{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.993762993762994,
  "eval_steps": 50,
  "global_step": 360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04158004158004158,
      "grad_norm": 60.93168963253538,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7216269969940186,
      "logits/rejected": -2.719656467437744,
      "logps/chosen": -279.2793884277344,
      "logps/rejected": -217.73977661132812,
      "loss": 0.6903,
      "rewards/accuracies": 0.3187499940395355,
      "rewards/chosen": 0.02068181149661541,
      "rewards/margins": 0.008613892830908298,
      "rewards/rejected": 0.012067919597029686,
      "step": 5
    },
    {
      "epoch": 0.08316008316008316,
      "grad_norm": 56.89106976222382,
      "learning_rate": 1e-06,
      "logits/chosen": -2.6707825660705566,
      "logits/rejected": -2.6669201850891113,
      "logps/chosen": -256.79901123046875,
      "logps/rejected": -231.99880981445312,
      "loss": 0.655,
      "rewards/accuracies": 0.6312500238418579,
      "rewards/chosen": 0.5268703103065491,
      "rewards/margins": 0.16865657269954681,
      "rewards/rejected": 0.35821372270584106,
      "step": 10
    },
    {
      "epoch": 0.12474012474012475,
      "grad_norm": 63.52271997282497,
      "learning_rate": 9.994965332706572e-07,
      "logits/chosen": -2.513793468475342,
      "logits/rejected": -2.5043387413024902,
      "logps/chosen": -264.4665222167969,
      "logps/rejected": -220.4278564453125,
      "loss": 0.6536,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 1.3569260835647583,
      "rewards/margins": 0.6852619051933289,
      "rewards/rejected": 0.6716641187667847,
      "step": 15
    },
    {
      "epoch": 0.16632016632016633,
      "grad_norm": 43.71783786864763,
      "learning_rate": 9.979871469976195e-07,
      "logits/chosen": -2.389570713043213,
      "logits/rejected": -2.362312078475952,
      "logps/chosen": -237.1497802734375,
      "logps/rejected": -194.1293182373047,
      "loss": 0.5961,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 1.4138034582138062,
      "rewards/margins": 0.9070841073989868,
      "rewards/rejected": 0.5067193508148193,
      "step": 20
    },
    {
      "epoch": 0.2079002079002079,
      "grad_norm": 46.44405380817153,
      "learning_rate": 9.954748808839674e-07,
      "logits/chosen": -2.2999889850616455,
      "logits/rejected": -2.317258358001709,
      "logps/chosen": -260.94281005859375,
      "logps/rejected": -250.1718292236328,
      "loss": 0.6178,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 1.4349501132965088,
      "rewards/margins": 1.189349889755249,
      "rewards/rejected": 0.2456001341342926,
      "step": 25
    },
    {
      "epoch": 0.2494802494802495,
      "grad_norm": 42.03082336502749,
      "learning_rate": 9.919647942993147e-07,
      "logits/chosen": -2.3956313133239746,
      "logits/rejected": -2.3552544116973877,
      "logps/chosen": -241.47103881835938,
      "logps/rejected": -215.29940795898438,
      "loss": 0.6159,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.977073073387146,
      "rewards/margins": 1.020427942276001,
      "rewards/rejected": -0.04335489124059677,
      "step": 30
    },
    {
      "epoch": 0.2910602910602911,
      "grad_norm": 46.49384193628988,
      "learning_rate": 9.874639560909118e-07,
      "logits/chosen": -2.493499994277954,
      "logits/rejected": -2.4809508323669434,
      "logps/chosen": -251.2248992919922,
      "logps/rejected": -201.01251220703125,
      "loss": 0.6348,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.6344128847122192,
      "rewards/margins": 0.7719027400016785,
      "rewards/rejected": -0.13748976588249207,
      "step": 35
    },
    {
      "epoch": 0.33264033264033266,
      "grad_norm": 36.79345771283175,
      "learning_rate": 9.819814303479267e-07,
      "logits/chosen": -2.4675540924072266,
      "logits/rejected": -2.46543288230896,
      "logps/chosen": -261.19268798828125,
      "logps/rejected": -202.8013153076172,
      "loss": 0.5763,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.6426112651824951,
      "rewards/margins": 0.9732248187065125,
      "rewards/rejected": -0.33061355352401733,
      "step": 40
    },
    {
      "epoch": 0.37422037422037424,
      "grad_norm": 41.675486172068894,
      "learning_rate": 9.755282581475767e-07,
      "logits/chosen": -2.425602436065674,
      "logits/rejected": -2.4132189750671387,
      "logps/chosen": -243.4677734375,
      "logps/rejected": -185.2170867919922,
      "loss": 0.6023,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 0.7012235522270203,
      "rewards/margins": 0.7697752714157104,
      "rewards/rejected": -0.06855180859565735,
      "step": 45
    },
    {
      "epoch": 0.4158004158004158,
      "grad_norm": 41.93176697972712,
      "learning_rate": 9.681174353198686e-07,
      "logits/chosen": -2.4606122970581055,
      "logits/rejected": -2.451716184616089,
      "logps/chosen": -232.999267578125,
      "logps/rejected": -211.801025390625,
      "loss": 0.6087,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": 0.8648797869682312,
      "rewards/margins": 1.025213599205017,
      "rewards/rejected": -0.16033388674259186,
      "step": 50
    },
    {
      "epoch": 0.4158004158004158,
      "eval_logits/chosen": -2.5455965995788574,
      "eval_logits/rejected": -2.528465509414673,
      "eval_logps/chosen": -241.17100524902344,
      "eval_logps/rejected": -207.0612335205078,
      "eval_loss": 0.5778272747993469,
      "eval_rewards/accuracies": 0.7870370149612427,
      "eval_rewards/chosen": 0.7418816685676575,
      "eval_rewards/margins": 1.2493042945861816,
      "eval_rewards/rejected": -0.5074228644371033,
      "eval_runtime": 229.2387,
      "eval_samples_per_second": 14.919,
      "eval_steps_per_second": 0.236,
      "step": 50
    },
    {
      "epoch": 0.4573804573804574,
      "grad_norm": 37.969958241527884,
      "learning_rate": 9.597638862757253e-07,
      "logits/chosen": -2.5910000801086426,
      "logits/rejected": -2.5817267894744873,
      "logps/chosen": -235.71316528320312,
      "logps/rejected": -194.326171875,
      "loss": 0.5877,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.866040825843811,
      "rewards/margins": 1.1515779495239258,
      "rewards/rejected": -0.28553709387779236,
      "step": 55
    },
    {
      "epoch": 0.498960498960499,
      "grad_norm": 55.72037568304586,
      "learning_rate": 9.504844339512094e-07,
      "logits/chosen": -2.643787145614624,
      "logits/rejected": -2.643979787826538,
      "logps/chosen": -255.2833709716797,
      "logps/rejected": -219.94290161132812,
      "loss": 0.604,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 1.1731880903244019,
      "rewards/margins": 1.4057620763778687,
      "rewards/rejected": -0.23257413506507874,
      "step": 60
    },
    {
      "epoch": 0.5405405405405406,
      "grad_norm": 42.3606140855208,
      "learning_rate": 9.402977659283689e-07,
      "logits/chosen": -2.645413875579834,
      "logits/rejected": -2.6578667163848877,
      "logps/chosen": -233.7159423828125,
      "logps/rejected": -209.6158447265625,
      "loss": 0.556,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 1.1114470958709717,
      "rewards/margins": 1.3722739219665527,
      "rewards/rejected": -0.2608269453048706,
      "step": 65
    },
    {
      "epoch": 0.5821205821205822,
      "grad_norm": 36.14760649472159,
      "learning_rate": 9.29224396800933e-07,
      "logits/chosen": -2.5539443492889404,
      "logits/rejected": -2.582876682281494,
      "logps/chosen": -252.9326171875,
      "logps/rejected": -204.24783325195312,
      "loss": 0.5409,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.9803892970085144,
      "rewards/margins": 1.488160252571106,
      "rewards/rejected": -0.507770836353302,
      "step": 70
    },
    {
      "epoch": 0.6237006237006237,
      "grad_norm": 40.66552692069641,
      "learning_rate": 9.172866268606513e-07,
      "logits/chosen": -2.556718349456787,
      "logits/rejected": -2.5814919471740723,
      "logps/chosen": -252.392333984375,
      "logps/rejected": -230.672607421875,
      "loss": 0.5972,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.6901723146438599,
      "rewards/margins": 1.7339065074920654,
      "rewards/rejected": -1.0437343120574951,
      "step": 75
    },
    {
      "epoch": 0.6652806652806653,
      "grad_norm": 37.41696546929204,
      "learning_rate": 9.045084971874737e-07,
      "logits/chosen": -2.5521421432495117,
      "logits/rejected": -2.558589220046997,
      "logps/chosen": -238.3732452392578,
      "logps/rejected": -218.8362274169922,
      "loss": 0.5516,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": 0.7391923666000366,
      "rewards/margins": 1.9195365905761719,
      "rewards/rejected": -1.1803443431854248,
      "step": 80
    },
    {
      "epoch": 0.7068607068607069,
      "grad_norm": 32.554352102245765,
      "learning_rate": 8.909157412340149e-07,
      "logits/chosen": -2.5185506343841553,
      "logits/rejected": -2.5202035903930664,
      "logps/chosen": -244.92233276367188,
      "logps/rejected": -222.07748413085938,
      "loss": 0.5411,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.8945465087890625,
      "rewards/margins": 1.484235167503357,
      "rewards/rejected": -0.589688777923584,
      "step": 85
    },
    {
      "epoch": 0.7484407484407485,
      "grad_norm": 41.80775702009017,
      "learning_rate": 8.765357330018055e-07,
      "logits/chosen": -2.527893543243408,
      "logits/rejected": -2.4960551261901855,
      "logps/chosen": -244.02468872070312,
      "logps/rejected": -226.1230926513672,
      "loss": 0.5433,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": 0.863143265247345,
      "rewards/margins": 1.4539084434509277,
      "rewards/rejected": -0.5907653570175171,
      "step": 90
    },
    {
      "epoch": 0.7900207900207901,
      "grad_norm": 35.599529291827764,
      "learning_rate": 8.613974319136957e-07,
      "logits/chosen": -2.546417713165283,
      "logits/rejected": -2.5654358863830566,
      "logps/chosen": -240.69375610351562,
      "logps/rejected": -204.02964782714844,
      "loss": 0.4936,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": 1.335752248764038,
      "rewards/margins": 1.7393795251846313,
      "rewards/rejected": -0.40362730622291565,
      "step": 95
    },
    {
      "epoch": 0.8316008316008316,
      "grad_norm": 42.869388172320754,
      "learning_rate": 8.455313244934324e-07,
      "logits/chosen": -2.597480297088623,
      "logits/rejected": -2.5727450847625732,
      "logps/chosen": -242.670166015625,
      "logps/rejected": -230.5889129638672,
      "loss": 0.5039,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.7388527393341064,
      "rewards/margins": 1.7855724096298218,
      "rewards/rejected": -1.0467197895050049,
      "step": 100
    },
    {
      "epoch": 0.8316008316008316,
      "eval_logits/chosen": -2.6104719638824463,
      "eval_logits/rejected": -2.5932347774505615,
      "eval_logps/chosen": -245.64724731445312,
      "eval_logps/rejected": -217.64056396484375,
      "eval_loss": 0.5226317048072815,
      "eval_rewards/accuracies": 0.8055555820465088,
      "eval_rewards/chosen": 0.2942599654197693,
      "eval_rewards/margins": 1.8596160411834717,
      "eval_rewards/rejected": -1.5653562545776367,
      "eval_runtime": 229.0549,
      "eval_samples_per_second": 14.931,
      "eval_steps_per_second": 0.236,
      "step": 100
    },
    {
      "epoch": 0.8731808731808732,
      "grad_norm": 63.89552939379053,
      "learning_rate": 8.289693629698563e-07,
      "logits/chosen": -2.6105101108551025,
      "logits/rejected": -2.6002025604248047,
      "logps/chosen": -250.3823699951172,
      "logps/rejected": -225.3841094970703,
      "loss": 0.4941,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.06903688609600067,
      "rewards/margins": 1.6266019344329834,
      "rewards/rejected": -1.5575649738311768,
      "step": 105
    },
    {
      "epoch": 0.9147609147609148,
      "grad_norm": 37.036407240889716,
      "learning_rate": 8.117449009293668e-07,
      "logits/chosen": -2.5854508876800537,
      "logits/rejected": -2.5774972438812256,
      "logps/chosen": -241.08596801757812,
      "logps/rejected": -205.6424102783203,
      "loss": 0.5011,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": 0.4706959128379822,
      "rewards/margins": 1.9714444875717163,
      "rewards/rejected": -1.5007483959197998,
      "step": 110
    },
    {
      "epoch": 0.9563409563409564,
      "grad_norm": 45.59653542600363,
      "learning_rate": 7.938926261462365e-07,
      "logits/chosen": -2.5712599754333496,
      "logits/rejected": -2.5764217376708984,
      "logps/chosen": -255.2756805419922,
      "logps/rejected": -225.7050323486328,
      "loss": 0.4869,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.6608749628067017,
      "rewards/margins": 1.9760288000106812,
      "rewards/rejected": -1.3151538372039795,
      "step": 115
    },
    {
      "epoch": 0.997920997920998,
      "grad_norm": 37.9241344117565,
      "learning_rate": 7.754484907260512e-07,
      "logits/chosen": -2.586732864379883,
      "logits/rejected": -2.567298650741577,
      "logps/chosen": -245.2010040283203,
      "logps/rejected": -231.32858276367188,
      "loss": 0.4961,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.0780966505408287,
      "rewards/margins": 1.6841703653335571,
      "rewards/rejected": -1.6060733795166016,
      "step": 120
    },
    {
      "epoch": 1.0395010395010396,
      "grad_norm": 16.41778883731705,
      "learning_rate": 7.564496387029531e-07,
      "logits/chosen": -2.5841448307037354,
      "logits/rejected": -2.5717430114746094,
      "logps/chosen": -242.82479858398438,
      "logps/rejected": -215.3890838623047,
      "loss": 0.2602,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.1772260665893555,
      "rewards/margins": 2.905987501144409,
      "rewards/rejected": -1.7287614345550537,
      "step": 125
    },
    {
      "epoch": 1.0810810810810811,
      "grad_norm": 24.87295730566652,
      "learning_rate": 7.369343312364993e-07,
      "logits/chosen": -2.6122641563415527,
      "logits/rejected": -2.5943500995635986,
      "logps/chosen": -239.027587890625,
      "logps/rejected": -240.2664031982422,
      "loss": 0.2533,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 1.7674968242645264,
      "rewards/margins": 3.0236265659332275,
      "rewards/rejected": -1.2561296224594116,
      "step": 130
    },
    {
      "epoch": 1.1226611226611227,
      "grad_norm": 18.64451849458332,
      "learning_rate": 7.16941869558779e-07,
      "logits/chosen": -2.642425060272217,
      "logits/rejected": -2.637680768966675,
      "logps/chosen": -247.26467895507812,
      "logps/rejected": -230.23666381835938,
      "loss": 0.2304,
      "rewards/accuracies": 0.887499988079071,
      "rewards/chosen": 1.8214725255966187,
      "rewards/margins": 3.1670384407043457,
      "rewards/rejected": -1.3455660343170166,
      "step": 135
    },
    {
      "epoch": 1.1642411642411643,
      "grad_norm": 20.326985348096915,
      "learning_rate": 6.965125158269618e-07,
      "logits/chosen": -2.6554737091064453,
      "logits/rejected": -2.6334750652313232,
      "logps/chosen": -243.7236328125,
      "logps/rejected": -223.4951629638672,
      "loss": 0.2805,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.6105072498321533,
      "rewards/margins": 3.326134443283081,
      "rewards/rejected": -1.7156273126602173,
      "step": 140
    },
    {
      "epoch": 1.2058212058212059,
      "grad_norm": 27.096042442305265,
      "learning_rate": 6.756874120406714e-07,
      "logits/chosen": -2.642449378967285,
      "logits/rejected": -2.6202588081359863,
      "logps/chosen": -237.80410766601562,
      "logps/rejected": -210.8230743408203,
      "loss": 0.2502,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 1.115569829940796,
      "rewards/margins": 2.6836743354797363,
      "rewards/rejected": -1.5681045055389404,
      "step": 145
    },
    {
      "epoch": 1.2474012474012475,
      "grad_norm": 19.90387897042561,
      "learning_rate": 6.545084971874736e-07,
      "logits/chosen": -2.5857880115509033,
      "logits/rejected": -2.6010355949401855,
      "logps/chosen": -234.38729858398438,
      "logps/rejected": -240.0408172607422,
      "loss": 0.2346,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.4643443822860718,
      "rewards/margins": 3.7548365592956543,
      "rewards/rejected": -2.290492534637451,
      "step": 150
    },
    {
      "epoch": 1.2474012474012475,
      "eval_logits/chosen": -2.572941303253174,
      "eval_logits/rejected": -2.551093816757202,
      "eval_logps/chosen": -243.4110565185547,
      "eval_logps/rejected": -220.85662841796875,
      "eval_loss": 0.4850994944572449,
      "eval_rewards/accuracies": 0.8356481194496155,
      "eval_rewards/chosen": 0.5178763270378113,
      "eval_rewards/margins": 2.4048376083374023,
      "eval_rewards/rejected": -1.8869613409042358,
      "eval_runtime": 229.0882,
      "eval_samples_per_second": 14.929,
      "eval_steps_per_second": 0.236,
      "step": 150
    },
    {
      "epoch": 1.288981288981289,
      "grad_norm": 25.39171705275189,
      "learning_rate": 6.330184227833375e-07,
      "logits/chosen": -2.556548595428467,
      "logits/rejected": -2.525398015975952,
      "logps/chosen": -241.52359008789062,
      "logps/rejected": -228.35836791992188,
      "loss": 0.2878,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.039872169494629,
      "rewards/margins": 3.453341245651245,
      "rewards/rejected": -2.413468837738037,
      "step": 155
    },
    {
      "epoch": 1.3305613305613306,
      "grad_norm": 25.836216262296954,
      "learning_rate": 6.112604669781572e-07,
      "logits/chosen": -2.509721517562866,
      "logits/rejected": -2.480980396270752,
      "logps/chosen": -250.9086151123047,
      "logps/rejected": -219.1513214111328,
      "loss": 0.2634,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.049938440322876,
      "rewards/margins": 3.7404415607452393,
      "rewards/rejected": -2.690502643585205,
      "step": 160
    },
    {
      "epoch": 1.3721413721413722,
      "grad_norm": 23.998128784929428,
      "learning_rate": 5.892784473993183e-07,
      "logits/chosen": -2.479841709136963,
      "logits/rejected": -2.4423975944519043,
      "logps/chosen": -255.5587615966797,
      "logps/rejected": -238.94900512695312,
      "loss": 0.291,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.9023858904838562,
      "rewards/margins": 3.534865617752075,
      "rewards/rejected": -2.632479667663574,
      "step": 165
    },
    {
      "epoch": 1.4137214137214138,
      "grad_norm": 25.390125036750348,
      "learning_rate": 5.671166329088277e-07,
      "logits/chosen": -2.52769136428833,
      "logits/rejected": -2.507841110229492,
      "logps/chosen": -256.8693542480469,
      "logps/rejected": -250.11770629882812,
      "loss": 0.3037,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 0.9882985353469849,
      "rewards/margins": 3.3482158184051514,
      "rewards/rejected": -2.3599178791046143,
      "step": 170
    },
    {
      "epoch": 1.4553014553014554,
      "grad_norm": 22.126640358297035,
      "learning_rate": 5.448196544517167e-07,
      "logits/chosen": -2.577094554901123,
      "logits/rejected": -2.5509932041168213,
      "logps/chosen": -272.64605712890625,
      "logps/rejected": -238.14901733398438,
      "loss": 0.3057,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.49201500415802,
      "rewards/margins": 3.8457350730895996,
      "rewards/rejected": -2.3537204265594482,
      "step": 175
    },
    {
      "epoch": 1.496881496881497,
      "grad_norm": 22.272559333838274,
      "learning_rate": 5.224324151752575e-07,
      "logits/chosen": -2.63065767288208,
      "logits/rejected": -2.609647274017334,
      "logps/chosen": -250.0604705810547,
      "logps/rejected": -220.90280151367188,
      "loss": 0.3264,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 1.3409287929534912,
      "rewards/margins": 3.233898162841797,
      "rewards/rejected": -1.8929691314697266,
      "step": 180
    },
    {
      "epoch": 1.5384615384615383,
      "grad_norm": 27.539818610981982,
      "learning_rate": 5e-07,
      "logits/chosen": -2.6626243591308594,
      "logits/rejected": -2.65144681930542,
      "logps/chosen": -254.73739624023438,
      "logps/rejected": -242.0198516845703,
      "loss": 0.3146,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.082256555557251,
      "rewards/margins": 3.2332160472869873,
      "rewards/rejected": -2.1509594917297363,
      "step": 185
    },
    {
      "epoch": 1.5800415800415801,
      "grad_norm": 20.09074928627218,
      "learning_rate": 4.775675848247427e-07,
      "logits/chosen": -2.6996893882751465,
      "logits/rejected": -2.692633628845215,
      "logps/chosen": -238.6735382080078,
      "logps/rejected": -252.790771484375,
      "loss": 0.2521,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.9300445318222046,
      "rewards/margins": 3.3958065509796143,
      "rewards/rejected": -2.465762138366699,
      "step": 190
    },
    {
      "epoch": 1.6216216216216215,
      "grad_norm": 27.310137552572137,
      "learning_rate": 4.5518034554828327e-07,
      "logits/chosen": -2.6913933753967285,
      "logits/rejected": -2.692023992538452,
      "logps/chosen": -246.1246795654297,
      "logps/rejected": -250.9800262451172,
      "loss": 0.3032,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.2431519031524658,
      "rewards/margins": 3.4656929969787598,
      "rewards/rejected": -2.2225406169891357,
      "step": 195
    },
    {
      "epoch": 1.6632016632016633,
      "grad_norm": 19.508980724848932,
      "learning_rate": 4.328833670911724e-07,
      "logits/chosen": -2.6803059577941895,
      "logits/rejected": -2.6774773597717285,
      "logps/chosen": -247.5465850830078,
      "logps/rejected": -224.4514923095703,
      "loss": 0.26,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 1.1387991905212402,
      "rewards/margins": 3.076381206512451,
      "rewards/rejected": -1.9375823736190796,
      "step": 200
    },
    {
      "epoch": 1.6632016632016633,
      "eval_logits/chosen": -2.676579475402832,
      "eval_logits/rejected": -2.6650500297546387,
      "eval_logps/chosen": -240.44088745117188,
      "eval_logps/rejected": -219.10662841796875,
      "eval_loss": 0.4691601097583771,
      "eval_rewards/accuracies": 0.8263888955116272,
      "eval_rewards/chosen": 0.8148964643478394,
      "eval_rewards/margins": 2.5268585681915283,
      "eval_rewards/rejected": -1.711962103843689,
      "eval_runtime": 229.072,
      "eval_samples_per_second": 14.93,
      "eval_steps_per_second": 0.236,
      "step": 200
    },
    {
      "epoch": 1.7047817047817047,
      "grad_norm": 21.69325339786789,
      "learning_rate": 4.107215526006817e-07,
      "logits/chosen": -2.6705069541931152,
      "logits/rejected": -2.671739101409912,
      "logps/chosen": -255.0220184326172,
      "logps/rejected": -251.6207275390625,
      "loss": 0.3021,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.3921308517456055,
      "rewards/margins": 3.475137710571289,
      "rewards/rejected": -2.0830068588256836,
      "step": 205
    },
    {
      "epoch": 1.7463617463617465,
      "grad_norm": 18.09505177508146,
      "learning_rate": 3.8873953302184283e-07,
      "logits/chosen": -2.655519485473633,
      "logits/rejected": -2.64316463470459,
      "logps/chosen": -244.9538116455078,
      "logps/rejected": -227.49459838867188,
      "loss": 0.2406,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.577366828918457,
      "rewards/margins": 4.07238245010376,
      "rewards/rejected": -2.4950149059295654,
      "step": 210
    },
    {
      "epoch": 1.7879417879417878,
      "grad_norm": 18.41796059615618,
      "learning_rate": 3.669815772166625e-07,
      "logits/chosen": -2.6481261253356934,
      "logits/rejected": -2.6404080390930176,
      "logps/chosen": -245.0645294189453,
      "logps/rejected": -244.56564331054688,
      "loss": 0.2737,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.9804733395576477,
      "rewards/margins": 3.504237413406372,
      "rewards/rejected": -2.5237643718719482,
      "step": 215
    },
    {
      "epoch": 1.8295218295218296,
      "grad_norm": 18.16045294910453,
      "learning_rate": 3.454915028125263e-07,
      "logits/chosen": -2.6570143699645996,
      "logits/rejected": -2.648756980895996,
      "logps/chosen": -232.47579956054688,
      "logps/rejected": -233.9488525390625,
      "loss": 0.2505,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 0.658257782459259,
      "rewards/margins": 3.45349383354187,
      "rewards/rejected": -2.795236349105835,
      "step": 220
    },
    {
      "epoch": 1.871101871101871,
      "grad_norm": 24.59443441003638,
      "learning_rate": 3.243125879593286e-07,
      "logits/chosen": -2.672626495361328,
      "logits/rejected": -2.6666738986968994,
      "logps/chosen": -251.75491333007812,
      "logps/rejected": -237.9063262939453,
      "loss": 0.2856,
      "rewards/accuracies": 0.8687499761581421,
      "rewards/chosen": 0.6436173915863037,
      "rewards/margins": 3.0495057106018066,
      "rewards/rejected": -2.405888080596924,
      "step": 225
    },
    {
      "epoch": 1.9126819126819128,
      "grad_norm": 21.71857811928758,
      "learning_rate": 3.034874841730382e-07,
      "logits/chosen": -2.655111789703369,
      "logits/rejected": -2.647090435028076,
      "logps/chosen": -254.3959503173828,
      "logps/rejected": -257.03790283203125,
      "loss": 0.2676,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 0.7714903950691223,
      "rewards/margins": 3.8913707733154297,
      "rewards/rejected": -3.119880199432373,
      "step": 230
    },
    {
      "epoch": 1.9542619542619541,
      "grad_norm": 24.999281656170186,
      "learning_rate": 2.8305813044122093e-07,
      "logits/chosen": -2.662872791290283,
      "logits/rejected": -2.6576457023620605,
      "logps/chosen": -253.2953643798828,
      "logps/rejected": -244.046630859375,
      "loss": 0.2667,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.3658444881439209,
      "rewards/margins": 3.359410047531128,
      "rewards/rejected": -2.993565320968628,
      "step": 235
    },
    {
      "epoch": 1.995841995841996,
      "grad_norm": 17.557850245563603,
      "learning_rate": 2.6306566876350067e-07,
      "logits/chosen": -2.6753599643707275,
      "logits/rejected": -2.6530680656433105,
      "logps/chosen": -236.2633056640625,
      "logps/rejected": -236.9930419921875,
      "loss": 0.3106,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 0.48787975311279297,
      "rewards/margins": 3.501561403274536,
      "rewards/rejected": -3.013681411743164,
      "step": 240
    },
    {
      "epoch": 2.0374220374220373,
      "grad_norm": 13.42453839113005,
      "learning_rate": 2.4355036129704696e-07,
      "logits/chosen": -2.675271987915039,
      "logits/rejected": -2.6660828590393066,
      "logps/chosen": -259.6005859375,
      "logps/rejected": -255.7154083251953,
      "loss": 0.1495,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.9453225135803223,
      "rewards/margins": 3.974649429321289,
      "rewards/rejected": -3.0293262004852295,
      "step": 245
    },
    {
      "epoch": 2.079002079002079,
      "grad_norm": 19.574331273158283,
      "learning_rate": 2.2455150927394878e-07,
      "logits/chosen": -2.6700210571289062,
      "logits/rejected": -2.6747846603393555,
      "logps/chosen": -254.7049560546875,
      "logps/rejected": -240.4936981201172,
      "loss": 0.1628,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.2429602146148682,
      "rewards/margins": 4.115931987762451,
      "rewards/rejected": -2.872972011566162,
      "step": 250
    },
    {
      "epoch": 2.079002079002079,
      "eval_logits/chosen": -2.6929080486297607,
      "eval_logits/rejected": -2.680229663848877,
      "eval_logps/chosen": -246.06834411621094,
      "eval_logps/rejected": -226.5530242919922,
      "eval_loss": 0.4653815031051636,
      "eval_rewards/accuracies": 0.8263888955116272,
      "eval_rewards/chosen": 0.25215068459510803,
      "eval_rewards/margins": 2.7087526321411133,
      "eval_rewards/rejected": -2.456601858139038,
      "eval_runtime": 229.1641,
      "eval_samples_per_second": 14.924,
      "eval_steps_per_second": 0.236,
      "step": 250
    },
    {
      "epoch": 2.1205821205821205,
      "grad_norm": 13.176592126973915,
      "learning_rate": 2.0610737385376348e-07,
      "logits/chosen": -2.7003557682037354,
      "logits/rejected": -2.6898903846740723,
      "logps/chosen": -231.67977905273438,
      "logps/rejected": -234.54928588867188,
      "loss": 0.1524,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 0.8713158369064331,
      "rewards/margins": 3.862780809402466,
      "rewards/rejected": -2.9914650917053223,
      "step": 255
    },
    {
      "epoch": 2.1621621621621623,
      "grad_norm": 13.531997132878638,
      "learning_rate": 1.8825509907063326e-07,
      "logits/chosen": -2.7138633728027344,
      "logits/rejected": -2.712902545928955,
      "logps/chosen": -250.47555541992188,
      "logps/rejected": -216.4556427001953,
      "loss": 0.1555,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 1.3138505220413208,
      "rewards/margins": 4.086125373840332,
      "rewards/rejected": -2.77227520942688,
      "step": 260
    },
    {
      "epoch": 2.2037422037422036,
      "grad_norm": 15.615532113289877,
      "learning_rate": 1.710306370301437e-07,
      "logits/chosen": -2.713639736175537,
      "logits/rejected": -2.7185206413269043,
      "logps/chosen": -238.790771484375,
      "logps/rejected": -243.9224090576172,
      "loss": 0.1611,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.4643341302871704,
      "rewards/margins": 4.29836368560791,
      "rewards/rejected": -2.8340296745300293,
      "step": 265
    },
    {
      "epoch": 2.2453222453222454,
      "grad_norm": 16.65673651133407,
      "learning_rate": 1.5446867550656767e-07,
      "logits/chosen": -2.7022347450256348,
      "logits/rejected": -2.689222812652588,
      "logps/chosen": -224.8180694580078,
      "logps/rejected": -242.55191040039062,
      "loss": 0.1539,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.495180368423462,
      "rewards/margins": 3.86863374710083,
      "rewards/rejected": -2.3734536170959473,
      "step": 270
    },
    {
      "epoch": 2.286902286902287,
      "grad_norm": 14.004772081888808,
      "learning_rate": 1.3860256808630427e-07,
      "logits/chosen": -2.6863961219787598,
      "logits/rejected": -2.6838507652282715,
      "logps/chosen": -235.44979858398438,
      "logps/rejected": -255.1193389892578,
      "loss": 0.155,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.9984718561172485,
      "rewards/margins": 4.826051235198975,
      "rewards/rejected": -2.8275790214538574,
      "step": 275
    },
    {
      "epoch": 2.3284823284823286,
      "grad_norm": 15.548014252517618,
      "learning_rate": 1.2346426699819456e-07,
      "logits/chosen": -2.6768174171447754,
      "logits/rejected": -2.67870831489563,
      "logps/chosen": -217.78097534179688,
      "logps/rejected": -228.59716796875,
      "loss": 0.1474,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.7014672756195068,
      "rewards/margins": 4.063814640045166,
      "rewards/rejected": -2.362347364425659,
      "step": 280
    },
    {
      "epoch": 2.37006237006237,
      "grad_norm": 15.669379708365364,
      "learning_rate": 1.090842587659851e-07,
      "logits/chosen": -2.6670119762420654,
      "logits/rejected": -2.658346652984619,
      "logps/chosen": -230.91799926757812,
      "logps/rejected": -255.7079620361328,
      "loss": 0.1414,
      "rewards/accuracies": 0.96875,
      "rewards/chosen": 1.8583574295043945,
      "rewards/margins": 4.283419132232666,
      "rewards/rejected": -2.4250617027282715,
      "step": 285
    },
    {
      "epoch": 2.4116424116424118,
      "grad_norm": 17.599310517059866,
      "learning_rate": 9.549150281252632e-08,
      "logits/chosen": -2.6570022106170654,
      "logits/rejected": -2.651043653488159,
      "logps/chosen": -244.97250366210938,
      "logps/rejected": -237.3176727294922,
      "loss": 0.1618,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.7519409656524658,
      "rewards/margins": 4.170150279998779,
      "rewards/rejected": -2.4182095527648926,
      "step": 290
    },
    {
      "epoch": 2.453222453222453,
      "grad_norm": 17.368023555799656,
      "learning_rate": 8.271337313934867e-08,
      "logits/chosen": -2.651902198791504,
      "logits/rejected": -2.6466565132141113,
      "logps/chosen": -238.5611114501953,
      "logps/rejected": -229.7045440673828,
      "loss": 0.1584,
      "rewards/accuracies": 0.9375,
      "rewards/chosen": 1.7158390283584595,
      "rewards/margins": 4.3892717361450195,
      "rewards/rejected": -2.6734321117401123,
      "step": 295
    },
    {
      "epoch": 2.494802494802495,
      "grad_norm": 19.25786544352152,
      "learning_rate": 7.077560319906694e-08,
      "logits/chosen": -2.657745838165283,
      "logits/rejected": -2.642855405807495,
      "logps/chosen": -224.6698760986328,
      "logps/rejected": -235.5657196044922,
      "loss": 0.1808,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.6349939107894897,
      "rewards/margins": 4.394418239593506,
      "rewards/rejected": -2.7594242095947266,
      "step": 300
    },
    {
      "epoch": 2.494802494802495,
      "eval_logits/chosen": -2.6527817249298096,
      "eval_logits/rejected": -2.6391522884368896,
      "eval_logps/chosen": -240.36119079589844,
      "eval_logps/rejected": -222.10067749023438,
      "eval_loss": 0.4721406400203705,
      "eval_rewards/accuracies": 0.8240740895271301,
      "eval_rewards/chosen": 0.822864830493927,
      "eval_rewards/margins": 2.834231376647949,
      "eval_rewards/rejected": -2.011366605758667,
      "eval_runtime": 229.0179,
      "eval_samples_per_second": 14.933,
      "eval_steps_per_second": 0.236,
      "step": 300
    },
    {
      "epoch": 2.5363825363825363,
      "grad_norm": 13.930273404809197,
      "learning_rate": 5.9702234071631e-08,
      "logits/chosen": -2.654966115951538,
      "logits/rejected": -2.6402499675750732,
      "logps/chosen": -227.8262481689453,
      "logps/rejected": -238.726318359375,
      "loss": 0.1607,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.5776718854904175,
      "rewards/margins": 4.08219575881958,
      "rewards/rejected": -2.504523754119873,
      "step": 305
    },
    {
      "epoch": 2.577962577962578,
      "grad_norm": 19.619559321496773,
      "learning_rate": 4.951556604879048e-08,
      "logits/chosen": -2.6543350219726562,
      "logits/rejected": -2.641832113265991,
      "logps/chosen": -251.985595703125,
      "logps/rejected": -252.0045623779297,
      "loss": 0.1511,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.890110969543457,
      "rewards/margins": 4.617300510406494,
      "rewards/rejected": -2.727189064025879,
      "step": 310
    },
    {
      "epoch": 2.6195426195426195,
      "grad_norm": 14.408775987552131,
      "learning_rate": 4.023611372427471e-08,
      "logits/chosen": -2.646648406982422,
      "logits/rejected": -2.647120952606201,
      "logps/chosen": -236.2719268798828,
      "logps/rejected": -235.2129669189453,
      "loss": 0.1665,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.3996338844299316,
      "rewards/margins": 4.081356048583984,
      "rewards/rejected": -2.681722402572632,
      "step": 315
    },
    {
      "epoch": 2.6611226611226613,
      "grad_norm": 18.476337042561774,
      "learning_rate": 3.188256468013139e-08,
      "logits/chosen": -2.641806125640869,
      "logits/rejected": -2.6388161182403564,
      "logps/chosen": -237.56344604492188,
      "logps/rejected": -257.267333984375,
      "loss": 0.1717,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.5857443809509277,
      "rewards/margins": 4.466619968414307,
      "rewards/rejected": -2.8808751106262207,
      "step": 320
    },
    {
      "epoch": 2.7027027027027026,
      "grad_norm": 16.101961992486956,
      "learning_rate": 2.4471741852423233e-08,
      "logits/chosen": -2.6497855186462402,
      "logits/rejected": -2.646726608276367,
      "logps/chosen": -252.9170684814453,
      "logps/rejected": -248.69717407226562,
      "loss": 0.1617,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.426393747329712,
      "rewards/margins": 4.511406898498535,
      "rewards/rejected": -3.0850136280059814,
      "step": 325
    },
    {
      "epoch": 2.7442827442827444,
      "grad_norm": 21.330859626282905,
      "learning_rate": 1.8018569652073378e-08,
      "logits/chosen": -2.653249979019165,
      "logits/rejected": -2.644749879837036,
      "logps/chosen": -243.327392578125,
      "logps/rejected": -233.7000732421875,
      "loss": 0.1707,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.2815972566604614,
      "rewards/margins": 3.986734390258789,
      "rewards/rejected": -2.705137014389038,
      "step": 330
    },
    {
      "epoch": 2.785862785862786,
      "grad_norm": 23.976094119041996,
      "learning_rate": 1.253604390908819e-08,
      "logits/chosen": -2.645918130874634,
      "logits/rejected": -2.6493842601776123,
      "logps/chosen": -240.95263671875,
      "logps/rejected": -252.9552764892578,
      "loss": 0.1545,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.5216797590255737,
      "rewards/margins": 4.699650764465332,
      "rewards/rejected": -3.177971124649048,
      "step": 335
    },
    {
      "epoch": 2.8274428274428276,
      "grad_norm": 18.590678935876824,
      "learning_rate": 8.035205700685165e-09,
      "logits/chosen": -2.6513190269470215,
      "logits/rejected": -2.6512649059295654,
      "logps/chosen": -253.2542724609375,
      "logps/rejected": -256.85882568359375,
      "loss": 0.1629,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.507804274559021,
      "rewards/margins": 4.7108473777771,
      "rewards/rejected": -3.2030436992645264,
      "step": 340
    },
    {
      "epoch": 2.869022869022869,
      "grad_norm": 16.529254599149706,
      "learning_rate": 4.5251191160326495e-09,
      "logits/chosen": -2.651240825653076,
      "logits/rejected": -2.634026288986206,
      "logps/chosen": -252.3319549560547,
      "logps/rejected": -246.98855590820312,
      "loss": 0.1717,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.2555811405181885,
      "rewards/margins": 4.030848503112793,
      "rewards/rejected": -2.7752671241760254,
      "step": 345
    },
    {
      "epoch": 2.9106029106029108,
      "grad_norm": 18.48041608280076,
      "learning_rate": 2.0128530023804656e-09,
      "logits/chosen": -2.6509900093078613,
      "logits/rejected": -2.6280765533447266,
      "logps/chosen": -248.33200073242188,
      "logps/rejected": -220.7266845703125,
      "loss": 0.1514,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.2462139129638672,
      "rewards/margins": 3.974480390548706,
      "rewards/rejected": -2.728266954421997,
      "step": 350
    },
    {
      "epoch": 2.9106029106029108,
      "eval_logits/chosen": -2.6499643325805664,
      "eval_logits/rejected": -2.6346051692962646,
      "eval_logps/chosen": -243.1808624267578,
      "eval_logps/rejected": -225.02040100097656,
      "eval_loss": 0.47360217571258545,
      "eval_rewards/accuracies": 0.8217592835426331,
      "eval_rewards/chosen": 0.5408979058265686,
      "eval_rewards/margins": 2.844236135482788,
      "eval_rewards/rejected": -2.303337812423706,
      "eval_runtime": 230.4659,
      "eval_samples_per_second": 14.84,
      "eval_steps_per_second": 0.234,
      "step": 350
    },
    {
      "epoch": 2.952182952182952,
      "grad_norm": 17.613598063550807,
      "learning_rate": 5.034667293427053e-10,
      "logits/chosen": -2.661424398422241,
      "logits/rejected": -2.625995635986328,
      "logps/chosen": -223.80783081054688,
      "logps/rejected": -229.94912719726562,
      "loss": 0.1525,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.005284070968628,
      "rewards/margins": 3.9230079650878906,
      "rewards/rejected": -2.9177234172821045,
      "step": 355
    },
    {
      "epoch": 2.993762993762994,
      "grad_norm": 20.660357349331896,
      "learning_rate": 0.0,
      "logits/chosen": -2.6404199600219727,
      "logits/rejected": -2.6347861289978027,
      "logps/chosen": -234.76513671875,
      "logps/rejected": -236.2282257080078,
      "loss": 0.1438,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 1.2399814128875732,
      "rewards/margins": 4.341221809387207,
      "rewards/rejected": -3.1012401580810547,
      "step": 360
    },
    {
      "epoch": 2.993762993762994,
      "step": 360,
      "total_flos": 4244978440077312.0,
      "train_loss": 0.33558651275104945,
      "train_runtime": 13885.8564,
      "train_samples_per_second": 6.648,
      "train_steps_per_second": 0.026
    }
  ],
  "logging_steps": 5,
  "max_steps": 360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 4244978440077312.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}