{
  "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": 14.92016418615593,
      "learning_rate": 3.125e-08,
      "logits/chosen": -2.845900058746338,
      "logits/rejected": -2.8750576972961426,
      "logps/chosen": -339.33447265625,
      "logps/pi_response": -84.25662994384766,
      "logps/ref_response": -84.25662994384766,
      "logps/rejected": -136.8517303466797,
      "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": 10.888267743448523,
      "learning_rate": 3.1249999999999997e-07,
      "logits/chosen": -2.778717041015625,
      "logits/rejected": -2.757580280303955,
      "logps/chosen": -226.78994750976562,
      "logps/pi_response": -69.23085021972656,
      "logps/ref_response": -68.98346710205078,
      "logps/rejected": -125.45468139648438,
      "loss": 0.692,
      "rewards/accuracies": 0.4930555522441864,
      "rewards/chosen": 0.0016165798297151923,
      "rewards/margins": 0.003002046374604106,
      "rewards/rejected": -0.0013854664284735918,
      "step": 10
    },
    {
      "epoch": 0.12539184952978055,
      "grad_norm": 9.519679472408939,
      "learning_rate": 4.990353313429303e-07,
      "logits/chosen": -2.7556426525115967,
      "logits/rejected": -2.740095615386963,
      "logps/chosen": -211.00942993164062,
      "logps/pi_response": -76.92619323730469,
      "logps/ref_response": -70.64810943603516,
      "logps/rejected": -106.86106872558594,
      "loss": 0.6637,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.006867959164083004,
      "rewards/margins": 0.04612569510936737,
      "rewards/rejected": -0.03925773501396179,
      "step": 20
    },
    {
      "epoch": 0.18808777429467086,
      "grad_norm": 7.098301149383853,
      "learning_rate": 4.882681251368548e-07,
      "logits/chosen": -2.6341042518615723,
      "logits/rejected": -2.6018269062042236,
      "logps/chosen": -237.74984741210938,
      "logps/pi_response": -98.3823013305664,
      "logps/ref_response": -65.06692504882812,
      "logps/rejected": -122.42607116699219,
      "loss": 0.618,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.08040433377027512,
      "rewards/margins": 0.18929488956928253,
      "rewards/rejected": -0.26969921588897705,
      "step": 30
    },
    {
      "epoch": 0.2507836990595611,
      "grad_norm": 9.130640915503477,
      "learning_rate": 4.6604720940421207e-07,
      "logits/chosen": -2.528998851776123,
      "logits/rejected": -2.5069429874420166,
      "logps/chosen": -276.9578857421875,
      "logps/pi_response": -141.6453857421875,
      "logps/ref_response": -69.38419342041016,
      "logps/rejected": -177.18133544921875,
      "loss": 0.5824,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.23818793892860413,
      "rewards/margins": 0.40282946825027466,
      "rewards/rejected": -0.6410173177719116,
      "step": 40
    },
    {
      "epoch": 0.31347962382445144,
      "grad_norm": 8.866163012880673,
      "learning_rate": 4.3344075855595097e-07,
      "logits/chosen": -2.5575942993164062,
      "logits/rejected": -2.532439708709717,
      "logps/chosen": -289.70233154296875,
      "logps/pi_response": -166.5010223388672,
      "logps/ref_response": -67.87925720214844,
      "logps/rejected": -180.68893432617188,
      "loss": 0.5712,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.36984091997146606,
      "rewards/margins": 0.5151764154434204,
      "rewards/rejected": -0.8850172758102417,
      "step": 50
    },
    {
      "epoch": 0.3761755485893417,
      "grad_norm": 13.516613178785937,
      "learning_rate": 3.920161866827889e-07,
      "logits/chosen": -2.553635835647583,
      "logits/rejected": -2.5055148601531982,
      "logps/chosen": -218.26220703125,
      "logps/pi_response": -163.2755584716797,
      "logps/ref_response": -64.52303314208984,
      "logps/rejected": -200.37921142578125,
      "loss": 0.5567,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.3870890140533447,
      "rewards/margins": 0.43599042296409607,
      "rewards/rejected": -0.8230794668197632,
      "step": 60
    },
    {
      "epoch": 0.438871473354232,
      "grad_norm": 11.23129211624929,
      "learning_rate": 3.4376480090239047e-07,
      "logits/chosen": -2.585047721862793,
      "logits/rejected": -2.546762228012085,
      "logps/chosen": -258.9521789550781,
      "logps/pi_response": -190.20260620117188,
      "logps/ref_response": -68.9866714477539,
      "logps/rejected": -210.7947235107422,
      "loss": 0.5237,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.46116265654563904,
      "rewards/margins": 0.6260020136833191,
      "rewards/rejected": -1.0871646404266357,
      "step": 70
    },
    {
      "epoch": 0.5015673981191222,
      "grad_norm": 14.241762559238571,
      "learning_rate": 2.910060778827554e-07,
      "logits/chosen": -2.6193995475769043,
      "logits/rejected": -2.5861072540283203,
      "logps/chosen": -253.4519500732422,
      "logps/pi_response": -183.41494750976562,
      "logps/ref_response": -72.47793579101562,
      "logps/rejected": -203.38510131835938,
      "loss": 0.5301,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.41766557097435,
      "rewards/margins": 0.5782700777053833,
      "rewards/rejected": -0.9959356188774109,
      "step": 80
    },
    {
      "epoch": 0.5642633228840125,
      "grad_norm": 12.061582086504794,
      "learning_rate": 2.3627616503391812e-07,
      "logits/chosen": -2.5625717639923096,
      "logits/rejected": -2.510937213897705,
      "logps/chosen": -296.7856140136719,
      "logps/pi_response": -197.53363037109375,
      "logps/ref_response": -69.97444152832031,
      "logps/rejected": -216.2652130126953,
      "loss": 0.5138,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -0.5185791850090027,
      "rewards/margins": 0.6481128931045532,
      "rewards/rejected": -1.1666921377182007,
      "step": 90
    },
    {
      "epoch": 0.6269592476489029,
      "grad_norm": 12.761867748586539,
      "learning_rate": 1.8220596619089573e-07,
      "logits/chosen": -2.5809009075164795,
      "logits/rejected": -2.5340776443481445,
      "logps/chosen": -302.91339111328125,
      "logps/pi_response": -206.8216552734375,
      "logps/ref_response": -67.3431625366211,
      "logps/rejected": -229.6073455810547,
      "loss": 0.5151,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.5213590860366821,
      "rewards/margins": 0.7751049399375916,
      "rewards/rejected": -1.2964640855789185,
      "step": 100
    },
    {
      "epoch": 0.6896551724137931,
      "grad_norm": 14.424999759406585,
      "learning_rate": 1.3139467229135998e-07,
      "logits/chosen": -2.5221171379089355,
      "logits/rejected": -2.4842729568481445,
      "logps/chosen": -236.73046875,
      "logps/pi_response": -175.51455688476562,
      "logps/ref_response": -60.56214141845703,
      "logps/rejected": -212.31582641601562,
      "loss": 0.5032,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": -0.48990482091903687,
      "rewards/margins": 0.5828049778938293,
      "rewards/rejected": -1.0727096796035767,
      "step": 110
    },
    {
      "epoch": 0.7523510971786834,
      "grad_norm": 45.50922725986238,
      "learning_rate": 8.628481651367875e-08,
      "logits/chosen": -2.5453531742095947,
      "logits/rejected": -2.499166488647461,
      "logps/chosen": -265.3163146972656,
      "logps/pi_response": -199.43765258789062,
      "logps/ref_response": -63.73808670043945,
      "logps/rejected": -221.1977996826172,
      "loss": 0.4952,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.47311264276504517,
      "rewards/margins": 0.8587131500244141,
      "rewards/rejected": -1.331825852394104,
      "step": 120
    },
    {
      "epoch": 0.8150470219435737,
      "grad_norm": 14.647279184181341,
      "learning_rate": 4.904486005914027e-08,
      "logits/chosen": -2.524919033050537,
      "logits/rejected": -2.497668743133545,
      "logps/chosen": -285.549072265625,
      "logps/pi_response": -207.68014526367188,
      "logps/ref_response": -70.44617462158203,
      "logps/rejected": -234.6571807861328,
      "loss": 0.4974,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.4377771019935608,
      "rewards/margins": 0.8497930765151978,
      "rewards/rejected": -1.2875702381134033,
      "step": 130
    },
    {
      "epoch": 0.877742946708464,
      "grad_norm": 18.06269511779979,
      "learning_rate": 2.1464952759020856e-08,
      "logits/chosen": -2.533571481704712,
      "logits/rejected": -2.4829599857330322,
      "logps/chosen": -296.2190246582031,
      "logps/pi_response": -209.37374877929688,
      "logps/ref_response": -70.29788208007812,
      "logps/rejected": -232.8479461669922,
      "loss": 0.5001,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.581600546836853,
      "rewards/margins": 0.6965141296386719,
      "rewards/rejected": -1.2781145572662354,
      "step": 140
    },
    {
      "epoch": 0.9404388714733543,
      "grad_norm": 15.171255483252263,
      "learning_rate": 4.8708793644441086e-09,
      "logits/chosen": -2.5554707050323486,
      "logits/rejected": -2.50273060798645,
      "logps/chosen": -289.8814392089844,
      "logps/pi_response": -220.58633422851562,
      "logps/ref_response": -73.47826385498047,
      "logps/rejected": -257.1234130859375,
      "loss": 0.5001,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.5721909403800964,
      "rewards/margins": 0.8123014569282532,
      "rewards/rejected": -1.3844925165176392,
      "step": 150
    },
    {
      "epoch": 0.9968652037617555,
      "step": 159,
      "total_flos": 0.0,
      "train_loss": 0.5464447999150498,
      "train_runtime": 3432.1011,
      "train_samples_per_second": 5.937,
      "train_steps_per_second": 0.046
    }
  ],
  "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
}