{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 327,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 37.56032423344454,
      "learning_rate": 1.5151515151515152e-08,
      "logits/chosen": -3.115293502807617,
      "logits/rejected": -3.0863265991210938,
      "logps/chosen": -870.7420043945312,
      "logps/rejected": -1334.502685546875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 37.49331944203386,
      "learning_rate": 1.5151515151515152e-07,
      "logits/chosen": -3.0526609420776367,
      "logits/rejected": -3.0749807357788086,
      "logps/chosen": -940.7243041992188,
      "logps/rejected": -1276.6510009765625,
      "loss": 0.6922,
      "rewards/accuracies": 0.5486111044883728,
      "rewards/chosen": 0.001147550530731678,
      "rewards/margins": 0.0018619541078805923,
      "rewards/rejected": -0.000714403809979558,
      "step": 10
    },
    {
      "epoch": 0.06,
      "grad_norm": 32.798777633158615,
      "learning_rate": 3.0303030303030305e-07,
      "logits/chosen": -3.0576233863830566,
      "logits/rejected": -3.0690040588378906,
      "logps/chosen": -956.8638916015625,
      "logps/rejected": -1231.0487060546875,
      "loss": 0.668,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.03549911081790924,
      "rewards/margins": 0.057550616562366486,
      "rewards/rejected": -0.022051498293876648,
      "step": 20
    },
    {
      "epoch": 0.09,
      "grad_norm": 31.200374512571734,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": -3.1981358528137207,
      "logits/rejected": -3.1817786693573,
      "logps/chosen": -1004.2825927734375,
      "logps/rejected": -1336.2808837890625,
      "loss": 0.6056,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.0012753400951623917,
      "rewards/margins": 0.2531731426715851,
      "rewards/rejected": -0.25444847345352173,
      "step": 30
    },
    {
      "epoch": 0.12,
      "grad_norm": 31.57664285278684,
      "learning_rate": 4.993009492952949e-07,
      "logits/chosen": -3.2853000164031982,
      "logits/rejected": -3.300828218460083,
      "logps/chosen": -1001.1776123046875,
      "logps/rejected": -1323.44287109375,
      "loss": 0.5156,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.10599758476018906,
      "rewards/margins": 0.550561785697937,
      "rewards/rejected": -0.656559407711029,
      "step": 40
    },
    {
      "epoch": 0.15,
      "grad_norm": 32.0572257728948,
      "learning_rate": 4.958864353026382e-07,
      "logits/chosen": -3.3941280841827393,
      "logits/rejected": -3.3549060821533203,
      "logps/chosen": -880.5036010742188,
      "logps/rejected": -1494.1220703125,
      "loss": 0.4348,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.06764356046915054,
      "rewards/margins": 1.6453561782836914,
      "rewards/rejected": -1.712999701499939,
      "step": 50
    },
    {
      "epoch": 0.18,
      "grad_norm": 37.653606745673876,
      "learning_rate": 4.896669632591651e-07,
      "logits/chosen": -3.4180691242218018,
      "logits/rejected": -3.3949356079101562,
      "logps/chosen": -997.1550903320312,
      "logps/rejected": -1486.9512939453125,
      "loss": 0.4241,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.191396564245224,
      "rewards/margins": 1.6561005115509033,
      "rewards/rejected": -1.847496747970581,
      "step": 60
    },
    {
      "epoch": 0.21,
      "grad_norm": 27.62213668513971,
      "learning_rate": 4.80713482008226e-07,
      "logits/chosen": -3.2756118774414062,
      "logits/rejected": -3.325742721557617,
      "logps/chosen": -1002.0143432617188,
      "logps/rejected": -1562.6756591796875,
      "loss": 0.3597,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.11527639627456665,
      "rewards/margins": 2.851576566696167,
      "rewards/rejected": -2.966853380203247,
      "step": 70
    },
    {
      "epoch": 0.24,
      "grad_norm": 24.96995353325617,
      "learning_rate": 4.691281286981619e-07,
      "logits/chosen": -3.2543506622314453,
      "logits/rejected": -3.2743306159973145,
      "logps/chosen": -1047.9954833984375,
      "logps/rejected": -1567.3741455078125,
      "loss": 0.3576,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -0.11167576164007187,
      "rewards/margins": 2.7660152912139893,
      "rewards/rejected": -2.8776910305023193,
      "step": 80
    },
    {
      "epoch": 0.28,
      "grad_norm": 27.750313403051702,
      "learning_rate": 4.550430636492389e-07,
      "logits/chosen": -3.152367115020752,
      "logits/rejected": -3.1291539669036865,
      "logps/chosen": -958.4727783203125,
      "logps/rejected": -1616.8697509765625,
      "loss": 0.3037,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -0.08375158160924911,
      "rewards/margins": 3.5008788108825684,
      "rewards/rejected": -3.5846304893493652,
      "step": 90
    },
    {
      "epoch": 0.31,
      "grad_norm": 30.508104431701753,
      "learning_rate": 4.386189627302036e-07,
      "logits/chosen": -3.1359643936157227,
      "logits/rejected": -3.0350069999694824,
      "logps/chosen": -1068.0445556640625,
      "logps/rejected": -1752.9798583984375,
      "loss": 0.2478,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.297347754240036,
      "rewards/margins": 4.142968654632568,
      "rewards/rejected": -4.440316200256348,
      "step": 100
    },
    {
      "epoch": 0.31,
      "eval_logits/chosen": -2.116163492202759,
      "eval_logits/rejected": -2.9199953079223633,
      "eval_logps/chosen": -214.01956176757812,
      "eval_logps/rejected": -996.0294189453125,
      "eval_loss": 0.03523317724466324,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -0.1739114671945572,
      "eval_rewards/margins": 4.2524919509887695,
      "eval_rewards/rejected": -4.426403999328613,
      "eval_runtime": 6.9945,
      "eval_samples_per_second": 9.15,
      "eval_steps_per_second": 0.286,
      "step": 100
    },
    {
      "epoch": 0.34,
      "grad_norm": 30.142384194174287,
      "learning_rate": 4.200431844427298e-07,
      "logits/chosen": -3.1088919639587402,
      "logits/rejected": -2.9363715648651123,
      "logps/chosen": -956.9268798828125,
      "logps/rejected": -1825.817138671875,
      "loss": 0.2479,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.4095672070980072,
      "rewards/margins": 4.6759538650512695,
      "rewards/rejected": -5.085520267486572,
      "step": 110
    },
    {
      "epoch": 0.37,
      "grad_norm": 30.548015460707564,
      "learning_rate": 3.9952763262280397e-07,
      "logits/chosen": -3.0125510692596436,
      "logits/rejected": -2.8323769569396973,
      "logps/chosen": -974.1087646484375,
      "logps/rejected": -1928.265380859375,
      "loss": 0.2222,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": -0.6024614572525024,
      "rewards/margins": 5.620959281921387,
      "rewards/rejected": -6.223421096801758,
      "step": 120
    },
    {
      "epoch": 0.4,
      "grad_norm": 23.95507006913498,
      "learning_rate": 3.7730633914035584e-07,
      "logits/chosen": -2.9912257194519043,
      "logits/rejected": -2.7818751335144043,
      "logps/chosen": -1056.4798583984375,
      "logps/rejected": -2160.14990234375,
      "loss": 0.1988,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.6903690099716187,
      "rewards/margins": 7.284422397613525,
      "rewards/rejected": -7.974791526794434,
      "step": 130
    },
    {
      "epoch": 0.43,
      "grad_norm": 42.81780370407222,
      "learning_rate": 3.5363279417257385e-07,
      "logits/chosen": -2.8955721855163574,
      "logits/rejected": -2.754878044128418,
      "logps/chosen": -1038.2923583984375,
      "logps/rejected": -1966.652587890625,
      "loss": 0.2092,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": -1.0339103937149048,
      "rewards/margins": 5.819826126098633,
      "rewards/rejected": -6.85373592376709,
      "step": 140
    },
    {
      "epoch": 0.46,
      "grad_norm": 31.239186243610785,
      "learning_rate": 3.287770545059052e-07,
      "logits/chosen": -2.8968818187713623,
      "logits/rejected": -2.668034076690674,
      "logps/chosen": -1137.552978515625,
      "logps/rejected": -2158.06298828125,
      "loss": 0.1807,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -1.3718111515045166,
      "rewards/margins": 7.208723545074463,
      "rewards/rejected": -8.580533981323242,
      "step": 150
    },
    {
      "epoch": 0.49,
      "grad_norm": 30.783011951421607,
      "learning_rate": 3.030226628538842e-07,
      "logits/chosen": -2.7924959659576416,
      "logits/rejected": -2.620363473892212,
      "logps/chosen": -1211.665771484375,
      "logps/rejected": -2208.79833984375,
      "loss": 0.1596,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -1.6732728481292725,
      "rewards/margins": 7.373621463775635,
      "rewards/rejected": -9.046895027160645,
      "step": 160
    },
    {
      "epoch": 0.52,
      "grad_norm": 38.33986073767423,
      "learning_rate": 2.7666341333378004e-07,
      "logits/chosen": -2.7492575645446777,
      "logits/rejected": -2.5758957862854004,
      "logps/chosen": -1068.8045654296875,
      "logps/rejected": -2013.82421875,
      "loss": 0.1765,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -1.2596815824508667,
      "rewards/margins": 6.303467750549316,
      "rewards/rejected": -7.563148498535156,
      "step": 170
    },
    {
      "epoch": 0.55,
      "grad_norm": 46.105813888725216,
      "learning_rate": 2.5e-07,
      "logits/chosen": -2.7488133907318115,
      "logits/rejected": -2.492980480194092,
      "logps/chosen": -1113.6063232421875,
      "logps/rejected": -2361.96923828125,
      "loss": 0.1225,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -1.5648447275161743,
      "rewards/margins": 8.788183212280273,
      "rewards/rejected": -10.353028297424316,
      "step": 180
    },
    {
      "epoch": 0.58,
      "grad_norm": 38.659892201318485,
      "learning_rate": 2.2333658666621994e-07,
      "logits/chosen": -2.716352939605713,
      "logits/rejected": -2.4929823875427246,
      "logps/chosen": -1151.252685546875,
      "logps/rejected": -2195.82958984375,
      "loss": 0.1587,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -1.521728277206421,
      "rewards/margins": 7.973954677581787,
      "rewards/rejected": -9.495683670043945,
      "step": 190
    },
    {
      "epoch": 0.61,
      "grad_norm": 34.038321670255556,
      "learning_rate": 1.9697733714611587e-07,
      "logits/chosen": -2.77030873298645,
      "logits/rejected": -2.5207982063293457,
      "logps/chosen": -1151.6956787109375,
      "logps/rejected": -2134.08349609375,
      "loss": 0.1385,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -1.476732850074768,
      "rewards/margins": 6.9633917808532715,
      "rewards/rejected": -8.440125465393066,
      "step": 200
    },
    {
      "epoch": 0.61,
      "eval_logits/chosen": -1.6287028789520264,
      "eval_logits/rejected": -2.333620548248291,
      "eval_logps/chosen": -230.22816467285156,
      "eval_logps/rejected": -1373.3590087890625,
      "eval_loss": 0.00411956338211894,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -0.3359977602958679,
      "eval_rewards/margins": 7.863701820373535,
      "eval_rewards/rejected": -8.199699401855469,
      "eval_runtime": 6.9561,
      "eval_samples_per_second": 9.201,
      "eval_steps_per_second": 0.288,
      "step": 200
    },
    {
      "epoch": 0.64,
      "grad_norm": 25.859030241191462,
      "learning_rate": 1.7122294549409482e-07,
      "logits/chosen": -2.7445123195648193,
      "logits/rejected": -2.477301597595215,
      "logps/chosen": -1209.62158203125,
      "logps/rejected": -2304.346435546875,
      "loss": 0.1239,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -1.5678670406341553,
      "rewards/margins": 8.935761451721191,
      "rewards/rejected": -10.503628730773926,
      "step": 210
    },
    {
      "epoch": 0.67,
      "grad_norm": 54.09588414480716,
      "learning_rate": 1.4636720582742613e-07,
      "logits/chosen": -2.6386165618896484,
      "logits/rejected": -2.3536062240600586,
      "logps/chosen": -1170.019287109375,
      "logps/rejected": -2326.010986328125,
      "loss": 0.1338,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -1.9397300481796265,
      "rewards/margins": 8.54932975769043,
      "rewards/rejected": -10.489059448242188,
      "step": 220
    },
    {
      "epoch": 0.7,
      "grad_norm": 35.229208104251256,
      "learning_rate": 1.2269366085964422e-07,
      "logits/chosen": -2.594449758529663,
      "logits/rejected": -2.3216192722320557,
      "logps/chosen": -1134.5150146484375,
      "logps/rejected": -2302.87158203125,
      "loss": 0.1241,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -2.0043346881866455,
      "rewards/margins": 8.11410903930664,
      "rewards/rejected": -10.118444442749023,
      "step": 230
    },
    {
      "epoch": 0.73,
      "grad_norm": 35.55573135898996,
      "learning_rate": 1.00472367377196e-07,
      "logits/chosen": -2.6050658226013184,
      "logits/rejected": -2.3415164947509766,
      "logps/chosen": -1212.265380859375,
      "logps/rejected": -2304.11669921875,
      "loss": 0.1046,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -1.8809270858764648,
      "rewards/margins": 8.250687599182129,
      "rewards/rejected": -10.131612777709961,
      "step": 240
    },
    {
      "epoch": 0.76,
      "grad_norm": 31.88590245463047,
      "learning_rate": 7.99568155572701e-08,
      "logits/chosen": -2.6101841926574707,
      "logits/rejected": -2.0731587409973145,
      "logps/chosen": -1139.8570556640625,
      "logps/rejected": -2845.157470703125,
      "loss": 0.0885,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -1.8957105875015259,
      "rewards/margins": 13.620719909667969,
      "rewards/rejected": -15.516430854797363,
      "step": 250
    },
    {
      "epoch": 0.8,
      "grad_norm": 51.7448993159194,
      "learning_rate": 6.13810372697964e-08,
      "logits/chosen": -2.5568299293518066,
      "logits/rejected": -2.152559757232666,
      "logps/chosen": -1236.4830322265625,
      "logps/rejected": -2579.609619140625,
      "loss": 0.1114,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": -2.0548171997070312,
      "rewards/margins": 11.762044906616211,
      "rewards/rejected": -13.816862106323242,
      "step": 260
    },
    {
      "epoch": 0.83,
      "grad_norm": 27.4784440806591,
      "learning_rate": 4.4956936350761005e-08,
      "logits/chosen": -2.598968267440796,
      "logits/rejected": -2.217064619064331,
      "logps/chosen": -1198.1903076171875,
      "logps/rejected": -2465.025634765625,
      "loss": 0.0966,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -2.0705723762512207,
      "rewards/margins": 9.766708374023438,
      "rewards/rejected": -11.837281227111816,
      "step": 270
    },
    {
      "epoch": 0.86,
      "grad_norm": 35.99094553945511,
      "learning_rate": 3.087187130183805e-08,
      "logits/chosen": -2.51914119720459,
      "logits/rejected": -2.214329481124878,
      "logps/chosen": -1253.900146484375,
      "logps/rejected": -2469.315185546875,
      "loss": 0.1108,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -2.140868663787842,
      "rewards/margins": 9.513227462768555,
      "rewards/rejected": -11.654096603393555,
      "step": 280
    },
    {
      "epoch": 0.89,
      "grad_norm": 62.565118319363144,
      "learning_rate": 1.928651799177408e-08,
      "logits/chosen": -2.5865471363067627,
      "logits/rejected": -2.248619556427002,
      "logps/chosen": -1134.03173828125,
      "logps/rejected": -2389.395751953125,
      "loss": 0.1085,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": -2.002607822418213,
      "rewards/margins": 8.705282211303711,
      "rewards/rejected": -10.707889556884766,
      "step": 290
    },
    {
      "epoch": 0.92,
      "grad_norm": 37.73877960542491,
      "learning_rate": 1.0333036740834855e-08,
      "logits/chosen": -2.552736520767212,
      "logits/rejected": -2.2399215698242188,
      "logps/chosen": -1204.527587890625,
      "logps/rejected": -2376.899169921875,
      "loss": 0.0899,
      "rewards/accuracies": 0.9937499761581421,
      "rewards/chosen": -2.1047260761260986,
      "rewards/margins": 8.565834045410156,
      "rewards/rejected": -10.670559883117676,
      "step": 300
    },
    {
      "epoch": 0.92,
      "eval_logits/chosen": -1.3416023254394531,
      "eval_logits/rejected": -2.055326461791992,
      "eval_logps/chosen": -241.41647338867188,
      "eval_logps/rejected": -1559.6263427734375,
      "eval_loss": 0.0018886247416958213,
      "eval_rewards/accuracies": 1.0,
      "eval_rewards/chosen": -0.44788068532943726,
      "eval_rewards/margins": 9.614492416381836,
      "eval_rewards/rejected": -10.062373161315918,
      "eval_runtime": 6.9591,
      "eval_samples_per_second": 9.197,
      "eval_steps_per_second": 0.287,
      "step": 300
    },
    {
      "epoch": 0.95,
      "grad_norm": 29.76747785442669,
      "learning_rate": 4.113564697361821e-09,
      "logits/chosen": -2.521284580230713,
      "logits/rejected": -2.1441428661346436,
      "logps/chosen": -1216.38916015625,
      "logps/rejected": -2597.3994140625,
      "loss": 0.0804,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -1.9038699865341187,
      "rewards/margins": 11.402238845825195,
      "rewards/rejected": -13.306109428405762,
      "step": 310
    },
    {
      "epoch": 0.98,
      "grad_norm": 32.64158882779933,
      "learning_rate": 6.990507047049676e-10,
      "logits/chosen": -2.489298105239868,
      "logits/rejected": -2.159824848175049,
      "logps/chosen": -1119.7908935546875,
      "logps/rejected": -2548.427001953125,
      "loss": 0.0939,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -2.2767906188964844,
      "rewards/margins": 10.406248092651367,
      "rewards/rejected": -12.683038711547852,
      "step": 320
    },
    {
      "epoch": 1.0,
      "step": 327,
      "total_flos": 0.0,
      "train_loss": 0.23720380818807385,
      "train_runtime": 5120.8034,
      "train_samples_per_second": 4.087,
      "train_steps_per_second": 0.064
    }
  ],
  "logging_steps": 10,
  "max_steps": 327,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}