{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9986824769433466,
  "eval_steps": 100,
  "global_step": 379,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 34.01362102288599,
      "learning_rate": 1.3157894736842104e-08,
      "logits/chosen": -4.685327529907227,
      "logits/rejected": -4.87608528137207,
      "logps/chosen": -207.7137451171875,
      "logps/rejected": -145.5098114013672,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 33.835137410082986,
      "learning_rate": 1.3157894736842104e-07,
      "logits/chosen": -4.499408721923828,
      "logits/rejected": -4.84108829498291,
      "logps/chosen": -223.5843048095703,
      "logps/rejected": -160.73016357421875,
      "loss": 0.6929,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.0008713232818990946,
      "rewards/margins": 3.9665635995334014e-05,
      "rewards/rejected": 0.000831657787784934,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 32.9122864599857,
      "learning_rate": 2.631578947368421e-07,
      "logits/chosen": -4.521907806396484,
      "logits/rejected": -4.8204779624938965,
      "logps/chosen": -220.248779296875,
      "logps/rejected": -173.30508422851562,
      "loss": 0.6818,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": 0.01710951328277588,
      "rewards/margins": 0.023774703964591026,
      "rewards/rejected": -0.006665193475782871,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 32.91997500077958,
      "learning_rate": 3.9473684210526315e-07,
      "logits/chosen": -4.713895320892334,
      "logits/rejected": -5.012445449829102,
      "logps/chosen": -223.3787841796875,
      "logps/rejected": -201.9126434326172,
      "loss": 0.6294,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -0.047253355383872986,
      "rewards/margins": 0.1270204484462738,
      "rewards/rejected": -0.1742737889289856,
      "step": 30
    },
    {
      "epoch": 0.11,
      "grad_norm": 44.205273980146465,
      "learning_rate": 4.999575626062319e-07,
      "logits/chosen": -4.862967491149902,
      "logits/rejected": -5.199351787567139,
      "logps/chosen": -284.15850830078125,
      "logps/rejected": -260.1568298339844,
      "loss": 0.5878,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.5039564371109009,
      "rewards/margins": 0.3390721082687378,
      "rewards/rejected": -0.8430284261703491,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 33.79573625552293,
      "learning_rate": 4.984737660598186e-07,
      "logits/chosen": -4.860326766967773,
      "logits/rejected": -5.1770920753479,
      "logps/chosen": -287.49212646484375,
      "logps/rejected": -275.1466979980469,
      "loss": 0.544,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": -0.6591841578483582,
      "rewards/margins": 0.3825286030769348,
      "rewards/rejected": -1.0417125225067139,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 37.5548220875479,
      "learning_rate": 4.948824853131236e-07,
      "logits/chosen": -5.250467777252197,
      "logits/rejected": -5.646960258483887,
      "logps/chosen": -355.8065185546875,
      "logps/rejected": -370.16064453125,
      "loss": 0.4873,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -1.3389923572540283,
      "rewards/margins": 0.6164921522140503,
      "rewards/rejected": -1.955484390258789,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 40.80845104625175,
      "learning_rate": 4.892141805936084e-07,
      "logits/chosen": -5.506089687347412,
      "logits/rejected": -5.9388532638549805,
      "logps/chosen": -373.0067443847656,
      "logps/rejected": -409.2164001464844,
      "loss": 0.4619,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -1.4274970293045044,
      "rewards/margins": 0.8630696535110474,
      "rewards/rejected": -2.290566921234131,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 62.51614299706751,
      "learning_rate": 4.81516928858564e-07,
      "logits/chosen": -5.957489967346191,
      "logits/rejected": -6.362034797668457,
      "logps/chosen": -431.8949279785156,
      "logps/rejected": -484.7469177246094,
      "loss": 0.4382,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -2.0529427528381348,
      "rewards/margins": 1.0657070875167847,
      "rewards/rejected": -3.118650197982788,
      "step": 80
    },
    {
      "epoch": 0.24,
      "grad_norm": 42.64563417225672,
      "learning_rate": 4.7185601601995784e-07,
      "logits/chosen": -5.974350929260254,
      "logits/rejected": -6.67104959487915,
      "logps/chosen": -387.2274475097656,
      "logps/rejected": -464.65826416015625,
      "loss": 0.4236,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -1.6722873449325562,
      "rewards/margins": 1.3039813041687012,
      "rewards/rejected": -2.976268768310547,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 49.657859383445725,
      "learning_rate": 4.603133832077953e-07,
      "logits/chosen": -6.59436559677124,
      "logits/rejected": -7.055686950683594,
      "logps/chosen": -443.9141540527344,
      "logps/rejected": -517.8334350585938,
      "loss": 0.4015,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -2.304152011871338,
      "rewards/margins": 1.163777232170105,
      "rewards/rejected": -3.4679291248321533,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": -6.410887718200684,
      "eval_logits/rejected": -6.793334007263184,
      "eval_logps/chosen": -826.5081176757812,
      "eval_logps/rejected": -987.8357543945312,
      "eval_loss": 0.9856035113334656,
      "eval_rewards/accuracies": 0.52734375,
      "eval_rewards/chosen": -4.360336780548096,
      "eval_rewards/margins": 0.3341439962387085,
      "eval_rewards/rejected": -4.694480895996094,
      "eval_runtime": 97.5721,
      "eval_samples_per_second": 20.498,
      "eval_steps_per_second": 0.328,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 59.660602657641974,
      "learning_rate": 4.4698693176863316e-07,
      "logits/chosen": -6.508014678955078,
      "logits/rejected": -6.995993137359619,
      "logps/chosen": -467.217041015625,
      "logps/rejected": -571.2991943359375,
      "loss": 0.3652,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -2.5293219089508057,
      "rewards/margins": 1.5391861200332642,
      "rewards/rejected": -4.068508148193359,
      "step": 110
    },
    {
      "epoch": 0.32,
      "grad_norm": 48.65436374701769,
      "learning_rate": 4.319896928940505e-07,
      "logits/chosen": -6.631407737731934,
      "logits/rejected": -7.224958896636963,
      "logps/chosen": -449.9879455566406,
      "logps/rejected": -551.552734375,
      "loss": 0.3856,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -2.370181083679199,
      "rewards/margins": 1.3927323818206787,
      "rewards/rejected": -3.762913465499878,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 41.65961698696447,
      "learning_rate": 4.1544886892205354e-07,
      "logits/chosen": -6.516014099121094,
      "logits/rejected": -7.032981872558594,
      "logps/chosen": -490.3934020996094,
      "logps/rejected": -610.3671264648438,
      "loss": 0.3736,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -2.659696578979492,
      "rewards/margins": 1.6181414127349854,
      "rewards/rejected": -4.277838230133057,
      "step": 130
    },
    {
      "epoch": 0.37,
      "grad_norm": 61.060769075088814,
      "learning_rate": 3.975047544428254e-07,
      "logits/chosen": -6.8892974853515625,
      "logits/rejected": -7.428150177001953,
      "logps/chosen": -479.0992736816406,
      "logps/rejected": -578.5732421875,
      "loss": 0.3918,
      "rewards/accuracies": 0.7875000238418579,
      "rewards/chosen": -2.728147506713867,
      "rewards/margins": 1.3954027891159058,
      "rewards/rejected": -4.123549938201904,
      "step": 140
    },
    {
      "epoch": 0.4,
      "grad_norm": 40.0921041292044,
      "learning_rate": 3.78309546359696e-07,
      "logits/chosen": -7.085695743560791,
      "logits/rejected": -7.593710422515869,
      "logps/chosen": -485.8174743652344,
      "logps/rejected": -611.1568603515625,
      "loss": 0.3821,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -2.8462297916412354,
      "rewards/margins": 1.527190089225769,
      "rewards/rejected": -4.373419761657715,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 45.94809301945095,
      "learning_rate": 3.580260529980584e-07,
      "logits/chosen": -6.79840087890625,
      "logits/rejected": -7.57172155380249,
      "logps/chosen": -476.08197021484375,
      "logps/rejected": -613.65576171875,
      "loss": 0.3533,
      "rewards/accuracies": 0.8062499761581421,
      "rewards/chosen": -2.6100144386291504,
      "rewards/margins": 1.7850373983383179,
      "rewards/rejected": -4.3950514793396,
      "step": 160
    },
    {
      "epoch": 0.45,
      "grad_norm": 49.925985141071024,
      "learning_rate": 3.36826313211205e-07,
      "logits/chosen": -7.27915096282959,
      "logits/rejected": -7.999810695648193,
      "logps/chosen": -464.27606201171875,
      "logps/rejected": -618.3897705078125,
      "loss": 0.3586,
      "rewards/accuracies": 0.8125,
      "rewards/chosen": -2.638946056365967,
      "rewards/margins": 1.9340379238128662,
      "rewards/rejected": -4.572983741760254,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 49.77263247276722,
      "learning_rate": 3.14890137195437e-07,
      "logits/chosen": -7.062252998352051,
      "logits/rejected": -7.757845878601074,
      "logps/chosen": -499.0604553222656,
      "logps/rejected": -631.4205322265625,
      "loss": 0.35,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -2.841132640838623,
      "rewards/margins": 1.6729885339736938,
      "rewards/rejected": -4.514122009277344,
      "step": 180
    },
    {
      "epoch": 0.5,
      "grad_norm": 47.32333832777734,
      "learning_rate": 2.9240358139084013e-07,
      "logits/chosen": -7.432755947113037,
      "logits/rejected": -8.177480697631836,
      "logps/chosen": -571.7742919921875,
      "logps/rejected": -753.3114013671875,
      "loss": 0.3584,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -3.5975048542022705,
      "rewards/margins": 2.1865601539611816,
      "rewards/rejected": -5.784065246582031,
      "step": 190
    },
    {
      "epoch": 0.53,
      "grad_norm": 47.64189353399637,
      "learning_rate": 2.695573704031885e-07,
      "logits/chosen": -6.939781188964844,
      "logits/rejected": -7.71111536026001,
      "logps/chosen": -500.54449462890625,
      "logps/rejected": -651.9711303710938,
      "loss": 0.3649,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -2.752312421798706,
      "rewards/margins": 1.9267911911010742,
      "rewards/rejected": -4.679104328155518,
      "step": 200
    },
    {
      "epoch": 0.53,
      "eval_logits/chosen": -7.152472496032715,
      "eval_logits/rejected": -7.637792110443115,
      "eval_logps/chosen": -878.0755615234375,
      "eval_logps/rejected": -1032.680908203125,
      "eval_loss": 1.1239182949066162,
      "eval_rewards/accuracies": 0.48828125,
      "eval_rewards/chosen": -4.876009941101074,
      "eval_rewards/margins": 0.26692283153533936,
      "eval_rewards/rejected": -5.142932891845703,
      "eval_runtime": 97.9349,
      "eval_samples_per_second": 20.422,
      "eval_steps_per_second": 0.327,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 54.30113996467665,
      "learning_rate": 2.465452793317865e-07,
      "logits/chosen": -6.918679237365723,
      "logits/rejected": -7.81919002532959,
      "logps/chosen": -490.4891662597656,
      "logps/rejected": -645.8287353515625,
      "loss": 0.351,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -2.56154727935791,
      "rewards/margins": 1.9791193008422852,
      "rewards/rejected": -4.5406670570373535,
      "step": 210
    },
    {
      "epoch": 0.58,
      "grad_norm": 48.87773515687126,
      "learning_rate": 2.2356249022388789e-07,
      "logits/chosen": -7.1236982345581055,
      "logits/rejected": -8.037015914916992,
      "logps/chosen": -493.3589782714844,
      "logps/rejected": -644.3180541992188,
      "loss": 0.3503,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -2.7893471717834473,
      "rewards/margins": 2.016085386276245,
      "rewards/rejected": -4.805432319641113,
      "step": 220
    },
    {
      "epoch": 0.61,
      "grad_norm": 42.654084351174774,
      "learning_rate": 2.0080393659578038e-07,
      "logits/chosen": -7.341279029846191,
      "logits/rejected": -8.349458694458008,
      "logps/chosen": -546.9940795898438,
      "logps/rejected": -735.5299072265625,
      "loss": 0.3336,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -3.2596652507781982,
      "rewards/margins": 2.3632309436798096,
      "rewards/rejected": -5.622895240783691,
      "step": 230
    },
    {
      "epoch": 0.63,
      "grad_norm": 48.836991831948275,
      "learning_rate": 1.7846265006183976e-07,
      "logits/chosen": -7.382364749908447,
      "logits/rejected": -8.096385955810547,
      "logps/chosen": -559.1055908203125,
      "logps/rejected": -690.3636474609375,
      "loss": 0.3239,
      "rewards/accuracies": 0.78125,
      "rewards/chosen": -3.478092670440674,
      "rewards/margins": 1.7500022649765015,
      "rewards/rejected": -5.228094577789307,
      "step": 240
    },
    {
      "epoch": 0.66,
      "grad_norm": 41.88272757196894,
      "learning_rate": 1.5672812309497722e-07,
      "logits/chosen": -7.3642730712890625,
      "logits/rejected": -8.137662887573242,
      "logps/chosen": -481.7484436035156,
      "logps/rejected": -658.2447509765625,
      "loss": 0.3435,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -2.742130756378174,
      "rewards/margins": 2.204786777496338,
      "rewards/rejected": -4.946917533874512,
      "step": 250
    },
    {
      "epoch": 0.69,
      "grad_norm": 42.22161557256122,
      "learning_rate": 1.357847018050843e-07,
      "logits/chosen": -7.285035133361816,
      "logits/rejected": -8.04423999786377,
      "logps/chosen": -495.21929931640625,
      "logps/rejected": -637.4512939453125,
      "loss": 0.3575,
      "rewards/accuracies": 0.856249988079071,
      "rewards/chosen": -2.5677530765533447,
      "rewards/margins": 1.859423041343689,
      "rewards/rejected": -4.427175521850586,
      "step": 260
    },
    {
      "epoch": 0.71,
      "grad_norm": 36.62148112990462,
      "learning_rate": 1.1581002236747328e-07,
      "logits/chosen": -7.350560188293457,
      "logits/rejected": -8.334383964538574,
      "logps/chosen": -470.9756774902344,
      "logps/rejected": -644.9085083007812,
      "loss": 0.3419,
      "rewards/accuracies": 0.831250011920929,
      "rewards/chosen": -2.7616236209869385,
      "rewards/margins": 2.1581368446350098,
      "rewards/rejected": -4.919760704040527,
      "step": 270
    },
    {
      "epoch": 0.74,
      "grad_norm": 42.46981787888043,
      "learning_rate": 9.697350436308427e-08,
      "logits/chosen": -7.353733062744141,
      "logits/rejected": -8.123218536376953,
      "logps/chosen": -547.2511596679688,
      "logps/rejected": -669.4581298828125,
      "loss": 0.3277,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -3.1215298175811768,
      "rewards/margins": 1.7408149242401123,
      "rewards/rejected": -4.862344264984131,
      "step": 280
    },
    {
      "epoch": 0.76,
      "grad_norm": 45.06982020026593,
      "learning_rate": 7.943491380952188e-08,
      "logits/chosen": -7.6978254318237305,
      "logits/rejected": -8.468889236450195,
      "logps/chosen": -488.8667907714844,
      "logps/rejected": -652.3041381835938,
      "loss": 0.3357,
      "rewards/accuracies": 0.862500011920929,
      "rewards/chosen": -2.835703134536743,
      "rewards/margins": 2.056410789489746,
      "rewards/rejected": -4.89211368560791,
      "step": 290
    },
    {
      "epoch": 0.79,
      "grad_norm": 50.46942762667779,
      "learning_rate": 6.334300807088508e-08,
      "logits/chosen": -7.3240485191345215,
      "logits/rejected": -8.329621315002441,
      "logps/chosen": -509.5638122558594,
      "logps/rejected": -673.7567749023438,
      "loss": 0.3506,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -3.0604350566864014,
      "rewards/margins": 2.1173126697540283,
      "rewards/rejected": -5.1777472496032715,
      "step": 300
    },
    {
      "epoch": 0.79,
      "eval_logits/chosen": -7.444587707519531,
      "eval_logits/rejected": -7.9900360107421875,
      "eval_logps/chosen": -951.4573974609375,
      "eval_logps/rejected": -1114.78076171875,
      "eval_loss": 1.1842519044876099,
      "eval_rewards/accuracies": 0.51171875,
      "eval_rewards/chosen": -5.609828948974609,
      "eval_rewards/margins": 0.35410135984420776,
      "eval_rewards/rejected": -5.963930606842041,
      "eval_runtime": 97.9073,
      "eval_samples_per_second": 20.427,
      "eval_steps_per_second": 0.327,
      "step": 300
    },
    {
      "epoch": 0.82,
      "grad_norm": 45.51913834484837,
      "learning_rate": 4.8834274139883084e-08,
      "logits/chosen": -7.380696773529053,
      "logits/rejected": -8.350247383117676,
      "logps/chosen": -504.37054443359375,
      "logps/rejected": -678.6265869140625,
      "loss": 0.3248,
      "rewards/accuracies": 0.84375,
      "rewards/chosen": -2.946779727935791,
      "rewards/margins": 2.2168285846710205,
      "rewards/rejected": -5.163609027862549,
      "step": 310
    },
    {
      "epoch": 0.84,
      "grad_norm": 53.031113784661194,
      "learning_rate": 3.60317709937693e-08,
      "logits/chosen": -7.483295440673828,
      "logits/rejected": -8.33633804321289,
      "logps/chosen": -519.6754150390625,
      "logps/rejected": -667.0064697265625,
      "loss": 0.334,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": -2.902660369873047,
      "rewards/margins": 2.0172414779663086,
      "rewards/rejected": -4.919901371002197,
      "step": 320
    },
    {
      "epoch": 0.87,
      "grad_norm": 41.68995945520798,
      "learning_rate": 2.5044085842905683e-08,
      "logits/chosen": -7.596086025238037,
      "logits/rejected": -8.42108154296875,
      "logps/chosen": -542.1593627929688,
      "logps/rejected": -761.1915893554688,
      "loss": 0.3439,
      "rewards/accuracies": 0.9312499761581421,
      "rewards/chosen": -3.237351894378662,
      "rewards/margins": 2.630371332168579,
      "rewards/rejected": -5.867722988128662,
      "step": 330
    },
    {
      "epoch": 0.9,
      "grad_norm": 44.38600213815557,
      "learning_rate": 1.5964413124758493e-08,
      "logits/chosen": -7.379315376281738,
      "logits/rejected": -8.234747886657715,
      "logps/chosen": -516.7742919921875,
      "logps/rejected": -690.0197143554688,
      "loss": 0.3394,
      "rewards/accuracies": 0.875,
      "rewards/chosen": -2.9615204334259033,
      "rewards/margins": 2.160275936126709,
      "rewards/rejected": -5.121796607971191,
      "step": 340
    },
    {
      "epoch": 0.92,
      "grad_norm": 49.890826040889124,
      "learning_rate": 8.869764055041501e-09,
      "logits/chosen": -7.525488376617432,
      "logits/rejected": -8.198974609375,
      "logps/chosen": -535.8231201171875,
      "logps/rejected": -737.1505737304688,
      "loss": 0.3403,
      "rewards/accuracies": 0.8374999761581421,
      "rewards/chosen": -3.1437525749206543,
      "rewards/margins": 2.3052563667297363,
      "rewards/rejected": -5.449008941650391,
      "step": 350
    },
    {
      "epoch": 0.95,
      "grad_norm": 41.28866479778979,
      "learning_rate": 3.82031344036729e-09,
      "logits/chosen": -7.295458793640137,
      "logits/rejected": -8.147361755371094,
      "logps/chosen": -541.268798828125,
      "logps/rejected": -692.4953002929688,
      "loss": 0.3191,
      "rewards/accuracies": 0.8187500238418579,
      "rewards/chosen": -3.197694778442383,
      "rewards/margins": 1.9286365509033203,
      "rewards/rejected": -5.126331329345703,
      "step": 360
    },
    {
      "epoch": 0.97,
      "grad_norm": 56.54843526627352,
      "learning_rate": 8.588892925590063e-10,
      "logits/chosen": -7.414445400238037,
      "logits/rejected": -8.447690963745117,
      "logps/chosen": -499.84197998046875,
      "logps/rejected": -670.2615966796875,
      "loss": 0.3342,
      "rewards/accuracies": 0.8500000238418579,
      "rewards/chosen": -2.8063206672668457,
      "rewards/margins": 2.2892918586730957,
      "rewards/rejected": -5.0956130027771,
      "step": 370
    },
    {
      "epoch": 1.0,
      "step": 379,
      "total_flos": 0.0,
      "train_loss": 0.39733357479830217,
      "train_runtime": 5845.1464,
      "train_samples_per_second": 8.303,
      "train_steps_per_second": 0.065
    }
  ],
  "logging_steps": 10,
  "max_steps": 379,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}