{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.803738317757009,
  "eval_steps": 50,
  "global_step": 300,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.04672897196261682,
      "grad_norm": 68.30027051845477,
      "learning_rate": 5e-07,
      "logits/chosen": -2.7332546710968018,
      "logits/rejected": -2.7345798015594482,
      "logps/chosen": -246.6896209716797,
      "logps/rejected": -212.1261749267578,
      "loss": 0.6905,
      "rewards/accuracies": 0.33125001192092896,
      "rewards/chosen": 0.01393141783773899,
      "rewards/margins": 0.007549271918833256,
      "rewards/rejected": 0.006382145918905735,
      "step": 5
    },
    {
      "epoch": 0.09345794392523364,
      "grad_norm": 47.12543235245713,
      "learning_rate": 1e-06,
      "logits/chosen": -2.690162181854248,
      "logits/rejected": -2.6820242404937744,
      "logps/chosen": -237.89517211914062,
      "logps/rejected": -218.36865234375,
      "loss": 0.6502,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.4055875837802887,
      "rewards/margins": 0.14588704705238342,
      "rewards/rejected": 0.2597005069255829,
      "step": 10
    },
    {
      "epoch": 0.14018691588785046,
      "grad_norm": 49.480284694367455,
      "learning_rate": 9.993623730611148e-07,
      "logits/chosen": -2.56203031539917,
      "logits/rejected": -2.5295283794403076,
      "logps/chosen": -215.8556365966797,
      "logps/rejected": -206.70339965820312,
      "loss": 0.6521,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 1.2055972814559937,
      "rewards/margins": 0.3723347783088684,
      "rewards/rejected": 0.8332624435424805,
      "step": 15
    },
    {
      "epoch": 0.18691588785046728,
      "grad_norm": 49.42485896222818,
      "learning_rate": 9.97451118516912e-07,
      "logits/chosen": -2.4396374225616455,
      "logits/rejected": -2.4257354736328125,
      "logps/chosen": -233.15951538085938,
      "logps/rejected": -192.3885955810547,
      "loss": 0.6234,
      "rewards/accuracies": 0.668749988079071,
      "rewards/chosen": 1.3042352199554443,
      "rewards/margins": 0.6797620058059692,
      "rewards/rejected": 0.6244731545448303,
      "step": 20
    },
    {
      "epoch": 0.2336448598130841,
      "grad_norm": 43.823143196477226,
      "learning_rate": 9.94271111036929e-07,
      "logits/chosen": -2.318066120147705,
      "logits/rejected": -2.304948091506958,
      "logps/chosen": -241.57180786132812,
      "logps/rejected": -191.6623992919922,
      "loss": 0.6435,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 1.0954402685165405,
      "rewards/margins": 0.5757324695587158,
      "rewards/rejected": 0.5197077393531799,
      "step": 25
    },
    {
      "epoch": 0.2803738317757009,
      "grad_norm": 39.78735079002501,
      "learning_rate": 9.898304612549066e-07,
      "logits/chosen": -2.3289685249328613,
      "logits/rejected": -2.326596736907959,
      "logps/chosen": -255.0546875,
      "logps/rejected": -215.7193145751953,
      "loss": 0.6003,
      "rewards/accuracies": 0.706250011920929,
      "rewards/chosen": 1.374457597732544,
      "rewards/margins": 1.0344688892364502,
      "rewards/rejected": 0.33998847007751465,
      "step": 30
    },
    {
      "epoch": 0.32710280373831774,
      "grad_norm": 43.394927490331185,
      "learning_rate": 9.841404950825536e-07,
      "logits/chosen": -2.3670554161071777,
      "logits/rejected": -2.3704476356506348,
      "logps/chosen": -242.648681640625,
      "logps/rejected": -209.353759765625,
      "loss": 0.5919,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.954569935798645,
      "rewards/margins": 1.0306309461593628,
      "rewards/rejected": -0.07606112957000732,
      "step": 35
    },
    {
      "epoch": 0.37383177570093457,
      "grad_norm": 46.37512759709472,
      "learning_rate": 9.77215724822721e-07,
      "logits/chosen": -2.437983274459839,
      "logits/rejected": -2.465634822845459,
      "logps/chosen": -254.10226440429688,
      "logps/rejected": -226.6571502685547,
      "loss": 0.648,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.6253764033317566,
      "rewards/margins": 0.8199795484542847,
      "rewards/rejected": -0.1946031153202057,
      "step": 40
    },
    {
      "epoch": 0.4205607476635514,
      "grad_norm": 44.696777081131444,
      "learning_rate": 9.69073812155662e-07,
      "logits/chosen": -2.4394752979278564,
      "logits/rejected": -2.4159750938415527,
      "logps/chosen": -246.6658172607422,
      "logps/rejected": -240.363525390625,
      "loss": 0.5918,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.7499775886535645,
      "rewards/margins": 1.0236563682556152,
      "rewards/rejected": -0.2736787497997284,
      "step": 45
    },
    {
      "epoch": 0.4672897196261682,
      "grad_norm": 38.923198317781896,
      "learning_rate": 9.597355230927788e-07,
      "logits/chosen": -2.4589362144470215,
      "logits/rejected": -2.412837505340576,
      "logps/chosen": -264.4033203125,
      "logps/rejected": -213.1973876953125,
      "loss": 0.576,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": 1.1025102138519287,
      "rewards/margins": 1.2783615589141846,
      "rewards/rejected": -0.175851508975029,
      "step": 50
    },
    {
      "epoch": 0.4672897196261682,
      "eval_logits/chosen": -2.496382474899292,
      "eval_logits/rejected": -2.4706027507781982,
      "eval_logps/chosen": -223.9072723388672,
      "eval_logps/rejected": -212.02557373046875,
      "eval_loss": 0.5725350379943848,
      "eval_rewards/accuracies": 0.7317708134651184,
      "eval_rewards/chosen": 0.8729944229125977,
      "eval_rewards/margins": 0.9930557608604431,
      "eval_rewards/rejected": -0.12006133049726486,
      "eval_runtime": 203.179,
      "eval_samples_per_second": 14.962,
      "eval_steps_per_second": 0.236,
      "step": 50
    },
    {
      "epoch": 0.514018691588785,
      "grad_norm": 32.87244621664002,
      "learning_rate": 9.4922467501275e-07,
      "logits/chosen": -2.5147769451141357,
      "logits/rejected": -2.4755141735076904,
      "logps/chosen": -248.0165557861328,
      "logps/rejected": -201.71705627441406,
      "loss": 0.5771,
      "rewards/accuracies": 0.7437499761581421,
      "rewards/chosen": 0.9974912405014038,
      "rewards/margins": 1.1693007946014404,
      "rewards/rejected": -0.17180955410003662,
      "step": 55
    },
    {
      "epoch": 0.5607476635514018,
      "grad_norm": 41.4567929981585,
      "learning_rate": 9.375680759151206e-07,
      "logits/chosen": -2.517082452774048,
      "logits/rejected": -2.506866931915283,
      "logps/chosen": -239.5508575439453,
      "logps/rejected": -215.45339965820312,
      "loss": 0.5856,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.7699365615844727,
      "rewards/margins": 1.1056894063949585,
      "rewards/rejected": -0.33575284481048584,
      "step": 60
    },
    {
      "epoch": 0.6074766355140186,
      "grad_norm": 42.95403443814742,
      "learning_rate": 9.247954560462927e-07,
      "logits/chosen": -2.575209140777588,
      "logits/rejected": -2.5410640239715576,
      "logps/chosen": -235.7403106689453,
      "logps/rejected": -221.4304656982422,
      "loss": 0.5815,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.6247920989990234,
      "rewards/margins": 1.079342246055603,
      "rewards/rejected": -0.454550176858902,
      "step": 65
    },
    {
      "epoch": 0.6542056074766355,
      "grad_norm": 39.08891854742411,
      "learning_rate": 9.109393920723001e-07,
      "logits/chosen": -2.584603786468506,
      "logits/rejected": -2.5690274238586426,
      "logps/chosen": -267.28839111328125,
      "logps/rejected": -231.07785034179688,
      "loss": 0.5626,
      "rewards/accuracies": 0.7562500238418579,
      "rewards/chosen": 0.6171663403511047,
      "rewards/margins": 1.1559643745422363,
      "rewards/rejected": -0.5387979745864868,
      "step": 70
    },
    {
      "epoch": 0.7009345794392523,
      "grad_norm": 44.455775233540244,
      "learning_rate": 8.960352239917699e-07,
      "logits/chosen": -2.566195011138916,
      "logits/rejected": -2.549044132232666,
      "logps/chosen": -246.45858764648438,
      "logps/rejected": -222.82357788085938,
      "loss": 0.5982,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.6170978546142578,
      "rewards/margins": 1.267897367477417,
      "rewards/rejected": -0.6507996916770935,
      "step": 75
    },
    {
      "epoch": 0.7476635514018691,
      "grad_norm": 32.16569720584001,
      "learning_rate": 8.801209650009814e-07,
      "logits/chosen": -2.4980108737945557,
      "logits/rejected": -2.4944815635681152,
      "logps/chosen": -251.861572265625,
      "logps/rejected": -227.2343292236328,
      "loss": 0.4956,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.6364807486534119,
      "rewards/margins": 1.5403988361358643,
      "rewards/rejected": -0.9039181470870972,
      "step": 80
    },
    {
      "epoch": 0.794392523364486,
      "grad_norm": 40.106812622105444,
      "learning_rate": 8.632372045409141e-07,
      "logits/chosen": -2.510608196258545,
      "logits/rejected": -2.50504994392395,
      "logps/chosen": -264.02777099609375,
      "logps/rejected": -228.0398406982422,
      "loss": 0.5527,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.6684669852256775,
      "rewards/margins": 1.134697675704956,
      "rewards/rejected": -0.4662308692932129,
      "step": 85
    },
    {
      "epoch": 0.8411214953271028,
      "grad_norm": 41.44913213337072,
      "learning_rate": 8.454270047735642e-07,
      "logits/chosen": -2.4945034980773926,
      "logits/rejected": -2.4969065189361572,
      "logps/chosen": -266.8227844238281,
      "logps/rejected": -218.293701171875,
      "loss": 0.5291,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": 0.4074540138244629,
      "rewards/margins": 1.6132476329803467,
      "rewards/rejected": -1.2057937383651733,
      "step": 90
    },
    {
      "epoch": 0.8878504672897196,
      "grad_norm": 38.304999982614596,
      "learning_rate": 8.267357907515661e-07,
      "logits/chosen": -2.4969687461853027,
      "logits/rejected": -2.4846863746643066,
      "logps/chosen": -253.90396118164062,
      "logps/rejected": -226.99136352539062,
      "loss": 0.531,
      "rewards/accuracies": 0.71875,
      "rewards/chosen": -0.14072401821613312,
      "rewards/margins": 1.3943620920181274,
      "rewards/rejected": -1.5350860357284546,
      "step": 95
    },
    {
      "epoch": 0.9345794392523364,
      "grad_norm": 35.63875623202816,
      "learning_rate": 8.072112345612433e-07,
      "logits/chosen": -2.468038558959961,
      "logits/rejected": -2.446315288543701,
      "logps/chosen": -242.86328125,
      "logps/rejected": -230.5182342529297,
      "loss": 0.508,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.023806383833289146,
      "rewards/margins": 1.7525684833526611,
      "rewards/rejected": -1.7287616729736328,
      "step": 100
    },
    {
      "epoch": 0.9345794392523364,
      "eval_logits/chosen": -2.460448980331421,
      "eval_logits/rejected": -2.4439258575439453,
      "eval_logps/chosen": -234.71864318847656,
      "eval_logps/rejected": -228.7583770751953,
      "eval_loss": 0.5507261753082275,
      "eval_rewards/accuracies": 0.7864583134651184,
      "eval_rewards/chosen": -0.2081437110900879,
      "eval_rewards/margins": 1.5851994752883911,
      "eval_rewards/rejected": -1.793343186378479,
      "eval_runtime": 202.7006,
      "eval_samples_per_second": 14.997,
      "eval_steps_per_second": 0.237,
      "step": 100
    },
    {
      "epoch": 0.9813084112149533,
      "grad_norm": 34.55937028630903,
      "learning_rate": 7.869031337345827e-07,
      "logits/chosen": -2.4676289558410645,
      "logits/rejected": -2.4437384605407715,
      "logps/chosen": -266.91009521484375,
      "logps/rejected": -226.27108764648438,
      "loss": 0.5141,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -0.0003374546649865806,
      "rewards/margins": 1.9741418361663818,
      "rewards/rejected": -1.9744793176651,
      "step": 105
    },
    {
      "epoch": 1.02803738317757,
      "grad_norm": 20.439479426811744,
      "learning_rate": 7.658632842402432e-07,
      "logits/chosen": -2.4328935146331787,
      "logits/rejected": -2.428864002227783,
      "logps/chosen": -264.23260498046875,
      "logps/rejected": -235.5162811279297,
      "loss": 0.3938,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 0.19133757054805756,
      "rewards/margins": 2.5184836387634277,
      "rewards/rejected": -2.327146053314209,
      "step": 110
    },
    {
      "epoch": 1.074766355140187,
      "grad_norm": 16.4772605014783,
      "learning_rate": 7.441453483775353e-07,
      "logits/chosen": -2.4627161026000977,
      "logits/rejected": -2.4542901515960693,
      "logps/chosen": -253.79190063476562,
      "logps/rejected": -240.01882934570312,
      "loss": 0.2137,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 0.7574207186698914,
      "rewards/margins": 2.823213577270508,
      "rewards/rejected": -2.0657927989959717,
      "step": 115
    },
    {
      "epoch": 1.1214953271028036,
      "grad_norm": 21.079385505903296,
      "learning_rate": 7.218047179103112e-07,
      "logits/chosen": -2.483931064605713,
      "logits/rejected": -2.4513020515441895,
      "logps/chosen": -274.6481018066406,
      "logps/rejected": -233.6184539794922,
      "loss": 0.2323,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": 1.5015299320220947,
      "rewards/margins": 3.6158969402313232,
      "rewards/rejected": -2.1143670082092285,
      "step": 120
    },
    {
      "epoch": 1.1682242990654206,
      "grad_norm": 18.700138401214947,
      "learning_rate": 6.988983727898413e-07,
      "logits/chosen": -2.4572913646698,
      "logits/rejected": -2.4601497650146484,
      "logps/chosen": -234.6405792236328,
      "logps/rejected": -239.22579956054688,
      "loss": 0.2211,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 1.3679006099700928,
      "rewards/margins": 3.5180695056915283,
      "rewards/rejected": -2.1501688957214355,
      "step": 125
    },
    {
      "epoch": 1.2149532710280373,
      "grad_norm": 25.923791581257113,
      "learning_rate": 6.754847358270066e-07,
      "logits/chosen": -2.475020408630371,
      "logits/rejected": -2.4622273445129395,
      "logps/chosen": -227.0565185546875,
      "logps/rejected": -223.29769897460938,
      "loss": 0.2316,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.2679039239883423,
      "rewards/margins": 3.164490222930908,
      "rewards/rejected": -1.8965866565704346,
      "step": 130
    },
    {
      "epoch": 1.2616822429906542,
      "grad_norm": 21.409246715009832,
      "learning_rate": 6.516235236844661e-07,
      "logits/chosen": -2.448514223098755,
      "logits/rejected": -2.443563222885132,
      "logps/chosen": -238.69821166992188,
      "logps/rejected": -244.6441650390625,
      "loss": 0.2412,
      "rewards/accuracies": 0.9125000238418579,
      "rewards/chosen": 0.7657096982002258,
      "rewards/margins": 3.521303176879883,
      "rewards/rejected": -2.7555935382843018,
      "step": 135
    },
    {
      "epoch": 1.308411214953271,
      "grad_norm": 23.212967684893474,
      "learning_rate": 6.273755945688457e-07,
      "logits/chosen": -2.41972017288208,
      "logits/rejected": -2.388028383255005,
      "logps/chosen": -248.5680389404297,
      "logps/rejected": -235.9073944091797,
      "loss": 0.2808,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 0.47816404700279236,
      "rewards/margins": 3.6754353046417236,
      "rewards/rejected": -3.1972711086273193,
      "step": 140
    },
    {
      "epoch": 1.355140186915888,
      "grad_norm": 21.50115612986297,
      "learning_rate": 6.02802793011411e-07,
      "logits/chosen": -2.3905227184295654,
      "logits/rejected": -2.355194091796875,
      "logps/chosen": -257.6103820800781,
      "logps/rejected": -261.6281433105469,
      "loss": 0.2369,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 0.6035176515579224,
      "rewards/margins": 3.675292491912842,
      "rewards/rejected": -3.071774482727051,
      "step": 145
    },
    {
      "epoch": 1.4018691588785046,
      "grad_norm": 24.279342554217575,
      "learning_rate": 5.779677921331093e-07,
      "logits/chosen": -2.3340132236480713,
      "logits/rejected": -2.3176941871643066,
      "logps/chosen": -258.11871337890625,
      "logps/rejected": -229.5651092529297,
      "loss": 0.2512,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.0050841569900513,
      "rewards/margins": 3.4988720417022705,
      "rewards/rejected": -2.4937877655029297,
      "step": 150
    },
    {
      "epoch": 1.4018691588785046,
      "eval_logits/chosen": -2.3323609828948975,
      "eval_logits/rejected": -2.297656297683716,
      "eval_logps/chosen": -230.61724853515625,
      "eval_logps/rejected": -228.846923828125,
      "eval_loss": 0.5608044266700745,
      "eval_rewards/accuracies": 0.7864583134651184,
      "eval_rewards/chosen": 0.20199893414974213,
      "eval_rewards/margins": 2.0041935443878174,
      "eval_rewards/rejected": -1.802194595336914,
      "eval_runtime": 202.9747,
      "eval_samples_per_second": 14.977,
      "eval_steps_per_second": 0.236,
      "step": 150
    },
    {
      "epoch": 1.4485981308411215,
      "grad_norm": 20.205361154395078,
      "learning_rate": 5.529339337962897e-07,
      "logits/chosen": -2.3143677711486816,
      "logits/rejected": -2.2800261974334717,
      "logps/chosen": -242.8279571533203,
      "logps/rejected": -219.5634002685547,
      "loss": 0.263,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.5413401126861572,
      "rewards/margins": 3.561903715133667,
      "rewards/rejected": -2.0205633640289307,
      "step": 155
    },
    {
      "epoch": 1.4953271028037383,
      "grad_norm": 23.02116586142561,
      "learning_rate": 5.277650670507915e-07,
      "logits/chosen": -2.253575086593628,
      "logits/rejected": -2.2178051471710205,
      "logps/chosen": -246.2871856689453,
      "logps/rejected": -227.64773559570312,
      "loss": 0.2479,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": 1.6365845203399658,
      "rewards/margins": 3.364820957183838,
      "rewards/rejected": -1.728236198425293,
      "step": 160
    },
    {
      "epoch": 1.542056074766355,
      "grad_norm": 19.86487212004718,
      "learning_rate": 5.025253852864471e-07,
      "logits/chosen": -2.2566065788269043,
      "logits/rejected": -2.225008964538574,
      "logps/chosen": -228.00927734375,
      "logps/rejected": -226.37344360351562,
      "loss": 0.3032,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.2359155416488647,
      "rewards/margins": 3.1283955574035645,
      "rewards/rejected": -1.8924801349639893,
      "step": 165
    },
    {
      "epoch": 1.588785046728972,
      "grad_norm": 24.791429950163543,
      "learning_rate": 4.77279262507344e-07,
      "logits/chosen": -2.318096876144409,
      "logits/rejected": -2.2841153144836426,
      "logps/chosen": -242.6847686767578,
      "logps/rejected": -228.2392578125,
      "loss": 0.2779,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.4863147735595703,
      "rewards/margins": 3.9804482460021973,
      "rewards/rejected": -2.494133234024048,
      "step": 170
    },
    {
      "epoch": 1.6355140186915889,
      "grad_norm": 19.264180326117685,
      "learning_rate": 4.5209108914542714e-07,
      "logits/chosen": -2.3531315326690674,
      "logits/rejected": -2.325747489929199,
      "logps/chosen": -252.4257049560547,
      "logps/rejected": -252.0369110107422,
      "loss": 0.2673,
      "rewards/accuracies": 0.90625,
      "rewards/chosen": 1.5227165222167969,
      "rewards/margins": 4.164400577545166,
      "rewards/rejected": -2.64168381690979,
      "step": 175
    },
    {
      "epoch": 1.6822429906542056,
      "grad_norm": 27.25921599934839,
      "learning_rate": 4.2702510783220475e-07,
      "logits/chosen": -2.366103410720825,
      "logits/rejected": -2.3553428649902344,
      "logps/chosen": -234.7184600830078,
      "logps/rejected": -236.8829345703125,
      "loss": 0.2941,
      "rewards/accuracies": 0.893750011920929,
      "rewards/chosen": 0.7034923434257507,
      "rewards/margins": 3.489102840423584,
      "rewards/rejected": -2.7856106758117676,
      "step": 180
    },
    {
      "epoch": 1.7289719626168223,
      "grad_norm": 23.78347204770169,
      "learning_rate": 4.0214524954741586e-07,
      "logits/chosen": -2.379379987716675,
      "logits/rejected": -2.3642616271972656,
      "logps/chosen": -255.665771484375,
      "logps/rejected": -259.4281005859375,
      "loss": 0.2928,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.0745766162872314,
      "rewards/margins": 3.858994722366333,
      "rewards/rejected": -2.7844185829162598,
      "step": 185
    },
    {
      "epoch": 1.7757009345794392,
      "grad_norm": 17.510555360031177,
      "learning_rate": 3.7751497056257305e-07,
      "logits/chosen": -2.3731751441955566,
      "logits/rejected": -2.372335195541382,
      "logps/chosen": -230.4172821044922,
      "logps/rejected": -212.8553009033203,
      "loss": 0.2656,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.1900562047958374,
      "rewards/margins": 3.306123733520508,
      "rewards/rejected": -2.116067409515381,
      "step": 190
    },
    {
      "epoch": 1.8224299065420562,
      "grad_norm": 21.273227533684853,
      "learning_rate": 3.531970905952478e-07,
      "logits/chosen": -2.386868715286255,
      "logits/rejected": -2.349654197692871,
      "logps/chosen": -224.6810302734375,
      "logps/rejected": -244.11325073242188,
      "loss": 0.2689,
      "rewards/accuracies": 0.8999999761581421,
      "rewards/chosen": 1.3505589962005615,
      "rewards/margins": 3.271749496459961,
      "rewards/rejected": -1.921190619468689,
      "step": 195
    },
    {
      "epoch": 1.8691588785046729,
      "grad_norm": 20.785587523647333,
      "learning_rate": 3.2925363258689553e-07,
      "logits/chosen": -2.3584108352661133,
      "logits/rejected": -2.341261386871338,
      "logps/chosen": -255.19729614257812,
      "logps/rejected": -238.4913330078125,
      "loss": 0.3125,
      "rewards/accuracies": 0.875,
      "rewards/chosen": 1.245615839958191,
      "rewards/margins": 3.1446993350982666,
      "rewards/rejected": -1.8990834951400757,
      "step": 200
    },
    {
      "epoch": 1.8691588785046729,
      "eval_logits/chosen": -2.3304452896118164,
      "eval_logits/rejected": -2.299372911453247,
      "eval_logps/chosen": -227.91493225097656,
      "eval_logps/rejected": -226.53717041015625,
      "eval_loss": 0.5447002649307251,
      "eval_rewards/accuracies": 0.8098958134651184,
      "eval_rewards/chosen": 0.4722280502319336,
      "eval_rewards/margins": 2.043447732925415,
      "eval_rewards/rejected": -1.57122004032135,
      "eval_runtime": 202.8621,
      "eval_samples_per_second": 14.986,
      "eval_steps_per_second": 0.237,
      "step": 200
    },
    {
      "epoch": 1.9158878504672896,
      "grad_norm": 23.047331819250495,
      "learning_rate": 3.0574566451286086e-07,
      "logits/chosen": -2.28798246383667,
      "logits/rejected": -2.283902883529663,
      "logps/chosen": -223.94827270507812,
      "logps/rejected": -218.6190643310547,
      "loss": 0.2768,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": 0.8499220609664917,
      "rewards/margins": 2.6679863929748535,
      "rewards/rejected": -1.8180643320083618,
      "step": 205
    },
    {
      "epoch": 1.9626168224299065,
      "grad_norm": 39.667190244371184,
      "learning_rate": 2.8273314362803333e-07,
      "logits/chosen": -2.284705638885498,
      "logits/rejected": -2.2555558681488037,
      "logps/chosen": -238.67947387695312,
      "logps/rejected": -241.21127319335938,
      "loss": 0.2995,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.3379483222961426,
      "rewards/margins": 3.8164286613464355,
      "rewards/rejected": -2.478480339050293,
      "step": 210
    },
    {
      "epoch": 2.0093457943925235,
      "grad_norm": 13.876171599034908,
      "learning_rate": 2.602747635454047e-07,
      "logits/chosen": -2.263437271118164,
      "logits/rejected": -2.2282402515411377,
      "logps/chosen": -250.09912109375,
      "logps/rejected": -219.9522705078125,
      "loss": 0.2074,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.3080263137817383,
      "rewards/margins": 3.6176400184631348,
      "rewards/rejected": -2.3096134662628174,
      "step": 215
    },
    {
      "epoch": 2.05607476635514,
      "grad_norm": 15.83494140899864,
      "learning_rate": 2.384278045375523e-07,
      "logits/chosen": -2.2789604663848877,
      "logits/rejected": -2.261378049850464,
      "logps/chosen": -234.61288452148438,
      "logps/rejected": -262.85968017578125,
      "loss": 0.1558,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.1120151281356812,
      "rewards/margins": 3.8855414390563965,
      "rewards/rejected": -2.773526668548584,
      "step": 220
    },
    {
      "epoch": 2.102803738317757,
      "grad_norm": 13.361516693109529,
      "learning_rate": 2.1724798744286071e-07,
      "logits/chosen": -2.267428398132324,
      "logits/rejected": -2.2524585723876953,
      "logps/chosen": -227.0860137939453,
      "logps/rejected": -238.20361328125,
      "loss": 0.1566,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.1956349611282349,
      "rewards/margins": 3.7189323902130127,
      "rewards/rejected": -2.52329683303833,
      "step": 225
    },
    {
      "epoch": 2.149532710280374,
      "grad_norm": 15.593655288939111,
      "learning_rate": 1.9678933154909095e-07,
      "logits/chosen": -2.3036351203918457,
      "logits/rejected": -2.285254955291748,
      "logps/chosen": -249.8123321533203,
      "logps/rejected": -263.1349182128906,
      "loss": 0.1615,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.5796983242034912,
      "rewards/margins": 4.481447696685791,
      "rewards/rejected": -2.901749849319458,
      "step": 230
    },
    {
      "epoch": 2.196261682242991,
      "grad_norm": 15.512242762429548,
      "learning_rate": 1.77104016816768e-07,
      "logits/chosen": -2.3310775756835938,
      "logits/rejected": -2.2966668605804443,
      "logps/chosen": -244.6162109375,
      "logps/rejected": -225.72958374023438,
      "loss": 0.1597,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.3985507488250732,
      "rewards/margins": 3.707745313644409,
      "rewards/rejected": -2.3091940879821777,
      "step": 235
    },
    {
      "epoch": 2.2429906542056073,
      "grad_norm": 13.04743880016478,
      "learning_rate": 1.5824225079378684e-07,
      "logits/chosen": -2.338270664215088,
      "logits/rejected": -2.332692861557007,
      "logps/chosen": -220.993408203125,
      "logps/rejected": -227.9103240966797,
      "loss": 0.1459,
      "rewards/accuracies": 0.9437500238418579,
      "rewards/chosen": 1.1042306423187256,
      "rewards/margins": 3.9524731636047363,
      "rewards/rejected": -2.8482425212860107,
      "step": 240
    },
    {
      "epoch": 2.289719626168224,
      "grad_norm": 13.734303457357665,
      "learning_rate": 1.4025214056067237e-07,
      "logits/chosen": -2.3562066555023193,
      "logits/rejected": -2.3316104412078857,
      "logps/chosen": -241.31936645507812,
      "logps/rejected": -255.2411651611328,
      "loss": 0.1576,
      "rewards/accuracies": 0.9750000238418579,
      "rewards/chosen": 1.400942087173462,
      "rewards/margins": 4.312483787536621,
      "rewards/rejected": -2.91154146194458,
      "step": 245
    },
    {
      "epoch": 2.336448598130841,
      "grad_norm": 19.868641423806476,
      "learning_rate": 1.2317957003309725e-07,
      "logits/chosen": -2.37241792678833,
      "logits/rejected": -2.331575870513916,
      "logps/chosen": -239.10018920898438,
      "logps/rejected": -236.6571502685547,
      "loss": 0.1519,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.6855798959732056,
      "rewards/margins": 4.394530296325684,
      "rewards/rejected": -2.7089507579803467,
      "step": 250
    },
    {
      "epoch": 2.336448598130841,
      "eval_logits/chosen": -2.3581583499908447,
      "eval_logits/rejected": -2.3302361965179443,
      "eval_logps/chosen": -230.7427215576172,
      "eval_logps/rejected": -231.17662048339844,
      "eval_loss": 0.5571395754814148,
      "eval_rewards/accuracies": 0.8046875,
      "eval_rewards/chosen": 0.1894473433494568,
      "eval_rewards/margins": 2.224613904953003,
      "eval_rewards/rejected": -2.0351665019989014,
      "eval_runtime": 202.849,
      "eval_samples_per_second": 14.987,
      "eval_steps_per_second": 0.237,
      "step": 250
    },
    {
      "epoch": 2.383177570093458,
      "grad_norm": 14.584541263712909,
      "learning_rate": 1.0706808293459873e-07,
      "logits/chosen": -2.3551063537597656,
      "logits/rejected": -2.3340346813201904,
      "logps/chosen": -230.0168914794922,
      "logps/rejected": -251.65060424804688,
      "loss": 0.151,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.2446879148483276,
      "rewards/margins": 4.14044713973999,
      "rewards/rejected": -2.895758867263794,
      "step": 255
    },
    {
      "epoch": 2.4299065420560746,
      "grad_norm": 15.86451522408118,
      "learning_rate": 9.195877173797534e-08,
      "logits/chosen": -2.3495898246765137,
      "logits/rejected": -2.33632230758667,
      "logps/chosen": -237.96395874023438,
      "logps/rejected": -236.162353515625,
      "loss": 0.1551,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.2900810241699219,
      "rewards/margins": 4.1981096267700195,
      "rewards/rejected": -2.9080288410186768,
      "step": 260
    },
    {
      "epoch": 2.4766355140186915,
      "grad_norm": 22.97374071251409,
      "learning_rate": 7.789017285861438e-08,
      "logits/chosen": -2.348433494567871,
      "logits/rejected": -2.3354239463806152,
      "logps/chosen": -234.0780792236328,
      "logps/rejected": -226.50698852539062,
      "loss": 0.1665,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 0.9046391248703003,
      "rewards/margins": 4.137303352355957,
      "rewards/rejected": -3.232664108276367,
      "step": 265
    },
    {
      "epoch": 2.5233644859813085,
      "grad_norm": 15.800484094844276,
      "learning_rate": 6.489816836706785e-08,
      "logits/chosen": -2.3266546726226807,
      "logits/rejected": -2.3040125370025635,
      "logps/chosen": -229.1071319580078,
      "logps/rejected": -239.10232543945312,
      "loss": 0.1421,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.0205641984939575,
      "rewards/margins": 4.1610260009765625,
      "rewards/rejected": -3.1404614448547363,
      "step": 270
    },
    {
      "epoch": 2.5700934579439254,
      "grad_norm": 16.2247229156408,
      "learning_rate": 5.3015894471550914e-08,
      "logits/chosen": -2.3282101154327393,
      "logits/rejected": -2.2819929122924805,
      "logps/chosen": -231.26358032226562,
      "logps/rejected": -228.27059936523438,
      "loss": 0.149,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": 1.1585379838943481,
      "rewards/margins": 4.134129047393799,
      "rewards/rejected": -2.9755911827087402,
      "step": 275
    },
    {
      "epoch": 2.616822429906542,
      "grad_norm": 14.76920977719455,
      "learning_rate": 4.227365700378799e-08,
      "logits/chosen": -2.3212122917175293,
      "logits/rejected": -2.287430763244629,
      "logps/chosen": -255.4926300048828,
      "logps/rejected": -250.0602569580078,
      "loss": 0.1517,
      "rewards/accuracies": 0.9624999761581421,
      "rewards/chosen": 1.5625426769256592,
      "rewards/margins": 4.667173385620117,
      "rewards/rejected": -3.104630947113037,
      "step": 280
    },
    {
      "epoch": 2.663551401869159,
      "grad_norm": 17.3445381269751,
      "learning_rate": 3.269885412375223e-08,
      "logits/chosen": -2.297814130783081,
      "logits/rejected": -2.291019916534424,
      "logps/chosen": -246.7578887939453,
      "logps/rejected": -254.56893920898438,
      "loss": 0.152,
      "rewards/accuracies": 0.949999988079071,
      "rewards/chosen": 1.0841681957244873,
      "rewards/margins": 4.465099334716797,
      "rewards/rejected": -3.380931854248047,
      "step": 285
    },
    {
      "epoch": 2.710280373831776,
      "grad_norm": 18.922584920713373,
      "learning_rate": 2.4315906440446952e-08,
      "logits/chosen": -2.3056201934814453,
      "logits/rejected": -2.2901225090026855,
      "logps/chosen": -267.8759765625,
      "logps/rejected": -242.197509765625,
      "loss": 0.1637,
      "rewards/accuracies": 0.925000011920929,
      "rewards/chosen": 1.5246771574020386,
      "rewards/margins": 4.00520658493042,
      "rewards/rejected": -2.480529308319092,
      "step": 290
    },
    {
      "epoch": 2.7570093457943923,
      "grad_norm": 16.82420233496269,
      "learning_rate": 1.7146194726952778e-08,
      "logits/chosen": -2.3040153980255127,
      "logits/rejected": -2.290541172027588,
      "logps/chosen": -245.0521697998047,
      "logps/rejected": -242.7193603515625,
      "loss": 0.156,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": 1.381598711013794,
      "rewards/margins": 4.1164751052856445,
      "rewards/rejected": -2.7348761558532715,
      "step": 295
    },
    {
      "epoch": 2.803738317757009,
      "grad_norm": 12.488604079614023,
      "learning_rate": 1.1208005388599951e-08,
      "logits/chosen": -2.305384397506714,
      "logits/rejected": -2.2505059242248535,
      "logps/chosen": -230.2703399658203,
      "logps/rejected": -246.6582794189453,
      "loss": 0.1708,
      "rewards/accuracies": 0.956250011920929,
      "rewards/chosen": 1.1310278177261353,
      "rewards/margins": 4.3728203773498535,
      "rewards/rejected": -3.2417922019958496,
      "step": 300
    },
    {
      "epoch": 2.803738317757009,
      "eval_logits/chosen": -2.3019378185272217,
      "eval_logits/rejected": -2.2672407627105713,
      "eval_logps/chosen": -232.6370849609375,
      "eval_logps/rejected": -234.4372100830078,
      "eval_loss": 0.5570527911186218,
      "eval_rewards/accuracies": 0.8072916865348816,
      "eval_rewards/chosen": 1.33514404296875e-05,
      "eval_rewards/margins": 2.3612377643585205,
      "eval_rewards/rejected": -2.36122465133667,
      "eval_runtime": 202.3124,
      "eval_samples_per_second": 15.026,
      "eval_steps_per_second": 0.237,
      "step": 300
    }
  ],
  "logging_steps": 5,
  "max_steps": 321,
  "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": false
      },
      "attributes": {}
    }
  },
  "total_flos": 3537404052897792.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}