuracies": 1.0, "rewards/chosen": 0.7692705392837524, "rewards/margins": 1.3775382041931152, "rewards/rejected": -0.6082676649093628, "step": 158 }, { "epoch": 1.5056818181818183, "grad_norm": 5.9001461393206505, "learning_rate": 2.6984126984126982e-08, "logits/chosen": -1.5756150484085083, "logits/rejected": -1.6618843078613281, "logps/chosen": -144.79986572265625, "logps/rejected": -142.03944396972656, "loss": 0.3529, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.27606311440467834, "rewards/margins": 0.5494292378425598, "rewards/rejected": -0.27336612343788147, "step": 159 }, { "epoch": 1.5151515151515151, "grad_norm": 5.919050512179857, "learning_rate": 2.6455026455026453e-08, "logits/chosen": -1.596430778503418, "logits/rejected": -1.7081763744354248, "logps/chosen": -381.9976806640625, "logps/rejected": -168.22308349609375, "loss": 0.3632, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.7063287496566772, "rewards/margins": 1.4756296873092651, "rewards/rejected": -0.7693011164665222, "step": 160 }, { "epoch": 1.5246212121212122, "grad_norm": 6.771452793232538, "learning_rate": 2.5925925925925923e-08, "logits/chosen": -1.6167709827423096, "logits/rejected": -1.702599287033081, "logps/chosen": -272.24017333984375, "logps/rejected": -205.1815185546875, "loss": 0.3699, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6321600675582886, "rewards/margins": 1.0702993869781494, "rewards/rejected": -0.43813928961753845, "step": 161 }, { "epoch": 1.5340909090909092, "grad_norm": 5.844973840956046, "learning_rate": 2.5396825396825393e-08, "logits/chosen": -1.5797332525253296, "logits/rejected": -1.6782172918319702, "logps/chosen": -294.10540771484375, "logps/rejected": -178.05677795410156, "loss": 0.4349, "rewards/accuracies": 0.8666666150093079, "rewards/chosen": 0.4381368160247803, "rewards/margins": 0.7244521975517273, "rewards/rejected": -0.28631535172462463, "step": 162 }, { "epoch": 1.543560606060606, "grad_norm": 5.9799282018427355, "learning_rate": 2.4867724867724866e-08, "logits/chosen": -1.5741981267929077, "logits/rejected": -1.7271267175674438, "logps/chosen": -279.97894287109375, "logps/rejected": -155.4124298095703, "loss": 0.3788, "rewards/accuracies": 1.0, "rewards/chosen": 0.5363057255744934, "rewards/margins": 1.2445316314697266, "rewards/rejected": -0.7082260251045227, "step": 163 }, { "epoch": 1.553030303030303, "grad_norm": 5.688137477051427, "learning_rate": 2.4338624338624337e-08, "logits/chosen": -1.5423920154571533, "logits/rejected": -1.7703033685684204, "logps/chosen": -318.68963623046875, "logps/rejected": -139.80978393554688, "loss": 0.3199, "rewards/accuracies": 1.0, "rewards/chosen": 0.7159373760223389, "rewards/margins": 1.3830876350402832, "rewards/rejected": -0.6671503186225891, "step": 164 }, { "epoch": 1.5625, "grad_norm": 5.159199147076709, "learning_rate": 2.3809523809523807e-08, "logits/chosen": -1.6055580377578735, "logits/rejected": -1.7235870361328125, "logps/chosen": -283.23931884765625, "logps/rejected": -170.2466278076172, "loss": 0.3879, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5190203785896301, "rewards/margins": 0.8587003946304321, "rewards/rejected": -0.3396799862384796, "step": 165 }, { "epoch": 1.571969696969697, "grad_norm": 5.51593434736744, "learning_rate": 2.3280423280423277e-08, "logits/chosen": -1.6276752948760986, "logits/rejected": -1.705373764038086, "logps/chosen": -361.263671875, "logps/rejected": -157.3756866455078, "loss": 0.3503, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.813951313495636, "rewards/margins": 1.6092281341552734, "rewards/rejected": -0.7952768206596375, "step": 166 }, { "epoch": 1.581439393939394, "grad_norm": 5.105458599961924, "learning_rate": 2.275132275132275e-08, "logits/chosen": -1.4909486770629883, "logits/rejected": -1.6180715560913086, "logps/chosen": -216.1019287109375, "logps/rejected": -78.71076965332031, "loss": 0.3386, "rewards/accuracies": 1.0, "rewards/chosen": 0.488716185092926, "rewards/margins": 1.0591917037963867, "rewards/rejected": -0.5704755187034607, "step": 167 }, { "epoch": 1.5909090909090908, "grad_norm": 4.489310539882785, "learning_rate": 2.222222222222222e-08, "logits/chosen": -1.5696783065795898, "logits/rejected": -1.6010246276855469, "logps/chosen": -306.27294921875, "logps/rejected": -123.8807373046875, "loss": 0.3769, "rewards/accuracies": 1.0, "rewards/chosen": 0.7523918151855469, "rewards/margins": 1.3328092098236084, "rewards/rejected": -0.5804173946380615, "step": 168 }, { "epoch": 1.6003787878787878, "grad_norm": 5.400154224053731, "learning_rate": 2.169312169312169e-08, "logits/chosen": -1.5251656770706177, "logits/rejected": -1.6274874210357666, "logps/chosen": -130.54373168945312, "logps/rejected": -54.27812576293945, "loss": 0.4268, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.27597469091415405, "rewards/margins": 0.5285931825637817, "rewards/rejected": -0.2526185214519501, "step": 169 }, { "epoch": 1.6098484848484849, "grad_norm": 4.692184736415839, "learning_rate": 2.116402116402116e-08, "logits/chosen": -1.5292972326278687, "logits/rejected": -1.6729373931884766, "logps/chosen": -360.6134948730469, "logps/rejected": -151.4227294921875, "loss": 0.3845, "rewards/accuracies": 1.0, "rewards/chosen": 0.8804464340209961, "rewards/margins": 1.6148141622543335, "rewards/rejected": -0.7343679666519165, "step": 170 }, { "epoch": 1.6193181818181817, "grad_norm": 5.938380627671962, "learning_rate": 2.0634920634920634e-08, "logits/chosen": -1.5955592393875122, "logits/rejected": -1.7036798000335693, "logps/chosen": -384.96380615234375, "logps/rejected": -170.1694793701172, "loss": 0.3205, "rewards/accuracies": 1.0, "rewards/chosen": 0.8083392977714539, "rewards/margins": 1.5802361965179443, "rewards/rejected": -0.7718968391418457, "step": 171 }, { "epoch": 1.628787878787879, "grad_norm": 5.0095603748386415, "learning_rate": 2.0105820105820104e-08, "logits/chosen": -1.559670329093933, "logits/rejected": -1.6990169286727905, "logps/chosen": -260.5107421875, "logps/rejected": -138.97021484375, "loss": 0.3619, "rewards/accuracies": 1.0, "rewards/chosen": 0.6676410436630249, "rewards/margins": 1.1652050018310547, "rewards/rejected": -0.49756401777267456, "step": 172 }, { "epoch": 1.6382575757575757, "grad_norm": 5.046726473189086, "learning_rate": 1.9576719576719575e-08, "logits/chosen": -1.5894399881362915, "logits/rejected": -1.711214303970337, "logps/chosen": -308.7052307128906, "logps/rejected": -193.82850646972656, "loss": 0.3758, "rewards/accuracies": 1.0, "rewards/chosen": 0.514888346195221, "rewards/margins": 1.0377848148345947, "rewards/rejected": -0.5228964686393738, "step": 173 }, { "epoch": 1.6477272727272727, "grad_norm": 10.53036659721855, "learning_rate": 1.9047619047619045e-08, "logits/chosen": -1.5739425420761108, "logits/rejected": -1.5461231470108032, "logps/chosen": -272.7663269042969, "logps/rejected": -152.17575073242188, "loss": 0.3848, "rewards/accuracies": 1.0, "rewards/chosen": 0.457489013671875, "rewards/margins": 1.1182512044906616, "rewards/rejected": -0.6607621908187866, "step": 174 }, { "epoch": 1.6571969696969697, "grad_norm": 5.78981346354209, "learning_rate": 1.8518518518518518e-08, "logits/chosen": -1.5584112405776978, "logits/rejected": -1.66403329372406, "logps/chosen": -247.25527954101562, "logps/rejected": -88.81005096435547, "loss": 0.3545, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.38863998651504517, "rewards/margins": 0.8395411372184753, "rewards/rejected": -0.45090118050575256, "step": 175 }, { "epoch": 1.6666666666666665, "grad_norm": 5.544069471962977, "learning_rate": 1.7989417989417988e-08, "logits/chosen": -1.5310989618301392, "logits/rejected": -1.6411716938018799, "logps/chosen": -177.37261962890625, "logps/rejected": -94.64937591552734, "loss": 0.3993, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.40063905715942383, "rewards/margins": 0.7527322769165039, "rewards/rejected": -0.35209327936172485, "step": 176 }, { "epoch": 1.6761363636363638, "grad_norm": 6.870515621075784, "learning_rate": 1.746031746031746e-08, "logits/chosen": -1.6042630672454834, "logits/rejected": -1.7060787677764893, "logps/chosen": -329.9851989746094, "logps/rejected": -170.91259765625, "loss": 0.3258, "rewards/accuracies": 1.0, "rewards/chosen": 0.6290345191955566, "rewards/margins": 1.0913975238800049, "rewards/rejected": -0.4623629152774811, "step": 177 }, { "epoch": 1.6856060606060606, "grad_norm": 5.283044248226327, "learning_rate": 1.693121693121693e-08, "logits/chosen": -1.5727237462997437, "logits/rejected": -1.670090675354004, "logps/chosen": -242.1526336669922, "logps/rejected": -123.24870300292969, "loss": 0.3156, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5657342076301575, "rewards/margins": 1.0815435647964478, "rewards/rejected": -0.5158092379570007, "step": 178 }, { "epoch": 1.6950757575757576, "grad_norm": 4.714665015837958, "learning_rate": 1.6402116402116402e-08, "logits/chosen": -1.6179364919662476, "logits/rejected": -1.6783673763275146, "logps/chosen": -268.26666259765625, "logps/rejected": -125.2500991821289, "loss": 0.3028, "rewards/accuracies": 1.0, "rewards/chosen": 0.6312516331672668, "rewards/margins": 1.2757152318954468, "rewards/rejected": -0.6444636583328247, "step": 179 }, { "epoch": 1.7045454545454546, "grad_norm": 5.48800386545785, "learning_rate": 1.5873015873015872e-08, "logits/chosen": -1.587045431137085, "logits/rejected": -1.5843976736068726, "logps/chosen": -314.2275085449219, "logps/rejected": -202.48590087890625, "loss": 0.3407, "rewards/accuracies": 1.0, "rewards/chosen": 0.4960121214389801, "rewards/margins": 1.1439507007598877, "rewards/rejected": -0.64793860912323, "step": 180 }, { "epoch": 1.7140151515151514, "grad_norm": 4.904790456188903, "learning_rate": 1.5343915343915342e-08, "logits/chosen": -1.5327612161636353, "logits/rejected": -1.6153557300567627, "logps/chosen": -337.2438659667969, "logps/rejected": -185.306396484375, "loss": 0.3592, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.5627692937850952, "rewards/margins": 1.312888741493225, "rewards/rejected": -0.7501195669174194, "step": 181 }, { "epoch": 1.7234848484848486, "grad_norm": 4.656789025606906, "learning_rate": 1.4814814814814813e-08, "logits/chosen": -1.6096194982528687, "logits/rejected": -1.7528839111328125, "logps/chosen": -336.3609313964844, "logps/rejected": -170.637939453125, "loss": 0.3605, "rewards/accuracies": 1.0, "rewards/chosen": 0.5875322818756104, "rewards/margins": 1.3081997632980347, "rewards/rejected": -0.7206674218177795, "step": 182 }, { "epoch": 1.7329545454545454, "grad_norm": 5.444360894971359, "learning_rate": 1.4285714285714284e-08, "logits/chosen": -1.5957181453704834, "logits/rejected": -1.64398992061615, "logps/chosen": -289.0057678222656, "logps/rejected": -167.2021484375, "loss": 0.3833, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6649573445320129, "rewards/margins": 1.3411014080047607, "rewards/rejected": -0.6761440634727478, "step": 183 }, { "epoch": 1.7424242424242424, "grad_norm": 4.598365388289265, "learning_rate": 1.3756613756613755e-08, "logits/chosen": -1.6330028772354126, "logits/rejected": -1.6998701095581055, "logps/chosen": -245.7683868408203, "logps/rejected": -142.4471435546875, "loss": 0.392, "rewards/accuracies": 0.73333340883255, "rewards/chosen": 0.36274439096450806, "rewards/margins": 0.7017933130264282, "rewards/rejected": -0.3390488922595978, "step": 184 }, { "epoch": 1.7518939393939394, "grad_norm": 5.234791929984342, "learning_rate": 1.3227513227513226e-08, "logits/chosen": -1.5748271942138672, "logits/rejected": -1.6001355648040771, "logps/chosen": -276.45074462890625, "logps/rejected": -114.4822006225586, "loss": 0.3666, "rewards/accuracies": 1.0, "rewards/chosen": 0.6184806227684021, "rewards/margins": 1.1440364122390747, "rewards/rejected": -0.5255557894706726, "step": 185 }, { "epoch": 1.7613636363636362, "grad_norm": 5.537848306393154, "learning_rate": 1.2698412698412696e-08, "logits/chosen": -1.5969727039337158, "logits/rejected": -1.673282265663147, "logps/chosen": -222.10464477539062, "logps/rejected": -115.9425048828125, "loss": 0.3751, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.4460936188697815, "rewards/margins": 0.7085092663764954, "rewards/rejected": -0.26241564750671387, "step": 186 }, { "epoch": 1.7708333333333335, "grad_norm": 4.835457356224698, "learning_rate": 1.2169312169312168e-08, "logits/chosen": -1.5568116903305054, "logits/rejected": -1.554927110671997, "logps/chosen": -256.4602966308594, "logps/rejected": -167.30160522460938, "loss": 0.3714, "rewards/accuracies": 1.0, "rewards/chosen": 0.4909042418003082, "rewards/margins": 1.2990154027938843, "rewards/rejected": -0.8081111907958984, "step": 187 }, { "epoch": 1.7803030303030303, "grad_norm": 5.606816215751534, "learning_rate": 1.1640211640211638e-08, "logits/chosen": -1.578005313873291, "logits/rejected": -1.5507481098175049, "logps/chosen": -302.0408630371094, "logps/rejected": -188.82948303222656, "loss": 0.3905, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.4071738123893738, "rewards/margins": 0.8548933267593384, "rewards/rejected": -0.44771942496299744, "step": 188 }, { "epoch": 1.7897727272727273, "grad_norm": 5.104068105653885, "learning_rate": 1.111111111111111e-08, "logits/chosen": -1.5117594003677368, "logits/rejected": -1.6626886129379272, "logps/chosen": -316.0799255371094, "logps/rejected": -152.69705200195312, "loss": 0.3749, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.49852848052978516, "rewards/margins": 0.9504121541976929, "rewards/rejected": -0.45188361406326294, "step": 189 }, { "epoch": 1.7992424242424243, "grad_norm": 4.472316490828338, "learning_rate": 1.058201058201058e-08, "logits/chosen": -1.62430739402771, "logits/rejected": -1.6985547542572021, "logps/chosen": -448.1697692871094, "logps/rejected": -224.6002655029297, "loss": 0.3446, "rewards/accuracies": 1.0, "rewards/chosen": 0.8359583616256714, "rewards/margins": 1.992840051651001, "rewards/rejected": -1.1568816900253296, "step": 190 }, { "epoch": 1.808712121212121, "grad_norm": 5.810880253949101, "learning_rate": 1.0052910052910052e-08, "logits/chosen": -1.5918437242507935, "logits/rejected": -1.765323281288147, "logps/chosen": -375.71429443359375, "logps/rejected": -159.8873291015625, "loss": 0.399, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.9120176434516907, "rewards/margins": 1.7272794246673584, "rewards/rejected": -0.8152618408203125, "step": 191 }, { "epoch": 1.8181818181818183, "grad_norm": 4.244845222552646, "learning_rate": 9.523809523809522e-09, "logits/chosen": -1.5941245555877686, "logits/rejected": -1.6600911617279053, "logps/chosen": -305.6556701660156, "logps/rejected": -134.87252807617188, "loss": 0.3493, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.5449376106262207, "rewards/margins": 1.1537270545959473, "rewards/rejected": -0.6087895631790161, "step": 192 }, { "epoch": 1.8276515151515151, "grad_norm": 4.738897289851384, "learning_rate": 8.994708994708994e-09, "logits/chosen": -1.5437757968902588, "logits/rejected": -1.6903842687606812, "logps/chosen": -321.27752685546875, "logps/rejected": -126.69343566894531, "loss": 0.3567, "rewards/accuracies": 1.0, "rewards/chosen": 0.6183031797409058, "rewards/margins": 1.4520056247711182, "rewards/rejected": -0.8337022662162781, "step": 193 }, { "epoch": 1.8371212121212122, "grad_norm": 5.490859875590591, "learning_rate": 8.465608465608464e-09, "logits/chosen": -1.6326143741607666, "logits/rejected": -1.7613766193389893, "logps/chosen": -266.3793029785156, "logps/rejected": -130.99026489257812, "loss": 0.4226, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5176450610160828, "rewards/margins": 0.859043300151825, "rewards/rejected": -0.3413982093334198, "step": 194 }, { "epoch": 1.8465909090909092, "grad_norm": 5.583102046304065, "learning_rate": 7.936507936507936e-09, "logits/chosen": -1.5793282985687256, "logits/rejected": -1.789536714553833, "logps/chosen": -386.55694580078125, "logps/rejected": -170.34373474121094, "loss": 0.3427, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5163098573684692, "rewards/margins": 0.9769344329833984, "rewards/rejected": -0.4606245458126068, "step": 195 }, { "epoch": 1.856060606060606, "grad_norm": 6.799122206347572, "learning_rate": 7.407407407407406e-09, "logits/chosen": -1.5655455589294434, "logits/rejected": -1.7494919300079346, "logps/chosen": -340.42059326171875, "logps/rejected": -156.17343139648438, "loss": 0.3914, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.6122487783432007, "rewards/margins": 1.0115610361099243, "rewards/rejected": -0.39931216835975647, "step": 196 }, { "epoch": 1.865530303030303, "grad_norm": 6.33489638516624, "learning_rate": 6.878306878306877e-09, "logits/chosen": -1.6181929111480713, "logits/rejected": -1.7337757349014282, "logps/chosen": -297.11236572265625, "logps/rejected": -133.5535430908203, "loss": 0.3948, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.5498533844947815, "rewards/margins": 1.0373098850250244, "rewards/rejected": -0.48745641112327576, "step": 197 }, { "epoch": 1.875, "grad_norm": 4.468769998980834, "learning_rate": 6.349206349206348e-09, "logits/chosen": -1.4640628099441528, "logits/rejected": -1.5838810205459595, "logps/chosen": -238.00106811523438, "logps/rejected": -104.63729095458984, "loss": 0.3595, "rewards/accuracies": 1.0, "rewards/chosen": 0.5341485142707825, "rewards/margins": 1.1648151874542236, "rewards/rejected": -0.6306666135787964, "step": 198 }, { "epoch": 1.884469696969697, "grad_norm": 5.080896081269075, "learning_rate": 5.820105820105819e-09, "logits/chosen": -1.5636723041534424, "logits/rejected": -1.658831238746643, "logps/chosen": -203.7349395751953, "logps/rejected": -110.94892883300781, "loss": 0.3676, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.3449416756629944, "rewards/margins": 0.7937332987785339, "rewards/rejected": -0.44879165291786194, "step": 199 }, { "epoch": 1.893939393939394, "grad_norm": 5.280192220580751, "learning_rate": 5.29100529100529e-09, "logits/chosen": -1.5569162368774414, "logits/rejected": -1.6692053079605103, "logps/chosen": -310.94293212890625, "logps/rejected": -180.21963500976562, "loss": 0.4127, "rewards/accuracies": 1.0, "rewards/chosen": 0.6343954801559448, "rewards/margins": 1.3330564498901367, "rewards/rejected": -0.6986608505249023, "step": 200 }, { "epoch": 1.9034090909090908, "grad_norm": 4.3144901310959956, "learning_rate": 4.761904761904761e-09, "logits/chosen": -1.5700764656066895, "logits/rejected": -1.5398668050765991, "logps/chosen": -123.7179946899414, "logps/rejected": -168.89608764648438, "loss": 0.3673, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.18818268179893494, "rewards/margins": 0.4049772620201111, "rewards/rejected": -0.21679461002349854, "step": 201 }, { "epoch": 1.9128787878787878, "grad_norm": 4.809421227753802, "learning_rate": 4.232804232804232e-09, "logits/chosen": -1.5965473651885986, "logits/rejected": -1.7248961925506592, "logps/chosen": -306.676025390625, "logps/rejected": -186.1800994873047, "loss": 0.363, "rewards/accuracies": 0.8666666746139526, "rewards/chosen": 0.401868999004364, "rewards/margins": 0.8128830194473267, "rewards/rejected": -0.41101402044296265, "step": 202 }, { "epoch": 1.9223484848484849, "grad_norm": 4.639189256023351, "learning_rate": 3.703703703703703e-09, "logits/chosen": -1.5983052253723145, "logits/rejected": -1.6595712900161743, "logps/chosen": -307.19183349609375, "logps/rejected": -195.84725952148438, "loss": 0.3743, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.6213973760604858, "rewards/margins": 1.057089924812317, "rewards/rejected": -0.4356924891471863, "step": 203 }, { "epoch": 1.9318181818181817, "grad_norm": 5.728537848527359, "learning_rate": 3.174603174603174e-09, "logits/chosen": -1.5698726177215576, "logits/rejected": -1.6646344661712646, "logps/chosen": -269.0412292480469, "logps/rejected": -137.50003051757812, "loss": 0.3975, "rewards/accuracies": 1.0, "rewards/chosen": 0.5319095849990845, "rewards/margins": 1.2130988836288452, "rewards/rejected": -0.6811891794204712, "step": 204 }, { "epoch": 1.941287878787879, "grad_norm": 5.416585887060536, "learning_rate": 2.645502645502645e-09, "logits/chosen": -1.6159827709197998, "logits/rejected": -1.8529045581817627, "logps/chosen": -324.2301025390625, "logps/rejected": -169.9268035888672, "loss": 0.4161, "rewards/accuracies": 0.8666666150093079, "rewards/chosen": 0.44274115562438965, "rewards/margins": 0.865648090839386, "rewards/rejected": -0.4229070544242859, "step": 205 }, { "epoch": 1.9507575757575757, "grad_norm": 5.021756713871844, "learning_rate": 2.116402116402116e-09, "logits/chosen": -1.6490414142608643, "logits/rejected": -1.701080083847046, "logps/chosen": -223.4647979736328, "logps/rejected": -112.0922622680664, "loss": 0.4041, "rewards/accuracies": 0.9333332777023315, "rewards/chosen": 0.25973638892173767, "rewards/margins": 0.5601609349250793, "rewards/rejected": -0.3004245162010193, "step": 206 }, { "epoch": 1.9602272727272727, "grad_norm": 5.226300428110887, "learning_rate": 1.587301587301587e-09, "logits/chosen": -1.6259441375732422, "logits/rejected": -1.6718814373016357, "logps/chosen": -314.91583251953125, "logps/rejected": -182.3265838623047, "loss": 0.3799, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.5016330480575562, "rewards/margins": 1.0548944473266602, "rewards/rejected": -0.553261399269104, "step": 207 }, { "epoch": 1.9696969696969697, "grad_norm": 4.5059277056393725, "learning_rate": 1.058201058201058e-09, "logits/chosen": -1.6556800603866577, "logits/rejected": -1.7395381927490234, "logps/chosen": -346.4697570800781, "logps/rejected": -191.90931701660156, "loss": 0.3702, "rewards/accuracies": 1.0, "rewards/chosen": 0.8226908445358276, "rewards/margins": 1.620936393737793, "rewards/rejected": -0.7982456684112549, "step": 208 }, { "epoch": 1.9791666666666665, "grad_norm": 4.466235222468724, "learning_rate": 5.29100529100529e-10, "logits/chosen": -1.5888211727142334, "logits/rejected": -1.6132078170776367, "logps/chosen": -215.74612426757812, "logps/rejected": -148.16468811035156, "loss": 0.3622, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.3967617154121399, "rewards/margins": 0.8830636143684387, "rewards/rejected": -0.48630183935165405, "step": 209 }, { "epoch": 1.9886363636363638, "grad_norm": 4.940010839228651, "learning_rate": 0.0, "logits/chosen": -1.6120548248291016, "logits/rejected": -1.675747275352478, "logps/chosen": -285.6683654785156, "logps/rejected": -142.8597869873047, "loss": 0.3871, "rewards/accuracies": 0.9333333969116211, "rewards/chosen": 0.4850710332393646, "rewards/margins": 1.1287145614624023, "rewards/rejected": -0.6436434984207153, "step": 210 } ], "logging_steps": 1.0, "max_steps": 210, "num_input_tokens_seen": 0, "num_train_epochs": 2, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 3, "trial_name": null, "trial_params": null }