{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.5823590019262644,
  "eval_steps": 1000,
  "global_step": 26000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.002239842315101017,
      "grad_norm": 71.10852813720703,
      "learning_rate": 0.0001,
      "loss": 3.5057,
      "step": 100
    },
    {
      "epoch": 0.004479684630202034,
      "grad_norm": 57.756534576416016,
      "learning_rate": 9.9998756572327e-05,
      "loss": 0.8254,
      "step": 200
    },
    {
      "epoch": 0.006719526945303051,
      "grad_norm": 44.61080551147461,
      "learning_rate": 9.999502635115246e-05,
      "loss": 0.6935,
      "step": 300
    },
    {
      "epoch": 0.008959369260404068,
      "grad_norm": 50.82619094848633,
      "learning_rate": 9.998880952200681e-05,
      "loss": 0.5972,
      "step": 400
    },
    {
      "epoch": 0.011199211575505085,
      "grad_norm": 44.26677703857422,
      "learning_rate": 9.998010639409713e-05,
      "loss": 0.5408,
      "step": 500
    },
    {
      "epoch": 0.013439053890606102,
      "grad_norm": 42.8045654296875,
      "learning_rate": 9.996891740029186e-05,
      "loss": 0.5509,
      "step": 600
    },
    {
      "epoch": 0.01567889620570712,
      "grad_norm": 30.989139556884766,
      "learning_rate": 9.995524309709913e-05,
      "loss": 0.4823,
      "step": 700
    },
    {
      "epoch": 0.017918738520808136,
      "grad_norm": 34.01952362060547,
      "learning_rate": 9.993908416463927e-05,
      "loss": 0.5111,
      "step": 800
    },
    {
      "epoch": 0.020158580835909153,
      "grad_norm": 34.079307556152344,
      "learning_rate": 9.992044140661079e-05,
      "loss": 0.4635,
      "step": 900
    },
    {
      "epoch": 0.02239842315101017,
      "grad_norm": 26.16071128845215,
      "learning_rate": 9.989931575025056e-05,
      "loss": 0.4883,
      "step": 1000
    },
    {
      "epoch": 0.02239842315101017,
      "eval_avg_non_pair_similarity": 0.0020343252948339737,
      "eval_avg_pair_similarity": 0.008852629057131708,
      "eval_loss": 0.5431402921676636,
      "eval_runtime": 19.6178,
      "eval_samples_per_second": 25.487,
      "eval_similarity_ratio": 4.35162905343228,
      "eval_steps_per_second": 0.816,
      "step": 1000
    },
    {
      "epoch": 0.024638265466111187,
      "grad_norm": 35.38695526123047,
      "learning_rate": 9.987570824628759e-05,
      "loss": 0.4655,
      "step": 1100
    },
    {
      "epoch": 0.026878107781212204,
      "grad_norm": 32.071346282958984,
      "learning_rate": 9.984962006889084e-05,
      "loss": 0.4342,
      "step": 1200
    },
    {
      "epoch": 0.029117950096313218,
      "grad_norm": 39.6610221862793,
      "learning_rate": 9.982105251561082e-05,
      "loss": 0.458,
      "step": 1300
    },
    {
      "epoch": 0.03135779241141424,
      "grad_norm": 31.493322372436523,
      "learning_rate": 9.979000700731491e-05,
      "loss": 0.4525,
      "step": 1400
    },
    {
      "epoch": 0.03359763472651525,
      "grad_norm": 34.453399658203125,
      "learning_rate": 9.975648508811693e-05,
      "loss": 0.41,
      "step": 1500
    },
    {
      "epoch": 0.03583747704161627,
      "grad_norm": 33.990074157714844,
      "learning_rate": 9.972048842530012e-05,
      "loss": 0.4097,
      "step": 1600
    },
    {
      "epoch": 0.038077319356717286,
      "grad_norm": 30.44228172302246,
      "learning_rate": 9.968201880923439e-05,
      "loss": 0.4257,
      "step": 1700
    },
    {
      "epoch": 0.040317161671818307,
      "grad_norm": 31.427162170410156,
      "learning_rate": 9.964107815328711e-05,
      "loss": 0.3821,
      "step": 1800
    },
    {
      "epoch": 0.04255700398691932,
      "grad_norm": 25.877887725830078,
      "learning_rate": 9.959766849372808e-05,
      "loss": 0.3788,
      "step": 1900
    },
    {
      "epoch": 0.04479684630202034,
      "grad_norm": 25.36798095703125,
      "learning_rate": 9.955179198962817e-05,
      "loss": 0.3854,
      "step": 2000
    },
    {
      "epoch": 0.04479684630202034,
      "eval_avg_non_pair_similarity": 0.0016289287904792565,
      "eval_avg_pair_similarity": 0.0032495629731565715,
      "eval_loss": 0.4394480586051941,
      "eval_runtime": 19.5064,
      "eval_samples_per_second": 25.633,
      "eval_similarity_ratio": 1.9949079371360972,
      "eval_steps_per_second": 0.82,
      "step": 2000
    },
    {
      "epoch": 0.047036688617121354,
      "grad_norm": 24.537384033203125,
      "learning_rate": 9.950345092275198e-05,
      "loss": 0.3886,
      "step": 2100
    },
    {
      "epoch": 0.049276530932222375,
      "grad_norm": 26.4981632232666,
      "learning_rate": 9.945264769744431e-05,
      "loss": 0.3926,
      "step": 2200
    },
    {
      "epoch": 0.05151637324732339,
      "grad_norm": 40.585941314697266,
      "learning_rate": 9.939938484051063e-05,
      "loss": 0.3796,
      "step": 2300
    },
    {
      "epoch": 0.05375621556242441,
      "grad_norm": 26.374149322509766,
      "learning_rate": 9.934366500109132e-05,
      "loss": 0.3723,
      "step": 2400
    },
    {
      "epoch": 0.05599605787752542,
      "grad_norm": 31.54728126525879,
      "learning_rate": 9.928549095053001e-05,
      "loss": 0.3639,
      "step": 2500
    },
    {
      "epoch": 0.058235900192626436,
      "grad_norm": 26.00455093383789,
      "learning_rate": 9.922486558223567e-05,
      "loss": 0.3387,
      "step": 2600
    },
    {
      "epoch": 0.060475742507727456,
      "grad_norm": 23.684057235717773,
      "learning_rate": 9.916179191153873e-05,
      "loss": 0.3494,
      "step": 2700
    },
    {
      "epoch": 0.06271558482282848,
      "grad_norm": 24.034656524658203,
      "learning_rate": 9.909627307554108e-05,
      "loss": 0.362,
      "step": 2800
    },
    {
      "epoch": 0.06495542713792948,
      "grad_norm": 31.07891082763672,
      "learning_rate": 9.902831233296009e-05,
      "loss": 0.3481,
      "step": 2900
    },
    {
      "epoch": 0.0671952694530305,
      "grad_norm": 24.183626174926758,
      "learning_rate": 9.895791306396644e-05,
      "loss": 0.3535,
      "step": 3000
    },
    {
      "epoch": 0.0671952694530305,
      "eval_avg_non_pair_similarity": -0.0007878901720614375,
      "eval_avg_pair_similarity": -0.008789425778668375,
      "eval_loss": 0.4187028110027313,
      "eval_runtime": 19.4839,
      "eval_samples_per_second": 25.662,
      "eval_similarity_ratio": 11.155648452463497,
      "eval_steps_per_second": 0.821,
      "step": 3000
    },
    {
      "epoch": 0.06943511176813152,
      "grad_norm": 20.178213119506836,
      "learning_rate": 9.888507877001616e-05,
      "loss": 0.3375,
      "step": 3100
    },
    {
      "epoch": 0.07167495408323254,
      "grad_norm": 31.757244110107422,
      "learning_rate": 9.880981307367627e-05,
      "loss": 0.3489,
      "step": 3200
    },
    {
      "epoch": 0.07391479639833355,
      "grad_norm": 27.889982223510742,
      "learning_rate": 9.873211971844477e-05,
      "loss": 0.3463,
      "step": 3300
    },
    {
      "epoch": 0.07615463871343457,
      "grad_norm": 23.282093048095703,
      "learning_rate": 9.865200256856437e-05,
      "loss": 0.3235,
      "step": 3400
    },
    {
      "epoch": 0.07839448102853559,
      "grad_norm": 24.573863983154297,
      "learning_rate": 9.856946560883034e-05,
      "loss": 0.3474,
      "step": 3500
    },
    {
      "epoch": 0.08063432334363661,
      "grad_norm": 20.7421932220459,
      "learning_rate": 9.848451294439224e-05,
      "loss": 0.3533,
      "step": 3600
    },
    {
      "epoch": 0.08287416565873762,
      "grad_norm": 22.051513671875,
      "learning_rate": 9.839714880054987e-05,
      "loss": 0.3277,
      "step": 3700
    },
    {
      "epoch": 0.08511400797383864,
      "grad_norm": 22.470027923583984,
      "learning_rate": 9.830737752254294e-05,
      "loss": 0.3261,
      "step": 3800
    },
    {
      "epoch": 0.08735385028893966,
      "grad_norm": 20.80890464782715,
      "learning_rate": 9.821520357533513e-05,
      "loss": 0.3474,
      "step": 3900
    },
    {
      "epoch": 0.08959369260404068,
      "grad_norm": 20.187280654907227,
      "learning_rate": 9.812063154339191e-05,
      "loss": 0.2984,
      "step": 4000
    },
    {
      "epoch": 0.08959369260404068,
      "eval_avg_non_pair_similarity": -0.0008396646829899625,
      "eval_avg_pair_similarity": -0.011315496074035763,
      "eval_loss": 0.32538020610809326,
      "eval_runtime": 19.4916,
      "eval_samples_per_second": 25.652,
      "eval_similarity_ratio": 13.476208185560939,
      "eval_steps_per_second": 0.821,
      "step": 4000
    },
    {
      "epoch": 0.09183353491914169,
      "grad_norm": 26.496036529541016,
      "learning_rate": 9.802366613045254e-05,
      "loss": 0.3326,
      "step": 4100
    },
    {
      "epoch": 0.09407337723424271,
      "grad_norm": 23.102359771728516,
      "learning_rate": 9.792431215929613e-05,
      "loss": 0.3341,
      "step": 4200
    },
    {
      "epoch": 0.09631321954934373,
      "grad_norm": 21.27369499206543,
      "learning_rate": 9.782257457150177e-05,
      "loss": 0.3254,
      "step": 4300
    },
    {
      "epoch": 0.09855306186444475,
      "grad_norm": 19.25406265258789,
      "learning_rate": 9.771845842720274e-05,
      "loss": 0.3334,
      "step": 4400
    },
    {
      "epoch": 0.10079290417954576,
      "grad_norm": 15.666335105895996,
      "learning_rate": 9.761196890483482e-05,
      "loss": 0.3064,
      "step": 4500
    },
    {
      "epoch": 0.10303274649464678,
      "grad_norm": 24.592592239379883,
      "learning_rate": 9.75031113008788e-05,
      "loss": 0.2902,
      "step": 4600
    },
    {
      "epoch": 0.1052725888097478,
      "grad_norm": 20.0572566986084,
      "learning_rate": 9.739189102959696e-05,
      "loss": 0.3121,
      "step": 4700
    },
    {
      "epoch": 0.10751243112484882,
      "grad_norm": 19.887725830078125,
      "learning_rate": 9.727831362276381e-05,
      "loss": 0.3014,
      "step": 4800
    },
    {
      "epoch": 0.10975227343994982,
      "grad_norm": 15.561097145080566,
      "learning_rate": 9.716238472939101e-05,
      "loss": 0.316,
      "step": 4900
    },
    {
      "epoch": 0.11199211575505084,
      "grad_norm": 17.96786880493164,
      "learning_rate": 9.704411011544629e-05,
      "loss": 0.3215,
      "step": 5000
    },
    {
      "epoch": 0.11199211575505084,
      "eval_avg_non_pair_similarity": 0.00024146916974524413,
      "eval_avg_pair_similarity": -0.009633154251612723,
      "eval_loss": 0.2951599061489105,
      "eval_runtime": 19.5025,
      "eval_samples_per_second": 25.638,
      "eval_similarity_ratio": -39.893930400207765,
      "eval_steps_per_second": 0.82,
      "step": 5000
    },
    {
      "epoch": 0.11423195807015186,
      "grad_norm": 14.672723770141602,
      "learning_rate": 9.692349566356677e-05,
      "loss": 0.2954,
      "step": 5100
    },
    {
      "epoch": 0.11647180038525287,
      "grad_norm": 17.881025314331055,
      "learning_rate": 9.680054737276638e-05,
      "loss": 0.2968,
      "step": 5200
    },
    {
      "epoch": 0.11871164270035389,
      "grad_norm": 16.55130958557129,
      "learning_rate": 9.667527135813737e-05,
      "loss": 0.2707,
      "step": 5300
    },
    {
      "epoch": 0.12095148501545491,
      "grad_norm": 21.650575637817383,
      "learning_rate": 9.654767385054627e-05,
      "loss": 0.3068,
      "step": 5400
    },
    {
      "epoch": 0.12319132733055593,
      "grad_norm": 20.990921020507812,
      "learning_rate": 9.641776119632397e-05,
      "loss": 0.3037,
      "step": 5500
    },
    {
      "epoch": 0.12543116964565695,
      "grad_norm": 21.97833824157715,
      "learning_rate": 9.628553985695005e-05,
      "loss": 0.3307,
      "step": 5600
    },
    {
      "epoch": 0.12767101196075797,
      "grad_norm": 20.353246688842773,
      "learning_rate": 9.61510164087314e-05,
      "loss": 0.2907,
      "step": 5700
    },
    {
      "epoch": 0.12991085427585897,
      "grad_norm": 16.254976272583008,
      "learning_rate": 9.601419754247514e-05,
      "loss": 0.3025,
      "step": 5800
    },
    {
      "epoch": 0.13215069659096,
      "grad_norm": 24.14662742614746,
      "learning_rate": 9.587509006315585e-05,
      "loss": 0.2676,
      "step": 5900
    },
    {
      "epoch": 0.134390538906061,
      "grad_norm": 15.861074447631836,
      "learning_rate": 9.573370088957712e-05,
      "loss": 0.2953,
      "step": 6000
    },
    {
      "epoch": 0.134390538906061,
      "eval_avg_non_pair_similarity": -0.00021786548842126337,
      "eval_avg_pair_similarity": -0.001965825233142823,
      "eval_loss": 0.31484127044677734,
      "eval_runtime": 19.5134,
      "eval_samples_per_second": 25.623,
      "eval_similarity_ratio": 9.023114433533939,
      "eval_steps_per_second": 0.82,
      "step": 6000
    },
    {
      "epoch": 0.13663038122116203,
      "grad_norm": 14.910375595092773,
      "learning_rate": 9.559003705402737e-05,
      "loss": 0.2846,
      "step": 6100
    },
    {
      "epoch": 0.13887022353626305,
      "grad_norm": 17.768138885498047,
      "learning_rate": 9.544410570193014e-05,
      "loss": 0.3031,
      "step": 6200
    },
    {
      "epoch": 0.14111006585136407,
      "grad_norm": 13.620220184326172,
      "learning_rate": 9.529591409148874e-05,
      "loss": 0.2716,
      "step": 6300
    },
    {
      "epoch": 0.1433499081664651,
      "grad_norm": 14.697577476501465,
      "learning_rate": 9.514546959332509e-05,
      "loss": 0.2759,
      "step": 6400
    },
    {
      "epoch": 0.1455897504815661,
      "grad_norm": 21.234310150146484,
      "learning_rate": 9.499277969011334e-05,
      "loss": 0.2845,
      "step": 6500
    },
    {
      "epoch": 0.1478295927966671,
      "grad_norm": 16.50038719177246,
      "learning_rate": 9.483785197620747e-05,
      "loss": 0.2471,
      "step": 6600
    },
    {
      "epoch": 0.15006943511176812,
      "grad_norm": 16.261537551879883,
      "learning_rate": 9.468069415726377e-05,
      "loss": 0.2978,
      "step": 6700
    },
    {
      "epoch": 0.15230927742686914,
      "grad_norm": 16.553571701049805,
      "learning_rate": 9.452131404985752e-05,
      "loss": 0.311,
      "step": 6800
    },
    {
      "epoch": 0.15454911974197016,
      "grad_norm": 18.445993423461914,
      "learning_rate": 9.43597195810941e-05,
      "loss": 0.2777,
      "step": 6900
    },
    {
      "epoch": 0.15678896205707119,
      "grad_norm": 13.675585746765137,
      "learning_rate": 9.419591878821496e-05,
      "loss": 0.2417,
      "step": 7000
    },
    {
      "epoch": 0.15678896205707119,
      "eval_avg_non_pair_similarity": 0.0003613333898595999,
      "eval_avg_pair_similarity": 0.005676110625499859,
      "eval_loss": 0.3019685447216034,
      "eval_runtime": 19.5219,
      "eval_samples_per_second": 25.612,
      "eval_similarity_ratio": 15.708790786551377,
      "eval_steps_per_second": 0.82,
      "step": 7000
    },
    {
      "epoch": 0.1590288043721722,
      "grad_norm": 15.780075073242188,
      "learning_rate": 9.402991981819758e-05,
      "loss": 0.2866,
      "step": 7100
    },
    {
      "epoch": 0.16126864668727323,
      "grad_norm": 17.34490966796875,
      "learning_rate": 9.386173092735051e-05,
      "loss": 0.2526,
      "step": 7200
    },
    {
      "epoch": 0.16350848900237425,
      "grad_norm": 14.431266784667969,
      "learning_rate": 9.36913604809026e-05,
      "loss": 0.2646,
      "step": 7300
    },
    {
      "epoch": 0.16574833131747524,
      "grad_norm": 11.20712661743164,
      "learning_rate": 9.351881695258693e-05,
      "loss": 0.2279,
      "step": 7400
    },
    {
      "epoch": 0.16798817363257626,
      "grad_norm": 15.549891471862793,
      "learning_rate": 9.334410892421945e-05,
      "loss": 0.2884,
      "step": 7500
    },
    {
      "epoch": 0.17022801594767728,
      "grad_norm": 14.635927200317383,
      "learning_rate": 9.316724508527205e-05,
      "loss": 0.2729,
      "step": 7600
    },
    {
      "epoch": 0.1724678582627783,
      "grad_norm": 12.841429710388184,
      "learning_rate": 9.298823423244038e-05,
      "loss": 0.271,
      "step": 7700
    },
    {
      "epoch": 0.17470770057787932,
      "grad_norm": 13.053755760192871,
      "learning_rate": 9.280708526920636e-05,
      "loss": 0.3094,
      "step": 7800
    },
    {
      "epoch": 0.17694754289298034,
      "grad_norm": 14.694463729858398,
      "learning_rate": 9.262380720539536e-05,
      "loss": 0.2687,
      "step": 7900
    },
    {
      "epoch": 0.17918738520808136,
      "grad_norm": 14.9842529296875,
      "learning_rate": 9.243840915672804e-05,
      "loss": 0.2539,
      "step": 8000
    },
    {
      "epoch": 0.17918738520808136,
      "eval_avg_non_pair_similarity": -0.0009227607365296818,
      "eval_avg_pair_similarity": -0.005276160409208387,
      "eval_loss": 0.22878731787204742,
      "eval_runtime": 19.5192,
      "eval_samples_per_second": 25.616,
      "eval_similarity_ratio": 5.717798992023619,
      "eval_steps_per_second": 0.82,
      "step": 8000
    },
    {
      "epoch": 0.18142722752318236,
      "grad_norm": 11.896175384521484,
      "learning_rate": 9.225090034436697e-05,
      "loss": 0.2739,
      "step": 8100
    },
    {
      "epoch": 0.18366706983828338,
      "grad_norm": 12.229938507080078,
      "learning_rate": 9.206129009445796e-05,
      "loss": 0.2467,
      "step": 8200
    },
    {
      "epoch": 0.1859069121533844,
      "grad_norm": 15.017464637756348,
      "learning_rate": 9.186958783766633e-05,
      "loss": 0.2668,
      "step": 8300
    },
    {
      "epoch": 0.18814675446848542,
      "grad_norm": 15.509882926940918,
      "learning_rate": 9.167580310870769e-05,
      "loss": 0.2614,
      "step": 8400
    },
    {
      "epoch": 0.19038659678358644,
      "grad_norm": 14.945472717285156,
      "learning_rate": 9.147994554587385e-05,
      "loss": 0.2266,
      "step": 8500
    },
    {
      "epoch": 0.19262643909868746,
      "grad_norm": 16.0211238861084,
      "learning_rate": 9.128202489055335e-05,
      "loss": 0.2428,
      "step": 8600
    },
    {
      "epoch": 0.19486628141378848,
      "grad_norm": 16.380727767944336,
      "learning_rate": 9.108205098674698e-05,
      "loss": 0.2578,
      "step": 8700
    },
    {
      "epoch": 0.1971061237288895,
      "grad_norm": 12.938665390014648,
      "learning_rate": 9.088003378057816e-05,
      "loss": 0.2445,
      "step": 8800
    },
    {
      "epoch": 0.1993459660439905,
      "grad_norm": 7.832601070404053,
      "learning_rate": 9.067598331979829e-05,
      "loss": 0.2284,
      "step": 8900
    },
    {
      "epoch": 0.2015858083590915,
      "grad_norm": 12.243515014648438,
      "learning_rate": 9.046990975328694e-05,
      "loss": 0.271,
      "step": 9000
    },
    {
      "epoch": 0.2015858083590915,
      "eval_avg_non_pair_similarity": -0.0005964781935742025,
      "eval_avg_pair_similarity": -0.0012636589542962611,
      "eval_loss": 0.24378302693367004,
      "eval_runtime": 19.5294,
      "eval_samples_per_second": 25.602,
      "eval_similarity_ratio": 2.1185333645211637,
      "eval_steps_per_second": 0.819,
      "step": 9000
    },
    {
      "epoch": 0.20382565067419253,
      "grad_norm": 14.057201385498047,
      "learning_rate": 9.026182333054714e-05,
      "loss": 0.2216,
      "step": 9100
    },
    {
      "epoch": 0.20606549298929355,
      "grad_norm": 9.724617958068848,
      "learning_rate": 9.005173440119555e-05,
      "loss": 0.2625,
      "step": 9200
    },
    {
      "epoch": 0.20830533530439457,
      "grad_norm": 13.209918975830078,
      "learning_rate": 8.983965341444769e-05,
      "loss": 0.2508,
      "step": 9300
    },
    {
      "epoch": 0.2105451776194956,
      "grad_norm": 12.0099458694458,
      "learning_rate": 8.962559091859828e-05,
      "loss": 0.2263,
      "step": 9400
    },
    {
      "epoch": 0.21278501993459661,
      "grad_norm": 11.473740577697754,
      "learning_rate": 8.940955756049658e-05,
      "loss": 0.256,
      "step": 9500
    },
    {
      "epoch": 0.21502486224969763,
      "grad_norm": 13.63650131225586,
      "learning_rate": 8.919156408501678e-05,
      "loss": 0.2594,
      "step": 9600
    },
    {
      "epoch": 0.21726470456479863,
      "grad_norm": 11.578638076782227,
      "learning_rate": 8.897162133452375e-05,
      "loss": 0.225,
      "step": 9700
    },
    {
      "epoch": 0.21950454687989965,
      "grad_norm": 8.985605239868164,
      "learning_rate": 8.874974024833351e-05,
      "loss": 0.2525,
      "step": 9800
    },
    {
      "epoch": 0.22174438919500067,
      "grad_norm": 15.842543601989746,
      "learning_rate": 8.852593186216942e-05,
      "loss": 0.239,
      "step": 9900
    },
    {
      "epoch": 0.2239842315101017,
      "grad_norm": 13.797926902770996,
      "learning_rate": 8.830020730761308e-05,
      "loss": 0.2133,
      "step": 10000
    },
    {
      "epoch": 0.2239842315101017,
      "eval_avg_non_pair_similarity": 3.824399728542623e-05,
      "eval_avg_pair_similarity": -0.010851175129879266,
      "eval_loss": 0.2623503506183624,
      "eval_runtime": 19.4973,
      "eval_samples_per_second": 25.645,
      "eval_similarity_ratio": -283.7353807159264,
      "eval_steps_per_second": 0.821,
      "step": 10000
    },
    {
      "epoch": 0.2262240738252027,
      "grad_norm": 13.315906524658203,
      "learning_rate": 8.807257781155081e-05,
      "loss": 0.2192,
      "step": 10100
    },
    {
      "epoch": 0.22846391614030373,
      "grad_norm": 11.403688430786133,
      "learning_rate": 8.784305469561519e-05,
      "loss": 0.2359,
      "step": 10200
    },
    {
      "epoch": 0.23070375845540475,
      "grad_norm": 11.676390647888184,
      "learning_rate": 8.761164937562199e-05,
      "loss": 0.224,
      "step": 10300
    },
    {
      "epoch": 0.23294360077050574,
      "grad_norm": 11.526522636413574,
      "learning_rate": 8.737837336100233e-05,
      "loss": 0.2288,
      "step": 10400
    },
    {
      "epoch": 0.23518344308560676,
      "grad_norm": 15.950361251831055,
      "learning_rate": 8.714323825423031e-05,
      "loss": 0.2388,
      "step": 10500
    },
    {
      "epoch": 0.23742328540070778,
      "grad_norm": 11.149569511413574,
      "learning_rate": 8.690625575024586e-05,
      "loss": 0.2067,
      "step": 10600
    },
    {
      "epoch": 0.2396631277158088,
      "grad_norm": 10.904093742370605,
      "learning_rate": 8.666743763587315e-05,
      "loss": 0.2354,
      "step": 10700
    },
    {
      "epoch": 0.24190297003090983,
      "grad_norm": 10.213909149169922,
      "learning_rate": 8.642679578923422e-05,
      "loss": 0.2236,
      "step": 10800
    },
    {
      "epoch": 0.24414281234601085,
      "grad_norm": 10.054357528686523,
      "learning_rate": 8.618434217915838e-05,
      "loss": 0.1987,
      "step": 10900
    },
    {
      "epoch": 0.24638265466111187,
      "grad_norm": 11.040483474731445,
      "learning_rate": 8.59400888645867e-05,
      "loss": 0.2471,
      "step": 11000
    },
    {
      "epoch": 0.24638265466111187,
      "eval_avg_non_pair_similarity": 0.00023486310084357872,
      "eval_avg_pair_similarity": 0.00031724373530596497,
      "eval_loss": 0.17524191737174988,
      "eval_runtime": 19.4456,
      "eval_samples_per_second": 25.713,
      "eval_similarity_ratio": 1.3507602265596101,
      "eval_steps_per_second": 0.823,
      "step": 11000
    },
    {
      "epoch": 0.2486224969762129,
      "grad_norm": 12.810904502868652,
      "learning_rate": 8.569404799397247e-05,
      "loss": 0.1974,
      "step": 11100
    },
    {
      "epoch": 0.2508623392913139,
      "grad_norm": 15.360380172729492,
      "learning_rate": 8.544623180467673e-05,
      "loss": 0.2204,
      "step": 11200
    },
    {
      "epoch": 0.2531021816064149,
      "grad_norm": 13.429972648620605,
      "learning_rate": 8.519665262235979e-05,
      "loss": 0.2479,
      "step": 11300
    },
    {
      "epoch": 0.25534202392151595,
      "grad_norm": 10.252517700195312,
      "learning_rate": 8.494532286036813e-05,
      "loss": 0.1886,
      "step": 11400
    },
    {
      "epoch": 0.25758186623661694,
      "grad_norm": 11.51858901977539,
      "learning_rate": 8.4692255019117e-05,
      "loss": 0.2191,
      "step": 11500
    },
    {
      "epoch": 0.25982170855171793,
      "grad_norm": 8.529491424560547,
      "learning_rate": 8.443746168546867e-05,
      "loss": 0.1856,
      "step": 11600
    },
    {
      "epoch": 0.262061550866819,
      "grad_norm": 7.846275806427002,
      "learning_rate": 8.418095553210641e-05,
      "loss": 0.1879,
      "step": 11700
    },
    {
      "epoch": 0.26430139318192,
      "grad_norm": 10.658947944641113,
      "learning_rate": 8.392274931690422e-05,
      "loss": 0.1998,
      "step": 11800
    },
    {
      "epoch": 0.266541235497021,
      "grad_norm": 8.623246192932129,
      "learning_rate": 8.366285588229223e-05,
      "loss": 0.2073,
      "step": 11900
    },
    {
      "epoch": 0.268781077812122,
      "grad_norm": 8.197981834411621,
      "learning_rate": 8.340128815461797e-05,
      "loss": 0.179,
      "step": 12000
    },
    {
      "epoch": 0.268781077812122,
      "eval_avg_non_pair_similarity": 5.694422117604282e-05,
      "eval_avg_pair_similarity": -0.00526782719604671,
      "eval_loss": 0.19077710807323456,
      "eval_runtime": 19.459,
      "eval_samples_per_second": 25.695,
      "eval_similarity_ratio": -92.50854761471308,
      "eval_steps_per_second": 0.822,
      "step": 12000
    },
    {
      "epoch": 0.27102092012722306,
      "grad_norm": 9.695572853088379,
      "learning_rate": 8.313805914350349e-05,
      "loss": 0.2022,
      "step": 12100
    },
    {
      "epoch": 0.27326076244232406,
      "grad_norm": 8.658490180969238,
      "learning_rate": 8.287318194119825e-05,
      "loss": 0.2011,
      "step": 12200
    },
    {
      "epoch": 0.2755006047574251,
      "grad_norm": 9.974045753479004,
      "learning_rate": 8.260666972192798e-05,
      "loss": 0.2163,
      "step": 12300
    },
    {
      "epoch": 0.2777404470725261,
      "grad_norm": 8.445626258850098,
      "learning_rate": 8.233853574123944e-05,
      "loss": 0.1947,
      "step": 12400
    },
    {
      "epoch": 0.2799802893876271,
      "grad_norm": 7.46968412399292,
      "learning_rate": 8.206879333534109e-05,
      "loss": 0.1709,
      "step": 12500
    },
    {
      "epoch": 0.28222013170272814,
      "grad_norm": 12.895581245422363,
      "learning_rate": 8.179745592043982e-05,
      "loss": 0.2281,
      "step": 12600
    },
    {
      "epoch": 0.28445997401782913,
      "grad_norm": 10.54282283782959,
      "learning_rate": 8.152453699207362e-05,
      "loss": 0.184,
      "step": 12700
    },
    {
      "epoch": 0.2866998163329302,
      "grad_norm": 7.9730634689331055,
      "learning_rate": 8.125005012444044e-05,
      "loss": 0.2525,
      "step": 12800
    },
    {
      "epoch": 0.2889396586480312,
      "grad_norm": 9.222777366638184,
      "learning_rate": 8.097400896972296e-05,
      "loss": 0.1912,
      "step": 12900
    },
    {
      "epoch": 0.2911795009631322,
      "grad_norm": 10.500025749206543,
      "learning_rate": 8.069642725740957e-05,
      "loss": 0.1948,
      "step": 13000
    },
    {
      "epoch": 0.2911795009631322,
      "eval_avg_non_pair_similarity": 0.0007876502425145761,
      "eval_avg_pair_similarity": -0.007292054696008563,
      "eval_loss": 0.18571369349956512,
      "eval_runtime": 19.4919,
      "eval_samples_per_second": 25.652,
      "eval_similarity_ratio": -9.257985718037302,
      "eval_steps_per_second": 0.821,
      "step": 13000
    },
    {
      "epoch": 0.2934193432782332,
      "grad_norm": 7.784255504608154,
      "learning_rate": 8.041731879361164e-05,
      "loss": 0.214,
      "step": 13100
    },
    {
      "epoch": 0.2956591855933342,
      "grad_norm": 9.356244087219238,
      "learning_rate": 8.013669746037662e-05,
      "loss": 0.168,
      "step": 13200
    },
    {
      "epoch": 0.29789902790843525,
      "grad_norm": 8.42744255065918,
      "learning_rate": 7.985457721499779e-05,
      "loss": 0.1759,
      "step": 13300
    },
    {
      "epoch": 0.30013887022353625,
      "grad_norm": 8.323335647583008,
      "learning_rate": 7.957097208931997e-05,
      "loss": 0.193,
      "step": 13400
    },
    {
      "epoch": 0.3023787125386373,
      "grad_norm": 10.35467529296875,
      "learning_rate": 7.928589618904158e-05,
      "loss": 0.244,
      "step": 13500
    },
    {
      "epoch": 0.3046185548537383,
      "grad_norm": 9.027731895446777,
      "learning_rate": 7.899936369301319e-05,
      "loss": 0.1878,
      "step": 13600
    },
    {
      "epoch": 0.30685839716883934,
      "grad_norm": 11.577035903930664,
      "learning_rate": 7.871138885253216e-05,
      "loss": 0.2033,
      "step": 13700
    },
    {
      "epoch": 0.30909823948394033,
      "grad_norm": 6.524970054626465,
      "learning_rate": 7.842198599063395e-05,
      "loss": 0.1792,
      "step": 13800
    },
    {
      "epoch": 0.3113380817990413,
      "grad_norm": 8.987520217895508,
      "learning_rate": 7.813116950137964e-05,
      "loss": 0.1936,
      "step": 13900
    },
    {
      "epoch": 0.31357792411414237,
      "grad_norm": 8.995404243469238,
      "learning_rate": 7.783895384914004e-05,
      "loss": 0.1831,
      "step": 14000
    },
    {
      "epoch": 0.31357792411414237,
      "eval_avg_non_pair_similarity": 0.000529102375272793,
      "eval_avg_pair_similarity": -0.001620051197707653,
      "eval_loss": 0.18957822024822235,
      "eval_runtime": 19.4615,
      "eval_samples_per_second": 25.692,
      "eval_similarity_ratio": -3.061886079933759,
      "eval_steps_per_second": 0.822,
      "step": 14000
    },
    {
      "epoch": 0.31581776642924336,
      "grad_norm": 6.4864068031311035,
      "learning_rate": 7.754535356787629e-05,
      "loss": 0.1976,
      "step": 14100
    },
    {
      "epoch": 0.3180576087443444,
      "grad_norm": 7.537520885467529,
      "learning_rate": 7.725038326041698e-05,
      "loss": 0.1935,
      "step": 14200
    },
    {
      "epoch": 0.3202974510594454,
      "grad_norm": 7.547581195831299,
      "learning_rate": 7.695405759773182e-05,
      "loss": 0.1706,
      "step": 14300
    },
    {
      "epoch": 0.32253729337454645,
      "grad_norm": 6.900126934051514,
      "learning_rate": 7.665639131820198e-05,
      "loss": 0.1479,
      "step": 14400
    },
    {
      "epoch": 0.32477713568964744,
      "grad_norm": 9.628849983215332,
      "learning_rate": 7.635739922688702e-05,
      "loss": 0.1649,
      "step": 14500
    },
    {
      "epoch": 0.3270169780047485,
      "grad_norm": 9.103193283081055,
      "learning_rate": 7.60570961947886e-05,
      "loss": 0.1672,
      "step": 14600
    },
    {
      "epoch": 0.3292568203198495,
      "grad_norm": 9.142544746398926,
      "learning_rate": 7.57554971581107e-05,
      "loss": 0.1831,
      "step": 14700
    },
    {
      "epoch": 0.3314966626349505,
      "grad_norm": 8.968326568603516,
      "learning_rate": 7.545261711751685e-05,
      "loss": 0.1566,
      "step": 14800
    },
    {
      "epoch": 0.3337365049500515,
      "grad_norm": 9.484015464782715,
      "learning_rate": 7.514847113738401e-05,
      "loss": 0.1949,
      "step": 14900
    },
    {
      "epoch": 0.3359763472651525,
      "grad_norm": 7.016854763031006,
      "learning_rate": 7.484307434505335e-05,
      "loss": 0.1419,
      "step": 15000
    },
    {
      "epoch": 0.3359763472651525,
      "eval_avg_non_pair_similarity": 8.844231431561901e-05,
      "eval_avg_pair_similarity": -0.0006939181396737695,
      "eval_loss": 0.1548442244529724,
      "eval_runtime": 19.513,
      "eval_samples_per_second": 25.624,
      "eval_similarity_ratio": -7.845997077794952,
      "eval_steps_per_second": 0.82,
      "step": 15000
    },
    {
      "epoch": 0.33821618958025357,
      "grad_norm": 7.563048362731934,
      "learning_rate": 7.453644193007774e-05,
      "loss": 0.1891,
      "step": 15100
    },
    {
      "epoch": 0.34045603189535456,
      "grad_norm": 7.630728244781494,
      "learning_rate": 7.422858914346641e-05,
      "loss": 0.1896,
      "step": 15200
    },
    {
      "epoch": 0.3426958742104556,
      "grad_norm": 7.264196395874023,
      "learning_rate": 7.391953129692633e-05,
      "loss": 0.1608,
      "step": 15300
    },
    {
      "epoch": 0.3449357165255566,
      "grad_norm": 10.433218955993652,
      "learning_rate": 7.360928376210064e-05,
      "loss": 0.1668,
      "step": 15400
    },
    {
      "epoch": 0.3471755588406576,
      "grad_norm": 7.074268817901611,
      "learning_rate": 7.329786196980417e-05,
      "loss": 0.1882,
      "step": 15500
    },
    {
      "epoch": 0.34941540115575864,
      "grad_norm": 8.925777435302734,
      "learning_rate": 7.298528140925589e-05,
      "loss": 0.1697,
      "step": 15600
    },
    {
      "epoch": 0.35165524347085964,
      "grad_norm": 9.438268661499023,
      "learning_rate": 7.267155762730855e-05,
      "loss": 0.1566,
      "step": 15700
    },
    {
      "epoch": 0.3538950857859607,
      "grad_norm": 8.304781913757324,
      "learning_rate": 7.235670622767546e-05,
      "loss": 0.1624,
      "step": 15800
    },
    {
      "epoch": 0.3561349281010617,
      "grad_norm": 6.194468975067139,
      "learning_rate": 7.204074287015433e-05,
      "loss": 0.1423,
      "step": 15900
    },
    {
      "epoch": 0.3583747704161627,
      "grad_norm": 10.31971549987793,
      "learning_rate": 7.172368326984846e-05,
      "loss": 0.15,
      "step": 16000
    },
    {
      "epoch": 0.3583747704161627,
      "eval_avg_non_pair_similarity": 0.0005423984809662916,
      "eval_avg_pair_similarity": -0.004194568347651512,
      "eval_loss": 0.16588900983333588,
      "eval_runtime": 19.5058,
      "eval_samples_per_second": 25.633,
      "eval_similarity_ratio": -7.7333703814561225,
      "eval_steps_per_second": 0.82,
      "step": 16000
    },
    {
      "epoch": 0.3606146127312637,
      "grad_norm": 6.6040449142456055,
      "learning_rate": 7.14055431963851e-05,
      "loss": 0.1929,
      "step": 16100
    },
    {
      "epoch": 0.3628544550463647,
      "grad_norm": 6.839012622833252,
      "learning_rate": 7.108633847313109e-05,
      "loss": 0.169,
      "step": 16200
    },
    {
      "epoch": 0.36509429736146576,
      "grad_norm": 8.55441665649414,
      "learning_rate": 7.076608497640588e-05,
      "loss": 0.1772,
      "step": 16300
    },
    {
      "epoch": 0.36733413967656675,
      "grad_norm": 7.641045093536377,
      "learning_rate": 7.044479863469189e-05,
      "loss": 0.1544,
      "step": 16400
    },
    {
      "epoch": 0.3695739819916678,
      "grad_norm": 9.20849895477295,
      "learning_rate": 7.012249542784223e-05,
      "loss": 0.1485,
      "step": 16500
    },
    {
      "epoch": 0.3718138243067688,
      "grad_norm": 6.943836212158203,
      "learning_rate": 6.979919138628598e-05,
      "loss": 0.1611,
      "step": 16600
    },
    {
      "epoch": 0.37405366662186984,
      "grad_norm": 8.36039924621582,
      "learning_rate": 6.94749025902308e-05,
      "loss": 0.1646,
      "step": 16700
    },
    {
      "epoch": 0.37629350893697083,
      "grad_norm": 7.000893592834473,
      "learning_rate": 6.914964516886324e-05,
      "loss": 0.1719,
      "step": 16800
    },
    {
      "epoch": 0.3785333512520719,
      "grad_norm": 8.178186416625977,
      "learning_rate": 6.882343529954641e-05,
      "loss": 0.1528,
      "step": 16900
    },
    {
      "epoch": 0.3807731935671729,
      "grad_norm": 7.016880512237549,
      "learning_rate": 6.849628920701549e-05,
      "loss": 0.1609,
      "step": 17000
    },
    {
      "epoch": 0.3807731935671729,
      "eval_avg_non_pair_similarity": -0.0001115500947464265,
      "eval_avg_pair_similarity": 0.00017968433955684303,
      "eval_loss": 0.13874633610248566,
      "eval_runtime": 19.4966,
      "eval_samples_per_second": 25.646,
      "eval_similarity_ratio": -1.6107950420418555,
      "eval_steps_per_second": 0.821,
      "step": 17000
    },
    {
      "epoch": 0.38301303588227387,
      "grad_norm": 6.156219005584717,
      "learning_rate": 6.816822316257066e-05,
      "loss": 0.1405,
      "step": 17100
    },
    {
      "epoch": 0.3852528781973749,
      "grad_norm": 6.131289482116699,
      "learning_rate": 6.783925348326782e-05,
      "loss": 0.167,
      "step": 17200
    },
    {
      "epoch": 0.3874927205124759,
      "grad_norm": 6.05889892578125,
      "learning_rate": 6.750939653110711e-05,
      "loss": 0.1819,
      "step": 17300
    },
    {
      "epoch": 0.38973256282757696,
      "grad_norm": 6.4408488273620605,
      "learning_rate": 6.717866871221901e-05,
      "loss": 0.1518,
      "step": 17400
    },
    {
      "epoch": 0.39197240514267795,
      "grad_norm": 7.979753494262695,
      "learning_rate": 6.684708647604843e-05,
      "loss": 0.1812,
      "step": 17500
    },
    {
      "epoch": 0.394212247457779,
      "grad_norm": 6.986364364624023,
      "learning_rate": 6.65146663145365e-05,
      "loss": 0.1447,
      "step": 17600
    },
    {
      "epoch": 0.39645208977288,
      "grad_norm": 6.019850730895996,
      "learning_rate": 6.618142476130032e-05,
      "loss": 0.1893,
      "step": 17700
    },
    {
      "epoch": 0.398691932087981,
      "grad_norm": 6.467818260192871,
      "learning_rate": 6.584737839081066e-05,
      "loss": 0.1751,
      "step": 17800
    },
    {
      "epoch": 0.40093177440308203,
      "grad_norm": 9.779342651367188,
      "learning_rate": 6.551254381756758e-05,
      "loss": 0.1524,
      "step": 17900
    },
    {
      "epoch": 0.403171616718183,
      "grad_norm": 6.93033504486084,
      "learning_rate": 6.517693769527402e-05,
      "loss": 0.1327,
      "step": 18000
    },
    {
      "epoch": 0.403171616718183,
      "eval_avg_non_pair_similarity": 0.0004680827924466593,
      "eval_avg_pair_similarity": 0.0004896340477280319,
      "eval_loss": 0.11614304035902023,
      "eval_runtime": 19.4704,
      "eval_samples_per_second": 25.68,
      "eval_similarity_ratio": 1.0460415457033243,
      "eval_steps_per_second": 0.822,
      "step": 18000
    },
    {
      "epoch": 0.40541145903328407,
      "grad_norm": 6.358455181121826,
      "learning_rate": 6.48405767160076e-05,
      "loss": 0.154,
      "step": 18100
    },
    {
      "epoch": 0.40765130134838506,
      "grad_norm": 7.965555191040039,
      "learning_rate": 6.450347760939031e-05,
      "loss": 0.1547,
      "step": 18200
    },
    {
      "epoch": 0.4098911436634861,
      "grad_norm": 8.082962036132812,
      "learning_rate": 6.416565714175642e-05,
      "loss": 0.1782,
      "step": 18300
    },
    {
      "epoch": 0.4121309859785871,
      "grad_norm": 6.913635730743408,
      "learning_rate": 6.382713211531868e-05,
      "loss": 0.1545,
      "step": 18400
    },
    {
      "epoch": 0.4143708282936881,
      "grad_norm": 7.295111656188965,
      "learning_rate": 6.348791936733252e-05,
      "loss": 0.1376,
      "step": 18500
    },
    {
      "epoch": 0.41661067060878915,
      "grad_norm": 7.109374523162842,
      "learning_rate": 6.314803576925865e-05,
      "loss": 0.1269,
      "step": 18600
    },
    {
      "epoch": 0.41885051292389014,
      "grad_norm": 6.2317328453063965,
      "learning_rate": 6.280749822592393e-05,
      "loss": 0.1545,
      "step": 18700
    },
    {
      "epoch": 0.4210903552389912,
      "grad_norm": 6.809806823730469,
      "learning_rate": 6.246632367468057e-05,
      "loss": 0.1367,
      "step": 18800
    },
    {
      "epoch": 0.4233301975540922,
      "grad_norm": 5.076844215393066,
      "learning_rate": 6.21245290845637e-05,
      "loss": 0.1353,
      "step": 18900
    },
    {
      "epoch": 0.42557003986919323,
      "grad_norm": 4.954884052276611,
      "learning_rate": 6.178213145544738e-05,
      "loss": 0.1507,
      "step": 19000
    },
    {
      "epoch": 0.42557003986919323,
      "eval_avg_non_pair_similarity": 0.0004066981645342978,
      "eval_avg_pair_similarity": -0.0005262972658965736,
      "eval_loss": 0.13725824654102325,
      "eval_runtime": 19.511,
      "eval_samples_per_second": 25.627,
      "eval_similarity_ratio": -1.2940733737985428,
      "eval_steps_per_second": 0.82,
      "step": 19000
    },
    {
      "epoch": 0.4278098821842942,
      "grad_norm": 7.648046493530273,
      "learning_rate": 6.143914781719913e-05,
      "loss": 0.1554,
      "step": 19100
    },
    {
      "epoch": 0.43004972449939527,
      "grad_norm": 6.501719951629639,
      "learning_rate": 6.10955952288328e-05,
      "loss": 0.1456,
      "step": 19200
    },
    {
      "epoch": 0.43228956681449626,
      "grad_norm": 6.05966854095459,
      "learning_rate": 6.0751490777660226e-05,
      "loss": 0.1348,
      "step": 19300
    },
    {
      "epoch": 0.43452940912959725,
      "grad_norm": 9.531146049499512,
      "learning_rate": 6.0406851578441305e-05,
      "loss": 0.1729,
      "step": 19400
    },
    {
      "epoch": 0.4367692514446983,
      "grad_norm": 4.878669261932373,
      "learning_rate": 6.0061694772532705e-05,
      "loss": 0.1265,
      "step": 19500
    },
    {
      "epoch": 0.4390090937597993,
      "grad_norm": 9.031988143920898,
      "learning_rate": 5.9716037527035416e-05,
      "loss": 0.1403,
      "step": 19600
    },
    {
      "epoch": 0.44124893607490034,
      "grad_norm": 10.188375473022461,
      "learning_rate": 5.93698970339408e-05,
      "loss": 0.1107,
      "step": 19700
    },
    {
      "epoch": 0.44348877839000134,
      "grad_norm": 5.289632320404053,
      "learning_rate": 5.902329050927557e-05,
      "loss": 0.1372,
      "step": 19800
    },
    {
      "epoch": 0.4457286207051024,
      "grad_norm": 6.766297340393066,
      "learning_rate": 5.867623519224551e-05,
      "loss": 0.1537,
      "step": 19900
    },
    {
      "epoch": 0.4479684630202034,
      "grad_norm": 6.518314361572266,
      "learning_rate": 5.8328748344378023e-05,
      "loss": 0.1797,
      "step": 20000
    },
    {
      "epoch": 0.4479684630202034,
      "eval_avg_non_pair_similarity": 1.2386919107971217e-05,
      "eval_avg_pair_similarity": -0.0007377016106620431,
      "eval_loss": 0.15548266470432281,
      "eval_runtime": 19.4908,
      "eval_samples_per_second": 25.653,
      "eval_similarity_ratio": -59.554890464031374,
      "eval_steps_per_second": 0.821,
      "step": 20000
    },
    {
      "epoch": 0.45020830533530437,
      "grad_norm": 7.09486722946167,
      "learning_rate": 5.7980847248663626e-05,
      "loss": 0.1493,
      "step": 20100
    },
    {
      "epoch": 0.4524481476504054,
      "grad_norm": 5.606503486633301,
      "learning_rate": 5.763254920869631e-05,
      "loss": 0.1242,
      "step": 20200
    },
    {
      "epoch": 0.4546879899655064,
      "grad_norm": 8.37496566772461,
      "learning_rate": 5.72838715478129e-05,
      "loss": 0.1732,
      "step": 20300
    },
    {
      "epoch": 0.45692783228060746,
      "grad_norm": 6.157011985778809,
      "learning_rate": 5.6934831608231544e-05,
      "loss": 0.1378,
      "step": 20400
    },
    {
      "epoch": 0.45916767459570845,
      "grad_norm": 6.028346061706543,
      "learning_rate": 5.658544675018901e-05,
      "loss": 0.1388,
      "step": 20500
    },
    {
      "epoch": 0.4614075169108095,
      "grad_norm": 5.829131603240967,
      "learning_rate": 5.623573435107735e-05,
      "loss": 0.1355,
      "step": 20600
    },
    {
      "epoch": 0.4636473592259105,
      "grad_norm": 4.870571136474609,
      "learning_rate": 5.5885711804579543e-05,
      "loss": 0.1424,
      "step": 20700
    },
    {
      "epoch": 0.4658872015410115,
      "grad_norm": 5.387896537780762,
      "learning_rate": 5.553539651980439e-05,
      "loss": 0.123,
      "step": 20800
    },
    {
      "epoch": 0.46812704385611253,
      "grad_norm": 6.630838871002197,
      "learning_rate": 5.51848059204207e-05,
      "loss": 0.1464,
      "step": 20900
    },
    {
      "epoch": 0.4703668861712135,
      "grad_norm": 6.7821526527404785,
      "learning_rate": 5.4833957443790586e-05,
      "loss": 0.1292,
      "step": 21000
    },
    {
      "epoch": 0.4703668861712135,
      "eval_avg_non_pair_similarity": 0.0003182698469606689,
      "eval_avg_pair_similarity": -0.00044986815843731163,
      "eval_loss": 0.12692449986934662,
      "eval_runtime": 19.486,
      "eval_samples_per_second": 25.659,
      "eval_similarity_ratio": -1.4134802989769413,
      "eval_steps_per_second": 0.821,
      "step": 21000
    },
    {
      "epoch": 0.4726067284863146,
      "grad_norm": 7.282751560211182,
      "learning_rate": 5.4482868540102236e-05,
      "loss": 0.1064,
      "step": 21100
    },
    {
      "epoch": 0.47484657080141557,
      "grad_norm": 6.863794326782227,
      "learning_rate": 5.4131556671502006e-05,
      "loss": 0.1167,
      "step": 21200
    },
    {
      "epoch": 0.4770864131165166,
      "grad_norm": 9.730511665344238,
      "learning_rate": 5.378003931122585e-05,
      "loss": 0.1936,
      "step": 21300
    },
    {
      "epoch": 0.4793262554316176,
      "grad_norm": 6.049126625061035,
      "learning_rate": 5.3428333942730304e-05,
      "loss": 0.1137,
      "step": 21400
    },
    {
      "epoch": 0.48156609774671866,
      "grad_norm": 6.33958625793457,
      "learning_rate": 5.3076458058822885e-05,
      "loss": 0.1135,
      "step": 21500
    },
    {
      "epoch": 0.48380594006181965,
      "grad_norm": 8.29135799407959,
      "learning_rate": 5.272442916079208e-05,
      "loss": 0.1355,
      "step": 21600
    },
    {
      "epoch": 0.48604578237692064,
      "grad_norm": 5.330431938171387,
      "learning_rate": 5.2372264757536796e-05,
      "loss": 0.1196,
      "step": 21700
    },
    {
      "epoch": 0.4882856246920217,
      "grad_norm": 4.634214401245117,
      "learning_rate": 5.2019982364695616e-05,
      "loss": 0.1437,
      "step": 21800
    },
    {
      "epoch": 0.4905254670071227,
      "grad_norm": 6.431977272033691,
      "learning_rate": 5.166759950377559e-05,
      "loss": 0.1234,
      "step": 21900
    },
    {
      "epoch": 0.49276530932222373,
      "grad_norm": 5.671669960021973,
      "learning_rate": 5.131513370128075e-05,
      "loss": 0.1102,
      "step": 22000
    },
    {
      "epoch": 0.49276530932222373,
      "eval_avg_non_pair_similarity": 7.926299255052003e-05,
      "eval_avg_pair_similarity": -0.0031928994464688,
      "eval_loss": 0.10579516738653183,
      "eval_runtime": 19.4888,
      "eval_samples_per_second": 25.656,
      "eval_similarity_ratio": -40.2823479624963,
      "eval_steps_per_second": 0.821,
      "step": 22000
    },
    {
      "epoch": 0.4950051516373247,
      "grad_norm": 5.917268753051758,
      "learning_rate": 5.09626024878404e-05,
      "loss": 0.0997,
      "step": 22100
    },
    {
      "epoch": 0.4972449939524258,
      "grad_norm": 6.4815993309021,
      "learning_rate": 5.0610023397337167e-05,
      "loss": 0.1541,
      "step": 22200
    },
    {
      "epoch": 0.49948483626752677,
      "grad_norm": 6.020129680633545,
      "learning_rate": 5.025741396603502e-05,
      "loss": 0.15,
      "step": 22300
    },
    {
      "epoch": 0.5017246785826278,
      "grad_norm": 5.328291416168213,
      "learning_rate": 4.990479173170691e-05,
      "loss": 0.1403,
      "step": 22400
    },
    {
      "epoch": 0.5039645208977288,
      "grad_norm": 4.73598051071167,
      "learning_rate": 4.955217423276261e-05,
      "loss": 0.1258,
      "step": 22500
    },
    {
      "epoch": 0.5062043632128298,
      "grad_norm": 5.456547737121582,
      "learning_rate": 4.91995790073764e-05,
      "loss": 0.1275,
      "step": 22600
    },
    {
      "epoch": 0.5084442055279308,
      "grad_norm": 6.9049153327941895,
      "learning_rate": 4.884702359261467e-05,
      "loss": 0.1477,
      "step": 22700
    },
    {
      "epoch": 0.5106840478430319,
      "grad_norm": 5.124217510223389,
      "learning_rate": 4.84945255235638e-05,
      "loss": 0.1157,
      "step": 22800
    },
    {
      "epoch": 0.5129238901581329,
      "grad_norm": 7.1598358154296875,
      "learning_rate": 4.814210233245791e-05,
      "loss": 0.1233,
      "step": 22900
    },
    {
      "epoch": 0.5151637324732339,
      "grad_norm": 6.744624137878418,
      "learning_rate": 4.778977154780698e-05,
      "loss": 0.1158,
      "step": 23000
    },
    {
      "epoch": 0.5151637324732339,
      "eval_avg_non_pair_similarity": 2.2767017136340985e-06,
      "eval_avg_pair_similarity": -0.0032549318633973597,
      "eval_loss": 0.13024543225765228,
      "eval_runtime": 19.7369,
      "eval_samples_per_second": 25.333,
      "eval_similarity_ratio": -1429.669878976723,
      "eval_steps_per_second": 0.811,
      "step": 23000
    },
    {
      "epoch": 0.5174035747883349,
      "grad_norm": 6.886088848114014,
      "learning_rate": 4.743755069352488e-05,
      "loss": 0.1127,
      "step": 23100
    },
    {
      "epoch": 0.5196434171034359,
      "grad_norm": 4.287947177886963,
      "learning_rate": 4.7085457288057925e-05,
      "loss": 0.1112,
      "step": 23200
    },
    {
      "epoch": 0.521883259418537,
      "grad_norm": 7.680822849273682,
      "learning_rate": 4.673350884351344e-05,
      "loss": 0.1437,
      "step": 23300
    },
    {
      "epoch": 0.524123101733638,
      "grad_norm": 9.121180534362793,
      "learning_rate": 4.6381722864788876e-05,
      "loss": 0.1441,
      "step": 23400
    },
    {
      "epoch": 0.526362944048739,
      "grad_norm": 9.648383140563965,
      "learning_rate": 4.603011684870104e-05,
      "loss": 0.1442,
      "step": 23500
    },
    {
      "epoch": 0.52860278636384,
      "grad_norm": 6.871197700500488,
      "learning_rate": 4.567870828311598e-05,
      "loss": 0.1007,
      "step": 23600
    },
    {
      "epoch": 0.530842628678941,
      "grad_norm": 6.548398971557617,
      "learning_rate": 4.5327514646079064e-05,
      "loss": 0.1391,
      "step": 23700
    },
    {
      "epoch": 0.533082470994042,
      "grad_norm": 4.919002532958984,
      "learning_rate": 4.4976553404945795e-05,
      "loss": 0.1413,
      "step": 23800
    },
    {
      "epoch": 0.535322313309143,
      "grad_norm": 5.23176383972168,
      "learning_rate": 4.462584201551291e-05,
      "loss": 0.1127,
      "step": 23900
    },
    {
      "epoch": 0.537562155624244,
      "grad_norm": 6.158422946929932,
      "learning_rate": 4.427539792115033e-05,
      "loss": 0.1297,
      "step": 24000
    },
    {
      "epoch": 0.537562155624244,
      "eval_avg_non_pair_similarity": 0.0003475586844268507,
      "eval_avg_pair_similarity": 0.0009371067539323121,
      "eval_loss": 0.12211039662361145,
      "eval_runtime": 19.4919,
      "eval_samples_per_second": 25.652,
      "eval_similarity_ratio": 2.6962547504104775,
      "eval_steps_per_second": 0.821,
      "step": 24000
    },
    {
      "epoch": 0.539801997939345,
      "grad_norm": 6.401062488555908,
      "learning_rate": 4.3925238551933406e-05,
      "loss": 0.1066,
      "step": 24100
    },
    {
      "epoch": 0.5420418402544461,
      "grad_norm": 6.289074420928955,
      "learning_rate": 4.357538132377615e-05,
      "loss": 0.1388,
      "step": 24200
    },
    {
      "epoch": 0.5442816825695471,
      "grad_norm": 6.771076679229736,
      "learning_rate": 4.32258436375649e-05,
      "loss": 0.123,
      "step": 24300
    },
    {
      "epoch": 0.5465215248846481,
      "grad_norm": 4.1949896812438965,
      "learning_rate": 4.2876642878292966e-05,
      "loss": 0.1005,
      "step": 24400
    },
    {
      "epoch": 0.5487613671997491,
      "grad_norm": 8.316376686096191,
      "learning_rate": 4.25277964141958e-05,
      "loss": 0.1194,
      "step": 24500
    },
    {
      "epoch": 0.5510012095148502,
      "grad_norm": 7.281992435455322,
      "learning_rate": 4.217932159588734e-05,
      "loss": 0.1275,
      "step": 24600
    },
    {
      "epoch": 0.5532410518299512,
      "grad_norm": 5.62957763671875,
      "learning_rate": 4.183123575549684e-05,
      "loss": 0.0989,
      "step": 24700
    },
    {
      "epoch": 0.5554808941450522,
      "grad_norm": 5.6208930015563965,
      "learning_rate": 4.148355620580699e-05,
      "loss": 0.1238,
      "step": 24800
    },
    {
      "epoch": 0.5577207364601532,
      "grad_norm": 5.27440071105957,
      "learning_rate": 4.113630023939269e-05,
      "loss": 0.1381,
      "step": 24900
    },
    {
      "epoch": 0.5599605787752542,
      "grad_norm": 7.504971981048584,
      "learning_rate": 4.0789485127761115e-05,
      "loss": 0.1047,
      "step": 25000
    },
    {
      "epoch": 0.5599605787752542,
      "eval_avg_non_pair_similarity": -0.0001179015729853973,
      "eval_avg_pair_similarity": 0.0010876374864019454,
      "eval_loss": 0.09219632297754288,
      "eval_runtime": 19.4855,
      "eval_samples_per_second": 25.66,
      "eval_similarity_ratio": -9.224961625716858,
      "eval_steps_per_second": 0.821,
      "step": 25000
    },
    {
      "epoch": 0.5622004210903553,
      "grad_norm": 9.146899223327637,
      "learning_rate": 4.044312812049252e-05,
      "loss": 0.1536,
      "step": 25100
    },
    {
      "epoch": 0.5644402634054563,
      "grad_norm": 5.325754642486572,
      "learning_rate": 4.009724644438243e-05,
      "loss": 0.1066,
      "step": 25200
    },
    {
      "epoch": 0.5666801057205573,
      "grad_norm": 6.888172626495361,
      "learning_rate": 3.975185730258473e-05,
      "loss": 0.11,
      "step": 25300
    },
    {
      "epoch": 0.5689199480356583,
      "grad_norm": 5.21387243270874,
      "learning_rate": 3.940697787375612e-05,
      "loss": 0.0928,
      "step": 25400
    },
    {
      "epoch": 0.5711597903507593,
      "grad_norm": 6.064445972442627,
      "learning_rate": 3.9062625311201595e-05,
      "loss": 0.1004,
      "step": 25500
    },
    {
      "epoch": 0.5733996326658604,
      "grad_norm": 10.20885944366455,
      "learning_rate": 3.871881674202141e-05,
      "loss": 0.1076,
      "step": 25600
    },
    {
      "epoch": 0.5756394749809614,
      "grad_norm": 5.855322360992432,
      "learning_rate": 3.8375569266259116e-05,
      "loss": 0.1287,
      "step": 25700
    },
    {
      "epoch": 0.5778793172960623,
      "grad_norm": 4.730558395385742,
      "learning_rate": 3.803289995605111e-05,
      "loss": 0.0928,
      "step": 25800
    },
    {
      "epoch": 0.5801191596111633,
      "grad_norm": 5.415820121765137,
      "learning_rate": 3.769082585477751e-05,
      "loss": 0.0993,
      "step": 25900
    },
    {
      "epoch": 0.5823590019262644,
      "grad_norm": 5.65316104888916,
      "learning_rate": 3.734936397621449e-05,
      "loss": 0.1053,
      "step": 26000
    },
    {
      "epoch": 0.5823590019262644,
      "eval_avg_non_pair_similarity": -7.923919997991759e-05,
      "eval_avg_pair_similarity": -0.0013436697054421529,
      "eval_loss": 0.09676006436347961,
      "eval_runtime": 19.4919,
      "eval_samples_per_second": 25.652,
      "eval_similarity_ratio": 16.957133663422812,
      "eval_steps_per_second": 0.821,
      "step": 26000
    }
  ],
  "logging_steps": 100,
  "max_steps": 44646,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}