{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 50,
  "global_step": 436,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.022935779816513763,
      "grad_norm": 5.594593346811413,
      "learning_rate": 1.1363636363636363e-07,
      "logits/chosen": -2.619579315185547,
      "logits/rejected": -2.552640676498413,
      "logps/chosen": -265.4322814941406,
      "logps/rejected": -236.1514434814453,
      "loss": 0.1942,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": 0.0001355366694042459,
      "rewards/margins": 0.0003500286547932774,
      "rewards/rejected": -0.00021449197083711624,
      "step": 10
    },
    {
      "epoch": 0.045871559633027525,
      "grad_norm": 5.257491443462003,
      "learning_rate": 2.2727272727272726e-07,
      "logits/chosen": -2.6575067043304443,
      "logits/rejected": -2.575622081756592,
      "logps/chosen": -298.8291320800781,
      "logps/rejected": -274.30499267578125,
      "loss": 0.1958,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": -0.0004384421627037227,
      "rewards/margins": 0.0015119289746508002,
      "rewards/rejected": -0.001950371079146862,
      "step": 20
    },
    {
      "epoch": 0.06880733944954129,
      "grad_norm": 4.9882040114909,
      "learning_rate": 3.4090909090909085e-07,
      "logits/chosen": -2.6754746437072754,
      "logits/rejected": -2.60184383392334,
      "logps/chosen": -290.3965759277344,
      "logps/rejected": -234.4066619873047,
      "loss": 0.1923,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.009362037293612957,
      "rewards/margins": 0.013609759509563446,
      "rewards/rejected": -0.004247722681611776,
      "step": 30
    },
    {
      "epoch": 0.09174311926605505,
      "grad_norm": 5.21855757124415,
      "learning_rate": 4.545454545454545e-07,
      "logits/chosen": -2.6603915691375732,
      "logits/rejected": -2.610896348953247,
      "logps/chosen": -281.05230712890625,
      "logps/rejected": -267.76873779296875,
      "loss": 0.1836,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.04116874188184738,
      "rewards/margins": 0.04092608764767647,
      "rewards/rejected": 0.00024265461252070963,
      "step": 40
    },
    {
      "epoch": 0.11467889908256881,
      "grad_norm": 5.5623018036583085,
      "learning_rate": 4.997110275491701e-07,
      "logits/chosen": -2.619241714477539,
      "logits/rejected": -2.6116530895233154,
      "logps/chosen": -293.481689453125,
      "logps/rejected": -303.96014404296875,
      "loss": 0.1732,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.02769497036933899,
      "rewards/margins": 0.07545463740825653,
      "rewards/rejected": -0.047759659588336945,
      "step": 50
    },
    {
      "epoch": 0.11467889908256881,
      "eval_logits/chosen": -2.5722098350524902,
      "eval_logits/rejected": -2.492361068725586,
      "eval_logps/chosen": -283.61468505859375,
      "eval_logps/rejected": -257.6287841796875,
      "eval_loss": 0.16397468745708466,
      "eval_rewards/accuracies": 0.7068965435028076,
      "eval_rewards/chosen": 0.014754011295735836,
      "eval_rewards/margins": 0.12581734359264374,
      "eval_rewards/rejected": -0.11106333881616592,
      "eval_runtime": 94.5224,
      "eval_samples_per_second": 19.234,
      "eval_steps_per_second": 0.307,
      "step": 50
    },
    {
      "epoch": 0.13761467889908258,
      "grad_norm": 7.227949877380974,
      "learning_rate": 4.979475034558115e-07,
      "logits/chosen": -2.5612680912017822,
      "logits/rejected": -2.5027718544006348,
      "logps/chosen": -293.4398193359375,
      "logps/rejected": -273.11016845703125,
      "loss": 0.1593,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.03955172747373581,
      "rewards/margins": 0.11598478257656097,
      "rewards/rejected": -0.155536487698555,
      "step": 60
    },
    {
      "epoch": 0.16055045871559634,
      "grad_norm": 7.344770125290469,
      "learning_rate": 4.945923025551788e-07,
      "logits/chosen": -2.451911449432373,
      "logits/rejected": -2.4021246433258057,
      "logps/chosen": -327.6131591796875,
      "logps/rejected": -284.70257568359375,
      "loss": 0.1564,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.00042213575216010213,
      "rewards/margins": 0.2325272113084793,
      "rewards/rejected": -0.23294934630393982,
      "step": 70
    },
    {
      "epoch": 0.1834862385321101,
      "grad_norm": 8.03952466935575,
      "learning_rate": 4.896669632591651e-07,
      "logits/chosen": -2.312901020050049,
      "logits/rejected": -2.1888813972473145,
      "logps/chosen": -298.2128601074219,
      "logps/rejected": -291.5619201660156,
      "loss": 0.1475,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.1236300840973854,
      "rewards/margins": 0.2654283344745636,
      "rewards/rejected": -0.3890584111213684,
      "step": 80
    },
    {
      "epoch": 0.20642201834862386,
      "grad_norm": 10.413197257520066,
      "learning_rate": 4.832031033425662e-07,
      "logits/chosen": -1.9742157459259033,
      "logits/rejected": -1.9053027629852295,
      "logps/chosen": -299.2644958496094,
      "logps/rejected": -292.5104064941406,
      "loss": 0.1436,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.11671074479818344,
      "rewards/margins": 0.34038084745407104,
      "rewards/rejected": -0.4570915102958679,
      "step": 90
    },
    {
      "epoch": 0.22935779816513763,
      "grad_norm": 8.350562582651383,
      "learning_rate": 4.752422169756047e-07,
      "logits/chosen": -1.6017181873321533,
      "logits/rejected": -1.5160282850265503,
      "logps/chosen": -289.7087097167969,
      "logps/rejected": -317.5405578613281,
      "loss": 0.1403,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": -0.17284031212329865,
      "rewards/margins": 0.28328073024749756,
      "rewards/rejected": -0.4561210572719574,
      "step": 100
    },
    {
      "epoch": 0.22935779816513763,
      "eval_logits/chosen": -2.056002616882324,
      "eval_logits/rejected": -1.8343770503997803,
      "eval_logps/chosen": -298.7873229980469,
      "eval_logps/rejected": -293.1888427734375,
      "eval_loss": 0.1361575871706009,
      "eval_rewards/accuracies": 0.693965494632721,
      "eval_rewards/chosen": -0.13697203993797302,
      "eval_rewards/margins": 0.3296918570995331,
      "eval_rewards/rejected": -0.4666639268398285,
      "eval_runtime": 95.6937,
      "eval_samples_per_second": 18.998,
      "eval_steps_per_second": 0.303,
      "step": 100
    },
    {
      "epoch": 0.25229357798165136,
      "grad_norm": 8.176356995673753,
      "learning_rate": 4.658354083558188e-07,
      "logits/chosen": -2.1571879386901855,
      "logits/rejected": -1.9664814472198486,
      "logps/chosen": -285.4144287109375,
      "logps/rejected": -280.06732177734375,
      "loss": 0.1397,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.2022629678249359,
      "rewards/margins": 0.2610785961151123,
      "rewards/rejected": -0.46334153413772583,
      "step": 110
    },
    {
      "epoch": 0.27522935779816515,
      "grad_norm": 11.47061183691114,
      "learning_rate": 4.550430636492389e-07,
      "logits/chosen": -1.4951406717300415,
      "logits/rejected": -1.340619444847107,
      "logps/chosen": -316.77838134765625,
      "logps/rejected": -329.01495361328125,
      "loss": 0.1386,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": -0.4095904231071472,
      "rewards/margins": 0.3659195899963379,
      "rewards/rejected": -0.7755100131034851,
      "step": 120
    },
    {
      "epoch": 0.2981651376146789,
      "grad_norm": 12.47670192079271,
      "learning_rate": 4.429344633468004e-07,
      "logits/chosen": -1.376880168914795,
      "logits/rejected": -1.1415055990219116,
      "logps/chosen": -303.68109130859375,
      "logps/rejected": -326.821533203125,
      "loss": 0.1383,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.47500982880592346,
      "rewards/margins": 0.3670799136161804,
      "rewards/rejected": -0.8420897722244263,
      "step": 130
    },
    {
      "epoch": 0.3211009174311927,
      "grad_norm": 10.567717397857333,
      "learning_rate": 4.2958733752443187e-07,
      "logits/chosen": -1.6405372619628906,
      "logits/rejected": -1.487768530845642,
      "logps/chosen": -310.8406677246094,
      "logps/rejected": -290.4248962402344,
      "loss": 0.1328,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.3465554118156433,
      "rewards/margins": 0.31472834944725037,
      "rewards/rejected": -0.6612837910652161,
      "step": 140
    },
    {
      "epoch": 0.3440366972477064,
      "grad_norm": 10.103632857127238,
      "learning_rate": 4.150873668617898e-07,
      "logits/chosen": -1.5787580013275146,
      "logits/rejected": -1.270272970199585,
      "logps/chosen": -316.16461181640625,
      "logps/rejected": -323.0017395019531,
      "loss": 0.1324,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.49621015787124634,
      "rewards/margins": 0.4019313454627991,
      "rewards/rejected": -0.8981413841247559,
      "step": 150
    },
    {
      "epoch": 0.3440366972477064,
      "eval_logits/chosen": -1.6554096937179565,
      "eval_logits/rejected": -1.2887206077575684,
      "eval_logps/chosen": -332.78277587890625,
      "eval_logps/rejected": -341.6122741699219,
      "eval_loss": 0.12861265242099762,
      "eval_rewards/accuracies": 0.7370689511299133,
      "eval_rewards/chosen": -0.47692668437957764,
      "eval_rewards/margins": 0.47397124767303467,
      "eval_rewards/rejected": -0.9508979916572571,
      "eval_runtime": 94.5745,
      "eval_samples_per_second": 19.223,
      "eval_steps_per_second": 0.307,
      "step": 150
    },
    {
      "epoch": 0.3669724770642202,
      "grad_norm": 12.610771394953652,
      "learning_rate": 3.9952763262280397e-07,
      "logits/chosen": -1.6438308954238892,
      "logits/rejected": -1.378089427947998,
      "logps/chosen": -342.8047790527344,
      "logps/rejected": -378.4410705566406,
      "loss": 0.125,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.46253666281700134,
      "rewards/margins": 0.4621877670288086,
      "rewards/rejected": -0.9247244000434875,
      "step": 160
    },
    {
      "epoch": 0.38990825688073394,
      "grad_norm": 7.3876281210799375,
      "learning_rate": 3.8300801912883414e-07,
      "logits/chosen": -1.8106731176376343,
      "logits/rejected": -1.5751054286956787,
      "logps/chosen": -300.5574035644531,
      "logps/rejected": -343.215576171875,
      "loss": 0.12,
      "rewards/accuracies": 0.6937500238418579,
      "rewards/chosen": -0.28956499695777893,
      "rewards/margins": 0.4419015049934387,
      "rewards/rejected": -0.73146653175354,
      "step": 170
    },
    {
      "epoch": 0.41284403669724773,
      "grad_norm": 9.731458591388959,
      "learning_rate": 3.6563457256020884e-07,
      "logits/chosen": -1.6891987323760986,
      "logits/rejected": -1.344957709312439,
      "logps/chosen": -351.0888671875,
      "logps/rejected": -328.3114013671875,
      "loss": 0.129,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.47730937600135803,
      "rewards/margins": 0.42085570096969604,
      "rewards/rejected": -0.8981650471687317,
      "step": 180
    },
    {
      "epoch": 0.43577981651376146,
      "grad_norm": 10.618153958266253,
      "learning_rate": 3.475188202022617e-07,
      "logits/chosen": -1.5754601955413818,
      "logits/rejected": -1.322837471961975,
      "logps/chosen": -299.17999267578125,
      "logps/rejected": -353.1037292480469,
      "loss": 0.1248,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.41065484285354614,
      "rewards/margins": 0.4769380986690521,
      "rewards/rejected": -0.8875927925109863,
      "step": 190
    },
    {
      "epoch": 0.45871559633027525,
      "grad_norm": 8.299502199250934,
      "learning_rate": 3.287770545059052e-07,
      "logits/chosen": -1.7796523571014404,
      "logits/rejected": -1.4286386966705322,
      "logps/chosen": -312.58306884765625,
      "logps/rejected": -324.02142333984375,
      "loss": 0.1249,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.3430142104625702,
      "rewards/margins": 0.4586402475833893,
      "rewards/rejected": -0.8016544580459595,
      "step": 200
    },
    {
      "epoch": 0.45871559633027525,
      "eval_logits/chosen": -1.8577966690063477,
      "eval_logits/rejected": -1.4797794818878174,
      "eval_logps/chosen": -314.017578125,
      "eval_logps/rejected": -322.6352233886719,
      "eval_loss": 0.12169274687767029,
      "eval_rewards/accuracies": 0.7241379022598267,
      "eval_rewards/chosen": -0.2892747223377228,
      "eval_rewards/margins": 0.4718529284000397,
      "eval_rewards/rejected": -0.7611277103424072,
      "eval_runtime": 95.5757,
      "eval_samples_per_second": 19.022,
      "eval_steps_per_second": 0.303,
      "step": 200
    },
    {
      "epoch": 0.481651376146789,
      "grad_norm": 8.358424851083543,
      "learning_rate": 3.0952958655864954e-07,
      "logits/chosen": -1.7499780654907227,
      "logits/rejected": -1.4417035579681396,
      "logps/chosen": -303.82586669921875,
      "logps/rejected": -331.75396728515625,
      "loss": 0.1201,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.30521491169929504,
      "rewards/margins": 0.46890267729759216,
      "rewards/rejected": -0.7741175293922424,
      "step": 210
    },
    {
      "epoch": 0.5045871559633027,
      "grad_norm": 9.67017691007605,
      "learning_rate": 2.898999737583448e-07,
      "logits/chosen": -1.6984879970550537,
      "logits/rejected": -1.103266954421997,
      "logps/chosen": -371.213623046875,
      "logps/rejected": -387.9047546386719,
      "loss": 0.1164,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -0.4853332042694092,
      "rewards/margins": 0.5440121293067932,
      "rewards/rejected": -1.0293452739715576,
      "step": 220
    },
    {
      "epoch": 0.5275229357798165,
      "grad_norm": 9.051099459635976,
      "learning_rate": 2.7001422664752333e-07,
      "logits/chosen": -1.5834382772445679,
      "logits/rejected": -1.3102777004241943,
      "logps/chosen": -317.35888671875,
      "logps/rejected": -356.3561096191406,
      "loss": 0.1216,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": -0.4941536784172058,
      "rewards/margins": 0.4132082462310791,
      "rewards/rejected": -0.9073619842529297,
      "step": 230
    },
    {
      "epoch": 0.5504587155963303,
      "grad_norm": 8.670604935851552,
      "learning_rate": 2.5e-07,
      "logits/chosen": -1.79386305809021,
      "logits/rejected": -1.4528472423553467,
      "logps/chosen": -328.12200927734375,
      "logps/rejected": -345.81500244140625,
      "loss": 0.1217,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.40590184926986694,
      "rewards/margins": 0.43874797224998474,
      "rewards/rejected": -0.8446499109268188,
      "step": 240
    },
    {
      "epoch": 0.573394495412844,
      "grad_norm": 8.325091240643497,
      "learning_rate": 2.2998577335247667e-07,
      "logits/chosen": -1.7739427089691162,
      "logits/rejected": -1.3005108833312988,
      "logps/chosen": -344.14129638671875,
      "logps/rejected": -352.2928771972656,
      "loss": 0.1189,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.4034748673439026,
      "rewards/margins": 0.515015721321106,
      "rewards/rejected": -0.9184905886650085,
      "step": 250
    },
    {
      "epoch": 0.573394495412844,
      "eval_logits/chosen": -1.6726821660995483,
      "eval_logits/rejected": -1.2226585149765015,
      "eval_logps/chosen": -327.7221374511719,
      "eval_logps/rejected": -344.0638427734375,
      "eval_loss": 0.11745046824216843,
      "eval_rewards/accuracies": 0.7629310488700867,
      "eval_rewards/chosen": -0.4263203740119934,
      "eval_rewards/margins": 0.5490937232971191,
      "eval_rewards/rejected": -0.9754140973091125,
      "eval_runtime": 95.1095,
      "eval_samples_per_second": 19.115,
      "eval_steps_per_second": 0.305,
      "step": 250
    },
    {
      "epoch": 0.5963302752293578,
      "grad_norm": 8.867617693309038,
      "learning_rate": 2.1010002624165524e-07,
      "logits/chosen": -1.6618998050689697,
      "logits/rejected": -1.3614200353622437,
      "logps/chosen": -315.7143249511719,
      "logps/rejected": -377.2685546875,
      "loss": 0.117,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.44999808073043823,
      "rewards/margins": 0.5787504315376282,
      "rewards/rejected": -1.0287485122680664,
      "step": 260
    },
    {
      "epoch": 0.6192660550458715,
      "grad_norm": 8.935506104756715,
      "learning_rate": 1.9047041344135043e-07,
      "logits/chosen": -1.6130893230438232,
      "logits/rejected": -1.4423035383224487,
      "logps/chosen": -308.60711669921875,
      "logps/rejected": -346.1998291015625,
      "loss": 0.1177,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.4254485070705414,
      "rewards/margins": 0.5180791020393372,
      "rewards/rejected": -0.9435275793075562,
      "step": 270
    },
    {
      "epoch": 0.6422018348623854,
      "grad_norm": 9.329726294195847,
      "learning_rate": 1.7122294549409482e-07,
      "logits/chosen": -1.6924329996109009,
      "logits/rejected": -1.2535055875778198,
      "logps/chosen": -311.0262451171875,
      "logps/rejected": -363.9076232910156,
      "loss": 0.1216,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.3802736699581146,
      "rewards/margins": 0.6208445429801941,
      "rewards/rejected": -1.0011183023452759,
      "step": 280
    },
    {
      "epoch": 0.6651376146788991,
      "grad_norm": 11.266324554606376,
      "learning_rate": 1.524811797977383e-07,
      "logits/chosen": -1.6415001153945923,
      "logits/rejected": -1.2690132856369019,
      "logps/chosen": -327.433837890625,
      "logps/rejected": -346.80560302734375,
      "loss": 0.1149,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.3970295786857605,
      "rewards/margins": 0.5318586826324463,
      "rewards/rejected": -0.928888201713562,
      "step": 290
    },
    {
      "epoch": 0.6880733944954128,
      "grad_norm": 9.171141301041647,
      "learning_rate": 1.3436542743979125e-07,
      "logits/chosen": -1.6442855596542358,
      "logits/rejected": -1.4366796016693115,
      "logps/chosen": -349.1554870605469,
      "logps/rejected": -343.85443115234375,
      "loss": 0.1252,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -0.4320616126060486,
      "rewards/margins": 0.4178841710090637,
      "rewards/rejected": -0.8499458432197571,
      "step": 300
    },
    {
      "epoch": 0.6880733944954128,
      "eval_logits/chosen": -1.663388729095459,
      "eval_logits/rejected": -1.1890878677368164,
      "eval_logps/chosen": -328.0691223144531,
      "eval_logps/rejected": -345.0453796386719,
      "eval_loss": 0.11537355184555054,
      "eval_rewards/accuracies": 0.7543103694915771,
      "eval_rewards/chosen": -0.4297903776168823,
      "eval_rewards/margins": 0.5554393529891968,
      "eval_rewards/rejected": -0.9852296710014343,
      "eval_runtime": 94.7636,
      "eval_samples_per_second": 19.185,
      "eval_steps_per_second": 0.306,
      "step": 300
    },
    {
      "epoch": 0.7110091743119266,
      "grad_norm": 8.786286019785855,
      "learning_rate": 1.1699198087116588e-07,
      "logits/chosen": -1.6734371185302734,
      "logits/rejected": -1.3057066202163696,
      "logps/chosen": -325.54986572265625,
      "logps/rejected": -359.43572998046875,
      "loss": 0.123,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.47584667801856995,
      "rewards/margins": 0.4540133476257324,
      "rewards/rejected": -0.92985999584198,
      "step": 310
    },
    {
      "epoch": 0.7339449541284404,
      "grad_norm": 9.586519281194345,
      "learning_rate": 1.00472367377196e-07,
      "logits/chosen": -1.605583906173706,
      "logits/rejected": -1.1544862985610962,
      "logps/chosen": -318.45654296875,
      "logps/rejected": -339.86163330078125,
      "loss": 0.1189,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.4422459006309509,
      "rewards/margins": 0.611468493938446,
      "rewards/rejected": -1.053714394569397,
      "step": 320
    },
    {
      "epoch": 0.7568807339449541,
      "grad_norm": 9.694294177484474,
      "learning_rate": 8.49126331382102e-08,
      "logits/chosen": -1.499674677848816,
      "logits/rejected": -1.1590025424957275,
      "logps/chosen": -327.3614807128906,
      "logps/rejected": -342.82550048828125,
      "loss": 0.1187,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.5283932685852051,
      "rewards/margins": 0.4288608431816101,
      "rewards/rejected": -0.9572542309761047,
      "step": 330
    },
    {
      "epoch": 0.7798165137614679,
      "grad_norm": 10.420796259880595,
      "learning_rate": 7.041266247556812e-08,
      "logits/chosen": -1.587501883506775,
      "logits/rejected": -1.274778127670288,
      "logps/chosen": -338.21112060546875,
      "logps/rejected": -354.18890380859375,
      "loss": 0.1144,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.5021848678588867,
      "rewards/margins": 0.439863920211792,
      "rewards/rejected": -0.9420488476753235,
      "step": 340
    },
    {
      "epoch": 0.8027522935779816,
      "grad_norm": 9.40520686958908,
      "learning_rate": 5.706553665319955e-08,
      "logits/chosen": -1.567800521850586,
      "logits/rejected": -1.0267796516418457,
      "logps/chosen": -327.0886535644531,
      "logps/rejected": -339.1602478027344,
      "loss": 0.1226,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.5259757041931152,
      "rewards/margins": 0.48901230096817017,
      "rewards/rejected": -1.0149879455566406,
      "step": 350
    },
    {
      "epoch": 0.8027522935779816,
      "eval_logits/chosen": -1.5758812427520752,
      "eval_logits/rejected": -1.0590085983276367,
      "eval_logps/chosen": -333.0171203613281,
      "eval_logps/rejected": -349.79791259765625,
      "eval_loss": 0.11367923766374588,
      "eval_rewards/accuracies": 0.7543103694915771,
      "eval_rewards/chosen": -0.47927024960517883,
      "eval_rewards/margins": 0.5534845590591431,
      "eval_rewards/rejected": -1.032754898071289,
      "eval_runtime": 94.1796,
      "eval_samples_per_second": 19.304,
      "eval_steps_per_second": 0.308,
      "step": 350
    },
    {
      "epoch": 0.8256880733944955,
      "grad_norm": 7.926122004749844,
      "learning_rate": 4.4956936350761005e-08,
      "logits/chosen": -1.530390977859497,
      "logits/rejected": -1.2470009326934814,
      "logps/chosen": -297.60394287109375,
      "logps/rejected": -349.6514587402344,
      "loss": 0.1145,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": -0.45283880829811096,
      "rewards/margins": 0.45830464363098145,
      "rewards/rejected": -0.91114342212677,
      "step": 360
    },
    {
      "epoch": 0.8486238532110092,
      "grad_norm": 8.501039777999448,
      "learning_rate": 3.416459164418123e-08,
      "logits/chosen": -1.637953519821167,
      "logits/rejected": -1.3401291370391846,
      "logps/chosen": -346.0074462890625,
      "logps/rejected": -358.09075927734375,
      "loss": 0.1125,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.46422967314720154,
      "rewards/margins": 0.4909321367740631,
      "rewards/rejected": -0.9551618695259094,
      "step": 370
    },
    {
      "epoch": 0.8715596330275229,
      "grad_norm": 10.294163163028395,
      "learning_rate": 2.475778302439524e-08,
      "logits/chosen": -1.6482412815093994,
      "logits/rejected": -1.1318132877349854,
      "logps/chosen": -337.7255859375,
      "logps/rejected": -354.24090576171875,
      "loss": 0.118,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.4134751260280609,
      "rewards/margins": 0.5523717403411865,
      "rewards/rejected": -0.965846836566925,
      "step": 380
    },
    {
      "epoch": 0.8944954128440367,
      "grad_norm": 9.29566160345761,
      "learning_rate": 1.6796896657433805e-08,
      "logits/chosen": -1.5356476306915283,
      "logits/rejected": -1.1930160522460938,
      "logps/chosen": -296.85870361328125,
      "logps/rejected": -320.10577392578125,
      "loss": 0.1227,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": -0.4490123391151428,
      "rewards/margins": 0.4589834213256836,
      "rewards/rejected": -0.9079957008361816,
      "step": 390
    },
    {
      "epoch": 0.9174311926605505,
      "grad_norm": 10.238826336953899,
      "learning_rate": 1.0333036740834855e-08,
      "logits/chosen": -1.455672264099121,
      "logits/rejected": -1.2008806467056274,
      "logps/chosen": -265.4068298339844,
      "logps/rejected": -317.6711120605469,
      "loss": 0.1206,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": -0.4084719121456146,
      "rewards/margins": 0.4760831892490387,
      "rewards/rejected": -0.8845551609992981,
      "step": 400
    },
    {
      "epoch": 0.9174311926605505,
      "eval_logits/chosen": -1.608026385307312,
      "eval_logits/rejected": -1.1351321935653687,
      "eval_logps/chosen": -328.656494140625,
      "eval_logps/rejected": -344.9654846191406,
      "eval_loss": 0.11354311555624008,
      "eval_rewards/accuracies": 0.75,
      "eval_rewards/chosen": -0.43566375970840454,
      "eval_rewards/margins": 0.5487664937973022,
      "eval_rewards/rejected": -0.984430193901062,
      "eval_runtime": 96.0499,
      "eval_samples_per_second": 18.928,
      "eval_steps_per_second": 0.302,
      "step": 400
    },
    {
      "epoch": 0.9403669724770642,
      "grad_norm": 10.26611326963808,
      "learning_rate": 5.4076974448211685e-09,
      "logits/chosen": -1.4635480642318726,
      "logits/rejected": -1.029957890510559,
      "logps/chosen": -317.2907409667969,
      "logps/rejected": -339.1365051269531,
      "loss": 0.1172,
      "rewards/accuracies": 0.75,
      "rewards/chosen": -0.4633886218070984,
      "rewards/margins": 0.5475090742111206,
      "rewards/rejected": -1.0108975172042847,
      "step": 410
    },
    {
      "epoch": 0.963302752293578,
      "grad_norm": 11.907377094577166,
      "learning_rate": 2.052496544188487e-09,
      "logits/chosen": -1.477085828781128,
      "logits/rejected": -0.9949935078620911,
      "logps/chosen": -308.0985412597656,
      "logps/rejected": -351.2358093261719,
      "loss": 0.1158,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.4865953028202057,
      "rewards/margins": 0.5669893026351929,
      "rewards/rejected": -1.0535845756530762,
      "step": 420
    },
    {
      "epoch": 0.9862385321100917,
      "grad_norm": 9.339603996948918,
      "learning_rate": 2.889724508297886e-10,
      "logits/chosen": -1.5852940082550049,
      "logits/rejected": -1.0293024778366089,
      "logps/chosen": -347.1506042480469,
      "logps/rejected": -338.89019775390625,
      "loss": 0.1161,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": -0.439751535654068,
      "rewards/margins": 0.5087853074073792,
      "rewards/rejected": -0.94853675365448,
      "step": 430
    },
    {
      "epoch": 1.0,
      "step": 436,
      "total_flos": 0.0,
      "train_loss": 0.13275344205012016,
      "train_runtime": 11631.0591,
      "train_samples_per_second": 4.794,
      "train_steps_per_second": 0.037
    }
  ],
  "logging_steps": 10,
  "max_steps": 436,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "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
}