{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": -1.7278180122375488,
      "logits/rejected": -1.7377450466156006,
      "logps/chosen": -29.553977966308594,
      "logps/rejected": -42.813133239746094,
      "loss": 25.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": -1.86653733253479,
      "logits/rejected": -1.8708631992340088,
      "logps/chosen": -36.98527526855469,
      "logps/rejected": -33.67702865600586,
      "loss": 24.351,
      "rewards/accuracies": 0.5555555820465088,
      "rewards/chosen": 0.0021321007516235113,
      "rewards/margins": 0.006827179808169603,
      "rewards/rejected": -0.004695079289376736,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": -1.9982036352157593,
      "logits/rejected": -2.0008480548858643,
      "logps/chosen": -29.6201114654541,
      "logps/rejected": -29.028329849243164,
      "loss": 25.1673,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": 0.002208721823990345,
      "rewards/margins": -0.001302852644585073,
      "rewards/rejected": 0.00351157458499074,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": -1.9204835891723633,
      "logits/rejected": -1.917790174484253,
      "logps/chosen": -31.419384002685547,
      "logps/rejected": -33.234046936035156,
      "loss": 24.926,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.00032898251083679497,
      "rewards/margins": 0.001229552668519318,
      "rewards/rejected": -0.0015585350338369608,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": -2.0177228450775146,
      "logits/rejected": -2.008981227874756,
      "logps/chosen": -32.599449157714844,
      "logps/rejected": -32.504905700683594,
      "loss": 25.3506,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.0022641182877123356,
      "rewards/margins": -0.0030405251309275627,
      "rewards/rejected": 0.0007764073088765144,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": -1.8622967004776,
      "logits/rejected": -1.8515253067016602,
      "logps/chosen": -33.548858642578125,
      "logps/rejected": -35.451934814453125,
      "loss": 24.9661,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.0008619143627583981,
      "rewards/margins": 0.0008210102096199989,
      "rewards/rejected": 4.09040367230773e-05,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": -1.941347360610962,
      "logits/rejected": -1.9432964324951172,
      "logps/chosen": -32.533042907714844,
      "logps/rejected": -33.22543716430664,
      "loss": 23.9783,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.006714025977998972,
      "rewards/margins": 0.012119540013372898,
      "rewards/rejected": -0.005405513569712639,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": -2.0721096992492676,
      "logits/rejected": -2.077080249786377,
      "logps/chosen": -33.98393249511719,
      "logps/rejected": -36.645912170410156,
      "loss": 24.3657,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.0004553424078039825,
      "rewards/margins": 0.008159191347658634,
      "rewards/rejected": -0.00770384818315506,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": -1.9330637454986572,
      "logits/rejected": -1.9362014532089233,
      "logps/chosen": -34.31563949584961,
      "logps/rejected": -34.64789581298828,
      "loss": 23.5462,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.011217002756893635,
      "rewards/margins": 0.01676579937338829,
      "rewards/rejected": -0.005548796150833368,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": -1.9414646625518799,
      "logits/rejected": -1.9459855556488037,
      "logps/chosen": -32.386444091796875,
      "logps/rejected": -32.36344528198242,
      "loss": 24.298,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.008097763173282146,
      "rewards/margins": 0.008551515638828278,
      "rewards/rejected": -0.0004537526401691139,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": -2.038987636566162,
      "logits/rejected": -2.03700590133667,
      "logps/chosen": -32.161869049072266,
      "logps/rejected": -31.31038475036621,
      "loss": 23.9156,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.007896887138485909,
      "rewards/margins": 0.012314335443079472,
      "rewards/rejected": -0.004417449701577425,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.2337892055511475,
      "eval_logits/rejected": -2.228935956954956,
      "eval_logps/chosen": -34.049957275390625,
      "eval_logps/rejected": -37.550052642822266,
      "eval_loss": 24.932252883911133,
      "eval_rewards/accuracies": 0.5282392501831055,
      "eval_rewards/chosen": -0.001540520810522139,
      "eval_rewards/margins": 0.0018024941673502326,
      "eval_rewards/rejected": -0.0033430152107030153,
      "eval_runtime": 145.9828,
      "eval_samples_per_second": 2.35,
      "eval_steps_per_second": 0.295,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": -1.9933013916015625,
      "logits/rejected": -1.9909241199493408,
      "logps/chosen": -33.13324737548828,
      "logps/rejected": -34.06130599975586,
      "loss": 24.1131,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.011056952178478241,
      "rewards/margins": 0.01425047405064106,
      "rewards/rejected": -0.003193522337824106,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": -2.0050952434539795,
      "logits/rejected": -1.9967548847198486,
      "logps/chosen": -32.3184814453125,
      "logps/rejected": -32.16117477416992,
      "loss": 23.9805,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.012642547488212585,
      "rewards/margins": 0.012609380297362804,
      "rewards/rejected": 3.316625952720642e-05,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": -2.033102512359619,
      "logits/rejected": -2.025144100189209,
      "logps/chosen": -30.341167449951172,
      "logps/rejected": -32.081993103027344,
      "loss": 23.735,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.013454675674438477,
      "rewards/margins": 0.016353728249669075,
      "rewards/rejected": -0.0028990507125854492,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": -1.9635530710220337,
      "logits/rejected": -1.9737846851348877,
      "logps/chosen": -31.221776962280273,
      "logps/rejected": -32.5506477355957,
      "loss": 23.1662,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.019788404926657677,
      "rewards/margins": 0.021364931017160416,
      "rewards/rejected": -0.0015765223652124405,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": -1.875128984451294,
      "logits/rejected": -1.876308798789978,
      "logps/chosen": -33.92096710205078,
      "logps/rejected": -34.793277740478516,
      "loss": 22.4884,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.027172502130270004,
      "rewards/margins": 0.030774494633078575,
      "rewards/rejected": -0.00360199436545372,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": -1.9267343282699585,
      "logits/rejected": -1.9233125448226929,
      "logps/chosen": -36.03553009033203,
      "logps/rejected": -32.72822570800781,
      "loss": 23.8244,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.013614905066788197,
      "rewards/margins": 0.014162260107696056,
      "rewards/rejected": -0.0005473563214763999,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": -2.0270447731018066,
      "logits/rejected": -2.0197136402130127,
      "logps/chosen": -33.48883819580078,
      "logps/rejected": -31.42250633239746,
      "loss": 21.7573,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.030902724713087082,
      "rewards/margins": 0.037968240678310394,
      "rewards/rejected": -0.007065513636916876,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": -2.0334360599517822,
      "logits/rejected": -2.038677215576172,
      "logps/chosen": -32.235355377197266,
      "logps/rejected": -32.45377731323242,
      "loss": 22.5681,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.030943090096116066,
      "rewards/margins": 0.027806812897324562,
      "rewards/rejected": 0.0031362746376544237,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": -2.034454345703125,
      "logits/rejected": -2.03169846534729,
      "logps/chosen": -31.2611141204834,
      "logps/rejected": -31.35662841796875,
      "loss": 22.7646,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.022784242406487465,
      "rewards/margins": 0.026625871658325195,
      "rewards/rejected": -0.0038416311144828796,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": -1.9043190479278564,
      "logits/rejected": -1.9089492559432983,
      "logps/chosen": -31.331798553466797,
      "logps/rejected": -32.8171501159668,
      "loss": 22.4879,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.027448315173387527,
      "rewards/margins": 0.030070018023252487,
      "rewards/rejected": -0.002621703315526247,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.2306509017944336,
      "eval_logits/rejected": -2.2258191108703613,
      "eval_logps/chosen": -34.05928039550781,
      "eval_logps/rejected": -37.57795333862305,
      "eval_loss": 24.79918098449707,
      "eval_rewards/accuracies": 0.5598006844520569,
      "eval_rewards/chosen": -0.0024726546835154295,
      "eval_rewards/margins": 0.0036605680361390114,
      "eval_rewards/rejected": -0.006133222486823797,
      "eval_runtime": 145.6285,
      "eval_samples_per_second": 2.355,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": -2.0171916484832764,
      "logits/rejected": -2.027843952178955,
      "logps/chosen": -31.748519897460938,
      "logps/rejected": -33.94535827636719,
      "loss": 22.396,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.022278117015957832,
      "rewards/margins": 0.030543917790055275,
      "rewards/rejected": -0.008265801705420017,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": -1.9094616174697876,
      "logits/rejected": -1.9241745471954346,
      "logps/chosen": -29.818714141845703,
      "logps/rejected": -31.6313533782959,
      "loss": 21.8415,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.029282480478286743,
      "rewards/margins": 0.0366000160574913,
      "rewards/rejected": -0.007317539304494858,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": -1.9662542343139648,
      "logits/rejected": -1.9702177047729492,
      "logps/chosen": -33.06914138793945,
      "logps/rejected": -31.627248764038086,
      "loss": 21.4931,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.03483276441693306,
      "rewards/margins": 0.042627494782209396,
      "rewards/rejected": -0.007794731762260199,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": -1.964835524559021,
      "logits/rejected": -1.9430347681045532,
      "logps/chosen": -33.808719635009766,
      "logps/rejected": -35.10743713378906,
      "loss": 20.8354,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.034502916038036346,
      "rewards/margins": 0.05016731098294258,
      "rewards/rejected": -0.015664398670196533,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": -2.0063602924346924,
      "logits/rejected": -2.0030438899993896,
      "logps/chosen": -32.721397399902344,
      "logps/rejected": -36.288963317871094,
      "loss": 22.7702,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.020366890355944633,
      "rewards/margins": 0.02677120640873909,
      "rewards/rejected": -0.006404316984117031,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": -1.8736040592193604,
      "logits/rejected": -1.8712146282196045,
      "logps/chosen": -33.984458923339844,
      "logps/rejected": -35.55829620361328,
      "loss": 22.8113,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.020393703132867813,
      "rewards/margins": 0.02576116845011711,
      "rewards/rejected": -0.0053674690425395966,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": -1.8587570190429688,
      "logits/rejected": -1.8563568592071533,
      "logps/chosen": -34.19550704956055,
      "logps/rejected": -31.841039657592773,
      "loss": 22.7599,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.020208846777677536,
      "rewards/margins": 0.027179840952157974,
      "rewards/rejected": -0.00697099044919014,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": -1.9618046283721924,
      "logits/rejected": -1.9513158798217773,
      "logps/chosen": -34.9988899230957,
      "logps/rejected": -31.8876895904541,
      "loss": 21.4726,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.03577522560954094,
      "rewards/margins": 0.040445104241371155,
      "rewards/rejected": -0.0046698772348463535,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": -2.057340621948242,
      "logits/rejected": -2.042445659637451,
      "logps/chosen": -30.718120574951172,
      "logps/rejected": -32.65331268310547,
      "loss": 23.4433,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.020600156858563423,
      "rewards/margins": 0.020748335868120193,
      "rewards/rejected": -0.00014818087220191956,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": -1.928430199623108,
      "logits/rejected": -1.9259220361709595,
      "logps/chosen": -32.38062286376953,
      "logps/rejected": -30.925670623779297,
      "loss": 19.8329,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.05464861914515495,
      "rewards/margins": 0.06553500890731812,
      "rewards/rejected": -0.010886380448937416,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.227757215499878,
      "eval_logits/rejected": -2.222930908203125,
      "eval_logps/chosen": -34.10100555419922,
      "eval_logps/rejected": -37.605979919433594,
      "eval_loss": 24.95565414428711,
      "eval_rewards/accuracies": 0.5215947031974792,
      "eval_rewards/chosen": -0.006645896937698126,
      "eval_rewards/margins": 0.0022900840267539024,
      "eval_rewards/rejected": -0.008935980498790741,
      "eval_runtime": 145.783,
      "eval_samples_per_second": 2.353,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": -1.9124844074249268,
      "logits/rejected": -1.9092193841934204,
      "logps/chosen": -31.315048217773438,
      "logps/rejected": -33.801414489746094,
      "loss": 22.1909,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.028830338269472122,
      "rewards/margins": 0.034757114946842194,
      "rewards/rejected": -0.005926776677370071,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": -1.963141679763794,
      "logits/rejected": -1.9509350061416626,
      "logps/chosen": -34.32793426513672,
      "logps/rejected": -33.663185119628906,
      "loss": 21.758,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.026814326643943787,
      "rewards/margins": 0.03924902528524399,
      "rewards/rejected": -0.012434699572622776,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": -1.9982473850250244,
      "logits/rejected": -1.9968185424804688,
      "logps/chosen": -33.15186309814453,
      "logps/rejected": -32.56055450439453,
      "loss": 21.5888,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.03339282050728798,
      "rewards/margins": 0.0414896085858345,
      "rewards/rejected": -0.008096789941191673,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": -2.084803819656372,
      "logits/rejected": -2.069124698638916,
      "logps/chosen": -33.807334899902344,
      "logps/rejected": -33.127708435058594,
      "loss": 21.7884,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.03720540553331375,
      "rewards/margins": 0.0381726399064064,
      "rewards/rejected": -0.0009672341984696686,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": -1.9572391510009766,
      "logits/rejected": -1.9564039707183838,
      "logps/chosen": -32.83594512939453,
      "logps/rejected": -32.575740814208984,
      "loss": 20.605,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.04461529105901718,
      "rewards/margins": 0.05567503720521927,
      "rewards/rejected": -0.011059742420911789,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": -1.9132989645004272,
      "logits/rejected": -1.923563003540039,
      "logps/chosen": -31.88213539123535,
      "logps/rejected": -35.31838607788086,
      "loss": 22.111,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.03225552663207054,
      "rewards/margins": 0.03521796688437462,
      "rewards/rejected": -0.0029624411836266518,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": -2.0528197288513184,
      "logits/rejected": -2.046360492706299,
      "logps/chosen": -33.33294677734375,
      "logps/rejected": -29.250751495361328,
      "loss": 22.0205,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.03190212696790695,
      "rewards/margins": 0.03433450311422348,
      "rewards/rejected": -0.0024323747493326664,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": -1.9122035503387451,
      "logits/rejected": -1.9144351482391357,
      "logps/chosen": -33.86780548095703,
      "logps/rejected": -30.96420669555664,
      "loss": 21.1141,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": 0.03712456300854683,
      "rewards/margins": 0.046282440423965454,
      "rewards/rejected": -0.009157879278063774,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 22.846377404943688,
      "train_runtime": 3252.1936,
      "train_samples_per_second": 0.947,
      "train_steps_per_second": 0.118
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}