{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 396,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.007575757575757576,
      "grad_norm": 15.596037687491558,
      "learning_rate": 1.25e-08,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.578125,
      "logps/chosen": -166.0,
      "logps/rejected": -140.0,
      "loss": 0.6914,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.07575757575757576,
      "grad_norm": 15.63845059404282,
      "learning_rate": 1.25e-07,
      "logits/chosen": -1.59375,
      "logits/rejected": -1.546875,
      "logps/chosen": -146.0,
      "logps/rejected": -132.0,
      "loss": 0.6918,
      "rewards/accuracies": 0.1805555522441864,
      "rewards/chosen": -0.006622314453125,
      "rewards/margins": -0.0020904541015625,
      "rewards/rejected": -0.0045166015625,
      "step": 10
    },
    {
      "epoch": 0.15151515151515152,
      "grad_norm": 15.541972149316718,
      "learning_rate": 2.5e-07,
      "logits/chosen": -1.5703125,
      "logits/rejected": -1.5703125,
      "logps/chosen": -135.0,
      "logps/rejected": -139.0,
      "loss": 0.6929,
      "rewards/accuracies": 0.1875,
      "rewards/chosen": -0.0137939453125,
      "rewards/margins": -0.015625,
      "rewards/rejected": 0.00188446044921875,
      "step": 20
    },
    {
      "epoch": 0.22727272727272727,
      "grad_norm": 16.167428018856896,
      "learning_rate": 3.75e-07,
      "logits/chosen": -1.578125,
      "logits/rejected": -1.5625,
      "logps/chosen": -127.5,
      "logps/rejected": -119.0,
      "loss": 0.6891,
      "rewards/accuracies": 0.4124999940395355,
      "rewards/chosen": -0.006256103515625,
      "rewards/margins": 0.010986328125,
      "rewards/rejected": -0.0172119140625,
      "step": 30
    },
    {
      "epoch": 0.30303030303030304,
      "grad_norm": 15.146059864609068,
      "learning_rate": 5e-07,
      "logits/chosen": -1.5859375,
      "logits/rejected": -1.5,
      "logps/chosen": -137.0,
      "logps/rejected": -129.0,
      "loss": 0.6851,
      "rewards/accuracies": 0.2874999940395355,
      "rewards/chosen": -0.03564453125,
      "rewards/margins": -0.00186920166015625,
      "rewards/rejected": -0.033935546875,
      "step": 40
    },
    {
      "epoch": 0.3787878787878788,
      "grad_norm": 16.049039283963904,
      "learning_rate": 4.859550561797752e-07,
      "logits/chosen": -1.6328125,
      "logits/rejected": -1.5546875,
      "logps/chosen": -166.0,
      "logps/rejected": -138.0,
      "loss": 0.6814,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.04931640625,
      "rewards/margins": 0.04736328125,
      "rewards/rejected": -0.0966796875,
      "step": 50
    },
    {
      "epoch": 0.45454545454545453,
      "grad_norm": 16.507555023132472,
      "learning_rate": 4.7191011235955054e-07,
      "logits/chosen": -1.5625,
      "logits/rejected": -1.5,
      "logps/chosen": -146.0,
      "logps/rejected": -131.0,
      "loss": 0.671,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.0888671875,
      "rewards/margins": 0.047119140625,
      "rewards/rejected": -0.1357421875,
      "step": 60
    },
    {
      "epoch": 0.5303030303030303,
      "grad_norm": 15.118251247338051,
      "learning_rate": 4.5786516853932584e-07,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.5390625,
      "logps/chosen": -146.0,
      "logps/rejected": -129.0,
      "loss": 0.6756,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.1279296875,
      "rewards/margins": 0.049072265625,
      "rewards/rejected": -0.1767578125,
      "step": 70
    },
    {
      "epoch": 0.6060606060606061,
      "grad_norm": 14.88927933446616,
      "learning_rate": 4.438202247191011e-07,
      "logits/chosen": -1.546875,
      "logits/rejected": -1.5390625,
      "logps/chosen": -135.0,
      "logps/rejected": -129.0,
      "loss": 0.6703,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.1318359375,
      "rewards/margins": 0.04248046875,
      "rewards/rejected": -0.173828125,
      "step": 80
    },
    {
      "epoch": 0.6818181818181818,
      "grad_norm": 14.799638412304567,
      "learning_rate": 4.297752808988764e-07,
      "logits/chosen": -1.53125,
      "logits/rejected": -1.5703125,
      "logps/chosen": -151.0,
      "logps/rejected": -142.0,
      "loss": 0.6579,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.138671875,
      "rewards/margins": 0.06689453125,
      "rewards/rejected": -0.2060546875,
      "step": 90
    },
    {
      "epoch": 0.7575757575757576,
      "grad_norm": 14.614214826436204,
      "learning_rate": 4.157303370786517e-07,
      "logits/chosen": -1.546875,
      "logits/rejected": -1.5625,
      "logps/chosen": -146.0,
      "logps/rejected": -142.0,
      "loss": 0.6571,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.181640625,
      "rewards/margins": 0.0791015625,
      "rewards/rejected": -0.259765625,
      "step": 100
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 14.469109228087035,
      "learning_rate": 4.0168539325842696e-07,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.5625,
      "logps/chosen": -143.0,
      "logps/rejected": -143.0,
      "loss": 0.6619,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.2353515625,
      "rewards/margins": 0.05029296875,
      "rewards/rejected": -0.28515625,
      "step": 110
    },
    {
      "epoch": 0.9090909090909091,
      "grad_norm": 14.828181252975622,
      "learning_rate": 3.876404494382022e-07,
      "logits/chosen": -1.5,
      "logits/rejected": -1.5078125,
      "logps/chosen": -139.0,
      "logps/rejected": -130.0,
      "loss": 0.6482,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.302734375,
      "rewards/margins": 0.0732421875,
      "rewards/rejected": -0.376953125,
      "step": 120
    },
    {
      "epoch": 0.9848484848484849,
      "grad_norm": 14.708424671919467,
      "learning_rate": 3.735955056179775e-07,
      "logits/chosen": -1.484375,
      "logits/rejected": -1.46875,
      "logps/chosen": -130.0,
      "logps/rejected": -119.0,
      "loss": 0.6459,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.248046875,
      "rewards/margins": 0.1552734375,
      "rewards/rejected": -0.40234375,
      "step": 130
    },
    {
      "epoch": 1.0,
      "eval_logits/chosen": -1.484375,
      "eval_logits/rejected": -1.4765625,
      "eval_logps/chosen": -141.0,
      "eval_logps/rejected": -129.0,
      "eval_loss": 0.6321874856948853,
      "eval_rewards/accuracies": 0.6428571343421936,
      "eval_rewards/chosen": -0.265625,
      "eval_rewards/margins": 0.1767578125,
      "eval_rewards/rejected": -0.443359375,
      "eval_runtime": 12.2418,
      "eval_samples_per_second": 16.337,
      "eval_steps_per_second": 0.572,
      "step": 132
    },
    {
      "epoch": 1.0606060606060606,
      "grad_norm": 12.113429780384491,
      "learning_rate": 3.5955056179775277e-07,
      "logits/chosen": -1.515625,
      "logits/rejected": -1.453125,
      "logps/chosen": -147.0,
      "logps/rejected": -141.0,
      "loss": 0.5898,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.2314453125,
      "rewards/margins": 0.236328125,
      "rewards/rejected": -0.466796875,
      "step": 140
    },
    {
      "epoch": 1.1363636363636362,
      "grad_norm": 11.88094972407373,
      "learning_rate": 3.4550561797752807e-07,
      "logits/chosen": -1.5703125,
      "logits/rejected": -1.515625,
      "logps/chosen": -149.0,
      "logps/rejected": -150.0,
      "loss": 0.5547,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.1630859375,
      "rewards/margins": 0.333984375,
      "rewards/rejected": -0.49609375,
      "step": 150
    },
    {
      "epoch": 1.2121212121212122,
      "grad_norm": 12.356915867599595,
      "learning_rate": 3.314606741573033e-07,
      "logits/chosen": -1.5703125,
      "logits/rejected": -1.5390625,
      "logps/chosen": -142.0,
      "logps/rejected": -130.0,
      "loss": 0.5473,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.2412109375,
      "rewards/margins": 0.369140625,
      "rewards/rejected": -0.609375,
      "step": 160
    },
    {
      "epoch": 1.2878787878787878,
      "grad_norm": 12.507248800803001,
      "learning_rate": 3.1741573033707863e-07,
      "logits/chosen": -1.5078125,
      "logits/rejected": -1.484375,
      "logps/chosen": -143.0,
      "logps/rejected": -131.0,
      "loss": 0.5301,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.33984375,
      "rewards/margins": 0.435546875,
      "rewards/rejected": -0.7734375,
      "step": 170
    },
    {
      "epoch": 1.3636363636363638,
      "grad_norm": 12.914463649706578,
      "learning_rate": 3.0337078651685393e-07,
      "logits/chosen": -1.5234375,
      "logits/rejected": -1.4921875,
      "logps/chosen": -152.0,
      "logps/rejected": -144.0,
      "loss": 0.5235,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.2470703125,
      "rewards/margins": 0.44140625,
      "rewards/rejected": -0.6875,
      "step": 180
    },
    {
      "epoch": 1.4393939393939394,
      "grad_norm": 14.04555120726488,
      "learning_rate": 2.893258426966292e-07,
      "logits/chosen": -1.515625,
      "logits/rejected": -1.5,
      "logps/chosen": -144.0,
      "logps/rejected": -137.0,
      "loss": 0.5319,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.515625,
      "rewards/margins": 0.423828125,
      "rewards/rejected": -0.9375,
      "step": 190
    },
    {
      "epoch": 1.5151515151515151,
      "grad_norm": 12.20920671777009,
      "learning_rate": 2.752808988764045e-07,
      "logits/chosen": -1.5234375,
      "logits/rejected": -1.5390625,
      "logps/chosen": -152.0,
      "logps/rejected": -155.0,
      "loss": 0.5072,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.40625,
      "rewards/margins": 0.5234375,
      "rewards/rejected": -0.9296875,
      "step": 200
    },
    {
      "epoch": 1.5909090909090908,
      "grad_norm": 12.279499640810553,
      "learning_rate": 2.612359550561798e-07,
      "logits/chosen": -1.4765625,
      "logits/rejected": -1.453125,
      "logps/chosen": -138.0,
      "logps/rejected": -137.0,
      "loss": 0.5292,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.447265625,
      "rewards/margins": 0.478515625,
      "rewards/rejected": -0.92578125,
      "step": 210
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 14.135915136852619,
      "learning_rate": 2.4719101123595505e-07,
      "logits/chosen": -1.5,
      "logits/rejected": -1.4609375,
      "logps/chosen": -142.0,
      "logps/rejected": -133.0,
      "loss": 0.5232,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.5625,
      "rewards/margins": 0.37109375,
      "rewards/rejected": -0.93359375,
      "step": 220
    },
    {
      "epoch": 1.7424242424242424,
      "grad_norm": 13.704058332324305,
      "learning_rate": 2.331460674157303e-07,
      "logits/chosen": -1.5546875,
      "logits/rejected": -1.546875,
      "logps/chosen": -147.0,
      "logps/rejected": -145.0,
      "loss": 0.5218,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.4375,
      "rewards/margins": 0.5390625,
      "rewards/rejected": -0.9765625,
      "step": 230
    },
    {
      "epoch": 1.8181818181818183,
      "grad_norm": 12.255502416027877,
      "learning_rate": 2.191011235955056e-07,
      "logits/chosen": -1.5859375,
      "logits/rejected": -1.5546875,
      "logps/chosen": -163.0,
      "logps/rejected": -152.0,
      "loss": 0.4985,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.349609375,
      "rewards/margins": 0.625,
      "rewards/rejected": -0.97265625,
      "step": 240
    },
    {
      "epoch": 1.893939393939394,
      "grad_norm": 13.196400764481162,
      "learning_rate": 2.0505617977528089e-07,
      "logits/chosen": -1.5546875,
      "logits/rejected": -1.453125,
      "logps/chosen": -162.0,
      "logps/rejected": -127.0,
      "loss": 0.4993,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.53125,
      "rewards/margins": 0.65234375,
      "rewards/rejected": -1.1796875,
      "step": 250
    },
    {
      "epoch": 1.9696969696969697,
      "grad_norm": 15.136911810706158,
      "learning_rate": 1.9101123595505617e-07,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.5078125,
      "logps/chosen": -148.0,
      "logps/rejected": -145.0,
      "loss": 0.5393,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.6328125,
      "rewards/margins": 0.46484375,
      "rewards/rejected": -1.09375,
      "step": 260
    },
    {
      "epoch": 2.0,
      "eval_logits/chosen": -1.484375,
      "eval_logits/rejected": -1.46875,
      "eval_logps/chosen": -144.0,
      "eval_logps/rejected": -133.0,
      "eval_loss": 0.6057812571525574,
      "eval_rewards/accuracies": 0.6607142686843872,
      "eval_rewards/chosen": -0.5625,
      "eval_rewards/margins": 0.3515625,
      "eval_rewards/rejected": -0.9140625,
      "eval_runtime": 14.4828,
      "eval_samples_per_second": 13.81,
      "eval_steps_per_second": 0.483,
      "step": 264
    },
    {
      "epoch": 2.0454545454545454,
      "grad_norm": 11.094892248853895,
      "learning_rate": 1.7696629213483144e-07,
      "logits/chosen": -1.5,
      "logits/rejected": -1.4453125,
      "logps/chosen": -148.0,
      "logps/rejected": -142.0,
      "loss": 0.4837,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.58984375,
      "rewards/margins": 0.62890625,
      "rewards/rejected": -1.21875,
      "step": 270
    },
    {
      "epoch": 2.121212121212121,
      "grad_norm": 10.64557888461514,
      "learning_rate": 1.6292134831460675e-07,
      "logits/chosen": -1.5546875,
      "logits/rejected": -1.4765625,
      "logps/chosen": -150.0,
      "logps/rejected": -134.0,
      "loss": 0.4358,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.5078125,
      "rewards/margins": 0.734375,
      "rewards/rejected": -1.2421875,
      "step": 280
    },
    {
      "epoch": 2.196969696969697,
      "grad_norm": 11.580320333983588,
      "learning_rate": 1.4887640449438203e-07,
      "logits/chosen": -1.5625,
      "logits/rejected": -1.5390625,
      "logps/chosen": -165.0,
      "logps/rejected": -160.0,
      "loss": 0.4185,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.546875,
      "rewards/margins": 0.765625,
      "rewards/rejected": -1.3125,
      "step": 290
    },
    {
      "epoch": 2.2727272727272725,
      "grad_norm": 15.056725052177573,
      "learning_rate": 1.3483146067415728e-07,
      "logits/chosen": -1.546875,
      "logits/rejected": -1.5078125,
      "logps/chosen": -158.0,
      "logps/rejected": -149.0,
      "loss": 0.458,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.435546875,
      "rewards/margins": 0.75,
      "rewards/rejected": -1.1875,
      "step": 300
    },
    {
      "epoch": 2.3484848484848486,
      "grad_norm": 12.006511122339738,
      "learning_rate": 1.2078651685393259e-07,
      "logits/chosen": -1.515625,
      "logits/rejected": -1.453125,
      "logps/chosen": -154.0,
      "logps/rejected": -144.0,
      "loss": 0.4472,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.5546875,
      "rewards/margins": 0.65625,
      "rewards/rejected": -1.2109375,
      "step": 310
    },
    {
      "epoch": 2.4242424242424243,
      "grad_norm": 13.746362722926833,
      "learning_rate": 1.0674157303370785e-07,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.5234375,
      "logps/chosen": -140.0,
      "logps/rejected": -145.0,
      "loss": 0.4728,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.58203125,
      "rewards/margins": 0.6015625,
      "rewards/rejected": -1.1796875,
      "step": 320
    },
    {
      "epoch": 2.5,
      "grad_norm": 10.459882513773048,
      "learning_rate": 9.269662921348314e-08,
      "logits/chosen": -1.515625,
      "logits/rejected": -1.46875,
      "logps/chosen": -153.0,
      "logps/rejected": -149.0,
      "loss": 0.4405,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -0.5703125,
      "rewards/margins": 0.79296875,
      "rewards/rejected": -1.3671875,
      "step": 330
    },
    {
      "epoch": 2.5757575757575757,
      "grad_norm": 13.750139706707387,
      "learning_rate": 7.865168539325842e-08,
      "logits/chosen": -1.46875,
      "logits/rejected": -1.5,
      "logps/chosen": -147.0,
      "logps/rejected": -143.0,
      "loss": 0.4422,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -0.515625,
      "rewards/margins": 0.69140625,
      "rewards/rejected": -1.2109375,
      "step": 340
    },
    {
      "epoch": 2.6515151515151514,
      "grad_norm": 10.53559766204164,
      "learning_rate": 6.460674157303371e-08,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.4921875,
      "logps/chosen": -155.0,
      "logps/rejected": -142.0,
      "loss": 0.4407,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.54296875,
      "rewards/margins": 0.7109375,
      "rewards/rejected": -1.25,
      "step": 350
    },
    {
      "epoch": 2.7272727272727275,
      "grad_norm": 10.932011907954834,
      "learning_rate": 5.056179775280899e-08,
      "logits/chosen": -1.5078125,
      "logits/rejected": -1.453125,
      "logps/chosen": -142.0,
      "logps/rejected": -139.0,
      "loss": 0.4395,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": -0.57421875,
      "rewards/margins": 0.87890625,
      "rewards/rejected": -1.453125,
      "step": 360
    },
    {
      "epoch": 2.8030303030303028,
      "grad_norm": 10.478220370160118,
      "learning_rate": 3.6516853932584266e-08,
      "logits/chosen": -1.53125,
      "logits/rejected": -1.5,
      "logps/chosen": -144.0,
      "logps/rejected": -142.0,
      "loss": 0.447,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.53515625,
      "rewards/margins": 0.71484375,
      "rewards/rejected": -1.25,
      "step": 370
    },
    {
      "epoch": 2.878787878787879,
      "grad_norm": 11.077008754645805,
      "learning_rate": 2.2471910112359548e-08,
      "logits/chosen": -1.546875,
      "logits/rejected": -1.484375,
      "logps/chosen": -150.0,
      "logps/rejected": -140.0,
      "loss": 0.4625,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.5234375,
      "rewards/margins": 0.76953125,
      "rewards/rejected": -1.2890625,
      "step": 380
    },
    {
      "epoch": 2.9545454545454546,
      "grad_norm": 10.906822523238372,
      "learning_rate": 8.42696629213483e-09,
      "logits/chosen": -1.5390625,
      "logits/rejected": -1.5390625,
      "logps/chosen": -156.0,
      "logps/rejected": -146.0,
      "loss": 0.4113,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.439453125,
      "rewards/margins": 0.82421875,
      "rewards/rejected": -1.265625,
      "step": 390
    },
    {
      "epoch": 3.0,
      "eval_logits/chosen": -1.4765625,
      "eval_logits/rejected": -1.46875,
      "eval_logps/chosen": -144.0,
      "eval_logps/rejected": -133.0,
      "eval_loss": 0.6071093678474426,
      "eval_rewards/accuracies": 0.6607142686843872,
      "eval_rewards/chosen": -0.490234375,
      "eval_rewards/margins": 0.396484375,
      "eval_rewards/rejected": -0.88671875,
      "eval_runtime": 14.7701,
      "eval_samples_per_second": 13.541,
      "eval_steps_per_second": 0.474,
      "step": 396
    },
    {
      "epoch": 3.0,
      "step": 396,
      "total_flos": 0.0,
      "train_loss": 0.5475334398674242,
      "train_runtime": 2411.1864,
      "train_samples_per_second": 5.224,
      "train_steps_per_second": 0.164
    }
  ],
  "logging_steps": 10,
  "max_steps": 396,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 8,
  "trial_name": null,
  "trial_params": null
}