{
  "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,
      "grad_norm": 26.125,
      "learning_rate": 1.282051282051282e-07,
      "logits/chosen": 88.18099975585938,
      "logits/rejected": 88.25153350830078,
      "logps/chosen": -29.073104858398438,
      "logps/rejected": -26.25731658935547,
      "loss": 1.0,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.03,
      "grad_norm": 21.75,
      "learning_rate": 1.282051282051282e-06,
      "logits/chosen": 81.06690979003906,
      "logits/rejected": 80.77025604248047,
      "logps/chosen": -34.31000900268555,
      "logps/rejected": -33.0362548828125,
      "loss": 0.9981,
      "rewards/accuracies": 0.4861111044883728,
      "rewards/chosen": -0.013656922616064548,
      "rewards/margins": 0.0019124278333038092,
      "rewards/rejected": -0.015569348819553852,
      "step": 10
    },
    {
      "epoch": 0.05,
      "grad_norm": 25.125,
      "learning_rate": 2.564102564102564e-06,
      "logits/chosen": 80.65390014648438,
      "logits/rejected": 80.54225158691406,
      "logps/chosen": -33.533935546875,
      "logps/rejected": -30.82293128967285,
      "loss": 0.9597,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.0311945378780365,
      "rewards/margins": 0.04028325527906418,
      "rewards/rejected": -0.009088722057640553,
      "step": 20
    },
    {
      "epoch": 0.08,
      "grad_norm": 24.25,
      "learning_rate": 3.846153846153847e-06,
      "logits/chosen": 82.50422668457031,
      "logits/rejected": 82.536376953125,
      "logps/chosen": -33.82068634033203,
      "logps/rejected": -31.21515464782715,
      "loss": 1.0033,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.03842673823237419,
      "rewards/margins": -0.0033095614053308964,
      "rewards/rejected": 0.04173629730939865,
      "step": 30
    },
    {
      "epoch": 0.1,
      "grad_norm": 21.25,
      "learning_rate": 4.999896948438434e-06,
      "logits/chosen": 81.09244537353516,
      "logits/rejected": 81.08795166015625,
      "logps/chosen": -32.703147888183594,
      "logps/rejected": -33.152015686035156,
      "loss": 0.9509,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.09418319165706635,
      "rewards/margins": 0.050881147384643555,
      "rewards/rejected": 0.043302033096551895,
      "step": 40
    },
    {
      "epoch": 0.13,
      "grad_norm": 16.875,
      "learning_rate": 4.987541037542187e-06,
      "logits/chosen": 78.85831451416016,
      "logits/rejected": 78.8629379272461,
      "logps/chosen": -30.369918823242188,
      "logps/rejected": -30.639278411865234,
      "loss": 0.9365,
      "rewards/accuracies": 0.550000011920929,
      "rewards/chosen": 0.16706591844558716,
      "rewards/margins": 0.08026467263698578,
      "rewards/rejected": 0.08680125325918198,
      "step": 50
    },
    {
      "epoch": 0.16,
      "grad_norm": 19.25,
      "learning_rate": 4.954691471941119e-06,
      "logits/chosen": 83.56826782226562,
      "logits/rejected": 83.62416076660156,
      "logps/chosen": -30.741540908813477,
      "logps/rejected": -29.298175811767578,
      "loss": 0.9795,
      "rewards/accuracies": 0.4000000059604645,
      "rewards/chosen": 0.08672446757555008,
      "rewards/margins": 0.028255227953195572,
      "rewards/rejected": 0.05846923589706421,
      "step": 60
    },
    {
      "epoch": 0.18,
      "grad_norm": 24.125,
      "learning_rate": 4.901618883413549e-06,
      "logits/chosen": 84.20392608642578,
      "logits/rejected": 84.23384857177734,
      "logps/chosen": -30.30936050415039,
      "logps/rejected": -32.60371398925781,
      "loss": 1.0229,
      "rewards/accuracies": 0.4375,
      "rewards/chosen": 0.06393321603536606,
      "rewards/margins": -0.022865157574415207,
      "rewards/rejected": 0.08679838478565216,
      "step": 70
    },
    {
      "epoch": 0.21,
      "grad_norm": 23.625,
      "learning_rate": 4.828760511501322e-06,
      "logits/chosen": 81.97456359863281,
      "logits/rejected": 81.94981384277344,
      "logps/chosen": -30.98190689086914,
      "logps/rejected": -30.73152732849121,
      "loss": 0.9071,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.10460755974054337,
      "rewards/margins": 0.09938390552997589,
      "rewards/rejected": 0.005223660264164209,
      "step": 80
    },
    {
      "epoch": 0.23,
      "grad_norm": 29.875,
      "learning_rate": 4.7367166013034295e-06,
      "logits/chosen": 78.75807189941406,
      "logits/rejected": 78.72810363769531,
      "logps/chosen": -32.24687957763672,
      "logps/rejected": -31.013330459594727,
      "loss": 0.9168,
      "rewards/accuracies": 0.574999988079071,
      "rewards/chosen": 0.07826016843318939,
      "rewards/margins": 0.09063171595335007,
      "rewards/rejected": -0.012371541932225227,
      "step": 90
    },
    {
      "epoch": 0.26,
      "grad_norm": 25.375,
      "learning_rate": 4.626245458345211e-06,
      "logits/chosen": 83.73965454101562,
      "logits/rejected": 83.75724792480469,
      "logps/chosen": -33.909358978271484,
      "logps/rejected": -31.719879150390625,
      "loss": 0.927,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": 0.0791260302066803,
      "rewards/margins": 0.07299245893955231,
      "rewards/rejected": 0.006133575923740864,
      "step": 100
    },
    {
      "epoch": 0.26,
      "eval_logits/chosen": 98.71903228759766,
      "eval_logits/rejected": 98.71150207519531,
      "eval_logps/chosen": -32.34075927734375,
      "eval_logps/rejected": -35.99679183959961,
      "eval_loss": 0.9733805060386658,
      "eval_rewards/accuracies": 0.5598006844520569,
      "eval_rewards/chosen": 0.0204838328063488,
      "eval_rewards/margins": 0.026536403223872185,
      "eval_rewards/rejected": -0.006052570417523384,
      "eval_runtime": 104.2721,
      "eval_samples_per_second": 3.289,
      "eval_steps_per_second": 0.412,
      "step": 100
    },
    {
      "epoch": 0.29,
      "grad_norm": 28.5,
      "learning_rate": 4.498257201263691e-06,
      "logits/chosen": 83.88168334960938,
      "logits/rejected": 83.75785827636719,
      "logps/chosen": -32.253421783447266,
      "logps/rejected": -32.89667510986328,
      "loss": 0.8099,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.15387704968452454,
      "rewards/margins": 0.202034592628479,
      "rewards/rejected": -0.04815753549337387,
      "step": 110
    },
    {
      "epoch": 0.31,
      "grad_norm": 26.25,
      "learning_rate": 4.353806263777678e-06,
      "logits/chosen": 83.8781967163086,
      "logits/rejected": 83.98338317871094,
      "logps/chosen": -28.098907470703125,
      "logps/rejected": -35.39207077026367,
      "loss": 0.8496,
      "rewards/accuracies": 0.6000000238418579,
      "rewards/chosen": 0.1661352962255478,
      "rewards/margins": 0.1527193784713745,
      "rewards/rejected": 0.013415923342108727,
      "step": 120
    },
    {
      "epoch": 0.34,
      "grad_norm": 18.875,
      "learning_rate": 4.1940827077152755e-06,
      "logits/chosen": 81.01908111572266,
      "logits/rejected": 81.05055236816406,
      "logps/chosen": -30.180002212524414,
      "logps/rejected": -31.972553253173828,
      "loss": 0.8597,
      "rewards/accuracies": 0.612500011920929,
      "rewards/chosen": 0.14547115564346313,
      "rewards/margins": 0.15365630388259888,
      "rewards/rejected": -0.008185150101780891,
      "step": 130
    },
    {
      "epoch": 0.36,
      "grad_norm": 22.375,
      "learning_rate": 4.0204024186666215e-06,
      "logits/chosen": 81.88218688964844,
      "logits/rejected": 81.89274597167969,
      "logps/chosen": -26.96360206604004,
      "logps/rejected": -32.96726608276367,
      "loss": 0.803,
      "rewards/accuracies": 0.6625000238418579,
      "rewards/chosen": 0.0967448502779007,
      "rewards/margins": 0.20552262663841248,
      "rewards/rejected": -0.10877779871225357,
      "step": 140
    },
    {
      "epoch": 0.39,
      "grad_norm": 24.375,
      "learning_rate": 3.834196265035119e-06,
      "logits/chosen": 80.27394104003906,
      "logits/rejected": 80.23723602294922,
      "logps/chosen": -28.884967803955078,
      "logps/rejected": -33.30558776855469,
      "loss": 0.77,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.10082833468914032,
      "rewards/margins": 0.2486453801393509,
      "rewards/rejected": -0.14781701564788818,
      "step": 150
    },
    {
      "epoch": 0.42,
      "grad_norm": 26.75,
      "learning_rate": 3.636998309800573e-06,
      "logits/chosen": 81.85964965820312,
      "logits/rejected": 81.88619232177734,
      "logps/chosen": -34.064693450927734,
      "logps/rejected": -30.97774887084961,
      "loss": 0.8017,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.014411324635148048,
      "rewards/margins": 0.21945340931415558,
      "rewards/rejected": -0.20504209399223328,
      "step": 160
    },
    {
      "epoch": 0.44,
      "grad_norm": 26.5,
      "learning_rate": 3.4304331721118078e-06,
      "logits/chosen": 82.6593246459961,
      "logits/rejected": 82.60637664794922,
      "logps/chosen": -31.141841888427734,
      "logps/rejected": -33.009368896484375,
      "loss": 0.8017,
      "rewards/accuracies": 0.6499999761581421,
      "rewards/chosen": 0.023489754647016525,
      "rewards/margins": 0.22135159373283386,
      "rewards/rejected": -0.19786186516284943,
      "step": 170
    },
    {
      "epoch": 0.47,
      "grad_norm": 20.0,
      "learning_rate": 3.2162026428305436e-06,
      "logits/chosen": 80.07731628417969,
      "logits/rejected": 80.05589294433594,
      "logps/chosen": -30.7539119720459,
      "logps/rejected": -32.01999282836914,
      "loss": 0.7558,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.08864770829677582,
      "rewards/margins": 0.2612043023109436,
      "rewards/rejected": -0.1725565642118454,
      "step": 180
    },
    {
      "epoch": 0.49,
      "grad_norm": 14.8125,
      "learning_rate": 2.996071664294641e-06,
      "logits/chosen": 81.69454193115234,
      "logits/rejected": 81.68315887451172,
      "logps/chosen": -30.46109962463379,
      "logps/rejected": -31.09275245666504,
      "loss": 0.8327,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.06106363609433174,
      "rewards/margins": 0.18600967526435852,
      "rewards/rejected": -0.12494603544473648,
      "step": 190
    },
    {
      "epoch": 0.52,
      "grad_norm": 21.125,
      "learning_rate": 2.7718537898066833e-06,
      "logits/chosen": 76.97463989257812,
      "logits/rejected": 76.91825866699219,
      "logps/chosen": -34.04479217529297,
      "logps/rejected": -33.270973205566406,
      "loss": 0.7448,
      "rewards/accuracies": 0.637499988079071,
      "rewards/chosen": 0.14158903062343597,
      "rewards/margins": 0.2971034646034241,
      "rewards/rejected": -0.1555144339799881,
      "step": 200
    },
    {
      "epoch": 0.52,
      "eval_logits/chosen": 98.2271499633789,
      "eval_logits/rejected": 98.20570373535156,
      "eval_logps/chosen": -32.86313247680664,
      "eval_logps/rejected": -36.650115966796875,
      "eval_loss": 0.9481998085975647,
      "eval_rewards/accuracies": 0.5307309031486511,
      "eval_rewards/chosen": -0.08399129658937454,
      "eval_rewards/margins": 0.05272620916366577,
      "eval_rewards/rejected": -0.13671749830245972,
      "eval_runtime": 103.8483,
      "eval_samples_per_second": 3.303,
      "eval_steps_per_second": 0.414,
      "step": 200
    },
    {
      "epoch": 0.55,
      "grad_norm": 34.0,
      "learning_rate": 2.5453962426402006e-06,
      "logits/chosen": 79.6450424194336,
      "logits/rejected": 79.561767578125,
      "logps/chosen": -33.472808837890625,
      "logps/rejected": -35.81511688232422,
      "loss": 0.7697,
      "rewards/accuracies": 0.675000011920929,
      "rewards/chosen": 0.09458215534687042,
      "rewards/margins": 0.2603868544101715,
      "rewards/rejected": -0.1658046841621399,
      "step": 210
    },
    {
      "epoch": 0.57,
      "grad_norm": 24.0,
      "learning_rate": 2.3185646976551794e-06,
      "logits/chosen": 81.76014709472656,
      "logits/rejected": 81.84996795654297,
      "logps/chosen": -31.052398681640625,
      "logps/rejected": -31.682384490966797,
      "loss": 0.681,
      "rewards/accuracies": 0.7749999761581421,
      "rewards/chosen": 0.17931082844734192,
      "rewards/margins": 0.3699965178966522,
      "rewards/rejected": -0.1906857192516327,
      "step": 220
    },
    {
      "epoch": 0.6,
      "grad_norm": 26.625,
      "learning_rate": 2.0932279108998323e-06,
      "logits/chosen": 78.84847259521484,
      "logits/rejected": 78.89868927001953,
      "logps/chosen": -32.36001205444336,
      "logps/rejected": -34.869354248046875,
      "loss": 0.7831,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.08914650976657867,
      "rewards/margins": 0.26058703660964966,
      "rewards/rejected": -0.17144052684307098,
      "step": 230
    },
    {
      "epoch": 0.62,
      "grad_norm": 26.125,
      "learning_rate": 1.8712423238279358e-06,
      "logits/chosen": 81.05337524414062,
      "logits/rejected": 81.3609848022461,
      "logps/chosen": -30.954391479492188,
      "logps/rejected": -32.368465423583984,
      "loss": 0.7122,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.14673486351966858,
      "rewards/margins": 0.3140195310115814,
      "rewards/rejected": -0.16728468239307404,
      "step": 240
    },
    {
      "epoch": 0.65,
      "grad_norm": 22.375,
      "learning_rate": 1.6544367689701824e-06,
      "logits/chosen": 79.7140884399414,
      "logits/rejected": 79.77289581298828,
      "logps/chosen": -27.334728240966797,
      "logps/rejected": -30.74212646484375,
      "loss": 0.8141,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.05223376676440239,
      "rewards/margins": 0.22051122784614563,
      "rewards/rejected": -0.16827745735645294,
      "step": 250
    },
    {
      "epoch": 0.68,
      "grad_norm": 20.5,
      "learning_rate": 1.4445974030621963e-06,
      "logits/chosen": 76.9362564086914,
      "logits/rejected": 77.0797119140625,
      "logps/chosen": -30.7308292388916,
      "logps/rejected": -37.249969482421875,
      "loss": 0.6632,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.1660328060388565,
      "rewards/margins": 0.3957008719444275,
      "rewards/rejected": -0.2296680212020874,
      "step": 260
    },
    {
      "epoch": 0.7,
      "grad_norm": 23.625,
      "learning_rate": 1.243452991757889e-06,
      "logits/chosen": 76.16326904296875,
      "logits/rejected": 76.19123840332031,
      "logps/chosen": -31.115991592407227,
      "logps/rejected": -32.44744110107422,
      "loss": 0.709,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.13258597254753113,
      "rewards/margins": 0.32099050283432007,
      "rewards/rejected": -0.18840453028678894,
      "step": 270
    },
    {
      "epoch": 0.73,
      "grad_norm": 35.5,
      "learning_rate": 1.0526606671603523e-06,
      "logits/chosen": 79.00858306884766,
      "logits/rejected": 78.7801513671875,
      "logps/chosen": -31.33221435546875,
      "logps/rejected": -30.144094467163086,
      "loss": 0.8422,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.09553596377372742,
      "rewards/margins": 0.19095517694950104,
      "rewards/rejected": -0.09541920572519302,
      "step": 280
    },
    {
      "epoch": 0.75,
      "grad_norm": 23.375,
      "learning_rate": 8.737922755071455e-07,
      "logits/chosen": 79.1707763671875,
      "logits/rejected": 79.09210205078125,
      "logps/chosen": -33.15503692626953,
      "logps/rejected": -33.348018646240234,
      "loss": 0.611,
      "rewards/accuracies": 0.824999988079071,
      "rewards/chosen": 0.18984279036521912,
      "rewards/margins": 0.45607319474220276,
      "rewards/rejected": -0.26623034477233887,
      "step": 290
    },
    {
      "epoch": 0.78,
      "grad_norm": 23.5,
      "learning_rate": 7.08321427484816e-07,
      "logits/chosen": 74.80469512939453,
      "logits/rejected": 74.91273498535156,
      "logps/chosen": -32.506813049316406,
      "logps/rejected": -29.802093505859375,
      "loss": 0.7402,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.1797788441181183,
      "rewards/margins": 0.3197931945323944,
      "rewards/rejected": -0.14001430571079254,
      "step": 300
    },
    {
      "epoch": 0.78,
      "eval_logits/chosen": 98.21781921386719,
      "eval_logits/rejected": 98.18999481201172,
      "eval_logps/chosen": -32.84402084350586,
      "eval_logps/rejected": -36.679203033447266,
      "eval_loss": 0.9399670362472534,
      "eval_rewards/accuracies": 0.5622923374176025,
      "eval_rewards/chosen": -0.08016957342624664,
      "eval_rewards/margins": 0.062365565448999405,
      "eval_rewards/rejected": -0.14253515005111694,
      "eval_runtime": 104.0457,
      "eval_samples_per_second": 3.297,
      "eval_steps_per_second": 0.413,
      "step": 300
    },
    {
      "epoch": 0.81,
      "grad_norm": 23.25,
      "learning_rate": 5.576113578589035e-07,
      "logits/chosen": 82.0827865600586,
      "logits/rejected": 82.1087417602539,
      "logps/chosen": -30.205184936523438,
      "logps/rejected": -33.123497009277344,
      "loss": 0.7347,
      "rewards/accuracies": 0.75,
      "rewards/chosen": 0.13166961073875427,
      "rewards/margins": 0.31641754508018494,
      "rewards/rejected": -0.18474793434143066,
      "step": 310
    },
    {
      "epoch": 0.83,
      "grad_norm": 24.375,
      "learning_rate": 4.229036944380913e-07,
      "logits/chosen": 79.43605041503906,
      "logits/rejected": 79.44461822509766,
      "logps/chosen": -30.739145278930664,
      "logps/rejected": -29.764511108398438,
      "loss": 0.683,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.1718827486038208,
      "rewards/margins": 0.3408864438533783,
      "rewards/rejected": -0.1690036654472351,
      "step": 320
    },
    {
      "epoch": 0.86,
      "grad_norm": 18.375,
      "learning_rate": 3.053082288996112e-07,
      "logits/chosen": 76.54524230957031,
      "logits/rejected": 76.5948257446289,
      "logps/chosen": -29.108013153076172,
      "logps/rejected": -33.65383529663086,
      "loss": 0.5952,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.2489181011915207,
      "rewards/margins": 0.4527127742767334,
      "rewards/rejected": -0.2037946879863739,
      "step": 330
    },
    {
      "epoch": 0.88,
      "grad_norm": 27.25,
      "learning_rate": 2.0579377374915805e-07,
      "logits/chosen": 80.96257781982422,
      "logits/rejected": 81.00981140136719,
      "logps/chosen": -32.56684112548828,
      "logps/rejected": -34.382102966308594,
      "loss": 0.7363,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.1290624588727951,
      "rewards/margins": 0.3317454755306244,
      "rewards/rejected": -0.2026829719543457,
      "step": 340
    },
    {
      "epoch": 0.91,
      "grad_norm": 22.875,
      "learning_rate": 1.2518018074041684e-07,
      "logits/chosen": 79.97700500488281,
      "logits/rejected": 79.99113464355469,
      "logps/chosen": -32.53782272338867,
      "logps/rejected": -34.063255310058594,
      "loss": 0.6574,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": 0.22118382155895233,
      "rewards/margins": 0.41547471284866333,
      "rewards/rejected": -0.1942909061908722,
      "step": 350
    },
    {
      "epoch": 0.94,
      "grad_norm": 21.75,
      "learning_rate": 6.41315865106129e-08,
      "logits/chosen": 81.52592468261719,
      "logits/rejected": 81.56230163574219,
      "logps/chosen": -28.674850463867188,
      "logps/rejected": -32.256248474121094,
      "loss": 0.7516,
      "rewards/accuracies": 0.6875,
      "rewards/chosen": 0.1754719465970993,
      "rewards/margins": 0.29675185680389404,
      "rewards/rejected": -0.12127991020679474,
      "step": 360
    },
    {
      "epoch": 0.96,
      "grad_norm": 21.25,
      "learning_rate": 2.3150941078050325e-08,
      "logits/chosen": 80.93915557861328,
      "logits/rejected": 80.9583740234375,
      "logps/chosen": -32.185455322265625,
      "logps/rejected": -36.000953674316406,
      "loss": 0.7515,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.13814154267311096,
      "rewards/margins": 0.31645822525024414,
      "rewards/rejected": -0.1783166527748108,
      "step": 370
    },
    {
      "epoch": 0.99,
      "grad_norm": 23.875,
      "learning_rate": 2.575864278703266e-09,
      "logits/chosen": 74.74893951416016,
      "logits/rejected": 74.6153564453125,
      "logps/chosen": -30.02016258239746,
      "logps/rejected": -29.047134399414062,
      "loss": 0.751,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": 0.11119310557842255,
      "rewards/margins": 0.28850510716438293,
      "rewards/rejected": -0.1773119866847992,
      "step": 380
    },
    {
      "epoch": 1.0,
      "step": 385,
      "total_flos": 0.0,
      "train_loss": 0.8039178501475941,
      "train_runtime": 2556.1517,
      "train_samples_per_second": 1.205,
      "train_steps_per_second": 0.151
    }
  ],
  "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
}