{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.968,
  "eval_steps": 100,
  "global_step": 248,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 2.0000000000000002e-07,
      "logits/chosen": 0.23443643748760223,
      "logits/rejected": 0.22047671675682068,
      "logps/chosen": -156.2631072998047,
      "logps/rejected": -63.140926361083984,
      "loss": 0.0011,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16,
      "learning_rate": 2.0000000000000003e-06,
      "logits/chosen": -0.1701306700706482,
      "logits/rejected": 0.23002521693706512,
      "logps/chosen": -134.1260223388672,
      "logps/rejected": -77.3587417602539,
      "loss": 0.0013,
      "rewards/accuracies": 0.3611111044883728,
      "rewards/chosen": -0.0012936631683260202,
      "rewards/margins": -0.001161300460807979,
      "rewards/rejected": -0.0001323625911027193,
      "step": 10
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.000000000000001e-06,
      "logits/chosen": -0.1295575499534607,
      "logits/rejected": 0.013991622254252434,
      "logps/chosen": -158.0836181640625,
      "logps/rejected": -93.96044158935547,
      "loss": 0.0014,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": -0.0005890157772228122,
      "rewards/margins": -0.0010765515035018325,
      "rewards/rejected": 0.0004875357262790203,
      "step": 20
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.993800445762451e-06,
      "logits/chosen": -0.055564045906066895,
      "logits/rejected": 0.03296099230647087,
      "logps/chosen": -152.33218383789062,
      "logps/rejected": -95.81327819824219,
      "loss": 0.0014,
      "rewards/accuracies": 0.4437499940395355,
      "rewards/chosen": -0.0007998603396117687,
      "rewards/margins": -0.0010081109357997775,
      "rewards/rejected": 0.00020825063984375447,
      "step": 30
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.944388344834205e-06,
      "logits/chosen": -0.07261649519205093,
      "logits/rejected": 0.052390582859516144,
      "logps/chosen": -139.1942901611328,
      "logps/rejected": -90.42237854003906,
      "loss": 0.0013,
      "rewards/accuracies": 0.4124999940395355,
      "rewards/chosen": 0.0016054753214120865,
      "rewards/margins": -0.0004346861387602985,
      "rewards/rejected": 0.002040161518380046,
      "step": 40
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.8465431931347904e-06,
      "logits/chosen": -0.18433263897895813,
      "logits/rejected": 0.1357397437095642,
      "logps/chosen": -134.5174560546875,
      "logps/rejected": -78.68228912353516,
      "loss": 0.0012,
      "rewards/accuracies": 0.4937500059604645,
      "rewards/chosen": 0.0027162893675267696,
      "rewards/margins": 0.0013110730797052383,
      "rewards/rejected": 0.0014052166370674968,
      "step": 50
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.702203692102539e-06,
      "logits/chosen": -0.13251043856143951,
      "logits/rejected": 0.08991677314043045,
      "logps/chosen": -131.57760620117188,
      "logps/rejected": -75.83609771728516,
      "loss": 0.0012,
      "rewards/accuracies": 0.5062500238418579,
      "rewards/chosen": 0.0032040614169090986,
      "rewards/margins": 0.00045879706158302724,
      "rewards/rejected": 0.0027452646754682064,
      "step": 60
    },
    {
      "epoch": 1.12,
      "learning_rate": 4.514229781074239e-06,
      "logits/chosen": -0.0031040043104439974,
      "logits/rejected": 0.02582806721329689,
      "logps/chosen": -146.350830078125,
      "logps/rejected": -91.60733795166016,
      "loss": 0.0012,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.002755067078396678,
      "rewards/margins": 0.0003604464582167566,
      "rewards/rejected": 0.002394620794802904,
      "step": 70
    },
    {
      "epoch": 1.28,
      "learning_rate": 4.286345970517195e-06,
      "logits/chosen": -0.14322714507579803,
      "logits/rejected": 0.1631973534822464,
      "logps/chosen": -147.42727661132812,
      "logps/rejected": -78.87086486816406,
      "loss": 0.0011,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.0060631753876805305,
      "rewards/margins": 0.003673513885587454,
      "rewards/rejected": 0.0023896615020930767,
      "step": 80
    },
    {
      "epoch": 1.44,
      "learning_rate": 4.023067544670082e-06,
      "logits/chosen": -0.13109929859638214,
      "logits/rejected": 0.06961780786514282,
      "logps/chosen": -156.49757385253906,
      "logps/rejected": -77.63360595703125,
      "loss": 0.001,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.0063803596422076225,
      "rewards/margins": 0.003957846201956272,
      "rewards/rejected": 0.0024225139059126377,
      "step": 90
    },
    {
      "epoch": 1.6,
      "learning_rate": 3.7296110958116845e-06,
      "logits/chosen": -0.14412106573581696,
      "logits/rejected": 0.08511891216039658,
      "logps/chosen": -135.07821655273438,
      "logps/rejected": -80.14738464355469,
      "loss": 0.0012,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.006507501006126404,
      "rewards/margins": 0.0025544913951307535,
      "rewards/rejected": 0.003953009843826294,
      "step": 100
    },
    {
      "epoch": 1.6,
      "eval_logits/chosen": -0.0783870667219162,
      "eval_logits/rejected": 0.018332112580537796,
      "eval_logps/chosen": -306.7934875488281,
      "eval_logps/rejected": -278.9591064453125,
      "eval_loss": 0.0017850365256890655,
      "eval_rewards/accuracies": 0.49399998784065247,
      "eval_rewards/chosen": -0.0022730445489287376,
      "eval_rewards/margins": -0.0012333833146840334,
      "eval_rewards/rejected": -0.0010396612342447042,
      "eval_runtime": 412.8211,
      "eval_samples_per_second": 4.845,
      "eval_steps_per_second": 1.211,
      "step": 100
    },
    {
      "epoch": 1.76,
      "learning_rate": 3.4117911628292944e-06,
      "logits/chosen": -0.10137580335140228,
      "logits/rejected": 0.1050504595041275,
      "logps/chosen": -142.63235473632812,
      "logps/rejected": -90.01012420654297,
      "loss": 0.0012,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.006990297231823206,
      "rewards/margins": 0.0029323671478778124,
      "rewards/rejected": 0.00405792985111475,
      "step": 110
    },
    {
      "epoch": 1.92,
      "learning_rate": 3.075905022087675e-06,
      "logits/chosen": -0.20957832038402557,
      "logits/rejected": 0.0603378526866436,
      "logps/chosen": -126.7168960571289,
      "logps/rejected": -94.3608169555664,
      "loss": 0.0011,
      "rewards/accuracies": 0.543749988079071,
      "rewards/chosen": 0.009139289148151875,
      "rewards/margins": 0.004026611801236868,
      "rewards/rejected": 0.005112676881253719,
      "step": 120
    },
    {
      "epoch": 2.08,
      "learning_rate": 2.728607913349464e-06,
      "logits/chosen": -0.14207085967063904,
      "logits/rejected": 0.05692986398935318,
      "logps/chosen": -135.62109375,
      "logps/rejected": -73.44181823730469,
      "loss": 0.0009,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.010310354642570019,
      "rewards/margins": 0.0062768300995230675,
      "rewards/rejected": 0.004033524077385664,
      "step": 130
    },
    {
      "epoch": 2.24,
      "learning_rate": 2.376781173017589e-06,
      "logits/chosen": -0.07452885806560516,
      "logits/rejected": 0.008148658089339733,
      "logps/chosen": -135.99551391601562,
      "logps/rejected": -86.43684387207031,
      "loss": 0.0011,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.00757336150854826,
      "rewards/margins": 0.0026444164104759693,
      "rewards/rejected": 0.004928944632411003,
      "step": 140
    },
    {
      "epoch": 2.4,
      "learning_rate": 2.0273958875043877e-06,
      "logits/chosen": -0.2039242684841156,
      "logits/rejected": 0.03630978241562843,
      "logps/chosen": -122.752685546875,
      "logps/rejected": -77.11092376708984,
      "loss": 0.0011,
      "rewards/accuracies": 0.5687500238418579,
      "rewards/chosen": 0.009455384686589241,
      "rewards/margins": 0.003300786018371582,
      "rewards/rejected": 0.006154598668217659,
      "step": 150
    },
    {
      "epoch": 2.56,
      "learning_rate": 1.6873747682962393e-06,
      "logits/chosen": -0.13353724777698517,
      "logits/rejected": 0.028349420055747032,
      "logps/chosen": -155.1117706298828,
      "logps/rejected": -95.18585968017578,
      "loss": 0.001,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.012836555950343609,
      "rewards/margins": 0.007292959839105606,
      "rewards/rejected": 0.005543597042560577,
      "step": 160
    },
    {
      "epoch": 2.72,
      "learning_rate": 1.363454985517803e-06,
      "logits/chosen": -0.1304449737071991,
      "logits/rejected": 0.12869992852210999,
      "logps/chosen": -150.98422241210938,
      "logps/rejected": -95.78520965576172,
      "loss": 0.0009,
      "rewards/accuracies": 0.606249988079071,
      "rewards/chosen": 0.012216785922646523,
      "rewards/margins": 0.007924379780888557,
      "rewards/rejected": 0.0042924038134515285,
      "step": 170
    },
    {
      "epoch": 2.88,
      "learning_rate": 1.062054677808238e-06,
      "logits/chosen": -0.2121671438217163,
      "logits/rejected": 0.03771474212408066,
      "logps/chosen": -137.01478576660156,
      "logps/rejected": -77.85432434082031,
      "loss": 0.0009,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.015273104421794415,
      "rewards/margins": 0.008995559997856617,
      "rewards/rejected": 0.006277543492615223,
      "step": 180
    },
    {
      "epoch": 3.04,
      "learning_rate": 7.891457834794711e-07,
      "logits/chosen": -0.12489555776119232,
      "logits/rejected": 0.01580502651631832,
      "logps/chosen": -139.43597412109375,
      "logps/rejected": -76.32955932617188,
      "loss": 0.0009,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": 0.012240779586136341,
      "rewards/margins": 0.006818818859755993,
      "rewards/rejected": 0.005421961657702923,
      "step": 190
    },
    {
      "epoch": 3.2,
      "learning_rate": 5.501357126768117e-07,
      "logits/chosen": -0.11466995626688004,
      "logits/rejected": 0.03905560448765755,
      "logps/chosen": -135.41317749023438,
      "logps/rejected": -85.0548324584961,
      "loss": 0.001,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.011853891424834728,
      "rewards/margins": 0.006260283291339874,
      "rewards/rejected": 0.005593608133494854,
      "step": 200
    },
    {
      "epoch": 3.2,
      "eval_logits/chosen": -0.04515805095434189,
      "eval_logits/rejected": 0.05387289822101593,
      "eval_logps/chosen": -307.02294921875,
      "eval_logps/rejected": -279.1457214355469,
      "eval_loss": 0.0018913293024525046,
      "eval_rewards/accuracies": 0.47749999165534973,
      "eval_rewards/chosen": -0.004567304626107216,
      "eval_rewards/margins": -0.001661244546994567,
      "eval_rewards/rejected": -0.002906060079112649,
      "eval_runtime": 413.2351,
      "eval_samples_per_second": 4.84,
      "eval_steps_per_second": 1.21,
      "step": 200
    },
    {
      "epoch": 3.36,
      "learning_rate": 3.4976020508682345e-07,
      "logits/chosen": -0.14853140711784363,
      "logits/rejected": 0.11701055616140366,
      "logps/chosen": -154.20809936523438,
      "logps/rejected": -87.95806884765625,
      "loss": 0.0009,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.014176271855831146,
      "rewards/margins": 0.007768501993268728,
      "rewards/rejected": 0.006407769862562418,
      "step": 210
    },
    {
      "epoch": 3.52,
      "learning_rate": 1.9198949610721273e-07,
      "logits/chosen": -0.16965876519680023,
      "logits/rejected": 0.030850818380713463,
      "logps/chosen": -149.52610778808594,
      "logps/rejected": -90.52479553222656,
      "loss": 0.001,
      "rewards/accuracies": 0.581250011920929,
      "rewards/chosen": 0.011471603065729141,
      "rewards/margins": 0.0050462381914258,
      "rewards/rejected": 0.006425364408642054,
      "step": 220
    },
    {
      "epoch": 3.68,
      "learning_rate": 7.994965069994143e-08,
      "logits/chosen": -0.1932154893875122,
      "logits/rejected": 0.003398424480110407,
      "logps/chosen": -119.58354187011719,
      "logps/rejected": -80.70613098144531,
      "loss": 0.0009,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.013346564956009388,
      "rewards/margins": 0.006659118924289942,
      "rewards/rejected": 0.006687444634735584,
      "step": 230
    },
    {
      "epoch": 3.84,
      "learning_rate": 1.5860623616664183e-08,
      "logits/chosen": -0.054938964545726776,
      "logits/rejected": 0.11343282461166382,
      "logps/chosen": -149.49655151367188,
      "logps/rejected": -88.27520751953125,
      "loss": 0.001,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.010542812757194042,
      "rewards/margins": 0.005555021110922098,
      "rewards/rejected": 0.004987792111933231,
      "step": 240
    },
    {
      "epoch": 3.97,
      "step": 248,
      "total_flos": 0.0,
      "train_loss": 0.00108777264728316,
      "train_runtime": 2657.7329,
      "train_samples_per_second": 1.505,
      "train_steps_per_second": 0.093
    }
  ],
  "logging_steps": 10,
  "max_steps": 248,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}