{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9968652037617555,
  "eval_steps": 500,
  "global_step": 159,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.006269592476489028,
      "grad_norm": 48.727538993129954,
      "learning_rate": 2.5e-08,
      "logits/chosen": -1.7884750366210938,
      "logits/rejected": -1.7773182392120361,
      "logps/chosen": -178.7047882080078,
      "logps/pi_response": -149.36985778808594,
      "logps/ref_response": -149.36985778808594,
      "logps/rejected": -327.8221435546875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.06269592476489028,
      "grad_norm": 44.27733622419657,
      "learning_rate": 2.5e-07,
      "logits/chosen": -1.925902009010315,
      "logits/rejected": -1.8669484853744507,
      "logps/chosen": -259.4127197265625,
      "logps/pi_response": -168.5211639404297,
      "logps/ref_response": -168.7109375,
      "logps/rejected": -363.8843994140625,
      "loss": 0.6877,
      "rewards/accuracies": 0.5208333134651184,
      "rewards/chosen": -0.007133293431252241,
      "rewards/margins": 0.008144436404109001,
      "rewards/rejected": -0.01527773030102253,
      "step": 10
    },
    {
      "epoch": 0.12539184952978055,
      "grad_norm": 77.06299399731483,
      "learning_rate": 3.9922826507434426e-07,
      "logits/chosen": -1.7760452032089233,
      "logits/rejected": -1.7109419107437134,
      "logps/chosen": -287.03887939453125,
      "logps/pi_response": -184.39227294921875,
      "logps/ref_response": -179.93296813964844,
      "logps/rejected": -452.4369201660156,
      "loss": 0.6483,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.2718208432197571,
      "rewards/margins": 0.36123865842819214,
      "rewards/rejected": -0.6330594420433044,
      "step": 20
    },
    {
      "epoch": 0.18808777429467086,
      "grad_norm": 37.12655893399871,
      "learning_rate": 3.9061450010948387e-07,
      "logits/chosen": -1.6445729732513428,
      "logits/rejected": -1.5510730743408203,
      "logps/chosen": -272.7137756347656,
      "logps/pi_response": -159.47608947753906,
      "logps/ref_response": -154.75852966308594,
      "logps/rejected": -406.7279968261719,
      "loss": 0.6095,
      "rewards/accuracies": 0.59375,
      "rewards/chosen": -0.29393836855888367,
      "rewards/margins": 0.42385944724082947,
      "rewards/rejected": -0.7177978157997131,
      "step": 30
    },
    {
      "epoch": 0.2507836990595611,
      "grad_norm": 43.17045113037534,
      "learning_rate": 3.7283776752336964e-07,
      "logits/chosen": -1.585517168045044,
      "logits/rejected": -1.5017468929290771,
      "logps/chosen": -241.1661376953125,
      "logps/pi_response": -170.72608947753906,
      "logps/ref_response": -165.39320373535156,
      "logps/rejected": -458.1199645996094,
      "loss": 0.5888,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.12789960205554962,
      "rewards/margins": 0.4231465756893158,
      "rewards/rejected": -0.5510461330413818,
      "step": 40
    },
    {
      "epoch": 0.31347962382445144,
      "grad_norm": 56.15552525880845,
      "learning_rate": 3.4675260684476076e-07,
      "logits/chosen": -1.31333327293396,
      "logits/rejected": -1.1443239450454712,
      "logps/chosen": -280.60247802734375,
      "logps/pi_response": -187.87725830078125,
      "logps/ref_response": -162.92739868164062,
      "logps/rejected": -448.25238037109375,
      "loss": 0.56,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": -0.38959115743637085,
      "rewards/margins": 0.5290592908859253,
      "rewards/rejected": -0.9186504483222961,
      "step": 50
    },
    {
      "epoch": 0.3761755485893417,
      "grad_norm": 59.31824188989313,
      "learning_rate": 3.1361294934623114e-07,
      "logits/chosen": -0.8963761329650879,
      "logits/rejected": -0.7264960408210754,
      "logps/chosen": -326.63250732421875,
      "logps/pi_response": -210.930419921875,
      "logps/ref_response": -159.56680297851562,
      "logps/rejected": -436.368408203125,
      "loss": 0.5461,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.676125168800354,
      "rewards/margins": 0.5084071755409241,
      "rewards/rejected": -1.1845324039459229,
      "step": 60
    },
    {
      "epoch": 0.438871473354232,
      "grad_norm": 48.22413613389395,
      "learning_rate": 2.750118407219124e-07,
      "logits/chosen": -0.902199923992157,
      "logits/rejected": -0.6483638882637024,
      "logps/chosen": -299.6310729980469,
      "logps/pi_response": -208.26123046875,
      "logps/ref_response": -157.38909912109375,
      "logps/rejected": -434.2081604003906,
      "loss": 0.5688,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.5099861025810242,
      "rewards/margins": 0.5872552394866943,
      "rewards/rejected": -1.0972412824630737,
      "step": 70
    },
    {
      "epoch": 0.5015673981191222,
      "grad_norm": 45.59894096049591,
      "learning_rate": 2.3280486230620434e-07,
      "logits/chosen": -0.756037175655365,
      "logits/rejected": -0.477339506149292,
      "logps/chosen": -285.1630554199219,
      "logps/pi_response": -204.59786987304688,
      "logps/ref_response": -150.49502563476562,
      "logps/rejected": -451.710693359375,
      "loss": 0.531,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.5814803242683411,
      "rewards/margins": 0.663509726524353,
      "rewards/rejected": -1.2449901103973389,
      "step": 80
    },
    {
      "epoch": 0.5642633228840125,
      "grad_norm": 51.60912561342815,
      "learning_rate": 1.890209320271345e-07,
      "logits/chosen": -0.6508065462112427,
      "logits/rejected": -0.365335613489151,
      "logps/chosen": -306.8770751953125,
      "logps/pi_response": -212.5884246826172,
      "logps/ref_response": -147.56716918945312,
      "logps/rejected": -485.7850036621094,
      "loss": 0.4977,
      "rewards/accuracies": 0.65625,
      "rewards/chosen": -0.6989418864250183,
      "rewards/margins": 0.8249226808547974,
      "rewards/rejected": -1.523864507675171,
      "step": 90
    },
    {
      "epoch": 0.6269592476489029,
      "grad_norm": 47.512990245630185,
      "learning_rate": 1.457647729527166e-07,
      "logits/chosen": -0.5415297746658325,
      "logits/rejected": -0.25003066658973694,
      "logps/chosen": -314.3260192871094,
      "logps/pi_response": -253.16586303710938,
      "logps/ref_response": -160.88829040527344,
      "logps/rejected": -611.5823974609375,
      "loss": 0.5027,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.8261867761611938,
      "rewards/margins": 1.2384675741195679,
      "rewards/rejected": -2.06465482711792,
      "step": 100
    },
    {
      "epoch": 0.6896551724137931,
      "grad_norm": 40.671499365551824,
      "learning_rate": 1.0511573783308798e-07,
      "logits/chosen": -0.45151621103286743,
      "logits/rejected": -0.2101474106311798,
      "logps/chosen": -330.63751220703125,
      "logps/pi_response": -239.79678344726562,
      "logps/ref_response": -149.11666870117188,
      "logps/rejected": -537.8972778320312,
      "loss": 0.4933,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.9507951736450195,
      "rewards/margins": 0.9308554530143738,
      "rewards/rejected": -1.881650686264038,
      "step": 110
    },
    {
      "epoch": 0.7523510971786834,
      "grad_norm": 49.168767905757875,
      "learning_rate": 6.9027853210943e-08,
      "logits/chosen": -0.5997277498245239,
      "logits/rejected": -0.22453825175762177,
      "logps/chosen": -349.99212646484375,
      "logps/pi_response": -280.00726318359375,
      "logps/ref_response": -178.89942932128906,
      "logps/rejected": -565.8890380859375,
      "loss": 0.5114,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.9609189033508301,
      "rewards/margins": 0.9585698843002319,
      "rewards/rejected": -1.9194889068603516,
      "step": 120
    },
    {
      "epoch": 0.8150470219435737,
      "grad_norm": 49.00943764848692,
      "learning_rate": 3.9235888047312215e-08,
      "logits/chosen": -0.6853595972061157,
      "logits/rejected": -0.2808631658554077,
      "logps/chosen": -327.33953857421875,
      "logps/pi_response": -256.1371154785156,
      "logps/ref_response": -166.0455780029297,
      "logps/rejected": -572.0056762695312,
      "loss": 0.5109,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.8519734144210815,
      "rewards/margins": 1.0642752647399902,
      "rewards/rejected": -1.9162486791610718,
      "step": 130
    },
    {
      "epoch": 0.877742946708464,
      "grad_norm": 65.70628626767385,
      "learning_rate": 1.7171962207216682e-08,
      "logits/chosen": -0.5600570440292358,
      "logits/rejected": -0.11410878598690033,
      "logps/chosen": -324.69439697265625,
      "logps/pi_response": -264.94036865234375,
      "logps/ref_response": -165.58712768554688,
      "logps/rejected": -562.1802978515625,
      "loss": 0.5045,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.8904104232788086,
      "rewards/margins": 0.9812215566635132,
      "rewards/rejected": -1.8716322183609009,
      "step": 140
    },
    {
      "epoch": 0.9404388714733543,
      "grad_norm": 44.83563629682317,
      "learning_rate": 3.8967034915552865e-09,
      "logits/chosen": -0.47401466965675354,
      "logits/rejected": -0.1014849916100502,
      "logps/chosen": -298.55377197265625,
      "logps/pi_response": -241.1999969482422,
      "logps/ref_response": -146.22021484375,
      "logps/rejected": -519.8775634765625,
      "loss": 0.5097,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.8489562273025513,
      "rewards/margins": 1.053095817565918,
      "rewards/rejected": -1.9020519256591797,
      "step": 150
    },
    {
      "epoch": 0.9968652037617555,
      "step": 159,
      "total_flos": 0.0,
      "train_loss": 0.5458380771133134,
      "train_runtime": 3385.2715,
      "train_samples_per_second": 6.02,
      "train_steps_per_second": 0.047
    }
  ],
  "logging_steps": 10,
  "max_steps": 159,
  "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
}