{
  "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-08,
      "logits/chosen": -1.7278180122375488,
      "logits/rejected": -1.7377450466156006,
      "logps/chosen": -29.553977966308594,
      "logps/rejected": -42.813133239746094,
      "loss": 0.3906,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.2820512820512818e-07,
      "logits/chosen": -1.8662992715835571,
      "logits/rejected": -1.8706117868423462,
      "logps/chosen": -36.97681427001953,
      "logps/rejected": -33.66523361206055,
      "loss": 0.3447,
      "rewards/accuracies": 0.5416666865348816,
      "rewards/chosen": 0.023829741403460503,
      "rewards/margins": 0.05195777863264084,
      "rewards/rejected": -0.028128040954470634,
      "step": 10
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.5641025641025636e-07,
      "logits/chosen": -1.997193694114685,
      "logits/rejected": -1.9998573064804077,
      "logps/chosen": -29.65359878540039,
      "logps/rejected": -29.054311752319336,
      "loss": 0.4534,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.009124360978603363,
      "rewards/margins": -0.016431041061878204,
      "rewards/rejected": 0.007306680083274841,
      "step": 20
    },
    {
      "epoch": 0.08,
      "learning_rate": 3.8461538461538463e-07,
      "logits/chosen": -1.9197280406951904,
      "logits/rejected": -1.9170429706573486,
      "logps/chosen": -31.414775848388672,
      "logps/rejected": -33.24064254760742,
      "loss": 0.4025,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.0010542668169364333,
      "rewards/margins": 0.018799806013703346,
      "rewards/rejected": -0.01774553768336773,
      "step": 30
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.999896948438433e-07,
      "logits/chosen": -2.0171010494232178,
      "logits/rejected": -2.008350372314453,
      "logps/chosen": -32.59648513793945,
      "logps/rejected": -32.50862121582031,
      "loss": 0.4532,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": -0.015740757808089256,
      "rewards/margins": -0.018979396671056747,
      "rewards/rejected": 0.0032386414241045713,
      "step": 40
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.987541037542186e-07,
      "logits/chosen": -1.8645904064178467,
      "logits/rejected": -1.8538070917129517,
      "logps/chosen": -33.542823791503906,
      "logps/rejected": -35.43744659423828,
      "loss": 0.4304,
      "rewards/accuracies": 0.5,
      "rewards/chosen": 0.011724205687642097,
      "rewards/margins": -0.00019102543592453003,
      "rewards/rejected": 0.011915231123566628,
      "step": 50
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.954691471941118e-07,
      "logits/chosen": -1.9461469650268555,
      "logits/rejected": -1.948094129562378,
      "logps/chosen": -32.587215423583984,
      "logps/rejected": -33.21445846557617,
      "loss": 0.3539,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.010374858975410461,
      "rewards/margins": 0.04483799636363983,
      "rewards/rejected": -0.03446313738822937,
      "step": 60
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.901618883413548e-07,
      "logits/chosen": -2.079591989517212,
      "logits/rejected": -2.0845742225646973,
      "logps/chosen": -34.00908279418945,
      "logps/rejected": -36.58150863647461,
      "loss": 0.4506,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.01648041605949402,
      "rewards/margins": -0.006370754446834326,
      "rewards/rejected": -0.010109663009643555,
      "step": 70
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.828760511501322e-07,
      "logits/chosen": -1.9423307180404663,
      "logits/rejected": -1.9454963207244873,
      "logps/chosen": -34.39698028564453,
      "logps/rejected": -34.59761428833008,
      "loss": 0.3889,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.024663064628839493,
      "rewards/margins": 0.028829574584960938,
      "rewards/rejected": -0.00416650902479887,
      "step": 80
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.736716601303429e-07,
      "logits/chosen": -1.9508873224258423,
      "logits/rejected": -1.9553953409194946,
      "logps/chosen": -32.48583221435547,
      "logps/rejected": -32.35867691040039,
      "loss": 0.4445,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": -0.014726865105330944,
      "rewards/margins": -0.014909917488694191,
      "rewards/rejected": 0.00018304325931239873,
      "step": 90
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.62624545834521e-07,
      "logits/chosen": -2.0491719245910645,
      "logits/rejected": -2.047179937362671,
      "logps/chosen": -32.22635269165039,
      "logps/rejected": -31.287487030029297,
      "loss": 0.3805,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.011584864929318428,
      "rewards/margins": 0.02860759198665619,
      "rewards/rejected": -0.01702272891998291,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -2.244148015975952,
      "eval_logits/rejected": -2.2392663955688477,
      "eval_logps/chosen": -34.01713943481445,
      "eval_logps/rejected": -37.514495849609375,
      "eval_loss": 0.4061162769794464,
      "eval_rewards/accuracies": 0.5394518375396729,
      "eval_rewards/chosen": 0.013928660191595554,
      "eval_rewards/margins": 0.012230273336172104,
      "eval_rewards/rejected": 0.0016983875539153814,
      "eval_runtime": 146.1091,
      "eval_samples_per_second": 2.348,
      "eval_steps_per_second": 0.294,
      "step": 100
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.4982572012636904e-07,
      "logits/chosen": -2.005580186843872,
      "logits/rejected": -2.0031564235687256,
      "logps/chosen": -33.24415588378906,
      "logps/rejected": -33.99993133544922,
      "loss": 0.4792,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.00026987865567207336,
      "rewards/margins": -0.02382112666964531,
      "rewards/rejected": 0.023551244288682938,
      "step": 110
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.353806263777677e-07,
      "logits/chosen": -2.01680326461792,
      "logits/rejected": -2.0084121227264404,
      "logps/chosen": -32.45465850830078,
      "logps/rejected": -32.1729850769043,
      "loss": 0.4286,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": -0.007803081069141626,
      "rewards/margins": 0.0013800703454762697,
      "rewards/rejected": -0.009183152578771114,
      "step": 120
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.194082707715275e-07,
      "logits/chosen": -2.0466856956481934,
      "logits/rejected": -2.0386409759521484,
      "logps/chosen": -30.494335174560547,
      "logps/rejected": -32.051963806152344,
      "loss": 0.4497,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.014895597472786903,
      "rewards/margins": -0.015729816630482674,
      "rewards/rejected": 0.0008342192741110921,
      "step": 130
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.020402418666621e-07,
      "logits/chosen": -1.977065086364746,
      "logits/rejected": -1.987343430519104,
      "logps/chosen": -31.37868309020996,
      "logps/rejected": -32.54730224609375,
      "loss": 0.3589,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.03278205543756485,
      "rewards/margins": 0.04271895810961723,
      "rewards/rejected": -0.009936909191310406,
      "step": 140
    },
    {
      "epoch": 0.39,
      "learning_rate": 3.8341962650351185e-07,
      "logits/chosen": -1.8916466236114502,
      "logits/rejected": -1.8927490711212158,
      "logps/chosen": -34.209651947021484,
      "logps/rejected": -34.766143798828125,
      "loss": 0.4325,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.013569498434662819,
      "rewards/margins": -0.00645996630191803,
      "rewards/rejected": -0.007109532598406076,
      "step": 150
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.636998309800572e-07,
      "logits/chosen": -1.9427303075790405,
      "logits/rejected": -1.939252495765686,
      "logps/chosen": -36.14452362060547,
      "logps/rejected": -32.73284149169922,
      "loss": 0.3796,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.021723434329032898,
      "rewards/margins": 0.02979486621916294,
      "rewards/rejected": -0.008071433752775192,
      "step": 160
    },
    {
      "epoch": 0.44,
      "learning_rate": 3.430433172111807e-07,
      "logits/chosen": -2.0424513816833496,
      "logits/rejected": -2.035060405731201,
      "logps/chosen": -33.786170959472656,
      "logps/rejected": -31.34820556640625,
      "loss": 0.4172,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": 0.009350698441267014,
      "rewards/margins": 0.006434415467083454,
      "rewards/rejected": 0.0029162843711674213,
      "step": 170
    },
    {
      "epoch": 0.47,
      "learning_rate": 3.216202642830543e-07,
      "logits/chosen": -2.0475661754608154,
      "logits/rejected": -2.052834987640381,
      "logps/chosen": -32.528114318847656,
      "logps/rejected": -32.50902557373047,
      "loss": 0.3766,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": 0.013338183984160423,
      "rewards/margins": 0.03244578838348389,
      "rewards/rejected": -0.019107606261968613,
      "step": 180
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.9960716642946403e-07,
      "logits/chosen": -2.048661231994629,
      "logits/rejected": -2.045872926712036,
      "logps/chosen": -31.496755599975586,
      "logps/rejected": -31.314464569091797,
      "loss": 0.4475,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.00624021515250206,
      "rewards/margins": -0.009242123924195766,
      "rewards/rejected": 0.003001909703016281,
      "step": 190
    },
    {
      "epoch": 0.52,
      "learning_rate": 2.771853789806683e-07,
      "logits/chosen": -1.9193611145019531,
      "logits/rejected": -1.9240529537200928,
      "logps/chosen": -31.579212188720703,
      "logps/rejected": -32.80603790283203,
      "loss": 0.371,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.02165621519088745,
      "rewards/margins": 0.033739686012268066,
      "rewards/rejected": -0.012083468958735466,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": -2.243856191635132,
      "eval_logits/rejected": -2.2389791011810303,
      "eval_logps/chosen": -34.023136138916016,
      "eval_logps/rejected": -37.501853942871094,
      "eval_loss": 0.4322855770587921,
      "eval_rewards/accuracies": 0.5253322720527649,
      "eval_rewards/chosen": 0.009133166633546352,
      "eval_rewards/margins": -0.002679171971976757,
      "eval_rewards/rejected": 0.01181233860552311,
      "eval_runtime": 145.9618,
      "eval_samples_per_second": 2.35,
      "eval_steps_per_second": 0.295,
      "step": 200
    },
    {
      "epoch": 0.55,
      "learning_rate": 2.5453962426402e-07,
      "logits/chosen": -2.0320448875427246,
      "logits/rejected": -2.0427393913269043,
      "logps/chosen": -31.947372436523438,
      "logps/rejected": -33.899864196777344,
      "loss": 0.3622,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.019143681973218918,
      "rewards/margins": 0.04887578636407852,
      "rewards/rejected": -0.029732098802924156,
      "step": 210
    },
    {
      "epoch": 0.57,
      "learning_rate": 2.318564697655179e-07,
      "logits/chosen": -1.9257709980010986,
      "logits/rejected": -1.940639853477478,
      "logps/chosen": -30.079341888427734,
      "logps/rejected": -31.5616397857666,
      "loss": 0.3884,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.025758206844329834,
      "rewards/margins": 0.028527002781629562,
      "rewards/rejected": -0.0027687971014529467,
      "step": 220
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.093227910899832e-07,
      "logits/chosen": -1.9829959869384766,
      "logits/rejected": -1.9869670867919922,
      "logps/chosen": -33.4053955078125,
      "logps/rejected": -31.562353134155273,
      "loss": 0.3918,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.009657363407313824,
      "rewards/margins": 0.02010050043463707,
      "rewards/rejected": -0.010443134233355522,
      "step": 230
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8712423238279356e-07,
      "logits/chosen": -1.9836658239364624,
      "logits/rejected": -1.9616800546646118,
      "logps/chosen": -34.175201416015625,
      "logps/rejected": -34.966102600097656,
      "loss": 0.4346,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": -0.017159918323159218,
      "rewards/margins": -0.004913450218737125,
      "rewards/rejected": -0.012246469967067242,
      "step": 240
    },
    {
      "epoch": 0.65,
      "learning_rate": 1.654436768970182e-07,
      "logits/chosen": -2.0244596004486084,
      "logits/rejected": -2.021144390106201,
      "logps/chosen": -32.932373046875,
      "logps/rejected": -36.2264289855957,
      "loss": 0.4397,
      "rewards/accuracies": 0.512499988079071,
      "rewards/chosen": -0.005845203995704651,
      "rewards/margins": -0.004640025552362204,
      "rewards/rejected": -0.001205177279189229,
      "step": 250
    },
    {
      "epoch": 0.68,
      "learning_rate": 1.444597403062196e-07,
      "logits/chosen": -1.891283392906189,
      "logits/rejected": -1.8888483047485352,
      "logps/chosen": -34.20048522949219,
      "logps/rejected": -35.507022857666016,
      "loss": 0.4441,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.009669994935393333,
      "rewards/margins": -0.0077504729852080345,
      "rewards/rejected": -0.0019195213681086898,
      "step": 260
    },
    {
      "epoch": 0.7,
      "learning_rate": 1.2434529917578887e-07,
      "logits/chosen": -1.8760111331939697,
      "logits/rejected": -1.8734772205352783,
      "logps/chosen": -34.39020919799805,
      "logps/rejected": -31.75579261779785,
      "loss": 0.4295,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": 0.0059113698080182076,
      "rewards/margins": -0.006519269198179245,
      "rewards/rejected": 0.012430639937520027,
      "step": 270
    },
    {
      "epoch": 0.73,
      "learning_rate": 1.0526606671603521e-07,
      "logits/chosen": -1.9797089099884033,
      "logits/rejected": -1.9690707921981812,
      "logps/chosen": -35.31258773803711,
      "logps/rejected": -31.837697982788086,
      "loss": 0.3868,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.03524213656783104,
      "rewards/margins": 0.0326063297688961,
      "rewards/rejected": 0.0026358035393059254,
      "step": 280
    },
    {
      "epoch": 0.75,
      "learning_rate": 8.737922755071453e-08,
      "logits/chosen": -2.0755209922790527,
      "logits/rejected": -2.060497760772705,
      "logps/chosen": -30.90865135192871,
      "logps/rejected": -32.64521026611328,
      "loss": 0.41,
      "rewards/accuracies": 0.44999998807907104,
      "rewards/chosen": 0.012379061430692673,
      "rewards/margins": 0.007080032490193844,
      "rewards/rejected": 0.005299028940498829,
      "step": 290
    },
    {
      "epoch": 0.78,
      "learning_rate": 7.08321427484816e-08,
      "logits/chosen": -1.9463651180267334,
      "logits/rejected": -1.9438308477401733,
      "logps/chosen": -32.89823532104492,
      "logps/rejected": -30.81850814819336,
      "loss": 0.3839,
      "rewards/accuracies": 0.5874999761581421,
      "rewards/chosen": 0.02309919334948063,
      "rewards/margins": 0.024458223953843117,
      "rewards/rejected": -0.0013590289745479822,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": -2.2435548305511475,
      "eval_logits/rejected": -2.2386720180511475,
      "eval_logps/chosen": -34.023826599121094,
      "eval_logps/rejected": -37.50039291381836,
      "eval_loss": 0.4344586431980133,
      "eval_rewards/accuracies": 0.4808970093727112,
      "eval_rewards/chosen": 0.00858243927359581,
      "eval_rewards/margins": -0.004399063065648079,
      "eval_rewards/rejected": 0.012981505133211613,
      "eval_runtime": 145.9473,
      "eval_samples_per_second": 2.35,
      "eval_steps_per_second": 0.295,
      "step": 300
    },
    {
      "epoch": 0.81,
      "learning_rate": 5.576113578589034e-08,
      "logits/chosen": -1.9286285638809204,
      "logits/rejected": -1.9253568649291992,
      "logps/chosen": -31.583232879638672,
      "logps/rejected": -33.75123596191406,
      "loss": 0.3881,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.016091840341687202,
      "rewards/margins": 0.023359118029475212,
      "rewards/rejected": -0.007267280016094446,
      "step": 310
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.229036944380912e-08,
      "logits/chosen": -1.9803975820541382,
      "logits/rejected": -1.9680900573730469,
      "logps/chosen": -34.58079147338867,
      "logps/rejected": -33.575767517089844,
      "loss": 0.3778,
      "rewards/accuracies": 0.5375000238418579,
      "rewards/chosen": 0.012225466780364513,
      "rewards/margins": 0.041770923882722855,
      "rewards/rejected": -0.029545456171035767,
      "step": 320
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.053082288996112e-08,
      "logits/chosen": -2.015634298324585,
      "logits/rejected": -2.0141713619232178,
      "logps/chosen": -33.45996856689453,
      "logps/rejected": -32.48029327392578,
      "loss": 0.4032,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.020660031586885452,
      "rewards/margins": 0.021225089207291603,
      "rewards/rejected": -0.0005650619277730584,
      "step": 330
    },
    {
      "epoch": 0.88,
      "learning_rate": 2.05793773749158e-08,
      "logits/chosen": -2.1035873889923096,
      "logits/rejected": -2.0877768993377686,
      "logps/chosen": -34.162208557128906,
      "logps/rejected": -33.095733642578125,
      "loss": 0.439,
      "rewards/accuracies": 0.4625000059604645,
      "rewards/chosen": 0.013746557757258415,
      "rewards/margins": -0.004091509617865086,
      "rewards/rejected": 0.017838066443800926,
      "step": 340
    },
    {
      "epoch": 0.91,
      "learning_rate": 1.251801807404168e-08,
      "logits/chosen": -1.9746344089508057,
      "logits/rejected": -1.9736888408660889,
      "logps/chosen": -33.253448486328125,
      "logps/rejected": -32.458499908447266,
      "loss": 0.3954,
      "rewards/accuracies": 0.5249999761581421,
      "rewards/chosen": 0.02291743829846382,
      "rewards/margins": 0.017600122839212418,
      "rewards/rejected": 0.005317316390573978,
      "step": 350
    },
    {
      "epoch": 0.94,
      "learning_rate": 6.41315865106129e-09,
      "logits/chosen": -1.9309253692626953,
      "logits/rejected": -1.941303014755249,
      "logps/chosen": -32.21305465698242,
      "logps/rejected": -35.309688568115234,
      "loss": 0.4095,
      "rewards/accuracies": 0.42500001192092896,
      "rewards/chosen": -0.006690672133117914,
      "rewards/margins": 0.010053041391074657,
      "rewards/rejected": -0.01674371212720871,
      "step": 360
    },
    {
      "epoch": 0.96,
      "learning_rate": 2.3150941078050324e-09,
      "logits/chosen": -2.069108009338379,
      "logits/rejected": -2.0625574588775635,
      "logps/chosen": -33.637176513671875,
      "logps/rejected": -29.221187591552734,
      "loss": 0.4059,
      "rewards/accuracies": 0.48750001192092896,
      "rewards/chosen": 0.011833530850708485,
      "rewards/margins": 0.00764369685202837,
      "rewards/rejected": 0.004189834464341402,
      "step": 370
    },
    {
      "epoch": 0.99,
      "learning_rate": 2.575864278703266e-10,
      "logits/chosen": -1.9292182922363281,
      "logits/rejected": -1.9313886165618896,
      "logps/chosen": -34.24732971191406,
      "logps/rejected": -30.893402099609375,
      "loss": 0.4086,
      "rewards/accuracies": 0.4749999940395355,
      "rewards/chosen": -0.006622877903282642,
      "rewards/margins": 0.009996414184570312,
      "rewards/rejected": -0.01661929115653038,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.4092799360101873,
      "train_runtime": 3256.9261,
      "train_samples_per_second": 0.945,
      "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
}