{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 385,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": -1.7278180122375488,
      "logits/rejected": -1.7377450466156006,
      "logps/chosen": -29.553977966308594,
      "logps/rejected": -42.813133239746094,
      "loss": 0.5,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": -1.8667322397232056,
      "logits/rejected": -1.8710299730300903,
      "logps/chosen": -36.98433303833008,
      "logps/rejected": -33.65640640258789,
      "loss": 0.4948,
      "rewards/accuracies": 0.5277777910232544,
      "rewards/chosen": 0.015585740096867085,
      "rewards/margins": 0.034019216895103455,
      "rewards/rejected": -0.018433474004268646,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": -1.9972655773162842,
      "logits/rejected": -1.9999065399169922,
      "logps/chosen": -29.6416072845459,
      "logps/rejected": -29.046981811523438,
      "loss": 0.5013,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.00041136163054034114,
      "rewards/margins": -0.011113673448562622,
      "rewards/rejected": 0.011525033973157406,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": -1.9205191135406494,
      "logits/rejected": -1.9178133010864258,
      "logps/chosen": -31.385412216186523,
      "logps/rejected": -33.23365020751953,
      "loss": 0.4942,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.02147640287876129,
      "rewards/margins": 0.03211076930165291,
      "rewards/rejected": -0.010634368285536766,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": -2.017784833908081,
      "logits/rejected": -2.009059429168701,
      "logps/chosen": -32.547088623046875,
      "logps/rejected": -32.473243713378906,
      "loss": 0.5024,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.020800117403268814,
      "rewards/margins": -0.006797379348427057,
      "rewards/rejected": 0.027597496286034584,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": -1.863996148109436,
      "logits/rejected": -1.8532311916351318,
      "logps/chosen": -33.490657806396484,
      "logps/rejected": -35.415367126464844,
      "loss": 0.4957,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.046776313334703445,
      "rewards/margins": 0.02089320495724678,
      "rewards/rejected": 0.025883108377456665,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": -1.9439821243286133,
      "logits/rejected": -1.9459091424942017,
      "logps/chosen": -32.50756072998047,
      "logps/rejected": -33.15720748901367,
      "loss": 0.489,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.06483479589223862,
      "rewards/margins": 0.05491489917039871,
      "rewards/rejected": 0.009919902309775352,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": -2.075462579727173,
      "logits/rejected": -2.0804481506347656,
      "logps/chosen": -33.86588668823242,
      "logps/rejected": -36.501365661621094,
      "loss": 0.491,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.08581922948360443,
      "rewards/margins": 0.038563162088394165,
      "rewards/rejected": 0.04725607484579086,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": -1.9374949932098389,
      "logits/rejected": -1.9405946731567383,
      "logps/chosen": -34.200557708740234,
      "logps/rejected": -34.543209075927734,
      "loss": 0.4719,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.15907494723796844,
      "rewards/margins": 0.1246359720826149,
      "rewards/rejected": 0.03443896025419235,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": -1.9453551769256592,
      "logits/rejected": -1.9498746395111084,
      "logps/chosen": -32.2825927734375,
      "logps/rejected": -32.25676727294922,
      "loss": 0.4858,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.12937995791435242,
      "rewards/margins": 0.05788033455610275,
      "rewards/rejected": 0.07149962335824966,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": -2.04329776763916,
      "logits/rejected": -2.0412919521331787,
      "logps/chosen": -32.015525817871094,
      "logps/rejected": -31.169103622436523,
      "loss": 0.4799,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.15771625936031342,
      "rewards/margins": 0.08974156528711319,
      "rewards/rejected": 0.06797470152378082,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.238145351409912,
      "eval_logits/rejected": -2.2333030700683594,
      "eval_logps/chosen": -33.86069107055664,
      "eval_logps/rejected": -37.356990814208984,
      "eval_loss": 0.49784979224205017,
      "eval_rewards/accuracies": 0.5340532064437866,
      "eval_rewards/chosen": 0.12170296907424927,
      "eval_rewards/margins": 0.009964452125132084,
      "eval_rewards/rejected": 0.11173851788043976,
      "eval_runtime": 145.9039,
      "eval_samples_per_second": 2.351,
      "eval_steps_per_second": 0.295,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": -1.9996016025543213,
      "logits/rejected": -1.9972642660140991,
      "logps/chosen": -32.95330810546875,
      "logps/rejected": -33.88172912597656,
      "loss": 0.4742,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.20335713028907776,
      "rewards/margins": 0.10000836849212646,
      "rewards/rejected": 0.10334876924753189,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": -2.0096044540405273,
      "logits/rejected": -2.0012660026550293,
      "logps/chosen": -32.153419494628906,
      "logps/rejected": -31.961782455444336,
      "loss": 0.4848,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.20403961837291718,
      "rewards/margins": 0.06423385441303253,
      "rewards/rejected": 0.13980577886104584,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": -2.038412570953369,
      "logits/rejected": -2.0304617881774902,
      "logps/chosen": -30.165945053100586,
      "logps/rejected": -31.880346298217773,
      "loss": 0.4787,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.21683910489082336,
      "rewards/margins": 0.0959777683019638,
      "rewards/rejected": 0.12086131423711777,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": -1.9683482646942139,
      "logits/rejected": -1.9784959554672241,
      "logps/chosen": -31.041671752929688,
      "logps/rejected": -32.388893127441406,
      "loss": 0.4621,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.264592707157135,
      "rewards/margins": 0.16240297257900238,
      "rewards/rejected": 0.10218973457813263,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": -1.8799880743026733,
      "logits/rejected": -1.8811382055282593,
      "logps/chosen": -33.686256408691406,
      "logps/rejected": -34.564491271972656,
      "loss": 0.4503,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.3545047640800476,
      "rewards/margins": 0.21957096457481384,
      "rewards/rejected": 0.134933739900589,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": -1.931627869606018,
      "logits/rejected": -1.9282808303833008,
      "logps/chosen": -35.77287292480469,
      "logps/rejected": -32.49742889404297,
      "loss": 0.4707,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.27916350960731506,
      "rewards/margins": 0.12143567949533463,
      "rewards/rejected": 0.15772780776023865,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": -2.0332443714141846,
      "logits/rejected": -2.0259475708007812,
      "logps/chosen": -33.251670837402344,
      "logps/rejected": -31.17826271057129,
      "loss": 0.439,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.382335364818573,
      "rewards/margins": 0.2608257234096527,
      "rewards/rejected": 0.1215096265077591,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": -2.040196657180786,
      "logits/rejected": -2.045389413833618,
      "logps/chosen": -31.95781898498535,
      "logps/rejected": -32.18687057495117,
      "loss": 0.4516,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.4108788073062897,
      "rewards/margins": 0.20208871364593506,
      "rewards/rejected": 0.20879006385803223,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": -2.040775775909424,
      "logits/rejected": -2.038043737411499,
      "logps/chosen": -31.056228637695312,
      "logps/rejected": -31.10147476196289,
      "loss": 0.4647,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.30290743708610535,
      "rewards/margins": 0.1511872559785843,
      "rewards/rejected": 0.15172019600868225,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": -1.9112863540649414,
      "logits/rejected": -1.9159523248672485,
      "logps/chosen": -31.08163070678711,
      "logps/rejected": -32.598487854003906,
      "loss": 0.4453,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.3672560155391693,
      "rewards/margins": 0.2325470894575119,
      "rewards/rejected": 0.1347089260816574,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.234708309173584,
      "eval_logits/rejected": -2.229902505874634,
      "eval_logps/chosen": -33.76850891113281,
      "eval_logps/rejected": -37.29365921020508,
      "eval_loss": 0.49275800585746765,
      "eval_rewards/accuracies": 0.5369601845741272,
      "eval_rewards/chosen": 0.18622951209545135,
      "eval_rewards/margins": 0.030155813321471214,
      "eval_rewards/rejected": 0.1560736745595932,
      "eval_runtime": 145.5287,
      "eval_samples_per_second": 2.357,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": -2.022939443588257,
      "logits/rejected": -2.033546209335327,
      "logps/chosen": -31.5457706451416,
      "logps/rejected": -33.67595672607422,
      "loss": 0.4628,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.2978750169277191,
      "rewards/margins": 0.16715146601200104,
      "rewards/rejected": 0.13072356581687927,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": -1.9158331155776978,
      "logits/rejected": -1.9305419921875,
      "logps/chosen": -29.57406997680664,
      "logps/rejected": -31.42013168334961,
      "loss": 0.4363,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.37622833251953125,
      "rewards/margins": 0.2795966863632202,
      "rewards/rejected": 0.09663165360689163,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": -1.9733517169952393,
      "logits/rejected": -1.9773433208465576,
      "logps/chosen": -32.83649444580078,
      "logps/rejected": -31.40326499938965,
      "loss": 0.4311,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.40668267011642456,
      "rewards/margins": 0.304456502199173,
      "rewards/rejected": 0.10222617536783218,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": -1.9713729619979858,
      "logits/rejected": -1.9496545791625977,
      "logps/chosen": -33.5602912902832,
      "logps/rejected": -34.85026550292969,
      "loss": 0.4227,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.41541916131973267,
      "rewards/margins": 0.3450462222099304,
      "rewards/rejected": 0.07037289440631866,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": -2.012058734893799,
      "logits/rejected": -2.0087602138519287,
      "logps/chosen": -32.46921157836914,
      "logps/rejected": -36.012245178222656,
      "loss": 0.4613,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.31910061836242676,
      "rewards/margins": 0.17022350430488586,
      "rewards/rejected": 0.1488770991563797,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": -1.878584861755371,
      "logits/rejected": -1.876153588294983,
      "logps/chosen": -33.71690368652344,
      "logps/rejected": -35.26210403442383,
      "loss": 0.4631,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.33004888892173767,
      "rewards/margins": 0.1602855622768402,
      "rewards/rejected": 0.16976332664489746,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": -1.863705039024353,
      "logits/rejected": -1.8611875772476196,
      "logps/chosen": -33.94118881225586,
      "logps/rejected": -31.579448699951172,
      "loss": 0.4582,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.3194878101348877,
      "rewards/margins": 0.18517015874385834,
      "rewards/rejected": 0.13431766629219055,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": -1.96773362159729,
      "logits/rejected": -1.957313895225525,
      "logps/chosen": -34.72475814819336,
      "logps/rejected": -31.634963989257812,
      "loss": 0.4297,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.4423200190067291,
      "rewards/margins": 0.298100084066391,
      "rewards/rejected": 0.14421990513801575,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": -2.0632526874542236,
      "logits/rejected": -2.0484511852264404,
      "logps/chosen": -30.412399291992188,
      "logps/rejected": -32.339237213134766,
      "loss": 0.4682,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.3582096993923187,
      "rewards/margins": 0.13939058780670166,
      "rewards/rejected": 0.21881911158561707,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": -1.9345197677612305,
      "logits/rejected": -1.9320701360702515,
      "logps/chosen": -32.11776351928711,
      "logps/rejected": -30.692516326904297,
      "loss": 0.3947,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.5665432810783386,
      "rewards/margins": 0.4795381426811218,
      "rewards/rejected": 0.08700509369373322,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.2321932315826416,
      "eval_logits/rejected": -2.227381944656372,
      "eval_logps/chosen": -33.75507736206055,
      "eval_logps/rejected": -37.28935623168945,
      "eval_loss": 0.49099090695381165,
      "eval_rewards/accuracies": 0.5564784407615662,
      "eval_rewards/chosen": 0.1956319361925125,
      "eval_rewards/margins": 0.03654861077666283,
      "eval_rewards/rejected": 0.15908333659172058,
      "eval_runtime": 145.678,
      "eval_samples_per_second": 2.355,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": -1.918630838394165,
      "logits/rejected": -1.9154895544052124,
      "logps/chosen": -31.033864974975586,
      "logps/rejected": -33.57470703125,
      "loss": 0.4357,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.3986373543739319,
      "rewards/margins": 0.28142648935317993,
      "rewards/rejected": 0.11721093952655792,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": -1.9683278799057007,
      "logits/rejected": -1.9561859369277954,
      "logps/chosen": -34.02568817138672,
      "logps/rejected": -33.45148849487305,
      "loss": 0.4237,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.39926695823669434,
      "rewards/margins": 0.3381231129169464,
      "rewards/rejected": 0.06114383786916733,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": -2.0041966438293457,
      "logits/rejected": -2.0028293132781982,
      "logps/chosen": -32.882896423339844,
      "logps/rejected": -32.24889373779297,
      "loss": 0.4411,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.4220260679721832,
      "rewards/margins": 0.26054397225379944,
      "rewards/rejected": 0.16148212552070618,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": -2.0908923149108887,
      "logits/rejected": -2.0752511024475098,
      "logps/chosen": -33.49097442626953,
      "logps/rejected": -32.81463623046875,
      "loss": 0.4381,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.4818909764289856,
      "rewards/margins": 0.26951199769973755,
      "rewards/rejected": 0.21237893402576447,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": -1.9633783102035522,
      "logits/rejected": -1.9625694751739502,
      "logps/chosen": -32.55985641479492,
      "logps/rejected": -32.24340057373047,
      "loss": 0.4211,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": 0.5055657029151917,
      "rewards/margins": 0.3503406047821045,
      "rewards/rejected": 0.15522508323192596,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": -1.9187549352645874,
      "logits/rejected": -1.9290361404418945,
      "logps/chosen": -31.61210060119629,
      "logps/rejected": -35.0199089050293,
      "loss": 0.4489,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.41481319069862366,
      "rewards/margins": 0.22661848366260529,
      "rewards/rejected": 0.18819470703601837,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": -2.0583438873291016,
      "logits/rejected": -2.051846742630005,
      "logps/chosen": -33.05150604248047,
      "logps/rejected": -28.9710750579834,
      "loss": 0.4429,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.42032259702682495,
      "rewards/margins": 0.2415778934955597,
      "rewards/rejected": 0.17874471843242645,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": -1.9180545806884766,
      "logits/rejected": -1.92022705078125,
      "logps/chosen": -33.62676239013672,
      "logps/rejected": -30.719894409179688,
      "loss": 0.4281,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.4286006987094879,
      "rewards/margins": 0.3216874301433563,
      "rewards/rejected": 0.10691330581903458,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.45851393303313814,
      "train_runtime": 3252.0771,
      "train_samples_per_second": 0.947,
      "train_steps_per_second": 0.118
    }
  ],
  "logging_steps": 10,
  "max_steps": 385,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}