{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.807339449541285,
  "eval_steps": 100,
  "global_step": 400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.14678899082568808,
      "grad_norm": 8.3125,
      "learning_rate": 1.44e-06,
      "log_odds_chosen": -0.013934114947915077,
      "log_odds_ratio": -0.7180649042129517,
      "logits/chosen": 1.2355022430419922,
      "logits/rejected": 1.233736276626587,
      "logps/chosen": -0.22528842091560364,
      "logps/rejected": -0.2213267982006073,
      "loss": 0.2971,
      "nll_loss": 0.22528842091560364,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.022528842091560364,
      "rewards/margins": -0.0003961602342315018,
      "rewards/rejected": -0.02213267982006073,
      "step": 10
    },
    {
      "epoch": 0.29357798165137616,
      "grad_norm": 7.4375,
      "learning_rate": 3.0399999999999997e-06,
      "log_odds_chosen": 0.09635721147060394,
      "log_odds_ratio": -0.6635279655456543,
      "logits/chosen": 1.2206422090530396,
      "logits/rejected": 1.2412109375,
      "logps/chosen": -0.18985943496227264,
      "logps/rejected": -0.21236571669578552,
      "loss": 0.2562,
      "nll_loss": 0.18985943496227264,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": -0.018985943868756294,
      "rewards/margins": 0.002250631572678685,
      "rewards/rejected": -0.02123657427728176,
      "step": 20
    },
    {
      "epoch": 0.44036697247706424,
      "grad_norm": 7.0,
      "learning_rate": 4.64e-06,
      "log_odds_chosen": -0.0012740284437313676,
      "log_odds_ratio": -0.730600118637085,
      "logits/chosen": 1.202905535697937,
      "logits/rejected": 1.2013709545135498,
      "logps/chosen": -0.16266927123069763,
      "logps/rejected": -0.16249629855155945,
      "loss": 0.2357,
      "nll_loss": 0.16266925632953644,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": -0.016266927123069763,
      "rewards/margins": -1.729675568640232e-05,
      "rewards/rejected": -0.016249630600214005,
      "step": 30
    },
    {
      "epoch": 0.5871559633027523,
      "grad_norm": 5.4375,
      "learning_rate": 6.2399999999999995e-06,
      "log_odds_chosen": 0.09341854602098465,
      "log_odds_ratio": -0.6845951676368713,
      "logits/chosen": 1.1137419939041138,
      "logits/rejected": 1.1249034404754639,
      "logps/chosen": -0.1443159580230713,
      "logps/rejected": -0.1575850248336792,
      "loss": 0.2128,
      "nll_loss": 0.1443159580230713,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": -0.014431595802307129,
      "rewards/margins": 0.001326907193288207,
      "rewards/rejected": -0.01575850322842598,
      "step": 40
    },
    {
      "epoch": 0.7339449541284404,
      "grad_norm": 5.34375,
      "learning_rate": 7.84e-06,
      "log_odds_chosen": 0.08991637080907822,
      "log_odds_ratio": -0.686031699180603,
      "logits/chosen": 1.040159821510315,
      "logits/rejected": 1.037593126296997,
      "logps/chosen": -0.13130609691143036,
      "logps/rejected": -0.1397678554058075,
      "loss": 0.1999,
      "nll_loss": 0.13130609691143036,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": -0.01313060987740755,
      "rewards/margins": 0.0008461755933240056,
      "rewards/rejected": -0.013976785354316235,
      "step": 50
    },
    {
      "epoch": 0.8807339449541285,
      "grad_norm": 5.96875,
      "learning_rate": 7.364596943186588e-06,
      "log_odds_chosen": 0.17444108426570892,
      "log_odds_ratio": -0.642212450504303,
      "logits/chosen": 1.0531730651855469,
      "logits/rejected": 1.0564888715744019,
      "logps/chosen": -0.12667593359947205,
      "logps/rejected": -0.15231557190418243,
      "loss": 0.1909,
      "nll_loss": 0.12667593359947205,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": -0.01266759354621172,
      "rewards/margins": 0.0025639645755290985,
      "rewards/rejected": -0.015231555327773094,
      "step": 60
    },
    {
      "epoch": 1.0146788990825688,
      "grad_norm": 4.90625,
      "learning_rate": 6.810052246069988e-06,
      "log_odds_chosen": 0.27120134234428406,
      "log_odds_ratio": -0.6171866655349731,
      "logits/chosen": 0.9734293222427368,
      "logits/rejected": 0.978347659111023,
      "logps/chosen": -0.1305103749036789,
      "logps/rejected": -0.16737911105155945,
      "loss": 0.183,
      "nll_loss": 0.1305103749036789,
      "rewards/accuracies": 0.7260273694992065,
      "rewards/chosen": -0.01305103674530983,
      "rewards/margins": 0.003686875570565462,
      "rewards/rejected": -0.016737911850214005,
      "step": 70
    },
    {
      "epoch": 1.161467889908257,
      "grad_norm": 6.71875,
      "learning_rate": 6.36445827340584e-06,
      "log_odds_chosen": 0.5265471339225769,
      "log_odds_ratio": -0.4956350326538086,
      "logits/chosen": 0.8999161720275879,
      "logits/rejected": 0.9058361053466797,
      "logps/chosen": -0.11032483726739883,
      "logps/rejected": -0.16881521046161652,
      "loss": 0.1599,
      "nll_loss": 0.11032483726739883,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -0.011032484471797943,
      "rewards/margins": 0.005849036853760481,
      "rewards/rejected": -0.01688152179121971,
      "step": 80
    },
    {
      "epoch": 1.308256880733945,
      "grad_norm": 5.15625,
      "learning_rate": 5.996253511966892e-06,
      "log_odds_chosen": 0.4344564378261566,
      "log_odds_ratio": -0.537087082862854,
      "logits/chosen": 0.8690741658210754,
      "logits/rejected": 0.8722096681594849,
      "logps/chosen": -0.10178230702877045,
      "logps/rejected": -0.14738620817661285,
      "loss": 0.1555,
      "nll_loss": 0.10178232192993164,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.01017823163419962,
      "rewards/margins": 0.004560390021651983,
      "rewards/rejected": -0.014738621190190315,
      "step": 90
    },
    {
      "epoch": 1.455045871559633,
      "grad_norm": 5.65625,
      "learning_rate": 5.685352436149611e-06,
      "log_odds_chosen": 0.567080020904541,
      "log_odds_ratio": -0.48860812187194824,
      "logits/chosen": 0.7519216537475586,
      "logits/rejected": 0.7579536437988281,
      "logps/chosen": -0.11033564805984497,
      "logps/rejected": -0.17491480708122253,
      "loss": 0.1592,
      "nll_loss": 0.11033564805984497,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.011033564805984497,
      "rewards/margins": 0.006457914598286152,
      "rewards/rejected": -0.017491478472948074,
      "step": 100
    },
    {
      "epoch": 1.455045871559633,
      "eval_log_odds_chosen": 0.41258907318115234,
      "eval_log_odds_ratio": -0.5504921674728394,
      "eval_logits/chosen": 0.7631099820137024,
      "eval_logits/rejected": 0.7616429924964905,
      "eval_logps/chosen": -0.11344099044799805,
      "eval_logps/rejected": -0.16630388796329498,
      "eval_loss": 0.1684901863336563,
      "eval_nll_loss": 0.11344097554683685,
      "eval_rewards/accuracies": 0.7586206793785095,
      "eval_rewards/chosen": -0.01134409848600626,
      "eval_rewards/margins": 0.005286289844661951,
      "eval_rewards/rejected": -0.01663038693368435,
      "eval_runtime": 2.1536,
      "eval_samples_per_second": 13.466,
      "eval_steps_per_second": 13.466,
      "step": 100
    },
    {
      "epoch": 1.6018348623853211,
      "grad_norm": 6.09375,
      "learning_rate": 5.418283691828771e-06,
      "log_odds_chosen": 0.5374370217323303,
      "log_odds_ratio": -0.5145700573921204,
      "logits/chosen": 0.7113163471221924,
      "logits/rejected": 0.7021201252937317,
      "logps/chosen": -0.10811618715524673,
      "logps/rejected": -0.1586550772190094,
      "loss": 0.1596,
      "nll_loss": 0.10811618715524673,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -0.010811617597937584,
      "rewards/margins": 0.005053888075053692,
      "rewards/rejected": -0.015865508466959,
      "step": 110
    },
    {
      "epoch": 1.7486238532110092,
      "grad_norm": 6.8125,
      "learning_rate": 5.1856297884173145e-06,
      "log_odds_chosen": 0.5470713376998901,
      "log_odds_ratio": -0.5047305226325989,
      "logits/chosen": 0.7087436318397522,
      "logits/rejected": 0.723972499370575,
      "logps/chosen": -0.10061709582805634,
      "logps/rejected": -0.16079466044902802,
      "loss": 0.1511,
      "nll_loss": 0.10061709582805634,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.010061711072921753,
      "rewards/margins": 0.006017756648361683,
      "rewards/rejected": -0.01607946678996086,
      "step": 120
    },
    {
      "epoch": 1.8954128440366973,
      "grad_norm": 6.59375,
      "learning_rate": 4.980582450917523e-06,
      "log_odds_chosen": 0.47062650322914124,
      "log_odds_ratio": -0.5357787013053894,
      "logits/chosen": 0.7007481455802917,
      "logits/rejected": 0.7074240446090698,
      "logps/chosen": -0.10861317068338394,
      "logps/rejected": -0.15855737030506134,
      "loss": 0.1622,
      "nll_loss": 0.10861317068338394,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.010861317627131939,
      "rewards/margins": 0.004994419403374195,
      "rewards/rejected": -0.015855735167860985,
      "step": 130
    },
    {
      "epoch": 2.0293577981651376,
      "grad_norm": 5.0625,
      "learning_rate": 4.798081151232537e-06,
      "log_odds_chosen": 0.6022133231163025,
      "log_odds_ratio": -0.46961599588394165,
      "logits/chosen": 0.7364146709442139,
      "logits/rejected": 0.7421723008155823,
      "logps/chosen": -0.10186123847961426,
      "logps/rejected": -0.17022177577018738,
      "loss": 0.15,
      "nll_loss": 0.10186123847961426,
      "rewards/accuracies": 0.9041095972061157,
      "rewards/chosen": -0.010186124593019485,
      "rewards/margins": 0.00683605158701539,
      "rewards/rejected": -0.017022177577018738,
      "step": 140
    },
    {
      "epoch": 2.1761467889908257,
      "grad_norm": 4.71875,
      "learning_rate": 4.634275570907938e-06,
      "log_odds_chosen": 0.7667453289031982,
      "log_odds_ratio": -0.4221932888031006,
      "logits/chosen": 0.6935725212097168,
      "logits/rejected": 0.6918830275535583,
      "logps/chosen": -0.08588626235723495,
      "logps/rejected": -0.1634286344051361,
      "loss": 0.1281,
      "nll_loss": 0.08588626235723495,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.00858862604945898,
      "rewards/margins": 0.00775423739105463,
      "rewards/rejected": -0.01634286530315876,
      "step": 150
    },
    {
      "epoch": 2.322935779816514,
      "grad_norm": 6.84375,
      "learning_rate": 4.4861772327363545e-06,
      "log_odds_chosen": 0.7993400692939758,
      "log_odds_ratio": -0.40405169129371643,
      "logits/chosen": 0.6566668748855591,
      "logits/rejected": 0.6548924446105957,
      "logps/chosen": -0.08073868602514267,
      "logps/rejected": -0.1553017944097519,
      "loss": 0.1211,
      "nll_loss": 0.08073868602514267,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.008073869161307812,
      "rewards/margins": 0.0074563100934028625,
      "rewards/rejected": -0.0155301783233881,
      "step": 160
    },
    {
      "epoch": 2.469724770642202,
      "grad_norm": 6.15625,
      "learning_rate": 4.3514263457633695e-06,
      "log_odds_chosen": 0.8461767435073853,
      "log_odds_ratio": -0.4002578854560852,
      "logits/chosen": 0.5603212714195251,
      "logits/rejected": 0.5709437727928162,
      "logps/chosen": -0.09075653553009033,
      "logps/rejected": -0.18311923742294312,
      "loss": 0.1308,
      "nll_loss": 0.09075653553009033,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": -0.009075653739273548,
      "rewards/margins": 0.009236268699169159,
      "rewards/rejected": -0.018311921507120132,
      "step": 170
    },
    {
      "epoch": 2.61651376146789,
      "grad_norm": 6.8125,
      "learning_rate": 4.228131380653519e-06,
      "log_odds_chosen": 0.8294868469238281,
      "log_odds_ratio": -0.4138098359107971,
      "logits/chosen": 0.5700741410255432,
      "logits/rejected": 0.5716632008552551,
      "logps/chosen": -0.09730932861566544,
      "logps/rejected": -0.18193913996219635,
      "loss": 0.1387,
      "nll_loss": 0.09730932861566544,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.009730933234095573,
      "rewards/margins": 0.008462979458272457,
      "rewards/rejected": -0.018193911761045456,
      "step": 180
    },
    {
      "epoch": 2.763302752293578,
      "grad_norm": 6.375,
      "learning_rate": 4.1147559989891175e-06,
      "log_odds_chosen": 0.9309076070785522,
      "log_odds_ratio": -0.36566996574401855,
      "logits/chosen": 0.5632237195968628,
      "logits/rejected": 0.5597234964370728,
      "logps/chosen": -0.08929087221622467,
      "logps/rejected": -0.1875963658094406,
      "loss": 0.1259,
      "nll_loss": 0.08929087221622467,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.008929086849093437,
      "rewards/margins": 0.00983054842799902,
      "rewards/rejected": -0.01875963807106018,
      "step": 190
    },
    {
      "epoch": 2.910091743119266,
      "grad_norm": 6.03125,
      "learning_rate": 4.010037656936684e-06,
      "log_odds_chosen": 0.858248233795166,
      "log_odds_ratio": -0.41156458854675293,
      "logits/chosen": 0.5531570315361023,
      "logits/rejected": 0.5494219660758972,
      "logps/chosen": -0.11003482341766357,
      "logps/rejected": -0.2141534388065338,
      "loss": 0.1512,
      "nll_loss": 0.11003480106592178,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -0.011003481224179268,
      "rewards/margins": 0.010411862283945084,
      "rewards/rejected": -0.02141534350812435,
      "step": 200
    },
    {
      "epoch": 2.910091743119266,
      "eval_log_odds_chosen": 0.5659250020980835,
      "eval_log_odds_ratio": -0.4923820197582245,
      "eval_logits/chosen": 0.5788495540618896,
      "eval_logits/rejected": 0.5684294700622559,
      "eval_logps/chosen": -0.10609915852546692,
      "eval_logps/rejected": -0.17641222476959229,
      "eval_loss": 0.1553373485803604,
      "eval_nll_loss": 0.10609915852546692,
      "eval_rewards/accuracies": 0.7931034564971924,
      "eval_rewards/chosen": -0.010609916411340237,
      "eval_rewards/margins": 0.007031304761767387,
      "eval_rewards/rejected": -0.01764122024178505,
      "eval_runtime": 2.1179,
      "eval_samples_per_second": 13.693,
      "eval_steps_per_second": 13.693,
      "step": 200
    },
    {
      "epoch": 3.0440366972477064,
      "grad_norm": 5.125,
      "learning_rate": 3.9129279043561475e-06,
      "log_odds_chosen": 0.8942320346832275,
      "log_odds_ratio": -0.37385156750679016,
      "logits/chosen": 0.5164369344711304,
      "logits/rejected": 0.5234641432762146,
      "logps/chosen": -0.08054827898740768,
      "logps/rejected": -0.17158031463623047,
      "loss": 0.1281,
      "nll_loss": 0.08054827898740768,
      "rewards/accuracies": 0.9452054500579834,
      "rewards/chosen": -0.008054828271269798,
      "rewards/margins": 0.009103202261030674,
      "rewards/rejected": -0.017158031463623047,
      "step": 210
    },
    {
      "epoch": 3.1908256880733945,
      "grad_norm": 4.90625,
      "learning_rate": 3.822547860539972e-06,
      "log_odds_chosen": 1.0804862976074219,
      "log_odds_ratio": -0.3342968225479126,
      "logits/chosen": 0.5213353633880615,
      "logits/rejected": 0.5297050476074219,
      "logps/chosen": -0.07486046850681305,
      "logps/rejected": -0.17718155682086945,
      "loss": 0.1083,
      "nll_loss": 0.07486046850681305,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.0074860467575490475,
      "rewards/margins": 0.01023210771381855,
      "rewards/rejected": -0.017718154937028885,
      "step": 220
    },
    {
      "epoch": 3.3376146788990826,
      "grad_norm": 6.09375,
      "learning_rate": 3.738154508127982e-06,
      "log_odds_chosen": 0.8557016253471375,
      "log_odds_ratio": -0.39284101128578186,
      "logits/chosen": 0.512273907661438,
      "logits/rejected": 0.5162364840507507,
      "logps/chosen": -0.08021821081638336,
      "logps/rejected": -0.16372254490852356,
      "loss": 0.1195,
      "nll_loss": 0.08021820336580276,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": -0.008021821267902851,
      "rewards/margins": 0.0083504319190979,
      "rewards/rejected": -0.016372252255678177,
      "step": 230
    },
    {
      "epoch": 3.4844036697247707,
      "grad_norm": 5.78125,
      "learning_rate": 3.6591148299707855e-06,
      "log_odds_chosen": 1.0299314260482788,
      "log_odds_ratio": -0.3464810252189636,
      "logits/chosen": 0.47674304246902466,
      "logits/rejected": 0.472658634185791,
      "logps/chosen": -0.08960604667663574,
      "logps/rejected": -0.205189511179924,
      "loss": 0.1243,
      "nll_loss": 0.08960604667663574,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.008960604667663574,
      "rewards/margins": 0.011558346450328827,
      "rewards/rejected": -0.0205189511179924,
      "step": 240
    },
    {
      "epoch": 3.6311926605504588,
      "grad_norm": 5.5,
      "learning_rate": 3.584885719585766e-06,
      "log_odds_chosen": 1.0607478618621826,
      "log_odds_ratio": -0.3343815505504608,
      "logits/chosen": 0.44886499643325806,
      "logits/rejected": 0.453238308429718,
      "logps/chosen": -0.08937816321849823,
      "logps/rejected": -0.20491468906402588,
      "loss": 0.1228,
      "nll_loss": 0.08937815576791763,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.008937815204262733,
      "rewards/margins": 0.011553652584552765,
      "rewards/rejected": -0.020491469651460648,
      "step": 250
    },
    {
      "epoch": 3.777981651376147,
      "grad_norm": 5.53125,
      "learning_rate": 3.514998201309974e-06,
      "log_odds_chosen": 1.089552640914917,
      "log_odds_ratio": -0.3337121903896332,
      "logits/chosen": 0.4593326151371002,
      "logits/rejected": 0.4618263244628906,
      "logps/chosen": -0.0794382244348526,
      "logps/rejected": -0.1857403665781021,
      "loss": 0.1128,
      "nll_loss": 0.0794382244348526,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.00794382207095623,
      "rewards/margins": 0.010630212724208832,
      "rewards/rejected": -0.018574032932519913,
      "step": 260
    },
    {
      "epoch": 3.924770642201835,
      "grad_norm": 5.34375,
      "learning_rate": 3.4490449084738148e-06,
      "log_odds_chosen": 1.1249830722808838,
      "log_odds_ratio": -0.31604427099227905,
      "logits/chosen": 0.4885838031768799,
      "logits/rejected": 0.4889053702354431,
      "logps/chosen": -0.08302140235900879,
      "logps/rejected": -0.2039506435394287,
      "loss": 0.1146,
      "nll_loss": 0.08302140235900879,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": -0.008302140049636364,
      "rewards/margins": 0.012092925608158112,
      "rewards/rejected": -0.02039506658911705,
      "step": 270
    },
    {
      "epoch": 4.058715596330275,
      "grad_norm": 5.84375,
      "learning_rate": 3.3866700533384128e-06,
      "log_odds_chosen": 1.3261604309082031,
      "log_odds_ratio": -0.2692450284957886,
      "logits/chosen": 0.45250293612480164,
      "logits/rejected": 0.4294807016849518,
      "logps/chosen": -0.0632709264755249,
      "logps/rejected": -0.1957898885011673,
      "loss": 0.1003,
      "nll_loss": 0.0632709264755249,
      "rewards/accuracies": 0.9726027250289917,
      "rewards/chosen": -0.0063270931132137775,
      "rewards/margins": 0.0132518969476223,
      "rewards/rejected": -0.01957898959517479,
      "step": 280
    },
    {
      "epoch": 4.205504587155963,
      "grad_norm": 6.34375,
      "learning_rate": 3.3275613232308115e-06,
      "log_odds_chosen": 1.2198822498321533,
      "log_odds_ratio": -0.30882710218429565,
      "logits/chosen": 0.4520139694213867,
      "logits/rejected": 0.44239917397499084,
      "logps/chosen": -0.07900197803974152,
      "logps/rejected": -0.20268447697162628,
      "loss": 0.1099,
      "nll_loss": 0.07900197803974152,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.007900198921561241,
      "rewards/margins": 0.012368248775601387,
      "rewards/rejected": -0.02026844583451748,
      "step": 290
    },
    {
      "epoch": 4.352293577981651,
      "grad_norm": 5.40625,
      "learning_rate": 3.2714432804381223e-06,
      "log_odds_chosen": 1.1661269664764404,
      "log_odds_ratio": -0.3138650059700012,
      "logits/chosen": 0.42225146293640137,
      "logits/rejected": 0.42416587471961975,
      "logps/chosen": -0.07937826961278915,
      "logps/rejected": -0.2012827843427658,
      "loss": 0.1108,
      "nll_loss": 0.07937826961278915,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": -0.007937826216220856,
      "rewards/margins": 0.012190451845526695,
      "rewards/rejected": -0.0201282761991024,
      "step": 300
    },
    {
      "epoch": 4.352293577981651,
      "eval_log_odds_chosen": 0.6708114147186279,
      "eval_log_odds_ratio": -0.45528367161750793,
      "eval_logits/chosen": 0.4614703953266144,
      "eval_logits/rejected": 0.4433332681655884,
      "eval_logps/chosen": -0.10493747889995575,
      "eval_logps/rejected": -0.19022339582443237,
      "eval_loss": 0.15046584606170654,
      "eval_nll_loss": 0.10493748635053635,
      "eval_rewards/accuracies": 0.8965517282485962,
      "eval_rewards/chosen": -0.010493747889995575,
      "eval_rewards/margins": 0.008528591133654118,
      "eval_rewards/rejected": -0.019022339954972267,
      "eval_runtime": 2.1131,
      "eval_samples_per_second": 13.724,
      "eval_steps_per_second": 13.724,
      "step": 300
    },
    {
      "epoch": 4.4990825688073395,
      "grad_norm": 6.5625,
      "learning_rate": 3.2180719468365186e-06,
      "log_odds_chosen": 1.1266820430755615,
      "log_odds_ratio": -0.3179430067539215,
      "logits/chosen": 0.43243640661239624,
      "logits/rejected": 0.4324881434440613,
      "logps/chosen": -0.075148805975914,
      "logps/rejected": -0.18810391426086426,
      "loss": 0.1069,
      "nll_loss": 0.075148805975914,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.007514881435781717,
      "rewards/margins": 0.011295510455965996,
      "rewards/rejected": -0.018810391426086426,
      "step": 310
    },
    {
      "epoch": 4.645871559633028,
      "grad_norm": 5.9375,
      "learning_rate": 3.167230329859583e-06,
      "log_odds_chosen": 1.277369737625122,
      "log_odds_ratio": -0.2844545245170593,
      "logits/chosen": 0.4020269513130188,
      "logits/rejected": 0.3961581587791443,
      "logps/chosen": -0.07351918518543243,
      "logps/rejected": -0.21050676703453064,
      "loss": 0.102,
      "nll_loss": 0.07351918518543243,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.007351918611675501,
      "rewards/margins": 0.013698759488761425,
      "rewards/rejected": -0.021050678566098213,
      "step": 320
    },
    {
      "epoch": 4.792660550458716,
      "grad_norm": 4.78125,
      "learning_rate": 3.118724702341802e-06,
      "log_odds_chosen": 1.1947705745697021,
      "log_odds_ratio": -0.299292653799057,
      "logits/chosen": 0.427717387676239,
      "logits/rejected": 0.4366043508052826,
      "logps/chosen": -0.07970800250768661,
      "logps/rejected": -0.20539803802967072,
      "loss": 0.1096,
      "nll_loss": 0.07970799505710602,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.007970800623297691,
      "rewards/margins": 0.012569005601108074,
      "rewards/rejected": -0.02053980901837349,
      "step": 330
    },
    {
      "epoch": 4.939449541284404,
      "grad_norm": 5.46875,
      "learning_rate": 3.072381490570274e-06,
      "log_odds_chosen": 1.329350471496582,
      "log_odds_ratio": -0.2779151499271393,
      "logits/chosen": 0.38692814111709595,
      "logits/rejected": 0.3954050540924072,
      "logps/chosen": -0.06956486403942108,
      "logps/rejected": -0.20327439904212952,
      "loss": 0.0974,
      "nll_loss": 0.06956486403942108,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.006956486497074366,
      "rewards/margins": 0.01337094884365797,
      "rewards/rejected": -0.020327437669038773,
      "step": 340
    },
    {
      "epoch": 5.073394495412844,
      "grad_norm": 6.4375,
      "learning_rate": 3.028044656417858e-06,
      "log_odds_chosen": 1.2671765089035034,
      "log_odds_ratio": -0.28358665108680725,
      "logits/chosen": 0.4236637055873871,
      "logits/rejected": 0.41844019293785095,
      "logps/chosen": -0.07393307238817215,
      "logps/rejected": -0.20131486654281616,
      "loss": 0.1046,
      "nll_loss": 0.07393307238817215,
      "rewards/accuracies": 0.9589040875434875,
      "rewards/chosen": -0.0073933070525527,
      "rewards/margins": 0.012738179415464401,
      "rewards/rejected": -0.020131487399339676,
      "step": 350
    },
    {
      "epoch": 5.220183486238533,
      "grad_norm": 6.53125,
      "learning_rate": 2.9855734834483035e-06,
      "log_odds_chosen": 1.3660424947738647,
      "log_odds_ratio": -0.2688550353050232,
      "logits/chosen": 0.3870636224746704,
      "logits/rejected": 0.37437111139297485,
      "logps/chosen": -0.07374311983585358,
      "logps/rejected": -0.2196178138256073,
      "loss": 0.1006,
      "nll_loss": 0.07374311238527298,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.007374311797320843,
      "rewards/margins": 0.014587472192943096,
      "rewards/rejected": -0.02196178399026394,
      "step": 360
    },
    {
      "epoch": 5.36697247706422,
      "grad_norm": 4.8125,
      "learning_rate": 2.944840695329241e-06,
      "log_odds_chosen": 1.2474989891052246,
      "log_odds_ratio": -0.29633039236068726,
      "logits/chosen": 0.4045291841030121,
      "logits/rejected": 0.4085766673088074,
      "logps/chosen": -0.0789073258638382,
      "logps/rejected": -0.21004538238048553,
      "loss": 0.1085,
      "nll_loss": 0.0789073258638382,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.00789073295891285,
      "rewards/margins": 0.013113808818161488,
      "rewards/rejected": -0.021004538983106613,
      "step": 370
    },
    {
      "epoch": 5.513761467889909,
      "grad_norm": 7.125,
      "learning_rate": 2.9057308491685917e-06,
      "log_odds_chosen": 1.2928500175476074,
      "log_odds_ratio": -0.28314128518104553,
      "logits/chosen": 0.3611066937446594,
      "logits/rejected": 0.36763590574264526,
      "logps/chosen": -0.0696084052324295,
      "logps/rejected": -0.19842052459716797,
      "loss": 0.0979,
      "nll_loss": 0.06960839778184891,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": -0.006960839964449406,
      "rewards/margins": 0.012881213799118996,
      "rewards/rejected": -0.019842052832245827,
      "step": 380
    },
    {
      "epoch": 5.660550458715596,
      "grad_norm": 6.875,
      "learning_rate": 2.868138957528299e-06,
      "log_odds_chosen": 1.308441400527954,
      "log_odds_ratio": -0.2882232964038849,
      "logits/chosen": 0.38956940174102783,
      "logits/rejected": 0.38561898469924927,
      "logps/chosen": -0.08248557895421982,
      "logps/rejected": -0.22931762039661407,
      "loss": 0.1113,
      "nll_loss": 0.08248557150363922,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": -0.008248557336628437,
      "rewards/margins": 0.01468320470303297,
      "rewards/rejected": -0.022931763902306557,
      "step": 390
    },
    {
      "epoch": 5.807339449541285,
      "grad_norm": 5.90625,
      "learning_rate": 2.8319693016191545e-06,
      "log_odds_chosen": 1.3143831491470337,
      "log_odds_ratio": -0.27059316635131836,
      "logits/chosen": 0.31519538164138794,
      "logits/rejected": 0.31572309136390686,
      "logps/chosen": -0.0752774104475975,
      "logps/rejected": -0.2211039513349533,
      "loss": 0.1023,
      "nll_loss": 0.0752774104475975,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": -0.007527741603553295,
      "rewards/margins": 0.014582650735974312,
      "rewards/rejected": -0.02211039327085018,
      "step": 400
    },
    {
      "epoch": 5.807339449541285,
      "eval_log_odds_chosen": 0.7261541485786438,
      "eval_log_odds_ratio": -0.4359833896160126,
      "eval_logits/chosen": 0.41933315992355347,
      "eval_logits/rejected": 0.3965945541858673,
      "eval_logps/chosen": -0.10455876588821411,
      "eval_logps/rejected": -0.19739307463169098,
      "eval_loss": 0.14815710484981537,
      "eval_nll_loss": 0.10455876588821411,
      "eval_rewards/accuracies": 0.8965517282485962,
      "eval_rewards/chosen": -0.010455875657498837,
      "eval_rewards/margins": 0.009283430874347687,
      "eval_rewards/rejected": -0.019739307463169098,
      "eval_runtime": 2.1188,
      "eval_samples_per_second": 13.687,
      "eval_steps_per_second": 13.687,
      "step": 400
    },
    {
      "epoch": 5.807339449541285,
      "step": 400,
      "total_flos": 0.0,
      "train_loss": 0.14153494492173194,
      "train_runtime": 748.5587,
      "train_samples_per_second": 4.275,
      "train_steps_per_second": 0.534
    }
  ],
  "logging_steps": 10,
  "max_steps": 400,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 400,
  "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": 1,
  "trial_name": null,
  "trial_params": null
}