Magpie-Align/MagpieLM-4B-Chat-v0.1
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity
Files
bff6bc7c7fa710accd8fafdf6313e8fe0f94812e
MagpieLM-4B-Chat-v0.1/trainer_state.json
ModelHub XC bff6bc7c7f 初始化项目，由ModelHub XC社区提供模型 
Model: Magpie-Align/MagpieLM-4B-Chat-v0.1
Source: Original Platform
2026-05-23 13:14:13 +08:00

23248 lines
788 KiB
JSON
Raw Blame History

{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.9999183606825047,
"eval_steps": 100,
"global_step": 1531,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.000653114539962446,
"grad_norm": 10.320163557955956,
"learning_rate": 9.74025974025974e-10,
"logits/chosen": -1.751611590385437,
"logits/rejected": -1.8014392852783203,
"logps/chosen": -475.2503662109375,
"logps/rejected": -473.9908752441406,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.001306229079924892,
"grad_norm": 9.57458924359402,
"learning_rate": 1.948051948051948e-09,
"logits/chosen": -1.6125147342681885,
"logits/rejected": -1.6155394315719604,
"logps/chosen": -510.954833984375,
"logps/rejected": -456.25836181640625,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 2
},
{
"epoch": 0.0019593436198873378,
"grad_norm": 7.386703525361656,
"learning_rate": 2.9220779220779217e-09,
"logits/chosen": -1.6310054063796997,
"logits/rejected": -1.664994239807129,
"logps/chosen": -526.9005737304688,
"logps/rejected": -502.1066589355469,
"loss": 0.6936,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0020557926036417484,
"rewards/margins": -0.0028404928743839264,
"rewards/rejected": 0.0007846998050808907,
"step": 3
},
{
"epoch": 0.002612458159849784,
"grad_norm": 8.184972062975513,
"learning_rate": 3.896103896103896e-09,
"logits/chosen": -1.6248514652252197,
"logits/rejected": -1.6408921480178833,
"logps/chosen": -452.23651123046875,
"logps/rejected": -464.2245178222656,
"loss": 0.6927,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0005986166652292013,
"rewards/margins": 0.0015139534370973706,
"rewards/rejected": -0.0009153365390375257,
"step": 4
},
{
"epoch": 0.0032655726998122294,
"grad_norm": 11.574533999877016,
"learning_rate": 4.8701298701298695e-09,
"logits/chosen": -1.7286796569824219,
"logits/rejected": -1.7585710287094116,
"logps/chosen": -462.3341979980469,
"logps/rejected": -441.8200378417969,
"loss": 0.6928,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00044421182246878743,
"rewards/margins": 0.00013704760931432247,
"rewards/rejected": -0.0005812598392367363,
"step": 5
},
{
"epoch": 0.0039186872397746755,
"grad_norm": 9.259170331718803,
"learning_rate": 5.844155844155843e-09,
"logits/chosen": -1.7418220043182373,
"logits/rejected": -1.7892532348632812,
"logps/chosen": -522.937744140625,
"logps/rejected": -552.8088989257812,
"loss": 0.6929,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0012052917154505849,
"rewards/margins": -0.002461440395563841,
"rewards/rejected": 0.0012561489129438996,
"step": 6
},
{
"epoch": 0.0045718017797371216,
"grad_norm": 7.63789930024362,
"learning_rate": 6.818181818181818e-09,
"logits/chosen": -1.7751970291137695,
"logits/rejected": -1.8038901090621948,
"logps/chosen": -533.4199829101562,
"logps/rejected": -480.9632263183594,
"loss": 0.6925,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0024563027545809746,
"rewards/margins": 0.0012150906259194016,
"rewards/rejected": 0.0012412117794156075,
"step": 7
},
{
"epoch": 0.005224916319699568,
"grad_norm": 9.961517065455444,
"learning_rate": 7.792207792207793e-09,
"logits/chosen": -1.7411315441131592,
"logits/rejected": -1.754071593284607,
"logps/chosen": -495.657470703125,
"logps/rejected": -475.14959716796875,
"loss": 0.693,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.00112921220716089,
"rewards/margins": -0.0019524528179317713,
"rewards/rejected": 0.0008232402033172548,
"step": 8
},
{
"epoch": 0.005878030859662013,
"grad_norm": 6.85854758907971,
"learning_rate": 8.766233766233765e-09,
"logits/chosen": -1.6711037158966064,
"logits/rejected": -1.7036432027816772,
"logps/chosen": -506.71331787109375,
"logps/rejected": -518.3823852539062,
"loss": 0.6932,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.00047096985508687794,
"rewards/margins": -0.0014907550066709518,
"rewards/rejected": 0.0019617248326539993,
"step": 9
},
{
"epoch": 0.006531145399624459,
"grad_norm": 11.552115994362305,
"learning_rate": 9.740259740259739e-09,
"logits/chosen": -1.6863927841186523,
"logits/rejected": -1.6270835399627686,
"logps/chosen": -527.3156127929688,
"logps/rejected": -595.2274169921875,
"loss": 0.6923,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.0007424141513183713,
"rewards/margins": 0.0002746771788224578,
"rewards/rejected": -0.0010170910973101854,
"step": 10
},
{
"epoch": 0.007184259939586905,
"grad_norm": 9.285086199342572,
"learning_rate": 1.0714285714285713e-08,
"logits/chosen": -1.6245068311691284,
"logits/rejected": -1.6487013101577759,
"logps/chosen": -552.9703369140625,
"logps/rejected": -526.51025390625,
"loss": 0.6929,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0022101732902228832,
"rewards/margins": 0.0029705429915338755,
"rewards/rejected": -0.0007603692938573658,
"step": 11
},
{
"epoch": 0.007837374479549351,
"grad_norm": 6.699656683031416,
"learning_rate": 1.1688311688311687e-08,
"logits/chosen": -1.7146100997924805,
"logits/rejected": -1.7387487888336182,
"logps/chosen": -512.6212158203125,
"logps/rejected": -510.3314514160156,
"loss": 0.6931,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.001821169862523675,
"rewards/margins": 0.0011536835227161646,
"rewards/rejected": -0.0029748533852398396,
"step": 12
},
{
"epoch": 0.008490489019511797,
"grad_norm": 7.205708657995355,
"learning_rate": 1.2662337662337662e-08,
"logits/chosen": -1.7469621896743774,
"logits/rejected": -1.7776381969451904,
"logps/chosen": -585.2794799804688,
"logps/rejected": -501.048828125,
"loss": 0.6933,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0012821245472878218,
"rewards/margins": 0.001307644764892757,
"rewards/rejected": -2.552039222791791e-05,
"step": 13
},
{
"epoch": 0.009143603559474243,
"grad_norm": 17.744991773610803,
"learning_rate": 1.3636363636363636e-08,
"logits/chosen": -1.704886555671692,
"logits/rejected": -1.7239997386932373,
"logps/chosen": -522.1668090820312,
"logps/rejected": -519.49951171875,
"loss": 0.6931,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.00018573057604953647,
"rewards/margins": 0.0024611069820821285,
"rewards/rejected": -0.0022753761149942875,
"step": 14
},
{
"epoch": 0.00979671809943669,
"grad_norm": 6.553103071836206,
"learning_rate": 1.461038961038961e-08,
"logits/chosen": -1.715399980545044,
"logits/rejected": -1.724109411239624,
"logps/chosen": -551.634765625,
"logps/rejected": -474.11785888671875,
"loss": 0.6937,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.001085147843696177,
"rewards/margins": 0.0005957795074209571,
"rewards/rejected": 0.0004893685108982027,
"step": 15
},
{
"epoch": 0.010449832639399135,
"grad_norm": 10.474583885683455,
"learning_rate": 1.5584415584415586e-08,
"logits/chosen": -1.791701078414917,
"logits/rejected": -1.7533316612243652,
"logps/chosen": -495.9596252441406,
"logps/rejected": -524.2186279296875,
"loss": 0.6935,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0018585537327453494,
"rewards/margins": -0.001228995155543089,
"rewards/rejected": -0.0006295585772022605,
"step": 16
},
{
"epoch": 0.011102947179361581,
"grad_norm": 8.585521664394983,
"learning_rate": 1.6558441558441556e-08,
"logits/chosen": -1.689960241317749,
"logits/rejected": -1.660653829574585,
"logps/chosen": -493.1070251464844,
"logps/rejected": -558.66748046875,
"loss": 0.6931,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.0004895019810646772,
"rewards/margins": 0.0019447661470621824,
"rewards/rejected": -0.0014552639331668615,
"step": 17
},
{
"epoch": 0.011756061719324026,
"grad_norm": 7.688284862178414,
"learning_rate": 1.753246753246753e-08,
"logits/chosen": -1.634385108947754,
"logits/rejected": -1.6527695655822754,
"logps/chosen": -476.8695068359375,
"logps/rejected": -481.3186950683594,
"loss": 0.6935,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0008362793596461415,
"rewards/margins": -0.0002491784398443997,
"rewards/rejected": 0.0010854577412828803,
"step": 18
},
{
"epoch": 0.012409176259286472,
"grad_norm": 11.300638272495242,
"learning_rate": 1.8506493506493504e-08,
"logits/chosen": -1.7336674928665161,
"logits/rejected": -1.7424148321151733,
"logps/chosen": -555.4613037109375,
"logps/rejected": -531.6415405273438,
"loss": 0.693,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0010435438016429543,
"rewards/margins": -0.0002537918044254184,
"rewards/rejected": 0.001297335489653051,
"step": 19
},
{
"epoch": 0.013062290799248918,
"grad_norm": 7.209810609202922,
"learning_rate": 1.9480519480519478e-08,
"logits/chosen": -1.6538033485412598,
"logits/rejected": -1.676405668258667,
"logps/chosen": -535.2705688476562,
"logps/rejected": -540.2841186523438,
"loss": 0.6926,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.000612707226537168,
"rewards/margins": -0.0002046869631158188,
"rewards/rejected": -0.0004080201033502817,
"step": 20
},
{
"epoch": 0.013715405339211364,
"grad_norm": 9.239564322245995,
"learning_rate": 2.0454545454545452e-08,
"logits/chosen": -1.6935447454452515,
"logits/rejected": -1.7362101078033447,
"logps/chosen": -483.9548034667969,
"logps/rejected": -443.07958984375,
"loss": 0.693,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.00011293424176983535,
"rewards/margins": 0.0016996811609715223,
"rewards/rejected": -0.0018126153154298663,
"step": 21
},
{
"epoch": 0.01436851987917381,
"grad_norm": 11.953662843213952,
"learning_rate": 2.1428571428571426e-08,
"logits/chosen": -1.649803638458252,
"logits/rejected": -1.6609901189804077,
"logps/chosen": -556.7703247070312,
"logps/rejected": -528.4605102539062,
"loss": 0.6941,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0003633833257481456,
"rewards/margins": -0.001797966891899705,
"rewards/rejected": 0.0021613501012325287,
"step": 22
},
{
"epoch": 0.015021634419136256,
"grad_norm": 5.585424980323363,
"learning_rate": 2.24025974025974e-08,
"logits/chosen": -1.7381466627120972,
"logits/rejected": -1.7497470378875732,
"logps/chosen": -470.3625183105469,
"logps/rejected": -458.3432312011719,
"loss": 0.6932,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0006353663047775626,
"rewards/margins": -0.0016739725833758712,
"rewards/rejected": 0.002309338888153434,
"step": 23
},
{
"epoch": 0.015674748959098702,
"grad_norm": 12.282080216373183,
"learning_rate": 2.3376623376623374e-08,
"logits/chosen": -1.6787421703338623,
"logits/rejected": -1.6916463375091553,
"logps/chosen": -500.0643615722656,
"logps/rejected": -530.3701171875,
"loss": 0.6923,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.00024698246852494776,
"rewards/margins": 0.0018526006024330854,
"rewards/rejected": -0.0016056179301813245,
"step": 24
},
{
"epoch": 0.016327863499061148,
"grad_norm": 6.151170677406452,
"learning_rate": 2.435064935064935e-08,
"logits/chosen": -1.730494499206543,
"logits/rejected": -1.7640714645385742,
"logps/chosen": -567.3126831054688,
"logps/rejected": -524.96630859375,
"loss": 0.694,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.002176732989028096,
"rewards/margins": -0.0018334074411541224,
"rewards/rejected": 0.004010140895843506,
"step": 25
},
{
"epoch": 0.016980978039023594,
"grad_norm": 6.027535384301359,
"learning_rate": 2.5324675324675325e-08,
"logits/chosen": -1.700589895248413,
"logits/rejected": -1.655698537826538,
"logps/chosen": -525.3576049804688,
"logps/rejected": -536.69384765625,
"loss": 0.6926,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.002966861240565777,
"rewards/margins": -0.00015424739103764296,
"rewards/rejected": 0.0031211089808493853,
"step": 26
},
{
"epoch": 0.01763409257898604,
"grad_norm": 10.92870471034886,
"learning_rate": 2.62987012987013e-08,
"logits/chosen": -1.7824229001998901,
"logits/rejected": -1.7060723304748535,
"logps/chosen": -499.6011962890625,
"logps/rejected": -459.31390380859375,
"loss": 0.6928,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0014267514925450087,
"rewards/margins": -0.000579934217967093,
"rewards/rejected": -0.0008468173909932375,
"step": 27
},
{
"epoch": 0.018287207118948486,
"grad_norm": 10.356153154714054,
"learning_rate": 2.7272727272727272e-08,
"logits/chosen": -1.7105991840362549,
"logits/rejected": -1.704366683959961,
"logps/chosen": -532.1188354492188,
"logps/rejected": -546.0477905273438,
"loss": 0.6924,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.0009431360522285104,
"rewards/margins": 0.0037088487297296524,
"rewards/rejected": -0.00276571256108582,
"step": 28
},
{
"epoch": 0.018940321658910932,
"grad_norm": 6.73381910894591,
"learning_rate": 2.8246753246753246e-08,
"logits/chosen": -1.651279330253601,
"logits/rejected": -1.6885946989059448,
"logps/chosen": -498.8572998046875,
"logps/rejected": -454.43701171875,
"loss": 0.6935,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0029693078249692917,
"rewards/margins": 0.00039883877616375685,
"rewards/rejected": 0.0025704693980515003,
"step": 29
},
{
"epoch": 0.01959343619887338,
"grad_norm": 9.815644247246649,
"learning_rate": 2.922077922077922e-08,
"logits/chosen": -1.659732699394226,
"logits/rejected": -1.6510050296783447,
"logps/chosen": -527.5221557617188,
"logps/rejected": -572.2091064453125,
"loss": 0.6924,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.00042819028021767735,
"rewards/margins": 0.0010157201904803514,
"rewards/rejected": -0.0014439105289056897,
"step": 30
},
{
"epoch": 0.020246550738835824,
"grad_norm": 6.574072733494995,
"learning_rate": 3.01948051948052e-08,
"logits/chosen": -1.7301993370056152,
"logits/rejected": -1.7147951126098633,
"logps/chosen": -495.9303894042969,
"logps/rejected": -487.27630615234375,
"loss": 0.6937,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.001899483148008585,
"rewards/margins": 0.0021509858779609203,
"rewards/rejected": -0.00025150307919830084,
"step": 31
},
{
"epoch": 0.02089966527879827,
"grad_norm": 6.336712001931794,
"learning_rate": 3.116883116883117e-08,
"logits/chosen": -1.769822597503662,
"logits/rejected": -1.7909101247787476,
"logps/chosen": -466.05523681640625,
"logps/rejected": -443.83001708984375,
"loss": 0.6939,
"rewards/accuracies": 0.34375,
"rewards/chosen": -0.002961869351565838,
"rewards/margins": -0.004024825058877468,
"rewards/rejected": 0.0010629557073116302,
"step": 32
},
{
"epoch": 0.021552779818760717,
"grad_norm": 7.984443377717489,
"learning_rate": 3.214285714285714e-08,
"logits/chosen": -1.6552331447601318,
"logits/rejected": -1.6343588829040527,
"logps/chosen": -473.5636901855469,
"logps/rejected": -490.8223876953125,
"loss": 0.6935,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.0031329344492405653,
"rewards/margins": 0.0010508847190067172,
"rewards/rejected": -0.004183819051831961,
"step": 33
},
{
"epoch": 0.022205894358723163,
"grad_norm": 6.978522846761633,
"learning_rate": 3.311688311688311e-08,
"logits/chosen": -1.685523271560669,
"logits/rejected": -1.6947942972183228,
"logps/chosen": -488.18597412109375,
"logps/rejected": -450.8877258300781,
"loss": 0.6926,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0016193913761526346,
"rewards/margins": 0.001421906752511859,
"rewards/rejected": 0.00019748439081013203,
"step": 34
},
{
"epoch": 0.02285900889868561,
"grad_norm": 7.493471887712862,
"learning_rate": 3.4090909090909086e-08,
"logits/chosen": -1.711535096168518,
"logits/rejected": -1.7027740478515625,
"logps/chosen": -518.72314453125,
"logps/rejected": -475.7657165527344,
"loss": 0.6931,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0004046607355121523,
"rewards/margins": 0.001060183160007,
"rewards/rejected": -0.0006555224535986781,
"step": 35
},
{
"epoch": 0.02351212343864805,
"grad_norm": 11.205930629830023,
"learning_rate": 3.506493506493506e-08,
"logits/chosen": -1.7367273569107056,
"logits/rejected": -1.762798547744751,
"logps/chosen": -453.77130126953125,
"logps/rejected": -481.04730224609375,
"loss": 0.6929,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.00015169614925980568,
"rewards/margins": 0.0007772659882903099,
"rewards/rejected": -0.0009289622539654374,
"step": 36
},
{
"epoch": 0.024165237978610497,
"grad_norm": 9.467273081236305,
"learning_rate": 3.6038961038961034e-08,
"logits/chosen": -1.7246737480163574,
"logits/rejected": -1.7480111122131348,
"logps/chosen": -466.54998779296875,
"logps/rejected": -427.57373046875,
"loss": 0.6931,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0015318728983402252,
"rewards/margins": 0.0017306231893599033,
"rewards/rejected": -0.00019875055295415223,
"step": 37
},
{
"epoch": 0.024818352518572943,
"grad_norm": 9.156288457593233,
"learning_rate": 3.701298701298701e-08,
"logits/chosen": -1.6512861251831055,
"logits/rejected": -1.690797209739685,
"logps/chosen": -509.2376708984375,
"logps/rejected": -528.0012817382812,
"loss": 0.6931,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0029702354222536087,
"rewards/margins": 0.0016618178924545646,
"rewards/rejected": 0.0013084171805530787,
"step": 38
},
{
"epoch": 0.02547146705853539,
"grad_norm": 17.95636045906593,
"learning_rate": 3.798701298701298e-08,
"logits/chosen": -1.6484848260879517,
"logits/rejected": -1.6626923084259033,
"logps/chosen": -444.7795715332031,
"logps/rejected": -453.7108154296875,
"loss": 0.6934,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.000330042967107147,
"rewards/margins": 0.0012718366924673319,
"rewards/rejected": -0.001601879601366818,
"step": 39
},
{
"epoch": 0.026124581598497836,
"grad_norm": 8.845607047635447,
"learning_rate": 3.8961038961038956e-08,
"logits/chosen": -1.6774462461471558,
"logits/rejected": -1.6941050291061401,
"logps/chosen": -490.5712890625,
"logps/rejected": -565.9147338867188,
"loss": 0.6937,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0004610727773979306,
"rewards/margins": -0.0016047237440943718,
"rewards/rejected": 0.002065796870738268,
"step": 40
},
{
"epoch": 0.02677769613846028,
"grad_norm": 9.919258253862095,
"learning_rate": 3.993506493506493e-08,
"logits/chosen": -1.6783130168914795,
"logits/rejected": -1.6832352876663208,
"logps/chosen": -424.7172546386719,
"logps/rejected": -463.73919677734375,
"loss": 0.6929,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0029250048100948334,
"rewards/margins": 0.00039243214996531606,
"rewards/rejected": 0.0025325727183371782,
"step": 41
},
{
"epoch": 0.027430810678422728,
"grad_norm": 7.033725232378199,
"learning_rate": 4.0909090909090904e-08,
"logits/chosen": -1.7260679006576538,
"logits/rejected": -1.7201330661773682,
"logps/chosen": -540.1764526367188,
"logps/rejected": -501.5211486816406,
"loss": 0.6926,
"rewards/accuracies": 0.5625,
"rewards/chosen": 2.3270258679986e-06,
"rewards/margins": 0.0007927274564281106,
"rewards/rejected": -0.0007904003723524511,
"step": 42
},
{
"epoch": 0.028083925218385174,
"grad_norm": 8.639855374503854,
"learning_rate": 4.188311688311688e-08,
"logits/chosen": -1.7189686298370361,
"logits/rejected": -1.7004797458648682,
"logps/chosen": -505.0670471191406,
"logps/rejected": -522.9998168945312,
"loss": 0.6933,
"rewards/accuracies": 0.34375,
"rewards/chosen": 0.0024530505761504173,
"rewards/margins": -0.0023120976984500885,
"rewards/rejected": 0.004765148274600506,
"step": 43
},
{
"epoch": 0.02873703975834762,
"grad_norm": 7.855142409694213,
"learning_rate": 4.285714285714285e-08,
"logits/chosen": -1.8125633001327515,
"logits/rejected": -1.8358653783798218,
"logps/chosen": -449.2396545410156,
"logps/rejected": -412.31622314453125,
"loss": 0.692,
"rewards/accuracies": 0.625,
"rewards/chosen": -7.799867307767272e-05,
"rewards/margins": 0.0013592124450951815,
"rewards/rejected": -0.0014372109435498714,
"step": 44
},
{
"epoch": 0.029390154298310066,
"grad_norm": 14.0615352021285,
"learning_rate": 4.3831168831168825e-08,
"logits/chosen": -1.6212176084518433,
"logits/rejected": -1.7029989957809448,
"logps/chosen": -485.98822021484375,
"logps/rejected": -451.6901550292969,
"loss": 0.6925,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.002106280066072941,
"rewards/margins": -0.00013485189992934465,
"rewards/rejected": 0.0022411320824176073,
"step": 45
},
{
"epoch": 0.030043268838272512,
"grad_norm": 13.882948188534005,
"learning_rate": 4.48051948051948e-08,
"logits/chosen": -1.7054741382598877,
"logits/rejected": -1.6707385778427124,
"logps/chosen": -523.1839599609375,
"logps/rejected": -538.1018676757812,
"loss": 0.6925,
"rewards/accuracies": 0.4375,
"rewards/chosen": 6.611342541873455e-05,
"rewards/margins": -0.0016670847544446588,
"rewards/rejected": 0.0017331981798633933,
"step": 46
},
{
"epoch": 0.030696383378234958,
"grad_norm": 9.843375032358303,
"learning_rate": 4.577922077922077e-08,
"logits/chosen": -1.7931544780731201,
"logits/rejected": -1.8134924173355103,
"logps/chosen": -576.3263549804688,
"logps/rejected": -531.6103515625,
"loss": 0.6925,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.00019430162501521409,
"rewards/margins": -0.0015074944822117686,
"rewards/rejected": 0.0013131927698850632,
"step": 47
},
{
"epoch": 0.031349497918197404,
"grad_norm": 13.08440244244893,
"learning_rate": 4.675324675324675e-08,
"logits/chosen": -1.728432536125183,
"logits/rejected": -1.687523365020752,
"logps/chosen": -424.1823425292969,
"logps/rejected": -436.7662353515625,
"loss": 0.6932,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0026053618639707565,
"rewards/margins": 0.0011907146545127034,
"rewards/rejected": 0.0014146470930427313,
"step": 48
},
{
"epoch": 0.03200261245815985,
"grad_norm": 9.954241633042399,
"learning_rate": 4.772727272727273e-08,
"logits/chosen": -1.7366923093795776,
"logits/rejected": -1.7875515222549438,
"logps/chosen": -482.4845886230469,
"logps/rejected": -462.21734619140625,
"loss": 0.6931,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.002501377835869789,
"rewards/margins": 0.0017617797711864114,
"rewards/rejected": 0.0007395982975140214,
"step": 49
},
{
"epoch": 0.032655726998122296,
"grad_norm": 7.516649010433926,
"learning_rate": 4.87012987012987e-08,
"logits/chosen": -1.643943190574646,
"logits/rejected": -1.6239250898361206,
"logps/chosen": -571.3704223632812,
"logps/rejected": -552.835205078125,
"loss": 0.6933,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0005948520265519619,
"rewards/margins": -0.001149489893577993,
"rewards/rejected": 0.0017443416872993112,
"step": 50
},
{
"epoch": 0.03330884153808474,
"grad_norm": 8.673266186430727,
"learning_rate": 4.9675324675324675e-08,
"logits/chosen": -1.654160499572754,
"logits/rejected": -1.6642738580703735,
"logps/chosen": -568.924072265625,
"logps/rejected": -502.4623107910156,
"loss": 0.6928,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.003631243482232094,
"rewards/margins": 0.002229990903288126,
"rewards/rejected": 0.0014012528117746115,
"step": 51
},
{
"epoch": 0.03396195607804719,
"grad_norm": 5.760323905203282,
"learning_rate": 5.064935064935065e-08,
"logits/chosen": -1.7572216987609863,
"logits/rejected": -1.746352195739746,
"logps/chosen": -500.9621276855469,
"logps/rejected": -516.9683837890625,
"loss": 0.6923,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0025606201961636543,
"rewards/margins": 0.0012761401012539864,
"rewards/rejected": 0.0012844798620790243,
"step": 52
},
{
"epoch": 0.034615070618009634,
"grad_norm": 13.745858132264178,
"learning_rate": 5.162337662337662e-08,
"logits/chosen": -1.6922714710235596,
"logits/rejected": -1.6829943656921387,
"logps/chosen": -521.5164794921875,
"logps/rejected": -511.12115478515625,
"loss": 0.6929,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0012258647475391626,
"rewards/margins": 0.0017541146371513605,
"rewards/rejected": -0.0005282496567815542,
"step": 53
},
{
"epoch": 0.03526818515797208,
"grad_norm": 9.901929656057623,
"learning_rate": 5.25974025974026e-08,
"logits/chosen": -1.6156866550445557,
"logits/rejected": -1.6489068269729614,
"logps/chosen": -504.8508605957031,
"logps/rejected": -510.78070068359375,
"loss": 0.6923,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0022437286097556353,
"rewards/margins": 0.0005393887404352427,
"rewards/rejected": 0.0017043398693203926,
"step": 54
},
{
"epoch": 0.035921299697934526,
"grad_norm": 7.628193867040592,
"learning_rate": 5.357142857142857e-08,
"logits/chosen": -1.7694371938705444,
"logits/rejected": -1.7641582489013672,
"logps/chosen": -497.0212707519531,
"logps/rejected": -577.7604370117188,
"loss": 0.693,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.00022474752040579915,
"rewards/margins": 0.0001437664614059031,
"rewards/rejected": 8.098152466118336e-05,
"step": 55
},
{
"epoch": 0.03657441423789697,
"grad_norm": 6.526308952187627,
"learning_rate": 5.4545454545454545e-08,
"logits/chosen": -1.6863549947738647,
"logits/rejected": -1.726689100265503,
"logps/chosen": -570.279052734375,
"logps/rejected": -529.3751831054688,
"loss": 0.693,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.0028358842246234417,
"rewards/margins": -0.0016305134631693363,
"rewards/rejected": -0.00120537041220814,
"step": 56
},
{
"epoch": 0.03722752877785942,
"grad_norm": 7.971501408541291,
"learning_rate": 5.551948051948052e-08,
"logits/chosen": -1.7459444999694824,
"logits/rejected": -1.8165135383605957,
"logps/chosen": -500.9310302734375,
"logps/rejected": -476.9623107910156,
"loss": 0.6933,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.002534267958253622,
"rewards/margins": -0.00011749513214454055,
"rewards/rejected": -0.0024167723022401333,
"step": 57
},
{
"epoch": 0.037880643317821865,
"grad_norm": 19.843826487104433,
"learning_rate": 5.649350649350649e-08,
"logits/chosen": -1.6915953159332275,
"logits/rejected": -1.7043017148971558,
"logps/chosen": -437.893310546875,
"logps/rejected": -444.86895751953125,
"loss": 0.6934,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0005522965220734477,
"rewards/margins": -0.0005562017904594541,
"rewards/rejected": 3.905253834091127e-06,
"step": 58
},
{
"epoch": 0.03853375785778431,
"grad_norm": 17.45299457584043,
"learning_rate": 5.7467532467532466e-08,
"logits/chosen": -1.7828876972198486,
"logits/rejected": -1.7247426509857178,
"logps/chosen": -489.2809753417969,
"logps/rejected": -575.0067138671875,
"loss": 0.6935,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.004004344809800386,
"rewards/margins": -0.001498160301707685,
"rewards/rejected": -0.0025061843916773796,
"step": 59
},
{
"epoch": 0.03918687239774676,
"grad_norm": 7.706775067267686,
"learning_rate": 5.844155844155844e-08,
"logits/chosen": -1.7627689838409424,
"logits/rejected": -1.7395163774490356,
"logps/chosen": -543.8458251953125,
"logps/rejected": -535.2295532226562,
"loss": 0.6929,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.002326030982658267,
"rewards/margins": 0.0011458850931376219,
"rewards/rejected": 0.0011801458895206451,
"step": 60
},
{
"epoch": 0.0398399869377092,
"grad_norm": 6.253034693514966,
"learning_rate": 5.9415584415584414e-08,
"logits/chosen": -1.6752774715423584,
"logits/rejected": -1.6769983768463135,
"logps/chosen": -513.5790405273438,
"logps/rejected": -491.4671630859375,
"loss": 0.6931,
"rewards/accuracies": 0.53125,
"rewards/chosen": 4.048121627420187e-05,
"rewards/margins": 0.00097788090351969,
"rewards/rejected": -0.0009373998618684709,
"step": 61
},
{
"epoch": 0.04049310147767165,
"grad_norm": 8.25357941061994,
"learning_rate": 6.03896103896104e-08,
"logits/chosen": -1.7725433111190796,
"logits/rejected": -1.8016797304153442,
"logps/chosen": -504.8186950683594,
"logps/rejected": -534.6757202148438,
"loss": 0.6929,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.0038966606371104717,
"rewards/margins": 0.0019743391312658787,
"rewards/rejected": 0.0019223212730139494,
"step": 62
},
{
"epoch": 0.041146216017634095,
"grad_norm": 7.603828708440927,
"learning_rate": 6.136363636363636e-08,
"logits/chosen": -1.784754991531372,
"logits/rejected": -1.7742725610733032,
"logps/chosen": -547.2255249023438,
"logps/rejected": -546.972412109375,
"loss": 0.6936,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.002894229721277952,
"rewards/margins": -0.0006492708926089108,
"rewards/rejected": 0.0035435007885098457,
"step": 63
},
{
"epoch": 0.04179933055759654,
"grad_norm": 15.969939708450756,
"learning_rate": 6.233766233766234e-08,
"logits/chosen": -1.654356837272644,
"logits/rejected": -1.6166218519210815,
"logps/chosen": -488.8111877441406,
"logps/rejected": -569.5529174804688,
"loss": 0.692,
"rewards/accuracies": 0.71875,
"rewards/chosen": 0.003502319101244211,
"rewards/margins": 0.005036606453359127,
"rewards/rejected": -0.0015342880506068468,
"step": 64
},
{
"epoch": 0.04245244509755899,
"grad_norm": 22.983780138981498,
"learning_rate": 6.331168831168831e-08,
"logits/chosen": -1.7115808725357056,
"logits/rejected": -1.7262012958526611,
"logps/chosen": -499.15399169921875,
"logps/rejected": -488.51214599609375,
"loss": 0.6929,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0016056394670158625,
"rewards/margins": 0.0008727168897166848,
"rewards/rejected": 0.0007329225772991776,
"step": 65
},
{
"epoch": 0.04310555963752143,
"grad_norm": 11.044651160280129,
"learning_rate": 6.428571428571428e-08,
"logits/chosen": -1.795534610748291,
"logits/rejected": -1.799638271331787,
"logps/chosen": -533.3612670898438,
"logps/rejected": -621.8697509765625,
"loss": 0.6923,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.001967411022633314,
"rewards/margins": 0.003664374118670821,
"rewards/rejected": -0.001696963096037507,
"step": 66
},
{
"epoch": 0.04375867417748388,
"grad_norm": 25.764603545634195,
"learning_rate": 6.525974025974026e-08,
"logits/chosen": -1.7212399244308472,
"logits/rejected": -1.6900732517242432,
"logps/chosen": -433.75335693359375,
"logps/rejected": -483.9370422363281,
"loss": 0.6929,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.005031700246036053,
"rewards/margins": 0.00277211656793952,
"rewards/rejected": 0.002259582979604602,
"step": 67
},
{
"epoch": 0.044411788717446325,
"grad_norm": 7.6456802404761985,
"learning_rate": 6.623376623376622e-08,
"logits/chosen": -1.8252774477005005,
"logits/rejected": -1.7665594816207886,
"logps/chosen": -505.1371765136719,
"logps/rejected": -519.2654418945312,
"loss": 0.693,
"rewards/accuracies": 0.53125,
"rewards/chosen": -7.071479922160506e-05,
"rewards/margins": 0.0002764108357951045,
"rewards/rejected": -0.00034712543128989637,
"step": 68
},
{
"epoch": 0.04506490325740877,
"grad_norm": 22.249196941026028,
"learning_rate": 6.72077922077922e-08,
"logits/chosen": -1.7685558795928955,
"logits/rejected": -1.754807710647583,
"logps/chosen": -498.36309814453125,
"logps/rejected": -479.595947265625,
"loss": 0.6927,
"rewards/accuracies": 0.59375,
"rewards/chosen": 0.0044998289085924625,
"rewards/margins": 0.0033680484630167484,
"rewards/rejected": 0.0011317802127450705,
"step": 69
},
{
"epoch": 0.04571801779737122,
"grad_norm": 18.41888656875299,
"learning_rate": 6.818181818181817e-08,
"logits/chosen": -1.6847811937332153,
"logits/rejected": -1.7150285243988037,
"logps/chosen": -521.6116333007812,
"logps/rejected": -501.6766662597656,
"loss": 0.6926,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.00515484344214201,
"rewards/margins": 0.0027162095066159964,
"rewards/rejected": 0.0024386332370340824,
"step": 70
},
{
"epoch": 0.046371132337333656,
"grad_norm": 7.5323060443272425,
"learning_rate": 6.915584415584415e-08,
"logits/chosen": -1.6845817565917969,
"logits/rejected": -1.6868298053741455,
"logps/chosen": -509.47174072265625,
"logps/rejected": -497.8253173828125,
"loss": 0.6923,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.006278076209127903,
"rewards/margins": 0.003210592083632946,
"rewards/rejected": 0.0030674838926643133,
"step": 71
},
{
"epoch": 0.0470242468772961,
"grad_norm": 11.518462131512745,
"learning_rate": 7.012987012987012e-08,
"logits/chosen": -1.689769983291626,
"logits/rejected": -1.6752269268035889,
"logps/chosen": -508.0777587890625,
"logps/rejected": -497.0141906738281,
"loss": 0.6925,
"rewards/accuracies": 0.4375,
"rewards/chosen": 0.0023371409624814987,
"rewards/margins": -0.0009353661444038153,
"rewards/rejected": 0.0032725068740546703,
"step": 72
},
{
"epoch": 0.04767736141725855,
"grad_norm": 13.464519389398141,
"learning_rate": 7.11038961038961e-08,
"logits/chosen": -1.7256083488464355,
"logits/rejected": -1.7064285278320312,
"logps/chosen": -528.0186157226562,
"logps/rejected": -521.5746459960938,
"loss": 0.693,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0029949708841741085,
"rewards/margins": 0.0013601587852463126,
"rewards/rejected": 0.0016348123317584395,
"step": 73
},
{
"epoch": 0.048330475957220995,
"grad_norm": 20.376786509212838,
"learning_rate": 7.207792207792207e-08,
"logits/chosen": -1.641348123550415,
"logits/rejected": -1.6459161043167114,
"logps/chosen": -482.44732666015625,
"logps/rejected": -473.39190673828125,
"loss": 0.6926,
"rewards/accuracies": 0.40625,
"rewards/chosen": 0.0051408931612968445,
"rewards/margins": -0.0015343116829171777,
"rewards/rejected": 0.006675205193459988,
"step": 74
},
{
"epoch": 0.04898359049718344,
"grad_norm": 6.970678593146312,
"learning_rate": 7.305194805194805e-08,
"logits/chosen": -1.7044446468353271,
"logits/rejected": -1.68734610080719,
"logps/chosen": -502.9562072753906,
"logps/rejected": -517.5908813476562,
"loss": 0.6922,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0027034569066017866,
"rewards/margins": 0.0027098727878183126,
"rewards/rejected": -6.4158812165260315e-06,
"step": 75
},
{
"epoch": 0.04963670503714589,
"grad_norm": 9.6526922150323,
"learning_rate": 7.402597402597402e-08,
"logits/chosen": -1.7046608924865723,
"logits/rejected": -1.7354328632354736,
"logps/chosen": -535.493408203125,
"logps/rejected": -594.682861328125,
"loss": 0.6927,
"rewards/accuracies": 0.34375,
"rewards/chosen": -0.004995846655219793,
"rewards/margins": -0.0013705159071832895,
"rewards/rejected": -0.0036253309808671474,
"step": 76
},
{
"epoch": 0.05028981957710833,
"grad_norm": 7.329104460905438,
"learning_rate": 7.5e-08,
"logits/chosen": -1.7437448501586914,
"logits/rejected": -1.786738395690918,
"logps/chosen": -503.44281005859375,
"logps/rejected": -479.0059814453125,
"loss": 0.6907,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.0025321291759610176,
"rewards/margins": 0.007668951991945505,
"rewards/rejected": -0.005136823281645775,
"step": 77
},
{
"epoch": 0.05094293411707078,
"grad_norm": 12.493258401725456,
"learning_rate": 7.597402597402596e-08,
"logits/chosen": -1.7119678258895874,
"logits/rejected": -1.7332830429077148,
"logps/chosen": -488.6138916015625,
"logps/rejected": -485.7553405761719,
"loss": 0.6921,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.004445738159120083,
"rewards/margins": 0.002824036870151758,
"rewards/rejected": 0.001621701754629612,
"step": 78
},
{
"epoch": 0.051596048657033225,
"grad_norm": 8.01136858731629,
"learning_rate": 7.694805194805194e-08,
"logits/chosen": -1.7600123882293701,
"logits/rejected": -1.7657560110092163,
"logps/chosen": -527.957763671875,
"logps/rejected": -547.0194702148438,
"loss": 0.6924,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.0005399225628934801,
"rewards/margins": 0.0007603027625009418,
"rewards/rejected": -0.00022038002498447895,
"step": 79
},
{
"epoch": 0.05224916319699567,
"grad_norm": 15.784986319963485,
"learning_rate": 7.792207792207791e-08,
"logits/chosen": -1.7285094261169434,
"logits/rejected": -1.712064504623413,
"logps/chosen": -521.34814453125,
"logps/rejected": -482.8376159667969,
"loss": 0.6921,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0010725975735113025,
"rewards/margins": -0.0014756344025954604,
"rewards/rejected": 0.00040303694549947977,
"step": 80
},
{
"epoch": 0.05290227773695812,
"grad_norm": 8.657892032924215,
"learning_rate": 7.889610389610389e-08,
"logits/chosen": -1.6536437273025513,
"logits/rejected": -1.6625897884368896,
"logps/chosen": -549.460205078125,
"logps/rejected": -534.7042236328125,
"loss": 0.6927,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.0003026010235771537,
"rewards/margins": 0.0008853006293065846,
"rewards/rejected": -0.0005827000131830573,
"step": 81
},
{
"epoch": 0.05355539227692056,
"grad_norm": 10.870568795609778,
"learning_rate": 7.987012987012986e-08,
"logits/chosen": -1.6692464351654053,
"logits/rejected": -1.6596240997314453,
"logps/chosen": -506.01922607421875,
"logps/rejected": -524.4590454101562,
"loss": 0.6917,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.0017397881019860506,
"rewards/margins": 0.005637907888740301,
"rewards/rejected": -0.003898120019584894,
"step": 82
},
{
"epoch": 0.05420850681688301,
"grad_norm": 5.851977319869193,
"learning_rate": 8.084415584415584e-08,
"logits/chosen": -1.738106608390808,
"logits/rejected": -1.7303545475006104,
"logps/chosen": -482.5537414550781,
"logps/rejected": -484.3572998046875,
"loss": 0.6922,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.0022105362731963396,
"rewards/margins": 0.001744122477248311,
"rewards/rejected": 0.00046641333028674126,
"step": 83
},
{
"epoch": 0.054861621356845455,
"grad_norm": 5.9816373124178375,
"learning_rate": 8.181818181818181e-08,
"logits/chosen": -1.7219839096069336,
"logits/rejected": -1.7243647575378418,
"logps/chosen": -531.4603881835938,
"logps/rejected": -624.1871337890625,
"loss": 0.6912,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.001187248039059341,
"rewards/margins": 0.008130445145070553,
"rewards/rejected": -0.009317693300545216,
"step": 84
},
{
"epoch": 0.0555147358968079,
"grad_norm": 17.14950100054283,
"learning_rate": 8.279220779220779e-08,
"logits/chosen": -1.7397611141204834,
"logits/rejected": -1.7358020544052124,
"logps/chosen": -475.4699401855469,
"logps/rejected": -454.5489501953125,
"loss": 0.6923,
"rewards/accuracies": 0.46875,
"rewards/chosen": 0.004636278375983238,
"rewards/margins": 0.0002992916852235794,
"rewards/rejected": 0.004336986690759659,
"step": 85
},
{
"epoch": 0.05616785043677035,
"grad_norm": 10.798912022369151,
"learning_rate": 8.376623376623376e-08,
"logits/chosen": -1.6888103485107422,
"logits/rejected": -1.6987148523330688,
"logps/chosen": -572.60400390625,
"logps/rejected": -599.0338745117188,
"loss": 0.691,
"rewards/accuracies": 0.6875,
"rewards/chosen": 0.0021805190481245518,
"rewards/margins": 0.006904205307364464,
"rewards/rejected": -0.004723686724901199,
"step": 86
},
{
"epoch": 0.05682096497673279,
"grad_norm": 5.848424010532979,
"learning_rate": 8.474025974025974e-08,
"logits/chosen": -1.7512600421905518,
"logits/rejected": -1.77037513256073,
"logps/chosen": -575.0525512695312,
"logps/rejected": -551.678955078125,
"loss": 0.692,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.000223713053856045,
"rewards/margins": 6.933440454304218e-05,
"rewards/rejected": 0.00015437835827469826,
"step": 87
},
{
"epoch": 0.05747407951669524,
"grad_norm": 9.540765629514766,
"learning_rate": 8.57142857142857e-08,
"logits/chosen": -1.6761497259140015,
"logits/rejected": -1.6493499279022217,
"logps/chosen": -565.0869140625,
"logps/rejected": -575.0073852539062,
"loss": 0.6916,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.001043493626639247,
"rewards/margins": 0.004106073174625635,
"rewards/rejected": -0.0051495665684342384,
"step": 88
},
{
"epoch": 0.058127194056657686,
"grad_norm": 10.17514292490149,
"learning_rate": 8.668831168831168e-08,
"logits/chosen": -1.8012816905975342,
"logits/rejected": -1.7821853160858154,
"logps/chosen": -503.46295166015625,
"logps/rejected": -492.86474609375,
"loss": 0.6923,
"rewards/accuracies": 0.375,
"rewards/chosen": 0.0033976598642766476,
"rewards/margins": -0.0017642759485170245,
"rewards/rejected": 0.005161936394870281,
"step": 89
},
{
"epoch": 0.05878030859662013,
"grad_norm": 9.087384783413304,
"learning_rate": 8.766233766233765e-08,
"logits/chosen": -1.6532089710235596,
"logits/rejected": -1.657583475112915,
"logps/chosen": -515.2310180664062,
"logps/rejected": -509.0292053222656,
"loss": 0.6916,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0025243638083338737,
"rewards/margins": 0.0019023300847038627,
"rewards/rejected": 0.0006220341892912984,
"step": 90
},
{
"epoch": 0.05943342313658258,
"grad_norm": 11.600432401053613,
"learning_rate": 8.863636363636363e-08,
"logits/chosen": -1.7180678844451904,
"logits/rejected": -1.702775478363037,
"logps/chosen": -534.7321166992188,
"logps/rejected": -540.93994140625,
"loss": 0.6914,
"rewards/accuracies": 0.625,
"rewards/chosen": 0.004910244606435299,
"rewards/margins": 0.003610987449064851,
"rewards/rejected": 0.0012992569245398045,
"step": 91
},
{
"epoch": 0.060086537676545024,
"grad_norm": 19.03876328807656,
"learning_rate": 8.96103896103896e-08,
"logits/chosen": -1.7077628374099731,
"logits/rejected": -1.7351174354553223,
"logps/chosen": -529.73681640625,
"logps/rejected": -508.55828857421875,
"loss": 0.6917,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.003127522300928831,
"rewards/margins": 0.00047443623770959675,
"rewards/rejected": -0.0036019585095345974,
"step": 92
},
{
"epoch": 0.06073965221650747,
"grad_norm": 16.589555191285804,
"learning_rate": 9.058441558441558e-08,
"logits/chosen": -1.6833391189575195,
"logits/rejected": -1.674033284187317,
"logps/chosen": -477.49798583984375,
"logps/rejected": -464.76336669921875,
"loss": 0.6915,
"rewards/accuracies": 0.53125,
"rewards/chosen": 0.001920571201480925,
"rewards/margins": 0.0017885612323880196,
"rewards/rejected": 0.00013200979446992278,
"step": 93
},
{
"epoch": 0.061392766756469916,
"grad_norm": 21.483285457055274,
"learning_rate": 9.155844155844155e-08,
"logits/chosen": -1.6899855136871338,
"logits/rejected": -1.7081849575042725,
"logps/chosen": -472.30670166015625,
"logps/rejected": -473.02532958984375,
"loss": 0.6906,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0001794814015738666,
"rewards/margins": 0.0037227103020995855,
"rewards/rejected": -0.0039021919947117567,
"step": 94
},
{
"epoch": 0.06204588129643236,
"grad_norm": 8.584522992242459,
"learning_rate": 9.253246753246754e-08,
"logits/chosen": -1.6826878786087036,
"logits/rejected": -1.6474887132644653,
"logps/chosen": -469.18511962890625,
"logps/rejected": -475.0853576660156,
"loss": 0.6914,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.0038090678863227367,
"rewards/margins": 0.005096444860100746,
"rewards/rejected": -0.00890551321208477,
"step": 95
},
{
"epoch": 0.06269899583639481,
"grad_norm": 10.611257258997737,
"learning_rate": 9.35064935064935e-08,
"logits/chosen": -1.7167426347732544,
"logits/rejected": -1.745957374572754,
"logps/chosen": -415.5352478027344,
"logps/rejected": -406.41021728515625,
"loss": 0.6919,
"rewards/accuracies": 0.5625,
"rewards/chosen": -8.110757335089147e-05,
"rewards/margins": 0.005273091606795788,
"rewards/rejected": -0.005354199092835188,
"step": 96
},
{
"epoch": 0.06335211037635725,
"grad_norm": 17.637332082165788,
"learning_rate": 9.448051948051949e-08,
"logits/chosen": -1.7745977640151978,
"logits/rejected": -1.7798304557800293,
"logps/chosen": -471.1458435058594,
"logps/rejected": -464.43621826171875,
"loss": 0.6916,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0014066218864172697,
"rewards/margins": 0.002926683286204934,
"rewards/rejected": -0.004333305172622204,
"step": 97
},
{
"epoch": 0.0640052249163197,
"grad_norm": 15.475362634607388,
"learning_rate": 9.545454545454546e-08,
"logits/chosen": -1.757310390472412,
"logits/rejected": -1.7441529035568237,
"logps/chosen": -543.7720336914062,
"logps/rejected": -521.256591796875,
"loss": 0.6905,
"rewards/accuracies": 0.5625,
"rewards/chosen": 0.0013191296020522714,
"rewards/margins": 0.006374385207891464,
"rewards/rejected": -0.005055255722254515,
"step": 98
},
{
"epoch": 0.06465833945628215,
"grad_norm": 8.90308809277182,
"learning_rate": 9.642857142857144e-08,
"logits/chosen": -1.7599557638168335,
"logits/rejected": -1.7538748979568481,
"logps/chosen": -488.3480224609375,
"logps/rejected": -493.222900390625,
"loss": 0.6907,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.005196585785597563,
"rewards/margins": 0.00629068398848176,
"rewards/rejected": -0.0010940982028841972,
"step": 99
},
{
"epoch": 0.06531145399624459,
"grad_norm": 8.599502814641067,
"learning_rate": 9.74025974025974e-08,
"logits/chosen": -1.6944297552108765,
"logits/rejected": -1.6814221143722534,
"logps/chosen": -491.616943359375,
"logps/rejected": -464.1621398925781,
"loss": 0.6911,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.0056236740201711655,
"rewards/margins": 0.0013095784233883023,
"rewards/rejected": -0.006933252792805433,
"step": 100
},
{
"epoch": 0.06531145399624459,
"eval_logits/chosen": -1.7781137228012085,
"eval_logits/rejected": -1.783394694328308,
"eval_logps/chosen": -510.6041564941406,
"eval_logps/rejected": -502.9036865234375,
"eval_loss": 0.6912173628807068,
"eval_rewards/accuracies": 0.5640000104904175,
"eval_rewards/chosen": -0.0025881431065499783,
"eval_rewards/margins": 0.004054033197462559,
"eval_rewards/rejected": -0.006642176769673824,
"eval_runtime": 309.9462,
"eval_samples_per_second": 12.905,
"eval_steps_per_second": 0.807,
"step": 100
},
{
"epoch": 0.06596456853620704,
"grad_norm": 11.975392598780218,
"learning_rate": 9.837662337662338e-08,
"logits/chosen": -1.8105459213256836,
"logits/rejected": -1.7418413162231445,
"logps/chosen": -495.683837890625,
"logps/rejected": -558.681884765625,
"loss": 0.6916,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.013931737281382084,
"rewards/margins": 0.0046969749964773655,
"rewards/rejected": -0.018628710880875587,
"step": 101
},
{
"epoch": 0.06661768307616948,
"grad_norm": 19.08100215691415,
"learning_rate": 9.935064935064935e-08,
"logits/chosen": -1.7734951972961426,
"logits/rejected": -1.785585880279541,
"logps/chosen": -521.8323974609375,
"logps/rejected": -493.7707824707031,
"loss": 0.69,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.0010571195743978024,
"rewards/margins": 0.004120311699807644,
"rewards/rejected": -0.0051774317398667336,
"step": 102
},
{
"epoch": 0.06727079761613193,
"grad_norm": 16.407227405047756,
"learning_rate": 1.0032467532467532e-07,
"logits/chosen": -1.813089370727539,
"logits/rejected": -1.8418606519699097,
"logps/chosen": -535.99658203125,
"logps/rejected": -547.38134765625,
"loss": 0.6894,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.0009367464226670563,
"rewards/margins": 0.004543146584182978,
"rewards/rejected": -0.005479893181473017,
"step": 103
},
{
"epoch": 0.06792391215609438,
"grad_norm": 10.784387451558391,
"learning_rate": 1.012987012987013e-07,
"logits/chosen": -1.6748157739639282,
"logits/rejected": -1.6749082803726196,
"logps/chosen": -544.4282836914062,
"logps/rejected": -502.47491455078125,
"loss": 0.6909,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.005649018567055464,
"rewards/margins": 0.002750987885519862,
"rewards/rejected": -0.008400006219744682,
"step": 104
},
{
"epoch": 0.06857702669605682,
"grad_norm": 7.366213696928153,
"learning_rate": 1.0227272727272727e-07,
"logits/chosen": -1.8434463739395142,
"logits/rejected": -1.8662176132202148,
"logps/chosen": -499.53240966796875,
"logps/rejected": -481.79345703125,
"loss": 0.6913,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.008984744548797607,
"rewards/margins": 0.001937179360538721,
"rewards/rejected": -0.010921923443675041,
"step": 105
},
{
"epoch": 0.06923014123601927,
"grad_norm": 8.12342504564889,
"learning_rate": 1.0324675324675325e-07,
"logits/chosen": -1.7362070083618164,
"logits/rejected": -1.7785553932189941,
"logps/chosen": -524.1248779296875,
"logps/rejected": -506.4564514160156,
"loss": 0.6903,
"rewards/accuracies": 0.65625,
"rewards/chosen": 0.001606869394890964,
"rewards/margins": 0.002641630358994007,
"rewards/rejected": -0.0010347607312723994,
"step": 106
},
{
"epoch": 0.06988325577598171,
"grad_norm": 6.5649883506893385,
"learning_rate": 1.0422077922077921e-07,
"logits/chosen": -1.695350170135498,
"logits/rejected": -1.6904462575912476,
"logps/chosen": -556.2769775390625,
"logps/rejected": -519.2373046875,
"loss": 0.6906,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.012179381214082241,
"rewards/margins": 0.008277284912765026,
"rewards/rejected": -0.020456667989492416,
"step": 107
},
{
"epoch": 0.07053637031594416,
"grad_norm": 13.47303429680314,
"learning_rate": 1.051948051948052e-07,
"logits/chosen": -1.6657894849777222,
"logits/rejected": -1.6693402528762817,
"logps/chosen": -535.307373046875,
"logps/rejected": -526.8478393554688,
"loss": 0.6906,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.006545607931911945,
"rewards/margins": 0.008519239723682404,
"rewards/rejected": -0.015064846724271774,
"step": 108
},
{
"epoch": 0.0711894848559066,
"grad_norm": 16.04827613099803,
"learning_rate": 1.0616883116883116e-07,
"logits/chosen": -1.6586048603057861,
"logits/rejected": -1.6781889200210571,
"logps/chosen": -530.2203979492188,
"logps/rejected": -543.2777709960938,
"loss": 0.6887,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.0075052459724247456,
"rewards/margins": 0.00435210857540369,
"rewards/rejected": -0.011857354082167149,
"step": 109
},
{
"epoch": 0.07184259939586905,
"grad_norm": 6.9887079248570245,
"learning_rate": 1.0714285714285714e-07,
"logits/chosen": -1.719894289970398,
"logits/rejected": -1.73370361328125,
"logps/chosen": -461.96893310546875,
"logps/rejected": -460.13861083984375,
"loss": 0.6903,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.003742053173482418,
"rewards/margins": 0.008006452582776546,
"rewards/rejected": -0.011748505756258965,
"step": 110
},
{
"epoch": 0.0724957139358315,
"grad_norm": 17.35796671683516,
"learning_rate": 1.0811688311688311e-07,
"logits/chosen": -1.7226269245147705,
"logits/rejected": -1.70708429813385,
"logps/chosen": -531.7066650390625,
"logps/rejected": -529.4414672851562,
"loss": 0.6929,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.013616499491035938,
"rewards/margins": 0.0025032046250998974,
"rewards/rejected": -0.01611970365047455,
"step": 111
},
{
"epoch": 0.07314882847579394,
"grad_norm": 8.237208524897573,
"learning_rate": 1.0909090909090909e-07,
"logits/chosen": -1.6860663890838623,
"logits/rejected": -1.6681212186813354,
"logps/chosen": -412.7210693359375,
"logps/rejected": -408.09210205078125,
"loss": 0.6898,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.010377769358456135,
"rewards/margins": 0.0005328749539330602,
"rewards/rejected": -0.010910644195973873,
"step": 112
},
{
"epoch": 0.07380194301575639,
"grad_norm": 14.11440807458263,
"learning_rate": 1.1006493506493506e-07,
"logits/chosen": -1.7322160005569458,
"logits/rejected": -1.7938529253005981,
"logps/chosen": -524.374755859375,
"logps/rejected": -477.0540771484375,
"loss": 0.6909,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.010047688148915768,
"rewards/margins": 0.0021818396635353565,
"rewards/rejected": -0.012229528278112411,
"step": 113
},
{
"epoch": 0.07445505755571884,
"grad_norm": 19.41135088594217,
"learning_rate": 1.1103896103896104e-07,
"logits/chosen": -1.7305679321289062,
"logits/rejected": -1.687849521636963,
"logps/chosen": -497.699462890625,
"logps/rejected": -470.3168640136719,
"loss": 0.69,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.013295488432049751,
"rewards/margins": 0.0023743584752082825,
"rewards/rejected": -0.015669845044612885,
"step": 114
},
{
"epoch": 0.07510817209568128,
"grad_norm": 16.441138226102247,
"learning_rate": 1.12012987012987e-07,
"logits/chosen": -1.7420533895492554,
"logits/rejected": -1.7339985370635986,
"logps/chosen": -581.6981201171875,
"logps/rejected": -568.88623046875,
"loss": 0.6899,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.014405852183699608,
"rewards/margins": 0.004987373016774654,
"rewards/rejected": -0.019393224269151688,
"step": 115
},
{
"epoch": 0.07576128663564373,
"grad_norm": 8.586529812749172,
"learning_rate": 1.1298701298701299e-07,
"logits/chosen": -1.7932806015014648,
"logits/rejected": -1.7588953971862793,
"logps/chosen": -532.9600830078125,
"logps/rejected": -497.33770751953125,
"loss": 0.6905,
"rewards/accuracies": 0.5,
"rewards/chosen": -8.44545429572463e-05,
"rewards/margins": 0.00019355083350092173,
"rewards/rejected": -0.0002780060167424381,
"step": 116
},
{
"epoch": 0.07641440117560618,
"grad_norm": 11.94419407737389,
"learning_rate": 1.1396103896103895e-07,
"logits/chosen": -1.7573413848876953,
"logits/rejected": -1.7641370296478271,
"logps/chosen": -529.7691650390625,
"logps/rejected": -494.5425109863281,
"loss": 0.6892,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.009526774287223816,
"rewards/margins": 0.004622914828360081,
"rewards/rejected": -0.014149690046906471,
"step": 117
},
{
"epoch": 0.07706751571556862,
"grad_norm": 8.288500408438793,
"learning_rate": 1.1493506493506493e-07,
"logits/chosen": -1.6377328634262085,
"logits/rejected": -1.7309472560882568,
"logps/chosen": -479.968505859375,
"logps/rejected": -474.0822448730469,
"loss": 0.6911,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0036013172939419746,
"rewards/margins": 0.0017371103167533875,
"rewards/rejected": -0.005338427610695362,
"step": 118
},
{
"epoch": 0.07772063025553107,
"grad_norm": 7.65494131386099,
"learning_rate": 1.159090909090909e-07,
"logits/chosen": -1.720550298690796,
"logits/rejected": -1.7268420457839966,
"logps/chosen": -511.4947814941406,
"logps/rejected": -518.540283203125,
"loss": 0.6886,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.002116803778335452,
"rewards/margins": 0.014177341014146805,
"rewards/rejected": -0.016294144093990326,
"step": 119
},
{
"epoch": 0.07837374479549351,
"grad_norm": 7.87167890173762,
"learning_rate": 1.1688311688311688e-07,
"logits/chosen": -1.6774556636810303,
"logits/rejected": -1.663172960281372,
"logps/chosen": -486.4458923339844,
"logps/rejected": -449.217529296875,
"loss": 0.6893,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.014286703430116177,
"rewards/margins": 0.00825961772352457,
"rewards/rejected": -0.022546321153640747,
"step": 120
},
{
"epoch": 0.07902685933545596,
"grad_norm": 6.91850162308767,
"learning_rate": 1.1785714285714285e-07,
"logits/chosen": -1.694111943244934,
"logits/rejected": -1.7094212770462036,
"logps/chosen": -547.680419921875,
"logps/rejected": -498.9229736328125,
"loss": 0.6889,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.003534555435180664,
"rewards/margins": 0.00987851619720459,
"rewards/rejected": -0.013413071632385254,
"step": 121
},
{
"epoch": 0.0796799738754184,
"grad_norm": 9.143990922203699,
"learning_rate": 1.1883116883116883e-07,
"logits/chosen": -1.6494725942611694,
"logits/rejected": -1.6601428985595703,
"logps/chosen": -484.3493347167969,
"logps/rejected": -487.05059814453125,
"loss": 0.6897,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.014209108427166939,
"rewards/margins": 0.0037215554621070623,
"rewards/rejected": -0.017930667847394943,
"step": 122
},
{
"epoch": 0.08033308841538085,
"grad_norm": 16.41656810337495,
"learning_rate": 1.198051948051948e-07,
"logits/chosen": -1.7612290382385254,
"logits/rejected": -1.7043516635894775,
"logps/chosen": -504.9853210449219,
"logps/rejected": -622.8348999023438,
"loss": 0.6853,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.017526667565107346,
"rewards/margins": 0.025080684572458267,
"rewards/rejected": -0.04260735213756561,
"step": 123
},
{
"epoch": 0.0809862029553433,
"grad_norm": 25.637817852859747,
"learning_rate": 1.207792207792208e-07,
"logits/chosen": -1.6937708854675293,
"logits/rejected": -1.6883544921875,
"logps/chosen": -384.4920959472656,
"logps/rejected": -507.6640930175781,
"loss": 0.6885,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.024008184671401978,
"rewards/margins": 0.0192131157964468,
"rewards/rejected": -0.04322130233049393,
"step": 124
},
{
"epoch": 0.08163931749530574,
"grad_norm": 22.15806914900568,
"learning_rate": 1.2175324675324674e-07,
"logits/chosen": -1.5668435096740723,
"logits/rejected": -1.5370216369628906,
"logps/chosen": -491.66278076171875,
"logps/rejected": -550.0761108398438,
"loss": 0.6886,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.030099207535386086,
"rewards/margins": 0.012964273802936077,
"rewards/rejected": -0.04306348040699959,
"step": 125
},
{
"epoch": 0.08229243203526819,
"grad_norm": 21.70535951365313,
"learning_rate": 1.2272727272727272e-07,
"logits/chosen": -1.718205451965332,
"logits/rejected": -1.7234244346618652,
"logps/chosen": -532.9760131835938,
"logps/rejected": -553.8611450195312,
"loss": 0.6876,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.028685756027698517,
"rewards/margins": 0.005702130496501923,
"rewards/rejected": -0.03438788652420044,
"step": 126
},
{
"epoch": 0.08294554657523064,
"grad_norm": 16.27604402538048,
"learning_rate": 1.237012987012987e-07,
"logits/chosen": -1.7105181217193604,
"logits/rejected": -1.7274971008300781,
"logps/chosen": -452.66326904296875,
"logps/rejected": -424.0040588378906,
"loss": 0.6882,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.023135703057050705,
"rewards/margins": 0.00477581936866045,
"rewards/rejected": -0.02791152149438858,
"step": 127
},
{
"epoch": 0.08359866111519308,
"grad_norm": 8.360473540141022,
"learning_rate": 1.2467532467532469e-07,
"logits/chosen": -1.7243715524673462,
"logits/rejected": -1.719879388809204,
"logps/chosen": -577.6858520507812,
"logps/rejected": -574.778564453125,
"loss": 0.6875,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.01625608652830124,
"rewards/margins": 0.01795053295791149,
"rewards/rejected": -0.03420662134885788,
"step": 128
},
{
"epoch": 0.08425177565515553,
"grad_norm": 7.426631483908922,
"learning_rate": 1.2564935064935064e-07,
"logits/chosen": -1.6296292543411255,
"logits/rejected": -1.645986557006836,
"logps/chosen": -514.0096435546875,
"logps/rejected": -507.37237548828125,
"loss": 0.6876,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.00854131206870079,
"rewards/margins": 0.007741453126072884,
"rewards/rejected": -0.016282765194773674,
"step": 129
},
{
"epoch": 0.08490489019511797,
"grad_norm": 12.61939534454806,
"learning_rate": 1.2662337662337662e-07,
"logits/chosen": -1.7628737688064575,
"logits/rejected": -1.784444808959961,
"logps/chosen": -515.55419921875,
"logps/rejected": -445.7431640625,
"loss": 0.6873,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.026726404204964638,
"rewards/margins": 0.007864664308726788,
"rewards/rejected": -0.03459106758236885,
"step": 130
},
{
"epoch": 0.08555800473508042,
"grad_norm": 6.940242974613805,
"learning_rate": 1.275974025974026e-07,
"logits/chosen": -1.723580002784729,
"logits/rejected": -1.7397780418395996,
"logps/chosen": -490.61956787109375,
"logps/rejected": -519.4144287109375,
"loss": 0.686,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.02750040590763092,
"rewards/margins": 0.028302742168307304,
"rewards/rejected": -0.055803146213293076,
"step": 131
},
{
"epoch": 0.08621111927504287,
"grad_norm": 9.176637689468706,
"learning_rate": 1.2857142857142855e-07,
"logits/chosen": -1.7846028804779053,
"logits/rejected": -1.7753349542617798,
"logps/chosen": -537.8035278320312,
"logps/rejected": -657.7811889648438,
"loss": 0.6844,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.039286255836486816,
"rewards/margins": 0.02002215012907982,
"rewards/rejected": -0.059308405965566635,
"step": 132
},
{
"epoch": 0.08686423381500531,
"grad_norm": 8.420310076033147,
"learning_rate": 1.2954545454545453e-07,
"logits/chosen": -1.7075575590133667,
"logits/rejected": -1.7227400541305542,
"logps/chosen": -505.66326904296875,
"logps/rejected": -475.65240478515625,
"loss": 0.6876,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.03044586256146431,
"rewards/margins": 0.011737257242202759,
"rewards/rejected": -0.04218312352895737,
"step": 133
},
{
"epoch": 0.08751734835496776,
"grad_norm": 17.816465764608324,
"learning_rate": 1.3051948051948052e-07,
"logits/chosen": -1.6720120906829834,
"logits/rejected": -1.6279743909835815,
"logps/chosen": -497.46038818359375,
"logps/rejected": -475.72442626953125,
"loss": 0.6897,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.02933063544332981,
"rewards/margins": 0.009470460936427116,
"rewards/rejected": -0.03880109637975693,
"step": 134
},
{
"epoch": 0.0881704628949302,
"grad_norm": 8.48602261690419,
"learning_rate": 1.314935064935065e-07,
"logits/chosen": -1.6978920698165894,
"logits/rejected": -1.7289825677871704,
"logps/chosen": -530.339111328125,
"logps/rejected": -493.29913330078125,
"loss": 0.6867,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.02855011634528637,
"rewards/margins": 0.0154922716319561,
"rewards/rejected": -0.04404238611459732,
"step": 135
},
{
"epoch": 0.08882357743489265,
"grad_norm": 9.366599761861622,
"learning_rate": 1.3246753246753245e-07,
"logits/chosen": -1.7024781703948975,
"logits/rejected": -1.6904100179672241,
"logps/chosen": -410.39215087890625,
"logps/rejected": -436.4376525878906,
"loss": 0.6859,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.012004725635051727,
"rewards/margins": 0.023064523935317993,
"rewards/rejected": -0.03506924957036972,
"step": 136
},
{
"epoch": 0.0894766919748551,
"grad_norm": 17.561997807750693,
"learning_rate": 1.3344155844155843e-07,
"logits/chosen": -1.6883275508880615,
"logits/rejected": -1.6822022199630737,
"logps/chosen": -470.7867736816406,
"logps/rejected": -476.1307067871094,
"loss": 0.6862,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.04406014457345009,
"rewards/margins": 0.020365647971630096,
"rewards/rejected": -0.06442578881978989,
"step": 137
},
{
"epoch": 0.09012980651481754,
"grad_norm": 6.311518850983097,
"learning_rate": 1.344155844155844e-07,
"logits/chosen": -1.6198405027389526,
"logits/rejected": -1.6445224285125732,
"logps/chosen": -577.4093627929688,
"logps/rejected": -524.4210815429688,
"loss": 0.6889,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.06365535408258438,
"rewards/margins": -0.0076623717322945595,
"rewards/rejected": -0.0559929758310318,
"step": 138
},
{
"epoch": 0.09078292105477999,
"grad_norm": 7.902620448969978,
"learning_rate": 1.353896103896104e-07,
"logits/chosen": -1.626230001449585,
"logits/rejected": -1.6832174062728882,
"logps/chosen": -452.2579650878906,
"logps/rejected": -457.4361267089844,
"loss": 0.6848,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.046750400215387344,
"rewards/margins": 0.018618889153003693,
"rewards/rejected": -0.06536928564310074,
"step": 139
},
{
"epoch": 0.09143603559474243,
"grad_norm": 9.061462627285719,
"learning_rate": 1.3636363636363635e-07,
"logits/chosen": -1.786929965019226,
"logits/rejected": -1.748306393623352,
"logps/chosen": -477.7335510253906,
"logps/rejected": -470.07623291015625,
"loss": 0.6823,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.05674988403916359,
"rewards/margins": 0.019451485946774483,
"rewards/rejected": -0.07620137184858322,
"step": 140
},
{
"epoch": 0.09208915013470488,
"grad_norm": 8.27691403538538,
"learning_rate": 1.3733766233766233e-07,
"logits/chosen": -1.7523396015167236,
"logits/rejected": -1.750685691833496,
"logps/chosen": -544.35888671875,
"logps/rejected": -523.8983154296875,
"loss": 0.6861,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.04826396703720093,
"rewards/margins": 0.019393280148506165,
"rewards/rejected": -0.06765724718570709,
"step": 141
},
{
"epoch": 0.09274226467466731,
"grad_norm": 40.62514167248459,
"learning_rate": 1.383116883116883e-07,
"logits/chosen": -1.6654901504516602,
"logits/rejected": -1.6239811182022095,
"logps/chosen": -479.25079345703125,
"logps/rejected": -476.67572021484375,
"loss": 0.6897,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.045496560633182526,
"rewards/margins": 0.007325804326683283,
"rewards/rejected": -0.05282236635684967,
"step": 142
},
{
"epoch": 0.09339537921462976,
"grad_norm": 28.906389537639807,
"learning_rate": 1.392857142857143e-07,
"logits/chosen": -1.805418848991394,
"logits/rejected": -1.7966678142547607,
"logps/chosen": -521.4232788085938,
"logps/rejected": -504.65008544921875,
"loss": 0.6839,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.043921858072280884,
"rewards/margins": 0.02651369199156761,
"rewards/rejected": -0.0704355537891388,
"step": 143
},
{
"epoch": 0.0940484937545922,
"grad_norm": 7.759027403093762,
"learning_rate": 1.4025974025974024e-07,
"logits/chosen": -1.808532953262329,
"logits/rejected": -1.7720710039138794,
"logps/chosen": -530.4207763671875,
"logps/rejected": -543.4842529296875,
"loss": 0.6823,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.058577775955200195,
"rewards/margins": 0.01986054703593254,
"rewards/rejected": -0.07843831926584244,
"step": 144
},
{
"epoch": 0.09470160829455465,
"grad_norm": 19.65781688946105,
"learning_rate": 1.4123376623376622e-07,
"logits/chosen": -1.7204667329788208,
"logits/rejected": -1.6994764804840088,
"logps/chosen": -600.3792724609375,
"logps/rejected": -596.1019897460938,
"loss": 0.6858,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.07191239297389984,
"rewards/margins": 0.02992035634815693,
"rewards/rejected": -0.10183274745941162,
"step": 145
},
{
"epoch": 0.0953547228345171,
"grad_norm": 27.083668342567048,
"learning_rate": 1.422077922077922e-07,
"logits/chosen": -1.6221792697906494,
"logits/rejected": -1.6598668098449707,
"logps/chosen": -526.0675048828125,
"logps/rejected": -521.8135986328125,
"loss": 0.6908,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.042592283338308334,
"rewards/margins": 0.012567641213536263,
"rewards/rejected": -0.055159930139780045,
"step": 146
},
{
"epoch": 0.09600783737447954,
"grad_norm": 8.808375433626706,
"learning_rate": 1.4318181818181818e-07,
"logits/chosen": -1.8164383172988892,
"logits/rejected": -1.8282561302185059,
"logps/chosen": -614.2815551757812,
"logps/rejected": -643.0169067382812,
"loss": 0.6865,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.07687513530254364,
"rewards/margins": 0.012751526199281216,
"rewards/rejected": -0.08962665498256683,
"step": 147
},
{
"epoch": 0.09666095191444199,
"grad_norm": 7.09253863840217,
"learning_rate": 1.4415584415584414e-07,
"logits/chosen": -1.6832506656646729,
"logits/rejected": -1.7021249532699585,
"logps/chosen": -498.71942138671875,
"logps/rejected": -501.34869384765625,
"loss": 0.6838,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.054647862911224365,
"rewards/margins": 0.026497341692447662,
"rewards/rejected": -0.08114521205425262,
"step": 148
},
{
"epoch": 0.09731406645440444,
"grad_norm": 16.781059240998136,
"learning_rate": 1.4512987012987012e-07,
"logits/chosen": -1.715110182762146,
"logits/rejected": -1.7118535041809082,
"logps/chosen": -474.4345703125,
"logps/rejected": -454.8081970214844,
"loss": 0.6855,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.062017131596803665,
"rewards/margins": 0.004093126859515905,
"rewards/rejected": -0.06611025333404541,
"step": 149
},
{
"epoch": 0.09796718099436688,
"grad_norm": 20.190214518687835,
"learning_rate": 1.461038961038961e-07,
"logits/chosen": -1.6403260231018066,
"logits/rejected": -1.601685881614685,
"logps/chosen": -516.1248168945312,
"logps/rejected": -536.0068969726562,
"loss": 0.6845,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.05097360908985138,
"rewards/margins": 0.021962041035294533,
"rewards/rejected": -0.07293565571308136,
"step": 150
},
{
"epoch": 0.09862029553432933,
"grad_norm": 12.205029148852612,
"learning_rate": 1.4707792207792208e-07,
"logits/chosen": -1.6476508378982544,
"logits/rejected": -1.7027806043624878,
"logps/chosen": -488.97259521484375,
"logps/rejected": -461.6432800292969,
"loss": 0.681,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.03291236236691475,
"rewards/margins": 0.01708255708217621,
"rewards/rejected": -0.04999491944909096,
"step": 151
},
{
"epoch": 0.09927341007429177,
"grad_norm": 22.040695719974778,
"learning_rate": 1.4805194805194803e-07,
"logits/chosen": -1.7001389265060425,
"logits/rejected": -1.7121036052703857,
"logps/chosen": -473.0397033691406,
"logps/rejected": -513.0086059570312,
"loss": 0.6773,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.05476941913366318,
"rewards/margins": 0.05117000639438629,
"rewards/rejected": -0.10593942552804947,
"step": 152
},
{
"epoch": 0.09992652461425422,
"grad_norm": 17.736405376818148,
"learning_rate": 1.4902597402597404e-07,
"logits/chosen": -1.7151176929473877,
"logits/rejected": -1.6587436199188232,
"logps/chosen": -528.3651733398438,
"logps/rejected": -519.2221069335938,
"loss": 0.6872,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.050955355167388916,
"rewards/margins": 0.02759256586432457,
"rewards/rejected": -0.07854791730642319,
"step": 153
},
{
"epoch": 0.10057963915421667,
"grad_norm": 20.0575565981101,
"learning_rate": 1.5e-07,
"logits/chosen": -1.7852433919906616,
"logits/rejected": -1.7886905670166016,
"logps/chosen": -482.718017578125,
"logps/rejected": -482.28759765625,
"loss": 0.6837,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.07595521211624146,
"rewards/margins": 0.016735542565584183,
"rewards/rejected": -0.09269075095653534,
"step": 154
},
{
"epoch": 0.10123275369417911,
"grad_norm": 6.032126948811811,
"learning_rate": 1.499998048075819e-07,
"logits/chosen": -1.649161696434021,
"logits/rejected": -1.6406943798065186,
"logps/chosen": -509.6943359375,
"logps/rejected": -504.8258361816406,
"loss": 0.6872,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.07708139717578888,
"rewards/margins": 0.013607650063931942,
"rewards/rejected": -0.0906890481710434,
"step": 155
},
{
"epoch": 0.10188586823414156,
"grad_norm": 8.923286598087802,
"learning_rate": 1.4999921923134367e-07,
"logits/chosen": -1.7759987115859985,
"logits/rejected": -1.731602668762207,
"logps/chosen": -508.0286865234375,
"logps/rejected": -549.663330078125,
"loss": 0.6864,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.04344085976481438,
"rewards/margins": 0.024824578315019608,
"rewards/rejected": -0.06826544553041458,
"step": 156
},
{
"epoch": 0.102538982774104,
"grad_norm": 18.440082543584566,
"learning_rate": 1.499982432743333e-07,
"logits/chosen": -1.7433996200561523,
"logits/rejected": -1.7404325008392334,
"logps/chosen": -483.1097717285156,
"logps/rejected": -518.4947509765625,
"loss": 0.682,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.07219300419092178,
"rewards/margins": 0.0072326865047216415,
"rewards/rejected": -0.07942568510770798,
"step": 157
},
{
"epoch": 0.10319209731406645,
"grad_norm": 26.950148340749994,
"learning_rate": 1.4999687694163071e-07,
"logits/chosen": -1.6927975416183472,
"logits/rejected": -1.7593554258346558,
"logps/chosen": -587.9935302734375,
"logps/rejected": -547.7039184570312,
"loss": 0.6831,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.05381292477250099,
"rewards/margins": 0.02101278305053711,
"rewards/rejected": -0.0748257115483284,
"step": 158
},
{
"epoch": 0.1038452118540289,
"grad_norm": 7.542663745310761,
"learning_rate": 1.499951202403479e-07,
"logits/chosen": -1.7588621377944946,
"logits/rejected": -1.703225016593933,
"logps/chosen": -505.0411376953125,
"logps/rejected": -599.8470458984375,
"loss": 0.6763,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.05922378972172737,
"rewards/margins": 0.053594160825014114,
"rewards/rejected": -0.11281795799732208,
"step": 159
},
{
"epoch": 0.10449832639399134,
"grad_norm": 17.13212774750112,
"learning_rate": 1.4999297317962876e-07,
"logits/chosen": -1.7104802131652832,
"logits/rejected": -1.7329843044281006,
"logps/chosen": -514.36279296875,
"logps/rejected": -533.5743408203125,
"loss": 0.6821,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.07126364856958389,
"rewards/margins": 0.031774699687957764,
"rewards/rejected": -0.10303835570812225,
"step": 160
},
{
"epoch": 0.10515144093395379,
"grad_norm": 10.645728177870987,
"learning_rate": 1.4999043577064894e-07,
"logits/chosen": -1.826465129852295,
"logits/rejected": -1.8719959259033203,
"logps/chosen": -616.4337768554688,
"logps/rejected": -609.4326171875,
"loss": 0.6779,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.07613222301006317,
"rewards/margins": 0.02693980187177658,
"rewards/rejected": -0.10307201743125916,
"step": 161
},
{
"epoch": 0.10580455547391623,
"grad_norm": 31.5882679445136,
"learning_rate": 1.4998750802661605e-07,
"logits/chosen": -1.727925419807434,
"logits/rejected": -1.724989414215088,
"logps/chosen": -599.0576171875,
"logps/rejected": -563.6280517578125,
"loss": 0.683,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.08880521357059479,
"rewards/margins": 0.025629587471485138,
"rewards/rejected": -0.11443479359149933,
"step": 162
},
{
"epoch": 0.10645767001387868,
"grad_norm": 12.592298356038635,
"learning_rate": 1.4998418996276933e-07,
"logits/chosen": -1.738313913345337,
"logits/rejected": -1.753507375717163,
"logps/chosen": -579.961669921875,
"logps/rejected": -594.0586547851562,
"loss": 0.6819,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.05988356098532677,
"rewards/margins": 0.03945886343717575,
"rewards/rejected": -0.09934242069721222,
"step": 163
},
{
"epoch": 0.10711078455384113,
"grad_norm": 19.46009999388665,
"learning_rate": 1.499804815963798e-07,
"logits/chosen": -1.7987678050994873,
"logits/rejected": -1.807891845703125,
"logps/chosen": -550.672119140625,
"logps/rejected": -561.8424072265625,
"loss": 0.677,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.040339332073926926,
"rewards/margins": 0.04034976288676262,
"rewards/rejected": -0.08068908751010895,
"step": 164
},
{
"epoch": 0.10776389909380357,
"grad_norm": 27.510358333266502,
"learning_rate": 1.4997638294674996e-07,
"logits/chosen": -1.7597366571426392,
"logits/rejected": -1.7664936780929565,
"logps/chosen": -517.5831909179688,
"logps/rejected": -524.0874633789062,
"loss": 0.6807,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.10509812831878662,
"rewards/margins": 0.02205245941877365,
"rewards/rejected": -0.12715059518814087,
"step": 165
},
{
"epoch": 0.10841701363376602,
"grad_norm": 26.359728514103736,
"learning_rate": 1.499718940352138e-07,
"logits/chosen": -1.660881519317627,
"logits/rejected": -1.6824941635131836,
"logps/chosen": -526.3524169921875,
"logps/rejected": -505.15814208984375,
"loss": 0.6795,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.09450910985469818,
"rewards/margins": 0.04056902229785919,
"rewards/rejected": -0.13507813215255737,
"step": 166
},
{
"epoch": 0.10907012817372846,
"grad_norm": 13.379229704663794,
"learning_rate": 1.499670148851367e-07,
"logits/chosen": -1.6919353008270264,
"logits/rejected": -1.7081630229949951,
"logps/chosen": -462.4615783691406,
"logps/rejected": -441.3831787109375,
"loss": 0.6793,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.047392651438713074,
"rewards/margins": 0.02708030305802822,
"rewards/rejected": -0.07447294890880585,
"step": 167
},
{
"epoch": 0.10972324271369091,
"grad_norm": 23.765383175487347,
"learning_rate": 1.4996174552191534e-07,
"logits/chosen": -1.6622581481933594,
"logits/rejected": -1.679819941520691,
"logps/chosen": -464.3515930175781,
"logps/rejected": -449.2407531738281,
"loss": 0.6781,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.059018295258283615,
"rewards/margins": 0.03328193724155426,
"rewards/rejected": -0.09230024367570877,
"step": 168
},
{
"epoch": 0.11037635725365336,
"grad_norm": 7.18145864960063,
"learning_rate": 1.4995608597297736e-07,
"logits/chosen": -1.7335408926010132,
"logits/rejected": -1.7119437456130981,
"logps/chosen": -517.2203979492188,
"logps/rejected": -603.5662841796875,
"loss": 0.6748,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.06651943922042847,
"rewards/margins": 0.05973542109131813,
"rewards/rejected": -0.1262548714876175,
"step": 169
},
{
"epoch": 0.1110294717936158,
"grad_norm": 33.768142398887186,
"learning_rate": 1.4995003626778149e-07,
"logits/chosen": -1.7551608085632324,
"logits/rejected": -1.718156099319458,
"logps/chosen": -503.027099609375,
"logps/rejected": -557.4558715820312,
"loss": 0.685,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.06958004832267761,
"rewards/margins": 0.021374018862843513,
"rewards/rejected": -0.09095406532287598,
"step": 170
},
{
"epoch": 0.11168258633357825,
"grad_norm": 13.332558234776263,
"learning_rate": 1.4994359643781725e-07,
"logits/chosen": -1.6308660507202148,
"logits/rejected": -1.6401119232177734,
"logps/chosen": -491.20263671875,
"logps/rejected": -492.4757995605469,
"loss": 0.6805,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.08171894401311874,
"rewards/margins": 0.03456515446305275,
"rewards/rejected": -0.1162840873003006,
"step": 171
},
{
"epoch": 0.1123357008735407,
"grad_norm": 11.374539796821198,
"learning_rate": 1.4993676651660479e-07,
"logits/chosen": -1.7294381856918335,
"logits/rejected": -1.7581063508987427,
"logps/chosen": -510.65618896484375,
"logps/rejected": -494.7225341796875,
"loss": 0.6791,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.07346140593290329,
"rewards/margins": 0.016454335302114487,
"rewards/rejected": -0.08991573750972748,
"step": 172
},
{
"epoch": 0.11298881541350314,
"grad_norm": 10.782302320724707,
"learning_rate": 1.4992954653969473e-07,
"logits/chosen": -1.7590150833129883,
"logits/rejected": -1.7065110206604004,
"logps/chosen": -579.6332397460938,
"logps/rejected": -642.480224609375,
"loss": 0.6756,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.16745489835739136,
"rewards/margins": 0.04030359163880348,
"rewards/rejected": -0.20775848627090454,
"step": 173
},
{
"epoch": 0.11364192995346559,
"grad_norm": 30.030650307666996,
"learning_rate": 1.4992193654466804e-07,
"logits/chosen": -1.7184665203094482,
"logits/rejected": -1.751639485359192,
"logps/chosen": -505.05078125,
"logps/rejected": -476.3111572265625,
"loss": 0.6774,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.07360526919364929,
"rewards/margins": 0.02522566169500351,
"rewards/rejected": -0.0988309308886528,
"step": 174
},
{
"epoch": 0.11429504449342803,
"grad_norm": 6.801707032164358,
"learning_rate": 1.4991393657113566e-07,
"logits/chosen": -1.7240628004074097,
"logits/rejected": -1.693426489830017,
"logps/chosen": -513.940673828125,
"logps/rejected": -593.656494140625,
"loss": 0.6822,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.09224756807088852,
"rewards/margins": 0.05458322912454605,
"rewards/rejected": -0.14683078229427338,
"step": 175
},
{
"epoch": 0.11494815903339048,
"grad_norm": 6.667305623368847,
"learning_rate": 1.499055466607386e-07,
"logits/chosen": -1.6916465759277344,
"logits/rejected": -1.7375673055648804,
"logps/chosen": -533.5892944335938,
"logps/rejected": -507.4823913574219,
"loss": 0.675,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.080963134765625,
"rewards/margins": 0.053236886858940125,
"rewards/rejected": -0.13420002162456512,
"step": 176
},
{
"epoch": 0.11560127357335293,
"grad_norm": 6.771588986223236,
"learning_rate": 1.498967668571474e-07,
"logits/chosen": -1.735141396522522,
"logits/rejected": -1.7742522954940796,
"logps/chosen": -510.868408203125,
"logps/rejected": -452.6504211425781,
"loss": 0.6749,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.10077373683452606,
"rewards/margins": 0.003773924894630909,
"rewards/rejected": -0.1045476570725441,
"step": 177
},
{
"epoch": 0.11625438811331537,
"grad_norm": 9.953482315244722,
"learning_rate": 1.4988759720606207e-07,
"logits/chosen": -1.7570782899856567,
"logits/rejected": -1.8162400722503662,
"logps/chosen": -554.0537719726562,
"logps/rejected": -541.5122680664062,
"loss": 0.6801,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.10425091534852982,
"rewards/margins": 0.060477737337350845,
"rewards/rejected": -0.16472867131233215,
"step": 178
},
{
"epoch": 0.11690750265327782,
"grad_norm": 5.973916986801873,
"learning_rate": 1.4987803775521184e-07,
"logits/chosen": -1.690962791442871,
"logits/rejected": -1.689450979232788,
"logps/chosen": -643.9844360351562,
"logps/rejected": -589.926025390625,
"loss": 0.6829,
"rewards/accuracies": 0.34375,
"rewards/chosen": -0.12561464309692383,
"rewards/margins": -0.008268720470368862,
"rewards/rejected": -0.11734593659639359,
"step": 179
},
{
"epoch": 0.11756061719324026,
"grad_norm": 14.891949152028998,
"learning_rate": 1.4986808855435498e-07,
"logits/chosen": -1.6953084468841553,
"logits/rejected": -1.6965618133544922,
"logps/chosen": -465.82086181640625,
"logps/rejected": -438.37799072265625,
"loss": 0.6767,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.08036669343709946,
"rewards/margins": 0.017360081896185875,
"rewards/rejected": -0.09772677719593048,
"step": 180
},
{
"epoch": 0.11821373173320271,
"grad_norm": 16.11153067050427,
"learning_rate": 1.498577496552783e-07,
"logits/chosen": -1.7526249885559082,
"logits/rejected": -1.7759355306625366,
"logps/chosen": -482.6202697753906,
"logps/rejected": -492.8947448730469,
"loss": 0.6659,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.09437385201454163,
"rewards/margins": 0.037018630653619766,
"rewards/rejected": -0.1313924789428711,
"step": 181
},
{
"epoch": 0.11886684627316516,
"grad_norm": 22.729237496040554,
"learning_rate": 1.4984702111179715e-07,
"logits/chosen": -1.6132816076278687,
"logits/rejected": -1.655545949935913,
"logps/chosen": -527.8369750976562,
"logps/rejected": -527.8175048828125,
"loss": 0.6727,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.0954374372959137,
"rewards/margins": 0.05533631891012192,
"rewards/rejected": -0.1507737785577774,
"step": 182
},
{
"epoch": 0.1195199608131276,
"grad_norm": 26.231353893622398,
"learning_rate": 1.4983590297975505e-07,
"logits/chosen": -1.6354296207427979,
"logits/rejected": -1.6777453422546387,
"logps/chosen": -504.5555419921875,
"logps/rejected": -538.9359741210938,
"loss": 0.6843,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.12271232157945633,
"rewards/margins": 0.04234550893306732,
"rewards/rejected": -0.16505783796310425,
"step": 183
},
{
"epoch": 0.12017307535309005,
"grad_norm": 7.114753478180882,
"learning_rate": 1.498243953170233e-07,
"logits/chosen": -1.69046950340271,
"logits/rejected": -1.7043442726135254,
"logps/chosen": -530.9853515625,
"logps/rejected": -529.0479125976562,
"loss": 0.677,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.06343412399291992,
"rewards/margins": 0.035453617572784424,
"rewards/rejected": -0.09888774156570435,
"step": 184
},
{
"epoch": 0.1208261898930525,
"grad_norm": 21.787798884018414,
"learning_rate": 1.498124981835008e-07,
"logits/chosen": -1.7766292095184326,
"logits/rejected": -1.74836266040802,
"logps/chosen": -475.3532409667969,
"logps/rejected": -489.02923583984375,
"loss": 0.6776,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.07969984412193298,
"rewards/margins": 0.041919078677892685,
"rewards/rejected": -0.12161892652511597,
"step": 185
},
{
"epoch": 0.12147930443301494,
"grad_norm": 23.709784617273577,
"learning_rate": 1.4980021164111366e-07,
"logits/chosen": -1.6516242027282715,
"logits/rejected": -1.673815369606018,
"logps/chosen": -455.3662109375,
"logps/rejected": -492.9649963378906,
"loss": 0.677,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.14689013361930847,
"rewards/margins": 0.02547987923026085,
"rewards/rejected": -0.1723700314760208,
"step": 186
},
{
"epoch": 0.12213241897297739,
"grad_norm": 12.348486008384063,
"learning_rate": 1.4978753575381498e-07,
"logits/chosen": -1.7523754835128784,
"logits/rejected": -1.745179533958435,
"logps/chosen": -503.9389343261719,
"logps/rejected": -510.8143615722656,
"loss": 0.6702,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.11973577737808228,
"rewards/margins": 0.05588344484567642,
"rewards/rejected": -0.1756192147731781,
"step": 187
},
{
"epoch": 0.12278553351293983,
"grad_norm": 13.48324499414794,
"learning_rate": 1.4977447058758439e-07,
"logits/chosen": -1.7630393505096436,
"logits/rejected": -1.7760398387908936,
"logps/chosen": -571.8206787109375,
"logps/rejected": -541.6646118164062,
"loss": 0.6801,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.11507945507764816,
"rewards/margins": 0.029867036268115044,
"rewards/rejected": -0.14494650065898895,
"step": 188
},
{
"epoch": 0.12343864805290228,
"grad_norm": 10.415485511799975,
"learning_rate": 1.4976101621042783e-07,
"logits/chosen": -1.684414267539978,
"logits/rejected": -1.6862590312957764,
"logps/chosen": -535.666015625,
"logps/rejected": -548.4473876953125,
"loss": 0.6756,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.1297360062599182,
"rewards/margins": 0.05931150168180466,
"rewards/rejected": -0.18904751539230347,
"step": 189
},
{
"epoch": 0.12409176259286472,
"grad_norm": 6.350273320282756,
"learning_rate": 1.4974717269237708e-07,
"logits/chosen": -1.6381902694702148,
"logits/rejected": -1.6570154428482056,
"logps/chosen": -497.8411560058594,
"logps/rejected": -489.9413757324219,
"loss": 0.6632,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.11118555068969727,
"rewards/margins": 0.06526104360818863,
"rewards/rejected": -0.17644661664962769,
"step": 190
},
{
"epoch": 0.12474487713282717,
"grad_norm": 6.038127824462166,
"learning_rate": 1.4973294010548946e-07,
"logits/chosen": -1.6311917304992676,
"logits/rejected": -1.6577972173690796,
"logps/chosen": -518.0223388671875,
"logps/rejected": -520.6466674804688,
"loss": 0.677,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.15784503519535065,
"rewards/margins": 0.01725524663925171,
"rewards/rejected": -0.17510026693344116,
"step": 191
},
{
"epoch": 0.12539799167278962,
"grad_norm": 26.218558847948582,
"learning_rate": 1.4971831852384745e-07,
"logits/chosen": -1.6489263772964478,
"logits/rejected": -1.6216522455215454,
"logps/chosen": -570.0574951171875,
"logps/rejected": -592.376708984375,
"loss": 0.6842,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.17664462327957153,
"rewards/margins": 0.014460130594670773,
"rewards/rejected": -0.1911047399044037,
"step": 192
},
{
"epoch": 0.12605110621275206,
"grad_norm": 14.358619994205673,
"learning_rate": 1.497033080235583e-07,
"logits/chosen": -1.754329800605774,
"logits/rejected": -1.7733534574508667,
"logps/chosen": -515.3424682617188,
"logps/rejected": -505.7689208984375,
"loss": 0.6687,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.12092112749814987,
"rewards/margins": 0.024929339066147804,
"rewards/rejected": -0.14585046470165253,
"step": 193
},
{
"epoch": 0.1267042207527145,
"grad_norm": 17.358492156799695,
"learning_rate": 1.4968790868275365e-07,
"logits/chosen": -1.666093349456787,
"logits/rejected": -1.6654276847839355,
"logps/chosen": -510.59320068359375,
"logps/rejected": -497.73431396484375,
"loss": 0.6717,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.13687899708747864,
"rewards/margins": 0.009451786056160927,
"rewards/rejected": -0.1463308036327362,
"step": 194
},
{
"epoch": 0.12735733529267695,
"grad_norm": 33.350981012916925,
"learning_rate": 1.4967212058158908e-07,
"logits/chosen": -1.7516664266586304,
"logits/rejected": -1.7618868350982666,
"logps/chosen": -574.4312744140625,
"logps/rejected": -562.3271484375,
"loss": 0.687,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.1302211880683899,
"rewards/margins": 0.03780483454465866,
"rewards/rejected": -0.16802600026130676,
"step": 195
},
{
"epoch": 0.1280104498326394,
"grad_norm": 12.460824021943772,
"learning_rate": 1.4965594380224373e-07,
"logits/chosen": -1.6539489030838013,
"logits/rejected": -1.6902060508728027,
"logps/chosen": -540.94873046875,
"logps/rejected": -570.3787231445312,
"loss": 0.6737,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.16215386986732483,
"rewards/margins": 0.05639547482132912,
"rewards/rejected": -0.21854937076568604,
"step": 196
},
{
"epoch": 0.12866356437260185,
"grad_norm": 7.277120819141452,
"learning_rate": 1.4963937842891983e-07,
"logits/chosen": -1.7623552083969116,
"logits/rejected": -1.777740240097046,
"logps/chosen": -562.0401611328125,
"logps/rejected": -549.0805053710938,
"loss": 0.6625,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.15102756023406982,
"rewards/margins": 0.02262779325246811,
"rewards/rejected": -0.17365534603595734,
"step": 197
},
{
"epoch": 0.1293166789125643,
"grad_norm": 6.4071063075287045,
"learning_rate": 1.4962242454784235e-07,
"logits/chosen": -1.7132699489593506,
"logits/rejected": -1.6834791898727417,
"logps/chosen": -473.0800476074219,
"logps/rejected": -574.1148681640625,
"loss": 0.6639,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.13342300057411194,
"rewards/margins": 0.09166872501373291,
"rewards/rejected": -0.22509171068668365,
"step": 198
},
{
"epoch": 0.12996979345252674,
"grad_norm": 16.16942335499691,
"learning_rate": 1.4960508224725845e-07,
"logits/chosen": -1.697445273399353,
"logits/rejected": -1.737441897392273,
"logps/chosen": -491.8898010253906,
"logps/rejected": -465.43133544921875,
"loss": 0.6753,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.12092840671539307,
"rewards/margins": 0.01965322345495224,
"rewards/rejected": -0.1405816376209259,
"step": 199
},
{
"epoch": 0.13062290799248918,
"grad_norm": 9.236224313907128,
"learning_rate": 1.495873516174371e-07,
"logits/chosen": -1.818574070930481,
"logits/rejected": -1.7804887294769287,
"logps/chosen": -583.37548828125,
"logps/rejected": -581.4929809570312,
"loss": 0.6703,
"rewards/accuracies": 0.375,
"rewards/chosen": -0.20678864419460297,
"rewards/margins": 0.013942277058959007,
"rewards/rejected": -0.22073093056678772,
"step": 200
},
{
"epoch": 0.13062290799248918,
"eval_logits/chosen": -1.7592895030975342,
"eval_logits/rejected": -1.7686011791229248,
"eval_logps/chosen": -524.639404296875,
"eval_logps/rejected": -522.0521240234375,
"eval_loss": 0.6713127493858337,
"eval_rewards/accuracies": 0.6380000114440918,
"eval_rewards/chosen": -0.1429399996995926,
"eval_rewards/margins": 0.05518652871251106,
"eval_rewards/rejected": -0.19812652468681335,
"eval_runtime": 300.5974,
"eval_samples_per_second": 13.307,
"eval_steps_per_second": 0.832,
"step": 200
},
{
"epoch": 0.13127602253245163,
"grad_norm": 8.660547603879296,
"learning_rate": 1.4956923275066855e-07,
"logits/chosen": -1.7169371843338013,
"logits/rejected": -1.7529581785202026,
"logps/chosen": -491.5006408691406,
"logps/rejected": -438.540283203125,
"loss": 0.6728,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.12306279689073563,
"rewards/margins": 0.035101909190416336,
"rewards/rejected": -0.15816472470760345,
"step": 201
},
{
"epoch": 0.13192913707241408,
"grad_norm": 9.941469801759618,
"learning_rate": 1.4955072574126383e-07,
"logits/chosen": -1.6325197219848633,
"logits/rejected": -1.6774578094482422,
"logps/chosen": -501.1746826171875,
"logps/rejected": -474.1672668457031,
"loss": 0.6778,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.1591545194387436,
"rewards/margins": -0.0037236525677144527,
"rewards/rejected": -0.155430868268013,
"step": 202
},
{
"epoch": 0.13258225161237652,
"grad_norm": 6.629751952322596,
"learning_rate": 1.4953183068555444e-07,
"logits/chosen": -1.6159188747406006,
"logits/rejected": -1.655371069908142,
"logps/chosen": -511.7205810546875,
"logps/rejected": -497.550537109375,
"loss": 0.6755,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.17229245603084564,
"rewards/margins": 0.036124035716056824,
"rewards/rejected": -0.20841647684574127,
"step": 203
},
{
"epoch": 0.13323536615233897,
"grad_norm": 14.86037901564549,
"learning_rate": 1.4951254768189153e-07,
"logits/chosen": -1.7175514698028564,
"logits/rejected": -1.7305848598480225,
"logps/chosen": -521.8203735351562,
"logps/rejected": -473.41363525390625,
"loss": 0.6698,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.1745811253786087,
"rewards/margins": 0.05103900283575058,
"rewards/rejected": -0.22562013566493988,
"step": 204
},
{
"epoch": 0.13388848069230141,
"grad_norm": 9.81928586229961,
"learning_rate": 1.4949287683064572e-07,
"logits/chosen": -1.7592101097106934,
"logits/rejected": -1.7103134393692017,
"logps/chosen": -506.8877258300781,
"logps/rejected": -512.57568359375,
"loss": 0.6752,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.12894892692565918,
"rewards/margins": 0.058204442262649536,
"rewards/rejected": -0.1871533840894699,
"step": 205
},
{
"epoch": 0.13454159523226386,
"grad_norm": 12.70868918654191,
"learning_rate": 1.4947281823420636e-07,
"logits/chosen": -1.74403715133667,
"logits/rejected": -1.7773162126541138,
"logps/chosen": -549.9198608398438,
"logps/rejected": -537.7161865234375,
"loss": 0.6726,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.15765821933746338,
"rewards/margins": 0.024023467674851418,
"rewards/rejected": -0.18168169260025024,
"step": 206
},
{
"epoch": 0.1351947097722263,
"grad_norm": 9.464198406615232,
"learning_rate": 1.4945237199698105e-07,
"logits/chosen": -1.7218544483184814,
"logits/rejected": -1.7245031595230103,
"logps/chosen": -575.4227294921875,
"logps/rejected": -613.4913330078125,
"loss": 0.6628,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2187643200159073,
"rewards/margins": 0.05670395866036415,
"rewards/rejected": -0.27546826004981995,
"step": 207
},
{
"epoch": 0.13584782431218875,
"grad_norm": 40.46016052410099,
"learning_rate": 1.4943153822539518e-07,
"logits/chosen": -1.7167556285858154,
"logits/rejected": -1.7805320024490356,
"logps/chosen": -536.693603515625,
"logps/rejected": -531.85986328125,
"loss": 0.6732,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.18354004621505737,
"rewards/margins": 0.048775896430015564,
"rewards/rejected": -0.23231592774391174,
"step": 208
},
{
"epoch": 0.1365009388521512,
"grad_norm": 22.288492214742288,
"learning_rate": 1.4941031702789123e-07,
"logits/chosen": -1.6213892698287964,
"logits/rejected": -1.63565993309021,
"logps/chosen": -565.4679565429688,
"logps/rejected": -557.27685546875,
"loss": 0.6666,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.13010086119174957,
"rewards/margins": 0.07792194187641144,
"rewards/rejected": -0.20802278816699982,
"step": 209
},
{
"epoch": 0.13715405339211365,
"grad_norm": 15.197653036128015,
"learning_rate": 1.4938870851492834e-07,
"logits/chosen": -1.6974363327026367,
"logits/rejected": -1.7302565574645996,
"logps/chosen": -496.8112487792969,
"logps/rejected": -477.2409362792969,
"loss": 0.6659,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.15533235669136047,
"rewards/margins": 0.022654909640550613,
"rewards/rejected": -0.1779872477054596,
"step": 210
},
{
"epoch": 0.1378071679320761,
"grad_norm": 6.3314279991311055,
"learning_rate": 1.4936671279898162e-07,
"logits/chosen": -1.7678776979446411,
"logits/rejected": -1.8228353261947632,
"logps/chosen": -473.0333557128906,
"logps/rejected": -480.84259033203125,
"loss": 0.6704,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.2336454689502716,
"rewards/margins": 0.01017429493367672,
"rewards/rejected": -0.24381977319717407,
"step": 211
},
{
"epoch": 0.13846028247203854,
"grad_norm": 33.62952471238388,
"learning_rate": 1.493443299945417e-07,
"logits/chosen": -1.674032211303711,
"logits/rejected": -1.6666409969329834,
"logps/chosen": -487.1645202636719,
"logps/rejected": -511.31976318359375,
"loss": 0.6672,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.17207323014736176,
"rewards/margins": 0.0362030491232872,
"rewards/rejected": -0.20827627182006836,
"step": 212
},
{
"epoch": 0.13911339701200098,
"grad_norm": 17.559552038629228,
"learning_rate": 1.4932156021811393e-07,
"logits/chosen": -1.6960467100143433,
"logits/rejected": -1.6814875602722168,
"logps/chosen": -448.228515625,
"logps/rejected": -523.8681640625,
"loss": 0.6702,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.19846728444099426,
"rewards/margins": 0.07104050368070602,
"rewards/rejected": -0.2695077657699585,
"step": 213
},
{
"epoch": 0.13976651155196343,
"grad_norm": 21.300381341216802,
"learning_rate": 1.492984035882181e-07,
"logits/chosen": -1.7540203332901,
"logits/rejected": -1.6938629150390625,
"logps/chosen": -510.3170166015625,
"logps/rejected": -555.1364135742188,
"loss": 0.6838,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.20609605312347412,
"rewards/margins": 0.04784277081489563,
"rewards/rejected": -0.25393882393836975,
"step": 214
},
{
"epoch": 0.14041962609192588,
"grad_norm": 18.999047527818195,
"learning_rate": 1.4927486022538743e-07,
"logits/chosen": -1.6708929538726807,
"logits/rejected": -1.712494969367981,
"logps/chosen": -605.5905151367188,
"logps/rejected": -571.502685546875,
"loss": 0.6654,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.21136337518692017,
"rewards/margins": 0.03826094791293144,
"rewards/rejected": -0.24962429702281952,
"step": 215
},
{
"epoch": 0.14107274063188832,
"grad_norm": 21.481084161977915,
"learning_rate": 1.4925093025216822e-07,
"logits/chosen": -1.6800730228424072,
"logits/rejected": -1.6440744400024414,
"logps/chosen": -561.3792724609375,
"logps/rejected": -543.5944213867188,
"loss": 0.6691,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.18766431510448456,
"rewards/margins": 0.08474057167768478,
"rewards/rejected": -0.27240487933158875,
"step": 216
},
{
"epoch": 0.14172585517185077,
"grad_norm": 36.19470917799039,
"learning_rate": 1.4922661379311916e-07,
"logits/chosen": -1.7101688385009766,
"logits/rejected": -1.7070766687393188,
"logps/chosen": -513.0838012695312,
"logps/rejected": -580.56982421875,
"loss": 0.6719,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.19527553021907806,
"rewards/margins": 0.07064682245254517,
"rewards/rejected": -0.26592230796813965,
"step": 217
},
{
"epoch": 0.1423789697118132,
"grad_norm": 19.07151886221254,
"learning_rate": 1.4920191097481055e-07,
"logits/chosen": -1.6426142454147339,
"logits/rejected": -1.6799873113632202,
"logps/chosen": -544.11083984375,
"logps/rejected": -474.7537536621094,
"loss": 0.6781,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.24504226446151733,
"rewards/margins": -0.005615321919322014,
"rewards/rejected": -0.23942697048187256,
"step": 218
},
{
"epoch": 0.14303208425177566,
"grad_norm": 16.177094066689747,
"learning_rate": 1.4917682192582382e-07,
"logits/chosen": -1.6813910007476807,
"logits/rejected": -1.7110719680786133,
"logps/chosen": -583.3339233398438,
"logps/rejected": -526.6600341796875,
"loss": 0.6832,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2589135766029358,
"rewards/margins": -0.009151825681328773,
"rewards/rejected": -0.24976176023483276,
"step": 219
},
{
"epoch": 0.1436851987917381,
"grad_norm": 7.819380233896718,
"learning_rate": 1.4915134677675075e-07,
"logits/chosen": -1.6574629545211792,
"logits/rejected": -1.6859042644500732,
"logps/chosen": -482.84991455078125,
"logps/rejected": -488.5449523925781,
"loss": 0.6614,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.13272731006145477,
"rewards/margins": 0.06106055900454521,
"rewards/rejected": -0.1937878578901291,
"step": 220
},
{
"epoch": 0.14433831333170055,
"grad_norm": 24.184875753596092,
"learning_rate": 1.4912548566019288e-07,
"logits/chosen": -1.730996012687683,
"logits/rejected": -1.7523002624511719,
"logps/chosen": -550.131103515625,
"logps/rejected": -511.6734313964844,
"loss": 0.6859,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.16892360150814056,
"rewards/margins": 0.009203894063830376,
"rewards/rejected": -0.1781274825334549,
"step": 221
},
{
"epoch": 0.144991427871663,
"grad_norm": 40.392554453725786,
"learning_rate": 1.4909923871076067e-07,
"logits/chosen": -1.7876818180084229,
"logits/rejected": -1.8064486980438232,
"logps/chosen": -545.5269775390625,
"logps/rejected": -553.6494140625,
"loss": 0.663,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.21288356184959412,
"rewards/margins": 0.10687703639268875,
"rewards/rejected": -0.31976059079170227,
"step": 222
},
{
"epoch": 0.14564454241162544,
"grad_norm": 26.592793442117337,
"learning_rate": 1.4907260606507294e-07,
"logits/chosen": -1.7836247682571411,
"logits/rejected": -1.7736763954162598,
"logps/chosen": -491.8907470703125,
"logps/rejected": -510.7823181152344,
"loss": 0.6533,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.17148932814598083,
"rewards/margins": 0.10181709378957748,
"rewards/rejected": -0.27330639958381653,
"step": 223
},
{
"epoch": 0.1462976569515879,
"grad_norm": 7.602991933565122,
"learning_rate": 1.490455878617561e-07,
"logits/chosen": -1.7201324701309204,
"logits/rejected": -1.6427278518676758,
"logps/chosen": -493.15155029296875,
"logps/rejected": -565.255615234375,
"loss": 0.6665,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.16769693791866302,
"rewards/margins": 0.1270834058523178,
"rewards/rejected": -0.29478034377098083,
"step": 224
},
{
"epoch": 0.14695077149155034,
"grad_norm": 19.894237277380455,
"learning_rate": 1.4901818424144348e-07,
"logits/chosen": -1.7538783550262451,
"logits/rejected": -1.7509037256240845,
"logps/chosen": -613.3535766601562,
"logps/rejected": -623.0008544921875,
"loss": 0.6655,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.22273880243301392,
"rewards/margins": 0.08194239437580109,
"rewards/rejected": -0.3046812117099762,
"step": 225
},
{
"epoch": 0.14760388603151278,
"grad_norm": 27.20495808215994,
"learning_rate": 1.4899039534677446e-07,
"logits/chosen": -1.678489327430725,
"logits/rejected": -1.598201870918274,
"logps/chosen": -493.9471435546875,
"logps/rejected": -593.92333984375,
"loss": 0.6653,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.23504018783569336,
"rewards/margins": 0.11126020550727844,
"rewards/rejected": -0.3463003933429718,
"step": 226
},
{
"epoch": 0.14825700057147523,
"grad_norm": 30.195854234112794,
"learning_rate": 1.489622213223939e-07,
"logits/chosen": -1.7436487674713135,
"logits/rejected": -1.712781310081482,
"logps/chosen": -514.0474243164062,
"logps/rejected": -565.7003784179688,
"loss": 0.6667,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.1994318962097168,
"rewards/margins": 0.1081019788980484,
"rewards/rejected": -0.307533860206604,
"step": 227
},
{
"epoch": 0.14891011511143767,
"grad_norm": 6.64214876236842,
"learning_rate": 1.4893366231495133e-07,
"logits/chosen": -1.6693284511566162,
"logits/rejected": -1.6889541149139404,
"logps/chosen": -538.619873046875,
"logps/rejected": -530.3651123046875,
"loss": 0.6658,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.18542669713497162,
"rewards/margins": 0.06551727652549744,
"rewards/rejected": -0.25094395875930786,
"step": 228
},
{
"epoch": 0.14956322965140012,
"grad_norm": 19.211074462031995,
"learning_rate": 1.489047184731001e-07,
"logits/chosen": -1.784908413887024,
"logits/rejected": -1.801963210105896,
"logps/chosen": -545.5239868164062,
"logps/rejected": -563.0906372070312,
"loss": 0.6648,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.20398400723934174,
"rewards/margins": 0.05913097783923149,
"rewards/rejected": -0.26311495900154114,
"step": 229
},
{
"epoch": 0.15021634419136257,
"grad_norm": 18.555655165809004,
"learning_rate": 1.488753899474967e-07,
"logits/chosen": -1.7692968845367432,
"logits/rejected": -1.7724251747131348,
"logps/chosen": -505.1445617675781,
"logps/rejected": -516.6723022460938,
"loss": 0.6724,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.21376638114452362,
"rewards/margins": 0.06437156349420547,
"rewards/rejected": -0.2781379520893097,
"step": 230
},
{
"epoch": 0.150869458731325,
"grad_norm": 12.537841664973817,
"learning_rate": 1.4884567689079993e-07,
"logits/chosen": -1.7249011993408203,
"logits/rejected": -1.7343446016311646,
"logps/chosen": -559.3654174804688,
"logps/rejected": -483.27349853515625,
"loss": 0.6834,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.23856452107429504,
"rewards/margins": -0.02549094706773758,
"rewards/rejected": -0.21307358145713806,
"step": 231
},
{
"epoch": 0.15152257327128746,
"grad_norm": 7.974565094982971,
"learning_rate": 1.4881557945767017e-07,
"logits/chosen": -1.6963640451431274,
"logits/rejected": -1.6877349615097046,
"logps/chosen": -534.5089111328125,
"logps/rejected": -530.7501220703125,
"loss": 0.6613,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.2620963156223297,
"rewards/margins": 0.11971013993024826,
"rewards/rejected": -0.38180649280548096,
"step": 232
},
{
"epoch": 0.1521756878112499,
"grad_norm": 18.70437584528313,
"learning_rate": 1.4878509780476852e-07,
"logits/chosen": -1.6992989778518677,
"logits/rejected": -1.7136512994766235,
"logps/chosen": -494.5487060546875,
"logps/rejected": -466.22174072265625,
"loss": 0.6548,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.1514841765165329,
"rewards/margins": 0.03264841437339783,
"rewards/rejected": -0.18413259088993073,
"step": 233
},
{
"epoch": 0.15282880235121235,
"grad_norm": 6.4709344020100925,
"learning_rate": 1.4875423209075598e-07,
"logits/chosen": -1.7042124271392822,
"logits/rejected": -1.6889500617980957,
"logps/chosen": -625.1463623046875,
"logps/rejected": -610.1585083007812,
"loss": 0.6638,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.2791493237018585,
"rewards/margins": 0.07801420241594315,
"rewards/rejected": -0.35716351866722107,
"step": 234
},
{
"epoch": 0.1534819168911748,
"grad_norm": 6.93021377573884,
"learning_rate": 1.4872298247629262e-07,
"logits/chosen": -1.7167174816131592,
"logits/rejected": -1.7292131185531616,
"logps/chosen": -540.9368286132812,
"logps/rejected": -528.7276611328125,
"loss": 0.6536,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.21813733875751495,
"rewards/margins": 0.07377970963716507,
"rewards/rejected": -0.29191702604293823,
"step": 235
},
{
"epoch": 0.15413503143113724,
"grad_norm": 17.987164953454663,
"learning_rate": 1.486913491240368e-07,
"logits/chosen": -1.7713466882705688,
"logits/rejected": -1.7571359872817993,
"logps/chosen": -569.6904907226562,
"logps/rejected": -590.6726684570312,
"loss": 0.6518,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.27502891421318054,
"rewards/margins": 0.07631243020296097,
"rewards/rejected": -0.3513413667678833,
"step": 236
},
{
"epoch": 0.1547881459710997,
"grad_norm": 13.606732916452634,
"learning_rate": 1.4865933219864426e-07,
"logits/chosen": -1.687245488166809,
"logits/rejected": -1.6311473846435547,
"logps/chosen": -535.9907836914062,
"logps/rejected": -575.3447875976562,
"loss": 0.6569,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.24033436179161072,
"rewards/margins": 0.06762813031673431,
"rewards/rejected": -0.3079625368118286,
"step": 237
},
{
"epoch": 0.15544126051106213,
"grad_norm": 19.472827108008165,
"learning_rate": 1.4862693186676727e-07,
"logits/chosen": -1.6632615327835083,
"logits/rejected": -1.7188615798950195,
"logps/chosen": -570.304931640625,
"logps/rejected": -528.0740966796875,
"loss": 0.6658,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.2309207022190094,
"rewards/margins": 0.0879175215959549,
"rewards/rejected": -0.3188382387161255,
"step": 238
},
{
"epoch": 0.15609437505102458,
"grad_norm": 22.88007198498636,
"learning_rate": 1.4859414829705384e-07,
"logits/chosen": -1.6684261560440063,
"logits/rejected": -1.7101047039031982,
"logps/chosen": -567.6971435546875,
"logps/rejected": -546.453369140625,
"loss": 0.6719,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.19033075869083405,
"rewards/margins": 0.06966198980808258,
"rewards/rejected": -0.2599927484989166,
"step": 239
},
{
"epoch": 0.15674748959098703,
"grad_norm": 36.841939632438105,
"learning_rate": 1.4856098166014676e-07,
"logits/chosen": -1.7202922105789185,
"logits/rejected": -1.6927356719970703,
"logps/chosen": -574.1094970703125,
"logps/rejected": -533.2037963867188,
"loss": 0.6703,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.25585952401161194,
"rewards/margins": 0.0038572316989302635,
"rewards/rejected": -0.2597167491912842,
"step": 240
},
{
"epoch": 0.15740060413094947,
"grad_norm": 21.236611473334218,
"learning_rate": 1.4852743212868267e-07,
"logits/chosen": -1.653743028640747,
"logits/rejected": -1.6406570672988892,
"logps/chosen": -491.4779968261719,
"logps/rejected": -504.90289306640625,
"loss": 0.6619,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.26939359307289124,
"rewards/margins": -0.0027835238724946976,
"rewards/rejected": -0.2666100263595581,
"step": 241
},
{
"epoch": 0.15805371867091192,
"grad_norm": 22.64341132140898,
"learning_rate": 1.4849349987729134e-07,
"logits/chosen": -1.6768407821655273,
"logits/rejected": -1.707101821899414,
"logps/chosen": -508.9664306640625,
"logps/rejected": -484.92047119140625,
"loss": 0.6723,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.22562530636787415,
"rewards/margins": 0.022399617359042168,
"rewards/rejected": -0.24802491068840027,
"step": 242
},
{
"epoch": 0.15870683321087437,
"grad_norm": 38.25143501681014,
"learning_rate": 1.4845918508259456e-07,
"logits/chosen": -1.5299750566482544,
"logits/rejected": -1.5509228706359863,
"logps/chosen": -515.6123046875,
"logps/rejected": -503.2989196777344,
"loss": 0.6704,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.2918733060359955,
"rewards/margins": 0.03851493448019028,
"rewards/rejected": -0.33038821816444397,
"step": 243
},
{
"epoch": 0.1593599477508368,
"grad_norm": 42.26475756226102,
"learning_rate": 1.4842448792320532e-07,
"logits/chosen": -1.6829333305358887,
"logits/rejected": -1.6792709827423096,
"logps/chosen": -551.5542602539062,
"logps/rejected": -623.4794921875,
"loss": 0.6633,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.2852548360824585,
"rewards/margins": 0.11484323441982269,
"rewards/rejected": -0.4000980854034424,
"step": 244
},
{
"epoch": 0.16001306229079926,
"grad_norm": 6.6637868943907135,
"learning_rate": 1.4838940857972694e-07,
"logits/chosen": -1.7213988304138184,
"logits/rejected": -1.7653172016143799,
"logps/chosen": -597.0615234375,
"logps/rejected": -580.1837158203125,
"loss": 0.6619,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.2760111391544342,
"rewards/margins": 0.08438215404748917,
"rewards/rejected": -0.3603932559490204,
"step": 245
},
{
"epoch": 0.1606661768307617,
"grad_norm": 7.466033172797767,
"learning_rate": 1.4835394723475195e-07,
"logits/chosen": -1.7462300062179565,
"logits/rejected": -1.7599581480026245,
"logps/chosen": -624.055908203125,
"logps/rejected": -556.203369140625,
"loss": 0.6582,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.32625612616539,
"rewards/margins": 0.019012071192264557,
"rewards/rejected": -0.3452681601047516,
"step": 246
},
{
"epoch": 0.16131929137072415,
"grad_norm": 18.36532918357786,
"learning_rate": 1.4831810407286132e-07,
"logits/chosen": -1.7185373306274414,
"logits/rejected": -1.6984236240386963,
"logps/chosen": -497.5492858886719,
"logps/rejected": -549.6527709960938,
"loss": 0.657,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.21095693111419678,
"rewards/margins": 0.10894811898469925,
"rewards/rejected": -0.31990501284599304,
"step": 247
},
{
"epoch": 0.1619724059106866,
"grad_norm": 6.712430351467606,
"learning_rate": 1.4828187928062343e-07,
"logits/chosen": -1.677220344543457,
"logits/rejected": -1.6865521669387817,
"logps/chosen": -496.3159484863281,
"logps/rejected": -491.939208984375,
"loss": 0.6605,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.2620304822921753,
"rewards/margins": 0.05196765065193176,
"rewards/rejected": -0.31399816274642944,
"step": 248
},
{
"epoch": 0.16262552045064904,
"grad_norm": 16.60390065867787,
"learning_rate": 1.4824527304659303e-07,
"logits/chosen": -1.8142789602279663,
"logits/rejected": -1.808518648147583,
"logps/chosen": -573.7063598632812,
"logps/rejected": -586.904541015625,
"loss": 0.656,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.2939229905605316,
"rewards/margins": 0.20337671041488647,
"rewards/rejected": -0.4972996711730957,
"step": 249
},
{
"epoch": 0.1632786349906115,
"grad_norm": 8.526623591061165,
"learning_rate": 1.4820828556131042e-07,
"logits/chosen": -1.682909607887268,
"logits/rejected": -1.6931588649749756,
"logps/chosen": -553.196044921875,
"logps/rejected": -606.1085205078125,
"loss": 0.6407,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3215971887111664,
"rewards/margins": 0.13771557807922363,
"rewards/rejected": -0.4593127965927124,
"step": 250
},
{
"epoch": 0.16393174953057393,
"grad_norm": 31.575444561187965,
"learning_rate": 1.4817091701730025e-07,
"logits/chosen": -1.7506418228149414,
"logits/rejected": -1.7645210027694702,
"logps/chosen": -504.40216064453125,
"logps/rejected": -466.3576965332031,
"loss": 0.6607,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2621554732322693,
"rewards/margins": 0.02839628979563713,
"rewards/rejected": -0.2905517518520355,
"step": 251
},
{
"epoch": 0.16458486407053638,
"grad_norm": 33.198936730829594,
"learning_rate": 1.4813316760907073e-07,
"logits/chosen": -1.7571806907653809,
"logits/rejected": -1.7107410430908203,
"logps/chosen": -535.119140625,
"logps/rejected": -570.49658203125,
"loss": 0.6653,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2747470736503601,
"rewards/margins": 0.10117386281490326,
"rewards/rejected": -0.37592095136642456,
"step": 252
},
{
"epoch": 0.16523797861049883,
"grad_norm": 8.023556183248012,
"learning_rate": 1.480950375331125e-07,
"logits/chosen": -1.724591851234436,
"logits/rejected": -1.724915623664856,
"logps/chosen": -533.9208374023438,
"logps/rejected": -578.927734375,
"loss": 0.6578,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.2872743010520935,
"rewards/margins": 0.1008455902338028,
"rewards/rejected": -0.3881198763847351,
"step": 253
},
{
"epoch": 0.16589109315046127,
"grad_norm": 17.29562845944618,
"learning_rate": 1.4805652698789758e-07,
"logits/chosen": -1.6281641721725464,
"logits/rejected": -1.6535223722457886,
"logps/chosen": -501.57318115234375,
"logps/rejected": -623.32568359375,
"loss": 0.6373,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.29188841581344604,
"rewards/margins": 0.24822545051574707,
"rewards/rejected": -0.5401138663291931,
"step": 254
},
{
"epoch": 0.16654420769042372,
"grad_norm": 12.955908095771617,
"learning_rate": 1.480176361738784e-07,
"logits/chosen": -1.6974416971206665,
"logits/rejected": -1.7031437158584595,
"logps/chosen": -536.0557250976562,
"logps/rejected": -505.6365661621094,
"loss": 0.6586,
"rewards/accuracies": 0.46875,
"rewards/chosen": -0.32581400871276855,
"rewards/margins": 0.018291521817445755,
"rewards/rejected": -0.3441055715084076,
"step": 255
},
{
"epoch": 0.16719732223038616,
"grad_norm": 10.398012268043638,
"learning_rate": 1.4797836529348678e-07,
"logits/chosen": -1.6877224445343018,
"logits/rejected": -1.6973309516906738,
"logps/chosen": -539.1455688476562,
"logps/rejected": -539.2821044921875,
"loss": 0.6554,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.30231714248657227,
"rewards/margins": 0.07891248911619186,
"rewards/rejected": -0.3812296390533447,
"step": 256
},
{
"epoch": 0.1678504367703486,
"grad_norm": 13.095086309931633,
"learning_rate": 1.4793871455113277e-07,
"logits/chosen": -1.7343683242797852,
"logits/rejected": -1.7324395179748535,
"logps/chosen": -509.0965576171875,
"logps/rejected": -560.6080322265625,
"loss": 0.6479,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.30541422963142395,
"rewards/margins": 0.18525291979312897,
"rewards/rejected": -0.4906671643257141,
"step": 257
},
{
"epoch": 0.16850355131031106,
"grad_norm": 8.776278689782488,
"learning_rate": 1.478986841532037e-07,
"logits/chosen": -1.6799153089523315,
"logits/rejected": -1.6792922019958496,
"logps/chosen": -507.31787109375,
"logps/rejected": -538.5675659179688,
"loss": 0.6616,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.2834991216659546,
"rewards/margins": 0.1000271737575531,
"rewards/rejected": -0.3835262954235077,
"step": 258
},
{
"epoch": 0.1691566658502735,
"grad_norm": 12.19544220249078,
"learning_rate": 1.4785827430806304e-07,
"logits/chosen": -1.6284384727478027,
"logits/rejected": -1.5995004177093506,
"logps/chosen": -496.8951416015625,
"logps/rejected": -489.3295593261719,
"loss": 0.6491,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.344193696975708,
"rewards/margins": 0.09732924401760101,
"rewards/rejected": -0.4415229558944702,
"step": 259
},
{
"epoch": 0.16980978039023595,
"grad_norm": 34.82104290571965,
"learning_rate": 1.4781748522604932e-07,
"logits/chosen": -1.7195165157318115,
"logits/rejected": -1.7009164094924927,
"logps/chosen": -552.0281982421875,
"logps/rejected": -630.2988891601562,
"loss": 0.6517,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3577840328216553,
"rewards/margins": 0.2054346352815628,
"rewards/rejected": -0.5632186532020569,
"step": 260
},
{
"epoch": 0.1704628949301984,
"grad_norm": 30.979244743569765,
"learning_rate": 1.4777631711947508e-07,
"logits/chosen": -1.6629853248596191,
"logits/rejected": -1.6167453527450562,
"logps/chosen": -505.7708740234375,
"logps/rejected": -532.100830078125,
"loss": 0.6485,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3221421241760254,
"rewards/margins": 0.06837349385023117,
"rewards/rejected": -0.39051565527915955,
"step": 261
},
{
"epoch": 0.17111600947016084,
"grad_norm": 41.43490075038442,
"learning_rate": 1.4773477020262572e-07,
"logits/chosen": -1.728777527809143,
"logits/rejected": -1.726311445236206,
"logps/chosen": -620.5463256835938,
"logps/rejected": -633.969970703125,
"loss": 0.6527,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.4750226140022278,
"rewards/margins": 0.15798604488372803,
"rewards/rejected": -0.6330086588859558,
"step": 262
},
{
"epoch": 0.1717691240101233,
"grad_norm": 23.108934696505543,
"learning_rate": 1.4769284469175835e-07,
"logits/chosen": -1.7010533809661865,
"logits/rejected": -1.700863242149353,
"logps/chosen": -496.9283752441406,
"logps/rejected": -496.18096923828125,
"loss": 0.6404,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.2904779613018036,
"rewards/margins": 0.08895879238843918,
"rewards/rejected": -0.37943676114082336,
"step": 263
},
{
"epoch": 0.17242223855008573,
"grad_norm": 29.393486906363,
"learning_rate": 1.476505408051008e-07,
"logits/chosen": -1.7014069557189941,
"logits/rejected": -1.785143494606018,
"logps/chosen": -577.2012939453125,
"logps/rejected": -608.8991088867188,
"loss": 0.6534,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3605682849884033,
"rewards/margins": 0.11243538558483124,
"rewards/rejected": -0.47300365567207336,
"step": 264
},
{
"epoch": 0.17307535309004818,
"grad_norm": 6.776931632936762,
"learning_rate": 1.476078587628503e-07,
"logits/chosen": -1.7219668626785278,
"logits/rejected": -1.7316367626190186,
"logps/chosen": -613.4144287109375,
"logps/rejected": -600.9342041015625,
"loss": 0.6394,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.4629693329334259,
"rewards/margins": 0.09507772326469421,
"rewards/rejected": -0.5580470561981201,
"step": 265
},
{
"epoch": 0.17372846763001062,
"grad_norm": 15.838565452388137,
"learning_rate": 1.4756479878717254e-07,
"logits/chosen": -1.6845935583114624,
"logits/rejected": -1.697797417640686,
"logps/chosen": -502.57818603515625,
"logps/rejected": -501.9476318359375,
"loss": 0.6348,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.3347086012363434,
"rewards/margins": 0.14087814092636108,
"rewards/rejected": -0.4755867123603821,
"step": 266
},
{
"epoch": 0.17438158216997307,
"grad_norm": 22.80772385019741,
"learning_rate": 1.4752136110220027e-07,
"logits/chosen": -1.689018964767456,
"logits/rejected": -1.68304443359375,
"logps/chosen": -510.1197509765625,
"logps/rejected": -546.5357666015625,
"loss": 0.6512,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.30356815457344055,
"rewards/margins": 0.14922784268856049,
"rewards/rejected": -0.45279598236083984,
"step": 267
},
{
"epoch": 0.17503469670993552,
"grad_norm": 23.730722811526665,
"learning_rate": 1.4747754593403243e-07,
"logits/chosen": -1.6035892963409424,
"logits/rejected": -1.63158118724823,
"logps/chosen": -526.643798828125,
"logps/rejected": -575.2221069335938,
"loss": 0.6536,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.3458973467350006,
"rewards/margins": 0.19965365529060364,
"rewards/rejected": -0.5455510020256042,
"step": 268
},
{
"epoch": 0.17568781124989796,
"grad_norm": 7.0753534582972,
"learning_rate": 1.4743335351073263e-07,
"logits/chosen": -1.7131729125976562,
"logits/rejected": -1.716842532157898,
"logps/chosen": -589.8107299804688,
"logps/rejected": -586.211669921875,
"loss": 0.6484,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.36673662066459656,
"rewards/margins": 0.08891665935516357,
"rewards/rejected": -0.45565328001976013,
"step": 269
},
{
"epoch": 0.1763409257898604,
"grad_norm": 13.47915067878089,
"learning_rate": 1.4738878406232824e-07,
"logits/chosen": -1.6974538564682007,
"logits/rejected": -1.7051836252212524,
"logps/chosen": -576.85302734375,
"logps/rejected": -595.0402221679688,
"loss": 0.6588,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.4583105444908142,
"rewards/margins": 0.09714774042367935,
"rewards/rejected": -0.5554581880569458,
"step": 270
},
{
"epoch": 0.17699404032982285,
"grad_norm": 26.554664114423755,
"learning_rate": 1.4734383782080914e-07,
"logits/chosen": -1.682262659072876,
"logits/rejected": -1.653159499168396,
"logps/chosen": -584.1964721679688,
"logps/rejected": -617.1187744140625,
"loss": 0.6637,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.42176809906959534,
"rewards/margins": 0.07791588455438614,
"rewards/rejected": -0.4996839761734009,
"step": 271
},
{
"epoch": 0.1776471548697853,
"grad_norm": 10.821406826613458,
"learning_rate": 1.4729851502012636e-07,
"logits/chosen": -1.711740255355835,
"logits/rejected": -1.7466825246810913,
"logps/chosen": -522.8740844726562,
"logps/rejected": -510.05743408203125,
"loss": 0.6331,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.388105183839798,
"rewards/margins": 0.08373910188674927,
"rewards/rejected": -0.47184431552886963,
"step": 272
},
{
"epoch": 0.17830026940974775,
"grad_norm": 6.675378991764519,
"learning_rate": 1.4725281589619103e-07,
"logits/chosen": -1.6703485250473022,
"logits/rejected": -1.6685247421264648,
"logps/chosen": -530.2496337890625,
"logps/rejected": -535.6815795898438,
"loss": 0.6426,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.42745375633239746,
"rewards/margins": 0.18275927007198334,
"rewards/rejected": -0.6102129817008972,
"step": 273
},
{
"epoch": 0.1789533839497102,
"grad_norm": 9.179293872139016,
"learning_rate": 1.4720674068687308e-07,
"logits/chosen": -1.7541323900222778,
"logits/rejected": -1.811903476715088,
"logps/chosen": -548.3256225585938,
"logps/rejected": -549.988525390625,
"loss": 0.6477,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.4320800304412842,
"rewards/margins": 0.04719913750886917,
"rewards/rejected": -0.47927919030189514,
"step": 274
},
{
"epoch": 0.17960649848967264,
"grad_norm": 19.694787175563757,
"learning_rate": 1.4716028963200005e-07,
"logits/chosen": -1.692307472229004,
"logits/rejected": -1.6705561876296997,
"logps/chosen": -493.9141845703125,
"logps/rejected": -527.4869384765625,
"loss": 0.6533,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.3439943492412567,
"rewards/margins": 0.14113494753837585,
"rewards/rejected": -0.48512929677963257,
"step": 275
},
{
"epoch": 0.18025961302963509,
"grad_norm": 20.421232540328838,
"learning_rate": 1.4711346297335575e-07,
"logits/chosen": -1.7493505477905273,
"logits/rejected": -1.7484819889068604,
"logps/chosen": -562.100341796875,
"logps/rejected": -544.1605224609375,
"loss": 0.6586,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.39796724915504456,
"rewards/margins": 0.1226586103439331,
"rewards/rejected": -0.5206258296966553,
"step": 276
},
{
"epoch": 0.18091272756959753,
"grad_norm": 7.813830449111613,
"learning_rate": 1.4706626095467905e-07,
"logits/chosen": -1.6537861824035645,
"logits/rejected": -1.6490662097930908,
"logps/chosen": -615.1405029296875,
"logps/rejected": -640.3934326171875,
"loss": 0.6364,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.39063894748687744,
"rewards/margins": 0.22695013880729675,
"rewards/rejected": -0.6175890564918518,
"step": 277
},
{
"epoch": 0.18156584210955998,
"grad_norm": 18.137788034555644,
"learning_rate": 1.470186838216627e-07,
"logits/chosen": -1.7237962484359741,
"logits/rejected": -1.7260386943817139,
"logps/chosen": -560.9794921875,
"logps/rejected": -544.9048461914062,
"loss": 0.6549,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.4301608204841614,
"rewards/margins": 0.10491985827684402,
"rewards/rejected": -0.5350806713104248,
"step": 278
},
{
"epoch": 0.18221895664952242,
"grad_norm": 6.203981970076561,
"learning_rate": 1.469707318219519e-07,
"logits/chosen": -1.7497467994689941,
"logits/rejected": -1.7625311613082886,
"logps/chosen": -476.921875,
"logps/rejected": -468.9197692871094,
"loss": 0.6288,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.39907845854759216,
"rewards/margins": 0.09824170172214508,
"rewards/rejected": -0.49732017517089844,
"step": 279
},
{
"epoch": 0.18287207118948487,
"grad_norm": 30.605405460520963,
"learning_rate": 1.4692240520514308e-07,
"logits/chosen": -1.692575216293335,
"logits/rejected": -1.7037385702133179,
"logps/chosen": -459.114990234375,
"logps/rejected": -475.6065368652344,
"loss": 0.6475,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.37590938806533813,
"rewards/margins": 0.06421739608049393,
"rewards/rejected": -0.4401267468929291,
"step": 280
},
{
"epoch": 0.18352518572944732,
"grad_norm": 6.890672032716978,
"learning_rate": 1.4687370422278264e-07,
"logits/chosen": -1.6996791362762451,
"logits/rejected": -1.7280397415161133,
"logps/chosen": -567.3660278320312,
"logps/rejected": -556.817138671875,
"loss": 0.6557,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.4747771620750427,
"rewards/margins": 0.10796495527029037,
"rewards/rejected": -0.5827420949935913,
"step": 281
},
{
"epoch": 0.18417830026940976,
"grad_norm": 7.575771612722592,
"learning_rate": 1.4682462912836556e-07,
"logits/chosen": -1.6831636428833008,
"logits/rejected": -1.659497857093811,
"logps/chosen": -591.215576171875,
"logps/rejected": -606.3766479492188,
"loss": 0.6464,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.5020853281021118,
"rewards/margins": 0.08574617654085159,
"rewards/rejected": -0.5878314971923828,
"step": 282
},
{
"epoch": 0.18483141480937218,
"grad_norm": 14.228779893218709,
"learning_rate": 1.4677518017733416e-07,
"logits/chosen": -1.620701789855957,
"logits/rejected": -1.6582701206207275,
"logps/chosen": -493.6382141113281,
"logps/rejected": -494.93731689453125,
"loss": 0.652,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.4052881896495819,
"rewards/margins": 0.05283729359507561,
"rewards/rejected": -0.4581254720687866,
"step": 283
},
{
"epoch": 0.18548452934933463,
"grad_norm": 9.401784297751998,
"learning_rate": 1.467253576270767e-07,
"logits/chosen": -1.741771936416626,
"logits/rejected": -1.75984787940979,
"logps/chosen": -491.3598937988281,
"logps/rejected": -527.6268310546875,
"loss": 0.6431,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.45621737837791443,
"rewards/margins": 0.12358233332633972,
"rewards/rejected": -0.5797997713088989,
"step": 284
},
{
"epoch": 0.18613764388929707,
"grad_norm": 8.738330524484025,
"learning_rate": 1.466751617369261e-07,
"logits/chosen": -1.7441930770874023,
"logits/rejected": -1.8137414455413818,
"logps/chosen": -600.4554443359375,
"logps/rejected": -585.8037719726562,
"loss": 0.6408,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5598416924476624,
"rewards/margins": 0.06251634657382965,
"rewards/rejected": -0.6223580241203308,
"step": 285
},
{
"epoch": 0.18679075842925952,
"grad_norm": 7.921733780275398,
"learning_rate": 1.4662459276815857e-07,
"logits/chosen": -1.740809679031372,
"logits/rejected": -1.6921014785766602,
"logps/chosen": -521.9088745117188,
"logps/rejected": -613.4127807617188,
"loss": 0.6222,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.5204059481620789,
"rewards/margins": 0.23885881900787354,
"rewards/rejected": -0.7592648267745972,
"step": 286
},
{
"epoch": 0.18744387296922196,
"grad_norm": 14.993635596662422,
"learning_rate": 1.4657365098399217e-07,
"logits/chosen": -1.6475145816802979,
"logits/rejected": -1.6759244203567505,
"logps/chosen": -588.921142578125,
"logps/rejected": -621.540771484375,
"loss": 0.6506,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.5357760787010193,
"rewards/margins": 0.17871497571468353,
"rewards/rejected": -0.714491069316864,
"step": 287
},
{
"epoch": 0.1880969875091844,
"grad_norm": 13.969578435371153,
"learning_rate": 1.4652233664958564e-07,
"logits/chosen": -1.6996246576309204,
"logits/rejected": -1.6887569427490234,
"logps/chosen": -524.6732177734375,
"logps/rejected": -578.5980224609375,
"loss": 0.6505,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.513081431388855,
"rewards/margins": 0.24069830775260925,
"rewards/rejected": -0.7537796497344971,
"step": 288
},
{
"epoch": 0.18875010204914686,
"grad_norm": 11.610438307536514,
"learning_rate": 1.4647065003203673e-07,
"logits/chosen": -1.629154086112976,
"logits/rejected": -1.675700306892395,
"logps/chosen": -522.1669921875,
"logps/rejected": -537.0850830078125,
"loss": 0.6589,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5520819425582886,
"rewards/margins": 0.10711211711168289,
"rewards/rejected": -0.6591941118240356,
"step": 289
},
{
"epoch": 0.1894032165891093,
"grad_norm": 10.283639619093217,
"learning_rate": 1.4641859140038115e-07,
"logits/chosen": -1.7125153541564941,
"logits/rejected": -1.7900097370147705,
"logps/chosen": -481.0557861328125,
"logps/rejected": -466.9371643066406,
"loss": 0.6447,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.4833194613456726,
"rewards/margins": 0.10249477624893188,
"rewards/rejected": -0.5858142375946045,
"step": 290
},
{
"epoch": 0.19005633112907175,
"grad_norm": 10.863781043680522,
"learning_rate": 1.4636616102559085e-07,
"logits/chosen": -1.7732642889022827,
"logits/rejected": -1.7560861110687256,
"logps/chosen": -497.37286376953125,
"logps/rejected": -492.2978515625,
"loss": 0.6447,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5158191919326782,
"rewards/margins": 0.12314166873693466,
"rewards/rejected": -0.6389608383178711,
"step": 291
},
{
"epoch": 0.1907094456690342,
"grad_norm": 8.437201886799969,
"learning_rate": 1.4631335918057284e-07,
"logits/chosen": -1.755825161933899,
"logits/rejected": -1.7563961744308472,
"logps/chosen": -545.761962890625,
"logps/rejected": -605.7576904296875,
"loss": 0.6288,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.5955057740211487,
"rewards/margins": 0.1695830523967743,
"rewards/rejected": -0.7650887966156006,
"step": 292
},
{
"epoch": 0.19136256020899664,
"grad_norm": 8.28332489348213,
"learning_rate": 1.4626018614016762e-07,
"logits/chosen": -1.6458687782287598,
"logits/rejected": -1.7265384197235107,
"logps/chosen": -524.3062744140625,
"logps/rejected": -535.7152709960938,
"loss": 0.6534,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.5473222136497498,
"rewards/margins": 0.09474081546068192,
"rewards/rejected": -0.6420629620552063,
"step": 293
},
{
"epoch": 0.1920156747489591,
"grad_norm": 12.121969382291516,
"learning_rate": 1.4620664218114785e-07,
"logits/chosen": -1.7038286924362183,
"logits/rejected": -1.7111061811447144,
"logps/chosen": -608.0626220703125,
"logps/rejected": -610.8870849609375,
"loss": 0.6474,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.6269205212593079,
"rewards/margins": 0.192575603723526,
"rewards/rejected": -0.8194961547851562,
"step": 294
},
{
"epoch": 0.19266878928892153,
"grad_norm": 8.914872873902894,
"learning_rate": 1.4615272758221687e-07,
"logits/chosen": -1.6704249382019043,
"logits/rejected": -1.6967017650604248,
"logps/chosen": -571.5804443359375,
"logps/rejected": -568.3060302734375,
"loss": 0.6579,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.5785356163978577,
"rewards/margins": 0.08196890354156494,
"rewards/rejected": -0.6605044603347778,
"step": 295
},
{
"epoch": 0.19332190382888398,
"grad_norm": 7.935449781786862,
"learning_rate": 1.4609844262400722e-07,
"logits/chosen": -1.7343144416809082,
"logits/rejected": -1.7453346252441406,
"logps/chosen": -595.2299194335938,
"logps/rejected": -627.33984375,
"loss": 0.6345,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6429135799407959,
"rewards/margins": 0.16354820132255554,
"rewards/rejected": -0.8064618110656738,
"step": 296
},
{
"epoch": 0.19397501836884642,
"grad_norm": 7.485459727998883,
"learning_rate": 1.4604378758907928e-07,
"logits/chosen": -1.717028021812439,
"logits/rejected": -1.7182062864303589,
"logps/chosen": -603.7417602539062,
"logps/rejected": -610.2848510742188,
"loss": 0.6507,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.6892397999763489,
"rewards/margins": 0.20615926384925842,
"rewards/rejected": -0.8953990936279297,
"step": 297
},
{
"epoch": 0.19462813290880887,
"grad_norm": 11.03240678642255,
"learning_rate": 1.459887627619196e-07,
"logits/chosen": -1.7350809574127197,
"logits/rejected": -1.7344433069229126,
"logps/chosen": -579.1885986328125,
"logps/rejected": -620.0966796875,
"loss": 0.6361,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6911602020263672,
"rewards/margins": 0.14496688544750214,
"rewards/rejected": -0.836126983165741,
"step": 298
},
{
"epoch": 0.19528124744877132,
"grad_norm": 8.53488849066449,
"learning_rate": 1.4593336842893963e-07,
"logits/chosen": -1.745806336402893,
"logits/rejected": -1.6635572910308838,
"logps/chosen": -607.8656616210938,
"logps/rejected": -672.4678344726562,
"loss": 0.6208,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.6869965195655823,
"rewards/margins": 0.2365533858537674,
"rewards/rejected": -0.9235499501228333,
"step": 299
},
{
"epoch": 0.19593436198873376,
"grad_norm": 9.49489920339701,
"learning_rate": 1.458776048784742e-07,
"logits/chosen": -1.6529350280761719,
"logits/rejected": -1.6770867109298706,
"logps/chosen": -532.1659545898438,
"logps/rejected": -558.4205322265625,
"loss": 0.6306,
"rewards/accuracies": 0.5625,
"rewards/chosen": -0.6781768798828125,
"rewards/margins": 0.10716618597507477,
"rewards/rejected": -0.7853431701660156,
"step": 300
},
{
"epoch": 0.19593436198873376,
"eval_logits/chosen": -1.7436009645462036,
"eval_logits/rejected": -1.7536306381225586,
"eval_logps/chosen": -574.737548828125,
"eval_logps/rejected": -584.3355712890625,
"eval_loss": 0.6346647143363953,
"eval_rewards/accuracies": 0.6840000152587891,
"eval_rewards/chosen": -0.6439220309257507,
"eval_rewards/margins": 0.17703963816165924,
"eval_rewards/rejected": -0.8209616541862488,
"eval_runtime": 297.2957,
"eval_samples_per_second": 13.455,
"eval_steps_per_second": 0.841,
"step": 300
},
{
"epoch": 0.1965874765286962,
"grad_norm": 8.98774554330102,
"learning_rate": 1.4582147240077982e-07,
"logits/chosen": -1.7660400867462158,
"logits/rejected": -1.7696928977966309,
"logps/chosen": -551.3262939453125,
"logps/rejected": -584.052001953125,
"loss": 0.6399,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.6249464154243469,
"rewards/margins": 0.09611751139163971,
"rewards/rejected": -0.7210639715194702,
"step": 301
},
{
"epoch": 0.19724059106865865,
"grad_norm": 7.863471372182064,
"learning_rate": 1.4576497128803348e-07,
"logits/chosen": -1.7265328168869019,
"logits/rejected": -1.6853750944137573,
"logps/chosen": -536.3348999023438,
"logps/rejected": -628.0272827148438,
"loss": 0.625,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6308943033218384,
"rewards/margins": 0.23925556242465973,
"rewards/rejected": -0.8701498508453369,
"step": 302
},
{
"epoch": 0.1978937056086211,
"grad_norm": 28.45275146429454,
"learning_rate": 1.4570810183433083e-07,
"logits/chosen": -1.6958410739898682,
"logits/rejected": -1.7414443492889404,
"logps/chosen": -663.8955078125,
"logps/rejected": -632.41015625,
"loss": 0.6593,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.8960934281349182,
"rewards/margins": 0.07879979908466339,
"rewards/rejected": -0.9748932123184204,
"step": 303
},
{
"epoch": 0.19854682014858355,
"grad_norm": 11.410487006682061,
"learning_rate": 1.4565086433568487e-07,
"logits/chosen": -1.7225995063781738,
"logits/rejected": -1.6888340711593628,
"logps/chosen": -581.2233276367188,
"logps/rejected": -675.6259155273438,
"loss": 0.6429,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.6933416724205017,
"rewards/margins": 0.3561021387577057,
"rewards/rejected": -1.0494438409805298,
"step": 304
},
{
"epoch": 0.199199934688546,
"grad_norm": 8.270002666629233,
"learning_rate": 1.4559325909002424e-07,
"logits/chosen": -1.7664427757263184,
"logits/rejected": -1.7638676166534424,
"logps/chosen": -560.244873046875,
"logps/rejected": -626.065185546875,
"loss": 0.6395,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6132575869560242,
"rewards/margins": 0.3563953936100006,
"rewards/rejected": -0.9696530103683472,
"step": 305
},
{
"epoch": 0.19985304922850844,
"grad_norm": 12.608968024472125,
"learning_rate": 1.4553528639719185e-07,
"logits/chosen": -1.6360845565795898,
"logits/rejected": -1.7036012411117554,
"logps/chosen": -494.1507568359375,
"logps/rejected": -567.8990478515625,
"loss": 0.6366,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.6473857760429382,
"rewards/margins": 0.21676453948020935,
"rewards/rejected": -0.8641502857208252,
"step": 306
},
{
"epoch": 0.20050616376847089,
"grad_norm": 17.135231726028064,
"learning_rate": 1.4547694655894313e-07,
"logits/chosen": -1.7396022081375122,
"logits/rejected": -1.6808857917785645,
"logps/chosen": -533.951416015625,
"logps/rejected": -562.3763427734375,
"loss": 0.6188,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6712486743927002,
"rewards/margins": 0.19814179837703705,
"rewards/rejected": -0.8693904876708984,
"step": 307
},
{
"epoch": 0.20115927830843333,
"grad_norm": 6.928796136463469,
"learning_rate": 1.454182398789446e-07,
"logits/chosen": -1.776479721069336,
"logits/rejected": -1.7864151000976562,
"logps/chosen": -642.5709228515625,
"logps/rejected": -602.6436157226562,
"loss": 0.6323,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.7122976779937744,
"rewards/margins": 0.10113891959190369,
"rewards/rejected": -0.8134365677833557,
"step": 308
},
{
"epoch": 0.20181239284839578,
"grad_norm": 6.579505553675459,
"learning_rate": 1.4535916666277225e-07,
"logits/chosen": -1.6873722076416016,
"logits/rejected": -1.683058500289917,
"logps/chosen": -625.4521484375,
"logps/rejected": -681.3677978515625,
"loss": 0.6223,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.7979632616043091,
"rewards/margins": 0.28991833329200745,
"rewards/rejected": -1.0878815650939941,
"step": 309
},
{
"epoch": 0.20246550738835822,
"grad_norm": 7.215238256234268,
"learning_rate": 1.4529972721790987e-07,
"logits/chosen": -1.7267826795578003,
"logits/rejected": -1.6847716569900513,
"logps/chosen": -607.7223510742188,
"logps/rejected": -689.0610961914062,
"loss": 0.5934,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.6881787776947021,
"rewards/margins": 0.390371173620224,
"rewards/rejected": -1.0785499811172485,
"step": 310
},
{
"epoch": 0.20311862192832067,
"grad_norm": 8.708876826384516,
"learning_rate": 1.4523992185374762e-07,
"logits/chosen": -1.6836024522781372,
"logits/rejected": -1.7239104509353638,
"logps/chosen": -581.9091796875,
"logps/rejected": -602.4833374023438,
"loss": 0.6503,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.6792923212051392,
"rewards/margins": 0.17103339731693268,
"rewards/rejected": -0.8503257632255554,
"step": 311
},
{
"epoch": 0.20377173646828312,
"grad_norm": 15.693301460696688,
"learning_rate": 1.4517975088158024e-07,
"logits/chosen": -1.6466374397277832,
"logits/rejected": -1.7001001834869385,
"logps/chosen": -522.967041015625,
"logps/rejected": -504.9639892578125,
"loss": 0.6303,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.6082721948623657,
"rewards/margins": 0.13548986613750458,
"rewards/rejected": -0.7437620759010315,
"step": 312
},
{
"epoch": 0.20442485100824556,
"grad_norm": 9.45995269518211,
"learning_rate": 1.4511921461460552e-07,
"logits/chosen": -1.6590129137039185,
"logits/rejected": -1.7052656412124634,
"logps/chosen": -619.462158203125,
"logps/rejected": -658.5916748046875,
"loss": 0.6425,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.8983247876167297,
"rewards/margins": 0.1738293468952179,
"rewards/rejected": -1.0721540451049805,
"step": 313
},
{
"epoch": 0.205077965548208,
"grad_norm": 8.619266517151214,
"learning_rate": 1.4505831336792268e-07,
"logits/chosen": -1.6681677103042603,
"logits/rejected": -1.6365997791290283,
"logps/chosen": -623.3759155273438,
"logps/rejected": -652.1790771484375,
"loss": 0.6061,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.817943811416626,
"rewards/margins": 0.2616656422615051,
"rewards/rejected": -1.0796093940734863,
"step": 314
},
{
"epoch": 0.20573108008817045,
"grad_norm": 16.305545075033027,
"learning_rate": 1.449970474585307e-07,
"logits/chosen": -1.7769198417663574,
"logits/rejected": -1.7549265623092651,
"logps/chosen": -508.7225646972656,
"logps/rejected": -541.2574462890625,
"loss": 0.6153,
"rewards/accuracies": 0.59375,
"rewards/chosen": -0.8023353815078735,
"rewards/margins": 0.1428290456533432,
"rewards/rejected": -0.9451643228530884,
"step": 315
},
{
"epoch": 0.2063841946281329,
"grad_norm": 16.881338125964565,
"learning_rate": 1.4493541720532666e-07,
"logits/chosen": -1.7054510116577148,
"logits/rejected": -1.7318084239959717,
"logps/chosen": -577.06787109375,
"logps/rejected": -674.53857421875,
"loss": 0.5824,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.7155259251594543,
"rewards/margins": 0.4451526999473572,
"rewards/rejected": -1.1606786251068115,
"step": 316
},
{
"epoch": 0.20703730916809535,
"grad_norm": 8.857782623326043,
"learning_rate": 1.4487342292910414e-07,
"logits/chosen": -1.6867567300796509,
"logits/rejected": -1.7076420783996582,
"logps/chosen": -634.0771484375,
"logps/rejected": -662.7860107421875,
"loss": 0.6135,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.9092274904251099,
"rewards/margins": 0.23942092061042786,
"rewards/rejected": -1.1486485004425049,
"step": 317
},
{
"epoch": 0.2076904237080578,
"grad_norm": 32.895897939433254,
"learning_rate": 1.4481106495255145e-07,
"logits/chosen": -1.6250176429748535,
"logits/rejected": -1.6629047393798828,
"logps/chosen": -638.931640625,
"logps/rejected": -675.3734130859375,
"loss": 0.6102,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.9061939120292664,
"rewards/margins": 0.18294726312160492,
"rewards/rejected": -1.0891411304473877,
"step": 318
},
{
"epoch": 0.20834353824802024,
"grad_norm": 35.209821508774134,
"learning_rate": 1.4474834360025005e-07,
"logits/chosen": -1.7239129543304443,
"logits/rejected": -1.7509957551956177,
"logps/chosen": -539.3623046875,
"logps/rejected": -585.4423828125,
"loss": 0.6255,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7921657562255859,
"rewards/margins": 0.20182442665100098,
"rewards/rejected": -0.9939901232719421,
"step": 319
},
{
"epoch": 0.20899665278798268,
"grad_norm": 22.542720186905754,
"learning_rate": 1.446852591986728e-07,
"logits/chosen": -1.7275651693344116,
"logits/rejected": -1.7619481086730957,
"logps/chosen": -574.9896850585938,
"logps/rejected": -537.2247924804688,
"loss": 0.6222,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.7382320761680603,
"rewards/margins": 0.09026205539703369,
"rewards/rejected": -0.8284941911697388,
"step": 320
},
{
"epoch": 0.20964976732794513,
"grad_norm": 11.689355163454417,
"learning_rate": 1.4462181207618226e-07,
"logits/chosen": -1.6684459447860718,
"logits/rejected": -1.664050817489624,
"logps/chosen": -612.190673828125,
"logps/rejected": -627.1231079101562,
"loss": 0.6396,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.9567988514900208,
"rewards/margins": 0.15958093106746674,
"rewards/rejected": -1.1163798570632935,
"step": 321
},
{
"epoch": 0.21030288186790758,
"grad_norm": 8.28817528253119,
"learning_rate": 1.445580025630291e-07,
"logits/chosen": -1.696157455444336,
"logits/rejected": -1.693131446838379,
"logps/chosen": -616.43017578125,
"logps/rejected": -605.6445922851562,
"loss": 0.6273,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.8394571542739868,
"rewards/margins": 0.19782942533493042,
"rewards/rejected": -1.0372865200042725,
"step": 322
},
{
"epoch": 0.21095599640787002,
"grad_norm": 9.732936932214093,
"learning_rate": 1.444938309913501e-07,
"logits/chosen": -1.6943405866622925,
"logits/rejected": -1.7072209119796753,
"logps/chosen": -584.8599853515625,
"logps/rejected": -641.0568237304688,
"loss": 0.6046,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.7837256193161011,
"rewards/margins": 0.3857865333557129,
"rewards/rejected": -1.1695120334625244,
"step": 323
},
{
"epoch": 0.21160911094783247,
"grad_norm": 11.744162208267507,
"learning_rate": 1.444292976951668e-07,
"logits/chosen": -1.6355708837509155,
"logits/rejected": -1.627254843711853,
"logps/chosen": -610.7489013671875,
"logps/rejected": -618.13330078125,
"loss": 0.5927,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.8486420512199402,
"rewards/margins": 0.18311645090579987,
"rewards/rejected": -1.0317585468292236,
"step": 324
},
{
"epoch": 0.21226222548779491,
"grad_norm": 8.36583130670152,
"learning_rate": 1.4436440301038337e-07,
"logits/chosen": -1.6780744791030884,
"logits/rejected": -1.681298017501831,
"logps/chosen": -618.8885498046875,
"logps/rejected": -628.7433471679688,
"loss": 0.6279,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.9191569089889526,
"rewards/margins": 0.21737900376319885,
"rewards/rejected": -1.1365360021591187,
"step": 325
},
{
"epoch": 0.21291534002775736,
"grad_norm": 7.011876209354938,
"learning_rate": 1.4429914727478526e-07,
"logits/chosen": -1.693228840827942,
"logits/rejected": -1.6980655193328857,
"logps/chosen": -560.333984375,
"logps/rejected": -595.471435546875,
"loss": 0.6033,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.7711371779441833,
"rewards/margins": 0.37155580520629883,
"rewards/rejected": -1.142693042755127,
"step": 326
},
{
"epoch": 0.2135684545677198,
"grad_norm": 8.197929978948176,
"learning_rate": 1.4423353082803705e-07,
"logits/chosen": -1.6932696104049683,
"logits/rejected": -1.7340316772460938,
"logps/chosen": -592.4081420898438,
"logps/rejected": -626.7181396484375,
"loss": 0.6079,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.9063585996627808,
"rewards/margins": 0.23381030559539795,
"rewards/rejected": -1.1401689052581787,
"step": 327
},
{
"epoch": 0.21422156910768225,
"grad_norm": 21.483660249834884,
"learning_rate": 1.44167554011681e-07,
"logits/chosen": -1.6718993186950684,
"logits/rejected": -1.6670811176300049,
"logps/chosen": -570.2409057617188,
"logps/rejected": -584.1919555664062,
"loss": 0.6376,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.6868249177932739,
"rewards/margins": 0.22840677201747894,
"rewards/rejected": -0.9152317643165588,
"step": 328
},
{
"epoch": 0.2148746836476447,
"grad_norm": 16.130012234299603,
"learning_rate": 1.4410121716913508e-07,
"logits/chosen": -1.6927289962768555,
"logits/rejected": -1.70664644241333,
"logps/chosen": -617.483642578125,
"logps/rejected": -620.2203369140625,
"loss": 0.6259,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.9101148247718811,
"rewards/margins": 0.1278352588415146,
"rewards/rejected": -1.0379501581192017,
"step": 329
},
{
"epoch": 0.21552779818760714,
"grad_norm": 8.367739313272205,
"learning_rate": 1.4403452064569127e-07,
"logits/chosen": -1.6701712608337402,
"logits/rejected": -1.7202551364898682,
"logps/chosen": -661.21435546875,
"logps/rejected": -602.512451171875,
"loss": 0.6385,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.1165893077850342,
"rewards/margins": 0.08732178062200546,
"rewards/rejected": -1.2039111852645874,
"step": 330
},
{
"epoch": 0.2161809127275696,
"grad_norm": 9.035063267419801,
"learning_rate": 1.439674647885137e-07,
"logits/chosen": -1.6763889789581299,
"logits/rejected": -1.6751902103424072,
"logps/chosen": -613.0372314453125,
"logps/rejected": -630.5709838867188,
"loss": 0.6134,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.0509929656982422,
"rewards/margins": 0.17839497327804565,
"rewards/rejected": -1.2293879985809326,
"step": 331
},
{
"epoch": 0.21683402726753204,
"grad_norm": 9.470243202935096,
"learning_rate": 1.439000499466369e-07,
"logits/chosen": -1.7101950645446777,
"logits/rejected": -1.693764328956604,
"logps/chosen": -592.8642578125,
"logps/rejected": -629.7588500976562,
"loss": 0.6173,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.9428937435150146,
"rewards/margins": 0.1368747353553772,
"rewards/rejected": -1.0797685384750366,
"step": 332
},
{
"epoch": 0.21748714180749448,
"grad_norm": 11.187409550489294,
"learning_rate": 1.4383227647096393e-07,
"logits/chosen": -1.651065707206726,
"logits/rejected": -1.6519496440887451,
"logps/chosen": -613.1717529296875,
"logps/rejected": -579.97265625,
"loss": 0.624,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.9745721817016602,
"rewards/margins": 0.17566874623298645,
"rewards/rejected": -1.1502408981323242,
"step": 333
},
{
"epoch": 0.21814025634745693,
"grad_norm": 49.216595878368345,
"learning_rate": 1.4376414471426472e-07,
"logits/chosen": -1.7066453695297241,
"logits/rejected": -1.7159579992294312,
"logps/chosen": -645.9775390625,
"logps/rejected": -741.1705932617188,
"loss": 0.6203,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.0998951196670532,
"rewards/margins": 0.478588730096817,
"rewards/rejected": -1.5784838199615479,
"step": 334
},
{
"epoch": 0.21879337088741938,
"grad_norm": 16.168657963184085,
"learning_rate": 1.436956550311739e-07,
"logits/chosen": -1.7134239673614502,
"logits/rejected": -1.7491358518600464,
"logps/chosen": -618.333251953125,
"logps/rejected": -704.8095092773438,
"loss": 0.613,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.038926124572754,
"rewards/margins": 0.38716238737106323,
"rewards/rejected": -1.426088571548462,
"step": 335
},
{
"epoch": 0.21944648542738182,
"grad_norm": 15.399404851620021,
"learning_rate": 1.4362680777818932e-07,
"logits/chosen": -1.6947367191314697,
"logits/rejected": -1.7013921737670898,
"logps/chosen": -616.1701049804688,
"logps/rejected": -646.965576171875,
"loss": 0.6476,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.009097695350647,
"rewards/margins": 0.12043975293636322,
"rewards/rejected": -1.1295373439788818,
"step": 336
},
{
"epoch": 0.22009959996734427,
"grad_norm": 12.377040896457764,
"learning_rate": 1.435576033136699e-07,
"logits/chosen": -1.72658109664917,
"logits/rejected": -1.7111766338348389,
"logps/chosen": -691.18701171875,
"logps/rejected": -746.902587890625,
"loss": 0.6108,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.1553032398223877,
"rewards/margins": 0.3523827791213989,
"rewards/rejected": -1.507685899734497,
"step": 337
},
{
"epoch": 0.2207527145073067,
"grad_norm": 21.708794576351895,
"learning_rate": 1.4348804199783397e-07,
"logits/chosen": -1.7045814990997314,
"logits/rejected": -1.733116865158081,
"logps/chosen": -591.8544921875,
"logps/rejected": -598.6890869140625,
"loss": 0.6173,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.881680428981781,
"rewards/margins": 0.18818916380405426,
"rewards/rejected": -1.0698695182800293,
"step": 338
},
{
"epoch": 0.22140582904726916,
"grad_norm": 12.368369376917613,
"learning_rate": 1.4341812419275735e-07,
"logits/chosen": -1.7070914506912231,
"logits/rejected": -1.7090568542480469,
"logps/chosen": -598.9154663085938,
"logps/rejected": -546.4701538085938,
"loss": 0.6502,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.9902876019477844,
"rewards/margins": 0.08627556264400482,
"rewards/rejected": -1.0765631198883057,
"step": 339
},
{
"epoch": 0.2220589435872316,
"grad_norm": 7.332074698031059,
"learning_rate": 1.4334785026237135e-07,
"logits/chosen": -1.757274866104126,
"logits/rejected": -1.7839263677597046,
"logps/chosen": -611.2498779296875,
"logps/rejected": -621.130859375,
"loss": 0.5915,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.0745751857757568,
"rewards/margins": 0.25732651352882385,
"rewards/rejected": -1.3319017887115479,
"step": 340
},
{
"epoch": 0.22271205812719405,
"grad_norm": 16.45318119474388,
"learning_rate": 1.43277220572461e-07,
"logits/chosen": -1.5772120952606201,
"logits/rejected": -1.5908713340759277,
"logps/chosen": -537.1170043945312,
"logps/rejected": -550.2281494140625,
"loss": 0.6306,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.9023160338401794,
"rewards/margins": 0.1352323591709137,
"rewards/rejected": -1.037548303604126,
"step": 341
},
{
"epoch": 0.2233651726671565,
"grad_norm": 37.88300298632188,
"learning_rate": 1.4320623549066308e-07,
"logits/chosen": -1.7355406284332275,
"logits/rejected": -1.7195450067520142,
"logps/chosen": -686.9090576171875,
"logps/rejected": -739.3055419921875,
"loss": 0.579,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.1977063417434692,
"rewards/margins": 0.5113601088523865,
"rewards/rejected": -1.7090665102005005,
"step": 342
},
{
"epoch": 0.22401828720711894,
"grad_norm": 22.776462467379197,
"learning_rate": 1.4313489538646427e-07,
"logits/chosen": -1.7235167026519775,
"logits/rejected": -1.7320938110351562,
"logps/chosen": -663.5242309570312,
"logps/rejected": -628.5629272460938,
"loss": 0.6169,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.0688875913619995,
"rewards/margins": 0.16213612258434296,
"rewards/rejected": -1.2310236692428589,
"step": 343
},
{
"epoch": 0.2246714017470814,
"grad_norm": 8.488895496301154,
"learning_rate": 1.4306320063119916e-07,
"logits/chosen": -1.6490345001220703,
"logits/rejected": -1.6467983722686768,
"logps/chosen": -686.3357543945312,
"logps/rejected": -679.5280151367188,
"loss": 0.6362,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.28061044216156,
"rewards/margins": 0.09073764830827713,
"rewards/rejected": -1.371348261833191,
"step": 344
},
{
"epoch": 0.22532451628704384,
"grad_norm": 22.168028050033644,
"learning_rate": 1.4299115159804836e-07,
"logits/chosen": -1.7651727199554443,
"logits/rejected": -1.756993293762207,
"logps/chosen": -707.9656372070312,
"logps/rejected": -787.037109375,
"loss": 0.6355,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.312680959701538,
"rewards/margins": 0.39120543003082275,
"rewards/rejected": -1.7038863897323608,
"step": 345
},
{
"epoch": 0.22597763082700628,
"grad_norm": 7.58865714123304,
"learning_rate": 1.4291874866203655e-07,
"logits/chosen": -1.659616231918335,
"logits/rejected": -1.637772798538208,
"logps/chosen": -658.1541137695312,
"logps/rejected": -791.5831909179688,
"loss": 0.5703,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.1292518377304077,
"rewards/margins": 0.6298790574073792,
"rewards/rejected": -1.759130835533142,
"step": 346
},
{
"epoch": 0.22663074536696873,
"grad_norm": 26.071692354910912,
"learning_rate": 1.428459922000305e-07,
"logits/chosen": -1.5885182619094849,
"logits/rejected": -1.6136213541030884,
"logps/chosen": -555.8087158203125,
"logps/rejected": -621.08349609375,
"loss": 0.5983,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.9678431749343872,
"rewards/margins": 0.3450620472431183,
"rewards/rejected": -1.312905192375183,
"step": 347
},
{
"epoch": 0.22728385990693117,
"grad_norm": 7.8840447097778545,
"learning_rate": 1.4277288259073708e-07,
"logits/chosen": -1.568705439567566,
"logits/rejected": -1.5865546464920044,
"logps/chosen": -589.7760009765625,
"logps/rejected": -573.239990234375,
"loss": 0.6277,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.1354511976242065,
"rewards/margins": 0.18996107578277588,
"rewards/rejected": -1.3254122734069824,
"step": 348
},
{
"epoch": 0.22793697444689362,
"grad_norm": 10.92884715777762,
"learning_rate": 1.4269942021470148e-07,
"logits/chosen": -1.6641845703125,
"logits/rejected": -1.6769914627075195,
"logps/chosen": -672.136474609375,
"logps/rejected": -664.6466674804688,
"loss": 0.6121,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.0376440286636353,
"rewards/margins": 0.1383248120546341,
"rewards/rejected": -1.1759687662124634,
"step": 349
},
{
"epoch": 0.22859008898685607,
"grad_norm": 31.26872455787033,
"learning_rate": 1.4262560545430495e-07,
"logits/chosen": -1.5913841724395752,
"logits/rejected": -1.6360561847686768,
"logps/chosen": -617.2501831054688,
"logps/rejected": -550.8568115234375,
"loss": 0.6164,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.979924738407135,
"rewards/margins": 0.24141384661197662,
"rewards/rejected": -1.2213386297225952,
"step": 350
},
{
"epoch": 0.2292432035268185,
"grad_norm": 8.368119693346522,
"learning_rate": 1.4255143869376301e-07,
"logits/chosen": -1.7337816953659058,
"logits/rejected": -1.6971759796142578,
"logps/chosen": -652.0234375,
"logps/rejected": -777.8805541992188,
"loss": 0.5913,
"rewards/accuracies": 0.875,
"rewards/chosen": -1.1779980659484863,
"rewards/margins": 0.6400733590126038,
"rewards/rejected": -1.8180711269378662,
"step": 351
},
{
"epoch": 0.22989631806678096,
"grad_norm": 7.04954036164927,
"learning_rate": 1.424769203191234e-07,
"logits/chosen": -1.706908941268921,
"logits/rejected": -1.7061164379119873,
"logps/chosen": -648.4498291015625,
"logps/rejected": -606.88623046875,
"loss": 0.6278,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.2371093034744263,
"rewards/margins": 0.11428666114807129,
"rewards/rejected": -1.3513959646224976,
"step": 352
},
{
"epoch": 0.2305494326067434,
"grad_norm": 24.647751119336064,
"learning_rate": 1.42402050718264e-07,
"logits/chosen": -1.6741199493408203,
"logits/rejected": -1.743821620941162,
"logps/chosen": -677.9364624023438,
"logps/rejected": -710.4935302734375,
"loss": 0.5985,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.0869948863983154,
"rewards/margins": 0.1914815604686737,
"rewards/rejected": -1.278476357460022,
"step": 353
},
{
"epoch": 0.23120254714670585,
"grad_norm": 32.99725275222979,
"learning_rate": 1.4232683028089092e-07,
"logits/chosen": -1.6567903757095337,
"logits/rejected": -1.6555206775665283,
"logps/chosen": -587.80029296875,
"logps/rejected": -598.2508544921875,
"loss": 0.6307,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.024768352508545,
"rewards/margins": 0.3201058804988861,
"rewards/rejected": -1.3448742628097534,
"step": 354
},
{
"epoch": 0.2318556616866683,
"grad_norm": 15.301036178500334,
"learning_rate": 1.4225125939853637e-07,
"logits/chosen": -1.5989353656768799,
"logits/rejected": -1.6277306079864502,
"logps/chosen": -566.171875,
"logps/rejected": -590.93505859375,
"loss": 0.6273,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.9104868173599243,
"rewards/margins": 0.22141122817993164,
"rewards/rejected": -1.1318979263305664,
"step": 355
},
{
"epoch": 0.23250877622663074,
"grad_norm": 8.150144656099924,
"learning_rate": 1.4217533846455675e-07,
"logits/chosen": -1.6726055145263672,
"logits/rejected": -1.6693233251571655,
"logps/chosen": -643.8858642578125,
"logps/rejected": -628.0254516601562,
"loss": 0.6466,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.2304469347000122,
"rewards/margins": 0.19707579910755157,
"rewards/rejected": -1.427522897720337,
"step": 356
},
{
"epoch": 0.2331618907665932,
"grad_norm": 33.25010241380852,
"learning_rate": 1.4209906787413047e-07,
"logits/chosen": -1.6271083354949951,
"logits/rejected": -1.6530542373657227,
"logps/chosen": -599.4658813476562,
"logps/rejected": -608.0745849609375,
"loss": 0.5938,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.1141527891159058,
"rewards/margins": 0.19912773370742798,
"rewards/rejected": -1.3132805824279785,
"step": 357
},
{
"epoch": 0.23381500530655563,
"grad_norm": 14.540966438844277,
"learning_rate": 1.420224480242559e-07,
"logits/chosen": -1.5631732940673828,
"logits/rejected": -1.582556128501892,
"logps/chosen": -584.55712890625,
"logps/rejected": -616.7429809570312,
"loss": 0.6118,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.0560500621795654,
"rewards/margins": 0.2983832359313965,
"rewards/rejected": -1.354433298110962,
"step": 358
},
{
"epoch": 0.23446811984651808,
"grad_norm": 11.291749985668568,
"learning_rate": 1.4194547931374948e-07,
"logits/chosen": -1.6868391036987305,
"logits/rejected": -1.7105621099472046,
"logps/chosen": -600.7924194335938,
"logps/rejected": -677.1724853515625,
"loss": 0.6779,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.23578941822052,
"rewards/margins": 0.33331653475761414,
"rewards/rejected": -1.569106101989746,
"step": 359
},
{
"epoch": 0.23512123438648053,
"grad_norm": 22.901901301811037,
"learning_rate": 1.418681621432434e-07,
"logits/chosen": -1.574560523033142,
"logits/rejected": -1.599095106124878,
"logps/chosen": -599.9710693359375,
"logps/rejected": -652.0952758789062,
"loss": 0.6254,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.1917017698287964,
"rewards/margins": 0.3461110591888428,
"rewards/rejected": -1.5378127098083496,
"step": 360
},
{
"epoch": 0.23577434892644297,
"grad_norm": 25.59056348020549,
"learning_rate": 1.417904969151837e-07,
"logits/chosen": -1.7425224781036377,
"logits/rejected": -1.7415771484375,
"logps/chosen": -704.6986694335938,
"logps/rejected": -693.380126953125,
"loss": 0.6119,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.2282774448394775,
"rewards/margins": 0.26026061177253723,
"rewards/rejected": -1.488538146018982,
"step": 361
},
{
"epoch": 0.23642746346640542,
"grad_norm": 8.463533236278087,
"learning_rate": 1.4171248403382806e-07,
"logits/chosen": -1.6078510284423828,
"logits/rejected": -1.700717568397522,
"logps/chosen": -629.0587768554688,
"logps/rejected": -598.7468872070312,
"loss": 0.6081,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.0717506408691406,
"rewards/margins": 0.29170456528663635,
"rewards/rejected": -1.3634551763534546,
"step": 362
},
{
"epoch": 0.23708057800636786,
"grad_norm": 27.979305395002925,
"learning_rate": 1.4163412390524378e-07,
"logits/chosen": -1.6555612087249756,
"logits/rejected": -1.701103925704956,
"logps/chosen": -515.6534423828125,
"logps/rejected": -615.6427001953125,
"loss": 0.5931,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.1157307624816895,
"rewards/margins": 0.35381343960762024,
"rewards/rejected": -1.4695440530776978,
"step": 363
},
{
"epoch": 0.2377336925463303,
"grad_norm": 29.822643828567546,
"learning_rate": 1.4155541693730556e-07,
"logits/chosen": -1.6813297271728516,
"logits/rejected": -1.6906770467758179,
"logps/chosen": -571.6758422851562,
"logps/rejected": -599.9418334960938,
"loss": 0.5808,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.0341238975524902,
"rewards/margins": 0.39114344120025635,
"rewards/rejected": -1.425267219543457,
"step": 364
},
{
"epoch": 0.23838680708629276,
"grad_norm": 8.151382967050962,
"learning_rate": 1.414763635396935e-07,
"logits/chosen": -1.6014394760131836,
"logits/rejected": -1.5656381845474243,
"logps/chosen": -540.3904418945312,
"logps/rejected": -593.1724853515625,
"loss": 0.6379,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.1642670631408691,
"rewards/margins": 0.3421732187271118,
"rewards/rejected": -1.506440281867981,
"step": 365
},
{
"epoch": 0.2390399216262552,
"grad_norm": 57.74463360143876,
"learning_rate": 1.4139696412389096e-07,
"logits/chosen": -1.7312068939208984,
"logits/rejected": -1.7102075815200806,
"logps/chosen": -643.19091796875,
"logps/rejected": -672.47509765625,
"loss": 0.5735,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.1425721645355225,
"rewards/margins": 0.4072319269180298,
"rewards/rejected": -1.5498042106628418,
"step": 366
},
{
"epoch": 0.23969303616621765,
"grad_norm": 12.197736908947663,
"learning_rate": 1.4131721910318227e-07,
"logits/chosen": -1.8108876943588257,
"logits/rejected": -1.7958606481552124,
"logps/chosen": -733.1993408203125,
"logps/rejected": -763.6776733398438,
"loss": 0.5912,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.482933759689331,
"rewards/margins": 0.2555669844150543,
"rewards/rejected": -1.7385008335113525,
"step": 367
},
{
"epoch": 0.2403461507061801,
"grad_norm": 19.29177394775571,
"learning_rate": 1.4123712889265072e-07,
"logits/chosen": -1.6916940212249756,
"logits/rejected": -1.7217150926589966,
"logps/chosen": -671.9423217773438,
"logps/rejected": -675.45361328125,
"loss": 0.5719,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.3399765491485596,
"rewards/margins": 0.20047269761562347,
"rewards/rejected": -1.5404491424560547,
"step": 368
},
{
"epoch": 0.24099926524614254,
"grad_norm": 25.260138658321655,
"learning_rate": 1.4115669390917636e-07,
"logits/chosen": -1.6250600814819336,
"logits/rejected": -1.6512022018432617,
"logps/chosen": -625.4688720703125,
"logps/rejected": -614.0571899414062,
"loss": 0.5783,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.2662508487701416,
"rewards/margins": 0.20612508058547974,
"rewards/rejected": -1.4723761081695557,
"step": 369
},
{
"epoch": 0.241652379786105,
"grad_norm": 9.803497546984937,
"learning_rate": 1.4107591457143383e-07,
"logits/chosen": -1.5895332098007202,
"logits/rejected": -1.602332592010498,
"logps/chosen": -675.6551513671875,
"logps/rejected": -709.5709838867188,
"loss": 0.5924,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.2006454467773438,
"rewards/margins": 0.3013772666454315,
"rewards/rejected": -1.5020227432250977,
"step": 370
},
{
"epoch": 0.24230549432606743,
"grad_norm": 8.545956638134795,
"learning_rate": 1.409947912998902e-07,
"logits/chosen": -1.6708649396896362,
"logits/rejected": -1.713352918624878,
"logps/chosen": -680.72607421875,
"logps/rejected": -674.447265625,
"loss": 0.6127,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.2759678363800049,
"rewards/margins": 0.18677686154842377,
"rewards/rejected": -1.4627448320388794,
"step": 371
},
{
"epoch": 0.24295860886602988,
"grad_norm": 19.844846501689318,
"learning_rate": 1.4091332451680267e-07,
"logits/chosen": -1.7042357921600342,
"logits/rejected": -1.7265393733978271,
"logps/chosen": -675.3814086914062,
"logps/rejected": -644.4542846679688,
"loss": 0.6427,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.4431321620941162,
"rewards/margins": 0.07462039589881897,
"rewards/rejected": -1.5177525281906128,
"step": 372
},
{
"epoch": 0.24361172340599233,
"grad_norm": 7.920082953904519,
"learning_rate": 1.408315146462166e-07,
"logits/chosen": -1.721234679222107,
"logits/rejected": -1.701398253440857,
"logps/chosen": -589.4611206054688,
"logps/rejected": -710.4976806640625,
"loss": 0.581,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.1915465593338013,
"rewards/margins": 0.42795515060424805,
"rewards/rejected": -1.6195017099380493,
"step": 373
},
{
"epoch": 0.24426483794595477,
"grad_norm": 12.858984927661352,
"learning_rate": 1.407493621139631e-07,
"logits/chosen": -1.713492751121521,
"logits/rejected": -1.7290925979614258,
"logps/chosen": -641.5909423828125,
"logps/rejected": -650.439697265625,
"loss": 0.5713,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.2154250144958496,
"rewards/margins": 0.2901778817176819,
"rewards/rejected": -1.5056030750274658,
"step": 374
},
{
"epoch": 0.24491795248591722,
"grad_norm": 15.690657976850302,
"learning_rate": 1.406668673476568e-07,
"logits/chosen": -1.6640945672988892,
"logits/rejected": -1.6718759536743164,
"logps/chosen": -649.737060546875,
"logps/rejected": -686.3532104492188,
"loss": 0.568,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.3824304342269897,
"rewards/margins": 0.3402037024497986,
"rewards/rejected": -1.722634196281433,
"step": 375
},
{
"epoch": 0.24557106702587966,
"grad_norm": 25.733728918193314,
"learning_rate": 1.4058403077669386e-07,
"logits/chosen": -1.671245813369751,
"logits/rejected": -1.6539918184280396,
"logps/chosen": -620.85791015625,
"logps/rejected": -613.2889404296875,
"loss": 0.5943,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.2888745069503784,
"rewards/margins": 0.22225916385650635,
"rewards/rejected": -1.5111336708068848,
"step": 376
},
{
"epoch": 0.2462241815658421,
"grad_norm": 9.763117553021695,
"learning_rate": 1.4050085283224946e-07,
"logits/chosen": -1.5626500844955444,
"logits/rejected": -1.57936429977417,
"logps/chosen": -653.4628295898438,
"logps/rejected": -741.0684814453125,
"loss": 0.6487,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4431822299957275,
"rewards/margins": 0.3858519494533539,
"rewards/rejected": -1.8290340900421143,
"step": 377
},
{
"epoch": 0.24687729610580456,
"grad_norm": 14.715453079588329,
"learning_rate": 1.4041733394727567e-07,
"logits/chosen": -1.6787821054458618,
"logits/rejected": -1.675290584564209,
"logps/chosen": -656.50830078125,
"logps/rejected": -641.9553833007812,
"loss": 0.5851,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.2514023780822754,
"rewards/margins": 0.16022822260856628,
"rewards/rejected": -1.411630630493164,
"step": 378
},
{
"epoch": 0.247530410645767,
"grad_norm": 17.584914736775964,
"learning_rate": 1.403334745564993e-07,
"logits/chosen": -1.6352498531341553,
"logits/rejected": -1.6839510202407837,
"logps/chosen": -743.545654296875,
"logps/rejected": -792.6776123046875,
"loss": 0.587,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.677843689918518,
"rewards/margins": 0.3219238221645355,
"rewards/rejected": -1.999767541885376,
"step": 379
},
{
"epoch": 0.24818352518572945,
"grad_norm": 35.19811670438595,
"learning_rate": 1.4024927509641947e-07,
"logits/chosen": -1.7035257816314697,
"logits/rejected": -1.6846370697021484,
"logps/chosen": -731.2345581054688,
"logps/rejected": -747.091064453125,
"loss": 0.5723,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.520798921585083,
"rewards/margins": 0.359272301197052,
"rewards/rejected": -1.8800714015960693,
"step": 380
},
{
"epoch": 0.2488366397256919,
"grad_norm": 51.54152025955344,
"learning_rate": 1.401647360053054e-07,
"logits/chosen": -1.6718703508377075,
"logits/rejected": -1.7071422338485718,
"logps/chosen": -684.9619140625,
"logps/rejected": -685.8736572265625,
"loss": 0.6127,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.1913833618164062,
"rewards/margins": 0.4190147817134857,
"rewards/rejected": -1.6103980541229248,
"step": 381
},
{
"epoch": 0.24948975426565434,
"grad_norm": 29.30838787108767,
"learning_rate": 1.4007985772319414e-07,
"logits/chosen": -1.58506441116333,
"logits/rejected": -1.576029896736145,
"logps/chosen": -592.5045166015625,
"logps/rejected": -606.4868774414062,
"loss": 0.57,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.3507953882217407,
"rewards/margins": 0.4049099385738373,
"rewards/rejected": -1.7557053565979004,
"step": 382
},
{
"epoch": 0.25014286880561676,
"grad_norm": 30.296371441246823,
"learning_rate": 1.3999464069188827e-07,
"logits/chosen": -1.511520504951477,
"logits/rejected": -1.5346118211746216,
"logps/chosen": -610.6934814453125,
"logps/rejected": -583.274658203125,
"loss": 0.5848,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.2899129390716553,
"rewards/margins": 0.24026429653167725,
"rewards/rejected": -1.5301774740219116,
"step": 383
},
{
"epoch": 0.25079598334557923,
"grad_norm": 22.072613902976816,
"learning_rate": 1.3990908535495366e-07,
"logits/chosen": -1.6665383577346802,
"logits/rejected": -1.7260501384735107,
"logps/chosen": -649.5576782226562,
"logps/rejected": -632.3207397460938,
"loss": 0.5902,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.3579076528549194,
"rewards/margins": 0.13580122590065002,
"rewards/rejected": -1.493708848953247,
"step": 384
},
{
"epoch": 0.25144909788554165,
"grad_norm": 11.159048199299994,
"learning_rate": 1.39823192157717e-07,
"logits/chosen": -1.7021327018737793,
"logits/rejected": -1.7216562032699585,
"logps/chosen": -702.6673583984375,
"logps/rejected": -677.152587890625,
"loss": 0.6105,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.3912357091903687,
"rewards/margins": 0.28019946813583374,
"rewards/rejected": -1.6714351177215576,
"step": 385
},
{
"epoch": 0.2521022124255041,
"grad_norm": 12.67256793359313,
"learning_rate": 1.3973696154726372e-07,
"logits/chosen": -1.652302622795105,
"logits/rejected": -1.6712439060211182,
"logps/chosen": -615.6753540039062,
"logps/rejected": -612.9768676757812,
"loss": 0.535,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.381615400314331,
"rewards/margins": 0.35143721103668213,
"rewards/rejected": -1.7330526113510132,
"step": 386
},
{
"epoch": 0.25275532696546654,
"grad_norm": 9.156751796221311,
"learning_rate": 1.396503939724354e-07,
"logits/chosen": -1.687638521194458,
"logits/rejected": -1.6907660961151123,
"logps/chosen": -679.770263671875,
"logps/rejected": -704.907470703125,
"loss": 0.581,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.3277437686920166,
"rewards/margins": 0.3244020938873291,
"rewards/rejected": -1.6521457433700562,
"step": 387
},
{
"epoch": 0.253408441505429,
"grad_norm": 26.505065205334525,
"learning_rate": 1.3956348988382757e-07,
"logits/chosen": -1.621063470840454,
"logits/rejected": -1.6041771173477173,
"logps/chosen": -596.8775634765625,
"logps/rejected": -675.8204345703125,
"loss": 0.5536,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4424089193344116,
"rewards/margins": 0.5358285903930664,
"rewards/rejected": -1.9782376289367676,
"step": 388
},
{
"epoch": 0.25406155604539143,
"grad_norm": 7.262480595814242,
"learning_rate": 1.394762497337875e-07,
"logits/chosen": -1.6179628372192383,
"logits/rejected": -1.6096928119659424,
"logps/chosen": -653.6219482421875,
"logps/rejected": -699.051513671875,
"loss": 0.6154,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.4057137966156006,
"rewards/margins": 0.2138241082429886,
"rewards/rejected": -1.6195377111434937,
"step": 389
},
{
"epoch": 0.2547146705853539,
"grad_norm": 17.339973588325055,
"learning_rate": 1.393886739764116e-07,
"logits/chosen": -1.6001747846603394,
"logits/rejected": -1.6313351392745972,
"logps/chosen": -626.1659545898438,
"logps/rejected": -644.4179077148438,
"loss": 0.5469,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.3461098670959473,
"rewards/margins": 0.41640228033065796,
"rewards/rejected": -1.76251220703125,
"step": 390
},
{
"epoch": 0.2553677851253163,
"grad_norm": 14.339393954095954,
"learning_rate": 1.3930076306754315e-07,
"logits/chosen": -1.5895658731460571,
"logits/rejected": -1.6347410678863525,
"logps/chosen": -611.0591430664062,
"logps/rejected": -595.1077880859375,
"loss": 0.5883,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.2730367183685303,
"rewards/margins": 0.2380126416683197,
"rewards/rejected": -1.5110492706298828,
"step": 391
},
{
"epoch": 0.2560208996652788,
"grad_norm": 10.18368249263564,
"learning_rate": 1.3921251746476998e-07,
"logits/chosen": -1.7359168529510498,
"logits/rejected": -1.7397726774215698,
"logps/chosen": -588.6852416992188,
"logps/rejected": -697.5007934570312,
"loss": 0.5732,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.3043936491012573,
"rewards/margins": 0.6203251481056213,
"rewards/rejected": -1.9247188568115234,
"step": 392
},
{
"epoch": 0.2566740142052412,
"grad_norm": 18.38247483748203,
"learning_rate": 1.39123937627422e-07,
"logits/chosen": -1.6276944875717163,
"logits/rejected": -1.5889580249786377,
"logps/chosen": -620.385498046875,
"logps/rejected": -635.3296508789062,
"loss": 0.6444,
"rewards/accuracies": 0.5,
"rewards/chosen": -1.3040493726730347,
"rewards/margins": 0.10177471488714218,
"rewards/rejected": -1.4058241844177246,
"step": 393
},
{
"epoch": 0.2573271287452037,
"grad_norm": 36.73565911642949,
"learning_rate": 1.390350240165689e-07,
"logits/chosen": -1.6098779439926147,
"logits/rejected": -1.7416176795959473,
"logps/chosen": -642.3756713867188,
"logps/rejected": -643.9348754882812,
"loss": 0.5877,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.407647728919983,
"rewards/margins": 0.2774192690849304,
"rewards/rejected": -1.685067057609558,
"step": 394
},
{
"epoch": 0.2579802432851661,
"grad_norm": 16.023490486119584,
"learning_rate": 1.3894577709501766e-07,
"logits/chosen": -1.692483901977539,
"logits/rejected": -1.7297693490982056,
"logps/chosen": -797.02880859375,
"logps/rejected": -657.3602294921875,
"loss": 0.7438,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8514511585235596,
"rewards/margins": -0.006901424378156662,
"rewards/rejected": -1.8445496559143066,
"step": 395
},
{
"epoch": 0.2586333578251286,
"grad_norm": 16.55592977846572,
"learning_rate": 1.3885619732731024e-07,
"logits/chosen": -1.6773359775543213,
"logits/rejected": -1.7305378913879395,
"logps/chosen": -627.8666381835938,
"logps/rejected": -695.2501220703125,
"loss": 0.5468,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.4438410997390747,
"rewards/margins": 0.509320855140686,
"rewards/rejected": -1.9531618356704712,
"step": 396
},
{
"epoch": 0.259286472365091,
"grad_norm": 31.880923432118284,
"learning_rate": 1.3876628517972106e-07,
"logits/chosen": -1.5321552753448486,
"logits/rejected": -1.52942955493927,
"logps/chosen": -627.0950317382812,
"logps/rejected": -674.76025390625,
"loss": 0.5836,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.53360915184021,
"rewards/margins": 0.2759949266910553,
"rewards/rejected": -1.8096040487289429,
"step": 397
},
{
"epoch": 0.2599395869050535,
"grad_norm": 13.538949029727982,
"learning_rate": 1.3867604112025465e-07,
"logits/chosen": -1.6481291055679321,
"logits/rejected": -1.6133708953857422,
"logps/chosen": -633.3253173828125,
"logps/rejected": -659.7344970703125,
"loss": 0.6281,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.4847848415374756,
"rewards/margins": 0.1952800452709198,
"rewards/rejected": -1.6800646781921387,
"step": 398
},
{
"epoch": 0.2605927014450159,
"grad_norm": 41.823948195416314,
"learning_rate": 1.3858546561864315e-07,
"logits/chosen": -1.6834887266159058,
"logits/rejected": -1.668868064880371,
"logps/chosen": -664.4490966796875,
"logps/rejected": -718.4891357421875,
"loss": 0.5855,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5809217691421509,
"rewards/margins": 0.4111039936542511,
"rewards/rejected": -1.9920259714126587,
"step": 399
},
{
"epoch": 0.26124581598497837,
"grad_norm": 7.562874624386567,
"learning_rate": 1.3849455914634399e-07,
"logits/chosen": -1.6510210037231445,
"logits/rejected": -1.6854655742645264,
"logps/chosen": -694.1316528320312,
"logps/rejected": -653.6522216796875,
"loss": 0.5831,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.8670495748519897,
"rewards/margins": 0.05141986906528473,
"rewards/rejected": -1.9184695482254028,
"step": 400
},
{
"epoch": 0.26124581598497837,
"eval_logits/chosen": -1.6877070665359497,
"eval_logits/rejected": -1.6963342428207397,
"eval_logps/chosen": -661.8920288085938,
"eval_logps/rejected": -689.978759765625,
"eval_loss": 0.5931771397590637,
"eval_rewards/accuracies": 0.7070000171661377,
"eval_rewards/chosen": -1.515466332435608,
"eval_rewards/margins": 0.36192643642425537,
"eval_rewards/rejected": -1.8773927688598633,
"eval_runtime": 300.4429,
"eval_samples_per_second": 13.314,
"eval_steps_per_second": 0.832,
"step": 400
},
{
"epoch": 0.2618989305249408,
"grad_norm": 9.182566410300494,
"learning_rate": 1.3840332217653723e-07,
"logits/chosen": -1.673874020576477,
"logits/rejected": -1.680929183959961,
"logps/chosen": -692.4191284179688,
"logps/rejected": -704.0576171875,
"loss": 0.6425,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.688931941986084,
"rewards/margins": 0.1834324449300766,
"rewards/rejected": -1.8723644018173218,
"step": 401
},
{
"epoch": 0.26255204506490326,
"grad_norm": 8.101445637974606,
"learning_rate": 1.3831175518412327e-07,
"logits/chosen": -1.6620182991027832,
"logits/rejected": -1.6938483715057373,
"logps/chosen": -745.3438720703125,
"logps/rejected": -760.4389038085938,
"loss": 0.6006,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.6187502145767212,
"rewards/margins": 0.23492467403411865,
"rewards/rejected": -1.8536747694015503,
"step": 402
},
{
"epoch": 0.2632051596048657,
"grad_norm": 10.95532049184201,
"learning_rate": 1.3821985864572028e-07,
"logits/chosen": -1.6001207828521729,
"logits/rejected": -1.565683126449585,
"logps/chosen": -678.5340576171875,
"logps/rejected": -816.8855590820312,
"loss": 0.6223,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.7041336297988892,
"rewards/margins": 0.5798905491828918,
"rewards/rejected": -2.284024238586426,
"step": 403
},
{
"epoch": 0.26385827414482815,
"grad_norm": 19.544045940040125,
"learning_rate": 1.3812763303966186e-07,
"logits/chosen": -1.6810550689697266,
"logits/rejected": -1.6656105518341064,
"logps/chosen": -635.9982299804688,
"logps/rejected": -654.9970703125,
"loss": 0.5854,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.5096485614776611,
"rewards/margins": 0.24476324021816254,
"rewards/rejected": -1.7544115781784058,
"step": 404
},
{
"epoch": 0.26451138868479057,
"grad_norm": 8.792026250775443,
"learning_rate": 1.3803507884599438e-07,
"logits/chosen": -1.693314552307129,
"logits/rejected": -1.678006887435913,
"logps/chosen": -794.29443359375,
"logps/rejected": -808.3348388671875,
"loss": 0.5861,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.7267074584960938,
"rewards/margins": 0.24482649564743042,
"rewards/rejected": -1.9715338945388794,
"step": 405
},
{
"epoch": 0.26516450322475305,
"grad_norm": 39.2544432402473,
"learning_rate": 1.379421965464745e-07,
"logits/chosen": -1.672491431236267,
"logits/rejected": -1.6746456623077393,
"logps/chosen": -739.5106811523438,
"logps/rejected": -728.4205322265625,
"loss": 0.611,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.7484608888626099,
"rewards/margins": 0.24220341444015503,
"rewards/rejected": -1.9906642436981201,
"step": 406
},
{
"epoch": 0.26581761776471546,
"grad_norm": 61.06487102613164,
"learning_rate": 1.378489866245668e-07,
"logits/chosen": -1.6239569187164307,
"logits/rejected": -1.6054996252059937,
"logps/chosen": -597.54833984375,
"logps/rejected": -692.744384765625,
"loss": 0.5651,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.441511631011963,
"rewards/margins": 0.5937218070030212,
"rewards/rejected": -2.035233497619629,
"step": 407
},
{
"epoch": 0.26647073230467794,
"grad_norm": 9.653358684629563,
"learning_rate": 1.3775544956544115e-07,
"logits/chosen": -1.6666955947875977,
"logits/rejected": -1.6608633995056152,
"logps/chosen": -721.04541015625,
"logps/rejected": -931.44482421875,
"loss": 0.519,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.6330653429031372,
"rewards/margins": 1.069937825202942,
"rewards/rejected": -2.7030029296875,
"step": 408
},
{
"epoch": 0.26712384684464036,
"grad_norm": 37.20492293139169,
"learning_rate": 1.3766158585597024e-07,
"logits/chosen": -1.747741460800171,
"logits/rejected": -1.681113600730896,
"logps/chosen": -684.744384765625,
"logps/rejected": -716.8374633789062,
"loss": 0.5827,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.5082005262374878,
"rewards/margins": 0.3761386573314667,
"rewards/rejected": -1.8843392133712769,
"step": 409
},
{
"epoch": 0.26777696138460283,
"grad_norm": 14.02260372554619,
"learning_rate": 1.3756739598472692e-07,
"logits/chosen": -1.6259886026382446,
"logits/rejected": -1.6599675416946411,
"logps/chosen": -740.4105834960938,
"logps/rejected": -777.7703247070312,
"loss": 0.5891,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6748157739639282,
"rewards/margins": 0.48535382747650146,
"rewards/rejected": -2.1601696014404297,
"step": 410
},
{
"epoch": 0.26843007592456525,
"grad_norm": 11.140097310760751,
"learning_rate": 1.3747288044198186e-07,
"logits/chosen": -1.692323088645935,
"logits/rejected": -1.705490231513977,
"logps/chosen": -676.2076416015625,
"logps/rejected": -718.6063232421875,
"loss": 0.6076,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.5469157695770264,
"rewards/margins": 0.4593140482902527,
"rewards/rejected": -2.006229877471924,
"step": 411
},
{
"epoch": 0.2690831904645277,
"grad_norm": 21.86521082932002,
"learning_rate": 1.373780397197009e-07,
"logits/chosen": -1.5999939441680908,
"logits/rejected": -1.638975739479065,
"logps/chosen": -697.4719848632812,
"logps/rejected": -822.9923095703125,
"loss": 0.5562,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7497361898422241,
"rewards/margins": 0.5286187529563904,
"rewards/rejected": -2.278355121612549,
"step": 412
},
{
"epoch": 0.26973630500449014,
"grad_norm": 85.11109239770148,
"learning_rate": 1.3728287431154236e-07,
"logits/chosen": -1.65316903591156,
"logits/rejected": -1.6552778482437134,
"logps/chosen": -639.5845947265625,
"logps/rejected": -646.5022583007812,
"loss": 0.6713,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.5560939311981201,
"rewards/margins": 0.2231585830450058,
"rewards/rejected": -1.779252529144287,
"step": 413
},
{
"epoch": 0.2703894195444526,
"grad_norm": 27.726328984830257,
"learning_rate": 1.371873847128547e-07,
"logits/chosen": -1.7042299509048462,
"logits/rejected": -1.664905309677124,
"logps/chosen": -594.562744140625,
"logps/rejected": -746.6199951171875,
"loss": 0.5604,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.5040781497955322,
"rewards/margins": 0.6467331647872925,
"rewards/rejected": -2.1508116722106934,
"step": 414
},
{
"epoch": 0.27104253408441503,
"grad_norm": 7.783183095188105,
"learning_rate": 1.3709157142067383e-07,
"logits/chosen": -1.530868649482727,
"logits/rejected": -1.5461376905441284,
"logps/chosen": -656.3353881835938,
"logps/rejected": -642.3773193359375,
"loss": 0.6118,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.472480297088623,
"rewards/margins": 0.1450308859348297,
"rewards/rejected": -1.61751127243042,
"step": 415
},
{
"epoch": 0.2716956486243775,
"grad_norm": 19.144666486903013,
"learning_rate": 1.3699543493372047e-07,
"logits/chosen": -1.69258713722229,
"logits/rejected": -1.7538609504699707,
"logps/chosen": -628.3487548828125,
"logps/rejected": -619.4342651367188,
"loss": 0.5694,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.569846510887146,
"rewards/margins": 0.38913729786872864,
"rewards/rejected": -1.9589838981628418,
"step": 416
},
{
"epoch": 0.2723487631643399,
"grad_norm": 27.797390348868447,
"learning_rate": 1.3689897575239766e-07,
"logits/chosen": -1.5702205896377563,
"logits/rejected": -1.602936863899231,
"logps/chosen": -667.676513671875,
"logps/rejected": -693.8980712890625,
"loss": 0.5853,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.529921531677246,
"rewards/margins": 0.45492294430732727,
"rewards/rejected": -1.9848445653915405,
"step": 417
},
{
"epoch": 0.2730018777043024,
"grad_norm": 8.991724030392819,
"learning_rate": 1.3680219437878805e-07,
"logits/chosen": -1.645481824874878,
"logits/rejected": -1.6687901020050049,
"logps/chosen": -626.5555419921875,
"logps/rejected": -623.034912109375,
"loss": 0.6408,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.4581669569015503,
"rewards/margins": 0.075958251953125,
"rewards/rejected": -1.5341250896453857,
"step": 418
},
{
"epoch": 0.2736549922442648,
"grad_norm": 34.54942489548843,
"learning_rate": 1.3670509131665145e-07,
"logits/chosen": -1.6880152225494385,
"logits/rejected": -1.7365965843200684,
"logps/chosen": -684.6690673828125,
"logps/rejected": -660.1009521484375,
"loss": 0.6038,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.548675298690796,
"rewards/margins": 0.23877474665641785,
"rewards/rejected": -1.7874499559402466,
"step": 419
},
{
"epoch": 0.2743081067842273,
"grad_norm": 7.725802049377969,
"learning_rate": 1.36607667071422e-07,
"logits/chosen": -1.60299551486969,
"logits/rejected": -1.5734202861785889,
"logps/chosen": -566.0638427734375,
"logps/rejected": -681.6033325195312,
"loss": 0.5367,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.342764973640442,
"rewards/margins": 0.6366487741470337,
"rewards/rejected": -1.9794137477874756,
"step": 420
},
{
"epoch": 0.2749612213241897,
"grad_norm": 29.01805274332031,
"learning_rate": 1.3650992215020568e-07,
"logits/chosen": -1.6936869621276855,
"logits/rejected": -1.7092525959014893,
"logps/chosen": -627.9664306640625,
"logps/rejected": -691.7342529296875,
"loss": 0.5607,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.4970836639404297,
"rewards/margins": 0.4171164631843567,
"rewards/rejected": -1.9142000675201416,
"step": 421
},
{
"epoch": 0.2756143358641522,
"grad_norm": 16.315855308400074,
"learning_rate": 1.3641185706177758e-07,
"logits/chosen": -1.6888682842254639,
"logits/rejected": -1.749967336654663,
"logps/chosen": -697.9429321289062,
"logps/rejected": -695.5234985351562,
"loss": 0.5916,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.7964341640472412,
"rewards/margins": 0.21294526755809784,
"rewards/rejected": -2.0093793869018555,
"step": 422
},
{
"epoch": 0.2762674504041146,
"grad_norm": 14.532459170681967,
"learning_rate": 1.3631347231657941e-07,
"logits/chosen": -1.5714447498321533,
"logits/rejected": -1.5943671464920044,
"logps/chosen": -706.9442138671875,
"logps/rejected": -723.3374633789062,
"loss": 0.5472,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.754028558731079,
"rewards/margins": 0.30723437666893005,
"rewards/rejected": -2.061262845993042,
"step": 423
},
{
"epoch": 0.2769205649440771,
"grad_norm": 17.1924352317544,
"learning_rate": 1.3621476842671663e-07,
"logits/chosen": -1.5889177322387695,
"logits/rejected": -1.607007384300232,
"logps/chosen": -727.8897705078125,
"logps/rejected": -668.5709838867188,
"loss": 0.5935,
"rewards/accuracies": 0.53125,
"rewards/chosen": -1.8610236644744873,
"rewards/margins": 0.0675252377986908,
"rewards/rejected": -1.92854905128479,
"step": 424
},
{
"epoch": 0.2775736794840395,
"grad_norm": 40.939411419121505,
"learning_rate": 1.3611574590595592e-07,
"logits/chosen": -1.6908094882965088,
"logits/rejected": -1.6986981630325317,
"logps/chosen": -739.2138671875,
"logps/rejected": -840.9638061523438,
"loss": 0.5882,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.968212604522705,
"rewards/margins": 0.6422140002250671,
"rewards/rejected": -2.610426425933838,
"step": 425
},
{
"epoch": 0.27822679402400197,
"grad_norm": 7.3972481534822405,
"learning_rate": 1.3601640526972256e-07,
"logits/chosen": -1.6408872604370117,
"logits/rejected": -1.6618748903274536,
"logps/chosen": -625.466796875,
"logps/rejected": -732.3475341796875,
"loss": 0.56,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.500174641609192,
"rewards/margins": 0.6626629829406738,
"rewards/rejected": -2.162837505340576,
"step": 426
},
{
"epoch": 0.2788799085639644,
"grad_norm": 17.28581442914079,
"learning_rate": 1.3591674703509755e-07,
"logits/chosen": -1.6025550365447998,
"logits/rejected": -1.6443158388137817,
"logps/chosen": -708.13232421875,
"logps/rejected": -776.6422729492188,
"loss": 0.5629,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.7544456720352173,
"rewards/margins": 0.7180143594741821,
"rewards/rejected": -2.4724600315093994,
"step": 427
},
{
"epoch": 0.27953302310392686,
"grad_norm": 21.933643031973503,
"learning_rate": 1.3581677172081503e-07,
"logits/chosen": -1.705406904220581,
"logits/rejected": -1.7163548469543457,
"logps/chosen": -776.873291015625,
"logps/rejected": -821.5286865234375,
"loss": 0.5636,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.0922696590423584,
"rewards/margins": 0.3925325274467468,
"rewards/rejected": -2.484802484512329,
"step": 428
},
{
"epoch": 0.2801861376438893,
"grad_norm": 9.981810314745449,
"learning_rate": 1.3571647984725965e-07,
"logits/chosen": -1.5826154947280884,
"logits/rejected": -1.5522079467773438,
"logps/chosen": -658.821044921875,
"logps/rejected": -708.5677490234375,
"loss": 0.5996,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.5901877880096436,
"rewards/margins": 0.4858096241950989,
"rewards/rejected": -2.0759973526000977,
"step": 429
},
{
"epoch": 0.28083925218385175,
"grad_norm": 11.36893380136887,
"learning_rate": 1.3561587193646377e-07,
"logits/chosen": -1.5808284282684326,
"logits/rejected": -1.61090087890625,
"logps/chosen": -645.9747314453125,
"logps/rejected": -630.6502685546875,
"loss": 0.6277,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.67696213722229,
"rewards/margins": 0.11575008928775787,
"rewards/rejected": -1.7927122116088867,
"step": 430
},
{
"epoch": 0.28149236672381417,
"grad_norm": 65.85082801878511,
"learning_rate": 1.355149485121048e-07,
"logits/chosen": -1.618956446647644,
"logits/rejected": -1.6172845363616943,
"logps/chosen": -570.473876953125,
"logps/rejected": -637.30126953125,
"loss": 0.5954,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.5234228372573853,
"rewards/margins": 0.49459823966026306,
"rewards/rejected": -2.0180211067199707,
"step": 431
},
{
"epoch": 0.28214548126377664,
"grad_norm": 16.266142416859648,
"learning_rate": 1.3541371009950234e-07,
"logits/chosen": -1.552247166633606,
"logits/rejected": -1.580383539199829,
"logps/chosen": -720.5084228515625,
"logps/rejected": -765.5215454101562,
"loss": 0.5772,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.807952880859375,
"rewards/margins": 0.48564061522483826,
"rewards/rejected": -2.293593406677246,
"step": 432
},
{
"epoch": 0.28279859580373906,
"grad_norm": 42.15718871671079,
"learning_rate": 1.3531215722561562e-07,
"logits/chosen": -1.6510796546936035,
"logits/rejected": -1.6283471584320068,
"logps/chosen": -709.1431884765625,
"logps/rejected": -720.7800903320312,
"loss": 0.6278,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.7743120193481445,
"rewards/margins": 0.24123550951480865,
"rewards/rejected": -2.015547513961792,
"step": 433
},
{
"epoch": 0.28345171034370154,
"grad_norm": 19.658520116756474,
"learning_rate": 1.3521029041904067e-07,
"logits/chosen": -1.7421890497207642,
"logits/rejected": -1.7447417974472046,
"logps/chosen": -633.9954833984375,
"logps/rejected": -745.4783325195312,
"loss": 0.5583,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.7052303552627563,
"rewards/margins": 0.5718401074409485,
"rewards/rejected": -2.2770705223083496,
"step": 434
},
{
"epoch": 0.28410482488366395,
"grad_norm": 61.59792122699335,
"learning_rate": 1.351081102100076e-07,
"logits/chosen": -1.6574376821517944,
"logits/rejected": -1.6252573728561401,
"logps/chosen": -680.404541015625,
"logps/rejected": -763.9954833984375,
"loss": 0.5739,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.7275390625,
"rewards/margins": 0.38947176933288574,
"rewards/rejected": -2.117011070251465,
"step": 435
},
{
"epoch": 0.2847579394236264,
"grad_norm": 7.980201871732804,
"learning_rate": 1.3500561713037777e-07,
"logits/chosen": -1.6319736242294312,
"logits/rejected": -1.6381350755691528,
"logps/chosen": -719.6119995117188,
"logps/rejected": -745.8162841796875,
"loss": 0.5692,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.7299431562423706,
"rewards/margins": 0.5405523180961609,
"rewards/rejected": -2.270495653152466,
"step": 436
},
{
"epoch": 0.28541105396358885,
"grad_norm": 8.727533392873836,
"learning_rate": 1.3490281171364112e-07,
"logits/chosen": -1.6387797594070435,
"logits/rejected": -1.6695307493209839,
"logps/chosen": -695.0264892578125,
"logps/rejected": -734.383544921875,
"loss": 0.5652,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.7570171356201172,
"rewards/margins": 0.508963942527771,
"rewards/rejected": -2.2659811973571777,
"step": 437
},
{
"epoch": 0.2860641685035513,
"grad_norm": 28.528027325398355,
"learning_rate": 1.3479969449491332e-07,
"logits/chosen": -1.600303053855896,
"logits/rejected": -1.5852181911468506,
"logps/chosen": -685.3209838867188,
"logps/rejected": -876.8265380859375,
"loss": 0.5698,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7708958387374878,
"rewards/margins": 0.7778002619743347,
"rewards/rejected": -2.5486960411071777,
"step": 438
},
{
"epoch": 0.28671728304351374,
"grad_norm": 14.87235543409777,
"learning_rate": 1.3469626601093301e-07,
"logits/chosen": -1.5701959133148193,
"logits/rejected": -1.6426581144332886,
"logps/chosen": -664.4986572265625,
"logps/rejected": -663.914794921875,
"loss": 0.6061,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.8818445205688477,
"rewards/margins": 0.31387796998023987,
"rewards/rejected": -2.1957225799560547,
"step": 439
},
{
"epoch": 0.2873703975834762,
"grad_norm": 10.063735379517492,
"learning_rate": 1.34592526800059e-07,
"logits/chosen": -1.6531732082366943,
"logits/rejected": -1.7189072370529175,
"logps/chosen": -714.6017456054688,
"logps/rejected": -692.885498046875,
"loss": 0.5671,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.7326029539108276,
"rewards/margins": 0.28911852836608887,
"rewards/rejected": -2.021721363067627,
"step": 440
},
{
"epoch": 0.28802351212343863,
"grad_norm": 9.184629537387579,
"learning_rate": 1.3448847740226753e-07,
"logits/chosen": -1.4753351211547852,
"logits/rejected": -1.5424079895019531,
"logps/chosen": -621.2201538085938,
"logps/rejected": -637.651611328125,
"loss": 0.6113,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.6852307319641113,
"rewards/margins": 0.3220943510532379,
"rewards/rejected": -2.0073251724243164,
"step": 441
},
{
"epoch": 0.2886766266634011,
"grad_norm": 8.636425870814191,
"learning_rate": 1.3438411835914935e-07,
"logits/chosen": -1.689327597618103,
"logits/rejected": -1.6427044868469238,
"logps/chosen": -652.803466796875,
"logps/rejected": -845.2495727539062,
"loss": 0.5666,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.6894198656082153,
"rewards/margins": 0.8975628614425659,
"rewards/rejected": -2.5869827270507812,
"step": 442
},
{
"epoch": 0.2893297412033635,
"grad_norm": 10.50323658639023,
"learning_rate": 1.3427945021390695e-07,
"logits/chosen": -1.7043615579605103,
"logits/rejected": -1.6875221729278564,
"logps/chosen": -757.1970825195312,
"logps/rejected": -846.8417358398438,
"loss": 0.5605,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.8582996129989624,
"rewards/margins": 0.7164192795753479,
"rewards/rejected": -2.574718952178955,
"step": 443
},
{
"epoch": 0.289982855743326,
"grad_norm": 58.021997546888514,
"learning_rate": 1.3417447351135174e-07,
"logits/chosen": -1.6309140920639038,
"logits/rejected": -1.5757015943527222,
"logps/chosen": -656.53466796875,
"logps/rejected": -700.3538818359375,
"loss": 0.556,
"rewards/accuracies": 0.59375,
"rewards/chosen": -1.6346073150634766,
"rewards/margins": 0.3053905665874481,
"rewards/rejected": -1.939997911453247,
"step": 444
},
{
"epoch": 0.2906359702832884,
"grad_norm": 23.765137933971218,
"learning_rate": 1.3406918879790125e-07,
"logits/chosen": -1.604915976524353,
"logits/rejected": -1.5979593992233276,
"logps/chosen": -678.1114501953125,
"logps/rejected": -687.6436767578125,
"loss": 0.5872,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8041229248046875,
"rewards/margins": 0.2883044183254242,
"rewards/rejected": -2.0924274921417236,
"step": 445
},
{
"epoch": 0.2912890848232509,
"grad_norm": 9.50417097205747,
"learning_rate": 1.3396359662157621e-07,
"logits/chosen": -1.625780463218689,
"logits/rejected": -1.6541712284088135,
"logps/chosen": -595.81396484375,
"logps/rejected": -702.814453125,
"loss": 0.544,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.58511221408844,
"rewards/margins": 0.7033818364143372,
"rewards/rejected": -2.288494110107422,
"step": 446
},
{
"epoch": 0.2919421993632133,
"grad_norm": 18.981324007656344,
"learning_rate": 1.3385769753199778e-07,
"logits/chosen": -1.6669667959213257,
"logits/rejected": -1.6522274017333984,
"logps/chosen": -686.9908447265625,
"logps/rejected": -650.0782470703125,
"loss": 0.5556,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.9364535808563232,
"rewards/margins": 0.2179185450077057,
"rewards/rejected": -2.154372215270996,
"step": 447
},
{
"epoch": 0.2925953139031758,
"grad_norm": 10.169059802410512,
"learning_rate": 1.3375149208038454e-07,
"logits/chosen": -1.6317005157470703,
"logits/rejected": -1.6902613639831543,
"logps/chosen": -641.2767944335938,
"logps/rejected": -646.5491943359375,
"loss": 0.5492,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7388436794281006,
"rewards/margins": 0.43374961614608765,
"rewards/rejected": -2.172593355178833,
"step": 448
},
{
"epoch": 0.2932484284431382,
"grad_norm": 49.39528712371372,
"learning_rate": 1.3364498081954984e-07,
"logits/chosen": -1.6534464359283447,
"logits/rejected": -1.6478941440582275,
"logps/chosen": -652.5145263671875,
"logps/rejected": -719.8404541015625,
"loss": 0.5349,
"rewards/accuracies": 0.90625,
"rewards/chosen": -1.6955946683883667,
"rewards/margins": 0.6352224349975586,
"rewards/rejected": -2.3308169841766357,
"step": 449
},
{
"epoch": 0.29390154298310067,
"grad_norm": 22.612708126765373,
"learning_rate": 1.3353816430389877e-07,
"logits/chosen": -1.6020948886871338,
"logits/rejected": -1.5905299186706543,
"logps/chosen": -640.2467041015625,
"logps/rejected": -654.7265625,
"loss": 0.5702,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.7803701162338257,
"rewards/margins": 0.2993907928466797,
"rewards/rejected": -2.079760789871216,
"step": 450
},
{
"epoch": 0.2945546575230631,
"grad_norm": 53.155242423002186,
"learning_rate": 1.3343104308942526e-07,
"logits/chosen": -1.5806388854980469,
"logits/rejected": -1.6192755699157715,
"logps/chosen": -638.1632080078125,
"logps/rejected": -661.924072265625,
"loss": 0.5966,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.5359299182891846,
"rewards/margins": 0.4746638536453247,
"rewards/rejected": -2.0105934143066406,
"step": 451
},
{
"epoch": 0.29520777206302556,
"grad_norm": 27.43879161964594,
"learning_rate": 1.3332361773370933e-07,
"logits/chosen": -1.615262746810913,
"logits/rejected": -1.618033766746521,
"logps/chosen": -684.5283203125,
"logps/rejected": -777.6100463867188,
"loss": 0.5537,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.678364634513855,
"rewards/margins": 0.7497380971908569,
"rewards/rejected": -2.428102731704712,
"step": 452
},
{
"epoch": 0.295860886602988,
"grad_norm": 90.50268890044883,
"learning_rate": 1.3321588879591404e-07,
"logits/chosen": -1.6059372425079346,
"logits/rejected": -1.646728754043579,
"logps/chosen": -712.547119140625,
"logps/rejected": -678.2857666015625,
"loss": 0.6262,
"rewards/accuracies": 0.46875,
"rewards/chosen": -2.0923614501953125,
"rewards/margins": 0.0826776921749115,
"rewards/rejected": -2.175039291381836,
"step": 453
},
{
"epoch": 0.29651400114295046,
"grad_norm": 23.544890284443007,
"learning_rate": 1.331078568367826e-07,
"logits/chosen": -1.6115236282348633,
"logits/rejected": -1.6034373044967651,
"logps/chosen": -676.42529296875,
"logps/rejected": -732.385009765625,
"loss": 0.5661,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.7986388206481934,
"rewards/margins": 0.3738027811050415,
"rewards/rejected": -2.1724417209625244,
"step": 454
},
{
"epoch": 0.2971671156829129,
"grad_norm": 20.85817286901651,
"learning_rate": 1.3299952241863558e-07,
"logits/chosen": -1.547767996788025,
"logits/rejected": -1.5575647354125977,
"logps/chosen": -677.2283325195312,
"logps/rejected": -753.81201171875,
"loss": 0.5791,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.8230202198028564,
"rewards/margins": 0.49635809659957886,
"rewards/rejected": -2.319378137588501,
"step": 455
},
{
"epoch": 0.29782023022287535,
"grad_norm": 83.24183783338488,
"learning_rate": 1.3289088610536775e-07,
"logits/chosen": -1.704686164855957,
"logits/rejected": -1.7018077373504639,
"logps/chosen": -748.0383911132812,
"logps/rejected": -787.422607421875,
"loss": 0.5771,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.033522605895996,
"rewards/margins": 0.40751010179519653,
"rewards/rejected": -2.441032886505127,
"step": 456
},
{
"epoch": 0.29847334476283777,
"grad_norm": 9.918684990634835,
"learning_rate": 1.3278194846244547e-07,
"logits/chosen": -1.6032931804656982,
"logits/rejected": -1.631669282913208,
"logps/chosen": -650.1741943359375,
"logps/rejected": -619.4449462890625,
"loss": 0.5868,
"rewards/accuracies": 0.5,
"rewards/chosen": -2.0186896324157715,
"rewards/margins": 0.09990565478801727,
"rewards/rejected": -2.1185953617095947,
"step": 457
},
{
"epoch": 0.29912645930280024,
"grad_norm": 8.930483236743907,
"learning_rate": 1.326727100569034e-07,
"logits/chosen": -1.5541692972183228,
"logits/rejected": -1.5824774503707886,
"logps/chosen": -660.3080444335938,
"logps/rejected": -716.2559814453125,
"loss": 0.5453,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9189565181732178,
"rewards/margins": 0.3501700758934021,
"rewards/rejected": -2.2691266536712646,
"step": 458
},
{
"epoch": 0.29977957384276266,
"grad_norm": 27.815139403173614,
"learning_rate": 1.3256317145734176e-07,
"logits/chosen": -1.6188397407531738,
"logits/rejected": -1.6419901847839355,
"logps/chosen": -769.3443603515625,
"logps/rejected": -774.8818969726562,
"loss": 0.5426,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.362337350845337,
"rewards/margins": 0.2768968641757965,
"rewards/rejected": -2.6392343044281006,
"step": 459
},
{
"epoch": 0.30043268838272513,
"grad_norm": 32.322236847979745,
"learning_rate": 1.3245333323392333e-07,
"logits/chosen": -1.575210452079773,
"logits/rejected": -1.5853873491287231,
"logps/chosen": -663.4458618164062,
"logps/rejected": -697.9381103515625,
"loss": 0.5825,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.7677443027496338,
"rewards/margins": 0.5210894346237183,
"rewards/rejected": -2.2888338565826416,
"step": 460
},
{
"epoch": 0.30108580292268755,
"grad_norm": 39.148266264726594,
"learning_rate": 1.3234319595837053e-07,
"logits/chosen": -1.6241929531097412,
"logits/rejected": -1.633462905883789,
"logps/chosen": -698.6934814453125,
"logps/rejected": -783.1587524414062,
"loss": 0.5855,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.971913456916809,
"rewards/margins": 0.39414018392562866,
"rewards/rejected": -2.366053581237793,
"step": 461
},
{
"epoch": 0.30173891746265,
"grad_norm": 11.953653868367251,
"learning_rate": 1.3223276020396224e-07,
"logits/chosen": -1.6563340425491333,
"logits/rejected": -1.6382731199264526,
"logps/chosen": -750.8363647460938,
"logps/rejected": -918.314453125,
"loss": 0.5138,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.174654722213745,
"rewards/margins": 0.9784154891967773,
"rewards/rejected": -3.1530702114105225,
"step": 462
},
{
"epoch": 0.30239203200261244,
"grad_norm": 11.156052343735718,
"learning_rate": 1.3212202654553108e-07,
"logits/chosen": -1.6498849391937256,
"logits/rejected": -1.611333966255188,
"logps/chosen": -697.7623291015625,
"logps/rejected": -808.7626953125,
"loss": 0.5996,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.0645065307617188,
"rewards/margins": 0.5361623167991638,
"rewards/rejected": -2.6006689071655273,
"step": 463
},
{
"epoch": 0.3030451465425749,
"grad_norm": 9.56202472673271,
"learning_rate": 1.3201099555946027e-07,
"logits/chosen": -1.5961620807647705,
"logits/rejected": -1.6157381534576416,
"logps/chosen": -695.4110717773438,
"logps/rejected": -687.124755859375,
"loss": 0.5877,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.75066339969635,
"rewards/margins": 0.38458389043807983,
"rewards/rejected": -2.135247230529785,
"step": 464
},
{
"epoch": 0.30369826108253734,
"grad_norm": 16.37722781419385,
"learning_rate": 1.3189966782368067e-07,
"logits/chosen": -1.478019118309021,
"logits/rejected": -1.5088088512420654,
"logps/chosen": -654.6697998046875,
"logps/rejected": -701.0975952148438,
"loss": 0.5274,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9592812061309814,
"rewards/margins": 0.3919811546802521,
"rewards/rejected": -2.351262331008911,
"step": 465
},
{
"epoch": 0.3043513756224998,
"grad_norm": 48.09748648700373,
"learning_rate": 1.3178804391766773e-07,
"logits/chosen": -1.6397638320922852,
"logits/rejected": -1.6730685234069824,
"logps/chosen": -689.5550537109375,
"logps/rejected": -693.6464233398438,
"loss": 0.5573,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.6381131410598755,
"rewards/margins": 0.5276268720626831,
"rewards/rejected": -2.1657400131225586,
"step": 466
},
{
"epoch": 0.3050044901624622,
"grad_norm": 38.21088887373832,
"learning_rate": 1.3167612442243849e-07,
"logits/chosen": -1.6109600067138672,
"logits/rejected": -1.588073968887329,
"logps/chosen": -679.6129760742188,
"logps/rejected": -752.5172119140625,
"loss": 0.5441,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.768452525138855,
"rewards/margins": 0.5461674332618713,
"rewards/rejected": -2.314620018005371,
"step": 467
},
{
"epoch": 0.3056576047024247,
"grad_norm": 33.132628684957666,
"learning_rate": 1.3156390992054862e-07,
"logits/chosen": -1.6505910158157349,
"logits/rejected": -1.6523258686065674,
"logps/chosen": -707.0516357421875,
"logps/rejected": -742.3071899414062,
"loss": 0.5618,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.175845146179199,
"rewards/margins": 0.5027600526809692,
"rewards/rejected": -2.678605318069458,
"step": 468
},
{
"epoch": 0.3063107192423871,
"grad_norm": 13.439285792051546,
"learning_rate": 1.3145140099608933e-07,
"logits/chosen": -1.6851032972335815,
"logits/rejected": -1.6651477813720703,
"logps/chosen": -708.4338989257812,
"logps/rejected": -813.6752319335938,
"loss": 0.5774,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.052492380142212,
"rewards/margins": 0.5105604529380798,
"rewards/rejected": -2.5630528926849365,
"step": 469
},
{
"epoch": 0.3069638337823496,
"grad_norm": 21.620672217002696,
"learning_rate": 1.3133859823468433e-07,
"logits/chosen": -1.5591509342193604,
"logits/rejected": -1.620843768119812,
"logps/chosen": -712.449951171875,
"logps/rejected": -747.5558471679688,
"loss": 0.561,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.18332576751709,
"rewards/margins": 0.5399474501609802,
"rewards/rejected": -2.7232730388641357,
"step": 470
},
{
"epoch": 0.307616948322312,
"grad_norm": 7.80149433228074,
"learning_rate": 1.3122550222348676e-07,
"logits/chosen": -1.6145565509796143,
"logits/rejected": -1.612226128578186,
"logps/chosen": -602.789794921875,
"logps/rejected": -650.0497436523438,
"loss": 0.5333,
"rewards/accuracies": 0.625,
"rewards/chosen": -1.8238171339035034,
"rewards/margins": 0.18020805716514587,
"rewards/rejected": -2.0040249824523926,
"step": 471
},
{
"epoch": 0.3082700628622745,
"grad_norm": 12.541386961248598,
"learning_rate": 1.3111211355117625e-07,
"logits/chosen": -1.7079181671142578,
"logits/rejected": -1.6550383567810059,
"logps/chosen": -733.6851806640625,
"logps/rejected": -804.6192626953125,
"loss": 0.5624,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.264955520629883,
"rewards/margins": 0.5682797431945801,
"rewards/rejected": -2.833235263824463,
"step": 472
},
{
"epoch": 0.3089231774022369,
"grad_norm": 49.77231044522242,
"learning_rate": 1.3099843280795564e-07,
"logits/chosen": -1.5972816944122314,
"logits/rejected": -1.5726184844970703,
"logps/chosen": -697.912109375,
"logps/rejected": -767.2205810546875,
"loss": 0.5352,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.8209738731384277,
"rewards/margins": 0.34553292393684387,
"rewards/rejected": -2.16650652885437,
"step": 473
},
{
"epoch": 0.3095762919421994,
"grad_norm": 9.964229839524728,
"learning_rate": 1.3088446058554813e-07,
"logits/chosen": -1.5458612442016602,
"logits/rejected": -1.533179759979248,
"logps/chosen": -707.8756103515625,
"logps/rejected": -787.054443359375,
"loss": 0.5863,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.043527364730835,
"rewards/margins": 0.6540895700454712,
"rewards/rejected": -2.6976165771484375,
"step": 474
},
{
"epoch": 0.3102294064821618,
"grad_norm": 45.959503738015734,
"learning_rate": 1.3077019747719412e-07,
"logits/chosen": -1.595198154449463,
"logits/rejected": -1.591080904006958,
"logps/chosen": -645.2706298828125,
"logps/rejected": -711.2510986328125,
"loss": 0.5782,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.0279712677001953,
"rewards/margins": 0.44245079159736633,
"rewards/rejected": -2.4704222679138184,
"step": 475
},
{
"epoch": 0.31088252102212427,
"grad_norm": 44.73324911306505,
"learning_rate": 1.3065564407764802e-07,
"logits/chosen": -1.7171953916549683,
"logits/rejected": -1.707771897315979,
"logps/chosen": -780.519775390625,
"logps/rejected": -787.8773193359375,
"loss": 0.6095,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.1023976802825928,
"rewards/margins": 0.306552529335022,
"rewards/rejected": -2.4089503288269043,
"step": 476
},
{
"epoch": 0.3115356355620867,
"grad_norm": 48.62571818947248,
"learning_rate": 1.3054080098317535e-07,
"logits/chosen": -1.6242291927337646,
"logits/rejected": -1.604690432548523,
"logps/chosen": -631.8426513671875,
"logps/rejected": -682.6934204101562,
"loss": 0.5734,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8018213510513306,
"rewards/margins": 0.6165929436683655,
"rewards/rejected": -2.418414354324341,
"step": 477
},
{
"epoch": 0.31218875010204916,
"grad_norm": 12.529349998561317,
"learning_rate": 1.3042566879154942e-07,
"logits/chosen": -1.6056309938430786,
"logits/rejected": -1.5826270580291748,
"logps/chosen": -729.5185546875,
"logps/rejected": -871.41064453125,
"loss": 0.5912,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.2389609813690186,
"rewards/margins": 0.8720102310180664,
"rewards/rejected": -3.110971450805664,
"step": 478
},
{
"epoch": 0.3128418646420116,
"grad_norm": 25.761326367725754,
"learning_rate": 1.3031024810204844e-07,
"logits/chosen": -1.652268648147583,
"logits/rejected": -1.6630209684371948,
"logps/chosen": -703.5462036132812,
"logps/rejected": -724.381591796875,
"loss": 0.5784,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.003835439682007,
"rewards/margins": 0.37049388885498047,
"rewards/rejected": -2.3743293285369873,
"step": 479
},
{
"epoch": 0.31349497918197405,
"grad_norm": 33.561886992169335,
"learning_rate": 1.3019453951545222e-07,
"logits/chosen": -1.5987491607666016,
"logits/rejected": -1.5921955108642578,
"logps/chosen": -745.3104248046875,
"logps/rejected": -764.4356689453125,
"loss": 0.5442,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.0976104736328125,
"rewards/margins": 0.27280157804489136,
"rewards/rejected": -2.3704121112823486,
"step": 480
},
{
"epoch": 0.31414809372193647,
"grad_norm": 36.917374177404064,
"learning_rate": 1.3007854363403912e-07,
"logits/chosen": -1.4915146827697754,
"logits/rejected": -1.4882618188858032,
"logps/chosen": -751.360107421875,
"logps/rejected": -766.1044311523438,
"loss": 0.5453,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.2802181243896484,
"rewards/margins": 0.4815993905067444,
"rewards/rejected": -2.761817693710327,
"step": 481
},
{
"epoch": 0.31480120826189895,
"grad_norm": 13.711790140200893,
"learning_rate": 1.2996226106158292e-07,
"logits/chosen": -1.5975637435913086,
"logits/rejected": -1.6035072803497314,
"logps/chosen": -755.90576171875,
"logps/rejected": -734.3273315429688,
"loss": 0.5692,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.2815115451812744,
"rewards/margins": 0.35683876276016235,
"rewards/rejected": -2.638350486755371,
"step": 482
},
{
"epoch": 0.31545432280186136,
"grad_norm": 10.225083803183653,
"learning_rate": 1.2984569240334968e-07,
"logits/chosen": -1.6066551208496094,
"logits/rejected": -1.6023541688919067,
"logps/chosen": -683.2215576171875,
"logps/rejected": -708.5989379882812,
"loss": 0.643,
"rewards/accuracies": 0.5,
"rewards/chosen": -2.3892295360565186,
"rewards/margins": 0.2522014081478119,
"rewards/rejected": -2.6414308547973633,
"step": 483
},
{
"epoch": 0.31610743734182384,
"grad_norm": 43.21606295702285,
"learning_rate": 1.297288382660945e-07,
"logits/chosen": -1.5631223917007446,
"logits/rejected": -1.5718114376068115,
"logps/chosen": -689.9915771484375,
"logps/rejected": -718.431396484375,
"loss": 0.6106,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.1257877349853516,
"rewards/margins": 0.2984415590763092,
"rewards/rejected": -2.424229145050049,
"step": 484
},
{
"epoch": 0.31676055188178626,
"grad_norm": 27.88872175196333,
"learning_rate": 1.2961169925805854e-07,
"logits/chosen": -1.5951167345046997,
"logits/rejected": -1.6255998611450195,
"logps/chosen": -788.0588989257812,
"logps/rejected": -845.193603515625,
"loss": 0.5597,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4089677333831787,
"rewards/margins": 0.6144492626190186,
"rewards/rejected": -3.0234172344207764,
"step": 485
},
{
"epoch": 0.31741366642174873,
"grad_norm": 20.90987797781216,
"learning_rate": 1.294942759889657e-07,
"logits/chosen": -1.675968050956726,
"logits/rejected": -1.7186676263809204,
"logps/chosen": -748.28466796875,
"logps/rejected": -773.9339599609375,
"loss": 0.5613,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.358844757080078,
"rewards/margins": 0.4122333824634552,
"rewards/rejected": -2.771078109741211,
"step": 486
},
{
"epoch": 0.31806678096171115,
"grad_norm": 55.17649113640136,
"learning_rate": 1.2937656907001946e-07,
"logits/chosen": -1.641244649887085,
"logits/rejected": -1.6599280834197998,
"logps/chosen": -761.1217041015625,
"logps/rejected": -830.0103149414062,
"loss": 0.5631,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4944777488708496,
"rewards/margins": 0.6374845504760742,
"rewards/rejected": -3.131962299346924,
"step": 487
},
{
"epoch": 0.3187198955016736,
"grad_norm": 102.83841154795559,
"learning_rate": 1.2925857911389977e-07,
"logits/chosen": -1.5266590118408203,
"logits/rejected": -1.5372052192687988,
"logps/chosen": -713.9895629882812,
"logps/rejected": -758.6173706054688,
"loss": 0.5926,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.067730188369751,
"rewards/margins": 0.3173166811466217,
"rewards/rejected": -2.3850467205047607,
"step": 488
},
{
"epoch": 0.31937301004163604,
"grad_norm": 16.29290504478057,
"learning_rate": 1.2914030673475987e-07,
"logits/chosen": -1.6052075624465942,
"logits/rejected": -1.6125444173812866,
"logps/chosen": -788.452392578125,
"logps/rejected": -915.9051513671875,
"loss": 0.5288,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4514353275299072,
"rewards/margins": 0.7415469884872437,
"rewards/rejected": -3.1929819583892822,
"step": 489
},
{
"epoch": 0.3200261245815985,
"grad_norm": 10.250010376773696,
"learning_rate": 1.29021752548223e-07,
"logits/chosen": -1.5654120445251465,
"logits/rejected": -1.5245580673217773,
"logps/chosen": -656.7188720703125,
"logps/rejected": -755.3834838867188,
"loss": 0.5413,
"rewards/accuracies": 0.75,
"rewards/chosen": -1.9858925342559814,
"rewards/margins": 0.6740937829017639,
"rewards/rejected": -2.6599864959716797,
"step": 490
},
{
"epoch": 0.32067923912156093,
"grad_norm": 26.08722651165643,
"learning_rate": 1.2890291717137919e-07,
"logits/chosen": -1.6398707628250122,
"logits/rejected": -1.6489192247390747,
"logps/chosen": -740.1271362304688,
"logps/rejected": -744.15576171875,
"loss": 0.5919,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.0526180267333984,
"rewards/margins": 0.4571646749973297,
"rewards/rejected": -2.509782552719116,
"step": 491
},
{
"epoch": 0.3213323536615234,
"grad_norm": 11.326643733648055,
"learning_rate": 1.287838012227822e-07,
"logits/chosen": -1.4910850524902344,
"logits/rejected": -1.5168672800064087,
"logps/chosen": -666.2679443359375,
"logps/rejected": -719.11328125,
"loss": 0.5579,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.113301992416382,
"rewards/margins": 0.3083459138870239,
"rewards/rejected": -2.4216480255126953,
"step": 492
},
{
"epoch": 0.3219854682014858,
"grad_norm": 10.301719458410284,
"learning_rate": 1.2866440532244618e-07,
"logits/chosen": -1.6052420139312744,
"logits/rejected": -1.5893734693527222,
"logps/chosen": -680.7243041992188,
"logps/rejected": -705.3173828125,
"loss": 0.5927,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.155557632446289,
"rewards/margins": 0.13128241896629333,
"rewards/rejected": -2.2868399620056152,
"step": 493
},
{
"epoch": 0.3226385827414483,
"grad_norm": 23.898710768355055,
"learning_rate": 1.2854473009184242e-07,
"logits/chosen": -1.582448959350586,
"logits/rejected": -1.6095366477966309,
"logps/chosen": -674.9298095703125,
"logps/rejected": -756.9193725585938,
"loss": 0.5823,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.0278987884521484,
"rewards/margins": 0.5040706992149353,
"rewards/rejected": -2.5319693088531494,
"step": 494
},
{
"epoch": 0.3232916972814107,
"grad_norm": 38.15944101714436,
"learning_rate": 1.2842477615389622e-07,
"logits/chosen": -1.582186222076416,
"logits/rejected": -1.589294195175171,
"logps/chosen": -693.418701171875,
"logps/rejected": -733.3966064453125,
"loss": 0.5397,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.0407207012176514,
"rewards/margins": 0.5694603323936462,
"rewards/rejected": -2.6101813316345215,
"step": 495
},
{
"epoch": 0.3239448118213732,
"grad_norm": 14.734546679223891,
"learning_rate": 1.2830454413298353e-07,
"logits/chosen": -1.6934702396392822,
"logits/rejected": -1.692880630493164,
"logps/chosen": -768.6975708007812,
"logps/rejected": -769.83984375,
"loss": 0.5351,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.1913037300109863,
"rewards/margins": 0.30963531136512756,
"rewards/rejected": -2.500938892364502,
"step": 496
},
{
"epoch": 0.3245979263613356,
"grad_norm": 57.674097527721926,
"learning_rate": 1.2818403465492783e-07,
"logits/chosen": -1.5970431566238403,
"logits/rejected": -1.6256811618804932,
"logps/chosen": -704.041748046875,
"logps/rejected": -755.9532470703125,
"loss": 0.5673,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.2154181003570557,
"rewards/margins": 0.48861750960350037,
"rewards/rejected": -2.704035520553589,
"step": 497
},
{
"epoch": 0.3252510409012981,
"grad_norm": 51.51237394300094,
"learning_rate": 1.280632483469967e-07,
"logits/chosen": -1.601000189781189,
"logits/rejected": -1.615774393081665,
"logps/chosen": -765.2921142578125,
"logps/rejected": -783.3229370117188,
"loss": 0.5797,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.198168992996216,
"rewards/margins": 0.4726821184158325,
"rewards/rejected": -2.670850992202759,
"step": 498
},
{
"epoch": 0.3259041554412605,
"grad_norm": 52.55743751092252,
"learning_rate": 1.2794218583789876e-07,
"logits/chosen": -1.6243011951446533,
"logits/rejected": -1.6285080909729004,
"logps/chosen": -646.337890625,
"logps/rejected": -663.8938598632812,
"loss": 0.5854,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.0927515029907227,
"rewards/margins": 0.4221267104148865,
"rewards/rejected": -2.514878273010254,
"step": 499
},
{
"epoch": 0.326557269981223,
"grad_norm": 40.80464561018148,
"learning_rate": 1.278208477577802e-07,
"logits/chosen": -1.6202638149261475,
"logits/rejected": -1.6449174880981445,
"logps/chosen": -826.0245361328125,
"logps/rejected": -890.9912109375,
"loss": 0.5447,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.2344555854797363,
"rewards/margins": 0.7165102958679199,
"rewards/rejected": -2.9509658813476562,
"step": 500
},
{
"epoch": 0.326557269981223,
"eval_logits/chosen": -1.6206858158111572,
"eval_logits/rejected": -1.6248681545257568,
"eval_logps/chosen": -728.922119140625,
"eval_logps/rejected": -772.7636108398438,
"eval_loss": 0.5644757151603699,
"eval_rewards/accuracies": 0.7110000252723694,
"eval_rewards/chosen": -2.1857683658599854,
"eval_rewards/margins": 0.5194733142852783,
"eval_rewards/rejected": -2.7052416801452637,
"eval_runtime": 296.3263,
"eval_samples_per_second": 13.499,
"eval_steps_per_second": 0.844,
"step": 500
},
{
"epoch": 0.3272103845211854,
"grad_norm": 18.6916196942294,
"learning_rate": 1.276992347382217e-07,
"logits/chosen": -1.6296683549880981,
"logits/rejected": -1.6314842700958252,
"logps/chosen": -768.372314453125,
"logps/rejected": -827.50634765625,
"loss": 0.5824,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.622889280319214,
"rewards/margins": 0.44725632667541504,
"rewards/rejected": -3.070145606994629,
"step": 501
},
{
"epoch": 0.32786349906114787,
"grad_norm": 38.99333241282542,
"learning_rate": 1.2757734741223494e-07,
"logits/chosen": -1.598349690437317,
"logits/rejected": -1.59029221534729,
"logps/chosen": -643.4448852539062,
"logps/rejected": -724.112548828125,
"loss": 0.584,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.8339576721191406,
"rewards/margins": 0.5225554704666138,
"rewards/rejected": -2.356513023376465,
"step": 502
},
{
"epoch": 0.3285166136011103,
"grad_norm": 57.964045081569594,
"learning_rate": 1.2745518641425945e-07,
"logits/chosen": -1.5579354763031006,
"logits/rejected": -1.5548033714294434,
"logps/chosen": -681.3307495117188,
"logps/rejected": -820.7464599609375,
"loss": 0.559,
"rewards/accuracies": 0.84375,
"rewards/chosen": -1.9576513767242432,
"rewards/margins": 0.9468222856521606,
"rewards/rejected": -2.9044735431671143,
"step": 503
},
{
"epoch": 0.32916972814107276,
"grad_norm": 26.127719689786453,
"learning_rate": 1.2733275238015923e-07,
"logits/chosen": -1.6055808067321777,
"logits/rejected": -1.630063772201538,
"logps/chosen": -629.9222412109375,
"logps/rejected": -704.1810302734375,
"loss": 0.5861,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.0422046184539795,
"rewards/margins": 0.5639029741287231,
"rewards/rejected": -2.606107234954834,
"step": 504
},
{
"epoch": 0.3298228426810352,
"grad_norm": 26.745925680747586,
"learning_rate": 1.272100459472195e-07,
"logits/chosen": -1.4966247081756592,
"logits/rejected": -1.51353120803833,
"logps/chosen": -733.5635986328125,
"logps/rejected": -757.8365478515625,
"loss": 0.6147,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.2747292518615723,
"rewards/margins": 0.3070646822452545,
"rewards/rejected": -2.581793785095215,
"step": 505
},
{
"epoch": 0.33047595722099765,
"grad_norm": 15.726026277688065,
"learning_rate": 1.2708706775414333e-07,
"logits/chosen": -1.562856912612915,
"logits/rejected": -1.5949100255966187,
"logps/chosen": -742.255859375,
"logps/rejected": -773.4927978515625,
"loss": 0.5832,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.044926643371582,
"rewards/margins": 0.5169533491134644,
"rewards/rejected": -2.561879873275757,
"step": 506
},
{
"epoch": 0.33112907176096007,
"grad_norm": 18.237745759610128,
"learning_rate": 1.269638184410483e-07,
"logits/chosen": -1.6007072925567627,
"logits/rejected": -1.627900242805481,
"logps/chosen": -760.51708984375,
"logps/rejected": -827.0796508789062,
"loss": 0.5619,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.159515619277954,
"rewards/margins": 0.579525887966156,
"rewards/rejected": -2.739041566848755,
"step": 507
},
{
"epoch": 0.33178218630092254,
"grad_norm": 10.09970310506052,
"learning_rate": 1.2684029864946334e-07,
"logits/chosen": -1.4610607624053955,
"logits/rejected": -1.4745755195617676,
"logps/chosen": -622.4526977539062,
"logps/rejected": -639.34814453125,
"loss": 0.5861,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.053276538848877,
"rewards/margins": 0.30753716826438904,
"rewards/rejected": -2.360813617706299,
"step": 508
},
{
"epoch": 0.33243530084088496,
"grad_norm": 29.561202029180375,
"learning_rate": 1.2671650902232512e-07,
"logits/chosen": -1.5951420068740845,
"logits/rejected": -1.6463465690612793,
"logps/chosen": -701.150634765625,
"logps/rejected": -748.044921875,
"loss": 0.5395,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.1120245456695557,
"rewards/margins": 0.4306832551956177,
"rewards/rejected": -2.542707681655884,
"step": 509
},
{
"epoch": 0.33308841538084744,
"grad_norm": 20.140045301365248,
"learning_rate": 1.2659245020397487e-07,
"logits/chosen": -1.6009819507598877,
"logits/rejected": -1.6235804557800293,
"logps/chosen": -767.361083984375,
"logps/rejected": -796.4589233398438,
"loss": 0.6337,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.34623646736145,
"rewards/margins": 0.62890625,
"rewards/rejected": -2.975142478942871,
"step": 510
},
{
"epoch": 0.33374152992080985,
"grad_norm": 20.816657867687876,
"learning_rate": 1.2646812284015502e-07,
"logits/chosen": -1.5827473402023315,
"logits/rejected": -1.5781760215759277,
"logps/chosen": -730.8731079101562,
"logps/rejected": -745.44775390625,
"loss": 0.5493,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.1604835987091064,
"rewards/margins": 0.3516397774219513,
"rewards/rejected": -2.5121235847473145,
"step": 511
},
{
"epoch": 0.33439464446077233,
"grad_norm": 27.37026270313324,
"learning_rate": 1.263435275780058e-07,
"logits/chosen": -1.5628294944763184,
"logits/rejected": -1.5843901634216309,
"logps/chosen": -666.028076171875,
"logps/rejected": -709.0813598632812,
"loss": 0.646,
"rewards/accuracies": 0.5625,
"rewards/chosen": -1.901466965675354,
"rewards/margins": 0.44154587388038635,
"rewards/rejected": -2.343013048171997,
"step": 512
},
{
"epoch": 0.33504775900073475,
"grad_norm": 35.16457922567049,
"learning_rate": 1.262186650660619e-07,
"logits/chosen": -1.5470932722091675,
"logits/rejected": -1.5630905628204346,
"logps/chosen": -718.0603637695312,
"logps/rejected": -732.3851928710938,
"loss": 0.5527,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.389463424682617,
"rewards/margins": 0.3482608497142792,
"rewards/rejected": -2.7377243041992188,
"step": 513
},
{
"epoch": 0.3357008735406972,
"grad_norm": 70.3813885396455,
"learning_rate": 1.2609353595424905e-07,
"logits/chosen": -1.5562270879745483,
"logits/rejected": -1.5387725830078125,
"logps/chosen": -601.1290893554688,
"logps/rejected": -723.7789916992188,
"loss": 0.5382,
"rewards/accuracies": 0.78125,
"rewards/chosen": -1.8300285339355469,
"rewards/margins": 0.7093777060508728,
"rewards/rejected": -2.5394062995910645,
"step": 514
},
{
"epoch": 0.33635398808065964,
"grad_norm": 40.988995422153955,
"learning_rate": 1.2596814089388074e-07,
"logits/chosen": -1.646407127380371,
"logits/rejected": -1.622480869293213,
"logps/chosen": -724.177001953125,
"logps/rejected": -750.5733032226562,
"loss": 0.5101,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.0493416786193848,
"rewards/margins": 0.4716246724128723,
"rewards/rejected": -2.520966053009033,
"step": 515
},
{
"epoch": 0.3370071026206221,
"grad_norm": 22.872898048232333,
"learning_rate": 1.2584248053765463e-07,
"logits/chosen": -1.668154239654541,
"logits/rejected": -1.6877217292785645,
"logps/chosen": -788.1798095703125,
"logps/rejected": -882.4114990234375,
"loss": 0.5467,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.5617947578430176,
"rewards/margins": 0.7210710048675537,
"rewards/rejected": -3.2828657627105713,
"step": 516
},
{
"epoch": 0.33766021716058453,
"grad_norm": 62.010832881435796,
"learning_rate": 1.257165555396494e-07,
"logits/chosen": -1.6311254501342773,
"logits/rejected": -1.603393793106079,
"logps/chosen": -683.9799194335938,
"logps/rejected": -818.5030517578125,
"loss": 0.5689,
"rewards/accuracies": 0.8125,
"rewards/chosen": -1.8681172132492065,
"rewards/margins": 0.7638627290725708,
"rewards/rejected": -2.6319799423217773,
"step": 517
},
{
"epoch": 0.338313331700547,
"grad_norm": 47.93407261910419,
"learning_rate": 1.2559036655532116e-07,
"logits/chosen": -1.639786958694458,
"logits/rejected": -1.6281380653381348,
"logps/chosen": -655.11083984375,
"logps/rejected": -711.305419921875,
"loss": 0.6042,
"rewards/accuracies": 0.65625,
"rewards/chosen": -1.9371042251586914,
"rewards/margins": 0.29598766565322876,
"rewards/rejected": -2.2330920696258545,
"step": 518
},
{
"epoch": 0.3389664462405094,
"grad_norm": 8.481962150515601,
"learning_rate": 1.2546391424150015e-07,
"logits/chosen": -1.5316431522369385,
"logits/rejected": -1.5820224285125732,
"logps/chosen": -723.28173828125,
"logps/rejected": -765.1573486328125,
"loss": 0.5594,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.011814594268799,
"rewards/margins": 0.4189469814300537,
"rewards/rejected": -2.4307615756988525,
"step": 519
},
{
"epoch": 0.3396195607804719,
"grad_norm": 16.110076519841456,
"learning_rate": 1.2533719925638722e-07,
"logits/chosen": -1.561875820159912,
"logits/rejected": -1.563469409942627,
"logps/chosen": -703.0348510742188,
"logps/rejected": -769.7527465820312,
"loss": 0.5457,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.9231442213058472,
"rewards/margins": 0.762036144733429,
"rewards/rejected": -2.685180187225342,
"step": 520
},
{
"epoch": 0.3402726753204343,
"grad_norm": 12.232146508628508,
"learning_rate": 1.2521022225955051e-07,
"logits/chosen": -1.6638747453689575,
"logits/rejected": -1.675110101699829,
"logps/chosen": -702.4622802734375,
"logps/rejected": -841.2349853515625,
"loss": 0.5551,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.236776351928711,
"rewards/margins": 0.7252943515777588,
"rewards/rejected": -2.962070941925049,
"step": 521
},
{
"epoch": 0.3409257898603968,
"grad_norm": 20.099246846498655,
"learning_rate": 1.2508298391192192e-07,
"logits/chosen": -1.6344259977340698,
"logits/rejected": -1.6060516834259033,
"logps/chosen": -690.3515625,
"logps/rejected": -725.8463134765625,
"loss": 0.5615,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.1555328369140625,
"rewards/margins": 0.24002605676651,
"rewards/rejected": -2.3955588340759277,
"step": 522
},
{
"epoch": 0.3415789044003592,
"grad_norm": 28.678293250926668,
"learning_rate": 1.2495548487579377e-07,
"logits/chosen": -1.6345038414001465,
"logits/rejected": -1.6450937986373901,
"logps/chosen": -744.6649780273438,
"logps/rejected": -786.7367553710938,
"loss": 0.561,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.3111023902893066,
"rewards/margins": 0.4379115700721741,
"rewards/rejected": -2.749014377593994,
"step": 523
},
{
"epoch": 0.3422320189403217,
"grad_norm": 55.26079414913828,
"learning_rate": 1.248277258148152e-07,
"logits/chosen": -1.632555365562439,
"logits/rejected": -1.6509549617767334,
"logps/chosen": -782.6264038085938,
"logps/rejected": -826.6570434570312,
"loss": 0.5775,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.1081671714782715,
"rewards/margins": 0.5822569131851196,
"rewards/rejected": -2.6904244422912598,
"step": 524
},
{
"epoch": 0.3428851334802841,
"grad_norm": 25.386727503372587,
"learning_rate": 1.2469970739398895e-07,
"logits/chosen": -1.5985530614852905,
"logits/rejected": -1.6135672330856323,
"logps/chosen": -738.0941772460938,
"logps/rejected": -705.9267578125,
"loss": 0.6116,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.2364561557769775,
"rewards/margins": 0.3205564022064209,
"rewards/rejected": -2.5570127964019775,
"step": 525
},
{
"epoch": 0.3435382480202466,
"grad_norm": 24.51284417618824,
"learning_rate": 1.2457143027966763e-07,
"logits/chosen": -1.634655475616455,
"logits/rejected": -1.6211440563201904,
"logps/chosen": -750.4459228515625,
"logps/rejected": -730.4688720703125,
"loss": 0.5521,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.091447353363037,
"rewards/margins": 0.35747280716896057,
"rewards/rejected": -2.448920249938965,
"step": 526
},
{
"epoch": 0.344191362560209,
"grad_norm": 44.66836450950333,
"learning_rate": 1.2444289513955052e-07,
"logits/chosen": -1.5693323612213135,
"logits/rejected": -1.6260815858840942,
"logps/chosen": -779.1264038085938,
"logps/rejected": -830.4940185546875,
"loss": 0.5846,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.496044635772705,
"rewards/margins": 0.4798283278942108,
"rewards/rejected": -2.975872755050659,
"step": 527
},
{
"epoch": 0.34484447710017146,
"grad_norm": 58.69879433030782,
"learning_rate": 1.2431410264267977e-07,
"logits/chosen": -1.604842185974121,
"logits/rejected": -1.604213833808899,
"logps/chosen": -802.912109375,
"logps/rejected": -868.535888671875,
"loss": 0.5993,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.3447303771972656,
"rewards/margins": 0.5476406216621399,
"rewards/rejected": -2.8923707008361816,
"step": 528
},
{
"epoch": 0.3454975916401339,
"grad_norm": 36.25638810204293,
"learning_rate": 1.2418505345943732e-07,
"logits/chosen": -1.5913817882537842,
"logits/rejected": -1.631279468536377,
"logps/chosen": -760.42138671875,
"logps/rejected": -691.488037109375,
"loss": 0.582,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.126579999923706,
"rewards/margins": 0.14462479948997498,
"rewards/rejected": -2.2712044715881348,
"step": 529
},
{
"epoch": 0.34615070618009636,
"grad_norm": 22.96730047945058,
"learning_rate": 1.24055748261541e-07,
"logits/chosen": -1.6423602104187012,
"logits/rejected": -1.6510244607925415,
"logps/chosen": -752.2402954101562,
"logps/rejected": -776.55810546875,
"loss": 0.5723,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.1990549564361572,
"rewards/margins": 0.4823690950870514,
"rewards/rejected": -2.681424140930176,
"step": 530
},
{
"epoch": 0.3468038207200588,
"grad_norm": 12.704637412947715,
"learning_rate": 1.2392618772204144e-07,
"logits/chosen": -1.5454316139221191,
"logits/rejected": -1.5524265766143799,
"logps/chosen": -704.6710815429688,
"logps/rejected": -745.8167114257812,
"loss": 0.5883,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.09757924079895,
"rewards/margins": 0.410552978515625,
"rewards/rejected": -2.508132219314575,
"step": 531
},
{
"epoch": 0.34745693526002125,
"grad_norm": 39.764198564369345,
"learning_rate": 1.2379637251531815e-07,
"logits/chosen": -1.6019591093063354,
"logits/rejected": -1.6283773183822632,
"logps/chosen": -743.0405883789062,
"logps/rejected": -784.7008666992188,
"loss": 0.5559,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.1056554317474365,
"rewards/margins": 0.5337550044059753,
"rewards/rejected": -2.6394104957580566,
"step": 532
},
{
"epoch": 0.34811004979998367,
"grad_norm": 34.128906613498735,
"learning_rate": 1.2366630331707633e-07,
"logits/chosen": -1.6114816665649414,
"logits/rejected": -1.6253386735916138,
"logps/chosen": -733.2994995117188,
"logps/rejected": -764.6762084960938,
"loss": 0.5313,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.0889058113098145,
"rewards/margins": 0.47615137696266174,
"rewards/rejected": -2.5650570392608643,
"step": 533
},
{
"epoch": 0.34876316433994614,
"grad_norm": 10.132934728994826,
"learning_rate": 1.2353598080434324e-07,
"logits/chosen": -1.632277250289917,
"logits/rejected": -1.6106319427490234,
"logps/chosen": -824.1144409179688,
"logps/rejected": -907.9232177734375,
"loss": 0.556,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.394355297088623,
"rewards/margins": 0.591341495513916,
"rewards/rejected": -2.985696792602539,
"step": 534
},
{
"epoch": 0.34941627887990856,
"grad_norm": 27.01945814074106,
"learning_rate": 1.234054056554646e-07,
"logits/chosen": -1.5513118505477905,
"logits/rejected": -1.5469458103179932,
"logps/chosen": -851.3035278320312,
"logps/rejected": -767.6931762695312,
"loss": 0.598,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.3710482120513916,
"rewards/margins": 0.34137019515037537,
"rewards/rejected": -2.7124183177948,
"step": 535
},
{
"epoch": 0.35006939341987103,
"grad_norm": 20.633383781514684,
"learning_rate": 1.2327457855010123e-07,
"logits/chosen": -1.5809993743896484,
"logits/rejected": -1.5902115106582642,
"logps/chosen": -829.6868896484375,
"logps/rejected": -818.66455078125,
"loss": 0.5935,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.530834913253784,
"rewards/margins": 0.49049264192581177,
"rewards/rejected": -3.021327495574951,
"step": 536
},
{
"epoch": 0.35072250795983345,
"grad_norm": 8.274651730353373,
"learning_rate": 1.2314350016922534e-07,
"logits/chosen": -1.634692907333374,
"logits/rejected": -1.593540072441101,
"logps/chosen": -737.755126953125,
"logps/rejected": -739.923583984375,
"loss": 0.5132,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.248922348022461,
"rewards/margins": 0.5101175308227539,
"rewards/rejected": -2.759039878845215,
"step": 537
},
{
"epoch": 0.3513756224997959,
"grad_norm": 10.192458514719585,
"learning_rate": 1.2301217119511708e-07,
"logits/chosen": -1.5304259061813354,
"logits/rejected": -1.5202865600585938,
"logps/chosen": -765.9959106445312,
"logps/rejected": -765.4609375,
"loss": 0.5613,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.386655330657959,
"rewards/margins": 0.377580851316452,
"rewards/rejected": -2.7642359733581543,
"step": 538
},
{
"epoch": 0.35202873703975834,
"grad_norm": 34.040881919948525,
"learning_rate": 1.2288059231136108e-07,
"logits/chosen": -1.532372236251831,
"logits/rejected": -1.5646611452102661,
"logps/chosen": -741.4403686523438,
"logps/rejected": -813.6516723632812,
"loss": 0.5147,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.1854164600372314,
"rewards/margins": 0.7029819488525391,
"rewards/rejected": -2.8883986473083496,
"step": 539
},
{
"epoch": 0.3526818515797208,
"grad_norm": 12.432751143916052,
"learning_rate": 1.2274876420284258e-07,
"logits/chosen": -1.6108044385910034,
"logits/rejected": -1.6093095541000366,
"logps/chosen": -760.4302978515625,
"logps/rejected": -788.9061889648438,
"loss": 0.5564,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.301989793777466,
"rewards/margins": 0.3381144404411316,
"rewards/rejected": -2.640104293823242,
"step": 540
},
{
"epoch": 0.35333496611968324,
"grad_norm": 30.749409766082465,
"learning_rate": 1.2261668755574421e-07,
"logits/chosen": -1.5781821012496948,
"logits/rejected": -1.6185979843139648,
"logps/chosen": -819.251220703125,
"logps/rejected": -834.0662841796875,
"loss": 0.5351,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3528733253479004,
"rewards/margins": 0.5704101324081421,
"rewards/rejected": -2.923283338546753,
"step": 541
},
{
"epoch": 0.3539880806596457,
"grad_norm": 29.16774400285802,
"learning_rate": 1.2248436305754222e-07,
"logits/chosen": -1.648099422454834,
"logits/rejected": -1.6576378345489502,
"logps/chosen": -712.8369140625,
"logps/rejected": -755.3017578125,
"loss": 0.5369,
"rewards/accuracies": 0.71875,
"rewards/chosen": -1.99370276927948,
"rewards/margins": 0.4913850426673889,
"rewards/rejected": -2.4850876331329346,
"step": 542
},
{
"epoch": 0.35464119519960813,
"grad_norm": 20.69027911099315,
"learning_rate": 1.2235179139700304e-07,
"logits/chosen": -1.6414737701416016,
"logits/rejected": -1.624894142150879,
"logps/chosen": -813.76611328125,
"logps/rejected": -831.1404418945312,
"loss": 0.5981,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.3536837100982666,
"rewards/margins": 0.42595234513282776,
"rewards/rejected": -2.7796361446380615,
"step": 543
},
{
"epoch": 0.3552943097395706,
"grad_norm": 29.949634840397668,
"learning_rate": 1.222189732641795e-07,
"logits/chosen": -1.5087977647781372,
"logits/rejected": -1.4944639205932617,
"logps/chosen": -627.0595703125,
"logps/rejected": -755.531005859375,
"loss": 0.5464,
"rewards/accuracies": 0.90625,
"rewards/chosen": -1.9218859672546387,
"rewards/margins": 1.039351463317871,
"rewards/rejected": -2.9612374305725098,
"step": 544
},
{
"epoch": 0.355947424279533,
"grad_norm": 14.943355173162043,
"learning_rate": 1.220859093504074e-07,
"logits/chosen": -1.5890822410583496,
"logits/rejected": -1.5994880199432373,
"logps/chosen": -857.3690185546875,
"logps/rejected": -871.0678100585938,
"loss": 0.5952,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.6291110515594482,
"rewards/margins": 0.46089449524879456,
"rewards/rejected": -3.09000563621521,
"step": 545
},
{
"epoch": 0.3566005388194955,
"grad_norm": 15.565475336026891,
"learning_rate": 1.2195260034830187e-07,
"logits/chosen": -1.6251795291900635,
"logits/rejected": -1.6245146989822388,
"logps/chosen": -726.9799194335938,
"logps/rejected": -732.9411010742188,
"loss": 0.5526,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.323512077331543,
"rewards/margins": 0.2341621220111847,
"rewards/rejected": -2.5576741695404053,
"step": 546
},
{
"epoch": 0.3572536533594579,
"grad_norm": 23.991791905870468,
"learning_rate": 1.2181904695175374e-07,
"logits/chosen": -1.4712308645248413,
"logits/rejected": -1.4625868797302246,
"logps/chosen": -717.213134765625,
"logps/rejected": -820.203857421875,
"loss": 0.6012,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.27494740486145,
"rewards/margins": 0.7880808115005493,
"rewards/rejected": -3.063028335571289,
"step": 547
},
{
"epoch": 0.3579067678994204,
"grad_norm": 35.7010085328891,
"learning_rate": 1.2168524985592597e-07,
"logits/chosen": -1.5415699481964111,
"logits/rejected": -1.5597350597381592,
"logps/chosen": -772.77978515625,
"logps/rejected": -746.6201782226562,
"loss": 0.5508,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.421182155609131,
"rewards/margins": 0.254733145236969,
"rewards/rejected": -2.675915241241455,
"step": 548
},
{
"epoch": 0.3585598824393828,
"grad_norm": 14.830303633870002,
"learning_rate": 1.2155120975724996e-07,
"logits/chosen": -1.5214571952819824,
"logits/rejected": -1.5628215074539185,
"logps/chosen": -706.9215087890625,
"logps/rejected": -791.664794921875,
"loss": 0.5483,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.0701212882995605,
"rewards/margins": 0.7649804949760437,
"rewards/rejected": -2.83510160446167,
"step": 549
},
{
"epoch": 0.3592129969793453,
"grad_norm": 11.60924722942488,
"learning_rate": 1.214169273534221e-07,
"logits/chosen": -1.5903964042663574,
"logits/rejected": -1.6238888502120972,
"logps/chosen": -754.437255859375,
"logps/rejected": -765.71435546875,
"loss": 0.5506,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.2985119819641113,
"rewards/margins": 0.33853238821029663,
"rewards/rejected": -2.6370441913604736,
"step": 550
},
{
"epoch": 0.3598661115193077,
"grad_norm": 19.208616054146393,
"learning_rate": 1.2128240334339978e-07,
"logits/chosen": -1.6252080202102661,
"logits/rejected": -1.6387659311294556,
"logps/chosen": -827.2951049804688,
"logps/rejected": -865.0391845703125,
"loss": 0.5597,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.646684408187866,
"rewards/margins": 0.5018330812454224,
"rewards/rejected": -3.14851713180542,
"step": 551
},
{
"epoch": 0.36051922605927017,
"grad_norm": 15.97827516501005,
"learning_rate": 1.211476384273982e-07,
"logits/chosen": -1.6076312065124512,
"logits/rejected": -1.6076362133026123,
"logps/chosen": -735.5797119140625,
"logps/rejected": -755.5748291015625,
"loss": 0.5435,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.272061347961426,
"rewards/margins": 0.8418019413948059,
"rewards/rejected": -3.113863229751587,
"step": 552
},
{
"epoch": 0.3611723405992326,
"grad_norm": 12.277648543582611,
"learning_rate": 1.2101263330688638e-07,
"logits/chosen": -1.6942815780639648,
"logits/rejected": -1.6733278036117554,
"logps/chosen": -777.0508422851562,
"logps/rejected": -763.1177368164062,
"loss": 0.5924,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.428075075149536,
"rewards/margins": 0.363750159740448,
"rewards/rejected": -2.79182505607605,
"step": 553
},
{
"epoch": 0.36182545513919506,
"grad_norm": 10.465467914806046,
"learning_rate": 1.208773886845837e-07,
"logits/chosen": -1.604886531829834,
"logits/rejected": -1.6047513484954834,
"logps/chosen": -739.2161865234375,
"logps/rejected": -860.6815185546875,
"loss": 0.511,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.230139970779419,
"rewards/margins": 0.676809549331665,
"rewards/rejected": -2.906949520111084,
"step": 554
},
{
"epoch": 0.3624785696791575,
"grad_norm": 21.364165430156397,
"learning_rate": 1.2074190526445616e-07,
"logits/chosen": -1.6185420751571655,
"logits/rejected": -1.6467700004577637,
"logps/chosen": -812.8031005859375,
"logps/rejected": -845.0140991210938,
"loss": 0.4957,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.4488282203674316,
"rewards/margins": 0.606843113899231,
"rewards/rejected": -3.055671453475952,
"step": 555
},
{
"epoch": 0.36313168421911995,
"grad_norm": 12.060857675840488,
"learning_rate": 1.2060618375171275e-07,
"logits/chosen": -1.582899808883667,
"logits/rejected": -1.5877315998077393,
"logps/chosen": -716.4563598632812,
"logps/rejected": -797.8826293945312,
"loss": 0.5325,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.498225688934326,
"rewards/margins": 0.3760990798473358,
"rewards/rejected": -2.8743247985839844,
"step": 556
},
{
"epoch": 0.3637847987590824,
"grad_norm": 13.492857611613953,
"learning_rate": 1.2047022485280168e-07,
"logits/chosen": -1.539080023765564,
"logits/rejected": -1.570099115371704,
"logps/chosen": -700.08203125,
"logps/rejected": -716.7232666015625,
"loss": 0.5849,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.2837417125701904,
"rewards/margins": 0.5059683322906494,
"rewards/rejected": -2.78971004486084,
"step": 557
},
{
"epoch": 0.36443791329904485,
"grad_norm": 33.571052467052354,
"learning_rate": 1.2033402927540688e-07,
"logits/chosen": -1.5120295286178589,
"logits/rejected": -1.4916774034500122,
"logps/chosen": -786.6419677734375,
"logps/rejected": -877.5504760742188,
"loss": 0.5212,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.3839404582977295,
"rewards/margins": 0.8281471133232117,
"rewards/rejected": -3.212087631225586,
"step": 558
},
{
"epoch": 0.36509102783900726,
"grad_norm": 49.87336711471519,
"learning_rate": 1.2019759772844423e-07,
"logits/chosen": -1.570830225944519,
"logits/rejected": -1.612029790878296,
"logps/chosen": -759.057373046875,
"logps/rejected": -869.3032836914062,
"loss": 0.608,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.1914572715759277,
"rewards/margins": 0.8840473890304565,
"rewards/rejected": -3.075504779815674,
"step": 559
},
{
"epoch": 0.36574414237896974,
"grad_norm": 74.5712473751751,
"learning_rate": 1.2006093092205777e-07,
"logits/chosen": -1.4767194986343384,
"logits/rejected": -1.4949413537979126,
"logps/chosen": -698.3121948242188,
"logps/rejected": -743.3089599609375,
"loss": 0.5547,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.0461912155151367,
"rewards/margins": 0.5971631407737732,
"rewards/rejected": -2.643354654312134,
"step": 560
},
{
"epoch": 0.36639725691893216,
"grad_norm": 44.74846351811676,
"learning_rate": 1.199240295676162e-07,
"logits/chosen": -1.5869390964508057,
"logits/rejected": -1.600525140762329,
"logps/chosen": -759.1281127929688,
"logps/rejected": -777.5281372070312,
"loss": 0.6048,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.285490036010742,
"rewards/margins": 0.29206350445747375,
"rewards/rejected": -2.5775535106658936,
"step": 561
},
{
"epoch": 0.36705037145889463,
"grad_norm": 10.999921054518422,
"learning_rate": 1.1978689437770896e-07,
"logits/chosen": -1.5283629894256592,
"logits/rejected": -1.5792925357818604,
"logps/chosen": -674.9652709960938,
"logps/rejected": -813.370849609375,
"loss": 0.5122,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.0200095176696777,
"rewards/margins": 0.8305485844612122,
"rewards/rejected": -2.850558280944824,
"step": 562
},
{
"epoch": 0.36770348599885705,
"grad_norm": 55.10455252064543,
"learning_rate": 1.1964952606614276e-07,
"logits/chosen": -1.6453619003295898,
"logits/rejected": -1.6702425479888916,
"logps/chosen": -794.0926513671875,
"logps/rejected": -870.9822998046875,
"loss": 0.5045,
"rewards/accuracies": 0.9375,
"rewards/chosen": -2.6258649826049805,
"rewards/margins": 0.731311559677124,
"rewards/rejected": -3.3571763038635254,
"step": 563
},
{
"epoch": 0.3683566005388195,
"grad_norm": 24.477947303915716,
"learning_rate": 1.1951192534793764e-07,
"logits/chosen": -1.5885945558547974,
"logits/rejected": -1.5889862775802612,
"logps/chosen": -740.5401611328125,
"logps/rejected": -770.5545654296875,
"loss": 0.5242,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.3373959064483643,
"rewards/margins": 0.3964667022228241,
"rewards/rejected": -2.7338626384735107,
"step": 564
},
{
"epoch": 0.36900971507878194,
"grad_norm": 18.91974104462604,
"learning_rate": 1.193740929393234e-07,
"logits/chosen": -1.5289589166641235,
"logits/rejected": -1.5133321285247803,
"logps/chosen": -695.8096923828125,
"logps/rejected": -892.4715576171875,
"loss": 0.5767,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.182143211364746,
"rewards/margins": 1.0010870695114136,
"rewards/rejected": -3.18323016166687,
"step": 565
},
{
"epoch": 0.36966282961874436,
"grad_norm": 96.53142007828465,
"learning_rate": 1.1923602955773583e-07,
"logits/chosen": -1.585618257522583,
"logits/rejected": -1.580026626586914,
"logps/chosen": -676.3846435546875,
"logps/rejected": -825.5516357421875,
"loss": 0.5286,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.03085994720459,
"rewards/margins": 0.9540601968765259,
"rewards/rejected": -2.984920024871826,
"step": 566
},
{
"epoch": 0.37031594415870683,
"grad_norm": 9.841984604340173,
"learning_rate": 1.1909773592181287e-07,
"logits/chosen": -1.6656473875045776,
"logits/rejected": -1.6531533002853394,
"logps/chosen": -783.5050659179688,
"logps/rejected": -795.6260986328125,
"loss": 0.5741,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.256021738052368,
"rewards/margins": 0.621117353439331,
"rewards/rejected": -2.877139091491699,
"step": 567
},
{
"epoch": 0.37096905869866925,
"grad_norm": 13.260118641189585,
"learning_rate": 1.189592127513911e-07,
"logits/chosen": -1.5660011768341064,
"logits/rejected": -1.5593037605285645,
"logps/chosen": -699.3264770507812,
"logps/rejected": -718.5197143554688,
"loss": 0.5123,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.3721680641174316,
"rewards/margins": 0.40681201219558716,
"rewards/rejected": -2.778980255126953,
"step": 568
},
{
"epoch": 0.3716221732386317,
"grad_norm": 66.7819908806299,
"learning_rate": 1.1882046076750176e-07,
"logits/chosen": -1.6180167198181152,
"logits/rejected": -1.6349365711212158,
"logps/chosen": -837.523681640625,
"logps/rejected": -804.4877319335938,
"loss": 0.5133,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.4943768978118896,
"rewards/margins": 0.5837184190750122,
"rewards/rejected": -3.078094959259033,
"step": 569
},
{
"epoch": 0.37227528777859414,
"grad_norm": 29.37833169883412,
"learning_rate": 1.186814806923671e-07,
"logits/chosen": -1.5915193557739258,
"logits/rejected": -1.5684032440185547,
"logps/chosen": -738.4844360351562,
"logps/rejected": -740.8411254882812,
"loss": 0.5667,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.2804222106933594,
"rewards/margins": 0.37493157386779785,
"rewards/rejected": -2.6553537845611572,
"step": 570
},
{
"epoch": 0.3729284023185566,
"grad_norm": 48.10084583273166,
"learning_rate": 1.1854227324939669e-07,
"logits/chosen": -1.612912654876709,
"logits/rejected": -1.5692781209945679,
"logps/chosen": -897.7774658203125,
"logps/rejected": -913.3131103515625,
"loss": 0.5835,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.936650276184082,
"rewards/margins": 0.400768518447876,
"rewards/rejected": -3.337418794631958,
"step": 571
},
{
"epoch": 0.37358151685851904,
"grad_norm": 11.716756841277345,
"learning_rate": 1.1840283916318347e-07,
"logits/chosen": -1.5072287321090698,
"logits/rejected": -1.4601385593414307,
"logps/chosen": -745.0501708984375,
"logps/rejected": -801.019775390625,
"loss": 0.5261,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.25480055809021,
"rewards/margins": 0.8588010668754578,
"rewards/rejected": -3.1136016845703125,
"step": 572
},
{
"epoch": 0.3742346313984815,
"grad_norm": 18.580715359087044,
"learning_rate": 1.1826317915950021e-07,
"logits/chosen": -1.5958278179168701,
"logits/rejected": -1.6139434576034546,
"logps/chosen": -767.4638061523438,
"logps/rejected": -891.0325927734375,
"loss": 0.5267,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.512181043624878,
"rewards/margins": 0.7721086144447327,
"rewards/rejected": -3.284290075302124,
"step": 573
},
{
"epoch": 0.37488774593844393,
"grad_norm": 24.286745313641156,
"learning_rate": 1.181232939652955e-07,
"logits/chosen": -1.5305031538009644,
"logits/rejected": -1.4823194742202759,
"logps/chosen": -748.1857299804688,
"logps/rejected": -847.1156005859375,
"loss": 0.5462,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.5048766136169434,
"rewards/margins": 0.7592265009880066,
"rewards/rejected": -3.2641031742095947,
"step": 574
},
{
"epoch": 0.3755408604784064,
"grad_norm": 29.652610575297764,
"learning_rate": 1.1798318430869012e-07,
"logits/chosen": -1.5794718265533447,
"logits/rejected": -1.5874780416488647,
"logps/chosen": -794.8291625976562,
"logps/rejected": -825.0755615234375,
"loss": 0.5033,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.382725238800049,
"rewards/margins": 0.7552610039710999,
"rewards/rejected": -3.137986183166504,
"step": 575
},
{
"epoch": 0.3761939750183688,
"grad_norm": 58.912701269068016,
"learning_rate": 1.1784285091897322e-07,
"logits/chosen": -1.6159145832061768,
"logits/rejected": -1.6184728145599365,
"logps/chosen": -892.6289672851562,
"logps/rejected": -935.57568359375,
"loss": 0.5889,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7877440452575684,
"rewards/margins": 0.5180625319480896,
"rewards/rejected": -3.3058066368103027,
"step": 576
},
{
"epoch": 0.3768470895583313,
"grad_norm": 92.84475707292661,
"learning_rate": 1.1770229452659852e-07,
"logits/chosen": -1.5818086862564087,
"logits/rejected": -1.583407998085022,
"logps/chosen": -678.228515625,
"logps/rejected": -849.5630493164062,
"loss": 0.5562,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.2353968620300293,
"rewards/margins": 1.1506848335266113,
"rewards/rejected": -3.386082172393799,
"step": 577
},
{
"epoch": 0.3775002040982937,
"grad_norm": 24.53245073109831,
"learning_rate": 1.1756151586318044e-07,
"logits/chosen": -1.5623910427093506,
"logits/rejected": -1.6067014932632446,
"logps/chosen": -806.4697875976562,
"logps/rejected": -860.4945068359375,
"loss": 0.5023,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.364492654800415,
"rewards/margins": 0.7060398459434509,
"rewards/rejected": -3.07053279876709,
"step": 578
},
{
"epoch": 0.3781533186382562,
"grad_norm": 18.85494518114228,
"learning_rate": 1.174205156614904e-07,
"logits/chosen": -1.4894192218780518,
"logits/rejected": -1.5204534530639648,
"logps/chosen": -829.0343017578125,
"logps/rejected": -848.0327758789062,
"loss": 0.5509,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7499516010284424,
"rewards/margins": 0.4761306643486023,
"rewards/rejected": -3.2260818481445312,
"step": 579
},
{
"epoch": 0.3788064331782186,
"grad_norm": 15.464999889697193,
"learning_rate": 1.1727929465545294e-07,
"logits/chosen": -1.5147112607955933,
"logits/rejected": -1.6048409938812256,
"logps/chosen": -769.5234985351562,
"logps/rejected": -810.0481567382812,
"loss": 0.6104,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.7371156215667725,
"rewards/margins": 0.32579949498176575,
"rewards/rejected": -3.062915325164795,
"step": 580
},
{
"epoch": 0.3794595477181811,
"grad_norm": 49.00611608700767,
"learning_rate": 1.1713785358014193e-07,
"logits/chosen": -1.603877305984497,
"logits/rejected": -1.6114190816879272,
"logps/chosen": -693.6171875,
"logps/rejected": -696.5467529296875,
"loss": 0.5624,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.458164691925049,
"rewards/margins": 0.3706192076206207,
"rewards/rejected": -2.8287839889526367,
"step": 581
},
{
"epoch": 0.3801126622581435,
"grad_norm": 41.687356680487866,
"learning_rate": 1.1699619317177668e-07,
"logits/chosen": -1.492578387260437,
"logits/rejected": -1.4849687814712524,
"logps/chosen": -711.063232421875,
"logps/rejected": -779.2449951171875,
"loss": 0.5001,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.399482250213623,
"rewards/margins": 0.7047311663627625,
"rewards/rejected": -3.104212999343872,
"step": 582
},
{
"epoch": 0.38076577679810597,
"grad_norm": 40.272642825045395,
"learning_rate": 1.1685431416771825e-07,
"logits/chosen": -1.5852625370025635,
"logits/rejected": -1.6129764318466187,
"logps/chosen": -813.8887939453125,
"logps/rejected": -817.9891357421875,
"loss": 0.5583,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.6035213470458984,
"rewards/margins": 0.5876621007919312,
"rewards/rejected": -3.19118332862854,
"step": 583
},
{
"epoch": 0.3814188913380684,
"grad_norm": 24.667173577610694,
"learning_rate": 1.1671221730646543e-07,
"logits/chosen": -1.5789833068847656,
"logits/rejected": -1.5653014183044434,
"logps/chosen": -698.559814453125,
"logps/rejected": -799.244384765625,
"loss": 0.5128,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.156489610671997,
"rewards/margins": 0.8030795454978943,
"rewards/rejected": -2.959568977355957,
"step": 584
},
{
"epoch": 0.38207200587803086,
"grad_norm": 10.056286338983814,
"learning_rate": 1.1656990332765101e-07,
"logits/chosen": -1.579622745513916,
"logits/rejected": -1.5801726579666138,
"logps/chosen": -727.0596923828125,
"logps/rejected": -783.4789428710938,
"loss": 0.5499,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.434051036834717,
"rewards/margins": 0.5680094957351685,
"rewards/rejected": -3.002060651779175,
"step": 585
},
{
"epoch": 0.3827251204179933,
"grad_norm": 10.639118531618282,
"learning_rate": 1.1642737297203792e-07,
"logits/chosen": -1.550283432006836,
"logits/rejected": -1.545919418334961,
"logps/chosen": -678.3670043945312,
"logps/rejected": -830.1644897460938,
"loss": 0.5517,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.09962797164917,
"rewards/margins": 0.9918718934059143,
"rewards/rejected": -3.0914998054504395,
"step": 586
},
{
"epoch": 0.38337823495795575,
"grad_norm": 20.197574332614796,
"learning_rate": 1.1628462698151538e-07,
"logits/chosen": -1.5244333744049072,
"logits/rejected": -1.5570902824401855,
"logps/chosen": -690.13720703125,
"logps/rejected": -777.196533203125,
"loss": 0.5574,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.285308837890625,
"rewards/margins": 0.6684877872467041,
"rewards/rejected": -2.95379638671875,
"step": 587
},
{
"epoch": 0.3840313494979182,
"grad_norm": 66.37641884436981,
"learning_rate": 1.1614166609909498e-07,
"logits/chosen": -1.5359015464782715,
"logits/rejected": -1.546217679977417,
"logps/chosen": -738.7581176757812,
"logps/rejected": -772.0260009765625,
"loss": 0.5286,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.2918949127197266,
"rewards/margins": 0.4515775740146637,
"rewards/rejected": -2.7434725761413574,
"step": 588
},
{
"epoch": 0.38468446403788065,
"grad_norm": 13.46028213535803,
"learning_rate": 1.1599849106890683e-07,
"logits/chosen": -1.527596116065979,
"logits/rejected": -1.5285677909851074,
"logps/chosen": -734.326171875,
"logps/rejected": -749.241455078125,
"loss": 0.568,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.2978053092956543,
"rewards/margins": 0.3799334466457367,
"rewards/rejected": -2.677738666534424,
"step": 589
},
{
"epoch": 0.38533757857784307,
"grad_norm": 22.97104425543304,
"learning_rate": 1.1585510263619577e-07,
"logits/chosen": -1.5912518501281738,
"logits/rejected": -1.5906767845153809,
"logps/chosen": -724.542236328125,
"logps/rejected": -752.7775268554688,
"loss": 0.6249,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.344860076904297,
"rewards/margins": 0.30979910492897034,
"rewards/rejected": -2.6546592712402344,
"step": 590
},
{
"epoch": 0.38599069311780554,
"grad_norm": 10.721634341469683,
"learning_rate": 1.157115015473174e-07,
"logits/chosen": -1.5702768564224243,
"logits/rejected": -1.5887250900268555,
"logps/chosen": -711.1195678710938,
"logps/rejected": -744.3549194335938,
"loss": 0.5678,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.271317958831787,
"rewards/margins": 0.5261479020118713,
"rewards/rejected": -2.7974658012390137,
"step": 591
},
{
"epoch": 0.38664380765776796,
"grad_norm": 20.96230357570753,
"learning_rate": 1.155676885497342e-07,
"logits/chosen": -1.587166666984558,
"logits/rejected": -1.5863206386566162,
"logps/chosen": -703.5427856445312,
"logps/rejected": -760.740234375,
"loss": 0.5131,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3268959522247314,
"rewards/margins": 0.47407492995262146,
"rewards/rejected": -2.800971031188965,
"step": 592
},
{
"epoch": 0.38729692219773043,
"grad_norm": 19.489795146613808,
"learning_rate": 1.154236643920117e-07,
"logits/chosen": -1.661864161491394,
"logits/rejected": -1.6856908798217773,
"logps/chosen": -724.410400390625,
"logps/rejected": -773.7425537109375,
"loss": 0.5241,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.489246368408203,
"rewards/margins": 0.5750184059143066,
"rewards/rejected": -3.064265012741089,
"step": 593
},
{
"epoch": 0.38795003673769285,
"grad_norm": 73.12725772218887,
"learning_rate": 1.1527942982381452e-07,
"logits/chosen": -1.5697174072265625,
"logits/rejected": -1.5859888792037964,
"logps/chosen": -754.874267578125,
"logps/rejected": -791.1784057617188,
"loss": 0.5575,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.533051013946533,
"rewards/margins": 0.5103416442871094,
"rewards/rejected": -3.0433928966522217,
"step": 594
},
{
"epoch": 0.3886031512776553,
"grad_norm": 50.458839536472475,
"learning_rate": 1.1513498559590252e-07,
"logits/chosen": -1.6159231662750244,
"logits/rejected": -1.5974745750427246,
"logps/chosen": -762.911376953125,
"logps/rejected": -743.5185546875,
"loss": 0.5855,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3125293254852295,
"rewards/margins": 0.30350756645202637,
"rewards/rejected": -2.616036891937256,
"step": 595
},
{
"epoch": 0.38925626581761774,
"grad_norm": 10.34566924087224,
"learning_rate": 1.1499033246012685e-07,
"logits/chosen": -1.5862224102020264,
"logits/rejected": -1.5608221292495728,
"logps/chosen": -864.4752807617188,
"logps/rejected": -983.65087890625,
"loss": 0.5411,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.47229266166687,
"rewards/margins": 0.9073572754859924,
"rewards/rejected": -3.379650115966797,
"step": 596
},
{
"epoch": 0.3899093803575802,
"grad_norm": 27.551259833581323,
"learning_rate": 1.1484547116942601e-07,
"logits/chosen": -1.6429082155227661,
"logits/rejected": -1.6100890636444092,
"logps/chosen": -772.03076171875,
"logps/rejected": -794.9556884765625,
"loss": 0.551,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.502126693725586,
"rewards/margins": 0.5208601951599121,
"rewards/rejected": -3.022986888885498,
"step": 597
},
{
"epoch": 0.39056249489754263,
"grad_norm": 21.20639182270992,
"learning_rate": 1.1470040247782205e-07,
"logits/chosen": -1.5553147792816162,
"logits/rejected": -1.5810749530792236,
"logps/chosen": -782.7449951171875,
"logps/rejected": -797.2606201171875,
"loss": 0.5487,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.488363742828369,
"rewards/margins": 0.4950363039970398,
"rewards/rejected": -2.983400344848633,
"step": 598
},
{
"epoch": 0.3912156094375051,
"grad_norm": 61.94478857020836,
"learning_rate": 1.1455512714041656e-07,
"logits/chosen": -1.567291021347046,
"logits/rejected": -1.5616893768310547,
"logps/chosen": -682.39306640625,
"logps/rejected": -779.3341674804688,
"loss": 0.5335,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.248433828353882,
"rewards/margins": 0.45859482884407043,
"rewards/rejected": -2.707028865814209,
"step": 599
},
{
"epoch": 0.3918687239774675,
"grad_norm": 27.096742126205903,
"learning_rate": 1.1440964591338669e-07,
"logits/chosen": -1.5842782258987427,
"logits/rejected": -1.588181734085083,
"logps/chosen": -848.2506713867188,
"logps/rejected": -845.3145141601562,
"loss": 0.5896,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.536600112915039,
"rewards/margins": 0.5013671517372131,
"rewards/rejected": -3.0379669666290283,
"step": 600
},
{
"epoch": 0.3918687239774675,
"eval_logits/chosen": -1.584659457206726,
"eval_logits/rejected": -1.5836009979248047,
"eval_logps/chosen": -748.0584106445312,
"eval_logps/rejected": -799.712158203125,
"eval_loss": 0.5452645421028137,
"eval_rewards/accuracies": 0.7179999947547913,
"eval_rewards/chosen": -2.3771302700042725,
"eval_rewards/margins": 0.5975968241691589,
"eval_rewards/rejected": -2.9747273921966553,
"eval_runtime": 300.4063,
"eval_samples_per_second": 13.315,
"eval_steps_per_second": 0.832,
"step": 600
},
{
"epoch": 0.39252183851743,
"grad_norm": 12.348978382810378,
"learning_rate": 1.142639595539813e-07,
"logits/chosen": -1.5698190927505493,
"logits/rejected": -1.529393196105957,
"logps/chosen": -787.7490234375,
"logps/rejected": -788.6868896484375,
"loss": 0.5584,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.350584030151367,
"rewards/margins": 0.6392349004745483,
"rewards/rejected": -2.989819049835205,
"step": 601
},
{
"epoch": 0.3931749530573924,
"grad_norm": 10.679850846766614,
"learning_rate": 1.1411806882051702e-07,
"logits/chosen": -1.5929275751113892,
"logits/rejected": -1.5805671215057373,
"logps/chosen": -723.1961669921875,
"logps/rejected": -713.7118530273438,
"loss": 0.5297,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.31701397895813,
"rewards/margins": 0.5196025371551514,
"rewards/rejected": -2.836616039276123,
"step": 602
},
{
"epoch": 0.3938280675973549,
"grad_norm": 35.708189094404084,
"learning_rate": 1.1397197447237423e-07,
"logits/chosen": -1.5436877012252808,
"logits/rejected": -1.562942624092102,
"logps/chosen": -728.4673461914062,
"logps/rejected": -798.8472290039062,
"loss": 0.5137,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6131293773651123,
"rewards/margins": 0.4990772604942322,
"rewards/rejected": -3.11220645904541,
"step": 603
},
{
"epoch": 0.3944811821373173,
"grad_norm": 82.94052901290438,
"learning_rate": 1.1382567726999319e-07,
"logits/chosen": -1.5964429378509521,
"logits/rejected": -1.5941182374954224,
"logps/chosen": -796.7896118164062,
"logps/rejected": -837.9287109375,
"loss": 0.5751,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.377279043197632,
"rewards/margins": 0.43824002146720886,
"rewards/rejected": -2.815519094467163,
"step": 604
},
{
"epoch": 0.3951342966772798,
"grad_norm": 48.03054462615425,
"learning_rate": 1.1367917797487002e-07,
"logits/chosen": -1.5520787239074707,
"logits/rejected": -1.5235430002212524,
"logps/chosen": -703.1958618164062,
"logps/rejected": -713.5035400390625,
"loss": 0.6487,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.3644354343414307,
"rewards/margins": 0.27925288677215576,
"rewards/rejected": -2.643688440322876,
"step": 605
},
{
"epoch": 0.3957874112172422,
"grad_norm": 32.26597426964153,
"learning_rate": 1.1353247734955275e-07,
"logits/chosen": -1.527002215385437,
"logits/rejected": -1.5140247344970703,
"logps/chosen": -783.9526977539062,
"logps/rejected": -775.091796875,
"loss": 0.5689,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.74333119392395,
"rewards/margins": 0.22180992364883423,
"rewards/rejected": -2.9651412963867188,
"step": 606
},
{
"epoch": 0.3964405257572047,
"grad_norm": 10.429513069602487,
"learning_rate": 1.133855761576374e-07,
"logits/chosen": -1.5365808010101318,
"logits/rejected": -1.5101385116577148,
"logps/chosen": -651.0933837890625,
"logps/rejected": -708.3961181640625,
"loss": 0.5409,
"rewards/accuracies": 0.6875,
"rewards/chosen": -1.9835056066513062,
"rewards/margins": 0.3382415473461151,
"rewards/rejected": -2.321747303009033,
"step": 607
},
{
"epoch": 0.3970936402971671,
"grad_norm": 19.72537569851452,
"learning_rate": 1.1323847516376392e-07,
"logits/chosen": -1.5134871006011963,
"logits/rejected": -1.545701026916504,
"logps/chosen": -684.7098999023438,
"logps/rejected": -666.0336303710938,
"loss": 0.5726,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.495856523513794,
"rewards/margins": 0.06803350895643234,
"rewards/rejected": -2.563889980316162,
"step": 608
},
{
"epoch": 0.39774675483712957,
"grad_norm": 78.12335940886078,
"learning_rate": 1.1309117513361228e-07,
"logits/chosen": -1.4830129146575928,
"logits/rejected": -1.4823014736175537,
"logps/chosen": -758.7376708984375,
"logps/rejected": -757.42138671875,
"loss": 0.5917,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.4975523948669434,
"rewards/margins": 0.32604148983955383,
"rewards/rejected": -2.823594093322754,
"step": 609
},
{
"epoch": 0.398399869377092,
"grad_norm": 23.094040308095767,
"learning_rate": 1.1294367683389848e-07,
"logits/chosen": -1.5845017433166504,
"logits/rejected": -1.556647539138794,
"logps/chosen": -739.6962280273438,
"logps/rejected": -748.73876953125,
"loss": 0.5258,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.4809250831604004,
"rewards/margins": 0.4438374638557434,
"rewards/rejected": -2.924762725830078,
"step": 610
},
{
"epoch": 0.39905298391705446,
"grad_norm": 39.1177725554315,
"learning_rate": 1.1279598103237047e-07,
"logits/chosen": -1.60092031955719,
"logits/rejected": -1.6122784614562988,
"logps/chosen": -759.567138671875,
"logps/rejected": -808.7623291015625,
"loss": 0.5337,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.235379695892334,
"rewards/margins": 0.5821799039840698,
"rewards/rejected": -2.8175594806671143,
"step": 611
},
{
"epoch": 0.3997060984570169,
"grad_norm": 14.877552981780648,
"learning_rate": 1.1264808849780429e-07,
"logits/chosen": -1.5760529041290283,
"logits/rejected": -1.5651963949203491,
"logps/chosen": -779.1673583984375,
"logps/rejected": -973.266845703125,
"loss": 0.5072,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3549463748931885,
"rewards/margins": 1.1115797758102417,
"rewards/rejected": -3.4665260314941406,
"step": 612
},
{
"epoch": 0.40035921299697935,
"grad_norm": 38.151466897675625,
"learning_rate": 1.125e-07,
"logits/chosen": -1.5994532108306885,
"logits/rejected": -1.62065589427948,
"logps/chosen": -740.2821655273438,
"logps/rejected": -788.3598022460938,
"loss": 0.5041,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.354248046875,
"rewards/margins": 0.6767839193344116,
"rewards/rejected": -3.031032085418701,
"step": 613
},
{
"epoch": 0.40101232753694177,
"grad_norm": 31.35161543065443,
"learning_rate": 1.123517163097776e-07,
"logits/chosen": -1.6029822826385498,
"logits/rejected": -1.6242347955703735,
"logps/chosen": -775.5657348632812,
"logps/rejected": -801.442138671875,
"loss": 0.5279,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.5471549034118652,
"rewards/margins": 0.6026932001113892,
"rewards/rejected": -3.149848222732544,
"step": 614
},
{
"epoch": 0.40166544207690424,
"grad_norm": 42.24561626020638,
"learning_rate": 1.1220323819897319e-07,
"logits/chosen": -1.525698184967041,
"logits/rejected": -1.528663992881775,
"logps/chosen": -715.5992431640625,
"logps/rejected": -701.003173828125,
"loss": 0.5406,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.155259609222412,
"rewards/margins": 0.44246095418930054,
"rewards/rejected": -2.5977203845977783,
"step": 615
},
{
"epoch": 0.40231855661686666,
"grad_norm": 43.53559559563413,
"learning_rate": 1.120545664404348e-07,
"logits/chosen": -1.5581973791122437,
"logits/rejected": -1.511064052581787,
"logps/chosen": -835.9921264648438,
"logps/rejected": -890.62744140625,
"loss": 0.4782,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.300487518310547,
"rewards/margins": 0.912307620048523,
"rewards/rejected": -3.2127950191497803,
"step": 616
},
{
"epoch": 0.40297167115682914,
"grad_norm": 22.99506036909689,
"learning_rate": 1.1190570180801842e-07,
"logits/chosen": -1.5422519445419312,
"logits/rejected": -1.580669641494751,
"logps/chosen": -782.707763671875,
"logps/rejected": -868.0604858398438,
"loss": 0.5118,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.45739483833313,
"rewards/margins": 0.641864538192749,
"rewards/rejected": -3.0992591381073,
"step": 617
},
{
"epoch": 0.40362478569679155,
"grad_norm": 9.096921333008568,
"learning_rate": 1.11756645076584e-07,
"logits/chosen": -1.5296615362167358,
"logits/rejected": -1.5493112802505493,
"logps/chosen": -742.4532470703125,
"logps/rejected": -760.0758666992188,
"loss": 0.4997,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.1916160583496094,
"rewards/margins": 0.4671185612678528,
"rewards/rejected": -2.6587345600128174,
"step": 618
},
{
"epoch": 0.40427790023675403,
"grad_norm": 19.544727760189303,
"learning_rate": 1.1160739702199136e-07,
"logits/chosen": -1.5654833316802979,
"logits/rejected": -1.6011905670166016,
"logps/chosen": -678.881103515625,
"logps/rejected": -688.1021728515625,
"loss": 0.605,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.180164098739624,
"rewards/margins": 0.3665446937084198,
"rewards/rejected": -2.546708583831787,
"step": 619
},
{
"epoch": 0.40493101477671645,
"grad_norm": 33.42410909057005,
"learning_rate": 1.1145795842109621e-07,
"logits/chosen": -1.4947848320007324,
"logits/rejected": -1.5099647045135498,
"logps/chosen": -757.1992797851562,
"logps/rejected": -829.0934448242188,
"loss": 0.6061,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.2935428619384766,
"rewards/margins": 0.46146607398986816,
"rewards/rejected": -2.7550089359283447,
"step": 620
},
{
"epoch": 0.4055841293166789,
"grad_norm": 18.79594230641745,
"learning_rate": 1.1130833005174605e-07,
"logits/chosen": -1.5316284894943237,
"logits/rejected": -1.5386699438095093,
"logps/chosen": -728.5615844726562,
"logps/rejected": -749.1942138671875,
"loss": 0.5481,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.2535488605499268,
"rewards/margins": 0.5478615760803223,
"rewards/rejected": -2.80141019821167,
"step": 621
},
{
"epoch": 0.40623724385664134,
"grad_norm": 8.286798076996332,
"learning_rate": 1.1115851269277616e-07,
"logits/chosen": -1.6122698783874512,
"logits/rejected": -1.6200590133666992,
"logps/chosen": -774.5802001953125,
"logps/rejected": -867.3875732421875,
"loss": 0.5414,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.146087884902954,
"rewards/margins": 0.6804053783416748,
"rewards/rejected": -2.82649302482605,
"step": 622
},
{
"epoch": 0.4068903583966038,
"grad_norm": 36.91218473973044,
"learning_rate": 1.1100850712400558e-07,
"logits/chosen": -1.392214298248291,
"logits/rejected": -1.4010385274887085,
"logps/chosen": -678.6796264648438,
"logps/rejected": -669.0376586914062,
"loss": 0.5327,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.425049066543579,
"rewards/margins": 0.2620934844017029,
"rewards/rejected": -2.6871426105499268,
"step": 623
},
{
"epoch": 0.40754347293656623,
"grad_norm": 7.985703028509244,
"learning_rate": 1.1085831412623295e-07,
"logits/chosen": -1.5507965087890625,
"logits/rejected": -1.5302854776382446,
"logps/chosen": -759.2698364257812,
"logps/rejected": -793.293701171875,
"loss": 0.5103,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.41506290435791,
"rewards/margins": 0.651451826095581,
"rewards/rejected": -3.0665149688720703,
"step": 624
},
{
"epoch": 0.4081965874765287,
"grad_norm": 96.15150960782869,
"learning_rate": 1.107079344812325e-07,
"logits/chosen": -1.5727829933166504,
"logits/rejected": -1.583585500717163,
"logps/chosen": -742.6324462890625,
"logps/rejected": -827.9069213867188,
"loss": 0.5681,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.324497938156128,
"rewards/margins": 0.5150927901268005,
"rewards/rejected": -2.839590549468994,
"step": 625
},
{
"epoch": 0.4088497020164911,
"grad_norm": 44.171131036436144,
"learning_rate": 1.1055736897175004e-07,
"logits/chosen": -1.5920262336730957,
"logits/rejected": -1.5330153703689575,
"logps/chosen": -811.8800048828125,
"logps/rejected": -813.5941772460938,
"loss": 0.5467,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.6739134788513184,
"rewards/margins": 0.49917757511138916,
"rewards/rejected": -3.173091173171997,
"step": 626
},
{
"epoch": 0.4095028165564536,
"grad_norm": 56.76814949785248,
"learning_rate": 1.1040661838149878e-07,
"logits/chosen": -1.5579776763916016,
"logits/rejected": -1.5597628355026245,
"logps/chosen": -807.7301025390625,
"logps/rejected": -894.228515625,
"loss": 0.4822,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6176559925079346,
"rewards/margins": 0.6038171648979187,
"rewards/rejected": -3.221472978591919,
"step": 627
},
{
"epoch": 0.410155931096416,
"grad_norm": 26.749818630716884,
"learning_rate": 1.1025568349515528e-07,
"logits/chosen": -1.4747806787490845,
"logits/rejected": -1.4980195760726929,
"logps/chosen": -683.0436401367188,
"logps/rejected": -686.6530151367188,
"loss": 0.5013,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.168447732925415,
"rewards/margins": 0.4241008460521698,
"rewards/rejected": -2.5925486087799072,
"step": 628
},
{
"epoch": 0.4108090456363785,
"grad_norm": 76.09203034611112,
"learning_rate": 1.1010456509835548e-07,
"logits/chosen": -1.6177140474319458,
"logits/rejected": -1.6009244918823242,
"logps/chosen": -720.8223876953125,
"logps/rejected": -741.0955810546875,
"loss": 0.5919,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.544870376586914,
"rewards/margins": 0.48653483390808105,
"rewards/rejected": -3.031405210494995,
"step": 629
},
{
"epoch": 0.4114621601763409,
"grad_norm": 52.43809285107618,
"learning_rate": 1.0995326397769042e-07,
"logits/chosen": -1.529783010482788,
"logits/rejected": -1.5328000783920288,
"logps/chosen": -664.4801025390625,
"logps/rejected": -768.4127807617188,
"loss": 0.4978,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.260840654373169,
"rewards/margins": 0.6211092472076416,
"rewards/rejected": -2.8819501399993896,
"step": 630
},
{
"epoch": 0.4121152747163034,
"grad_norm": 9.495957129740592,
"learning_rate": 1.0980178092070225e-07,
"logits/chosen": -1.578983187675476,
"logits/rejected": -1.5786319971084595,
"logps/chosen": -745.9150390625,
"logps/rejected": -808.736083984375,
"loss": 0.5322,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.5862929821014404,
"rewards/margins": 0.4649811387062073,
"rewards/rejected": -3.051273822784424,
"step": 631
},
{
"epoch": 0.4127683892562658,
"grad_norm": 40.60424346315586,
"learning_rate": 1.0965011671588021e-07,
"logits/chosen": -1.5746647119522095,
"logits/rejected": -1.554513931274414,
"logps/chosen": -797.9632568359375,
"logps/rejected": -785.4474487304688,
"loss": 0.601,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.463163375854492,
"rewards/margins": 0.24002833664417267,
"rewards/rejected": -2.7031917572021484,
"step": 632
},
{
"epoch": 0.4134215037962283,
"grad_norm": 52.831172156435784,
"learning_rate": 1.094982721526563e-07,
"logits/chosen": -1.540035367012024,
"logits/rejected": -1.5365304946899414,
"logps/chosen": -826.9459228515625,
"logps/rejected": -893.4970703125,
"loss": 0.4984,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0703601837158203,
"rewards/margins": 0.7675031423568726,
"rewards/rejected": -3.8378632068634033,
"step": 633
},
{
"epoch": 0.4140746183361907,
"grad_norm": 15.726901054602733,
"learning_rate": 1.0934624802140147e-07,
"logits/chosen": -1.5329731702804565,
"logits/rejected": -1.5518198013305664,
"logps/chosen": -763.7630615234375,
"logps/rejected": -806.2711791992188,
"loss": 0.5294,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.819103479385376,
"rewards/margins": 0.638033926486969,
"rewards/rejected": -3.457137107849121,
"step": 634
},
{
"epoch": 0.41472773287615317,
"grad_norm": 18.674657678831167,
"learning_rate": 1.0919404511342121e-07,
"logits/chosen": -1.5634597539901733,
"logits/rejected": -1.5584088563919067,
"logps/chosen": -730.9223022460938,
"logps/rejected": -818.416259765625,
"loss": 0.5022,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.1986281871795654,
"rewards/margins": 0.8553330898284912,
"rewards/rejected": -3.0539615154266357,
"step": 635
},
{
"epoch": 0.4153808474161156,
"grad_norm": 19.5510889690524,
"learning_rate": 1.0904166422095162e-07,
"logits/chosen": -1.5800487995147705,
"logits/rejected": -1.5099263191223145,
"logps/chosen": -743.9994506835938,
"logps/rejected": -761.8692016601562,
"loss": 0.5499,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.8127152919769287,
"rewards/margins": 0.35885605216026306,
"rewards/rejected": -3.1715714931488037,
"step": 636
},
{
"epoch": 0.41603396195607806,
"grad_norm": 29.4349268646558,
"learning_rate": 1.0888910613715523e-07,
"logits/chosen": -1.5855112075805664,
"logits/rejected": -1.5745254755020142,
"logps/chosen": -724.60888671875,
"logps/rejected": -817.6202392578125,
"loss": 0.5035,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.4678425788879395,
"rewards/margins": 0.9483436346054077,
"rewards/rejected": -3.416186571121216,
"step": 637
},
{
"epoch": 0.4166870764960405,
"grad_norm": 27.606512517928415,
"learning_rate": 1.0873637165611688e-07,
"logits/chosen": -1.5267822742462158,
"logits/rejected": -1.5166692733764648,
"logps/chosen": -756.2583618164062,
"logps/rejected": -905.69287109375,
"loss": 0.5459,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.325356960296631,
"rewards/margins": 0.8004791736602783,
"rewards/rejected": -3.125836133956909,
"step": 638
},
{
"epoch": 0.41734019103600295,
"grad_norm": 10.112999624727967,
"learning_rate": 1.085834615728396e-07,
"logits/chosen": -1.5627224445343018,
"logits/rejected": -1.589940071105957,
"logps/chosen": -757.5079345703125,
"logps/rejected": -797.4879150390625,
"loss": 0.5334,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.7449193000793457,
"rewards/margins": 0.47740480303764343,
"rewards/rejected": -3.2223243713378906,
"step": 639
},
{
"epoch": 0.41799330557596537,
"grad_norm": 49.554186475304675,
"learning_rate": 1.0843037668324037e-07,
"logits/chosen": -1.6192628145217896,
"logits/rejected": -1.6073408126831055,
"logps/chosen": -692.6961669921875,
"logps/rejected": -731.430419921875,
"loss": 0.5213,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.307335376739502,
"rewards/margins": 0.42022186517715454,
"rewards/rejected": -2.7275571823120117,
"step": 640
},
{
"epoch": 0.41864642011592784,
"grad_norm": 12.12872006128116,
"learning_rate": 1.0827711778414616e-07,
"logits/chosen": -1.5925276279449463,
"logits/rejected": -1.617337703704834,
"logps/chosen": -773.3049926757812,
"logps/rejected": -886.1497192382812,
"loss": 0.5054,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7378339767456055,
"rewards/margins": 0.9473193287849426,
"rewards/rejected": -3.6851532459259033,
"step": 641
},
{
"epoch": 0.41929953465589026,
"grad_norm": 9.420684381655802,
"learning_rate": 1.0812368567328965e-07,
"logits/chosen": -1.5554149150848389,
"logits/rejected": -1.5351797342300415,
"logps/chosen": -741.2017211914062,
"logps/rejected": -848.9796752929688,
"loss": 0.5536,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7187960147857666,
"rewards/margins": 0.7441635727882385,
"rewards/rejected": -3.4629597663879395,
"step": 642
},
{
"epoch": 0.41995264919585273,
"grad_norm": 17.59371870537502,
"learning_rate": 1.0797008114930504e-07,
"logits/chosen": -1.5738487243652344,
"logits/rejected": -1.5535151958465576,
"logps/chosen": -871.8766479492188,
"logps/rejected": -991.9896240234375,
"loss": 0.52,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.753142833709717,
"rewards/margins": 0.8435327410697937,
"rewards/rejected": -3.5966756343841553,
"step": 643
},
{
"epoch": 0.42060576373581515,
"grad_norm": 65.67234375195817,
"learning_rate": 1.078163050117241e-07,
"logits/chosen": -1.5654664039611816,
"logits/rejected": -1.5345994234085083,
"logps/chosen": -724.1328735351562,
"logps/rejected": -764.9933471679688,
"loss": 0.5788,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.66972279548645,
"rewards/margins": 0.42094868421554565,
"rewards/rejected": -3.0906717777252197,
"step": 644
},
{
"epoch": 0.4212588782757776,
"grad_norm": 29.19611600663847,
"learning_rate": 1.0766235806097172e-07,
"logits/chosen": -1.6037955284118652,
"logits/rejected": -1.552761197090149,
"logps/chosen": -850.3397216796875,
"logps/rejected": -891.075927734375,
"loss": 0.498,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.936112880706787,
"rewards/margins": 0.9705631136894226,
"rewards/rejected": -3.9066760540008545,
"step": 645
},
{
"epoch": 0.42191199281574004,
"grad_norm": 55.001913874158426,
"learning_rate": 1.0750824109836202e-07,
"logits/chosen": -1.63307785987854,
"logits/rejected": -1.5964423418045044,
"logps/chosen": -867.8717041015625,
"logps/rejected": -924.471435546875,
"loss": 0.577,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.914452314376831,
"rewards/margins": 0.48612168431282043,
"rewards/rejected": -3.40057373046875,
"step": 646
},
{
"epoch": 0.4225651073557025,
"grad_norm": 28.42526755883717,
"learning_rate": 1.0735395492609401e-07,
"logits/chosen": -1.477207899093628,
"logits/rejected": -1.4840601682662964,
"logps/chosen": -654.694091796875,
"logps/rejected": -722.9656982421875,
"loss": 0.5245,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.427155017852783,
"rewards/margins": 0.7215266227722168,
"rewards/rejected": -3.148681879043579,
"step": 647
},
{
"epoch": 0.42321822189566494,
"grad_norm": 20.153695342892174,
"learning_rate": 1.0719950034724741e-07,
"logits/chosen": -1.6053216457366943,
"logits/rejected": -1.58915114402771,
"logps/chosen": -833.958251953125,
"logps/rejected": -869.42236328125,
"loss": 0.5383,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.8524227142333984,
"rewards/margins": 0.6670438051223755,
"rewards/rejected": -3.5194666385650635,
"step": 648
},
{
"epoch": 0.4238713364356274,
"grad_norm": 40.45524219360296,
"learning_rate": 1.0704487816577857e-07,
"logits/chosen": -1.5915526151657104,
"logits/rejected": -1.5662232637405396,
"logps/chosen": -755.484619140625,
"logps/rejected": -798.5493774414062,
"loss": 0.5185,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.445694923400879,
"rewards/margins": 0.6455334424972534,
"rewards/rejected": -3.0912282466888428,
"step": 649
},
{
"epoch": 0.42452445097558983,
"grad_norm": 45.48878539730004,
"learning_rate": 1.0689008918651624e-07,
"logits/chosen": -1.4695310592651367,
"logits/rejected": -1.4832361936569214,
"logps/chosen": -884.2122802734375,
"logps/rejected": -879.8310546875,
"loss": 0.5647,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.048762559890747,
"rewards/margins": 0.2833816707134247,
"rewards/rejected": -3.332144260406494,
"step": 650
},
{
"epoch": 0.4251775655155523,
"grad_norm": 56.42516159244524,
"learning_rate": 1.0673513421515733e-07,
"logits/chosen": -1.6036771535873413,
"logits/rejected": -1.617321252822876,
"logps/chosen": -794.57421875,
"logps/rejected": -828.6959838867188,
"loss": 0.5217,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.69077467918396,
"rewards/margins": 0.46351078152656555,
"rewards/rejected": -3.154285430908203,
"step": 651
},
{
"epoch": 0.4258306800555147,
"grad_norm": 46.62254829895216,
"learning_rate": 1.0658001405826283e-07,
"logits/chosen": -1.5648261308670044,
"logits/rejected": -1.5797994136810303,
"logps/chosen": -772.7039184570312,
"logps/rejected": -819.6089477539062,
"loss": 0.5396,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7385077476501465,
"rewards/margins": 0.5473735332489014,
"rewards/rejected": -3.285881519317627,
"step": 652
},
{
"epoch": 0.4264837945954772,
"grad_norm": 59.34200258381232,
"learning_rate": 1.0642472952325346e-07,
"logits/chosen": -1.5929808616638184,
"logits/rejected": -1.6374764442443848,
"logps/chosen": -735.03662109375,
"logps/rejected": -850.94189453125,
"loss": 0.4824,
"rewards/accuracies": 0.9375,
"rewards/chosen": -2.357342004776001,
"rewards/margins": 1.2214264869689941,
"rewards/rejected": -3.578768253326416,
"step": 653
},
{
"epoch": 0.4271369091354396,
"grad_norm": 37.46472982323093,
"learning_rate": 1.062692814184056e-07,
"logits/chosen": -1.5046641826629639,
"logits/rejected": -1.4834293127059937,
"logps/chosen": -670.6259765625,
"logps/rejected": -814.8147583007812,
"loss": 0.5009,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.507924795150757,
"rewards/margins": 0.8526814579963684,
"rewards/rejected": -3.3606061935424805,
"step": 654
},
{
"epoch": 0.4277900236754021,
"grad_norm": 65.41811386558527,
"learning_rate": 1.0611367055284704e-07,
"logits/chosen": -1.5957469940185547,
"logits/rejected": -1.6134015321731567,
"logps/chosen": -910.1995239257812,
"logps/rejected": -958.0068969726562,
"loss": 0.5173,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.841102123260498,
"rewards/margins": 0.8195388913154602,
"rewards/rejected": -3.6606409549713135,
"step": 655
},
{
"epoch": 0.4284431382153645,
"grad_norm": 15.408196632262841,
"learning_rate": 1.0595789773655273e-07,
"logits/chosen": -1.5747413635253906,
"logits/rejected": -1.565165400505066,
"logps/chosen": -880.7881469726562,
"logps/rejected": -917.05419921875,
"loss": 0.5269,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.843327760696411,
"rewards/margins": 0.5438075661659241,
"rewards/rejected": -3.3871350288391113,
"step": 656
},
{
"epoch": 0.429096252755327,
"grad_norm": 12.34615445550627,
"learning_rate": 1.0580196378034061e-07,
"logits/chosen": -1.5093989372253418,
"logits/rejected": -1.528839111328125,
"logps/chosen": -804.33837890625,
"logps/rejected": -832.3118896484375,
"loss": 0.5247,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.883572816848755,
"rewards/margins": 0.4760682284832001,
"rewards/rejected": -3.3596410751342773,
"step": 657
},
{
"epoch": 0.4297493672952894,
"grad_norm": 20.79744618252237,
"learning_rate": 1.0564586949586736e-07,
"logits/chosen": -1.5655585527420044,
"logits/rejected": -1.5473308563232422,
"logps/chosen": -784.39208984375,
"logps/rejected": -780.73974609375,
"loss": 0.5298,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7032248973846436,
"rewards/margins": 0.7658544182777405,
"rewards/rejected": -3.46907901763916,
"step": 658
},
{
"epoch": 0.43040248183525187,
"grad_norm": 9.541823346620628,
"learning_rate": 1.0548961569562423e-07,
"logits/chosen": -1.551105260848999,
"logits/rejected": -1.5648680925369263,
"logps/chosen": -728.1370239257812,
"logps/rejected": -863.0052490234375,
"loss": 0.5929,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.4059126377105713,
"rewards/margins": 0.7981605529785156,
"rewards/rejected": -3.2040724754333496,
"step": 659
},
{
"epoch": 0.4310555963752143,
"grad_norm": 48.77651510525204,
"learning_rate": 1.0533320319293272e-07,
"logits/chosen": -1.521799087524414,
"logits/rejected": -1.5136668682098389,
"logps/chosen": -767.982421875,
"logps/rejected": -844.113037109375,
"loss": 0.5468,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.578664541244507,
"rewards/margins": 0.6596671342849731,
"rewards/rejected": -3.2383313179016113,
"step": 660
},
{
"epoch": 0.43170871091517676,
"grad_norm": 14.135228489391437,
"learning_rate": 1.0517663280194042e-07,
"logits/chosen": -1.541378140449524,
"logits/rejected": -1.5287246704101562,
"logps/chosen": -854.5809326171875,
"logps/rejected": -845.8003540039062,
"loss": 0.5743,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.7386510372161865,
"rewards/margins": 0.5117663145065308,
"rewards/rejected": -3.2504169940948486,
"step": 661
},
{
"epoch": 0.4323618254551392,
"grad_norm": 39.96027400878076,
"learning_rate": 1.050199053376168e-07,
"logits/chosen": -1.5787113904953003,
"logits/rejected": -1.5196938514709473,
"logps/chosen": -895.7177734375,
"logps/rejected": -902.5272827148438,
"loss": 0.5603,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.0833802223205566,
"rewards/margins": 0.44520139694213867,
"rewards/rejected": -3.5285820960998535,
"step": 662
},
{
"epoch": 0.43301493999510166,
"grad_norm": 9.336642307745894,
"learning_rate": 1.0486302161574876e-07,
"logits/chosen": -1.5741958618164062,
"logits/rejected": -1.566494345664978,
"logps/chosen": -747.9061889648438,
"logps/rejected": -897.3228759765625,
"loss": 0.5216,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.3504905700683594,
"rewards/margins": 1.2214593887329102,
"rewards/rejected": -3.5719499588012695,
"step": 663
},
{
"epoch": 0.4336680545350641,
"grad_norm": 34.74703628355522,
"learning_rate": 1.0470598245293676e-07,
"logits/chosen": -1.5200121402740479,
"logits/rejected": -1.537473201751709,
"logps/chosen": -792.9605102539062,
"logps/rejected": -799.4251708984375,
"loss": 0.5492,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.7280704975128174,
"rewards/margins": 0.4831348657608032,
"rewards/rejected": -3.211205244064331,
"step": 664
},
{
"epoch": 0.43432116907502655,
"grad_norm": 26.166178900391024,
"learning_rate": 1.0454878866659017e-07,
"logits/chosen": -1.591930627822876,
"logits/rejected": -1.5695099830627441,
"logps/chosen": -727.5706787109375,
"logps/rejected": -792.0450439453125,
"loss": 0.5506,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.6660943031311035,
"rewards/margins": 0.6367244124412537,
"rewards/rejected": -3.302818775177002,
"step": 665
},
{
"epoch": 0.43497428361498897,
"grad_norm": 10.372172666812576,
"learning_rate": 1.0439144107492328e-07,
"logits/chosen": -1.4827499389648438,
"logits/rejected": -1.4672393798828125,
"logps/chosen": -800.9483032226562,
"logps/rejected": -798.546875,
"loss": 0.5153,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8575448989868164,
"rewards/margins": 0.5324373841285706,
"rewards/rejected": -3.389982223510742,
"step": 666
},
{
"epoch": 0.43562739815495144,
"grad_norm": 11.965156702513355,
"learning_rate": 1.0423394049695095e-07,
"logits/chosen": -1.5964605808258057,
"logits/rejected": -1.5629247426986694,
"logps/chosen": -812.4568481445312,
"logps/rejected": -916.690185546875,
"loss": 0.5439,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.873971939086914,
"rewards/margins": 0.766517162322998,
"rewards/rejected": -3.640489101409912,
"step": 667
},
{
"epoch": 0.43628051269491386,
"grad_norm": 17.25391970334388,
"learning_rate": 1.0407628775248433e-07,
"logits/chosen": -1.5258575677871704,
"logits/rejected": -1.557664155960083,
"logps/chosen": -774.147216796875,
"logps/rejected": -935.6986083984375,
"loss": 0.4742,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.4668354988098145,
"rewards/margins": 1.3980622291564941,
"rewards/rejected": -3.8648977279663086,
"step": 668
},
{
"epoch": 0.43693362723487633,
"grad_norm": 14.901011998748888,
"learning_rate": 1.0391848366212666e-07,
"logits/chosen": -1.594072937965393,
"logits/rejected": -1.587380051612854,
"logps/chosen": -780.095703125,
"logps/rejected": -784.4081420898438,
"loss": 0.5749,
"rewards/accuracies": 0.5625,
"rewards/chosen": -2.8471426963806152,
"rewards/margins": 0.19949409365653992,
"rewards/rejected": -3.0466368198394775,
"step": 669
},
{
"epoch": 0.43758674177483875,
"grad_norm": 46.17531372330182,
"learning_rate": 1.0376052904726888e-07,
"logits/chosen": -1.6172115802764893,
"logits/rejected": -1.5498206615447998,
"logps/chosen": -820.4274291992188,
"logps/rejected": -908.1333618164062,
"loss": 0.5152,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.615933656692505,
"rewards/margins": 0.7912464737892151,
"rewards/rejected": -3.407179832458496,
"step": 670
},
{
"epoch": 0.4382398563148012,
"grad_norm": 51.65597788618139,
"learning_rate": 1.0360242473008551e-07,
"logits/chosen": -1.5762639045715332,
"logits/rejected": -1.594434380531311,
"logps/chosen": -824.281982421875,
"logps/rejected": -829.8516845703125,
"loss": 0.5618,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.047189712524414,
"rewards/margins": 0.6297323703765869,
"rewards/rejected": -3.67692232131958,
"step": 671
},
{
"epoch": 0.43889297085476364,
"grad_norm": 33.441136796480215,
"learning_rate": 1.0344417153353023e-07,
"logits/chosen": -1.4966257810592651,
"logits/rejected": -1.459989309310913,
"logps/chosen": -817.207763671875,
"logps/rejected": -897.0996704101562,
"loss": 0.5394,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9682440757751465,
"rewards/margins": 0.6141051054000854,
"rewards/rejected": -3.5823495388031006,
"step": 672
},
{
"epoch": 0.4395460853947261,
"grad_norm": 31.378606252739424,
"learning_rate": 1.0328577028133171e-07,
"logits/chosen": -1.5966362953186035,
"logits/rejected": -1.6483103036880493,
"logps/chosen": -732.49609375,
"logps/rejected": -800.7023315429688,
"loss": 0.5138,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.764153003692627,
"rewards/margins": 0.5777581930160522,
"rewards/rejected": -3.341911554336548,
"step": 673
},
{
"epoch": 0.44019919993468853,
"grad_norm": 34.56431138782349,
"learning_rate": 1.0312722179798924e-07,
"logits/chosen": -1.598524808883667,
"logits/rejected": -1.6316325664520264,
"logps/chosen": -787.64892578125,
"logps/rejected": -861.8873291015625,
"loss": 0.5273,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7730770111083984,
"rewards/margins": 0.7392612099647522,
"rewards/rejected": -3.5123379230499268,
"step": 674
},
{
"epoch": 0.440852314474651,
"grad_norm": 15.372795733191637,
"learning_rate": 1.0296852690876846e-07,
"logits/chosen": -1.5273460149765015,
"logits/rejected": -1.5593777894973755,
"logps/chosen": -790.3528442382812,
"logps/rejected": -815.5953979492188,
"loss": 0.5533,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.494476079940796,
"rewards/margins": 0.4577151834964752,
"rewards/rejected": -2.95219087600708,
"step": 675
},
{
"epoch": 0.4415054290146134,
"grad_norm": 36.80495254523571,
"learning_rate": 1.0280968643969706e-07,
"logits/chosen": -1.5882431268692017,
"logits/rejected": -1.5868853330612183,
"logps/chosen": -863.531005859375,
"logps/rejected": -894.946044921875,
"loss": 0.5153,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.805412769317627,
"rewards/margins": 0.8240199685096741,
"rewards/rejected": -3.629432201385498,
"step": 676
},
{
"epoch": 0.4421585435545759,
"grad_norm": 9.850541147286798,
"learning_rate": 1.0265070121756054e-07,
"logits/chosen": -1.5755228996276855,
"logits/rejected": -1.5751676559448242,
"logps/chosen": -729.009521484375,
"logps/rejected": -794.595947265625,
"loss": 0.5564,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.3728232383728027,
"rewards/margins": 0.473653107881546,
"rewards/rejected": -2.8464765548706055,
"step": 677
},
{
"epoch": 0.4428116580945383,
"grad_norm": 8.929995911050677,
"learning_rate": 1.0249157206989785e-07,
"logits/chosen": -1.5073692798614502,
"logits/rejected": -1.542021632194519,
"logps/chosen": -802.9788818359375,
"logps/rejected": -878.1073608398438,
"loss": 0.4953,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.467076301574707,
"rewards/margins": 0.7465323209762573,
"rewards/rejected": -3.213608503341675,
"step": 678
},
{
"epoch": 0.4434647726345008,
"grad_norm": 17.43336381715629,
"learning_rate": 1.0233229982499702e-07,
"logits/chosen": -1.5233006477355957,
"logits/rejected": -1.575763463973999,
"logps/chosen": -728.665771484375,
"logps/rejected": -826.36279296875,
"loss": 0.5105,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.66928768157959,
"rewards/margins": 0.6221798062324524,
"rewards/rejected": -3.2914676666259766,
"step": 679
},
{
"epoch": 0.4441178871744632,
"grad_norm": 8.33574898866346,
"learning_rate": 1.0217288531189101e-07,
"logits/chosen": -1.4891436100006104,
"logits/rejected": -1.487116813659668,
"logps/chosen": -790.6260375976562,
"logps/rejected": -806.4535522460938,
"loss": 0.5277,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7111411094665527,
"rewards/margins": 0.511049747467041,
"rewards/rejected": -3.2221908569335938,
"step": 680
},
{
"epoch": 0.4447710017144257,
"grad_norm": 93.25103994990552,
"learning_rate": 1.0201332936035328e-07,
"logits/chosen": -1.5428980588912964,
"logits/rejected": -1.5394231081008911,
"logps/chosen": -750.823486328125,
"logps/rejected": -812.9085083007812,
"loss": 0.6195,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.5228466987609863,
"rewards/margins": 0.6121708750724792,
"rewards/rejected": -3.1350173950195312,
"step": 681
},
{
"epoch": 0.4454241162543881,
"grad_norm": 24.723412198537826,
"learning_rate": 1.0185363280089346e-07,
"logits/chosen": -1.5300703048706055,
"logits/rejected": -1.5563037395477295,
"logps/chosen": -808.7451782226562,
"logps/rejected": -814.0613403320312,
"loss": 0.5547,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.553114891052246,
"rewards/margins": 0.2794783413410187,
"rewards/rejected": -2.8325929641723633,
"step": 682
},
{
"epoch": 0.4460772307943506,
"grad_norm": 18.06928652509424,
"learning_rate": 1.0169379646475307e-07,
"logits/chosen": -1.5730717182159424,
"logits/rejected": -1.553600549697876,
"logps/chosen": -827.7550048828125,
"logps/rejected": -824.0142211914062,
"loss": 0.5089,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.8725967407226562,
"rewards/margins": 0.5028645992279053,
"rewards/rejected": -3.3754611015319824,
"step": 683
},
{
"epoch": 0.446730345334313,
"grad_norm": 13.432734617314106,
"learning_rate": 1.0153382118390124e-07,
"logits/chosen": -1.517228364944458,
"logits/rejected": -1.5097568035125732,
"logps/chosen": -828.512451171875,
"logps/rejected": -960.2579345703125,
"loss": 0.5539,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.074080467224121,
"rewards/margins": 0.7247848510742188,
"rewards/rejected": -3.7988648414611816,
"step": 684
},
{
"epoch": 0.44738345987427547,
"grad_norm": 111.08651677833566,
"learning_rate": 1.0137370779103024e-07,
"logits/chosen": -1.6381279230117798,
"logits/rejected": -1.642504334449768,
"logps/chosen": -895.6856079101562,
"logps/rejected": -954.673828125,
"loss": 0.4931,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.942159652709961,
"rewards/margins": 0.6529674530029297,
"rewards/rejected": -3.5951271057128906,
"step": 685
},
{
"epoch": 0.4480365744142379,
"grad_norm": 13.010841457025625,
"learning_rate": 1.0121345711955134e-07,
"logits/chosen": -1.4719197750091553,
"logits/rejected": -1.4533613920211792,
"logps/chosen": -753.68408203125,
"logps/rejected": -819.1432495117188,
"loss": 0.5515,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.3883912563323975,
"rewards/margins": 0.776418149471283,
"rewards/rejected": -3.1648097038269043,
"step": 686
},
{
"epoch": 0.44868968895420036,
"grad_norm": 19.59305991487636,
"learning_rate": 1.0105307000359027e-07,
"logits/chosen": -1.5581806898117065,
"logits/rejected": -1.563573956489563,
"logps/chosen": -782.7869873046875,
"logps/rejected": -825.7442626953125,
"loss": 0.5426,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.6787917613983154,
"rewards/margins": 0.797275722026825,
"rewards/rejected": -3.4760677814483643,
"step": 687
},
{
"epoch": 0.4493428034941628,
"grad_norm": 41.30661441239489,
"learning_rate": 1.0089254727798299e-07,
"logits/chosen": -1.5456461906433105,
"logits/rejected": -1.5804781913757324,
"logps/chosen": -802.162841796875,
"logps/rejected": -830.0077514648438,
"loss": 0.5099,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7756364345550537,
"rewards/margins": 0.779686450958252,
"rewards/rejected": -3.5553228855133057,
"step": 688
},
{
"epoch": 0.44999591803412525,
"grad_norm": 43.257182510610015,
"learning_rate": 1.0073188977827134e-07,
"logits/chosen": -1.530835509300232,
"logits/rejected": -1.5210024118423462,
"logps/chosen": -774.5211791992188,
"logps/rejected": -830.6015625,
"loss": 0.4923,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.939332962036133,
"rewards/margins": 0.8855284452438354,
"rewards/rejected": -3.824861526489258,
"step": 689
},
{
"epoch": 0.45064903257408767,
"grad_norm": 28.06793800668904,
"learning_rate": 1.005710983406987e-07,
"logits/chosen": -1.5865156650543213,
"logits/rejected": -1.536832571029663,
"logps/chosen": -714.5545654296875,
"logps/rejected": -739.9314575195312,
"loss": 0.5408,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.3614490032196045,
"rewards/margins": 0.4486488997936249,
"rewards/rejected": -2.8100976943969727,
"step": 690
},
{
"epoch": 0.45130214711405015,
"grad_norm": 13.710061169800268,
"learning_rate": 1.0041017380220558e-07,
"logits/chosen": -1.6234124898910522,
"logits/rejected": -1.6429170370101929,
"logps/chosen": -942.2227783203125,
"logps/rejected": -983.3606567382812,
"loss": 0.5934,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9503729343414307,
"rewards/margins": 0.5114213824272156,
"rewards/rejected": -3.461794376373291,
"step": 691
},
{
"epoch": 0.45195526165401256,
"grad_norm": 42.74504837687908,
"learning_rate": 1.002491170004253e-07,
"logits/chosen": -1.5679419040679932,
"logits/rejected": -1.5443594455718994,
"logps/chosen": -842.61474609375,
"logps/rejected": -882.8880004882812,
"loss": 0.5011,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7981529235839844,
"rewards/margins": 0.8635019659996033,
"rewards/rejected": -3.6616549491882324,
"step": 692
},
{
"epoch": 0.45260837619397504,
"grad_norm": 51.01361393063693,
"learning_rate": 1.0008792877367964e-07,
"logits/chosen": -1.6048243045806885,
"logits/rejected": -1.5834286212921143,
"logps/chosen": -800.6567993164062,
"logps/rejected": -932.0943603515625,
"loss": 0.6234,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.7766685485839844,
"rewards/margins": 0.7954378128051758,
"rewards/rejected": -3.57210636138916,
"step": 693
},
{
"epoch": 0.45326149073393746,
"grad_norm": 26.952246598872957,
"learning_rate": 9.992660996097447e-08,
"logits/chosen": -1.487755537033081,
"logits/rejected": -1.475599765777588,
"logps/chosen": -764.441162109375,
"logps/rejected": -858.9654541015625,
"loss": 0.471,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6644558906555176,
"rewards/margins": 0.7122585773468018,
"rewards/rejected": -3.3767144680023193,
"step": 694
},
{
"epoch": 0.45391460527389993,
"grad_norm": 11.341690238802435,
"learning_rate": 9.976516140199535e-08,
"logits/chosen": -1.654636025428772,
"logits/rejected": -1.6279263496398926,
"logps/chosen": -713.5647583007812,
"logps/rejected": -748.8384399414062,
"loss": 0.5867,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.517662525177002,
"rewards/margins": 0.454134076833725,
"rewards/rejected": -2.971796989440918,
"step": 695
},
{
"epoch": 0.45456771981386235,
"grad_norm": 40.33369925086035,
"learning_rate": 9.960358393710321e-08,
"logits/chosen": -1.528939962387085,
"logits/rejected": -1.5497705936431885,
"logps/chosen": -764.209716796875,
"logps/rejected": -920.8458862304688,
"loss": 0.4745,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.5999767780303955,
"rewards/margins": 1.0393584966659546,
"rewards/rejected": -3.6393353939056396,
"step": 696
},
{
"epoch": 0.4552208343538248,
"grad_norm": 12.320269188700282,
"learning_rate": 9.944187840732994e-08,
"logits/chosen": -1.5356950759887695,
"logits/rejected": -1.5276272296905518,
"logps/chosen": -824.4510498046875,
"logps/rejected": -828.6773681640625,
"loss": 0.4832,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7461459636688232,
"rewards/margins": 0.7607426047325134,
"rewards/rejected": -3.5068888664245605,
"step": 697
},
{
"epoch": 0.45587394889378724,
"grad_norm": 20.693131756565563,
"learning_rate": 9.928004565437409e-08,
"logits/chosen": -1.5997885465621948,
"logits/rejected": -1.596414566040039,
"logps/chosen": -749.1096801757812,
"logps/rejected": -785.3474731445312,
"loss": 0.5544,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.6378190517425537,
"rewards/margins": 0.5456668138504028,
"rewards/rejected": -3.183485746383667,
"step": 698
},
{
"epoch": 0.4565270634337497,
"grad_norm": 31.393879848402715,
"learning_rate": 9.911808652059627e-08,
"logits/chosen": -1.5327155590057373,
"logits/rejected": -1.5094130039215088,
"logps/chosen": -713.3516235351562,
"logps/rejected": -709.5831298828125,
"loss": 0.565,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.3836562633514404,
"rewards/margins": 0.40900933742523193,
"rewards/rejected": -2.792665719985962,
"step": 699
},
{
"epoch": 0.45718017797371213,
"grad_norm": 44.56122798633119,
"learning_rate": 9.895600184901504e-08,
"logits/chosen": -1.6142834424972534,
"logits/rejected": -1.5989428758621216,
"logps/chosen": -814.4269409179688,
"logps/rejected": -822.01513671875,
"loss": 0.5342,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7036519050598145,
"rewards/margins": 0.4850347638130188,
"rewards/rejected": -3.1886863708496094,
"step": 700
},
{
"epoch": 0.45718017797371213,
"eval_logits/chosen": -1.5524276494979858,
"eval_logits/rejected": -1.5454105138778687,
"eval_logps/chosen": -772.6592407226562,
"eval_logps/rejected": -832.8743896484375,
"eval_loss": 0.5305144190788269,
"eval_rewards/accuracies": 0.7350000143051147,
"eval_rewards/chosen": -2.623138427734375,
"eval_rewards/margins": 0.6832101345062256,
"eval_rewards/rejected": -3.3063488006591797,
"eval_runtime": 296.7195,
"eval_samples_per_second": 13.481,
"eval_steps_per_second": 0.843,
"step": 700
},
{
"epoch": 0.4578332925136746,
"grad_norm": 8.9780374884061,
"learning_rate": 9.879379248330239e-08,
"logits/chosen": -1.5523990392684937,
"logits/rejected": -1.5356321334838867,
"logps/chosen": -745.2019653320312,
"logps/rejected": -832.36376953125,
"loss": 0.5029,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.4654135704040527,
"rewards/margins": 0.739882230758667,
"rewards/rejected": -3.2052958011627197,
"step": 701
},
{
"epoch": 0.458486407053637,
"grad_norm": 13.764196067118446,
"learning_rate": 9.863145926777934e-08,
"logits/chosen": -1.5704573392868042,
"logits/rejected": -1.5471737384796143,
"logps/chosen": -749.1688232421875,
"logps/rejected": -803.7322998046875,
"loss": 0.5578,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.7841954231262207,
"rewards/margins": 0.7516999840736389,
"rewards/rejected": -3.535895347595215,
"step": 702
},
{
"epoch": 0.4591395215935995,
"grad_norm": 28.637922274224316,
"learning_rate": 9.846900304741157e-08,
"logits/chosen": -1.5777729749679565,
"logits/rejected": -1.5776381492614746,
"logps/chosen": -775.2662353515625,
"logps/rejected": -831.3828735351562,
"loss": 0.5208,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6320080757141113,
"rewards/margins": 0.6068910360336304,
"rewards/rejected": -3.238898992538452,
"step": 703
},
{
"epoch": 0.4597926361335619,
"grad_norm": 19.220774272497522,
"learning_rate": 9.830642466780502e-08,
"logits/chosen": -1.6617802381515503,
"logits/rejected": -1.6172996759414673,
"logps/chosen": -792.265380859375,
"logps/rejected": -793.1304931640625,
"loss": 0.5726,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.6746013164520264,
"rewards/margins": 0.42366841435432434,
"rewards/rejected": -3.0982699394226074,
"step": 704
},
{
"epoch": 0.4604457506735244,
"grad_norm": 9.690502883669247,
"learning_rate": 9.814372497520143e-08,
"logits/chosen": -1.59645414352417,
"logits/rejected": -1.5628377199172974,
"logps/chosen": -759.610595703125,
"logps/rejected": -795.2174072265625,
"loss": 0.4766,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.5422470569610596,
"rewards/margins": 0.4729628562927246,
"rewards/rejected": -3.015209674835205,
"step": 705
},
{
"epoch": 0.4610988652134868,
"grad_norm": 29.25714731199898,
"learning_rate": 9.798090481647411e-08,
"logits/chosen": -1.5286840200424194,
"logits/rejected": -1.5571849346160889,
"logps/chosen": -745.8056640625,
"logps/rejected": -824.3461303710938,
"loss": 0.5433,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.467949867248535,
"rewards/margins": 0.9363622665405273,
"rewards/rejected": -3.4043126106262207,
"step": 706
},
{
"epoch": 0.4617519797534493,
"grad_norm": 11.95355469721655,
"learning_rate": 9.781796503912328e-08,
"logits/chosen": -1.5327837467193604,
"logits/rejected": -1.5477138757705688,
"logps/chosen": -788.408447265625,
"logps/rejected": -918.1892700195312,
"loss": 0.5018,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.641738176345825,
"rewards/margins": 0.9118731617927551,
"rewards/rejected": -3.5536112785339355,
"step": 707
},
{
"epoch": 0.4624050942934117,
"grad_norm": 9.057095892979188,
"learning_rate": 9.765490649127187e-08,
"logits/chosen": -1.58297860622406,
"logits/rejected": -1.5312621593475342,
"logps/chosen": -640.93798828125,
"logps/rejected": -716.5929565429688,
"loss": 0.5052,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.1810810565948486,
"rewards/margins": 0.6035025119781494,
"rewards/rejected": -2.784583568572998,
"step": 708
},
{
"epoch": 0.4630582088333742,
"grad_norm": 17.256671471200423,
"learning_rate": 9.749173002166101e-08,
"logits/chosen": -1.4686298370361328,
"logits/rejected": -1.5486806631088257,
"logps/chosen": -787.2227172851562,
"logps/rejected": -868.33056640625,
"loss": 0.5082,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.844477415084839,
"rewards/margins": 0.6099371314048767,
"rewards/rejected": -3.4544146060943604,
"step": 709
},
{
"epoch": 0.4637113233733366,
"grad_norm": 11.493776981683995,
"learning_rate": 9.732843647964563e-08,
"logits/chosen": -1.3998236656188965,
"logits/rejected": -1.3739439249038696,
"logps/chosen": -645.9183959960938,
"logps/rejected": -692.6560668945312,
"loss": 0.5008,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.068920135498047,
"rewards/margins": 0.3378119170665741,
"rewards/rejected": -2.4067318439483643,
"step": 710
},
{
"epoch": 0.46436443791329907,
"grad_norm": 19.407875344745552,
"learning_rate": 9.716502671519003e-08,
"logits/chosen": -1.594405174255371,
"logits/rejected": -1.5376571416854858,
"logps/chosen": -743.6107177734375,
"logps/rejected": -760.2347412109375,
"loss": 0.5606,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.476472854614258,
"rewards/margins": 0.6245778799057007,
"rewards/rejected": -3.101050853729248,
"step": 711
},
{
"epoch": 0.4650175524532615,
"grad_norm": 30.65022124829921,
"learning_rate": 9.700150157886345e-08,
"logits/chosen": -1.4861985445022583,
"logits/rejected": -1.4618767499923706,
"logps/chosen": -715.8246459960938,
"logps/rejected": -819.9178466796875,
"loss": 0.4772,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.780965805053711,
"rewards/margins": 0.6235308647155762,
"rewards/rejected": -3.404496669769287,
"step": 712
},
{
"epoch": 0.46567066699322396,
"grad_norm": 66.74879517812167,
"learning_rate": 9.683786192183569e-08,
"logits/chosen": -1.5658459663391113,
"logits/rejected": -1.5532565116882324,
"logps/chosen": -791.1273803710938,
"logps/rejected": -858.0960083007812,
"loss": 0.4781,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.74170184135437,
"rewards/margins": 0.8308650255203247,
"rewards/rejected": -3.5725669860839844,
"step": 713
},
{
"epoch": 0.4663237815331864,
"grad_norm": 58.623748201586416,
"learning_rate": 9.667410859587261e-08,
"logits/chosen": -1.5724753141403198,
"logits/rejected": -1.581168532371521,
"logps/chosen": -814.78125,
"logps/rejected": -851.2848510742188,
"loss": 0.5508,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.974850654602051,
"rewards/margins": 0.2711948752403259,
"rewards/rejected": -3.2460455894470215,
"step": 714
},
{
"epoch": 0.46697689607314885,
"grad_norm": 12.956291639260309,
"learning_rate": 9.651024245333177e-08,
"logits/chosen": -1.5934419631958008,
"logits/rejected": -1.581264615058899,
"logps/chosen": -838.2115478515625,
"logps/rejected": -835.7385864257812,
"loss": 0.5465,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9340994358062744,
"rewards/margins": 0.6708505749702454,
"rewards/rejected": -3.604949951171875,
"step": 715
},
{
"epoch": 0.46763001061311127,
"grad_norm": 11.679151198008364,
"learning_rate": 9.634626434715791e-08,
"logits/chosen": -1.5744848251342773,
"logits/rejected": -1.5612497329711914,
"logps/chosen": -834.1600341796875,
"logps/rejected": -887.748779296875,
"loss": 0.5068,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9672343730926514,
"rewards/margins": 0.5058416128158569,
"rewards/rejected": -3.4730756282806396,
"step": 716
},
{
"epoch": 0.46828312515307374,
"grad_norm": 16.13898005368633,
"learning_rate": 9.618217513087857e-08,
"logits/chosen": -1.5531257390975952,
"logits/rejected": -1.5761160850524902,
"logps/chosen": -858.4584350585938,
"logps/rejected": -849.4945678710938,
"loss": 0.5375,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.655836820602417,
"rewards/margins": 0.46926283836364746,
"rewards/rejected": -3.1250996589660645,
"step": 717
},
{
"epoch": 0.46893623969303616,
"grad_norm": 29.758612473106314,
"learning_rate": 9.601797565859966e-08,
"logits/chosen": -1.5403785705566406,
"logits/rejected": -1.5410677194595337,
"logps/chosen": -761.43896484375,
"logps/rejected": -858.6434936523438,
"loss": 0.508,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7271745204925537,
"rewards/margins": 0.8027867078781128,
"rewards/rejected": -3.529961585998535,
"step": 718
},
{
"epoch": 0.46958935423299863,
"grad_norm": 12.15073631227884,
"learning_rate": 9.585366678500099e-08,
"logits/chosen": -1.5667126178741455,
"logits/rejected": -1.5372388362884521,
"logps/chosen": -725.1177368164062,
"logps/rejected": -834.06494140625,
"loss": 0.528,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.6709213256835938,
"rewards/margins": 0.7990758419036865,
"rewards/rejected": -3.469996929168701,
"step": 719
},
{
"epoch": 0.47024246877296105,
"grad_norm": 24.52142611093719,
"learning_rate": 9.568924936533176e-08,
"logits/chosen": -1.5142229795455933,
"logits/rejected": -1.5096334218978882,
"logps/chosen": -788.1256713867188,
"logps/rejected": -853.0916748046875,
"loss": 0.4793,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.0224575996398926,
"rewards/margins": 0.8449983596801758,
"rewards/rejected": -3.8674559593200684,
"step": 720
},
{
"epoch": 0.4708955833129235,
"grad_norm": 12.298496353962463,
"learning_rate": 9.552472425540622e-08,
"logits/chosen": -1.4641120433807373,
"logits/rejected": -1.4405646324157715,
"logps/chosen": -814.8626708984375,
"logps/rejected": -852.576904296875,
"loss": 0.5757,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.7843127250671387,
"rewards/margins": 0.5765048265457153,
"rewards/rejected": -3.3608171939849854,
"step": 721
},
{
"epoch": 0.47154869785288595,
"grad_norm": 9.717229033768323,
"learning_rate": 9.536009231159913e-08,
"logits/chosen": -1.5470542907714844,
"logits/rejected": -1.5365655422210693,
"logps/chosen": -879.9993896484375,
"logps/rejected": -939.4688720703125,
"loss": 0.4599,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.9776620864868164,
"rewards/margins": 0.7803228497505188,
"rewards/rejected": -3.7579846382141113,
"step": 722
},
{
"epoch": 0.4722018123928484,
"grad_norm": 36.798678834078146,
"learning_rate": 9.519535439084134e-08,
"logits/chosen": -1.5727907419204712,
"logits/rejected": -1.5672595500946045,
"logps/chosen": -777.5218505859375,
"logps/rejected": -853.3557739257812,
"loss": 0.5406,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.753126859664917,
"rewards/margins": 0.6771417260169983,
"rewards/rejected": -3.4302685260772705,
"step": 723
},
{
"epoch": 0.47285492693281084,
"grad_norm": 25.32155426640716,
"learning_rate": 9.503051135061538e-08,
"logits/chosen": -1.5253040790557861,
"logits/rejected": -1.5683470964431763,
"logps/chosen": -827.9707641601562,
"logps/rejected": -843.636962890625,
"loss": 0.4839,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8832221031188965,
"rewards/margins": 0.7474449276924133,
"rewards/rejected": -3.630666732788086,
"step": 724
},
{
"epoch": 0.4735080414727733,
"grad_norm": 29.38483596431046,
"learning_rate": 9.486556404895083e-08,
"logits/chosen": -1.4884456396102905,
"logits/rejected": -1.4345768690109253,
"logps/chosen": -738.1644897460938,
"logps/rejected": -887.9599609375,
"loss": 0.5609,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.7885422706604004,
"rewards/margins": 1.1899745464324951,
"rewards/rejected": -3.9785170555114746,
"step": 725
},
{
"epoch": 0.47416115601273573,
"grad_norm": 10.62636396986744,
"learning_rate": 9.470051334442008e-08,
"logits/chosen": -1.477259874343872,
"logits/rejected": -1.5057060718536377,
"logps/chosen": -770.3033447265625,
"logps/rejected": -863.6229248046875,
"loss": 0.5272,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.945749044418335,
"rewards/margins": 0.7120295763015747,
"rewards/rejected": -3.657778739929199,
"step": 726
},
{
"epoch": 0.4748142705526982,
"grad_norm": 17.901405492456256,
"learning_rate": 9.453536009613367e-08,
"logits/chosen": -1.49770987033844,
"logits/rejected": -1.5056265592575073,
"logps/chosen": -741.2904052734375,
"logps/rejected": -932.2896118164062,
"loss": 0.4609,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.7706236839294434,
"rewards/margins": 0.9994168281555176,
"rewards/rejected": -3.770040988922119,
"step": 727
},
{
"epoch": 0.4754673850926606,
"grad_norm": 12.867336823169316,
"learning_rate": 9.437010516373592e-08,
"logits/chosen": -1.4868743419647217,
"logits/rejected": -1.5124531984329224,
"logps/chosen": -870.4682006835938,
"logps/rejected": -907.4706420898438,
"loss": 0.4915,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.0645503997802734,
"rewards/margins": 0.6983194351196289,
"rewards/rejected": -3.7628698348999023,
"step": 728
},
{
"epoch": 0.4761204996326231,
"grad_norm": 50.44249826662269,
"learning_rate": 9.420474940740042e-08,
"logits/chosen": -1.635010838508606,
"logits/rejected": -1.6417289972305298,
"logps/chosen": -834.1351318359375,
"logps/rejected": -948.0263061523438,
"loss": 0.5263,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.062831401824951,
"rewards/margins": 0.9943631887435913,
"rewards/rejected": -4.057194709777832,
"step": 729
},
{
"epoch": 0.4767736141725855,
"grad_norm": 14.64756449278543,
"learning_rate": 9.403929368782558e-08,
"logits/chosen": -1.5567909479141235,
"logits/rejected": -1.6151758432388306,
"logps/chosen": -768.6214599609375,
"logps/rejected": -870.6768798828125,
"loss": 0.4699,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.4706671237945557,
"rewards/margins": 0.9328286647796631,
"rewards/rejected": -3.4034957885742188,
"step": 730
},
{
"epoch": 0.477426728712548,
"grad_norm": 13.947518321539896,
"learning_rate": 9.387373886623012e-08,
"logits/chosen": -1.51140558719635,
"logits/rejected": -1.536016583442688,
"logps/chosen": -860.7777709960938,
"logps/rejected": -902.8041381835938,
"loss": 0.52,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.0753116607666016,
"rewards/margins": 0.5363348722457886,
"rewards/rejected": -3.6116464138031006,
"step": 731
},
{
"epoch": 0.4780798432525104,
"grad_norm": 16.851055770042308,
"learning_rate": 9.37080858043486e-08,
"logits/chosen": -1.4963881969451904,
"logits/rejected": -1.4485998153686523,
"logps/chosen": -813.0571899414062,
"logps/rejected": -845.43408203125,
"loss": 0.519,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9917490482330322,
"rewards/margins": 0.81999272108078,
"rewards/rejected": -3.811741352081299,
"step": 732
},
{
"epoch": 0.4787329577924729,
"grad_norm": 36.178464401201914,
"learning_rate": 9.354233536442691e-08,
"logits/chosen": -1.5164326429367065,
"logits/rejected": -1.5172199010849,
"logps/chosen": -872.368408203125,
"logps/rejected": -948.313720703125,
"loss": 0.5787,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.244874954223633,
"rewards/margins": 0.6580997705459595,
"rewards/rejected": -3.9029746055603027,
"step": 733
},
{
"epoch": 0.4793860723324353,
"grad_norm": 12.484932403976112,
"learning_rate": 9.337648840921784e-08,
"logits/chosen": -1.6157573461532593,
"logits/rejected": -1.5741209983825684,
"logps/chosen": -884.84912109375,
"logps/rejected": -1095.204345703125,
"loss": 0.4598,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.276892900466919,
"rewards/margins": 1.09320068359375,
"rewards/rejected": -4.370093822479248,
"step": 734
},
{
"epoch": 0.48003918687239777,
"grad_norm": 14.089376755304505,
"learning_rate": 9.321054580197656e-08,
"logits/chosen": -1.4722511768341064,
"logits/rejected": -1.474959135055542,
"logps/chosen": -730.8760375976562,
"logps/rejected": -825.123779296875,
"loss": 0.5249,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8380813598632812,
"rewards/margins": 0.6124694347381592,
"rewards/rejected": -3.4505507946014404,
"step": 735
},
{
"epoch": 0.4806923014123602,
"grad_norm": 16.35406356020656,
"learning_rate": 9.304450840645609e-08,
"logits/chosen": -1.4326503276824951,
"logits/rejected": -1.4325727224349976,
"logps/chosen": -839.5531005859375,
"logps/rejected": -894.9701538085938,
"loss": 0.5027,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.4309120178222656,
"rewards/margins": 0.5214735865592957,
"rewards/rejected": -3.952385663986206,
"step": 736
},
{
"epoch": 0.48134541595232266,
"grad_norm": 17.20276393860269,
"learning_rate": 9.287837708690284e-08,
"logits/chosen": -1.5666112899780273,
"logits/rejected": -1.524141550064087,
"logps/chosen": -820.8998413085938,
"logps/rejected": -992.0237426757812,
"loss": 0.4732,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9690146446228027,
"rewards/margins": 0.996252179145813,
"rewards/rejected": -3.9652669429779053,
"step": 737
},
{
"epoch": 0.4819985304922851,
"grad_norm": 43.29010354161253,
"learning_rate": 9.271215270805212e-08,
"logits/chosen": -1.596786379814148,
"logits/rejected": -1.5615063905715942,
"logps/chosen": -804.3712768554688,
"logps/rejected": -849.1760864257812,
"loss": 0.56,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8177123069763184,
"rewards/margins": 0.5579187273979187,
"rewards/rejected": -3.3756308555603027,
"step": 738
},
{
"epoch": 0.48265164503224756,
"grad_norm": 15.346572016217646,
"learning_rate": 9.254583613512365e-08,
"logits/chosen": -1.6470632553100586,
"logits/rejected": -1.6171380281448364,
"logps/chosen": -913.9931640625,
"logps/rejected": -999.4883422851562,
"loss": 0.4807,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2777669429779053,
"rewards/margins": 0.7959187626838684,
"rewards/rejected": -4.073686122894287,
"step": 739
},
{
"epoch": 0.48330475957221,
"grad_norm": 9.556534348064194,
"learning_rate": 9.237942823381696e-08,
"logits/chosen": -1.573829174041748,
"logits/rejected": -1.5064740180969238,
"logps/chosen": -777.1859741210938,
"logps/rejected": -792.814208984375,
"loss": 0.469,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.7544057369232178,
"rewards/margins": 0.8636622428894043,
"rewards/rejected": -3.618068218231201,
"step": 740
},
{
"epoch": 0.48395787411217245,
"grad_norm": 25.011865185818184,
"learning_rate": 9.221292987030702e-08,
"logits/chosen": -1.5379745960235596,
"logits/rejected": -1.4926586151123047,
"logps/chosen": -820.5584716796875,
"logps/rejected": -835.9033203125,
"loss": 0.4647,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9160995483398438,
"rewards/margins": 0.7988985180854797,
"rewards/rejected": -3.7149980068206787,
"step": 741
},
{
"epoch": 0.48461098865213487,
"grad_norm": 13.575787231912816,
"learning_rate": 9.204634191123965e-08,
"logits/chosen": -1.508238673210144,
"logits/rejected": -1.4783196449279785,
"logps/chosen": -902.4708251953125,
"logps/rejected": -893.9943237304688,
"loss": 0.5139,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.170164108276367,
"rewards/margins": 0.5966829061508179,
"rewards/rejected": -3.7668471336364746,
"step": 742
},
{
"epoch": 0.48526410319209734,
"grad_norm": 15.830542176100348,
"learning_rate": 9.187966522372705e-08,
"logits/chosen": -1.534610390663147,
"logits/rejected": -1.5084702968597412,
"logps/chosen": -787.0113525390625,
"logps/rejected": -803.7102661132812,
"loss": 0.5124,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9289166927337646,
"rewards/margins": 0.520931601524353,
"rewards/rejected": -3.4498486518859863,
"step": 743
},
{
"epoch": 0.48591721773205976,
"grad_norm": 21.140461323350923,
"learning_rate": 9.17129006753432e-08,
"logits/chosen": -1.525702714920044,
"logits/rejected": -1.5293265581130981,
"logps/chosen": -725.1962280273438,
"logps/rejected": -809.5682983398438,
"loss": 0.4683,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.3667490482330322,
"rewards/margins": 0.9049954414367676,
"rewards/rejected": -3.271744728088379,
"step": 744
},
{
"epoch": 0.48657033227202223,
"grad_norm": 38.71867984311598,
"learning_rate": 9.154604913411943e-08,
"logits/chosen": -1.5762661695480347,
"logits/rejected": -1.5938408374786377,
"logps/chosen": -876.135986328125,
"logps/rejected": -978.6011962890625,
"loss": 0.4793,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.281604528427124,
"rewards/margins": 0.8068190813064575,
"rewards/rejected": -4.088423252105713,
"step": 745
},
{
"epoch": 0.48722344681198465,
"grad_norm": 29.90445121131915,
"learning_rate": 9.137911146853995e-08,
"logits/chosen": -1.5684157609939575,
"logits/rejected": -1.5826231241226196,
"logps/chosen": -795.3310546875,
"logps/rejected": -863.130859375,
"loss": 0.572,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.2306532859802246,
"rewards/margins": 0.5794581770896912,
"rewards/rejected": -3.8101112842559814,
"step": 746
},
{
"epoch": 0.4878765613519471,
"grad_norm": 20.953402023971254,
"learning_rate": 9.121208854753716e-08,
"logits/chosen": -1.511757493019104,
"logits/rejected": -1.4341886043548584,
"logps/chosen": -716.0580444335938,
"logps/rejected": -796.005859375,
"loss": 0.5049,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9407427310943604,
"rewards/margins": 0.7499470114707947,
"rewards/rejected": -3.6906895637512207,
"step": 747
},
{
"epoch": 0.48852967589190954,
"grad_norm": 23.63758752296789,
"learning_rate": 9.10449812404873e-08,
"logits/chosen": -1.5069888830184937,
"logits/rejected": -1.537886142730713,
"logps/chosen": -781.0473022460938,
"logps/rejected": -866.6488037109375,
"loss": 0.4737,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.995260000228882,
"rewards/margins": 0.538629412651062,
"rewards/rejected": -3.5338892936706543,
"step": 748
},
{
"epoch": 0.489182790431872,
"grad_norm": 9.601439395523968,
"learning_rate": 9.087779041720581e-08,
"logits/chosen": -1.5230836868286133,
"logits/rejected": -1.4928765296936035,
"logps/chosen": -708.7151489257812,
"logps/rejected": -755.310302734375,
"loss": 0.4822,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.70585560798645,
"rewards/margins": 0.5378022789955139,
"rewards/rejected": -3.2436575889587402,
"step": 749
},
{
"epoch": 0.48983590497183443,
"grad_norm": 52.26590360034194,
"learning_rate": 9.071051694794283e-08,
"logits/chosen": -1.546759843826294,
"logits/rejected": -1.4855560064315796,
"logps/chosen": -800.2686767578125,
"logps/rejected": -858.6179809570312,
"loss": 0.5451,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.1868784427642822,
"rewards/margins": 0.29365602135658264,
"rewards/rejected": -3.480534553527832,
"step": 750
},
{
"epoch": 0.4904890195117969,
"grad_norm": 27.46647480470444,
"learning_rate": 9.054316170337872e-08,
"logits/chosen": -1.5644291639328003,
"logits/rejected": -1.5516630411148071,
"logps/chosen": -817.9088745117188,
"logps/rejected": -884.1804809570312,
"loss": 0.4977,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.8426513671875,
"rewards/margins": 0.5670284032821655,
"rewards/rejected": -3.409679412841797,
"step": 751
},
{
"epoch": 0.4911421340517593,
"grad_norm": 22.65417735488772,
"learning_rate": 9.037572555461949e-08,
"logits/chosen": -1.4849984645843506,
"logits/rejected": -1.5145663022994995,
"logps/chosen": -797.359619140625,
"logps/rejected": -906.737548828125,
"loss": 0.4895,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.670722007751465,
"rewards/margins": 0.6933446526527405,
"rewards/rejected": -3.3640666007995605,
"step": 752
},
{
"epoch": 0.4917952485917218,
"grad_norm": 13.411177355069162,
"learning_rate": 9.020820937319222e-08,
"logits/chosen": -1.5731292963027954,
"logits/rejected": -1.5484392642974854,
"logps/chosen": -884.296630859375,
"logps/rejected": -940.8453979492188,
"loss": 0.471,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.245163679122925,
"rewards/margins": 0.7693843245506287,
"rewards/rejected": -4.014548301696777,
"step": 753
},
{
"epoch": 0.4924483631316842,
"grad_norm": 22.100162564155266,
"learning_rate": 9.004061403104063e-08,
"logits/chosen": -1.5291342735290527,
"logits/rejected": -1.541631817817688,
"logps/chosen": -894.0836181640625,
"logps/rejected": -972.8955688476562,
"loss": 0.4913,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.270658016204834,
"rewards/margins": 0.9292981028556824,
"rewards/rejected": -4.19995641708374,
"step": 754
},
{
"epoch": 0.49310147767164664,
"grad_norm": 40.53277025479099,
"learning_rate": 8.987294040052048e-08,
"logits/chosen": -1.513358235359192,
"logits/rejected": -1.4985860586166382,
"logps/chosen": -815.3209838867188,
"logps/rejected": -845.5489501953125,
"loss": 0.5791,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.830094337463379,
"rewards/margins": 0.6605682373046875,
"rewards/rejected": -3.4906625747680664,
"step": 755
},
{
"epoch": 0.4937545922116091,
"grad_norm": 52.36436157875439,
"learning_rate": 8.970518935439494e-08,
"logits/chosen": -1.5312970876693726,
"logits/rejected": -1.536803126335144,
"logps/chosen": -841.9302978515625,
"logps/rejected": -1007.2651977539062,
"loss": 0.52,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.154250383377075,
"rewards/margins": 0.8153612017631531,
"rewards/rejected": -3.9696109294891357,
"step": 756
},
{
"epoch": 0.49440770675157153,
"grad_norm": 10.906915240973083,
"learning_rate": 8.953736176583024e-08,
"logits/chosen": -1.5253336429595947,
"logits/rejected": -1.533002495765686,
"logps/chosen": -844.7198486328125,
"logps/rejected": -904.4848022460938,
"loss": 0.5255,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.983484983444214,
"rewards/margins": 0.8254218101501465,
"rewards/rejected": -3.8089070320129395,
"step": 757
},
{
"epoch": 0.495060821291534,
"grad_norm": 21.62178976119015,
"learning_rate": 8.936945850839103e-08,
"logits/chosen": -1.5330315828323364,
"logits/rejected": -1.5863041877746582,
"logps/chosen": -836.3416137695312,
"logps/rejected": -919.7655029296875,
"loss": 0.5125,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.1585867404937744,
"rewards/margins": 0.9332132935523987,
"rewards/rejected": -4.091800212860107,
"step": 758
},
{
"epoch": 0.4957139358314964,
"grad_norm": 59.346179693747,
"learning_rate": 8.920148045603571e-08,
"logits/chosen": -1.4187573194503784,
"logits/rejected": -1.437163233757019,
"logps/chosen": -740.33837890625,
"logps/rejected": -822.4481811523438,
"loss": 0.4519,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.87211275100708,
"rewards/margins": 0.8624040484428406,
"rewards/rejected": -3.7345166206359863,
"step": 759
},
{
"epoch": 0.4963670503714589,
"grad_norm": 17.104210267847428,
"learning_rate": 8.903342848311213e-08,
"logits/chosen": -1.4946515560150146,
"logits/rejected": -1.511128306388855,
"logps/chosen": -787.0501098632812,
"logps/rejected": -832.3441162109375,
"loss": 0.5147,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0261549949645996,
"rewards/margins": 0.5119923949241638,
"rewards/rejected": -3.538147211074829,
"step": 760
},
{
"epoch": 0.4970201649114213,
"grad_norm": 37.01444427857606,
"learning_rate": 8.886530346435281e-08,
"logits/chosen": -1.5614135265350342,
"logits/rejected": -1.5467019081115723,
"logps/chosen": -809.6796264648438,
"logps/rejected": -850.0182495117188,
"loss": 0.5183,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.892712116241455,
"rewards/margins": 0.5590572357177734,
"rewards/rejected": -3.4517693519592285,
"step": 761
},
{
"epoch": 0.4976732794513838,
"grad_norm": 20.36864614388382,
"learning_rate": 8.869710627487057e-08,
"logits/chosen": -1.5446330308914185,
"logits/rejected": -1.5651966333389282,
"logps/chosen": -765.47412109375,
"logps/rejected": -840.524169921875,
"loss": 0.5331,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1215081214904785,
"rewards/margins": 0.7881425619125366,
"rewards/rejected": -3.9096505641937256,
"step": 762
},
{
"epoch": 0.4983263939913462,
"grad_norm": 88.03942880437029,
"learning_rate": 8.852883779015377e-08,
"logits/chosen": -1.5617398023605347,
"logits/rejected": -1.5589573383331299,
"logps/chosen": -719.5538940429688,
"logps/rejected": -807.90869140625,
"loss": 0.5644,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.8834686279296875,
"rewards/margins": 0.5054572820663452,
"rewards/rejected": -3.3889260292053223,
"step": 763
},
{
"epoch": 0.4989795085313087,
"grad_norm": 85.24504779473465,
"learning_rate": 8.836049888606199e-08,
"logits/chosen": -1.5608245134353638,
"logits/rejected": -1.5477367639541626,
"logps/chosen": -867.1178588867188,
"logps/rejected": -1046.8099365234375,
"loss": 0.4946,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.545870304107666,
"rewards/margins": 1.3517121076583862,
"rewards/rejected": -4.897582530975342,
"step": 764
},
{
"epoch": 0.4996326230712711,
"grad_norm": 103.8676814765994,
"learning_rate": 8.819209043882131e-08,
"logits/chosen": -1.5328567028045654,
"logits/rejected": -1.4881083965301514,
"logps/chosen": -879.6492919921875,
"logps/rejected": -995.6163940429688,
"loss": 0.5008,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1603341102600098,
"rewards/margins": 1.162267804145813,
"rewards/rejected": -4.322602272033691,
"step": 765
},
{
"epoch": 0.5002857376112335,
"grad_norm": 11.728242486539118,
"learning_rate": 8.802361332501978e-08,
"logits/chosen": -1.5301557779312134,
"logits/rejected": -1.5447125434875488,
"logps/chosen": -793.1903076171875,
"logps/rejected": -930.641357421875,
"loss": 0.4803,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.126751184463501,
"rewards/margins": 0.9962693452835083,
"rewards/rejected": -4.123020648956299,
"step": 766
},
{
"epoch": 0.500938852151196,
"grad_norm": 10.020322944474998,
"learning_rate": 8.785506842160285e-08,
"logits/chosen": -1.5191065073013306,
"logits/rejected": -1.5158594846725464,
"logps/chosen": -775.5346069335938,
"logps/rejected": -828.55126953125,
"loss": 0.4989,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.9595863819122314,
"rewards/margins": 0.6543839573860168,
"rewards/rejected": -3.6139702796936035,
"step": 767
},
{
"epoch": 0.5015919666911585,
"grad_norm": 19.19484359451946,
"learning_rate": 8.768645660586886e-08,
"logits/chosen": -1.5571017265319824,
"logits/rejected": -1.599165439605713,
"logps/chosen": -870.1515502929688,
"logps/rejected": -1006.7691650390625,
"loss": 0.4255,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.215618848800659,
"rewards/margins": 1.278153657913208,
"rewards/rejected": -4.493772506713867,
"step": 768
},
{
"epoch": 0.5022450812311209,
"grad_norm": 55.25909288858508,
"learning_rate": 8.751777875546442e-08,
"logits/chosen": -1.5235897302627563,
"logits/rejected": -1.5015480518341064,
"logps/chosen": -766.0463256835938,
"logps/rejected": -827.17919921875,
"loss": 0.5069,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9575929641723633,
"rewards/margins": 0.6264766454696655,
"rewards/rejected": -3.5840699672698975,
"step": 769
},
{
"epoch": 0.5028981957710833,
"grad_norm": 31.7480045035506,
"learning_rate": 8.734903574837985e-08,
"logits/chosen": -1.4985229969024658,
"logits/rejected": -1.4879060983657837,
"logps/chosen": -830.8861083984375,
"logps/rejected": -901.6783447265625,
"loss": 0.5273,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9937520027160645,
"rewards/margins": 0.953312337398529,
"rewards/rejected": -3.9470643997192383,
"step": 770
},
{
"epoch": 0.5035513103110458,
"grad_norm": 23.133647048008825,
"learning_rate": 8.718022846294466e-08,
"logits/chosen": -1.4812209606170654,
"logits/rejected": -1.4926509857177734,
"logps/chosen": -764.713134765625,
"logps/rejected": -867.8782348632812,
"loss": 0.4615,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.96504807472229,
"rewards/margins": 0.7924860715866089,
"rewards/rejected": -3.7575340270996094,
"step": 771
},
{
"epoch": 0.5042044248510082,
"grad_norm": 13.218818964781667,
"learning_rate": 8.701135777782291e-08,
"logits/chosen": -1.5664947032928467,
"logits/rejected": -1.5273160934448242,
"logps/chosen": -843.0235595703125,
"logps/rejected": -841.0691528320312,
"loss": 0.6184,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.2560126781463623,
"rewards/margins": 0.49974071979522705,
"rewards/rejected": -3.755753517150879,
"step": 772
},
{
"epoch": 0.5048575393909707,
"grad_norm": 29.96826439077636,
"learning_rate": 8.684242457200865e-08,
"logits/chosen": -1.4153635501861572,
"logits/rejected": -1.4303746223449707,
"logps/chosen": -740.8873291015625,
"logps/rejected": -893.1712646484375,
"loss": 0.4935,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7135632038116455,
"rewards/margins": 1.0459972620010376,
"rewards/rejected": -3.7595603466033936,
"step": 773
},
{
"epoch": 0.5055106539309331,
"grad_norm": 17.20275448114271,
"learning_rate": 8.667342972482136e-08,
"logits/chosen": -1.5214056968688965,
"logits/rejected": -1.5121796131134033,
"logps/chosen": -887.89794921875,
"logps/rejected": -939.7318115234375,
"loss": 0.5725,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.212088108062744,
"rewards/margins": 0.7043830156326294,
"rewards/rejected": -3.916471481323242,
"step": 774
},
{
"epoch": 0.5061637684708956,
"grad_norm": 44.42583289533232,
"learning_rate": 8.650437411590141e-08,
"logits/chosen": -1.5345828533172607,
"logits/rejected": -1.5515172481536865,
"logps/chosen": -841.9071044921875,
"logps/rejected": -872.3067626953125,
"loss": 0.5004,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2682528495788574,
"rewards/margins": 0.5741406679153442,
"rewards/rejected": -3.842393636703491,
"step": 775
},
{
"epoch": 0.506816883010858,
"grad_norm": 11.651543725419435,
"learning_rate": 8.633525862520538e-08,
"logits/chosen": -1.5557384490966797,
"logits/rejected": -1.5258511304855347,
"logps/chosen": -821.5269165039062,
"logps/rejected": -849.76611328125,
"loss": 0.5619,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.8014473915100098,
"rewards/margins": 0.6341366171836853,
"rewards/rejected": -3.435584306716919,
"step": 776
},
{
"epoch": 0.5074699975508205,
"grad_norm": 46.00900328788502,
"learning_rate": 8.616608413300162e-08,
"logits/chosen": -1.465898871421814,
"logits/rejected": -1.5084877014160156,
"logps/chosen": -825.7061767578125,
"logps/rejected": -899.7260131835938,
"loss": 0.5337,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.140413284301758,
"rewards/margins": 0.8182295560836792,
"rewards/rejected": -3.9586427211761475,
"step": 777
},
{
"epoch": 0.5081231120907829,
"grad_norm": 34.00237370143665,
"learning_rate": 8.599685151986555e-08,
"logits/chosen": -1.5190675258636475,
"logits/rejected": -1.5114282369613647,
"logps/chosen": -788.9437255859375,
"logps/rejected": -859.0122680664062,
"loss": 0.5124,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8215911388397217,
"rewards/margins": 0.813235878944397,
"rewards/rejected": -3.634827136993408,
"step": 778
},
{
"epoch": 0.5087762266307454,
"grad_norm": 11.243459420945996,
"learning_rate": 8.582756166667506e-08,
"logits/chosen": -1.523029088973999,
"logits/rejected": -1.5339922904968262,
"logps/chosen": -776.9263916015625,
"logps/rejected": -869.6576538085938,
"loss": 0.4985,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9392380714416504,
"rewards/margins": 0.5630586743354797,
"rewards/rejected": -3.5022971630096436,
"step": 779
},
{
"epoch": 0.5094293411707078,
"grad_norm": 11.246859304420749,
"learning_rate": 8.565821545460607e-08,
"logits/chosen": -1.5100888013839722,
"logits/rejected": -1.5363471508026123,
"logps/chosen": -822.7084350585938,
"logps/rejected": -870.5260009765625,
"loss": 0.4704,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.991995334625244,
"rewards/margins": 0.8592602610588074,
"rewards/rejected": -3.851255416870117,
"step": 780
},
{
"epoch": 0.5100824557106702,
"grad_norm": 22.710318086337125,
"learning_rate": 8.548881376512784e-08,
"logits/chosen": -1.55587899684906,
"logits/rejected": -1.5447403192520142,
"logps/chosen": -911.83935546875,
"logps/rejected": -1001.2164916992188,
"loss": 0.5482,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.194542407989502,
"rewards/margins": 1.0161552429199219,
"rewards/rejected": -4.210697174072266,
"step": 781
},
{
"epoch": 0.5107355702506327,
"grad_norm": 30.493025252100427,
"learning_rate": 8.531935747999837e-08,
"logits/chosen": -1.5483152866363525,
"logits/rejected": -1.5260441303253174,
"logps/chosen": -825.1983642578125,
"logps/rejected": -968.305419921875,
"loss": 0.4986,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9049713611602783,
"rewards/margins": 0.961846113204956,
"rewards/rejected": -3.8668179512023926,
"step": 782
},
{
"epoch": 0.5113886847905952,
"grad_norm": 11.07008824806609,
"learning_rate": 8.514984748125984e-08,
"logits/chosen": -1.4909794330596924,
"logits/rejected": -1.5248680114746094,
"logps/chosen": -831.9634399414062,
"logps/rejected": -922.552978515625,
"loss": 0.4796,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.975536346435547,
"rewards/margins": 1.0760196447372437,
"rewards/rejected": -4.051555633544922,
"step": 783
},
{
"epoch": 0.5120417993305576,
"grad_norm": 31.858032148945103,
"learning_rate": 8.498028465123402e-08,
"logits/chosen": -1.4506326913833618,
"logits/rejected": -1.4219865798950195,
"logps/chosen": -696.5535278320312,
"logps/rejected": -853.2267456054688,
"loss": 0.4794,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.721505641937256,
"rewards/margins": 1.1351335048675537,
"rewards/rejected": -3.8566393852233887,
"step": 784
},
{
"epoch": 0.51269491387052,
"grad_norm": 67.88266427137975,
"learning_rate": 8.48106698725177e-08,
"logits/chosen": -1.5079140663146973,
"logits/rejected": -1.5019490718841553,
"logps/chosen": -861.70458984375,
"logps/rejected": -1028.7357177734375,
"loss": 0.5223,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1499381065368652,
"rewards/margins": 1.0946592092514038,
"rewards/rejected": -4.244597434997559,
"step": 785
},
{
"epoch": 0.5133480284104824,
"grad_norm": 77.6759127035247,
"learning_rate": 8.464100402797803e-08,
"logits/chosen": -1.4545526504516602,
"logits/rejected": -1.4197547435760498,
"logps/chosen": -805.7166137695312,
"logps/rejected": -865.64501953125,
"loss": 0.5099,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2749907970428467,
"rewards/margins": 0.8093310594558716,
"rewards/rejected": -4.084321975708008,
"step": 786
},
{
"epoch": 0.514001142950445,
"grad_norm": 16.83703843955555,
"learning_rate": 8.4471288000748e-08,
"logits/chosen": -1.5386128425598145,
"logits/rejected": -1.5408875942230225,
"logps/chosen": -842.3564453125,
"logps/rejected": -944.2308959960938,
"loss": 0.5631,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.174443483352661,
"rewards/margins": 0.8851109147071838,
"rewards/rejected": -4.059554576873779,
"step": 787
},
{
"epoch": 0.5146542574904074,
"grad_norm": 31.817856468339304,
"learning_rate": 8.430152267422177e-08,
"logits/chosen": -1.5486061573028564,
"logits/rejected": -1.53328537940979,
"logps/chosen": -797.2845458984375,
"logps/rejected": -912.8262329101562,
"loss": 0.5719,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.162616729736328,
"rewards/margins": 0.9437724947929382,
"rewards/rejected": -4.106389045715332,
"step": 788
},
{
"epoch": 0.5153073720303698,
"grad_norm": 32.1546133597687,
"learning_rate": 8.413170893205015e-08,
"logits/chosen": -1.5652754306793213,
"logits/rejected": -1.5659431219100952,
"logps/chosen": -796.43994140625,
"logps/rejected": -850.5557861328125,
"loss": 0.5232,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.044846773147583,
"rewards/margins": 0.7479317784309387,
"rewards/rejected": -3.792778730392456,
"step": 789
},
{
"epoch": 0.5159604865703322,
"grad_norm": 11.020829449688339,
"learning_rate": 8.396184765813591e-08,
"logits/chosen": -1.5715382099151611,
"logits/rejected": -1.5400923490524292,
"logps/chosen": -811.490234375,
"logps/rejected": -932.505615234375,
"loss": 0.5092,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.054403781890869,
"rewards/margins": 1.0824800729751587,
"rewards/rejected": -4.136883735656738,
"step": 790
},
{
"epoch": 0.5166136011102948,
"grad_norm": 30.888894963547546,
"learning_rate": 8.379193973662927e-08,
"logits/chosen": -1.5172935724258423,
"logits/rejected": -1.5318307876586914,
"logps/chosen": -823.394287109375,
"logps/rejected": -943.768310546875,
"loss": 0.498,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9745614528656006,
"rewards/margins": 1.0555309057235718,
"rewards/rejected": -4.030092716217041,
"step": 791
},
{
"epoch": 0.5172667156502572,
"grad_norm": 29.8842541290539,
"learning_rate": 8.362198605192326e-08,
"logits/chosen": -1.557112693786621,
"logits/rejected": -1.5465399026870728,
"logps/chosen": -826.0409545898438,
"logps/rejected": -961.9547119140625,
"loss": 0.4962,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1106011867523193,
"rewards/margins": 0.760248601436615,
"rewards/rejected": -3.870850086212158,
"step": 792
},
{
"epoch": 0.5179198301902196,
"grad_norm": 25.848000115396033,
"learning_rate": 8.345198748864909e-08,
"logits/chosen": -1.4976153373718262,
"logits/rejected": -1.4590017795562744,
"logps/chosen": -834.5736083984375,
"logps/rejected": -815.7548217773438,
"loss": 0.5151,
"rewards/accuracies": 0.5625,
"rewards/chosen": -3.545506477355957,
"rewards/margins": 0.19579878449440002,
"rewards/rejected": -3.741305351257324,
"step": 793
},
{
"epoch": 0.518572944730182,
"grad_norm": 22.608186681230375,
"learning_rate": 8.328194493167156e-08,
"logits/chosen": -1.5127606391906738,
"logits/rejected": -1.5453428030014038,
"logps/chosen": -874.2019653320312,
"logps/rejected": -923.8192749023438,
"loss": 0.5177,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3078603744506836,
"rewards/margins": 0.7927409410476685,
"rewards/rejected": -4.1006011962890625,
"step": 794
},
{
"epoch": 0.5192260592701445,
"grad_norm": 14.489367562787846,
"learning_rate": 8.311185926608451e-08,
"logits/chosen": -1.50680673122406,
"logits/rejected": -1.484675645828247,
"logps/chosen": -844.3919067382812,
"logps/rejected": -819.8763427734375,
"loss": 0.5177,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.12322998046875,
"rewards/margins": 0.5924826860427856,
"rewards/rejected": -3.715712785720825,
"step": 795
},
{
"epoch": 0.519879173810107,
"grad_norm": 11.919265518405389,
"learning_rate": 8.29417313772061e-08,
"logits/chosen": -1.5627418756484985,
"logits/rejected": -1.580378770828247,
"logps/chosen": -855.9948120117188,
"logps/rejected": -1031.594970703125,
"loss": 0.5956,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3382222652435303,
"rewards/margins": 0.9997389912605286,
"rewards/rejected": -4.337961673736572,
"step": 796
},
{
"epoch": 0.5205322883500694,
"grad_norm": 28.37261621955881,
"learning_rate": 8.277156215057434e-08,
"logits/chosen": -1.4434901475906372,
"logits/rejected": -1.4327919483184814,
"logps/chosen": -895.87646484375,
"logps/rejected": -958.2759399414062,
"loss": 0.4678,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.354534864425659,
"rewards/margins": 0.7718441486358643,
"rewards/rejected": -4.126379013061523,
"step": 797
},
{
"epoch": 0.5211854028900318,
"grad_norm": 22.35879028875356,
"learning_rate": 8.260135247194235e-08,
"logits/chosen": -1.539190411567688,
"logits/rejected": -1.5733530521392822,
"logps/chosen": -710.5164794921875,
"logps/rejected": -795.2035522460938,
"loss": 0.4548,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.8853349685668945,
"rewards/margins": 0.540171205997467,
"rewards/rejected": -3.425506353378296,
"step": 798
},
{
"epoch": 0.5218385174299943,
"grad_norm": 10.507592796277946,
"learning_rate": 8.243110322727382e-08,
"logits/chosen": -1.5344185829162598,
"logits/rejected": -1.5373344421386719,
"logps/chosen": -755.857177734375,
"logps/rejected": -858.6672973632812,
"loss": 0.4848,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8476948738098145,
"rewards/margins": 0.9375979900360107,
"rewards/rejected": -3.785292863845825,
"step": 799
},
{
"epoch": 0.5224916319699567,
"grad_norm": 33.61581825469172,
"learning_rate": 8.226081530273843e-08,
"logits/chosen": -1.557815670967102,
"logits/rejected": -1.561926007270813,
"logps/chosen": -820.7285766601562,
"logps/rejected": -851.3828125,
"loss": 0.511,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0852344036102295,
"rewards/margins": 0.7549958825111389,
"rewards/rejected": -3.8402299880981445,
"step": 800
},
{
"epoch": 0.5224916319699567,
"eval_logits/chosen": -1.527337670326233,
"eval_logits/rejected": -1.5160092115402222,
"eval_logps/chosen": -815.5145263671875,
"eval_logps/rejected": -886.17138671875,
"eval_loss": 0.5177087187767029,
"eval_rewards/accuracies": 0.7400000095367432,
"eval_rewards/chosen": -3.0516912937164307,
"eval_rewards/margins": 0.7876277565956116,
"eval_rewards/rejected": -3.8393189907073975,
"eval_runtime": 296.7047,
"eval_samples_per_second": 13.481,
"eval_steps_per_second": 0.843,
"step": 800
},
{
"epoch": 0.5231447465099192,
"grad_norm": 29.72721689887477,
"learning_rate": 8.209048958470714e-08,
"logits/chosen": -1.5397412776947021,
"logits/rejected": -1.573317050933838,
"logps/chosen": -794.4456176757812,
"logps/rejected": -861.4932861328125,
"loss": 0.4681,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1979122161865234,
"rewards/margins": 0.44129082560539246,
"rewards/rejected": -3.6392035484313965,
"step": 801
},
{
"epoch": 0.5237978610498816,
"grad_norm": 24.93121116287881,
"learning_rate": 8.192012695974765e-08,
"logits/chosen": -1.440004587173462,
"logits/rejected": -1.4455904960632324,
"logps/chosen": -756.099365234375,
"logps/rejected": -812.0906372070312,
"loss": 0.4259,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.7917063236236572,
"rewards/margins": 0.8878428936004639,
"rewards/rejected": -3.6795494556427,
"step": 802
},
{
"epoch": 0.5244509755898441,
"grad_norm": 21.486488987183108,
"learning_rate": 8.174972831461975e-08,
"logits/chosen": -1.4720410108566284,
"logits/rejected": -1.4299757480621338,
"logps/chosen": -797.6776733398438,
"logps/rejected": -876.0466918945312,
"loss": 0.5424,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.303070545196533,
"rewards/margins": 0.6923160552978516,
"rewards/rejected": -3.9953863620758057,
"step": 803
},
{
"epoch": 0.5251040901298065,
"grad_norm": 15.956985029082132,
"learning_rate": 8.157929453627079e-08,
"logits/chosen": -1.528617262840271,
"logits/rejected": -1.543229579925537,
"logps/chosen": -814.8829345703125,
"logps/rejected": -840.0612182617188,
"loss": 0.5554,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0760016441345215,
"rewards/margins": 0.729838490486145,
"rewards/rejected": -3.805840253829956,
"step": 804
},
{
"epoch": 0.5257572046697689,
"grad_norm": 14.955725644267702,
"learning_rate": 8.140882651183087e-08,
"logits/chosen": -1.4945454597473145,
"logits/rejected": -1.50547194480896,
"logps/chosen": -847.5591430664062,
"logps/rejected": -860.2355346679688,
"loss": 0.5914,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3800101280212402,
"rewards/margins": 0.4141124486923218,
"rewards/rejected": -3.7941229343414307,
"step": 805
},
{
"epoch": 0.5264103192097314,
"grad_norm": 22.99418263109128,
"learning_rate": 8.123832512860848e-08,
"logits/chosen": -1.5119673013687134,
"logits/rejected": -1.5116000175476074,
"logps/chosen": -837.5704956054688,
"logps/rejected": -927.5317993164062,
"loss": 0.5061,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.0885889530181885,
"rewards/margins": 0.9447399377822876,
"rewards/rejected": -4.033329010009766,
"step": 806
},
{
"epoch": 0.5270634337496939,
"grad_norm": 18.701086163571972,
"learning_rate": 8.106779127408563e-08,
"logits/chosen": -1.4224164485931396,
"logits/rejected": -1.4035032987594604,
"logps/chosen": -733.4963989257812,
"logps/rejected": -846.0479736328125,
"loss": 0.5246,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.811711072921753,
"rewards/margins": 0.7736900448799133,
"rewards/rejected": -3.5854010581970215,
"step": 807
},
{
"epoch": 0.5277165482896563,
"grad_norm": 40.22777398303923,
"learning_rate": 8.08972258359134e-08,
"logits/chosen": -1.497331142425537,
"logits/rejected": -1.5132163763046265,
"logps/chosen": -791.0665893554688,
"logps/rejected": -848.4359130859375,
"loss": 0.5209,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.804387092590332,
"rewards/margins": 0.7790587544441223,
"rewards/rejected": -3.5834455490112305,
"step": 808
},
{
"epoch": 0.5283696628296187,
"grad_norm": 28.929510873590274,
"learning_rate": 8.07266297019073e-08,
"logits/chosen": -1.5387717485427856,
"logits/rejected": -1.5651496648788452,
"logps/chosen": -821.0679321289062,
"logps/rejected": -911.5535278320312,
"loss": 0.4894,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0287632942199707,
"rewards/margins": 0.7187597155570984,
"rewards/rejected": -3.747523307800293,
"step": 809
},
{
"epoch": 0.5290227773695811,
"grad_norm": 45.22369470855717,
"learning_rate": 8.055600376004255e-08,
"logits/chosen": -1.5283149480819702,
"logits/rejected": -1.520777940750122,
"logps/chosen": -786.1401977539062,
"logps/rejected": -925.03125,
"loss": 0.5102,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.684399366378784,
"rewards/margins": 1.1089178323745728,
"rewards/rejected": -3.7933173179626465,
"step": 810
},
{
"epoch": 0.5296758919095437,
"grad_norm": 42.19326253786495,
"learning_rate": 8.038534889844956e-08,
"logits/chosen": -1.6188210248947144,
"logits/rejected": -1.5729517936706543,
"logps/chosen": -868.0130004882812,
"logps/rejected": -917.4168090820312,
"loss": 0.5735,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.09293270111084,
"rewards/margins": 1.0333658456802368,
"rewards/rejected": -4.126298904418945,
"step": 811
},
{
"epoch": 0.5303290064495061,
"grad_norm": 117.37707476367844,
"learning_rate": 8.021466600540928e-08,
"logits/chosen": -1.5684051513671875,
"logits/rejected": -1.5472626686096191,
"logps/chosen": -852.613525390625,
"logps/rejected": -913.04833984375,
"loss": 0.5436,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4583096504211426,
"rewards/margins": 0.717620313167572,
"rewards/rejected": -4.175930023193359,
"step": 812
},
{
"epoch": 0.5309821209894685,
"grad_norm": 28.12903118620616,
"learning_rate": 8.004395596934856e-08,
"logits/chosen": -1.4949054718017578,
"logits/rejected": -1.4636915922164917,
"logps/chosen": -733.1012573242188,
"logps/rejected": -767.9586181640625,
"loss": 0.5379,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7585432529449463,
"rewards/margins": 0.46337318420410156,
"rewards/rejected": -3.221916675567627,
"step": 813
},
{
"epoch": 0.5316352355294309,
"grad_norm": 9.221461814471073,
"learning_rate": 7.987321967883549e-08,
"logits/chosen": -1.6841922998428345,
"logits/rejected": -1.6034355163574219,
"logps/chosen": -839.085205078125,
"logps/rejected": -876.251953125,
"loss": 0.5158,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.193796396255493,
"rewards/margins": 0.6729452610015869,
"rewards/rejected": -3.866741418838501,
"step": 814
},
{
"epoch": 0.5322883500693935,
"grad_norm": 10.618048036543714,
"learning_rate": 7.970245802257487e-08,
"logits/chosen": -1.6014914512634277,
"logits/rejected": -1.5899266004562378,
"logps/chosen": -847.2337646484375,
"logps/rejected": -908.861572265625,
"loss": 0.5192,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.173642158508301,
"rewards/margins": 0.662325382232666,
"rewards/rejected": -3.835967540740967,
"step": 815
},
{
"epoch": 0.5329414646093559,
"grad_norm": 24.838517311372605,
"learning_rate": 7.953167188940353e-08,
"logits/chosen": -1.531007170677185,
"logits/rejected": -1.5374797582626343,
"logps/chosen": -871.73291015625,
"logps/rejected": -1004.2645874023438,
"loss": 0.4973,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.170253276824951,
"rewards/margins": 0.9588162899017334,
"rewards/rejected": -4.1290693283081055,
"step": 816
},
{
"epoch": 0.5335945791493183,
"grad_norm": 101.58547120360855,
"learning_rate": 7.936086216828568e-08,
"logits/chosen": -1.5855270624160767,
"logits/rejected": -1.600976586341858,
"logps/chosen": -784.5303955078125,
"logps/rejected": -773.5015258789062,
"loss": 0.4903,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.953719139099121,
"rewards/margins": 0.48683837056159973,
"rewards/rejected": -3.4405574798583984,
"step": 817
},
{
"epoch": 0.5342476936892807,
"grad_norm": 39.27500759821802,
"learning_rate": 7.919002974830833e-08,
"logits/chosen": -1.577518105506897,
"logits/rejected": -1.5548105239868164,
"logps/chosen": -707.9842529296875,
"logps/rejected": -847.3580932617188,
"loss": 0.4995,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.563530921936035,
"rewards/margins": 0.9225165843963623,
"rewards/rejected": -3.4860472679138184,
"step": 818
},
{
"epoch": 0.5349008082292432,
"grad_norm": 51.172477176857036,
"learning_rate": 7.901917551867663e-08,
"logits/chosen": -1.5167884826660156,
"logits/rejected": -1.4933404922485352,
"logps/chosen": -764.8781127929688,
"logps/rejected": -829.58935546875,
"loss": 0.4601,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.603954792022705,
"rewards/margins": 0.4926825761795044,
"rewards/rejected": -3.09663724899292,
"step": 819
},
{
"epoch": 0.5355539227692057,
"grad_norm": 30.46290960655871,
"learning_rate": 7.884830036870922e-08,
"logits/chosen": -1.4870680570602417,
"logits/rejected": -1.4628260135650635,
"logps/chosen": -821.735595703125,
"logps/rejected": -956.1165771484375,
"loss": 0.484,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.0134003162384033,
"rewards/margins": 1.0060707330703735,
"rewards/rejected": -4.019470691680908,
"step": 820
},
{
"epoch": 0.5362070373091681,
"grad_norm": 13.505022242348229,
"learning_rate": 7.867740518783371e-08,
"logits/chosen": -1.5747795104980469,
"logits/rejected": -1.5648517608642578,
"logps/chosen": -851.3766479492188,
"logps/rejected": -903.6992797851562,
"loss": 0.5012,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9860377311706543,
"rewards/margins": 0.8135138750076294,
"rewards/rejected": -3.799551486968994,
"step": 821
},
{
"epoch": 0.5368601518491305,
"grad_norm": 19.48923466978227,
"learning_rate": 7.85064908655819e-08,
"logits/chosen": -1.4971075057983398,
"logits/rejected": -1.548054814338684,
"logps/chosen": -855.16748046875,
"logps/rejected": -904.4014282226562,
"loss": 0.4958,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.101994514465332,
"rewards/margins": 0.8241324424743652,
"rewards/rejected": -3.926126480102539,
"step": 822
},
{
"epoch": 0.537513266389093,
"grad_norm": 24.801244541119452,
"learning_rate": 7.833555829158527e-08,
"logits/chosen": -1.5962388515472412,
"logits/rejected": -1.5477280616760254,
"logps/chosen": -887.2816772460938,
"logps/rejected": -913.8333129882812,
"loss": 0.4767,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1752548217773438,
"rewards/margins": 0.6703548431396484,
"rewards/rejected": -3.845609664916992,
"step": 823
},
{
"epoch": 0.5381663809290554,
"grad_norm": 42.45784716129404,
"learning_rate": 7.816460835557028e-08,
"logits/chosen": -1.5705194473266602,
"logits/rejected": -1.5918525457382202,
"logps/chosen": -808.6838989257812,
"logps/rejected": -973.3630981445312,
"loss": 0.4975,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.012310743331909,
"rewards/margins": 1.2067666053771973,
"rewards/rejected": -4.2190775871276855,
"step": 824
},
{
"epoch": 0.5388194954690179,
"grad_norm": 18.2583142847124,
"learning_rate": 7.799364194735377e-08,
"logits/chosen": -1.5670945644378662,
"logits/rejected": -1.5296969413757324,
"logps/chosen": -861.2294311523438,
"logps/rejected": -866.001953125,
"loss": 0.5391,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.31349515914917,
"rewards/margins": 0.387903094291687,
"rewards/rejected": -3.7013983726501465,
"step": 825
},
{
"epoch": 0.5394726100089803,
"grad_norm": 14.894630917984315,
"learning_rate": 7.782265995683828e-08,
"logits/chosen": -1.539316177368164,
"logits/rejected": -1.5315918922424316,
"logps/chosen": -808.724609375,
"logps/rejected": -842.3665771484375,
"loss": 0.4964,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.908222198486328,
"rewards/margins": 0.6761490106582642,
"rewards/rejected": -3.584371328353882,
"step": 826
},
{
"epoch": 0.5401257245489428,
"grad_norm": 46.701075509344925,
"learning_rate": 7.765166327400754e-08,
"logits/chosen": -1.5789310932159424,
"logits/rejected": -1.552915334701538,
"logps/chosen": -835.5314331054688,
"logps/rejected": -890.4103393554688,
"loss": 0.4631,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.1613926887512207,
"rewards/margins": 0.716499924659729,
"rewards/rejected": -3.8778927326202393,
"step": 827
},
{
"epoch": 0.5407788390889052,
"grad_norm": 23.68366708316572,
"learning_rate": 7.748065278892171e-08,
"logits/chosen": -1.5369728803634644,
"logits/rejected": -1.5010067224502563,
"logps/chosen": -812.9715576171875,
"logps/rejected": -846.779052734375,
"loss": 0.5664,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.117812156677246,
"rewards/margins": 0.5145601630210876,
"rewards/rejected": -3.6323723793029785,
"step": 828
},
{
"epoch": 0.5414319536288676,
"grad_norm": 9.80288754353676,
"learning_rate": 7.730962939171278e-08,
"logits/chosen": -1.5014228820800781,
"logits/rejected": -1.503250241279602,
"logps/chosen": -921.4869384765625,
"logps/rejected": -941.43896484375,
"loss": 0.4775,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4559803009033203,
"rewards/margins": 0.8563462495803833,
"rewards/rejected": -4.312326431274414,
"step": 829
},
{
"epoch": 0.5420850681688301,
"grad_norm": 23.08260983849523,
"learning_rate": 7.713859397257995e-08,
"logits/chosen": -1.5203160047531128,
"logits/rejected": -1.4632571935653687,
"logps/chosen": -813.74267578125,
"logps/rejected": -845.8812255859375,
"loss": 0.5687,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1719284057617188,
"rewards/margins": 0.44786882400512695,
"rewards/rejected": -3.6197972297668457,
"step": 830
},
{
"epoch": 0.5427381827087926,
"grad_norm": 86.00587707569419,
"learning_rate": 7.696754742178503e-08,
"logits/chosen": -1.5021620988845825,
"logits/rejected": -1.5242396593093872,
"logps/chosen": -817.4739990234375,
"logps/rejected": -865.0052490234375,
"loss": 0.5095,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8962783813476562,
"rewards/margins": 0.689846932888031,
"rewards/rejected": -3.586124897003174,
"step": 831
},
{
"epoch": 0.543391297248755,
"grad_norm": 52.62733320349321,
"learning_rate": 7.679649062964774e-08,
"logits/chosen": -1.567655324935913,
"logits/rejected": -1.5549966096878052,
"logps/chosen": -735.2625732421875,
"logps/rejected": -834.2606811523438,
"loss": 0.459,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.9052071571350098,
"rewards/margins": 0.9503459930419922,
"rewards/rejected": -3.85555362701416,
"step": 832
},
{
"epoch": 0.5440444117887174,
"grad_norm": 34.035780541177466,
"learning_rate": 7.662542448654109e-08,
"logits/chosen": -1.5708342790603638,
"logits/rejected": -1.511175274848938,
"logps/chosen": -821.263671875,
"logps/rejected": -975.2786254882812,
"loss": 0.5008,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.831267833709717,
"rewards/margins": 1.1821346282958984,
"rewards/rejected": -4.013402462005615,
"step": 833
},
{
"epoch": 0.5446975263286798,
"grad_norm": 12.816837148569553,
"learning_rate": 7.645434988288683e-08,
"logits/chosen": -1.5968691110610962,
"logits/rejected": -1.5883498191833496,
"logps/chosen": -778.30712890625,
"logps/rejected": -840.7581176757812,
"loss": 0.5001,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8528571128845215,
"rewards/margins": 0.7486177086830139,
"rewards/rejected": -3.6014747619628906,
"step": 834
},
{
"epoch": 0.5453506408686424,
"grad_norm": 22.221139309815047,
"learning_rate": 7.628326770915069e-08,
"logits/chosen": -1.5044746398925781,
"logits/rejected": -1.4766840934753418,
"logps/chosen": -839.3434448242188,
"logps/rejected": -888.7236328125,
"loss": 0.4587,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.229734182357788,
"rewards/margins": 0.8066515922546387,
"rewards/rejected": -4.036386013031006,
"step": 835
},
{
"epoch": 0.5460037554086048,
"grad_norm": 21.464697851921866,
"learning_rate": 7.611217885583783e-08,
"logits/chosen": -1.6017417907714844,
"logits/rejected": -1.6247013807296753,
"logps/chosen": -887.736328125,
"logps/rejected": -987.9991455078125,
"loss": 0.5011,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1084957122802734,
"rewards/margins": 0.854011595249176,
"rewards/rejected": -3.9625072479248047,
"step": 836
},
{
"epoch": 0.5466568699485672,
"grad_norm": 39.79039092333048,
"learning_rate": 7.594108421348816e-08,
"logits/chosen": -1.4360780715942383,
"logits/rejected": -1.4754202365875244,
"logps/chosen": -759.8140258789062,
"logps/rejected": -813.8535766601562,
"loss": 0.5061,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8158435821533203,
"rewards/margins": 0.8641850352287292,
"rewards/rejected": -3.6800286769866943,
"step": 837
},
{
"epoch": 0.5473099844885296,
"grad_norm": 55.549125631484166,
"learning_rate": 7.576998467267174e-08,
"logits/chosen": -1.56089448928833,
"logits/rejected": -1.581824541091919,
"logps/chosen": -909.1053466796875,
"logps/rejected": -948.4317016601562,
"loss": 0.5306,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.213887929916382,
"rewards/margins": 0.46870189905166626,
"rewards/rejected": -3.6825897693634033,
"step": 838
},
{
"epoch": 0.5479630990284922,
"grad_norm": 55.14261568208239,
"learning_rate": 7.559888112398411e-08,
"logits/chosen": -1.4533392190933228,
"logits/rejected": -1.4627681970596313,
"logps/chosen": -726.8400268554688,
"logps/rejected": -814.5458374023438,
"loss": 0.5163,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.764704942703247,
"rewards/margins": 0.9564896821975708,
"rewards/rejected": -3.7211947441101074,
"step": 839
},
{
"epoch": 0.5486162135684546,
"grad_norm": 30.505010236460816,
"learning_rate": 7.542777445804171e-08,
"logits/chosen": -1.3663051128387451,
"logits/rejected": -1.3810725212097168,
"logps/chosen": -795.670654296875,
"logps/rejected": -962.4092407226562,
"loss": 0.4755,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.0239500999450684,
"rewards/margins": 1.2967259883880615,
"rewards/rejected": -4.320675849914551,
"step": 840
},
{
"epoch": 0.549269328108417,
"grad_norm": 14.300839480022828,
"learning_rate": 7.525666556547714e-08,
"logits/chosen": -1.6288928985595703,
"logits/rejected": -1.5527688264846802,
"logps/chosen": -799.6822509765625,
"logps/rejected": -940.3196411132812,
"loss": 0.494,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9739322662353516,
"rewards/margins": 0.9197661876678467,
"rewards/rejected": -3.893698215484619,
"step": 841
},
{
"epoch": 0.5499224426483794,
"grad_norm": 88.7947159228681,
"learning_rate": 7.508555533693462e-08,
"logits/chosen": -1.4860543012619019,
"logits/rejected": -1.4642175436019897,
"logps/chosen": -732.2152099609375,
"logps/rejected": -754.6222534179688,
"loss": 0.4724,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6171233654022217,
"rewards/margins": 0.7891316413879395,
"rewards/rejected": -3.4062552452087402,
"step": 842
},
{
"epoch": 0.550575557188342,
"grad_norm": 69.35311679506285,
"learning_rate": 7.49144446630654e-08,
"logits/chosen": -1.4915255308151245,
"logits/rejected": -1.4560573101043701,
"logps/chosen": -874.8602294921875,
"logps/rejected": -909.4222412109375,
"loss": 0.5181,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3296079635620117,
"rewards/margins": 0.6430115699768066,
"rewards/rejected": -3.9726200103759766,
"step": 843
},
{
"epoch": 0.5512286717283044,
"grad_norm": 14.122314075620803,
"learning_rate": 7.474333443452289e-08,
"logits/chosen": -1.5023688077926636,
"logits/rejected": -1.4814002513885498,
"logps/chosen": -812.0813598632812,
"logps/rejected": -892.1846923828125,
"loss": 0.4773,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.119580030441284,
"rewards/margins": 1.089004397392273,
"rewards/rejected": -4.208584785461426,
"step": 844
},
{
"epoch": 0.5518817862682668,
"grad_norm": 22.523095660476606,
"learning_rate": 7.45722255419583e-08,
"logits/chosen": -1.598945140838623,
"logits/rejected": -1.5560745000839233,
"logps/chosen": -848.6195678710938,
"logps/rejected": -891.3744506835938,
"loss": 0.5091,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.038963556289673,
"rewards/margins": 0.6308895349502563,
"rewards/rejected": -3.6698527336120605,
"step": 845
},
{
"epoch": 0.5525349008082292,
"grad_norm": 37.5554038949777,
"learning_rate": 7.44011188760159e-08,
"logits/chosen": -1.593327283859253,
"logits/rejected": -1.6003532409667969,
"logps/chosen": -984.0642700195312,
"logps/rejected": -945.6554565429688,
"loss": 0.5136,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.7509801387786865,
"rewards/margins": 0.44811174273490906,
"rewards/rejected": -4.199091911315918,
"step": 846
},
{
"epoch": 0.5531880153481917,
"grad_norm": 16.920523902644987,
"learning_rate": 7.423001532732826e-08,
"logits/chosen": -1.5089354515075684,
"logits/rejected": -1.468766212463379,
"logps/chosen": -713.7976684570312,
"logps/rejected": -817.9136962890625,
"loss": 0.5476,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9451818466186523,
"rewards/margins": 1.023890495300293,
"rewards/rejected": -3.9690723419189453,
"step": 847
},
{
"epoch": 0.5538411298881541,
"grad_norm": 11.301729003307928,
"learning_rate": 7.405891578651185e-08,
"logits/chosen": -1.5803592205047607,
"logits/rejected": -1.5427913665771484,
"logps/chosen": -875.8158569335938,
"logps/rejected": -872.2986450195312,
"loss": 0.5062,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.213369369506836,
"rewards/margins": 0.4956037104129791,
"rewards/rejected": -3.708972930908203,
"step": 848
},
{
"epoch": 0.5544942444281166,
"grad_norm": 74.39833754806011,
"learning_rate": 7.388782114416217e-08,
"logits/chosen": -1.6001735925674438,
"logits/rejected": -1.5752531290054321,
"logps/chosen": -795.3203125,
"logps/rejected": -860.407470703125,
"loss": 0.5166,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.002932548522949,
"rewards/margins": 0.9377602934837341,
"rewards/rejected": -3.9406933784484863,
"step": 849
},
{
"epoch": 0.555147358968079,
"grad_norm": 19.66725027995187,
"learning_rate": 7.371673229084931e-08,
"logits/chosen": -1.5427309274673462,
"logits/rejected": -1.5140000581741333,
"logps/chosen": -799.3212890625,
"logps/rejected": -892.1961669921875,
"loss": 0.5065,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.130077838897705,
"rewards/margins": 1.1122186183929443,
"rewards/rejected": -4.24229621887207,
"step": 850
},
{
"epoch": 0.5558004735080415,
"grad_norm": 40.531931496184235,
"learning_rate": 7.354565011711317e-08,
"logits/chosen": -1.4393893480300903,
"logits/rejected": -1.425200343132019,
"logps/chosen": -727.331787109375,
"logps/rejected": -858.167236328125,
"loss": 0.4825,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.9096834659576416,
"rewards/margins": 1.041924238204956,
"rewards/rejected": -3.9516077041625977,
"step": 851
},
{
"epoch": 0.5564535880480039,
"grad_norm": 16.984478852782892,
"learning_rate": 7.33745755134589e-08,
"logits/chosen": -1.5062000751495361,
"logits/rejected": -1.53849458694458,
"logps/chosen": -802.3837890625,
"logps/rejected": -890.3169555664062,
"loss": 0.5344,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2534284591674805,
"rewards/margins": 0.8640724420547485,
"rewards/rejected": -4.1175007820129395,
"step": 852
},
{
"epoch": 0.5571067025879664,
"grad_norm": 35.84235839938917,
"learning_rate": 7.320350937035228e-08,
"logits/chosen": -1.5342974662780762,
"logits/rejected": -1.5023765563964844,
"logps/chosen": -835.8595581054688,
"logps/rejected": -865.0036010742188,
"loss": 0.5255,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0694611072540283,
"rewards/margins": 0.7715376615524292,
"rewards/rejected": -3.840998411178589,
"step": 853
},
{
"epoch": 0.5577598171279288,
"grad_norm": 10.233512310662372,
"learning_rate": 7.303245257821498e-08,
"logits/chosen": -1.517195701599121,
"logits/rejected": -1.518413782119751,
"logps/chosen": -832.9429931640625,
"logps/rejected": -876.033447265625,
"loss": 0.5145,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8127410411834717,
"rewards/margins": 0.624780535697937,
"rewards/rejected": -3.437521457672119,
"step": 854
},
{
"epoch": 0.5584129316678913,
"grad_norm": 26.56768549861038,
"learning_rate": 7.286140602742005e-08,
"logits/chosen": -1.4718785285949707,
"logits/rejected": -1.4336018562316895,
"logps/chosen": -897.831298828125,
"logps/rejected": -885.6130981445312,
"loss": 0.5896,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3304998874664307,
"rewards/margins": 0.5319950580596924,
"rewards/rejected": -3.862494945526123,
"step": 855
},
{
"epoch": 0.5590660462078537,
"grad_norm": 184.22901412620044,
"learning_rate": 7.269037060828724e-08,
"logits/chosen": -1.501589298248291,
"logits/rejected": -1.5195332765579224,
"logps/chosen": -806.921875,
"logps/rejected": -836.3479614257812,
"loss": 0.5501,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9807791709899902,
"rewards/margins": 0.5941177606582642,
"rewards/rejected": -3.574897050857544,
"step": 856
},
{
"epoch": 0.5597191607478161,
"grad_norm": 67.4144741062901,
"learning_rate": 7.25193472110783e-08,
"logits/chosen": -1.5734254121780396,
"logits/rejected": -1.5101865530014038,
"logps/chosen": -823.78759765625,
"logps/rejected": -862.5943603515625,
"loss": 0.4974,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.210695266723633,
"rewards/margins": 0.569321870803833,
"rewards/rejected": -3.780017375946045,
"step": 857
},
{
"epoch": 0.5603722752877786,
"grad_norm": 18.871171186142504,
"learning_rate": 7.234833672599245e-08,
"logits/chosen": -1.580259919166565,
"logits/rejected": -1.5304930210113525,
"logps/chosen": -937.5279541015625,
"logps/rejected": -984.5499267578125,
"loss": 0.5273,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.529520034790039,
"rewards/margins": 0.9078921675682068,
"rewards/rejected": -4.437412261962891,
"step": 858
},
{
"epoch": 0.5610253898277411,
"grad_norm": 48.26102317175817,
"learning_rate": 7.217734004316172e-08,
"logits/chosen": -1.5722343921661377,
"logits/rejected": -1.5775549411773682,
"logps/chosen": -803.7421875,
"logps/rejected": -866.3418579101562,
"loss": 0.4969,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.7525508403778076,
"rewards/margins": 0.8512840270996094,
"rewards/rejected": -3.603835105895996,
"step": 859
},
{
"epoch": 0.5616785043677035,
"grad_norm": 13.00017996435541,
"learning_rate": 7.200635805264625e-08,
"logits/chosen": -1.551430583000183,
"logits/rejected": -1.4977049827575684,
"logps/chosen": -795.7646484375,
"logps/rejected": -836.5863037109375,
"loss": 0.4979,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.18217134475708,
"rewards/margins": 0.8377689123153687,
"rewards/rejected": -4.019940376281738,
"step": 860
},
{
"epoch": 0.5623316189076659,
"grad_norm": 9.231732491941768,
"learning_rate": 7.183539164442973e-08,
"logits/chosen": -1.530940055847168,
"logits/rejected": -1.5165448188781738,
"logps/chosen": -862.6856689453125,
"logps/rejected": -992.7938232421875,
"loss": 0.455,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.554490089416504,
"rewards/margins": 1.1885462999343872,
"rewards/rejected": -4.743036270141602,
"step": 861
},
{
"epoch": 0.5629847334476283,
"grad_norm": 62.61248101231807,
"learning_rate": 7.166444170841473e-08,
"logits/chosen": -1.4707496166229248,
"logits/rejected": -1.4739537239074707,
"logps/chosen": -821.8203125,
"logps/rejected": -850.6475219726562,
"loss": 0.5211,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7928149700164795,
"rewards/margins": 0.47822028398513794,
"rewards/rejected": -3.2710354328155518,
"step": 862
},
{
"epoch": 0.5636378479875909,
"grad_norm": 109.02226402387393,
"learning_rate": 7.149350913441809e-08,
"logits/chosen": -1.5233734846115112,
"logits/rejected": -1.5370923280715942,
"logps/chosen": -793.0894775390625,
"logps/rejected": -808.7146606445312,
"loss": 0.534,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8652713298797607,
"rewards/margins": 0.5744865536689758,
"rewards/rejected": -3.43975830078125,
"step": 863
},
{
"epoch": 0.5642909625275533,
"grad_norm": 29.93004571485592,
"learning_rate": 7.132259481216628e-08,
"logits/chosen": -1.543608546257019,
"logits/rejected": -1.496211051940918,
"logps/chosen": -812.6435546875,
"logps/rejected": -853.930908203125,
"loss": 0.4421,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.108489990234375,
"rewards/margins": 0.8767549991607666,
"rewards/rejected": -3.9852447509765625,
"step": 864
},
{
"epoch": 0.5649440770675157,
"grad_norm": 61.12660963375552,
"learning_rate": 7.115169963129076e-08,
"logits/chosen": -1.5221493244171143,
"logits/rejected": -1.4868009090423584,
"logps/chosen": -754.1113891601562,
"logps/rejected": -894.2318115234375,
"loss": 0.475,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.7326500415802,
"rewards/margins": 0.7210690975189209,
"rewards/rejected": -3.453718662261963,
"step": 865
},
{
"epoch": 0.5655971916074781,
"grad_norm": 17.649759346104574,
"learning_rate": 7.098082448132339e-08,
"logits/chosen": -1.4271429777145386,
"logits/rejected": -1.3975777626037598,
"logps/chosen": -795.72412109375,
"logps/rejected": -917.9866943359375,
"loss": 0.5119,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1587576866149902,
"rewards/margins": 0.951889157295227,
"rewards/rejected": -4.110646724700928,
"step": 866
},
{
"epoch": 0.5662503061474407,
"grad_norm": 8.92934245433921,
"learning_rate": 7.080997025169167e-08,
"logits/chosen": -1.5210411548614502,
"logits/rejected": -1.5380655527114868,
"logps/chosen": -813.1768798828125,
"logps/rejected": -933.7425537109375,
"loss": 0.4734,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.348355293273926,
"rewards/margins": 0.8711065649986267,
"rewards/rejected": -4.2194623947143555,
"step": 867
},
{
"epoch": 0.5669034206874031,
"grad_norm": 12.604264013248898,
"learning_rate": 7.063913783171431e-08,
"logits/chosen": -1.532928466796875,
"logits/rejected": -1.5313588380813599,
"logps/chosen": -769.3619995117188,
"logps/rejected": -937.7196044921875,
"loss": 0.4693,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.7876639366149902,
"rewards/margins": 1.2222365140914917,
"rewards/rejected": -4.0099005699157715,
"step": 868
},
{
"epoch": 0.5675565352273655,
"grad_norm": 78.29234208780828,
"learning_rate": 7.046832811059646e-08,
"logits/chosen": -1.5154472589492798,
"logits/rejected": -1.5125453472137451,
"logps/chosen": -837.0894165039062,
"logps/rejected": -932.6472778320312,
"loss": 0.4858,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.860597848892212,
"rewards/margins": 0.9204597473144531,
"rewards/rejected": -3.781057596206665,
"step": 869
},
{
"epoch": 0.5682096497673279,
"grad_norm": 180.9882527133902,
"learning_rate": 7.029754197742512e-08,
"logits/chosen": -1.5261096954345703,
"logits/rejected": -1.5237348079681396,
"logps/chosen": -826.7614135742188,
"logps/rejected": -1008.6333618164062,
"loss": 0.5309,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2434701919555664,
"rewards/margins": 1.0818681716918945,
"rewards/rejected": -4.325338363647461,
"step": 870
},
{
"epoch": 0.5688627643072904,
"grad_norm": 40.32414479490123,
"learning_rate": 7.01267803211645e-08,
"logits/chosen": -1.5508898496627808,
"logits/rejected": -1.527504801750183,
"logps/chosen": -790.9599609375,
"logps/rejected": -838.9732666015625,
"loss": 0.5444,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.0933451652526855,
"rewards/margins": 0.376648485660553,
"rewards/rejected": -3.469994306564331,
"step": 871
},
{
"epoch": 0.5695158788472529,
"grad_norm": 72.83509120194428,
"learning_rate": 6.995604403065144e-08,
"logits/chosen": -1.5775620937347412,
"logits/rejected": -1.5864499807357788,
"logps/chosen": -769.3580932617188,
"logps/rejected": -876.96484375,
"loss": 0.5465,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.112041473388672,
"rewards/margins": 0.8449998497962952,
"rewards/rejected": -3.957040786743164,
"step": 872
},
{
"epoch": 0.5701689933872153,
"grad_norm": 12.905974016118856,
"learning_rate": 6.978533399459071e-08,
"logits/chosen": -1.5712107419967651,
"logits/rejected": -1.5447235107421875,
"logps/chosen": -777.7793579101562,
"logps/rejected": -840.2425537109375,
"loss": 0.5643,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9859039783477783,
"rewards/margins": 0.5791632533073425,
"rewards/rejected": -3.5650668144226074,
"step": 873
},
{
"epoch": 0.5708221079271777,
"grad_norm": 21.726965969176625,
"learning_rate": 6.961465110155043e-08,
"logits/chosen": -1.512514352798462,
"logits/rejected": -1.4853906631469727,
"logps/chosen": -893.4528198242188,
"logps/rejected": -1067.001220703125,
"loss": 0.4754,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4161694049835205,
"rewards/margins": 1.114807367324829,
"rewards/rejected": -4.530977249145508,
"step": 874
},
{
"epoch": 0.5714752224671402,
"grad_norm": 22.280036018028355,
"learning_rate": 6.944399623995744e-08,
"logits/chosen": -1.5252811908721924,
"logits/rejected": -1.5037376880645752,
"logps/chosen": -850.0037841796875,
"logps/rejected": -850.1954345703125,
"loss": 0.5284,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4480960369110107,
"rewards/margins": 0.41501325368881226,
"rewards/rejected": -3.8631091117858887,
"step": 875
},
{
"epoch": 0.5721283370071026,
"grad_norm": 40.53249032061718,
"learning_rate": 6.92733702980927e-08,
"logits/chosen": -1.5603222846984863,
"logits/rejected": -1.5633220672607422,
"logps/chosen": -817.9495849609375,
"logps/rejected": -1105.2486572265625,
"loss": 0.442,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.2014811038970947,
"rewards/margins": 1.719805359840393,
"rewards/rejected": -4.921286106109619,
"step": 876
},
{
"epoch": 0.5727814515470651,
"grad_norm": 18.002814145085477,
"learning_rate": 6.910277416408661e-08,
"logits/chosen": -1.5658931732177734,
"logits/rejected": -1.5071964263916016,
"logps/chosen": -770.361083984375,
"logps/rejected": -793.378173828125,
"loss": 0.4778,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1433448791503906,
"rewards/margins": 0.5039821863174438,
"rewards/rejected": -3.647326946258545,
"step": 877
},
{
"epoch": 0.5734345660870275,
"grad_norm": 12.307347390907383,
"learning_rate": 6.89322087259144e-08,
"logits/chosen": -1.511487603187561,
"logits/rejected": -1.5128693580627441,
"logps/chosen": -846.0947875976562,
"logps/rejected": -906.3300170898438,
"loss": 0.4864,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2089366912841797,
"rewards/margins": 0.9323613047599792,
"rewards/rejected": -4.141297340393066,
"step": 878
},
{
"epoch": 0.57408768062699,
"grad_norm": 21.452756223686876,
"learning_rate": 6.876167487139154e-08,
"logits/chosen": -1.526938796043396,
"logits/rejected": -1.5194412469863892,
"logps/chosen": -1003.674560546875,
"logps/rejected": -1024.025390625,
"loss": 0.5129,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.6137397289276123,
"rewards/margins": 0.7001018524169922,
"rewards/rejected": -4.313840866088867,
"step": 879
},
{
"epoch": 0.5747407951669524,
"grad_norm": 12.942447833043328,
"learning_rate": 6.859117348816912e-08,
"logits/chosen": -1.4752848148345947,
"logits/rejected": -1.3929071426391602,
"logps/chosen": -797.0738525390625,
"logps/rejected": -978.4248046875,
"loss": 0.5142,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4416940212249756,
"rewards/margins": 1.1079531908035278,
"rewards/rejected": -4.549646854400635,
"step": 880
},
{
"epoch": 0.5753939097069148,
"grad_norm": 64.69859008439977,
"learning_rate": 6.842070546372922e-08,
"logits/chosen": -1.5244097709655762,
"logits/rejected": -1.486146092414856,
"logps/chosen": -969.6875,
"logps/rejected": -1005.9173583984375,
"loss": 0.6013,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.9753427505493164,
"rewards/margins": 0.6842749118804932,
"rewards/rejected": -4.659617900848389,
"step": 881
},
{
"epoch": 0.5760470242468773,
"grad_norm": 42.79091393495759,
"learning_rate": 6.825027168538024e-08,
"logits/chosen": -1.4888124465942383,
"logits/rejected": -1.4149153232574463,
"logps/chosen": -813.0571899414062,
"logps/rejected": -825.3543701171875,
"loss": 0.574,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.368730068206787,
"rewards/margins": 0.39747297763824463,
"rewards/rejected": -3.766202926635742,
"step": 882
},
{
"epoch": 0.5767001387868398,
"grad_norm": 18.008747424093183,
"learning_rate": 6.807987304025236e-08,
"logits/chosen": -1.4898674488067627,
"logits/rejected": -1.4935599565505981,
"logps/chosen": -851.6963500976562,
"logps/rejected": -859.6405639648438,
"loss": 0.5219,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.2997894287109375,
"rewards/margins": 0.487488329410553,
"rewards/rejected": -3.787277936935425,
"step": 883
},
{
"epoch": 0.5773532533268022,
"grad_norm": 41.196321721092666,
"learning_rate": 6.790951041529286e-08,
"logits/chosen": -1.5880982875823975,
"logits/rejected": -1.5457425117492676,
"logps/chosen": -958.2757568359375,
"logps/rejected": -986.9423828125,
"loss": 0.522,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4477083683013916,
"rewards/margins": 0.6854377388954163,
"rewards/rejected": -4.133145809173584,
"step": 884
},
{
"epoch": 0.5780063678667646,
"grad_norm": 30.703596533657475,
"learning_rate": 6.773918469726156e-08,
"logits/chosen": -1.480318546295166,
"logits/rejected": -1.4444735050201416,
"logps/chosen": -940.3262939453125,
"logps/rejected": -1004.0597534179688,
"loss": 0.5406,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.5821070671081543,
"rewards/margins": 0.554044246673584,
"rewards/rejected": -4.136151313781738,
"step": 885
},
{
"epoch": 0.578659482406727,
"grad_norm": 126.80730044989116,
"learning_rate": 6.756889677272617e-08,
"logits/chosen": -1.5217347145080566,
"logits/rejected": -1.498375415802002,
"logps/chosen": -762.6358642578125,
"logps/rejected": -902.398681640625,
"loss": 0.5009,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.058880567550659,
"rewards/margins": 0.8846085667610168,
"rewards/rejected": -3.943488597869873,
"step": 886
},
{
"epoch": 0.5793125969466896,
"grad_norm": 11.803823871913366,
"learning_rate": 6.739864752805765e-08,
"logits/chosen": -1.4541584253311157,
"logits/rejected": -1.426889181137085,
"logps/chosen": -766.4666137695312,
"logps/rejected": -813.1023559570312,
"loss": 0.4918,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2732667922973633,
"rewards/margins": 0.5391014218330383,
"rewards/rejected": -3.8123679161071777,
"step": 887
},
{
"epoch": 0.579965711486652,
"grad_norm": 21.000885515886726,
"learning_rate": 6.722843784942565e-08,
"logits/chosen": -1.4722752571105957,
"logits/rejected": -1.467155933380127,
"logps/chosen": -837.9805908203125,
"logps/rejected": -857.227294921875,
"loss": 0.6011,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.297391653060913,
"rewards/margins": 0.3573925197124481,
"rewards/rejected": -3.6547842025756836,
"step": 888
},
{
"epoch": 0.5806188260266144,
"grad_norm": 17.365098193739787,
"learning_rate": 6.705826862279391e-08,
"logits/chosen": -1.4612233638763428,
"logits/rejected": -1.4755150079727173,
"logps/chosen": -873.2598876953125,
"logps/rejected": -963.6185913085938,
"loss": 0.5842,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.1137642860412598,
"rewards/margins": 1.115809679031372,
"rewards/rejected": -4.229574203491211,
"step": 889
},
{
"epoch": 0.5812719405665768,
"grad_norm": 13.20416027815365,
"learning_rate": 6.688814073391551e-08,
"logits/chosen": -1.4795432090759277,
"logits/rejected": -1.528482437133789,
"logps/chosen": -914.3403930664062,
"logps/rejected": -992.7194213867188,
"loss": 0.5484,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.8117380142211914,
"rewards/margins": 0.9581300616264343,
"rewards/rejected": -4.769867897033691,
"step": 890
},
{
"epoch": 0.5819250551065394,
"grad_norm": 58.643014274399484,
"learning_rate": 6.671805506832844e-08,
"logits/chosen": -1.4377716779708862,
"logits/rejected": -1.4766364097595215,
"logps/chosen": -770.939697265625,
"logps/rejected": -905.95458984375,
"loss": 0.5286,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8314456939697266,
"rewards/margins": 0.9767945408821106,
"rewards/rejected": -3.8082404136657715,
"step": 891
},
{
"epoch": 0.5825781696465018,
"grad_norm": 55.6022723858616,
"learning_rate": 6.654801251135092e-08,
"logits/chosen": -1.6092873811721802,
"logits/rejected": -1.6193262338638306,
"logps/chosen": -859.6929321289062,
"logps/rejected": -937.2235107421875,
"loss": 0.5144,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.6701273918151855,
"rewards/margins": 0.6065517067909241,
"rewards/rejected": -4.276679039001465,
"step": 892
},
{
"epoch": 0.5832312841864642,
"grad_norm": 21.239795373704855,
"learning_rate": 6.637801394807675e-08,
"logits/chosen": -1.465186357498169,
"logits/rejected": -1.5041096210479736,
"logps/chosen": -876.3399047851562,
"logps/rejected": -936.5634765625,
"loss": 0.479,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.4567980766296387,
"rewards/margins": 0.6494026184082031,
"rewards/rejected": -4.106200695037842,
"step": 893
},
{
"epoch": 0.5838843987264266,
"grad_norm": 22.901871352735185,
"learning_rate": 6.620806026337073e-08,
"logits/chosen": -1.567824363708496,
"logits/rejected": -1.5391998291015625,
"logps/chosen": -831.55615234375,
"logps/rejected": -894.0813598632812,
"loss": 0.5542,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1797890663146973,
"rewards/margins": 0.6112327575683594,
"rewards/rejected": -3.7910218238830566,
"step": 894
},
{
"epoch": 0.5845375132663891,
"grad_norm": 18.99146292432215,
"learning_rate": 6.603815234186409e-08,
"logits/chosen": -1.5131785869598389,
"logits/rejected": -1.4995113611221313,
"logps/chosen": -811.837158203125,
"logps/rejected": -859.697265625,
"loss": 0.4643,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3386635780334473,
"rewards/margins": 1.0135873556137085,
"rewards/rejected": -4.352251052856445,
"step": 895
},
{
"epoch": 0.5851906278063516,
"grad_norm": 85.94411267727696,
"learning_rate": 6.586829106794986e-08,
"logits/chosen": -1.4760074615478516,
"logits/rejected": -1.5129234790802002,
"logps/chosen": -827.99560546875,
"logps/rejected": -902.9091796875,
"loss": 0.55,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.300877809524536,
"rewards/margins": 0.9479480981826782,
"rewards/rejected": -4.248825550079346,
"step": 896
},
{
"epoch": 0.585843742346314,
"grad_norm": 28.761272058111484,
"learning_rate": 6.569847732577822e-08,
"logits/chosen": -1.51106595993042,
"logits/rejected": -1.5085562467575073,
"logps/chosen": -797.22216796875,
"logps/rejected": -902.5498046875,
"loss": 0.4886,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.1795077323913574,
"rewards/margins": 0.7169556021690369,
"rewards/rejected": -3.896463394165039,
"step": 897
},
{
"epoch": 0.5864968568862764,
"grad_norm": 20.15314666563717,
"learning_rate": 6.5528711999252e-08,
"logits/chosen": -1.5399476289749146,
"logits/rejected": -1.527458906173706,
"logps/chosen": -862.6876831054688,
"logps/rejected": -874.1339111328125,
"loss": 0.5053,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.810636520385742,
"rewards/margins": 0.7659615278244019,
"rewards/rejected": -3.5765976905822754,
"step": 898
},
{
"epoch": 0.5871499714262389,
"grad_norm": 12.274080681893953,
"learning_rate": 6.535899597202195e-08,
"logits/chosen": -1.513730525970459,
"logits/rejected": -1.484087586402893,
"logps/chosen": -859.1988525390625,
"logps/rejected": -1115.99365234375,
"loss": 0.429,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3637874126434326,
"rewards/margins": 1.6439146995544434,
"rewards/rejected": -5.007702350616455,
"step": 899
},
{
"epoch": 0.5878030859662013,
"grad_norm": 46.88003480260202,
"learning_rate": 6.518933012748232e-08,
"logits/chosen": -1.4574429988861084,
"logits/rejected": -1.4671128988265991,
"logps/chosen": -768.5726928710938,
"logps/rejected": -817.6867065429688,
"loss": 0.5007,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.511812210083008,
"rewards/margins": 0.7634918689727783,
"rewards/rejected": -3.2753043174743652,
"step": 900
},
{
"epoch": 0.5878030859662013,
"eval_logits/chosen": -1.5143874883651733,
"eval_logits/rejected": -1.500697135925293,
"eval_logps/chosen": -819.5907592773438,
"eval_logps/rejected": -894.2119750976562,
"eval_loss": 0.5088227987289429,
"eval_rewards/accuracies": 0.7540000081062317,
"eval_rewards/chosen": -3.092453956604004,
"eval_rewards/margins": 0.8272719979286194,
"eval_rewards/rejected": -3.9197258949279785,
"eval_runtime": 300.2638,
"eval_samples_per_second": 13.322,
"eval_steps_per_second": 0.833,
"step": 900
},
{
"epoch": 0.5884562005061638,
"grad_norm": 15.611510087556148,
"learning_rate": 6.5019715348766e-08,
"logits/chosen": -1.4778497219085693,
"logits/rejected": -1.4426562786102295,
"logps/chosen": -764.3760986328125,
"logps/rejected": -894.08740234375,
"loss": 0.4707,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.829637289047241,
"rewards/margins": 1.0723354816436768,
"rewards/rejected": -3.901973247528076,
"step": 901
},
{
"epoch": 0.5891093150461262,
"grad_norm": 27.04744713063808,
"learning_rate": 6.485015251874019e-08,
"logits/chosen": -1.4524825811386108,
"logits/rejected": -1.4247350692749023,
"logps/chosen": -833.126708984375,
"logps/rejected": -851.63037109375,
"loss": 0.5002,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4548346996307373,
"rewards/margins": 0.45528683066368103,
"rewards/rejected": -3.910121440887451,
"step": 902
},
{
"epoch": 0.5897624295860887,
"grad_norm": 14.887394055487325,
"learning_rate": 6.468064252000163e-08,
"logits/chosen": -1.4548847675323486,
"logits/rejected": -1.441070318222046,
"logps/chosen": -803.3111572265625,
"logps/rejected": -800.1891479492188,
"loss": 0.5278,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.290768623352051,
"rewards/margins": 0.5420413613319397,
"rewards/rejected": -3.8328099250793457,
"step": 903
},
{
"epoch": 0.5904155441260511,
"grad_norm": 88.18212760967876,
"learning_rate": 6.451118623487215e-08,
"logits/chosen": -1.4934018850326538,
"logits/rejected": -1.491977334022522,
"logps/chosen": -763.4550170898438,
"logps/rejected": -997.37255859375,
"loss": 0.5368,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.9802534580230713,
"rewards/margins": 1.5265756845474243,
"rewards/rejected": -4.506828784942627,
"step": 904
},
{
"epoch": 0.5910686586660135,
"grad_norm": 24.118764808026647,
"learning_rate": 6.434178454539393e-08,
"logits/chosen": -1.4591268301010132,
"logits/rejected": -1.475930094718933,
"logps/chosen": -738.4482421875,
"logps/rejected": -801.5521850585938,
"loss": 0.4642,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7284841537475586,
"rewards/margins": 0.7377627491950989,
"rewards/rejected": -3.466247081756592,
"step": 905
},
{
"epoch": 0.591721773205976,
"grad_norm": 17.550624193334937,
"learning_rate": 6.417243833332495e-08,
"logits/chosen": -1.5250346660614014,
"logits/rejected": -1.4923107624053955,
"logps/chosen": -816.05419921875,
"logps/rejected": -943.5857543945312,
"loss": 0.5238,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.181175708770752,
"rewards/margins": 1.0177024602890015,
"rewards/rejected": -4.198878288269043,
"step": 906
},
{
"epoch": 0.5923748877459385,
"grad_norm": 23.27302229809888,
"learning_rate": 6.400314848013446e-08,
"logits/chosen": -1.5263398885726929,
"logits/rejected": -1.5403982400894165,
"logps/chosen": -898.93408203125,
"logps/rejected": -945.4384765625,
"loss": 0.5122,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.269822835922241,
"rewards/margins": 0.5661499500274658,
"rewards/rejected": -3.835972785949707,
"step": 907
},
{
"epoch": 0.5930280022859009,
"grad_norm": 56.67567956066173,
"learning_rate": 6.383391586699837e-08,
"logits/chosen": -1.5245555639266968,
"logits/rejected": -1.5307707786560059,
"logps/chosen": -879.91064453125,
"logps/rejected": -969.3151245117188,
"loss": 0.4661,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0276691913604736,
"rewards/margins": 0.8546069264411926,
"rewards/rejected": -3.8822762966156006,
"step": 908
},
{
"epoch": 0.5936811168258633,
"grad_norm": 18.262554233951846,
"learning_rate": 6.366474137479459e-08,
"logits/chosen": -1.5455853939056396,
"logits/rejected": -1.428063154220581,
"logps/chosen": -790.438720703125,
"logps/rejected": -823.03271484375,
"loss": 0.5309,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.124584674835205,
"rewards/margins": 0.8692245483398438,
"rewards/rejected": -3.993809223175049,
"step": 909
},
{
"epoch": 0.5943342313658257,
"grad_norm": 37.07411490442863,
"learning_rate": 6.349562588409858e-08,
"logits/chosen": -1.540130853652954,
"logits/rejected": -1.5190839767456055,
"logps/chosen": -887.24462890625,
"logps/rejected": -943.5260620117188,
"loss": 0.5406,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2311315536499023,
"rewards/margins": 0.5626290440559387,
"rewards/rejected": -3.7937607765197754,
"step": 910
},
{
"epoch": 0.5949873459057883,
"grad_norm": 81.1787368218721,
"learning_rate": 6.332657027517865e-08,
"logits/chosen": -1.531748652458191,
"logits/rejected": -1.566206932067871,
"logps/chosen": -858.2910766601562,
"logps/rejected": -1005.5297241210938,
"loss": 0.5071,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.267298698425293,
"rewards/margins": 1.0169984102249146,
"rewards/rejected": -4.284296989440918,
"step": 911
},
{
"epoch": 0.5956404604457507,
"grad_norm": 32.28904552141215,
"learning_rate": 6.315757542799137e-08,
"logits/chosen": -1.5486570596694946,
"logits/rejected": -1.530015230178833,
"logps/chosen": -837.4837036132812,
"logps/rejected": -862.83447265625,
"loss": 0.5273,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.196139335632324,
"rewards/margins": 0.7048149108886719,
"rewards/rejected": -3.900954246520996,
"step": 912
},
{
"epoch": 0.5962935749857131,
"grad_norm": 10.082427550315396,
"learning_rate": 6.29886422221771e-08,
"logits/chosen": -1.6052241325378418,
"logits/rejected": -1.5839667320251465,
"logps/chosen": -873.5725708007812,
"logps/rejected": -978.7376708984375,
"loss": 0.4844,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.3677515983581543,
"rewards/margins": 1.1688979864120483,
"rewards/rejected": -4.536649227142334,
"step": 913
},
{
"epoch": 0.5969466895256755,
"grad_norm": 49.67965132832666,
"learning_rate": 6.281977153705534e-08,
"logits/chosen": -1.5648092031478882,
"logits/rejected": -1.5402863025665283,
"logps/chosen": -851.8053588867188,
"logps/rejected": -900.38818359375,
"loss": 0.5503,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.964672327041626,
"rewards/margins": 0.8616347312927246,
"rewards/rejected": -3.8263068199157715,
"step": 914
},
{
"epoch": 0.5975998040656381,
"grad_norm": 19.092868780473093,
"learning_rate": 6.265096425162015e-08,
"logits/chosen": -1.4740824699401855,
"logits/rejected": -1.4873621463775635,
"logps/chosen": -818.6807250976562,
"logps/rejected": -898.1356201171875,
"loss": 0.5058,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7591428756713867,
"rewards/margins": 1.210620641708374,
"rewards/rejected": -3.9697635173797607,
"step": 915
},
{
"epoch": 0.5982529186056005,
"grad_norm": 47.5168167772122,
"learning_rate": 6.24822212445356e-08,
"logits/chosen": -1.4681650400161743,
"logits/rejected": -1.4547978639602661,
"logps/chosen": -792.4906616210938,
"logps/rejected": -850.310546875,
"loss": 0.4641,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.865018367767334,
"rewards/margins": 0.6488918662071228,
"rewards/rejected": -3.5139102935791016,
"step": 916
},
{
"epoch": 0.5989060331455629,
"grad_norm": 33.12140555034356,
"learning_rate": 6.231354339413116e-08,
"logits/chosen": -1.4651777744293213,
"logits/rejected": -1.473002314567566,
"logps/chosen": -830.7955932617188,
"logps/rejected": -866.82080078125,
"loss": 0.5181,
"rewards/accuracies": 0.5625,
"rewards/chosen": -3.1264560222625732,
"rewards/margins": 0.34810954332351685,
"rewards/rejected": -3.4745657444000244,
"step": 917
},
{
"epoch": 0.5995591476855253,
"grad_norm": 15.005549258653641,
"learning_rate": 6.214493157839716e-08,
"logits/chosen": -1.5664610862731934,
"logits/rejected": -1.541873574256897,
"logps/chosen": -868.6499633789062,
"logps/rejected": -926.1016845703125,
"loss": 0.444,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0149686336517334,
"rewards/margins": 0.7024274468421936,
"rewards/rejected": -3.7173960208892822,
"step": 918
},
{
"epoch": 0.6002122622254878,
"grad_norm": 31.0641454388364,
"learning_rate": 6.197638667498022e-08,
"logits/chosen": -1.5562750101089478,
"logits/rejected": -1.5133049488067627,
"logps/chosen": -773.593994140625,
"logps/rejected": -830.7118530273438,
"loss": 0.5931,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2983880043029785,
"rewards/margins": 0.45370548963546753,
"rewards/rejected": -3.7520933151245117,
"step": 919
},
{
"epoch": 0.6008653767654503,
"grad_norm": 20.956839284919962,
"learning_rate": 6.180790956117867e-08,
"logits/chosen": -1.467447280883789,
"logits/rejected": -1.4105029106140137,
"logps/chosen": -782.5927124023438,
"logps/rejected": -818.0179443359375,
"loss": 0.508,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.0541372299194336,
"rewards/margins": 0.4292081892490387,
"rewards/rejected": -3.4833457469940186,
"step": 920
},
{
"epoch": 0.6015184913054127,
"grad_norm": 14.668503955298801,
"learning_rate": 6.163950111393799e-08,
"logits/chosen": -1.5520689487457275,
"logits/rejected": -1.5361812114715576,
"logps/chosen": -864.1260375976562,
"logps/rejected": -910.3584594726562,
"loss": 0.5677,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.183807134628296,
"rewards/margins": 1.1038367748260498,
"rewards/rejected": -4.287644386291504,
"step": 921
},
{
"epoch": 0.6021716058453751,
"grad_norm": 27.305517777571882,
"learning_rate": 6.147116220984622e-08,
"logits/chosen": -1.4598233699798584,
"logits/rejected": -1.460257649421692,
"logps/chosen": -816.2798461914062,
"logps/rejected": -911.81591796875,
"loss": 0.4793,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3738749027252197,
"rewards/margins": 0.7180493474006653,
"rewards/rejected": -4.09192419052124,
"step": 922
},
{
"epoch": 0.6028247203853376,
"grad_norm": 41.739674304138276,
"learning_rate": 6.130289372512946e-08,
"logits/chosen": -1.444016933441162,
"logits/rejected": -1.474906325340271,
"logps/chosen": -754.188232421875,
"logps/rejected": -857.038818359375,
"loss": 0.5205,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8269124031066895,
"rewards/margins": 0.8509232997894287,
"rewards/rejected": -3.677835464477539,
"step": 923
},
{
"epoch": 0.6034778349253,
"grad_norm": 58.02110360239537,
"learning_rate": 6.113469653564719e-08,
"logits/chosen": -1.4690029621124268,
"logits/rejected": -1.4412765502929688,
"logps/chosen": -810.4243774414062,
"logps/rejected": -840.3269653320312,
"loss": 0.5251,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8800265789031982,
"rewards/margins": 0.6990075707435608,
"rewards/rejected": -3.5790340900421143,
"step": 924
},
{
"epoch": 0.6041309494652625,
"grad_norm": 61.758663895068544,
"learning_rate": 6.096657151688788e-08,
"logits/chosen": -1.4827535152435303,
"logits/rejected": -1.4380667209625244,
"logps/chosen": -844.2979736328125,
"logps/rejected": -990.4068603515625,
"loss": 0.5044,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.213895082473755,
"rewards/margins": 0.7906748056411743,
"rewards/rejected": -4.004570007324219,
"step": 925
},
{
"epoch": 0.6047840640052249,
"grad_norm": 47.91069541881108,
"learning_rate": 6.07985195439643e-08,
"logits/chosen": -1.5746287107467651,
"logits/rejected": -1.5688815116882324,
"logps/chosen": -938.1190185546875,
"logps/rejected": -984.5784912109375,
"loss": 0.4972,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.6475698947906494,
"rewards/margins": 0.6271919012069702,
"rewards/rejected": -4.274761199951172,
"step": 926
},
{
"epoch": 0.6054371785451874,
"grad_norm": 12.621896694201023,
"learning_rate": 6.063054149160899e-08,
"logits/chosen": -1.4833399057388306,
"logits/rejected": -1.5237452983856201,
"logps/chosen": -763.331298828125,
"logps/rejected": -738.412353515625,
"loss": 0.6048,
"rewards/accuracies": 0.40625,
"rewards/chosen": -2.918760299682617,
"rewards/margins": 0.033799026161432266,
"rewards/rejected": -2.952559232711792,
"step": 927
},
{
"epoch": 0.6060902930851498,
"grad_norm": 48.565935333486564,
"learning_rate": 6.046263823416975e-08,
"logits/chosen": -1.5271244049072266,
"logits/rejected": -1.4947071075439453,
"logps/chosen": -845.4649047851562,
"logps/rejected": -829.05615234375,
"loss": 0.5946,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.154777765274048,
"rewards/margins": 0.36189785599708557,
"rewards/rejected": -3.5166759490966797,
"step": 928
},
{
"epoch": 0.6067434076251123,
"grad_norm": 201.1566716246379,
"learning_rate": 6.029481064560507e-08,
"logits/chosen": -1.5659098625183105,
"logits/rejected": -1.4598219394683838,
"logps/chosen": -745.5149536132812,
"logps/rejected": -827.2947998046875,
"loss": 0.5307,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.176211357116699,
"rewards/margins": 0.9403245449066162,
"rewards/rejected": -4.116535663604736,
"step": 929
},
{
"epoch": 0.6073965221650747,
"grad_norm": 77.8419741018575,
"learning_rate": 6.012705959947953e-08,
"logits/chosen": -1.5268572568893433,
"logits/rejected": -1.5410304069519043,
"logps/chosen": -791.3049926757812,
"logps/rejected": -944.24365234375,
"loss": 0.5016,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.995246410369873,
"rewards/margins": 0.96639084815979,
"rewards/rejected": -3.961637258529663,
"step": 930
},
{
"epoch": 0.6080496367050372,
"grad_norm": 21.356823796166157,
"learning_rate": 5.995938596895936e-08,
"logits/chosen": -1.5197898149490356,
"logits/rejected": -1.5448129177093506,
"logps/chosen": -880.7378540039062,
"logps/rejected": -929.0924682617188,
"loss": 0.4803,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1356070041656494,
"rewards/margins": 0.8477343320846558,
"rewards/rejected": -3.9833414554595947,
"step": 931
},
{
"epoch": 0.6087027512449996,
"grad_norm": 78.1296697326321,
"learning_rate": 5.979179062680777e-08,
"logits/chosen": -1.5615133047103882,
"logits/rejected": -1.5305187702178955,
"logps/chosen": -805.41748046875,
"logps/rejected": -857.9087524414062,
"loss": 0.4887,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.8929922580718994,
"rewards/margins": 0.7500249743461609,
"rewards/rejected": -3.643017292022705,
"step": 932
},
{
"epoch": 0.609355865784962,
"grad_norm": 44.16786903754792,
"learning_rate": 5.96242744453805e-08,
"logits/chosen": -1.5862808227539062,
"logits/rejected": -1.500042200088501,
"logps/chosen": -957.4008178710938,
"logps/rejected": -1003.3927001953125,
"loss": 0.5375,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4804091453552246,
"rewards/margins": 0.9446480870246887,
"rewards/rejected": -4.425057411193848,
"step": 933
},
{
"epoch": 0.6100089803249245,
"grad_norm": 43.88738627099548,
"learning_rate": 5.945683829662129e-08,
"logits/chosen": -1.4918217658996582,
"logits/rejected": -1.4933407306671143,
"logps/chosen": -826.7015380859375,
"logps/rejected": -888.3599853515625,
"loss": 0.5092,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.8977818489074707,
"rewards/margins": 0.904629111289978,
"rewards/rejected": -3.80241060256958,
"step": 934
},
{
"epoch": 0.610662094864887,
"grad_norm": 84.54075685318386,
"learning_rate": 5.928948305205719e-08,
"logits/chosen": -1.6410014629364014,
"logits/rejected": -1.5713551044464111,
"logps/chosen": -862.2045288085938,
"logps/rejected": -862.973876953125,
"loss": 0.4944,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.225311517715454,
"rewards/margins": 0.5583637356758118,
"rewards/rejected": -3.7836754322052,
"step": 935
},
{
"epoch": 0.6113152094048494,
"grad_norm": 49.202199653752565,
"learning_rate": 5.912220958279421e-08,
"logits/chosen": -1.525040626525879,
"logits/rejected": -1.5112783908843994,
"logps/chosen": -898.3567504882812,
"logps/rejected": -948.94091796875,
"loss": 0.4935,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4913763999938965,
"rewards/margins": 0.7729541063308716,
"rewards/rejected": -4.2643303871154785,
"step": 936
},
{
"epoch": 0.6119683239448118,
"grad_norm": 19.051733763857207,
"learning_rate": 5.895501875951271e-08,
"logits/chosen": -1.5405813455581665,
"logits/rejected": -1.4980101585388184,
"logps/chosen": -818.7405395507812,
"logps/rejected": -821.7682495117188,
"loss": 0.5139,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.834625482559204,
"rewards/margins": 0.37921154499053955,
"rewards/rejected": -3.213836669921875,
"step": 937
},
{
"epoch": 0.6126214384847742,
"grad_norm": 39.98147203055881,
"learning_rate": 5.878791145246284e-08,
"logits/chosen": -1.552310585975647,
"logits/rejected": -1.4916616678237915,
"logps/chosen": -828.3511962890625,
"logps/rejected": -897.640380859375,
"loss": 0.4085,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9902727603912354,
"rewards/margins": 0.938800036907196,
"rewards/rejected": -3.929072856903076,
"step": 938
},
{
"epoch": 0.6132745530247368,
"grad_norm": 9.665640962792187,
"learning_rate": 5.862088853146006e-08,
"logits/chosen": -1.483540654182434,
"logits/rejected": -1.4673054218292236,
"logps/chosen": -868.4718627929688,
"logps/rejected": -904.5808715820312,
"loss": 0.4925,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.237854480743408,
"rewards/margins": 0.40152764320373535,
"rewards/rejected": -3.6393821239471436,
"step": 939
},
{
"epoch": 0.6139276675646992,
"grad_norm": 29.6041722046408,
"learning_rate": 5.8453950865880574e-08,
"logits/chosen": -1.5257904529571533,
"logits/rejected": -1.5109593868255615,
"logps/chosen": -839.3677368164062,
"logps/rejected": -890.6753540039062,
"loss": 0.4616,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.944354295730591,
"rewards/margins": 0.7917759418487549,
"rewards/rejected": -3.7361302375793457,
"step": 940
},
{
"epoch": 0.6145807821046616,
"grad_norm": 36.1361730170012,
"learning_rate": 5.82870993246568e-08,
"logits/chosen": -1.575820803642273,
"logits/rejected": -1.5428611040115356,
"logps/chosen": -984.8616943359375,
"logps/rejected": -1002.877685546875,
"loss": 0.5118,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.8942298889160156,
"rewards/margins": 0.8039427995681763,
"rewards/rejected": -4.6981730461120605,
"step": 941
},
{
"epoch": 0.615233896644624,
"grad_norm": 65.95471806082536,
"learning_rate": 5.812033477627295e-08,
"logits/chosen": -1.5245524644851685,
"logits/rejected": -1.4878289699554443,
"logps/chosen": -797.1603393554688,
"logps/rejected": -862.60400390625,
"loss": 0.5424,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.427908420562744,
"rewards/margins": 0.5726900100708008,
"rewards/rejected": -4.000597953796387,
"step": 942
},
{
"epoch": 0.6158870111845864,
"grad_norm": 42.35986198202249,
"learning_rate": 5.795365808876033e-08,
"logits/chosen": -1.4567804336547852,
"logits/rejected": -1.4847595691680908,
"logps/chosen": -761.0791625976562,
"logps/rejected": -862.4808349609375,
"loss": 0.4818,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.592480421066284,
"rewards/margins": 0.8631225228309631,
"rewards/rejected": -3.4556026458740234,
"step": 943
},
{
"epoch": 0.616540125724549,
"grad_norm": 116.76247999305836,
"learning_rate": 5.778707012969296e-08,
"logits/chosen": -1.5003118515014648,
"logits/rejected": -1.4539164304733276,
"logps/chosen": -782.5404052734375,
"logps/rejected": -837.072509765625,
"loss": 0.5626,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.003209352493286,
"rewards/margins": 0.6795889139175415,
"rewards/rejected": -3.682798385620117,
"step": 944
},
{
"epoch": 0.6171932402645114,
"grad_norm": 36.38150538725848,
"learning_rate": 5.762057176618306e-08,
"logits/chosen": -1.5234191417694092,
"logits/rejected": -1.5212476253509521,
"logps/chosen": -823.0789184570312,
"logps/rejected": -1108.3194580078125,
"loss": 0.4678,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8806896209716797,
"rewards/margins": 1.4171735048294067,
"rewards/rejected": -4.297863483428955,
"step": 945
},
{
"epoch": 0.6178463548044738,
"grad_norm": 24.621342657442526,
"learning_rate": 5.745416386487637e-08,
"logits/chosen": -1.4999808073043823,
"logits/rejected": -1.5050296783447266,
"logps/chosen": -884.189697265625,
"logps/rejected": -1065.35546875,
"loss": 0.4709,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3289954662323,
"rewards/margins": 1.0373826026916504,
"rewards/rejected": -4.366378307342529,
"step": 946
},
{
"epoch": 0.6184994693444362,
"grad_norm": 39.9664210404075,
"learning_rate": 5.728784729194788e-08,
"logits/chosen": -1.6225008964538574,
"logits/rejected": -1.5957049131393433,
"logps/chosen": -797.822021484375,
"logps/rejected": -842.7942504882812,
"loss": 0.459,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.0013129711151123,
"rewards/margins": 0.8668568134307861,
"rewards/rejected": -3.8681697845458984,
"step": 947
},
{
"epoch": 0.6191525838843988,
"grad_norm": 18.300722710956293,
"learning_rate": 5.712162291309717e-08,
"logits/chosen": -1.4809714555740356,
"logits/rejected": -1.4627764225006104,
"logps/chosen": -780.6005859375,
"logps/rejected": -837.5048828125,
"loss": 0.4601,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.6685848236083984,
"rewards/margins": 0.947687029838562,
"rewards/rejected": -3.61627197265625,
"step": 948
},
{
"epoch": 0.6198056984243612,
"grad_norm": 48.042619463612205,
"learning_rate": 5.695549159354392e-08,
"logits/chosen": -1.5010271072387695,
"logits/rejected": -1.5268317461013794,
"logps/chosen": -862.3072509765625,
"logps/rejected": -902.5770263671875,
"loss": 0.5344,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.204814910888672,
"rewards/margins": 0.5735858082771301,
"rewards/rejected": -3.7784006595611572,
"step": 949
},
{
"epoch": 0.6204588129643236,
"grad_norm": 12.02401612140606,
"learning_rate": 5.678945419802344e-08,
"logits/chosen": -1.5606863498687744,
"logits/rejected": -1.553232192993164,
"logps/chosen": -786.1539916992188,
"logps/rejected": -862.3563842773438,
"loss": 0.4561,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9571497440338135,
"rewards/margins": 0.9905616641044617,
"rewards/rejected": -3.94771146774292,
"step": 950
},
{
"epoch": 0.621111927504286,
"grad_norm": 23.86580730911811,
"learning_rate": 5.662351159078216e-08,
"logits/chosen": -1.552533745765686,
"logits/rejected": -1.5205721855163574,
"logps/chosen": -958.0302124023438,
"logps/rejected": -975.6397094726562,
"loss": 0.5031,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.6955161094665527,
"rewards/margins": 0.5410246253013611,
"rewards/rejected": -4.2365403175354,
"step": 951
},
{
"epoch": 0.6217650420442485,
"grad_norm": 135.43034670723952,
"learning_rate": 5.645766463557309e-08,
"logits/chosen": -1.4750163555145264,
"logits/rejected": -1.4967775344848633,
"logps/chosen": -814.1315307617188,
"logps/rejected": -898.7764282226562,
"loss": 0.554,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.208841562271118,
"rewards/margins": 0.698648989200592,
"rewards/rejected": -3.9074904918670654,
"step": 952
},
{
"epoch": 0.622418156584211,
"grad_norm": 60.25491022191491,
"learning_rate": 5.629191419565141e-08,
"logits/chosen": -1.471923828125,
"logits/rejected": -1.5026978254318237,
"logps/chosen": -836.269775390625,
"logps/rejected": -875.132080078125,
"loss": 0.4922,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.013134479522705,
"rewards/margins": 0.8672950267791748,
"rewards/rejected": -3.880429744720459,
"step": 953
},
{
"epoch": 0.6230712711241734,
"grad_norm": 70.1250850771653,
"learning_rate": 5.612626113376988e-08,
"logits/chosen": -1.415960431098938,
"logits/rejected": -1.4024447202682495,
"logps/chosen": -861.2704467773438,
"logps/rejected": -906.9395751953125,
"loss": 0.4583,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.9992175102233887,
"rewards/margins": 1.1700170040130615,
"rewards/rejected": -4.169234752655029,
"step": 954
},
{
"epoch": 0.6237243856641358,
"grad_norm": 118.63943863651916,
"learning_rate": 5.596070631217441e-08,
"logits/chosen": -1.5227341651916504,
"logits/rejected": -1.5267282724380493,
"logps/chosen": -795.377197265625,
"logps/rejected": -826.0686645507812,
"loss": 0.5701,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9778614044189453,
"rewards/margins": 0.6003108024597168,
"rewards/rejected": -3.578172206878662,
"step": 955
},
{
"epoch": 0.6243775002040983,
"grad_norm": 17.56854621129216,
"learning_rate": 5.579525059259957e-08,
"logits/chosen": -1.5488358736038208,
"logits/rejected": -1.573460340499878,
"logps/chosen": -817.08447265625,
"logps/rejected": -863.1941528320312,
"loss": 0.4524,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.050839900970459,
"rewards/margins": 0.9858266115188599,
"rewards/rejected": -4.0366668701171875,
"step": 956
},
{
"epoch": 0.6250306147440607,
"grad_norm": 69.88285733489641,
"learning_rate": 5.562989483626409e-08,
"logits/chosen": -1.4820830821990967,
"logits/rejected": -1.4609520435333252,
"logps/chosen": -727.6483154296875,
"logps/rejected": -839.3212280273438,
"loss": 0.4287,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8100788593292236,
"rewards/margins": 1.259989857673645,
"rewards/rejected": -4.070068836212158,
"step": 957
},
{
"epoch": 0.6256837292840232,
"grad_norm": 12.018768296537614,
"learning_rate": 5.546463990386634e-08,
"logits/chosen": -1.5005813837051392,
"logits/rejected": -1.522735357284546,
"logps/chosen": -787.5521850585938,
"logps/rejected": -897.03564453125,
"loss": 0.4824,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8902053833007812,
"rewards/margins": 1.399174451828003,
"rewards/rejected": -4.289380073547363,
"step": 958
},
{
"epoch": 0.6263368438239856,
"grad_norm": 33.17348323808573,
"learning_rate": 5.5299486655579924e-08,
"logits/chosen": -1.530156135559082,
"logits/rejected": -1.5278276205062866,
"logps/chosen": -807.750244140625,
"logps/rejected": -950.3114013671875,
"loss": 0.4757,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.519279956817627,
"rewards/margins": 1.149764060974121,
"rewards/rejected": -4.669044494628906,
"step": 959
},
{
"epoch": 0.6269899583639481,
"grad_norm": 15.26418942385675,
"learning_rate": 5.513443595104917e-08,
"logits/chosen": -1.4855040311813354,
"logits/rejected": -1.495133638381958,
"logps/chosen": -800.33837890625,
"logps/rejected": -888.4577026367188,
"loss": 0.4282,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.716895580291748,
"rewards/margins": 0.8394394516944885,
"rewards/rejected": -3.556334972381592,
"step": 960
},
{
"epoch": 0.6276430729039105,
"grad_norm": 18.55918206412983,
"learning_rate": 5.496948864938463e-08,
"logits/chosen": -1.564420461654663,
"logits/rejected": -1.5435676574707031,
"logps/chosen": -925.506591796875,
"logps/rejected": -1076.7103271484375,
"loss": 0.4707,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.589116096496582,
"rewards/margins": 1.4336352348327637,
"rewards/rejected": -5.022751331329346,
"step": 961
},
{
"epoch": 0.6282961874438729,
"grad_norm": 36.21241704409419,
"learning_rate": 5.480464560915865e-08,
"logits/chosen": -1.457442045211792,
"logits/rejected": -1.466660976409912,
"logps/chosen": -880.1464233398438,
"logps/rejected": -952.699462890625,
"loss": 0.5234,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.218412160873413,
"rewards/margins": 0.8761493563652039,
"rewards/rejected": -4.094561576843262,
"step": 962
},
{
"epoch": 0.6289493019838354,
"grad_norm": 59.3801136308343,
"learning_rate": 5.463990768840088e-08,
"logits/chosen": -1.5099765062332153,
"logits/rejected": -1.4892768859863281,
"logps/chosen": -755.8883056640625,
"logps/rejected": -835.78173828125,
"loss": 0.5088,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.6380348205566406,
"rewards/margins": 0.8317041993141174,
"rewards/rejected": -3.469738721847534,
"step": 963
},
{
"epoch": 0.6296024165237979,
"grad_norm": 47.301890427275495,
"learning_rate": 5.447527574459378e-08,
"logits/chosen": -1.522971272468567,
"logits/rejected": -1.533935546875,
"logps/chosen": -732.6190185546875,
"logps/rejected": -835.910888671875,
"loss": 0.4433,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6405673027038574,
"rewards/margins": 0.8309347033500671,
"rewards/rejected": -3.4715018272399902,
"step": 964
},
{
"epoch": 0.6302555310637603,
"grad_norm": 36.636882714332266,
"learning_rate": 5.431075063466824e-08,
"logits/chosen": -1.485040307044983,
"logits/rejected": -1.5213403701782227,
"logps/chosen": -858.669921875,
"logps/rejected": -968.9814453125,
"loss": 0.4807,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.340672254562378,
"rewards/margins": 0.8855950236320496,
"rewards/rejected": -4.226266860961914,
"step": 965
},
{
"epoch": 0.6309086456037227,
"grad_norm": 126.49507570203406,
"learning_rate": 5.4146333214998996e-08,
"logits/chosen": -1.5613657236099243,
"logits/rejected": -1.510394811630249,
"logps/chosen": -831.458740234375,
"logps/rejected": -846.1495971679688,
"loss": 0.5621,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.379094362258911,
"rewards/margins": 0.5405767560005188,
"rewards/rejected": -3.919670820236206,
"step": 966
},
{
"epoch": 0.6315617601436851,
"grad_norm": 34.41276919074804,
"learning_rate": 5.39820243414003e-08,
"logits/chosen": -1.5569851398468018,
"logits/rejected": -1.5476921796798706,
"logps/chosen": -832.99169921875,
"logps/rejected": -912.7144165039062,
"loss": 0.4608,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6668541431427,
"rewards/margins": 0.7351149916648865,
"rewards/rejected": -4.401968955993652,
"step": 967
},
{
"epoch": 0.6322148746836477,
"grad_norm": 16.388453341999888,
"learning_rate": 5.381782486912144e-08,
"logits/chosen": -1.5224254131317139,
"logits/rejected": -1.48786199092865,
"logps/chosen": -816.0333862304688,
"logps/rejected": -842.23486328125,
"loss": 0.5767,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.6304800510406494,
"rewards/margins": 0.6139175891876221,
"rewards/rejected": -4.24439811706543,
"step": 968
},
{
"epoch": 0.6328679892236101,
"grad_norm": 16.605006142190543,
"learning_rate": 5.365373565284211e-08,
"logits/chosen": -1.5403863191604614,
"logits/rejected": -1.54334557056427,
"logps/chosen": -795.1408081054688,
"logps/rejected": -888.0582275390625,
"loss": 0.5094,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9061810970306396,
"rewards/margins": 0.6548399925231934,
"rewards/rejected": -3.561021089553833,
"step": 969
},
{
"epoch": 0.6335211037635725,
"grad_norm": 86.2008006757627,
"learning_rate": 5.348975754666825e-08,
"logits/chosen": -1.4919219017028809,
"logits/rejected": -1.508528709411621,
"logps/chosen": -781.9150390625,
"logps/rejected": -914.5775146484375,
"loss": 0.5111,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3672971725463867,
"rewards/margins": 0.6925301551818848,
"rewards/rejected": -4.0598273277282715,
"step": 970
},
{
"epoch": 0.6341742183035349,
"grad_norm": 108.14966269733243,
"learning_rate": 5.33258914041274e-08,
"logits/chosen": -1.6157301664352417,
"logits/rejected": -1.5612612962722778,
"logps/chosen": -840.8732299804688,
"logps/rejected": -893.8431396484375,
"loss": 0.4703,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.353294610977173,
"rewards/margins": 0.7179979681968689,
"rewards/rejected": -4.071292877197266,
"step": 971
},
{
"epoch": 0.6348273328434975,
"grad_norm": 13.7781550358131,
"learning_rate": 5.316213807816432e-08,
"logits/chosen": -1.5367947816848755,
"logits/rejected": -1.4901845455169678,
"logps/chosen": -795.0319213867188,
"logps/rejected": -865.484375,
"loss": 0.4664,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.197481155395508,
"rewards/margins": 0.9921532869338989,
"rewards/rejected": -4.189634323120117,
"step": 972
},
{
"epoch": 0.6354804473834599,
"grad_norm": 22.603804513420233,
"learning_rate": 5.299849842113656e-08,
"logits/chosen": -1.5428191423416138,
"logits/rejected": -1.565401554107666,
"logps/chosen": -923.1882934570312,
"logps/rejected": -956.2698974609375,
"loss": 0.4838,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5026655197143555,
"rewards/margins": 0.8109586238861084,
"rewards/rejected": -4.313623905181885,
"step": 973
},
{
"epoch": 0.6361335619234223,
"grad_norm": 126.16238690733957,
"learning_rate": 5.283497328480998e-08,
"logits/chosen": -1.6220266819000244,
"logits/rejected": -1.613905906677246,
"logps/chosen": -949.4186401367188,
"logps/rejected": -993.9503784179688,
"loss": 0.5042,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.326920986175537,
"rewards/margins": 0.878463864326477,
"rewards/rejected": -4.205385208129883,
"step": 974
},
{
"epoch": 0.6367866764633847,
"grad_norm": 65.27147183382468,
"learning_rate": 5.267156352035437e-08,
"logits/chosen": -1.5607855319976807,
"logits/rejected": -1.5158913135528564,
"logps/chosen": -880.0157470703125,
"logps/rejected": -966.0748291015625,
"loss": 0.4936,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.369838237762451,
"rewards/margins": 0.9538223147392273,
"rewards/rejected": -4.323660850524902,
"step": 975
},
{
"epoch": 0.6374397910033472,
"grad_norm": 22.36577561540512,
"learning_rate": 5.250826997833899e-08,
"logits/chosen": -1.5453389883041382,
"logits/rejected": -1.540825366973877,
"logps/chosen": -879.3869018554688,
"logps/rejected": -885.604248046875,
"loss": 0.5292,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.171088218688965,
"rewards/margins": 0.3536863625049591,
"rewards/rejected": -3.5247745513916016,
"step": 976
},
{
"epoch": 0.6380929055433097,
"grad_norm": 10.578277029115368,
"learning_rate": 5.234509350872813e-08,
"logits/chosen": -1.5381667613983154,
"logits/rejected": -1.5094205141067505,
"logps/chosen": -812.7030639648438,
"logps/rejected": -927.2329711914062,
"loss": 0.5076,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.316643476486206,
"rewards/margins": 1.0832691192626953,
"rewards/rejected": -4.399912357330322,
"step": 977
},
{
"epoch": 0.6387460200832721,
"grad_norm": 117.19097325210925,
"learning_rate": 5.218203496087671e-08,
"logits/chosen": -1.555480718612671,
"logits/rejected": -1.554026484489441,
"logps/chosen": -885.7477416992188,
"logps/rejected": -906.7445068359375,
"loss": 0.6101,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1535983085632324,
"rewards/margins": 0.5752413272857666,
"rewards/rejected": -3.72883939743042,
"step": 978
},
{
"epoch": 0.6393991346232345,
"grad_norm": 43.06647390428903,
"learning_rate": 5.2019095183525886e-08,
"logits/chosen": -1.5017975568771362,
"logits/rejected": -1.5113730430603027,
"logps/chosen": -754.3181762695312,
"logps/rejected": -775.8157958984375,
"loss": 0.5283,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1152279376983643,
"rewards/margins": 0.6389954090118408,
"rewards/rejected": -3.754223346710205,
"step": 979
},
{
"epoch": 0.640052249163197,
"grad_norm": 41.64410725373992,
"learning_rate": 5.185627502479857e-08,
"logits/chosen": -1.5162409543991089,
"logits/rejected": -1.5376276969909668,
"logps/chosen": -834.6458740234375,
"logps/rejected": -990.2511596679688,
"loss": 0.5032,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.955130100250244,
"rewards/margins": 1.1836378574371338,
"rewards/rejected": -4.138767719268799,
"step": 980
},
{
"epoch": 0.6407053637031594,
"grad_norm": 58.564011363723154,
"learning_rate": 5.1693575332195006e-08,
"logits/chosen": -1.4699060916900635,
"logits/rejected": -1.4639524221420288,
"logps/chosen": -755.951416015625,
"logps/rejected": -867.30126953125,
"loss": 0.5015,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.9639952182769775,
"rewards/margins": 1.0040624141693115,
"rewards/rejected": -3.968057632446289,
"step": 981
},
{
"epoch": 0.6413584782431219,
"grad_norm": 87.46285821272616,
"learning_rate": 5.153099695258843e-08,
"logits/chosen": -1.4716684818267822,
"logits/rejected": -1.5059077739715576,
"logps/chosen": -923.4231567382812,
"logps/rejected": -1027.0228271484375,
"loss": 0.4797,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.596079111099243,
"rewards/margins": 1.1137962341308594,
"rewards/rejected": -4.709875583648682,
"step": 982
},
{
"epoch": 0.6420115927830843,
"grad_norm": 34.65500498760471,
"learning_rate": 5.1368540732220656e-08,
"logits/chosen": -1.5052711963653564,
"logits/rejected": -1.525931477546692,
"logps/chosen": -801.951171875,
"logps/rejected": -935.139404296875,
"loss": 0.5685,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9107887744903564,
"rewards/margins": 1.193021535873413,
"rewards/rejected": -4.1038103103637695,
"step": 983
},
{
"epoch": 0.6426647073230468,
"grad_norm": 15.45197836175366,
"learning_rate": 5.1206207516697614e-08,
"logits/chosen": -1.5689780712127686,
"logits/rejected": -1.4814107418060303,
"logps/chosen": -876.8812255859375,
"logps/rejected": -1007.7050170898438,
"loss": 0.5014,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.607487678527832,
"rewards/margins": 1.0582144260406494,
"rewards/rejected": -4.665701866149902,
"step": 984
},
{
"epoch": 0.6433178218630092,
"grad_norm": 29.42430407328828,
"learning_rate": 5.104399815098496e-08,
"logits/chosen": -1.5423859357833862,
"logits/rejected": -1.5497791767120361,
"logps/chosen": -816.120849609375,
"logps/rejected": -943.1589965820312,
"loss": 0.5109,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1182186603546143,
"rewards/margins": 0.9232460856437683,
"rewards/rejected": -4.041464805603027,
"step": 985
},
{
"epoch": 0.6439709364029716,
"grad_norm": 18.815807983433164,
"learning_rate": 5.088191347940375e-08,
"logits/chosen": -1.51715087890625,
"logits/rejected": -1.5343945026397705,
"logps/chosen": -813.7894287109375,
"logps/rejected": -986.6640014648438,
"loss": 0.413,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.2437820434570312,
"rewards/margins": 1.1290417909622192,
"rewards/rejected": -4.372823238372803,
"step": 986
},
{
"epoch": 0.6446240509429341,
"grad_norm": 25.895229217854762,
"learning_rate": 5.071995434562592e-08,
"logits/chosen": -1.4702000617980957,
"logits/rejected": -1.4505935907363892,
"logps/chosen": -854.226318359375,
"logps/rejected": -964.643310546875,
"loss": 0.4751,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3152987957000732,
"rewards/margins": 0.7256244421005249,
"rewards/rejected": -4.040923595428467,
"step": 987
},
{
"epoch": 0.6452771654828966,
"grad_norm": 33.724194436827155,
"learning_rate": 5.055812159267003e-08,
"logits/chosen": -1.5073872804641724,
"logits/rejected": -1.4952834844589233,
"logps/chosen": -855.9230346679688,
"logps/rejected": -1033.7293701171875,
"loss": 0.4451,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.2582755088806152,
"rewards/margins": 1.6962679624557495,
"rewards/rejected": -4.9545440673828125,
"step": 988
},
{
"epoch": 0.645930280022859,
"grad_norm": 11.66573176765737,
"learning_rate": 5.0396416062896766e-08,
"logits/chosen": -1.579730749130249,
"logits/rejected": -1.5733931064605713,
"logps/chosen": -906.1632080078125,
"logps/rejected": -925.17626953125,
"loss": 0.5891,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.705543041229248,
"rewards/margins": 0.6829342246055603,
"rewards/rejected": -4.388477325439453,
"step": 989
},
{
"epoch": 0.6465833945628214,
"grad_norm": 150.79764751923673,
"learning_rate": 5.023483859800463e-08,
"logits/chosen": -1.5521610975265503,
"logits/rejected": -1.5487536191940308,
"logps/chosen": -818.4269409179688,
"logps/rejected": -842.7174682617188,
"loss": 0.5022,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1259286403656006,
"rewards/margins": 0.5289927124977112,
"rewards/rejected": -3.654921293258667,
"step": 990
},
{
"epoch": 0.6472365091027839,
"grad_norm": 16.00702278017296,
"learning_rate": 5.0073390039025534e-08,
"logits/chosen": -1.443716049194336,
"logits/rejected": -1.415492057800293,
"logps/chosen": -841.180419921875,
"logps/rejected": -963.9603271484375,
"loss": 0.5006,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2827086448669434,
"rewards/margins": 0.9521893262863159,
"rewards/rejected": -4.234897613525391,
"step": 991
},
{
"epoch": 0.6478896236427464,
"grad_norm": 65.24190893051149,
"learning_rate": 4.991207122632035e-08,
"logits/chosen": -1.5591700077056885,
"logits/rejected": -1.5191017389297485,
"logps/chosen": -942.4129638671875,
"logps/rejected": -1009.4859619140625,
"loss": 0.5094,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.8302550315856934,
"rewards/margins": 0.7394087314605713,
"rewards/rejected": -4.5696635246276855,
"step": 992
},
{
"epoch": 0.6485427381827088,
"grad_norm": 41.277206672220316,
"learning_rate": 4.975088299957471e-08,
"logits/chosen": -1.557931661605835,
"logits/rejected": -1.555905818939209,
"logps/chosen": -836.3804931640625,
"logps/rejected": -922.718994140625,
"loss": 0.4929,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.1140992641448975,
"rewards/margins": 1.1032991409301758,
"rewards/rejected": -4.217398643493652,
"step": 993
},
{
"epoch": 0.6491958527226712,
"grad_norm": 14.725778545114895,
"learning_rate": 4.958982619779442e-08,
"logits/chosen": -1.5064477920532227,
"logits/rejected": -1.521809697151184,
"logps/chosen": -736.304931640625,
"logps/rejected": -813.41650390625,
"loss": 0.4734,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9339964389801025,
"rewards/margins": 0.7689220309257507,
"rewards/rejected": -3.702918291091919,
"step": 994
},
{
"epoch": 0.6498489672626336,
"grad_norm": 52.91945215629695,
"learning_rate": 4.942890165930129e-08,
"logits/chosen": -1.5186612606048584,
"logits/rejected": -1.5100464820861816,
"logps/chosen": -754.244140625,
"logps/rejected": -958.1654052734375,
"loss": 0.4943,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.929853916168213,
"rewards/margins": 1.4619183540344238,
"rewards/rejected": -4.391772270202637,
"step": 995
},
{
"epoch": 0.6505020818025962,
"grad_norm": 61.7741247551618,
"learning_rate": 4.926811022172866e-08,
"logits/chosen": -1.5684045553207397,
"logits/rejected": -1.464762568473816,
"logps/chosen": -862.2124633789062,
"logps/rejected": -940.2279052734375,
"loss": 0.4754,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.588654041290283,
"rewards/margins": 1.0604631900787354,
"rewards/rejected": -4.649117469787598,
"step": 996
},
{
"epoch": 0.6511551963425586,
"grad_norm": 85.06663516145058,
"learning_rate": 4.9107452722017015e-08,
"logits/chosen": -1.4936169385910034,
"logits/rejected": -1.497806429862976,
"logps/chosen": -827.88720703125,
"logps/rejected": -843.3629150390625,
"loss": 0.5419,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.254192352294922,
"rewards/margins": 0.3634593188762665,
"rewards/rejected": -3.6176517009735107,
"step": 997
},
{
"epoch": 0.651808310882521,
"grad_norm": 26.951955302254902,
"learning_rate": 4.894692999640973e-08,
"logits/chosen": -1.5051591396331787,
"logits/rejected": -1.4692599773406982,
"logps/chosen": -797.8718872070312,
"logps/rejected": -915.5535888671875,
"loss": 0.5042,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2658886909484863,
"rewards/margins": 0.9658613801002502,
"rewards/rejected": -4.231750011444092,
"step": 998
},
{
"epoch": 0.6524614254224834,
"grad_norm": 55.7608063216077,
"learning_rate": 4.8786542880448653e-08,
"logits/chosen": -1.537933349609375,
"logits/rejected": -1.4756388664245605,
"logps/chosen": -817.772705078125,
"logps/rejected": -852.2359008789062,
"loss": 0.5176,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.0438647270202637,
"rewards/margins": 0.5471150875091553,
"rewards/rejected": -3.590980052947998,
"step": 999
},
{
"epoch": 0.653114539962446,
"grad_norm": 25.848553185603418,
"learning_rate": 4.8626292208969733e-08,
"logits/chosen": -1.553949236869812,
"logits/rejected": -1.5505160093307495,
"logps/chosen": -906.37109375,
"logps/rejected": -967.929931640625,
"loss": 0.485,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.5798516273498535,
"rewards/margins": 0.8147382140159607,
"rewards/rejected": -4.394589900970459,
"step": 1000
},
{
"epoch": 0.653114539962446,
"eval_logits/chosen": -1.4997057914733887,
"eval_logits/rejected": -1.483424186706543,
"eval_logps/chosen": -823.3939819335938,
"eval_logps/rejected": -900.8679809570312,
"eval_loss": 0.5032714605331421,
"eval_rewards/accuracies": 0.7630000114440918,
"eval_rewards/chosen": -3.1304855346679688,
"eval_rewards/margins": 0.8558005094528198,
"eval_rewards/rejected": -3.986285924911499,
"eval_runtime": 296.6594,
"eval_samples_per_second": 13.483,
"eval_steps_per_second": 0.843,
"step": 1000
},
{
"epoch": 0.6537676545024084,
"grad_norm": 30.491811094126753,
"learning_rate": 4.846617881609876e-08,
"logits/chosen": -1.520873785018921,
"logits/rejected": -1.4895522594451904,
"logps/chosen": -772.0455322265625,
"logps/rejected": -870.1766357421875,
"loss": 0.4863,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9721617698669434,
"rewards/margins": 0.8731052875518799,
"rewards/rejected": -3.845266819000244,
"step": 1001
},
{
"epoch": 0.6544207690423708,
"grad_norm": 40.22401127775807,
"learning_rate": 4.8306203535246946e-08,
"logits/chosen": -1.5727264881134033,
"logits/rejected": -1.464941143989563,
"logps/chosen": -808.9293212890625,
"logps/rejected": -871.8182373046875,
"loss": 0.5355,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1198384761810303,
"rewards/margins": 0.9884458780288696,
"rewards/rejected": -4.108283996582031,
"step": 1002
},
{
"epoch": 0.6550738835823332,
"grad_norm": 11.806744900302544,
"learning_rate": 4.814636719910657e-08,
"logits/chosen": -1.5598303079605103,
"logits/rejected": -1.474313735961914,
"logps/chosen": -862.10400390625,
"logps/rejected": -857.1552124023438,
"loss": 0.546,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.641342878341675,
"rewards/margins": 0.5948621034622192,
"rewards/rejected": -4.236205101013184,
"step": 1003
},
{
"epoch": 0.6557269981222957,
"grad_norm": 15.051255942076827,
"learning_rate": 4.798667063964673e-08,
"logits/chosen": -1.5384275913238525,
"logits/rejected": -1.494019865989685,
"logps/chosen": -754.1705932617188,
"logps/rejected": -797.1336059570312,
"loss": 0.472,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.307189702987671,
"rewards/margins": 0.5865740180015564,
"rewards/rejected": -3.893763542175293,
"step": 1004
},
{
"epoch": 0.6563801126622582,
"grad_norm": 12.800756135654732,
"learning_rate": 4.7827114688108985e-08,
"logits/chosen": -1.478043556213379,
"logits/rejected": -1.4925929307937622,
"logps/chosen": -698.2744140625,
"logps/rejected": -822.2246704101562,
"loss": 0.5131,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.6346049308776855,
"rewards/margins": 0.8852910399436951,
"rewards/rejected": -3.5198960304260254,
"step": 1005
},
{
"epoch": 0.6570332272022206,
"grad_norm": 16.667891699953692,
"learning_rate": 4.7667700175002986e-08,
"logits/chosen": -1.4476077556610107,
"logits/rejected": -1.460363745689392,
"logps/chosen": -839.6358032226562,
"logps/rejected": -914.096923828125,
"loss": 0.4806,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.174222469329834,
"rewards/margins": 1.0781100988388062,
"rewards/rejected": -4.2523322105407715,
"step": 1006
},
{
"epoch": 0.657686341742183,
"grad_norm": 25.715571290386414,
"learning_rate": 4.750842793010217e-08,
"logits/chosen": -1.5329231023788452,
"logits/rejected": -1.5252712965011597,
"logps/chosen": -878.42041015625,
"logps/rejected": -985.8428955078125,
"loss": 0.5129,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.7149271965026855,
"rewards/margins": 0.733319878578186,
"rewards/rejected": -4.448246955871582,
"step": 1007
},
{
"epoch": 0.6583394562821455,
"grad_norm": 104.75961567327492,
"learning_rate": 4.7349298782439464e-08,
"logits/chosen": -1.488255500793457,
"logits/rejected": -1.4739494323730469,
"logps/chosen": -775.230224609375,
"logps/rejected": -847.8889770507812,
"loss": 0.4897,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7119245529174805,
"rewards/margins": 0.9239524602890015,
"rewards/rejected": -3.6358766555786133,
"step": 1008
},
{
"epoch": 0.6589925708221079,
"grad_norm": 52.144252857671894,
"learning_rate": 4.719031356030294e-08,
"logits/chosen": -1.5857702493667603,
"logits/rejected": -1.5610796213150024,
"logps/chosen": -816.1323852539062,
"logps/rejected": -970.305419921875,
"loss": 0.5158,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0045716762542725,
"rewards/margins": 1.0432796478271484,
"rewards/rejected": -4.0478515625,
"step": 1009
},
{
"epoch": 0.6596456853620704,
"grad_norm": 102.68956882940834,
"learning_rate": 4.703147309123156e-08,
"logits/chosen": -1.4632647037506104,
"logits/rejected": -1.4590816497802734,
"logps/chosen": -727.7660522460938,
"logps/rejected": -776.1034545898438,
"loss": 0.5347,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.654884099960327,
"rewards/margins": 0.699772298336029,
"rewards/rejected": -3.35465669631958,
"step": 1010
},
{
"epoch": 0.6602987999020328,
"grad_norm": 23.05602130979561,
"learning_rate": 4.687277820201077e-08,
"logits/chosen": -1.5591952800750732,
"logits/rejected": -1.5697981119155884,
"logps/chosen": -876.64599609375,
"logps/rejected": -1013.8555297851562,
"loss": 0.4686,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.513322353363037,
"rewards/margins": 0.9770915508270264,
"rewards/rejected": -4.490414142608643,
"step": 1011
},
{
"epoch": 0.6609519144419953,
"grad_norm": 36.01319445097211,
"learning_rate": 4.671422971866829e-08,
"logits/chosen": -1.5157580375671387,
"logits/rejected": -1.4674278497695923,
"logps/chosen": -769.1876220703125,
"logps/rejected": -776.2877197265625,
"loss": 0.5207,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.8552205562591553,
"rewards/margins": 0.6877336502075195,
"rewards/rejected": -3.542954444885254,
"step": 1012
},
{
"epoch": 0.6616050289819577,
"grad_norm": 88.36042423899895,
"learning_rate": 4.655582846646977e-08,
"logits/chosen": -1.4231505393981934,
"logits/rejected": -1.4381476640701294,
"logps/chosen": -826.669921875,
"logps/rejected": -876.0449829101562,
"loss": 0.5261,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.297783613204956,
"rewards/margins": 1.0941379070281982,
"rewards/rejected": -4.391921043395996,
"step": 1013
},
{
"epoch": 0.6622581435219201,
"grad_norm": 36.895557036458406,
"learning_rate": 4.6397575269914516e-08,
"logits/chosen": -1.5066965818405151,
"logits/rejected": -1.4606728553771973,
"logps/chosen": -728.8414306640625,
"logps/rejected": -851.2317504882812,
"loss": 0.4876,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7775468826293945,
"rewards/margins": 1.0805087089538574,
"rewards/rejected": -3.858055591583252,
"step": 1014
},
{
"epoch": 0.6629112580618826,
"grad_norm": 42.67356514700968,
"learning_rate": 4.6239470952731144e-08,
"logits/chosen": -1.4574358463287354,
"logits/rejected": -1.4584983587265015,
"logps/chosen": -798.5891723632812,
"logps/rejected": -916.482421875,
"loss": 0.5216,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7513608932495117,
"rewards/margins": 0.9783158898353577,
"rewards/rejected": -3.7296767234802246,
"step": 1015
},
{
"epoch": 0.6635643726018451,
"grad_norm": 103.36971446244533,
"learning_rate": 4.608151633787337e-08,
"logits/chosen": -1.4695475101470947,
"logits/rejected": -1.4544553756713867,
"logps/chosen": -803.3701171875,
"logps/rejected": -876.03369140625,
"loss": 0.4574,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9469823837280273,
"rewards/margins": 1.0371326208114624,
"rewards/rejected": -3.9841156005859375,
"step": 1016
},
{
"epoch": 0.6642174871418075,
"grad_norm": 53.366420843639595,
"learning_rate": 4.5923712247515675e-08,
"logits/chosen": -1.498404860496521,
"logits/rejected": -1.5131099224090576,
"logps/chosen": -852.6907958984375,
"logps/rejected": -1058.0159912109375,
"loss": 0.4487,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.187804698944092,
"rewards/margins": 1.617006778717041,
"rewards/rejected": -4.804811954498291,
"step": 1017
},
{
"epoch": 0.6648706016817699,
"grad_norm": 37.41870236327263,
"learning_rate": 4.5766059503049046e-08,
"logits/chosen": -1.485478401184082,
"logits/rejected": -1.501274824142456,
"logps/chosen": -788.52734375,
"logps/rejected": -801.43505859375,
"loss": 0.4773,
"rewards/accuracies": 0.625,
"rewards/chosen": -2.898345470428467,
"rewards/margins": 0.5790349245071411,
"rewards/rejected": -3.4773800373077393,
"step": 1018
},
{
"epoch": 0.6655237162217323,
"grad_norm": 36.10214381184078,
"learning_rate": 4.560855892507671e-08,
"logits/chosen": -1.5048822164535522,
"logits/rejected": -1.5141572952270508,
"logps/chosen": -830.2411499023438,
"logps/rejected": -920.7844848632812,
"loss": 0.5217,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3689942359924316,
"rewards/margins": 1.0679469108581543,
"rewards/rejected": -4.436941146850586,
"step": 1019
},
{
"epoch": 0.6661768307616949,
"grad_norm": 45.389879083458595,
"learning_rate": 4.5451211333409836e-08,
"logits/chosen": -1.4572731256484985,
"logits/rejected": -1.4573543071746826,
"logps/chosen": -800.939697265625,
"logps/rejected": -863.949951171875,
"loss": 0.5109,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.996425151824951,
"rewards/margins": 0.8323167562484741,
"rewards/rejected": -3.828742027282715,
"step": 1020
},
{
"epoch": 0.6668299453016573,
"grad_norm": 10.553936200597361,
"learning_rate": 4.5294017547063234e-08,
"logits/chosen": -1.5413297414779663,
"logits/rejected": -1.520641803741455,
"logps/chosen": -755.6485595703125,
"logps/rejected": -840.0621337890625,
"loss": 0.4168,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9807937145233154,
"rewards/margins": 1.0162277221679688,
"rewards/rejected": -3.9970216751098633,
"step": 1021
},
{
"epoch": 0.6674830598416197,
"grad_norm": 10.572170985494784,
"learning_rate": 4.513697838425122e-08,
"logits/chosen": -1.5927180051803589,
"logits/rejected": -1.6134157180786133,
"logps/chosen": -868.3167114257812,
"logps/rejected": -956.796875,
"loss": 0.5045,
"rewards/accuracies": 0.9375,
"rewards/chosen": -3.4438788890838623,
"rewards/margins": 1.0897140502929688,
"rewards/rejected": -4.533592700958252,
"step": 1022
},
{
"epoch": 0.6681361743815821,
"grad_norm": 20.81491049314488,
"learning_rate": 4.4980094662383206e-08,
"logits/chosen": -1.4901061058044434,
"logits/rejected": -1.4351348876953125,
"logps/chosen": -784.634765625,
"logps/rejected": -926.8602294921875,
"loss": 0.5169,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.037071704864502,
"rewards/margins": 1.026695966720581,
"rewards/rejected": -4.063767433166504,
"step": 1023
},
{
"epoch": 0.6687892889215447,
"grad_norm": 25.45661309370082,
"learning_rate": 4.4823367198059555e-08,
"logits/chosen": -1.513559103012085,
"logits/rejected": -1.5025659799575806,
"logps/chosen": -918.1492919921875,
"logps/rejected": -963.5728759765625,
"loss": 0.4967,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.492802381515503,
"rewards/margins": 0.798168957233429,
"rewards/rejected": -4.290971279144287,
"step": 1024
},
{
"epoch": 0.6694424034615071,
"grad_norm": 63.951136926876586,
"learning_rate": 4.466679680706727e-08,
"logits/chosen": -1.5798838138580322,
"logits/rejected": -1.4765055179595947,
"logps/chosen": -794.0283203125,
"logps/rejected": -891.6921997070312,
"loss": 0.5115,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.961698532104492,
"rewards/margins": 0.8770202994346619,
"rewards/rejected": -3.8387184143066406,
"step": 1025
},
{
"epoch": 0.6700955180014695,
"grad_norm": 57.70779325077905,
"learning_rate": 4.4510384304375773e-08,
"logits/chosen": -1.4599577188491821,
"logits/rejected": -1.5163075923919678,
"logps/chosen": -797.9346923828125,
"logps/rejected": -910.8660888671875,
"loss": 0.5434,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1370911598205566,
"rewards/margins": 0.9660441279411316,
"rewards/rejected": -4.103135108947754,
"step": 1026
},
{
"epoch": 0.6707486325414319,
"grad_norm": 160.10963139231885,
"learning_rate": 4.435413050413264e-08,
"logits/chosen": -1.5408748388290405,
"logits/rejected": -1.4838809967041016,
"logps/chosen": -880.136474609375,
"logps/rejected": -953.572509765625,
"loss": 0.496,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4785830974578857,
"rewards/margins": 0.9672165513038635,
"rewards/rejected": -4.445799350738525,
"step": 1027
},
{
"epoch": 0.6714017470813944,
"grad_norm": 30.04810670653747,
"learning_rate": 4.41980362196594e-08,
"logits/chosen": -1.5107932090759277,
"logits/rejected": -1.4906303882598877,
"logps/chosen": -787.1295166015625,
"logps/rejected": -931.1858520507812,
"loss": 0.5156,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.80479097366333,
"rewards/margins": 1.6828497648239136,
"rewards/rejected": -4.487640380859375,
"step": 1028
},
{
"epoch": 0.6720548616213569,
"grad_norm": 8.725100728079909,
"learning_rate": 4.4042102263447275e-08,
"logits/chosen": -1.4393929243087769,
"logits/rejected": -1.3799033164978027,
"logps/chosen": -751.6057739257812,
"logps/rejected": -883.99755859375,
"loss": 0.4621,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.040332317352295,
"rewards/margins": 1.2370002269744873,
"rewards/rejected": -4.277332782745361,
"step": 1029
},
{
"epoch": 0.6727079761613193,
"grad_norm": 10.666968801911986,
"learning_rate": 4.388632944715296e-08,
"logits/chosen": -1.474955677986145,
"logits/rejected": -1.4837840795516968,
"logps/chosen": -798.3235473632812,
"logps/rejected": -894.8926391601562,
"loss": 0.4808,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9682440757751465,
"rewards/margins": 0.7301187515258789,
"rewards/rejected": -3.6983628273010254,
"step": 1030
},
{
"epoch": 0.6733610907012817,
"grad_norm": 60.661967247074934,
"learning_rate": 4.37307185815944e-08,
"logits/chosen": -1.562424898147583,
"logits/rejected": -1.5712769031524658,
"logps/chosen": -810.6946411132812,
"logps/rejected": -843.1661987304688,
"loss": 0.5108,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2374746799468994,
"rewards/margins": 0.8579811453819275,
"rewards/rejected": -4.095455646514893,
"step": 1031
},
{
"epoch": 0.6740142052412442,
"grad_norm": 165.13796597577232,
"learning_rate": 4.3575270476746543e-08,
"logits/chosen": -1.4801855087280273,
"logits/rejected": -1.5016859769821167,
"logps/chosen": -857.4439697265625,
"logps/rejected": -974.478515625,
"loss": 0.515,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.03521990776062,
"rewards/margins": 0.830250084400177,
"rewards/rejected": -3.8654701709747314,
"step": 1032
},
{
"epoch": 0.6746673197812066,
"grad_norm": 97.99389725505694,
"learning_rate": 4.341998594173717e-08,
"logits/chosen": -1.4819018840789795,
"logits/rejected": -1.4765572547912598,
"logps/chosen": -779.4025268554688,
"logps/rejected": -837.5120239257812,
"loss": 0.4812,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9766929149627686,
"rewards/margins": 0.7156229019165039,
"rewards/rejected": -3.6923160552978516,
"step": 1033
},
{
"epoch": 0.6753204343211691,
"grad_norm": 24.808395132696045,
"learning_rate": 4.326486578484266e-08,
"logits/chosen": -1.4551912546157837,
"logits/rejected": -1.4724352359771729,
"logps/chosen": -780.5784301757812,
"logps/rejected": -889.181396484375,
"loss": 0.5369,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8798131942749023,
"rewards/margins": 0.8098124265670776,
"rewards/rejected": -3.6896255016326904,
"step": 1034
},
{
"epoch": 0.6759735488611315,
"grad_norm": 32.75717304070085,
"learning_rate": 4.310991081348376e-08,
"logits/chosen": -1.4434568881988525,
"logits/rejected": -1.4026532173156738,
"logps/chosen": -810.0265502929688,
"logps/rejected": -864.6695556640625,
"loss": 0.5423,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1554508209228516,
"rewards/margins": 0.6393953561782837,
"rewards/rejected": -3.7948460578918457,
"step": 1035
},
{
"epoch": 0.676626663401094,
"grad_norm": 95.45593023373124,
"learning_rate": 4.295512183422145e-08,
"logits/chosen": -1.4506603479385376,
"logits/rejected": -1.4620624780654907,
"logps/chosen": -711.5603637695312,
"logps/rejected": -851.3104858398438,
"loss": 0.5207,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.811810255050659,
"rewards/margins": 1.0253384113311768,
"rewards/rejected": -3.837148666381836,
"step": 1036
},
{
"epoch": 0.6772797779410564,
"grad_norm": 13.65496722074414,
"learning_rate": 4.280049965275261e-08,
"logits/chosen": -1.5078392028808594,
"logits/rejected": -1.4908549785614014,
"logps/chosen": -752.6651000976562,
"logps/rejected": -865.7678833007812,
"loss": 0.4159,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.94309663772583,
"rewards/margins": 0.8273873329162598,
"rewards/rejected": -3.77048397064209,
"step": 1037
},
{
"epoch": 0.6779328924810188,
"grad_norm": 23.66625717077645,
"learning_rate": 4.2646045073906e-08,
"logits/chosen": -1.4181371927261353,
"logits/rejected": -1.4423913955688477,
"logps/chosen": -730.9456176757812,
"logps/rejected": -790.7864990234375,
"loss": 0.4546,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9021615982055664,
"rewards/margins": 0.7823746800422668,
"rewards/rejected": -3.6845362186431885,
"step": 1038
},
{
"epoch": 0.6785860070209813,
"grad_norm": 38.74541697204639,
"learning_rate": 4.249175890163797e-08,
"logits/chosen": -1.531599521636963,
"logits/rejected": -1.5187830924987793,
"logps/chosen": -762.074951171875,
"logps/rejected": -845.5159912109375,
"loss": 0.4991,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8555479049682617,
"rewards/margins": 0.8805040121078491,
"rewards/rejected": -3.7360520362854004,
"step": 1039
},
{
"epoch": 0.6792391215609438,
"grad_norm": 85.38760955194438,
"learning_rate": 4.233764193902828e-08,
"logits/chosen": -1.547985553741455,
"logits/rejected": -1.5426965951919556,
"logps/chosen": -837.4592895507812,
"logps/rejected": -898.4917602539062,
"loss": 0.4809,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.8173162937164307,
"rewards/margins": 0.9812331795692444,
"rewards/rejected": -3.798549175262451,
"step": 1040
},
{
"epoch": 0.6798922361009062,
"grad_norm": 57.38295376577258,
"learning_rate": 4.2183694988275914e-08,
"logits/chosen": -1.5143656730651855,
"logits/rejected": -1.5319607257843018,
"logps/chosen": -795.1925048828125,
"logps/rejected": -825.0725708007812,
"loss": 0.5237,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.043870210647583,
"rewards/margins": 0.7048807144165039,
"rewards/rejected": -3.748751163482666,
"step": 1041
},
{
"epoch": 0.6805453506408686,
"grad_norm": 13.482382576237015,
"learning_rate": 4.2029918850694955e-08,
"logits/chosen": -1.5160479545593262,
"logits/rejected": -1.5341300964355469,
"logps/chosen": -728.4697875976562,
"logps/rejected": -833.2994384765625,
"loss": 0.516,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0110697746276855,
"rewards/margins": 0.9559275507926941,
"rewards/rejected": -3.9669976234436035,
"step": 1042
},
{
"epoch": 0.681198465180831,
"grad_norm": 13.93984327839821,
"learning_rate": 4.1876314326710367e-08,
"logits/chosen": -1.4439070224761963,
"logits/rejected": -1.4740190505981445,
"logps/chosen": -883.0184326171875,
"logps/rejected": -914.0108032226562,
"loss": 0.5132,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.1336121559143066,
"rewards/margins": 0.8373432159423828,
"rewards/rejected": -3.9709551334381104,
"step": 1043
},
{
"epoch": 0.6818515797207936,
"grad_norm": 93.33586037651152,
"learning_rate": 4.172288221585383e-08,
"logits/chosen": -1.4956300258636475,
"logits/rejected": -1.4777367115020752,
"logps/chosen": -819.32958984375,
"logps/rejected": -873.03955078125,
"loss": 0.5033,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.1138126850128174,
"rewards/margins": 0.7950230836868286,
"rewards/rejected": -3.9088354110717773,
"step": 1044
},
{
"epoch": 0.682504694260756,
"grad_norm": 24.399107508868735,
"learning_rate": 4.1569623316759634e-08,
"logits/chosen": -1.5415163040161133,
"logits/rejected": -1.5112028121948242,
"logps/chosen": -853.9760131835938,
"logps/rejected": -965.800537109375,
"loss": 0.5098,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.227302312850952,
"rewards/margins": 0.9138199687004089,
"rewards/rejected": -4.141121864318848,
"step": 1045
},
{
"epoch": 0.6831578088007184,
"grad_norm": 10.737613957729454,
"learning_rate": 4.1416538427160414e-08,
"logits/chosen": -1.508976936340332,
"logits/rejected": -1.4843047857284546,
"logps/chosen": -843.1103515625,
"logps/rejected": -909.9649658203125,
"loss": 0.5126,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.028240203857422,
"rewards/margins": 0.6666500568389893,
"rewards/rejected": -3.694890022277832,
"step": 1046
},
{
"epoch": 0.6838109233406808,
"grad_norm": 13.15409511965096,
"learning_rate": 4.126362834388311e-08,
"logits/chosen": -1.4586668014526367,
"logits/rejected": -1.450480580329895,
"logps/chosen": -823.8426513671875,
"logps/rejected": -831.0048828125,
"loss": 0.5012,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.807426691055298,
"rewards/margins": 0.6834732890129089,
"rewards/rejected": -3.4908998012542725,
"step": 1047
},
{
"epoch": 0.6844640378806434,
"grad_norm": 35.5691033339692,
"learning_rate": 4.11108938628448e-08,
"logits/chosen": -1.44548499584198,
"logits/rejected": -1.465252161026001,
"logps/chosen": -912.369140625,
"logps/rejected": -958.8450927734375,
"loss": 0.4735,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.652446746826172,
"rewards/margins": 0.8420040607452393,
"rewards/rejected": -4.494450569152832,
"step": 1048
},
{
"epoch": 0.6851171524206058,
"grad_norm": 55.79275790082709,
"learning_rate": 4.095833577904842e-08,
"logits/chosen": -1.5095837116241455,
"logits/rejected": -1.4833836555480957,
"logps/chosen": -766.8374633789062,
"logps/rejected": -825.0868530273438,
"loss": 0.4932,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.7053170204162598,
"rewards/margins": 0.8340969085693359,
"rewards/rejected": -3.5394136905670166,
"step": 1049
},
{
"epoch": 0.6857702669605682,
"grad_norm": 16.058246806646952,
"learning_rate": 4.0805954886578825e-08,
"logits/chosen": -1.4918146133422852,
"logits/rejected": -1.4603523015975952,
"logps/chosen": -789.8798217773438,
"logps/rejected": -847.2196655273438,
"loss": 0.5385,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.3735291957855225,
"rewards/margins": 0.4681362509727478,
"rewards/rejected": -3.841665744781494,
"step": 1050
},
{
"epoch": 0.6864233815005306,
"grad_norm": 25.38044398676047,
"learning_rate": 4.065375197859855e-08,
"logits/chosen": -1.5610326528549194,
"logits/rejected": -1.465526819229126,
"logps/chosen": -896.5455932617188,
"logps/rejected": -925.2203369140625,
"loss": 0.5014,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.476266622543335,
"rewards/margins": 0.7058266401290894,
"rewards/rejected": -4.182093620300293,
"step": 1051
},
{
"epoch": 0.6870764960404931,
"grad_norm": 27.871824279674087,
"learning_rate": 4.0501727847343706e-08,
"logits/chosen": -1.5372264385223389,
"logits/rejected": -1.521620512008667,
"logps/chosen": -966.5,
"logps/rejected": -982.9517822265625,
"loss": 0.5236,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.5799427032470703,
"rewards/margins": 0.6284856796264648,
"rewards/rejected": -4.208428382873535,
"step": 1052
},
{
"epoch": 0.6877296105804556,
"grad_norm": 58.973884464098276,
"learning_rate": 4.034988328411982e-08,
"logits/chosen": -1.4340426921844482,
"logits/rejected": -1.4606634378433228,
"logps/chosen": -855.3406372070312,
"logps/rejected": -920.906494140625,
"loss": 0.4464,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.8592772483825684,
"rewards/margins": 0.6489996314048767,
"rewards/rejected": -3.5082767009735107,
"step": 1053
},
{
"epoch": 0.688382725120418,
"grad_norm": 21.96563134850709,
"learning_rate": 4.0198219079297756e-08,
"logits/chosen": -1.4492591619491577,
"logits/rejected": -1.4421489238739014,
"logps/chosen": -826.43359375,
"logps/rejected": -959.9185180664062,
"loss": 0.5362,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1718485355377197,
"rewards/margins": 0.9998074769973755,
"rewards/rejected": -4.171655654907227,
"step": 1054
},
{
"epoch": 0.6890358396603804,
"grad_norm": 56.311096113880815,
"learning_rate": 4.004673602230961e-08,
"logits/chosen": -1.6020833253860474,
"logits/rejected": -1.5822880268096924,
"logps/chosen": -848.7802124023438,
"logps/rejected": -922.5611572265625,
"loss": 0.5243,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1218650341033936,
"rewards/margins": 0.7863477468490601,
"rewards/rejected": -3.908212661743164,
"step": 1055
},
{
"epoch": 0.6896889542003429,
"grad_norm": 17.48722643305911,
"learning_rate": 3.989543490164453e-08,
"logits/chosen": -1.4615482091903687,
"logits/rejected": -1.5211315155029297,
"logps/chosen": -854.398681640625,
"logps/rejected": -965.4825439453125,
"loss": 0.4684,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.055753231048584,
"rewards/margins": 0.7217795848846436,
"rewards/rejected": -3.7775330543518066,
"step": 1056
},
{
"epoch": 0.6903420687403053,
"grad_norm": 37.57075928693893,
"learning_rate": 3.974431650484468e-08,
"logits/chosen": -1.4383900165557861,
"logits/rejected": -1.3848916292190552,
"logps/chosen": -895.6553344726562,
"logps/rejected": -919.5377807617188,
"loss": 0.536,
"rewards/accuracies": 0.53125,
"rewards/chosen": -3.639094352722168,
"rewards/margins": 0.42604905366897583,
"rewards/rejected": -4.06514310836792,
"step": 1057
},
{
"epoch": 0.6909951832802678,
"grad_norm": 26.018606342144913,
"learning_rate": 3.95933816185012e-08,
"logits/chosen": -1.5629183053970337,
"logits/rejected": -1.528983235359192,
"logps/chosen": -896.5241088867188,
"logps/rejected": -974.3204345703125,
"loss": 0.5263,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3990674018859863,
"rewards/margins": 0.964748203754425,
"rewards/rejected": -4.363815784454346,
"step": 1058
},
{
"epoch": 0.6916482978202302,
"grad_norm": 137.70834817726936,
"learning_rate": 3.944263102824996e-08,
"logits/chosen": -1.4782737493515015,
"logits/rejected": -1.4820241928100586,
"logps/chosen": -844.2286987304688,
"logps/rejected": -879.0075073242188,
"loss": 0.4687,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.190678358078003,
"rewards/margins": 0.6480005979537964,
"rewards/rejected": -3.8386788368225098,
"step": 1059
},
{
"epoch": 0.6923014123601927,
"grad_norm": 135.63517874372047,
"learning_rate": 3.9292065518767495e-08,
"logits/chosen": -1.4735994338989258,
"logits/rejected": -1.461629033088684,
"logps/chosen": -815.2098388671875,
"logps/rejected": -904.7244873046875,
"loss": 0.5408,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.002563714981079,
"rewards/margins": 0.6686397194862366,
"rewards/rejected": -3.671203374862671,
"step": 1060
},
{
"epoch": 0.6929545269001551,
"grad_norm": 12.912624702236881,
"learning_rate": 3.914168587376706e-08,
"logits/chosen": -1.5482277870178223,
"logits/rejected": -1.5398210287094116,
"logps/chosen": -831.609130859375,
"logps/rejected": -957.0618286132812,
"loss": 0.494,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.257383108139038,
"rewards/margins": 0.9438177347183228,
"rewards/rejected": -4.201200485229492,
"step": 1061
},
{
"epoch": 0.6936076414401176,
"grad_norm": 19.51174100176863,
"learning_rate": 3.899149287599442e-08,
"logits/chosen": -1.4981271028518677,
"logits/rejected": -1.442700982093811,
"logps/chosen": -779.00927734375,
"logps/rejected": -808.9547729492188,
"loss": 0.5122,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.9186878204345703,
"rewards/margins": 0.47457852959632874,
"rewards/rejected": -3.3932666778564453,
"step": 1062
},
{
"epoch": 0.69426075598008,
"grad_norm": 71.96791759257684,
"learning_rate": 3.884148730722383e-08,
"logits/chosen": -1.5923326015472412,
"logits/rejected": -1.5967943668365479,
"logps/chosen": -838.6897583007812,
"logps/rejected": -877.890625,
"loss": 0.4842,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.1096723079681396,
"rewards/margins": 0.708088219165802,
"rewards/rejected": -3.817760467529297,
"step": 1063
},
{
"epoch": 0.6949138705200425,
"grad_norm": 16.78816497884493,
"learning_rate": 3.8691669948253964e-08,
"logits/chosen": -1.4334865808486938,
"logits/rejected": -1.4443024396896362,
"logps/chosen": -852.86572265625,
"logps/rejected": -981.8878173828125,
"loss": 0.483,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1702041625976562,
"rewards/margins": 1.1130633354187012,
"rewards/rejected": -4.283267974853516,
"step": 1064
},
{
"epoch": 0.6955669850600049,
"grad_norm": 19.842008322377158,
"learning_rate": 3.85420415789038e-08,
"logits/chosen": -1.4435492753982544,
"logits/rejected": -1.480208396911621,
"logps/chosen": -812.3143920898438,
"logps/rejected": -912.236572265625,
"loss": 0.475,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2056543827056885,
"rewards/margins": 0.799709141254425,
"rewards/rejected": -4.005363464355469,
"step": 1065
},
{
"epoch": 0.6962200995999673,
"grad_norm": 17.570135752432474,
"learning_rate": 3.839260297800864e-08,
"logits/chosen": -1.5509750843048096,
"logits/rejected": -1.542686939239502,
"logps/chosen": -918.3096923828125,
"logps/rejected": -1013.3355102539062,
"loss": 0.4643,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.234311103820801,
"rewards/margins": 1.1861119270324707,
"rewards/rejected": -4.42042350769043,
"step": 1066
},
{
"epoch": 0.6968732141399298,
"grad_norm": 13.333457525865867,
"learning_rate": 3.824335492341599e-08,
"logits/chosen": -1.5408436059951782,
"logits/rejected": -1.5043909549713135,
"logps/chosen": -823.61865234375,
"logps/rejected": -986.2550048828125,
"loss": 0.4637,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.8370096683502197,
"rewards/margins": 1.4054211378097534,
"rewards/rejected": -4.242431163787842,
"step": 1067
},
{
"epoch": 0.6975263286798923,
"grad_norm": 19.537814553814222,
"learning_rate": 3.8094298191981565e-08,
"logits/chosen": -1.4838995933532715,
"logits/rejected": -1.483154535293579,
"logps/chosen": -757.1932373046875,
"logps/rejected": -872.9207763671875,
"loss": 0.5317,
"rewards/accuracies": 0.96875,
"rewards/chosen": -2.696157217025757,
"rewards/margins": 1.4016393423080444,
"rewards/rejected": -4.09779691696167,
"step": 1068
},
{
"epoch": 0.6981794432198547,
"grad_norm": 9.688042362394432,
"learning_rate": 3.794543355956518e-08,
"logits/chosen": -1.4866911172866821,
"logits/rejected": -1.449821949005127,
"logps/chosen": -834.3885498046875,
"logps/rejected": -924.32470703125,
"loss": 0.4753,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.975139856338501,
"rewards/margins": 0.8105930685997009,
"rewards/rejected": -3.7857329845428467,
"step": 1069
},
{
"epoch": 0.6988325577598171,
"grad_norm": 13.451541283780028,
"learning_rate": 3.779676180102678e-08,
"logits/chosen": -1.5045047998428345,
"logits/rejected": -1.5032069683074951,
"logps/chosen": -714.1544189453125,
"logps/rejected": -732.0367431640625,
"loss": 0.5696,
"rewards/accuracies": 0.59375,
"rewards/chosen": -2.933642864227295,
"rewards/margins": 0.15253925323486328,
"rewards/rejected": -3.0861823558807373,
"step": 1070
},
{
"epoch": 0.6994856722997795,
"grad_norm": 12.44342691849441,
"learning_rate": 3.76482836902224e-08,
"logits/chosen": -1.572385311126709,
"logits/rejected": -1.5535383224487305,
"logps/chosen": -902.6881713867188,
"logps/rejected": -1089.9610595703125,
"loss": 0.4715,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.132718801498413,
"rewards/margins": 1.6019270420074463,
"rewards/rejected": -4.734645843505859,
"step": 1071
},
{
"epoch": 0.7001387868397421,
"grad_norm": 10.296422466002083,
"learning_rate": 3.750000000000002e-08,
"logits/chosen": -1.5108206272125244,
"logits/rejected": -1.477285623550415,
"logps/chosen": -779.607666015625,
"logps/rejected": -822.5366821289062,
"loss": 0.4943,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.6698358058929443,
"rewards/margins": 0.7314268946647644,
"rewards/rejected": -3.4012625217437744,
"step": 1072
},
{
"epoch": 0.7007919013797045,
"grad_norm": 19.95219987159032,
"learning_rate": 3.735191150219571e-08,
"logits/chosen": -1.5256969928741455,
"logits/rejected": -1.4763067960739136,
"logps/chosen": -837.349853515625,
"logps/rejected": -901.6431274414062,
"loss": 0.5116,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.106201648712158,
"rewards/margins": 0.7756415009498596,
"rewards/rejected": -3.881843328475952,
"step": 1073
},
{
"epoch": 0.7014450159196669,
"grad_norm": 35.78189318471355,
"learning_rate": 3.7204018967629534e-08,
"logits/chosen": -1.4913254976272583,
"logits/rejected": -1.523918628692627,
"logps/chosen": -725.7921142578125,
"logps/rejected": -901.6088256835938,
"loss": 0.496,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.752901315689087,
"rewards/margins": 1.2373815774917603,
"rewards/rejected": -3.9902822971343994,
"step": 1074
},
{
"epoch": 0.7020981304596293,
"grad_norm": 67.37326737845098,
"learning_rate": 3.7056323166101525e-08,
"logits/chosen": -1.5039902925491333,
"logits/rejected": -1.489875078201294,
"logps/chosen": -836.2699584960938,
"logps/rejected": -852.8065185546875,
"loss": 0.4636,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.870565891265869,
"rewards/margins": 0.6211225986480713,
"rewards/rejected": -3.4916884899139404,
"step": 1075
},
{
"epoch": 0.7027512449995919,
"grad_norm": 122.82544033335364,
"learning_rate": 3.690882486638771e-08,
"logits/chosen": -1.5150420665740967,
"logits/rejected": -1.4826247692108154,
"logps/chosen": -794.5693359375,
"logps/rejected": -939.4715576171875,
"loss": 0.4775,
"rewards/accuracies": 0.65625,
"rewards/chosen": -2.9887611865997314,
"rewards/margins": 1.1564222574234009,
"rewards/rejected": -4.145183563232422,
"step": 1076
},
{
"epoch": 0.7034043595395543,
"grad_norm": 29.770872322440407,
"learning_rate": 3.6761524836236085e-08,
"logits/chosen": -1.4067631959915161,
"logits/rejected": -1.3834965229034424,
"logps/chosen": -879.8606567382812,
"logps/rejected": -925.6470336914062,
"loss": 0.4719,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.328319787979126,
"rewards/margins": 0.6426284313201904,
"rewards/rejected": -3.9709482192993164,
"step": 1077
},
{
"epoch": 0.7040574740795167,
"grad_norm": 32.679662772650815,
"learning_rate": 3.6614423842362605e-08,
"logits/chosen": -1.4490833282470703,
"logits/rejected": -1.4676353931427002,
"logps/chosen": -785.7656860351562,
"logps/rejected": -959.332763671875,
"loss": 0.5489,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0860912799835205,
"rewards/margins": 0.828029453754425,
"rewards/rejected": -3.914120674133301,
"step": 1078
},
{
"epoch": 0.7047105886194791,
"grad_norm": 26.80045244487548,
"learning_rate": 3.646752265044725e-08,
"logits/chosen": -1.5650575160980225,
"logits/rejected": -1.5426418781280518,
"logps/chosen": -833.9342041015625,
"logps/rejected": -841.90283203125,
"loss": 0.4943,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.413437604904175,
"rewards/margins": 0.4831462800502777,
"rewards/rejected": -3.8965840339660645,
"step": 1079
},
{
"epoch": 0.7053637031594416,
"grad_norm": 128.50612923555713,
"learning_rate": 3.6320822025129986e-08,
"logits/chosen": -1.4550724029541016,
"logits/rejected": -1.4530410766601562,
"logps/chosen": -917.0040893554688,
"logps/rejected": -1088.79248046875,
"loss": 0.5566,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.368748903274536,
"rewards/margins": 1.5208771228790283,
"rewards/rejected": -4.889625549316406,
"step": 1080
},
{
"epoch": 0.706016817699404,
"grad_norm": 10.851702941973679,
"learning_rate": 3.6174322730006816e-08,
"logits/chosen": -1.5157017707824707,
"logits/rejected": -1.5042166709899902,
"logps/chosen": -775.1571044921875,
"logps/rejected": -932.432373046875,
"loss": 0.4836,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.8636438846588135,
"rewards/margins": 1.276392936706543,
"rewards/rejected": -4.1400370597839355,
"step": 1081
},
{
"epoch": 0.7066699322393665,
"grad_norm": 10.972406234728096,
"learning_rate": 3.6028025527625804e-08,
"logits/chosen": -1.513984203338623,
"logits/rejected": -1.451591968536377,
"logps/chosen": -830.9796142578125,
"logps/rejected": -924.893310546875,
"loss": 0.4536,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.566514015197754,
"rewards/margins": 0.943221390247345,
"rewards/rejected": -4.509735584259033,
"step": 1082
},
{
"epoch": 0.7073230467793289,
"grad_norm": 18.986648016914206,
"learning_rate": 3.588193117948301e-08,
"logits/chosen": -1.4808248281478882,
"logits/rejected": -1.5148813724517822,
"logps/chosen": -857.4286499023438,
"logps/rejected": -901.3809814453125,
"loss": 0.5308,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.144946575164795,
"rewards/margins": 0.7388260960578918,
"rewards/rejected": -3.883772611618042,
"step": 1083
},
{
"epoch": 0.7079761613192914,
"grad_norm": 45.59239900144737,
"learning_rate": 3.573604044601873e-08,
"logits/chosen": -1.4333078861236572,
"logits/rejected": -1.4482799768447876,
"logps/chosen": -845.7681884765625,
"logps/rejected": -967.3292846679688,
"loss": 0.4798,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3961985111236572,
"rewards/margins": 1.0223309993743896,
"rewards/rejected": -4.418529510498047,
"step": 1084
},
{
"epoch": 0.7086292758592538,
"grad_norm": 13.726428268406318,
"learning_rate": 3.559035408661334e-08,
"logits/chosen": -1.4743175506591797,
"logits/rejected": -1.467435598373413,
"logps/chosen": -852.9383544921875,
"logps/rejected": -932.0983276367188,
"loss": 0.4883,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2556493282318115,
"rewards/margins": 1.4109907150268555,
"rewards/rejected": -4.666640281677246,
"step": 1085
},
{
"epoch": 0.7092823903992163,
"grad_norm": 24.407459386863895,
"learning_rate": 3.544487285958346e-08,
"logits/chosen": -1.599566102027893,
"logits/rejected": -1.5549285411834717,
"logps/chosen": -739.4046020507812,
"logps/rejected": -724.448486328125,
"loss": 0.5617,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.502041816711426,
"rewards/margins": 0.5174904465675354,
"rewards/rejected": -3.0195322036743164,
"step": 1086
},
{
"epoch": 0.7099355049391787,
"grad_norm": 23.319894013252863,
"learning_rate": 3.5299597522177944e-08,
"logits/chosen": -1.571434497833252,
"logits/rejected": -1.5202834606170654,
"logps/chosen": -872.5787353515625,
"logps/rejected": -882.1563720703125,
"loss": 0.5185,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.388620138168335,
"rewards/margins": 0.7215232253074646,
"rewards/rejected": -4.110143184661865,
"step": 1087
},
{
"epoch": 0.7105886194791412,
"grad_norm": 83.13631229824816,
"learning_rate": 3.5154528830574e-08,
"logits/chosen": -1.5138779878616333,
"logits/rejected": -1.5081030130386353,
"logps/chosen": -740.7789916992188,
"logps/rejected": -852.57421875,
"loss": 0.4798,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.757505416870117,
"rewards/margins": 1.019026756286621,
"rewards/rejected": -3.7765321731567383,
"step": 1088
},
{
"epoch": 0.7112417340191036,
"grad_norm": 15.401125482802248,
"learning_rate": 3.500966753987317e-08,
"logits/chosen": -1.4526937007904053,
"logits/rejected": -1.4075860977172852,
"logps/chosen": -769.2919311523438,
"logps/rejected": -886.1649169921875,
"loss": 0.482,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.7922770977020264,
"rewards/margins": 0.9756352305412292,
"rewards/rejected": -3.7679123878479004,
"step": 1089
},
{
"epoch": 0.711894848559066,
"grad_norm": 22.983567681134193,
"learning_rate": 3.486501440409748e-08,
"logits/chosen": -1.5759086608886719,
"logits/rejected": -1.5538822412490845,
"logps/chosen": -872.93896484375,
"logps/rejected": -946.4339599609375,
"loss": 0.4708,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2252018451690674,
"rewards/margins": 0.8281301856040955,
"rewards/rejected": -4.0533318519592285,
"step": 1090
},
{
"epoch": 0.7125479630990285,
"grad_norm": 25.263230007411096,
"learning_rate": 3.472057017618547e-08,
"logits/chosen": -1.3890550136566162,
"logits/rejected": -1.3870556354522705,
"logps/chosen": -866.8782958984375,
"logps/rejected": -1034.20458984375,
"loss": 0.4849,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3459925651550293,
"rewards/margins": 1.1013741493225098,
"rewards/rejected": -4.447366237640381,
"step": 1091
},
{
"epoch": 0.713201077638991,
"grad_norm": 23.164092211586546,
"learning_rate": 3.4576335607988294e-08,
"logits/chosen": -1.556492567062378,
"logits/rejected": -1.5247762203216553,
"logps/chosen": -825.4620971679688,
"logps/rejected": -909.668701171875,
"loss": 0.452,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.7882065773010254,
"rewards/margins": 1.1564209461212158,
"rewards/rejected": -3.944627523422241,
"step": 1092
},
{
"epoch": 0.7138541921789534,
"grad_norm": 30.491984391282454,
"learning_rate": 3.44323114502658e-08,
"logits/chosen": -1.4933907985687256,
"logits/rejected": -1.4707281589508057,
"logps/chosen": -831.043701171875,
"logps/rejected": -888.13037109375,
"loss": 0.5366,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.5473837852478027,
"rewards/margins": 0.7864856719970703,
"rewards/rejected": -4.333869934082031,
"step": 1093
},
{
"epoch": 0.7145073067189158,
"grad_norm": 26.466125427075724,
"learning_rate": 3.42884984526826e-08,
"logits/chosen": -1.4837186336517334,
"logits/rejected": -1.4462792873382568,
"logps/chosen": -932.162353515625,
"logps/rejected": -1006.1118774414062,
"loss": 0.5344,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4717721939086914,
"rewards/margins": 0.9058358073234558,
"rewards/rejected": -4.377608299255371,
"step": 1094
},
{
"epoch": 0.7151604212588782,
"grad_norm": 22.114835756620792,
"learning_rate": 3.414489736380423e-08,
"logits/chosen": -1.5490305423736572,
"logits/rejected": -1.528594970703125,
"logps/chosen": -765.060791015625,
"logps/rejected": -864.4068603515625,
"loss": 0.4142,
"rewards/accuracies": 0.9375,
"rewards/chosen": -2.7475967407226562,
"rewards/margins": 1.1565359830856323,
"rewards/rejected": -3.904132843017578,
"step": 1095
},
{
"epoch": 0.7158135357988408,
"grad_norm": 13.258696810534204,
"learning_rate": 3.400150893109317e-08,
"logits/chosen": -1.4212427139282227,
"logits/rejected": -1.3869833946228027,
"logps/chosen": -761.0309448242188,
"logps/rejected": -820.0980224609375,
"loss": 0.4784,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.5319325923919678,
"rewards/margins": 0.7716701626777649,
"rewards/rejected": -4.30360221862793,
"step": 1096
},
{
"epoch": 0.7164666503388032,
"grad_norm": 153.0170543549745,
"learning_rate": 3.385833390090502e-08,
"logits/chosen": -1.4919387102127075,
"logits/rejected": -1.4864999055862427,
"logps/chosen": -876.7687377929688,
"logps/rejected": -904.2010498046875,
"loss": 0.544,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.105478525161743,
"rewards/margins": 0.4873059093952179,
"rewards/rejected": -3.5927841663360596,
"step": 1097
},
{
"epoch": 0.7171197648787656,
"grad_norm": 22.73385039619432,
"learning_rate": 3.3715373018484606e-08,
"logits/chosen": -1.5606061220169067,
"logits/rejected": -1.5803030729293823,
"logps/chosen": -864.9635009765625,
"logps/rejected": -979.0385131835938,
"loss": 0.5134,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1329944133758545,
"rewards/margins": 1.058984637260437,
"rewards/rejected": -4.19197940826416,
"step": 1098
},
{
"epoch": 0.717772879418728,
"grad_norm": 47.2766708425411,
"learning_rate": 3.357262702796206e-08,
"logits/chosen": -1.5417003631591797,
"logits/rejected": -1.5057967901229858,
"logps/chosen": -882.0269775390625,
"logps/rejected": -921.403076171875,
"loss": 0.4704,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.9814200401306152,
"rewards/margins": 0.637610912322998,
"rewards/rejected": -3.619030714035034,
"step": 1099
},
{
"epoch": 0.7184259939586906,
"grad_norm": 17.039889063919627,
"learning_rate": 3.343009667234898e-08,
"logits/chosen": -1.4841748476028442,
"logits/rejected": -1.5299304723739624,
"logps/chosen": -809.2041625976562,
"logps/rejected": -876.6251831054688,
"loss": 0.4307,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.4067025184631348,
"rewards/margins": 0.9841042757034302,
"rewards/rejected": -4.390806674957275,
"step": 1100
},
{
"epoch": 0.7184259939586906,
"eval_logits/chosen": -1.4911024570465088,
"eval_logits/rejected": -1.4728410243988037,
"eval_logps/chosen": -824.2158813476562,
"eval_logps/rejected": -903.2113037109375,
"eval_loss": 0.498925119638443,
"eval_rewards/accuracies": 0.7609999775886536,
"eval_rewards/chosen": -3.1387054920196533,
"eval_rewards/margins": 0.8710131645202637,
"eval_rewards/rejected": -4.009718418121338,
"eval_runtime": 300.1149,
"eval_samples_per_second": 13.328,
"eval_steps_per_second": 0.833,
"step": 1100
},
{
"epoch": 0.719079108498653,
"grad_norm": 16.11848033889973,
"learning_rate": 3.3287782693534566e-08,
"logits/chosen": -1.4810887575149536,
"logits/rejected": -1.4817496538162231,
"logps/chosen": -845.0581665039062,
"logps/rejected": -940.9647827148438,
"loss": 0.5442,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.3775806427001953,
"rewards/margins": 0.7508167624473572,
"rewards/rejected": -4.128397464752197,
"step": 1101
},
{
"epoch": 0.7197322230386154,
"grad_norm": 68.65354889922948,
"learning_rate": 3.3145685832281736e-08,
"logits/chosen": -1.5042979717254639,
"logits/rejected": -1.523160457611084,
"logps/chosen": -804.5962524414062,
"logps/rejected": -879.611572265625,
"loss": 0.437,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.8270790576934814,
"rewards/margins": 0.7390732765197754,
"rewards/rejected": -3.566152572631836,
"step": 1102
},
{
"epoch": 0.7203853375785778,
"grad_norm": 73.57913188242289,
"learning_rate": 3.30038068282233e-08,
"logits/chosen": -1.4858269691467285,
"logits/rejected": -1.440920352935791,
"logps/chosen": -834.6658935546875,
"logps/rejected": -878.7103271484375,
"loss": 0.4838,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6924118995666504,
"rewards/margins": 0.7402853965759277,
"rewards/rejected": -4.432697772979736,
"step": 1103
},
{
"epoch": 0.7210384521185403,
"grad_norm": 13.062198655888952,
"learning_rate": 3.286214641985807e-08,
"logits/chosen": -1.5750617980957031,
"logits/rejected": -1.491349458694458,
"logps/chosen": -816.414306640625,
"logps/rejected": -881.9208984375,
"loss": 0.4839,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.108093738555908,
"rewards/margins": 0.7648839950561523,
"rewards/rejected": -3.8729777336120605,
"step": 1104
},
{
"epoch": 0.7216915666585028,
"grad_norm": 14.08459741772447,
"learning_rate": 3.272070534454708e-08,
"logits/chosen": -1.5048151016235352,
"logits/rejected": -1.4695175886154175,
"logps/chosen": -810.2633056640625,
"logps/rejected": -856.9027099609375,
"loss": 0.4868,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.639930486679077,
"rewards/margins": 0.8205062747001648,
"rewards/rejected": -3.4604365825653076,
"step": 1105
},
{
"epoch": 0.7223446811984652,
"grad_norm": 9.792790670749572,
"learning_rate": 3.2579484338509616e-08,
"logits/chosen": -1.441813349723816,
"logits/rejected": -1.4555586576461792,
"logps/chosen": -759.58154296875,
"logps/rejected": -834.8800659179688,
"loss": 0.4439,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1834728717803955,
"rewards/margins": 0.8506563901901245,
"rewards/rejected": -4.0341291427612305,
"step": 1106
},
{
"epoch": 0.7229977957384276,
"grad_norm": 13.42092316503716,
"learning_rate": 3.2438484136819575e-08,
"logits/chosen": -1.5256390571594238,
"logits/rejected": -1.5091501474380493,
"logps/chosen": -882.526123046875,
"logps/rejected": -992.293212890625,
"loss": 0.4902,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3528499603271484,
"rewards/margins": 1.2070109844207764,
"rewards/rejected": -4.559861183166504,
"step": 1107
},
{
"epoch": 0.7236509102783901,
"grad_norm": 36.798480052905624,
"learning_rate": 3.22977054734015e-08,
"logits/chosen": -1.612870693206787,
"logits/rejected": -1.6026808023452759,
"logps/chosen": -824.7280883789062,
"logps/rejected": -989.9205322265625,
"loss": 0.5238,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1623008251190186,
"rewards/margins": 1.1107250452041626,
"rewards/rejected": -4.273025989532471,
"step": 1108
},
{
"epoch": 0.7243040248183525,
"grad_norm": 15.657306421690263,
"learning_rate": 3.215714908102678e-08,
"logits/chosen": -1.4942344427108765,
"logits/rejected": -1.447746753692627,
"logps/chosen": -790.576904296875,
"logps/rejected": -844.633056640625,
"loss": 0.5197,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0599842071533203,
"rewards/margins": 0.9251794815063477,
"rewards/rejected": -3.985163450241089,
"step": 1109
},
{
"epoch": 0.724957139358315,
"grad_norm": 10.52893767352228,
"learning_rate": 3.201681569130988e-08,
"logits/chosen": -1.4922351837158203,
"logits/rejected": -1.4611746072769165,
"logps/chosen": -874.3091430664062,
"logps/rejected": -943.7155151367188,
"loss": 0.4354,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.2651565074920654,
"rewards/margins": 0.8614932894706726,
"rewards/rejected": -4.126649379730225,
"step": 1110
},
{
"epoch": 0.7256102538982774,
"grad_norm": 54.5450174448529,
"learning_rate": 3.187670603470451e-08,
"logits/chosen": -1.474999189376831,
"logits/rejected": -1.5109246969223022,
"logps/chosen": -883.6320190429688,
"logps/rejected": -954.03369140625,
"loss": 0.4765,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4104151725769043,
"rewards/margins": 0.9843173027038574,
"rewards/rejected": -4.394732475280762,
"step": 1111
},
{
"epoch": 0.7262633684382399,
"grad_norm": 59.388463871448565,
"learning_rate": 3.173682084049979e-08,
"logits/chosen": -1.4505696296691895,
"logits/rejected": -1.458186388015747,
"logps/chosen": -664.27587890625,
"logps/rejected": -750.30078125,
"loss": 0.4635,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.536130428314209,
"rewards/margins": 0.9553031921386719,
"rewards/rejected": -3.49143385887146,
"step": 1112
},
{
"epoch": 0.7269164829782023,
"grad_norm": 137.01606102320702,
"learning_rate": 3.159716083681652e-08,
"logits/chosen": -1.6230467557907104,
"logits/rejected": -1.5941567420959473,
"logps/chosen": -926.1805419921875,
"logps/rejected": -998.4340209960938,
"loss": 0.5433,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.0903451442718506,
"rewards/margins": 1.0845158100128174,
"rewards/rejected": -4.174860954284668,
"step": 1113
},
{
"epoch": 0.7275695975181647,
"grad_norm": 29.04184199104772,
"learning_rate": 3.1457726750603317e-08,
"logits/chosen": -1.512034296989441,
"logits/rejected": -1.47452712059021,
"logps/chosen": -846.271728515625,
"logps/rejected": -941.9117431640625,
"loss": 0.4757,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.4414303302764893,
"rewards/margins": 1.0037970542907715,
"rewards/rejected": -4.44522762298584,
"step": 1114
},
{
"epoch": 0.7282227120581272,
"grad_norm": 67.67268164595161,
"learning_rate": 3.131851930763289e-08,
"logits/chosen": -1.51804518699646,
"logits/rejected": -1.497375726699829,
"logps/chosen": -890.821533203125,
"logps/rejected": -1002.13916015625,
"loss": 0.4851,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.418884515762329,
"rewards/margins": 0.7488983869552612,
"rewards/rejected": -4.167783260345459,
"step": 1115
},
{
"epoch": 0.7288758265980897,
"grad_norm": 20.304379891765954,
"learning_rate": 3.1179539232498276e-08,
"logits/chosen": -1.4011988639831543,
"logits/rejected": -1.400657296180725,
"logps/chosen": -691.755615234375,
"logps/rejected": -767.6619873046875,
"loss": 0.5219,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.7513551712036133,
"rewards/margins": 0.778099775314331,
"rewards/rejected": -3.5294547080993652,
"step": 1116
},
{
"epoch": 0.7295289411380521,
"grad_norm": 11.565056384039993,
"learning_rate": 3.104078724860892e-08,
"logits/chosen": -1.6112922430038452,
"logits/rejected": -1.5689594745635986,
"logps/chosen": -833.8433837890625,
"logps/rejected": -893.80615234375,
"loss": 0.514,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4891977310180664,
"rewards/margins": 0.7067195177078247,
"rewards/rejected": -4.195917129516602,
"step": 1117
},
{
"epoch": 0.7301820556780145,
"grad_norm": 40.26811950012431,
"learning_rate": 3.090226407818714e-08,
"logits/chosen": -1.5029704570770264,
"logits/rejected": -1.4893207550048828,
"logps/chosen": -858.901611328125,
"logps/rejected": -917.4168090820312,
"loss": 0.5033,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3486008644104004,
"rewards/margins": 0.908878743648529,
"rewards/rejected": -4.257479667663574,
"step": 1118
},
{
"epoch": 0.730835170217977,
"grad_norm": 16.207319176406262,
"learning_rate": 3.07639704422642e-08,
"logits/chosen": -1.4990010261535645,
"logits/rejected": -1.4996980428695679,
"logps/chosen": -875.76904296875,
"logps/rejected": -1025.61767578125,
"loss": 0.5226,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.398207902908325,
"rewards/margins": 1.0922547578811646,
"rewards/rejected": -4.490462303161621,
"step": 1119
},
{
"epoch": 0.7314882847579395,
"grad_norm": 33.13246656200166,
"learning_rate": 3.06259070606766e-08,
"logits/chosen": -1.4612716436386108,
"logits/rejected": -1.4551949501037598,
"logps/chosen": -897.0416259765625,
"logps/rejected": -981.119873046875,
"loss": 0.491,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.664212465286255,
"rewards/margins": 0.6127184629440308,
"rewards/rejected": -4.276930809020996,
"step": 1120
},
{
"epoch": 0.7321413992979019,
"grad_norm": 27.787496208253557,
"learning_rate": 3.048807465206237e-08,
"logits/chosen": -1.4162070751190186,
"logits/rejected": -1.4611150026321411,
"logps/chosen": -788.558349609375,
"logps/rejected": -879.7010498046875,
"loss": 0.5043,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.006227493286133,
"rewards/margins": 1.1710853576660156,
"rewards/rejected": -4.177312850952148,
"step": 1121
},
{
"epoch": 0.7327945138378643,
"grad_norm": 31.88982746049559,
"learning_rate": 3.035047393385725e-08,
"logits/chosen": -1.5180845260620117,
"logits/rejected": -1.5133084058761597,
"logps/chosen": -882.9698486328125,
"logps/rejected": -901.154052734375,
"loss": 0.5017,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.381559371948242,
"rewards/margins": 0.6179601550102234,
"rewards/rejected": -3.9995198249816895,
"step": 1122
},
{
"epoch": 0.7334476283778267,
"grad_norm": 14.275386018313844,
"learning_rate": 3.021310562229105e-08,
"logits/chosen": -1.424015998840332,
"logits/rejected": -1.391045331954956,
"logps/chosen": -808.16796875,
"logps/rejected": -890.8568115234375,
"loss": 0.4597,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.398470163345337,
"rewards/margins": 0.7546467781066895,
"rewards/rejected": -4.1531171798706055,
"step": 1123
},
{
"epoch": 0.7341007429177893,
"grad_norm": 31.589588156255736,
"learning_rate": 3.0075970432383824e-08,
"logits/chosen": -1.4777663946151733,
"logits/rejected": -1.4628345966339111,
"logps/chosen": -788.1425170898438,
"logps/rejected": -1018.20751953125,
"loss": 0.4613,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.831998348236084,
"rewards/margins": 1.624790906906128,
"rewards/rejected": -4.456789016723633,
"step": 1124
},
{
"epoch": 0.7347538574577517,
"grad_norm": 23.163673831394075,
"learning_rate": 2.993906907794223e-08,
"logits/chosen": -1.4283198118209839,
"logits/rejected": -1.4135866165161133,
"logps/chosen": -848.6114501953125,
"logps/rejected": -875.2060546875,
"loss": 0.4822,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.065795660018921,
"rewards/margins": 0.7749052047729492,
"rewards/rejected": -3.84070086479187,
"step": 1125
},
{
"epoch": 0.7354069719977141,
"grad_norm": 66.36390765427439,
"learning_rate": 2.980240227155578e-08,
"logits/chosen": -1.4410760402679443,
"logits/rejected": -1.4437801837921143,
"logps/chosen": -856.7933959960938,
"logps/rejected": -962.9047241210938,
"loss": 0.5124,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.395413875579834,
"rewards/margins": 0.9301176071166992,
"rewards/rejected": -4.325531959533691,
"step": 1126
},
{
"epoch": 0.7360600865376765,
"grad_norm": 81.45798747264044,
"learning_rate": 2.9665970724593113e-08,
"logits/chosen": -1.5262151956558228,
"logits/rejected": -1.4462474584579468,
"logps/chosen": -931.7483520507812,
"logps/rejected": -1020.103271484375,
"loss": 0.515,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.550004720687866,
"rewards/margins": 1.2063486576080322,
"rewards/rejected": -4.756353378295898,
"step": 1127
},
{
"epoch": 0.736713201077639,
"grad_norm": 35.012242513097036,
"learning_rate": 2.9529775147198323e-08,
"logits/chosen": -1.5010284185409546,
"logits/rejected": -1.511681079864502,
"logps/chosen": -828.48583984375,
"logps/rejected": -907.3237915039062,
"loss": 0.4117,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.0704474449157715,
"rewards/margins": 1.052125096321106,
"rewards/rejected": -4.122572898864746,
"step": 1128
},
{
"epoch": 0.7373663156176015,
"grad_norm": 12.390653146331033,
"learning_rate": 2.9393816248287257e-08,
"logits/chosen": -1.4855061769485474,
"logits/rejected": -1.5361738204956055,
"logps/chosen": -775.18359375,
"logps/rejected": -859.5416259765625,
"loss": 0.4782,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.886140823364258,
"rewards/margins": 0.9194518327713013,
"rewards/rejected": -3.8055927753448486,
"step": 1129
},
{
"epoch": 0.7380194301575639,
"grad_norm": 41.60514101169407,
"learning_rate": 2.925809473554382e-08,
"logits/chosen": -1.4579628705978394,
"logits/rejected": -1.4414145946502686,
"logps/chosen": -883.7291870117188,
"logps/rejected": -980.4385375976562,
"loss": 0.505,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3500232696533203,
"rewards/margins": 0.7552958726882935,
"rewards/rejected": -4.105319023132324,
"step": 1130
},
{
"epoch": 0.7386725446975263,
"grad_norm": 23.290126695482954,
"learning_rate": 2.9122611315416283e-08,
"logits/chosen": -1.5425716638565063,
"logits/rejected": -1.511645793914795,
"logps/chosen": -785.6270141601562,
"logps/rejected": -796.7711181640625,
"loss": 0.4878,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.8701109886169434,
"rewards/margins": 0.560005784034729,
"rewards/rejected": -3.430117130279541,
"step": 1131
},
{
"epoch": 0.7393256592374887,
"grad_norm": 14.618557708069593,
"learning_rate": 2.898736669311361e-08,
"logits/chosen": -1.5095089673995972,
"logits/rejected": -1.424134373664856,
"logps/chosen": -893.7913818359375,
"logps/rejected": -921.09912109375,
"loss": 0.5096,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3666210174560547,
"rewards/margins": 1.207306981086731,
"rewards/rejected": -4.573927879333496,
"step": 1132
},
{
"epoch": 0.7399787737774512,
"grad_norm": 25.284769898933742,
"learning_rate": 2.8852361572601798e-08,
"logits/chosen": -1.4017362594604492,
"logits/rejected": -1.4303860664367676,
"logps/chosen": -785.92041015625,
"logps/rejected": -907.3682861328125,
"loss": 0.4776,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.0821828842163086,
"rewards/margins": 1.0643155574798584,
"rewards/rejected": -4.146498680114746,
"step": 1133
},
{
"epoch": 0.7406318883174137,
"grad_norm": 83.88894947610976,
"learning_rate": 2.8717596656600207e-08,
"logits/chosen": -1.4224882125854492,
"logits/rejected": -1.4386307001113892,
"logps/chosen": -819.8443603515625,
"logps/rejected": -834.15234375,
"loss": 0.6099,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.7083494663238525,
"rewards/margins": 0.4391818642616272,
"rewards/rejected": -4.147531509399414,
"step": 1134
},
{
"epoch": 0.7412850028573761,
"grad_norm": 86.50212048050254,
"learning_rate": 2.8583072646577904e-08,
"logits/chosen": -1.5508708953857422,
"logits/rejected": -1.5168962478637695,
"logps/chosen": -878.55712890625,
"logps/rejected": -941.9532470703125,
"loss": 0.4686,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2938575744628906,
"rewards/margins": 0.9149394631385803,
"rewards/rejected": -4.208796977996826,
"step": 1135
},
{
"epoch": 0.7419381173973385,
"grad_norm": 46.81398792643234,
"learning_rate": 2.8448790242750002e-08,
"logits/chosen": -1.4926443099975586,
"logits/rejected": -1.4884159564971924,
"logps/chosen": -827.5214233398438,
"logps/rejected": -890.3485107421875,
"loss": 0.5022,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.1017537117004395,
"rewards/margins": 0.6141247749328613,
"rewards/rejected": -3.715878486633301,
"step": 1136
},
{
"epoch": 0.742591231937301,
"grad_norm": 73.85737704941202,
"learning_rate": 2.831475014407402e-08,
"logits/chosen": -1.5685663223266602,
"logits/rejected": -1.5168462991714478,
"logps/chosen": -825.589599609375,
"logps/rejected": -906.937744140625,
"loss": 0.5298,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.318887948989868,
"rewards/margins": 1.1284462213516235,
"rewards/rejected": -4.447334289550781,
"step": 1137
},
{
"epoch": 0.7432443464772635,
"grad_norm": 86.46807131648339,
"learning_rate": 2.8180953048246247e-08,
"logits/chosen": -1.4860353469848633,
"logits/rejected": -1.468380093574524,
"logps/chosen": -782.5503540039062,
"logps/rejected": -827.2178344726562,
"loss": 0.4897,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.859600067138672,
"rewards/margins": 0.6450478434562683,
"rewards/rejected": -3.5046474933624268,
"step": 1138
},
{
"epoch": 0.7438974610172259,
"grad_norm": 84.46407466008054,
"learning_rate": 2.8047399651698154e-08,
"logits/chosen": -1.4216095209121704,
"logits/rejected": -1.4595965147018433,
"logps/chosen": -844.6117553710938,
"logps/rejected": -825.7047729492188,
"loss": 0.4965,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.3110671043395996,
"rewards/margins": 0.38550424575805664,
"rewards/rejected": -3.6965713500976562,
"step": 1139
},
{
"epoch": 0.7445505755571883,
"grad_norm": 55.02374788341525,
"learning_rate": 2.791409064959262e-08,
"logits/chosen": -1.4333022832870483,
"logits/rejected": -1.3619767427444458,
"logps/chosen": -795.4210205078125,
"logps/rejected": -872.92333984375,
"loss": 0.4726,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1887168884277344,
"rewards/margins": 0.8975101113319397,
"rewards/rejected": -4.086226940155029,
"step": 1140
},
{
"epoch": 0.7452036900971508,
"grad_norm": 87.71895948569228,
"learning_rate": 2.7781026735820516e-08,
"logits/chosen": -1.5040805339813232,
"logits/rejected": -1.47743821144104,
"logps/chosen": -840.98583984375,
"logps/rejected": -963.7841796875,
"loss": 0.4646,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2599844932556152,
"rewards/margins": 1.1233594417572021,
"rewards/rejected": -4.3833441734313965,
"step": 1141
},
{
"epoch": 0.7458568046371132,
"grad_norm": 19.67396573430194,
"learning_rate": 2.7648208602996965e-08,
"logits/chosen": -1.5113935470581055,
"logits/rejected": -1.4451313018798828,
"logps/chosen": -839.17578125,
"logps/rejected": -976.3116455078125,
"loss": 0.4397,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.220781087875366,
"rewards/margins": 1.0840184688568115,
"rewards/rejected": -4.304799556732178,
"step": 1142
},
{
"epoch": 0.7465099191770757,
"grad_norm": 84.55591225773897,
"learning_rate": 2.751563694245776e-08,
"logits/chosen": -1.480290412902832,
"logits/rejected": -1.5081285238265991,
"logps/chosen": -809.2010498046875,
"logps/rejected": -877.5315551757812,
"loss": 0.4513,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.9877707958221436,
"rewards/margins": 0.8733527660369873,
"rewards/rejected": -3.8611233234405518,
"step": 1143
},
{
"epoch": 0.7471630337170381,
"grad_norm": 95.02644629261052,
"learning_rate": 2.7383312444255792e-08,
"logits/chosen": -1.5553948879241943,
"logits/rejected": -1.5468087196350098,
"logps/chosen": -876.2579345703125,
"logps/rejected": -1008.3162231445312,
"loss": 0.4952,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.5550930500030518,
"rewards/margins": 1.053817629814148,
"rewards/rejected": -4.60891056060791,
"step": 1144
},
{
"epoch": 0.7478161482570006,
"grad_norm": 81.78452279244114,
"learning_rate": 2.7251235797157426e-08,
"logits/chosen": -1.5420262813568115,
"logits/rejected": -1.5177980661392212,
"logps/chosen": -839.2437133789062,
"logps/rejected": -872.8123779296875,
"loss": 0.5823,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.291133403778076,
"rewards/margins": 0.3555396795272827,
"rewards/rejected": -3.6466729640960693,
"step": 1145
},
{
"epoch": 0.748469262796963,
"grad_norm": 18.413788007685703,
"learning_rate": 2.7119407688638925e-08,
"logits/chosen": -1.4686810970306396,
"logits/rejected": -1.434198021888733,
"logps/chosen": -887.4147338867188,
"logps/rejected": -914.2786865234375,
"loss": 0.5467,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.379523754119873,
"rewards/margins": 0.6430516839027405,
"rewards/rejected": -4.022575378417969,
"step": 1146
},
{
"epoch": 0.7491223773369254,
"grad_norm": 25.04209084197724,
"learning_rate": 2.6987828804882885e-08,
"logits/chosen": -1.473648190498352,
"logits/rejected": -1.4673340320587158,
"logps/chosen": -755.6442260742188,
"logps/rejected": -806.13818359375,
"loss": 0.5137,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.124025344848633,
"rewards/margins": 0.8462156653404236,
"rewards/rejected": -3.970241069793701,
"step": 1147
},
{
"epoch": 0.7497754918768879,
"grad_norm": 17.831323248171984,
"learning_rate": 2.6856499830774655e-08,
"logits/chosen": -1.5108214616775513,
"logits/rejected": -1.4946174621582031,
"logps/chosen": -833.5274658203125,
"logps/rejected": -902.3148193359375,
"loss": 0.533,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2335917949676514,
"rewards/margins": 1.1803102493286133,
"rewards/rejected": -4.413901329040527,
"step": 1148
},
{
"epoch": 0.7504286064168504,
"grad_norm": 55.940688397054416,
"learning_rate": 2.6725421449898775e-08,
"logits/chosen": -1.5316839218139648,
"logits/rejected": -1.4671372175216675,
"logps/chosen": -804.1160278320312,
"logps/rejected": -807.0818481445312,
"loss": 0.4974,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.2132842540740967,
"rewards/margins": 0.4933798611164093,
"rewards/rejected": -3.7066640853881836,
"step": 1149
},
{
"epoch": 0.7510817209568128,
"grad_norm": 12.040478011711151,
"learning_rate": 2.6594594344535416e-08,
"logits/chosen": -1.5430724620819092,
"logits/rejected": -1.4982068538665771,
"logps/chosen": -903.4131469726562,
"logps/rejected": -982.6273193359375,
"loss": 0.494,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6371841430664062,
"rewards/margins": 0.9791971445083618,
"rewards/rejected": -4.616381645202637,
"step": 1150
},
{
"epoch": 0.7517348354967752,
"grad_norm": 17.536012470211805,
"learning_rate": 2.646401919565679e-08,
"logits/chosen": -1.5547937154769897,
"logits/rejected": -1.578386664390564,
"logps/chosen": -779.9551391601562,
"logps/rejected": -914.7993774414062,
"loss": 0.4548,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0140175819396973,
"rewards/margins": 1.036094307899475,
"rewards/rejected": -4.050111770629883,
"step": 1151
},
{
"epoch": 0.7523879500367376,
"grad_norm": 13.483448697030374,
"learning_rate": 2.6333696682923677e-08,
"logits/chosen": -1.478268027305603,
"logits/rejected": -1.4786677360534668,
"logps/chosen": -848.6054077148438,
"logps/rejected": -985.2492065429688,
"loss": 0.3929,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.365396499633789,
"rewards/margins": 1.4922524690628052,
"rewards/rejected": -4.857649326324463,
"step": 1152
},
{
"epoch": 0.7530410645767002,
"grad_norm": 17.054706803705002,
"learning_rate": 2.6203627484681862e-08,
"logits/chosen": -1.5397868156433105,
"logits/rejected": -1.4064992666244507,
"logps/chosen": -779.4160766601562,
"logps/rejected": -861.3985595703125,
"loss": 0.5078,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.2905077934265137,
"rewards/margins": 1.0052298307418823,
"rewards/rejected": -4.295737266540527,
"step": 1153
},
{
"epoch": 0.7536941791166626,
"grad_norm": 32.79446433639373,
"learning_rate": 2.6073812277958565e-08,
"logits/chosen": -1.4748960733413696,
"logits/rejected": -1.4349013566970825,
"logps/chosen": -851.8651123046875,
"logps/rejected": -946.66357421875,
"loss": 0.4066,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2334001064300537,
"rewards/margins": 0.994164228439331,
"rewards/rejected": -4.227564334869385,
"step": 1154
},
{
"epoch": 0.754347293656625,
"grad_norm": 67.74543674863597,
"learning_rate": 2.5944251738458985e-08,
"logits/chosen": -1.4562938213348389,
"logits/rejected": -1.4436331987380981,
"logps/chosen": -778.5364990234375,
"logps/rejected": -855.3143920898438,
"loss": 0.4975,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.058043956756592,
"rewards/margins": 0.7909223437309265,
"rewards/rejected": -3.848966360092163,
"step": 1155
},
{
"epoch": 0.7550004081965874,
"grad_norm": 120.90210970935338,
"learning_rate": 2.58149465405627e-08,
"logits/chosen": -1.5155216455459595,
"logits/rejected": -1.5171725749969482,
"logps/chosen": -874.121826171875,
"logps/rejected": -1065.99609375,
"loss": 0.5047,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.261012554168701,
"rewards/margins": 1.1070443391799927,
"rewards/rejected": -4.368056774139404,
"step": 1156
},
{
"epoch": 0.75565352273655,
"grad_norm": 22.64305491156398,
"learning_rate": 2.5685897357320236e-08,
"logits/chosen": -1.5339815616607666,
"logits/rejected": -1.534791350364685,
"logps/chosen": -835.3990478515625,
"logps/rejected": -900.058349609375,
"loss": 0.4601,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2667551040649414,
"rewards/margins": 0.9697099328041077,
"rewards/rejected": -4.236464977264404,
"step": 1157
},
{
"epoch": 0.7563066372765124,
"grad_norm": 12.582513483094319,
"learning_rate": 2.555710486044951e-08,
"logits/chosen": -1.5386102199554443,
"logits/rejected": -1.504981279373169,
"logps/chosen": -791.6339721679688,
"logps/rejected": -953.6394653320312,
"loss": 0.4794,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0311837196350098,
"rewards/margins": 1.1218692064285278,
"rewards/rejected": -4.153052806854248,
"step": 1158
},
{
"epoch": 0.7569597518164748,
"grad_norm": 12.043718502977505,
"learning_rate": 2.542856972033237e-08,
"logits/chosen": -1.5250686407089233,
"logits/rejected": -1.502375602722168,
"logps/chosen": -846.478515625,
"logps/rejected": -939.0611572265625,
"loss": 0.4625,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.351508378982544,
"rewards/margins": 1.0537912845611572,
"rewards/rejected": -4.405299186706543,
"step": 1159
},
{
"epoch": 0.7576128663564372,
"grad_norm": 38.38806781229567,
"learning_rate": 2.5300292606011058e-08,
"logits/chosen": -1.334637999534607,
"logits/rejected": -1.3652501106262207,
"logps/chosen": -769.8375854492188,
"logps/rejected": -865.3778076171875,
"loss": 0.4443,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.9790191650390625,
"rewards/margins": 0.9127359986305237,
"rewards/rejected": -3.8917548656463623,
"step": 1160
},
{
"epoch": 0.7582659808963997,
"grad_norm": 67.81176562059466,
"learning_rate": 2.5172274185184795e-08,
"logits/chosen": -1.5681705474853516,
"logits/rejected": -1.5657100677490234,
"logps/chosen": -860.4470825195312,
"logps/rejected": -908.8406982421875,
"loss": 0.5013,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.02535080909729,
"rewards/margins": 0.7398912906646729,
"rewards/rejected": -3.765242099761963,
"step": 1161
},
{
"epoch": 0.7589190954363622,
"grad_norm": 18.366197171597655,
"learning_rate": 2.504451512420624e-08,
"logits/chosen": -1.439697265625,
"logits/rejected": -1.4873735904693604,
"logps/chosen": -856.3161010742188,
"logps/rejected": -1009.01708984375,
"loss": 0.4974,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3951454162597656,
"rewards/margins": 1.1207799911499023,
"rewards/rejected": -4.515925407409668,
"step": 1162
},
{
"epoch": 0.7595722099763246,
"grad_norm": 14.372222255787399,
"learning_rate": 2.491701608807807e-08,
"logits/chosen": -1.467217206954956,
"logits/rejected": -1.4398670196533203,
"logps/chosen": -864.7401123046875,
"logps/rejected": -916.0717163085938,
"loss": 0.5183,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.4642720222473145,
"rewards/margins": 0.79485023021698,
"rewards/rejected": -4.259122848510742,
"step": 1163
},
{
"epoch": 0.760225324516287,
"grad_norm": 51.64803091036702,
"learning_rate": 2.478977774044948e-08,
"logits/chosen": -1.4573464393615723,
"logits/rejected": -1.450842261314392,
"logps/chosen": -743.4122314453125,
"logps/rejected": -863.0440063476562,
"loss": 0.4459,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0900349617004395,
"rewards/margins": 0.9152263402938843,
"rewards/rejected": -4.005261421203613,
"step": 1164
},
{
"epoch": 0.7608784390562495,
"grad_norm": 22.157747485258298,
"learning_rate": 2.466280074361277e-08,
"logits/chosen": -1.5302906036376953,
"logits/rejected": -1.5034409761428833,
"logps/chosen": -872.1870727539062,
"logps/rejected": -945.05859375,
"loss": 0.5473,
"rewards/accuracies": 0.71875,
"rewards/chosen": -4.0566864013671875,
"rewards/margins": 0.5899438858032227,
"rewards/rejected": -4.646629810333252,
"step": 1165
},
{
"epoch": 0.7615315535962119,
"grad_norm": 61.748046100373145,
"learning_rate": 2.4536085758499845e-08,
"logits/chosen": -1.4410693645477295,
"logits/rejected": -1.4090842008590698,
"logps/chosen": -860.2272338867188,
"logps/rejected": -942.9466552734375,
"loss": 0.5124,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.1852986812591553,
"rewards/margins": 1.0137269496917725,
"rewards/rejected": -4.199025630950928,
"step": 1166
},
{
"epoch": 0.7621846681361744,
"grad_norm": 111.18614790178633,
"learning_rate": 2.4409633444678828e-08,
"logits/chosen": -1.5321191549301147,
"logits/rejected": -1.5227891206741333,
"logps/chosen": -827.511474609375,
"logps/rejected": -978.70458984375,
"loss": 0.493,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.226242780685425,
"rewards/margins": 1.047170877456665,
"rewards/rejected": -4.27341365814209,
"step": 1167
},
{
"epoch": 0.7628377826761368,
"grad_norm": 39.173609596636474,
"learning_rate": 2.42834444603506e-08,
"logits/chosen": -1.4213345050811768,
"logits/rejected": -1.4618169069290161,
"logps/chosen": -748.9151000976562,
"logps/rejected": -865.19775390625,
"loss": 0.5424,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.1750364303588867,
"rewards/margins": 0.8123236298561096,
"rewards/rejected": -3.9873600006103516,
"step": 1168
},
{
"epoch": 0.7634908972160993,
"grad_norm": 100.75678767672636,
"learning_rate": 2.4157519462345373e-08,
"logits/chosen": -1.494844675064087,
"logits/rejected": -1.3755334615707397,
"logps/chosen": -878.364990234375,
"logps/rejected": -937.23193359375,
"loss": 0.507,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4742939472198486,
"rewards/margins": 0.7716997861862183,
"rewards/rejected": -4.245993614196777,
"step": 1169
},
{
"epoch": 0.7641440117560617,
"grad_norm": 18.303452723748897,
"learning_rate": 2.4031859106119267e-08,
"logits/chosen": -1.4939923286437988,
"logits/rejected": -1.4547199010849,
"logps/chosen": -811.0922241210938,
"logps/rejected": -918.572021484375,
"loss": 0.5138,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.3449172973632812,
"rewards/margins": 1.024104356765747,
"rewards/rejected": -4.369021415710449,
"step": 1170
},
{
"epoch": 0.7647971262960241,
"grad_norm": 67.44858894155142,
"learning_rate": 2.3906464045750928e-08,
"logits/chosen": -1.5134905576705933,
"logits/rejected": -1.4933390617370605,
"logps/chosen": -844.356689453125,
"logps/rejected": -925.1680908203125,
"loss": 0.4776,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6389784812927246,
"rewards/margins": 0.7109842300415039,
"rewards/rejected": -4.34996223449707,
"step": 1171
},
{
"epoch": 0.7654502408359866,
"grad_norm": 60.64995050523796,
"learning_rate": 2.3781334933938094e-08,
"logits/chosen": -1.5297952890396118,
"logits/rejected": -1.520561695098877,
"logps/chosen": -983.57958984375,
"logps/rejected": -1131.6011962890625,
"loss": 0.4451,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.437025785446167,
"rewards/margins": 1.1520684957504272,
"rewards/rejected": -4.589094161987305,
"step": 1172
},
{
"epoch": 0.7661033553759491,
"grad_norm": 61.990214860015534,
"learning_rate": 2.3656472421994215e-08,
"logits/chosen": -1.447062373161316,
"logits/rejected": -1.447553038597107,
"logps/chosen": -779.4403076171875,
"logps/rejected": -957.9571533203125,
"loss": 0.5032,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.1557390689849854,
"rewards/margins": 1.0444045066833496,
"rewards/rejected": -4.200143814086914,
"step": 1173
},
{
"epoch": 0.7667564699159115,
"grad_norm": 16.149576622820327,
"learning_rate": 2.3531877159844986e-08,
"logits/chosen": -1.5438363552093506,
"logits/rejected": -1.5356152057647705,
"logps/chosen": -883.9613647460938,
"logps/rejected": -928.40087890625,
"loss": 0.5555,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.240826368331909,
"rewards/margins": 0.7265506386756897,
"rewards/rejected": -3.967376947402954,
"step": 1174
},
{
"epoch": 0.7674095844558739,
"grad_norm": 12.855340027433474,
"learning_rate": 2.3407549796025138e-08,
"logits/chosen": -1.444447636604309,
"logits/rejected": -1.4474413394927979,
"logps/chosen": -902.9122924804688,
"logps/rejected": -1022.8636474609375,
"loss": 0.4878,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.586104393005371,
"rewards/margins": 1.1032967567443848,
"rewards/rejected": -4.689400672912598,
"step": 1175
},
{
"epoch": 0.7680626989958363,
"grad_norm": 44.85596625562831,
"learning_rate": 2.3283490977674887e-08,
"logits/chosen": -1.47477126121521,
"logits/rejected": -1.473871111869812,
"logps/chosen": -713.3693237304688,
"logps/rejected": -823.8683471679688,
"loss": 0.4626,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.159616708755493,
"rewards/margins": 0.8683343529701233,
"rewards/rejected": -4.027950763702393,
"step": 1176
},
{
"epoch": 0.7687158135357989,
"grad_norm": 10.37144025860678,
"learning_rate": 2.3159701350536645e-08,
"logits/chosen": -1.4041333198547363,
"logits/rejected": -1.3923746347427368,
"logps/chosen": -832.6809692382812,
"logps/rejected": -847.46923828125,
"loss": 0.4712,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.0015857219696045,
"rewards/margins": 0.7252290844917297,
"rewards/rejected": -3.7268147468566895,
"step": 1177
},
{
"epoch": 0.7693689280757613,
"grad_norm": 14.75625387529424,
"learning_rate": 2.3036181558951672e-08,
"logits/chosen": -1.5356097221374512,
"logits/rejected": -1.5658423900604248,
"logps/chosen": -829.6064453125,
"logps/rejected": -899.4547729492188,
"loss": 0.4993,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.934075117111206,
"rewards/margins": 1.1105401515960693,
"rewards/rejected": -4.044615268707275,
"step": 1178
},
{
"epoch": 0.7700220426157237,
"grad_norm": 11.129241460176539,
"learning_rate": 2.2912932245856683e-08,
"logits/chosen": -1.5327390432357788,
"logits/rejected": -1.4865920543670654,
"logps/chosen": -914.9544067382812,
"logps/rejected": -1039.690185546875,
"loss": 0.431,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.7165744304656982,
"rewards/margins": 1.1967370510101318,
"rewards/rejected": -4.913311958312988,
"step": 1179
},
{
"epoch": 0.7706751571556861,
"grad_norm": 27.071377513200186,
"learning_rate": 2.278995405278051e-08,
"logits/chosen": -1.490029215812683,
"logits/rejected": -1.4955387115478516,
"logps/chosen": -788.3419799804688,
"logps/rejected": -908.0980224609375,
"loss": 0.4685,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.865569829940796,
"rewards/margins": 0.890224814414978,
"rewards/rejected": -3.7557942867279053,
"step": 1180
},
{
"epoch": 0.7713282716956487,
"grad_norm": 68.89217637019948,
"learning_rate": 2.266724761984077e-08,
"logits/chosen": -1.5287120342254639,
"logits/rejected": -1.5607788562774658,
"logps/chosen": -864.7396240234375,
"logps/rejected": -897.0245971679688,
"loss": 0.5442,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2756667137145996,
"rewards/margins": 0.8783707618713379,
"rewards/rejected": -4.1540374755859375,
"step": 1181
},
{
"epoch": 0.7719813862356111,
"grad_norm": 23.549000554913878,
"learning_rate": 2.2544813585740552e-08,
"logits/chosen": -1.5588055849075317,
"logits/rejected": -1.566838264465332,
"logps/chosen": -907.4659423828125,
"logps/rejected": -1011.3638916015625,
"loss": 0.4863,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.550849437713623,
"rewards/margins": 0.8633126616477966,
"rewards/rejected": -4.414162635803223,
"step": 1182
},
{
"epoch": 0.7726345007755735,
"grad_norm": 73.6862565187378,
"learning_rate": 2.242265258776505e-08,
"logits/chosen": -1.4739999771118164,
"logits/rejected": -1.4539254903793335,
"logps/chosen": -828.4271850585938,
"logps/rejected": -1103.9161376953125,
"loss": 0.4844,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.812699317932129,
"rewards/margins": 1.4099655151367188,
"rewards/rejected": -4.222664833068848,
"step": 1183
},
{
"epoch": 0.7732876153155359,
"grad_norm": 28.50045032866641,
"learning_rate": 2.2300765261778312e-08,
"logits/chosen": -1.4642502069473267,
"logits/rejected": -1.466383457183838,
"logps/chosen": -875.912841796875,
"logps/rejected": -906.390380859375,
"loss": 0.5413,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.574983596801758,
"rewards/margins": 0.3633178770542145,
"rewards/rejected": -3.9383015632629395,
"step": 1184
},
{
"epoch": 0.7739407298554984,
"grad_norm": 13.495729369988036,
"learning_rate": 2.2179152242219803e-08,
"logits/chosen": -1.5543307065963745,
"logits/rejected": -1.5425894260406494,
"logps/chosen": -870.7875366210938,
"logps/rejected": -942.445556640625,
"loss": 0.489,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3177812099456787,
"rewards/margins": 0.8684400320053101,
"rewards/rejected": -4.186221599578857,
"step": 1185
},
{
"epoch": 0.7745938443954609,
"grad_norm": 121.88098996369544,
"learning_rate": 2.205781416210126e-08,
"logits/chosen": -1.3570061922073364,
"logits/rejected": -1.3827579021453857,
"logps/chosen": -873.6181030273438,
"logps/rejected": -917.5654296875,
"loss": 0.4861,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.150343894958496,
"rewards/margins": 1.0735583305358887,
"rewards/rejected": -4.223902702331543,
"step": 1186
},
{
"epoch": 0.7752469589354233,
"grad_norm": 42.77475056245068,
"learning_rate": 2.1936751653003312e-08,
"logits/chosen": -1.525205135345459,
"logits/rejected": -1.5034763813018799,
"logps/chosen": -855.9130249023438,
"logps/rejected": -936.5135498046875,
"loss": 0.4463,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.344566822052002,
"rewards/margins": 0.8852987289428711,
"rewards/rejected": -4.229865550994873,
"step": 1187
},
{
"epoch": 0.7759000734753857,
"grad_norm": 22.279778341803244,
"learning_rate": 2.181596534507219e-08,
"logits/chosen": -1.5698206424713135,
"logits/rejected": -1.5208396911621094,
"logps/chosen": -858.2536010742188,
"logps/rejected": -927.1165771484375,
"loss": 0.4899,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2629714012145996,
"rewards/margins": 0.9225921630859375,
"rewards/rejected": -4.185564041137695,
"step": 1188
},
{
"epoch": 0.7765531880153482,
"grad_norm": 79.29971170882058,
"learning_rate": 2.169545586701647e-08,
"logits/chosen": -1.6312425136566162,
"logits/rejected": -1.5835908651351929,
"logps/chosen": -894.0716552734375,
"logps/rejected": -900.3629760742188,
"loss": 0.4934,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.885010242462158,
"rewards/margins": 0.54335618019104,
"rewards/rejected": -4.428366661071777,
"step": 1189
},
{
"epoch": 0.7772063025553106,
"grad_norm": 50.17561559841045,
"learning_rate": 2.157522384610379e-08,
"logits/chosen": -1.4537131786346436,
"logits/rejected": -1.4958133697509766,
"logps/chosen": -851.051513671875,
"logps/rejected": -874.2980346679688,
"loss": 0.5582,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4431140422821045,
"rewards/margins": 0.47253862023353577,
"rewards/rejected": -3.9156525135040283,
"step": 1190
},
{
"epoch": 0.7778594170952731,
"grad_norm": 16.156948341176815,
"learning_rate": 2.1455269908157583e-08,
"logits/chosen": -1.574291467666626,
"logits/rejected": -1.532688856124878,
"logps/chosen": -930.2337646484375,
"logps/rejected": -1081.06103515625,
"loss": 0.4525,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6473445892333984,
"rewards/margins": 1.3627281188964844,
"rewards/rejected": -5.010072708129883,
"step": 1191
},
{
"epoch": 0.7785125316352355,
"grad_norm": 18.980973183816936,
"learning_rate": 2.133559467755383e-08,
"logits/chosen": -1.6023908853530884,
"logits/rejected": -1.603822946548462,
"logps/chosen": -828.7919311523438,
"logps/rejected": -913.5147705078125,
"loss": 0.4728,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.939816951751709,
"rewards/margins": 0.9978110790252686,
"rewards/rejected": -3.9376275539398193,
"step": 1192
},
{
"epoch": 0.779165646175198,
"grad_norm": 22.58657566222872,
"learning_rate": 2.12161987772178e-08,
"logits/chosen": -1.5595180988311768,
"logits/rejected": -1.5140559673309326,
"logps/chosen": -898.070068359375,
"logps/rejected": -931.6812744140625,
"loss": 0.5008,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.8756892681121826,
"rewards/margins": 0.6614542603492737,
"rewards/rejected": -4.537143230438232,
"step": 1193
},
{
"epoch": 0.7798187607151604,
"grad_norm": 14.841539963082138,
"learning_rate": 2.1097082828620823e-08,
"logits/chosen": -1.5292487144470215,
"logits/rejected": -1.4865946769714355,
"logps/chosen": -779.00244140625,
"logps/rejected": -803.3477172851562,
"loss": 0.5534,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.951836347579956,
"rewards/margins": 0.46454131603240967,
"rewards/rejected": -3.4163777828216553,
"step": 1194
},
{
"epoch": 0.7804718752551228,
"grad_norm": 70.51401475102385,
"learning_rate": 2.0978247451777027e-08,
"logits/chosen": -1.4845449924468994,
"logits/rejected": -1.4704723358154297,
"logps/chosen": -838.1801147460938,
"logps/rejected": -939.9219970703125,
"loss": 0.4543,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.091367244720459,
"rewards/margins": 0.8508425354957581,
"rewards/rejected": -3.9422097206115723,
"step": 1195
},
{
"epoch": 0.7811249897950853,
"grad_norm": 39.49476961487304,
"learning_rate": 2.0859693265240133e-08,
"logits/chosen": -1.4192631244659424,
"logits/rejected": -1.4639571905136108,
"logps/chosen": -819.2432861328125,
"logps/rejected": -913.455078125,
"loss": 0.4721,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.472623109817505,
"rewards/margins": 0.8139183521270752,
"rewards/rejected": -4.28654146194458,
"step": 1196
},
{
"epoch": 0.7817781043350478,
"grad_norm": 38.127552444665085,
"learning_rate": 2.0741420886100226e-08,
"logits/chosen": -1.4315974712371826,
"logits/rejected": -1.3939564228057861,
"logps/chosen": -748.397705078125,
"logps/rejected": -789.2747802734375,
"loss": 0.609,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.52626895904541,
"rewards/margins": 0.5112132430076599,
"rewards/rejected": -4.037482261657715,
"step": 1197
},
{
"epoch": 0.7824312188750102,
"grad_norm": 92.0980308502603,
"learning_rate": 2.0623430929980555e-08,
"logits/chosen": -1.485008955001831,
"logits/rejected": -1.4899412393569946,
"logps/chosen": -1039.5450439453125,
"logps/rejected": -1029.2069091796875,
"loss": 0.5313,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.7229716777801514,
"rewards/margins": 0.5886335372924805,
"rewards/rejected": -4.311605453491211,
"step": 1198
},
{
"epoch": 0.7830843334149726,
"grad_norm": 23.40754516414652,
"learning_rate": 2.0505724011034305e-08,
"logits/chosen": -1.5619934797286987,
"logits/rejected": -1.5442955493927002,
"logps/chosen": -920.8180541992188,
"logps/rejected": -1025.9613037109375,
"loss": 0.4786,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.7618021965026855,
"rewards/margins": 0.9860894680023193,
"rewards/rejected": -4.747891426086426,
"step": 1199
},
{
"epoch": 0.783737447954935,
"grad_norm": 71.55959913562558,
"learning_rate": 2.0388300741941447e-08,
"logits/chosen": -1.4612364768981934,
"logits/rejected": -1.426287293434143,
"logps/chosen": -804.0591430664062,
"logps/rejected": -901.520751953125,
"loss": 0.5403,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.961287260055542,
"rewards/margins": 0.8116233944892883,
"rewards/rejected": -3.7729105949401855,
"step": 1200
},
{
"epoch": 0.783737447954935,
"eval_logits/chosen": -1.4822365045547485,
"eval_logits/rejected": -1.464072823524475,
"eval_logps/chosen": -844.524169921875,
"eval_logps/rejected": -927.9747314453125,
"eval_loss": 0.49636971950531006,
"eval_rewards/accuracies": 0.7620000243186951,
"eval_rewards/chosen": -3.3417880535125732,
"eval_rewards/margins": 0.9155648350715637,
"eval_rewards/rejected": -4.257352828979492,
"eval_runtime": 296.4203,
"eval_samples_per_second": 13.494,
"eval_steps_per_second": 0.843,
"step": 1200
},
{
"epoch": 0.7843905624948976,
"grad_norm": 38.14391865815223,
"learning_rate": 2.027116173390549e-08,
"logits/chosen": -1.4551496505737305,
"logits/rejected": -1.446352243423462,
"logps/chosen": -821.078369140625,
"logps/rejected": -918.0855712890625,
"loss": 0.4739,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5896596908569336,
"rewards/margins": 0.7162079215049744,
"rewards/rejected": -4.305867671966553,
"step": 1201
},
{
"epoch": 0.78504367703486,
"grad_norm": 59.161675430462715,
"learning_rate": 2.015430759665032e-08,
"logits/chosen": -1.5572847127914429,
"logits/rejected": -1.5632448196411133,
"logps/chosen": -812.02294921875,
"logps/rejected": -977.6295776367188,
"loss": 0.4483,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.5253820419311523,
"rewards/margins": 1.2742270231246948,
"rewards/rejected": -4.799609184265137,
"step": 1202
},
{
"epoch": 0.7856967915748224,
"grad_norm": 96.67396687187629,
"learning_rate": 2.003773893841706e-08,
"logits/chosen": -1.5836585760116577,
"logits/rejected": -1.5667706727981567,
"logps/chosen": -812.0299682617188,
"logps/rejected": -925.529052734375,
"loss": 0.4787,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8535349369049072,
"rewards/margins": 1.1299189329147339,
"rewards/rejected": -3.9834539890289307,
"step": 1203
},
{
"epoch": 0.7863499061147848,
"grad_norm": 10.591507916072793,
"learning_rate": 1.9921456365960856e-08,
"logits/chosen": -1.4587793350219727,
"logits/rejected": -1.4299750328063965,
"logps/chosen": -911.44384765625,
"logps/rejected": -1020.269287109375,
"loss": 0.5223,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.7610325813293457,
"rewards/margins": 0.6058591604232788,
"rewards/rejected": -4.366891860961914,
"step": 1204
},
{
"epoch": 0.7870030206547474,
"grad_norm": 92.25556529129581,
"learning_rate": 1.980546048454776e-08,
"logits/chosen": -1.4399176836013794,
"logits/rejected": -1.4096533060073853,
"logps/chosen": -771.2178955078125,
"logps/rejected": -856.9338989257812,
"loss": 0.5283,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.911550521850586,
"rewards/margins": 0.6725373268127441,
"rewards/rejected": -3.584087610244751,
"step": 1205
},
{
"epoch": 0.7876561351947098,
"grad_norm": 50.596463509645794,
"learning_rate": 1.9689751897951532e-08,
"logits/chosen": -1.4841548204421997,
"logits/rejected": -1.4760531187057495,
"logps/chosen": -828.1544799804688,
"logps/rejected": -881.6762084960938,
"loss": 0.5132,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.460951328277588,
"rewards/margins": 0.6589277982711792,
"rewards/rejected": -4.119879245758057,
"step": 1206
},
{
"epoch": 0.7883092497346722,
"grad_norm": 16.96643780611066,
"learning_rate": 1.9574331208450577e-08,
"logits/chosen": -1.4726507663726807,
"logits/rejected": -1.5015829801559448,
"logps/chosen": -860.7503051757812,
"logps/rejected": -979.796142578125,
"loss": 0.4893,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.2518157958984375,
"rewards/margins": 0.9525402784347534,
"rewards/rejected": -4.2043561935424805,
"step": 1207
},
{
"epoch": 0.7889623642746346,
"grad_norm": 17.37151480950942,
"learning_rate": 1.9459199016824668e-08,
"logits/chosen": -1.4686522483825684,
"logits/rejected": -1.4412596225738525,
"logps/chosen": -841.184326171875,
"logps/rejected": -1021.4664306640625,
"loss": 0.4921,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.811629295349121,
"rewards/margins": 1.1526672840118408,
"rewards/rejected": -4.964296340942383,
"step": 1208
},
{
"epoch": 0.7896154788145971,
"grad_norm": 15.472923440337157,
"learning_rate": 1.9344355922351986e-08,
"logits/chosen": -1.4205299615859985,
"logits/rejected": -1.4428074359893799,
"logps/chosen": -806.4320678710938,
"logps/rejected": -891.0250854492188,
"loss": 0.4432,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.923144578933716,
"rewards/margins": 1.4412071704864502,
"rewards/rejected": -4.364351272583008,
"step": 1209
},
{
"epoch": 0.7902685933545596,
"grad_norm": 67.06746939382018,
"learning_rate": 1.922980252280589e-08,
"logits/chosen": -1.4646856784820557,
"logits/rejected": -1.4149240255355835,
"logps/chosen": -835.4559936523438,
"logps/rejected": -962.0414428710938,
"loss": 0.4964,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.448225498199463,
"rewards/margins": 0.8466680645942688,
"rewards/rejected": -4.294894218444824,
"step": 1210
},
{
"epoch": 0.790921707894522,
"grad_norm": 11.275402565937414,
"learning_rate": 1.9115539414451864e-08,
"logits/chosen": -1.4719829559326172,
"logits/rejected": -1.5199016332626343,
"logps/chosen": -804.529541015625,
"logps/rejected": -926.13623046875,
"loss": 0.4676,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2839932441711426,
"rewards/margins": 1.0498989820480347,
"rewards/rejected": -4.333892345428467,
"step": 1211
},
{
"epoch": 0.7915748224344844,
"grad_norm": 27.08202842365883,
"learning_rate": 1.9001567192044367e-08,
"logits/chosen": -1.4914042949676514,
"logits/rejected": -1.438066840171814,
"logps/chosen": -764.7699584960938,
"logps/rejected": -883.679443359375,
"loss": 0.4568,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.755364418029785,
"rewards/margins": 0.9105137586593628,
"rewards/rejected": -3.6658782958984375,
"step": 1212
},
{
"epoch": 0.7922279369744469,
"grad_norm": 19.72267692530407,
"learning_rate": 1.888788644882376e-08,
"logits/chosen": -1.5127348899841309,
"logits/rejected": -1.491123080253601,
"logps/chosen": -755.7037963867188,
"logps/rejected": -753.5874633789062,
"loss": 0.507,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.000194549560547,
"rewards/margins": 0.5762927532196045,
"rewards/rejected": -3.5764873027801514,
"step": 1213
},
{
"epoch": 0.7928810515144094,
"grad_norm": 87.91830418107274,
"learning_rate": 1.8774497776513222e-08,
"logits/chosen": -1.4988164901733398,
"logits/rejected": -1.5042667388916016,
"logps/chosen": -846.1057739257812,
"logps/rejected": -972.7584228515625,
"loss": 0.4886,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.0713298320770264,
"rewards/margins": 0.8871229290962219,
"rewards/rejected": -3.9584529399871826,
"step": 1214
},
{
"epoch": 0.7935341660543718,
"grad_norm": 40.368754316154686,
"learning_rate": 1.8661401765315665e-08,
"logits/chosen": -1.430102825164795,
"logits/rejected": -1.4247267246246338,
"logps/chosen": -770.588623046875,
"logps/rejected": -842.4027099609375,
"loss": 0.4727,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.137661933898926,
"rewards/margins": 1.112958550453186,
"rewards/rejected": -4.2506208419799805,
"step": 1215
},
{
"epoch": 0.7941872805943342,
"grad_norm": 80.49162691211416,
"learning_rate": 1.8548599003910666e-08,
"logits/chosen": -1.4262512922286987,
"logits/rejected": -1.4448204040527344,
"logps/chosen": -881.8290405273438,
"logps/rejected": -979.1732788085938,
"loss": 0.4524,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.9086101055145264,
"rewards/margins": 1.3396742343902588,
"rewards/rejected": -5.248284339904785,
"step": 1216
},
{
"epoch": 0.7948403951342967,
"grad_norm": 16.198495509148298,
"learning_rate": 1.843609007945138e-08,
"logits/chosen": -1.4633352756500244,
"logits/rejected": -1.462742805480957,
"logps/chosen": -818.7781372070312,
"logps/rejected": -850.6954956054688,
"loss": 0.5104,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.175035238265991,
"rewards/margins": 0.5016341805458069,
"rewards/rejected": -3.676669120788574,
"step": 1217
},
{
"epoch": 0.7954935096742591,
"grad_norm": 30.16649596643903,
"learning_rate": 1.832387557756151e-08,
"logits/chosen": -1.4935240745544434,
"logits/rejected": -1.4943130016326904,
"logps/chosen": -749.455078125,
"logps/rejected": -783.4890747070312,
"loss": 0.5097,
"rewards/accuracies": 0.5625,
"rewards/chosen": -3.018079996109009,
"rewards/margins": 0.39417386054992676,
"rewards/rejected": -3.4122540950775146,
"step": 1218
},
{
"epoch": 0.7961466242142216,
"grad_norm": 63.990103383436264,
"learning_rate": 1.82119560823323e-08,
"logits/chosen": -1.5611546039581299,
"logits/rejected": -1.5044180154800415,
"logps/chosen": -828.6058349609375,
"logps/rejected": -971.0333862304688,
"loss": 0.5007,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2003324031829834,
"rewards/margins": 1.3617459535598755,
"rewards/rejected": -4.56207799911499,
"step": 1219
},
{
"epoch": 0.796799738754184,
"grad_norm": 54.853667872883626,
"learning_rate": 1.8100332176319338e-08,
"logits/chosen": -1.4820096492767334,
"logits/rejected": -1.5054271221160889,
"logps/chosen": -863.5888061523438,
"logps/rejected": -936.4130249023438,
"loss": 0.5115,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.6043381690979004,
"rewards/margins": 0.8906245827674866,
"rewards/rejected": -4.494962692260742,
"step": 1220
},
{
"epoch": 0.7974528532941465,
"grad_norm": 14.995625717970094,
"learning_rate": 1.798900444053972e-08,
"logits/chosen": -1.5272332429885864,
"logits/rejected": -1.516082763671875,
"logps/chosen": -958.9482421875,
"logps/rejected": -1058.3154296875,
"loss": 0.4977,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6495161056518555,
"rewards/margins": 0.7711564898490906,
"rewards/rejected": -4.420672416687012,
"step": 1221
},
{
"epoch": 0.7981059678341089,
"grad_norm": 17.49568930582209,
"learning_rate": 1.7877973454468918e-08,
"logits/chosen": -1.5633349418640137,
"logits/rejected": -1.5596158504486084,
"logps/chosen": -835.0476684570312,
"logps/rejected": -886.090087890625,
"loss": 0.4917,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4766788482666016,
"rewards/margins": 0.8626784682273865,
"rewards/rejected": -4.339357376098633,
"step": 1222
},
{
"epoch": 0.7987590823740713,
"grad_norm": 52.96090556956152,
"learning_rate": 1.7767239796037765e-08,
"logits/chosen": -1.4490398168563843,
"logits/rejected": -1.443628191947937,
"logps/chosen": -802.57373046875,
"logps/rejected": -889.3080444335938,
"loss": 0.4679,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.239607334136963,
"rewards/margins": 1.018710970878601,
"rewards/rejected": -4.258317947387695,
"step": 1223
},
{
"epoch": 0.7994121969140338,
"grad_norm": 150.8059871681939,
"learning_rate": 1.7656804041629487e-08,
"logits/chosen": -1.5590651035308838,
"logits/rejected": -1.5880131721496582,
"logps/chosen": -928.2868041992188,
"logps/rejected": -929.5325317382812,
"loss": 0.527,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.8474342823028564,
"rewards/margins": 0.5773695111274719,
"rewards/rejected": -4.424803733825684,
"step": 1224
},
{
"epoch": 0.8000653114539963,
"grad_norm": 35.966169524301534,
"learning_rate": 1.7546666766076656e-08,
"logits/chosen": -1.449623465538025,
"logits/rejected": -1.5213428735733032,
"logps/chosen": -757.2681884765625,
"logps/rejected": -924.322509765625,
"loss": 0.4752,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0022165775299072,
"rewards/margins": 0.9503800868988037,
"rewards/rejected": -3.952596664428711,
"step": 1225
},
{
"epoch": 0.8007184259939587,
"grad_norm": 56.42831654190206,
"learning_rate": 1.743682854265825e-08,
"logits/chosen": -1.4219781160354614,
"logits/rejected": -1.466795802116394,
"logps/chosen": -840.7289428710938,
"logps/rejected": -965.5621337890625,
"loss": 0.4994,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4919745922088623,
"rewards/margins": 0.8413702249526978,
"rewards/rejected": -4.333344459533691,
"step": 1226
},
{
"epoch": 0.8013715405339211,
"grad_norm": 73.37249453736422,
"learning_rate": 1.732728994309661e-08,
"logits/chosen": -1.5576728582382202,
"logits/rejected": -1.5214743614196777,
"logps/chosen": -763.7567749023438,
"logps/rejected": -810.2583618164062,
"loss": 0.4673,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8903627395629883,
"rewards/margins": 0.7463845610618591,
"rewards/rejected": -3.636747360229492,
"step": 1227
},
{
"epoch": 0.8020246550738835,
"grad_norm": 61.87692158071728,
"learning_rate": 1.7218051537554536e-08,
"logits/chosen": -1.5665578842163086,
"logits/rejected": -1.549164891242981,
"logps/chosen": -867.8675537109375,
"logps/rejected": -983.275634765625,
"loss": 0.4786,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3755083084106445,
"rewards/margins": 0.9794182777404785,
"rewards/rejected": -4.354927062988281,
"step": 1228
},
{
"epoch": 0.8026777696138461,
"grad_norm": 45.704835797476456,
"learning_rate": 1.7109113894632233e-08,
"logits/chosen": -1.4501968622207642,
"logits/rejected": -1.470442771911621,
"logps/chosen": -766.7777099609375,
"logps/rejected": -880.814453125,
"loss": 0.5157,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.088731288909912,
"rewards/margins": 0.6824835538864136,
"rewards/rejected": -3.7712154388427734,
"step": 1229
},
{
"epoch": 0.8033308841538085,
"grad_norm": 12.041002229288496,
"learning_rate": 1.700047758136443e-08,
"logits/chosen": -1.4775810241699219,
"logits/rejected": -1.4621860980987549,
"logps/chosen": -792.953857421875,
"logps/rejected": -907.03759765625,
"loss": 0.4903,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9522554874420166,
"rewards/margins": 1.064509391784668,
"rewards/rejected": -4.016765117645264,
"step": 1230
},
{
"epoch": 0.8039839986937709,
"grad_norm": 63.50098008198583,
"learning_rate": 1.689214316321739e-08,
"logits/chosen": -1.6399027109146118,
"logits/rejected": -1.5657185316085815,
"logps/chosen": -886.7124633789062,
"logps/rejected": -979.0396728515625,
"loss": 0.5165,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.5052478313446045,
"rewards/margins": 1.1115096807479858,
"rewards/rejected": -4.616757869720459,
"step": 1231
},
{
"epoch": 0.8046371132337333,
"grad_norm": 34.66004736818009,
"learning_rate": 1.678411120408595e-08,
"logits/chosen": -1.4924274682998657,
"logits/rejected": -1.4871619939804077,
"logps/chosen": -876.4576416015625,
"logps/rejected": -956.4109497070312,
"loss": 0.4692,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.6071510314941406,
"rewards/margins": 1.06039297580719,
"rewards/rejected": -4.667544364929199,
"step": 1232
},
{
"epoch": 0.8052902277736959,
"grad_norm": 43.61205378322202,
"learning_rate": 1.6676382266290647e-08,
"logits/chosen": -1.4874005317687988,
"logits/rejected": -1.5195080041885376,
"logps/chosen": -837.166015625,
"logps/rejected": -892.5218505859375,
"loss": 0.5412,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4187517166137695,
"rewards/margins": 0.5736494660377502,
"rewards/rejected": -3.992400884628296,
"step": 1233
},
{
"epoch": 0.8059433423136583,
"grad_norm": 98.81263517998161,
"learning_rate": 1.6568956910574712e-08,
"logits/chosen": -1.530608892440796,
"logits/rejected": -1.5279160737991333,
"logps/chosen": -837.9874877929688,
"logps/rejected": -845.3988647460938,
"loss": 0.5812,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.741583824157715,
"rewards/margins": 0.040613558143377304,
"rewards/rejected": -3.7821972370147705,
"step": 1234
},
{
"epoch": 0.8065964568536207,
"grad_norm": 78.86885176654913,
"learning_rate": 1.6461835696101227e-08,
"logits/chosen": -1.4850387573242188,
"logits/rejected": -1.4923676252365112,
"logps/chosen": -871.99658203125,
"logps/rejected": -949.8641357421875,
"loss": 0.4868,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.7873427867889404,
"rewards/margins": 0.7236688733100891,
"rewards/rejected": -4.511011600494385,
"step": 1235
},
{
"epoch": 0.8072495713935831,
"grad_norm": 35.00941246017448,
"learning_rate": 1.6355019180450148e-08,
"logits/chosen": -1.555824637413025,
"logits/rejected": -1.4870938062667847,
"logps/chosen": -837.19140625,
"logps/rejected": -921.0797729492188,
"loss": 0.4621,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3375821113586426,
"rewards/margins": 0.9209874272346497,
"rewards/rejected": -4.258569717407227,
"step": 1236
},
{
"epoch": 0.8079026859335456,
"grad_norm": 90.40912979516071,
"learning_rate": 1.6248507919615452e-08,
"logits/chosen": -1.4929189682006836,
"logits/rejected": -1.5122658014297485,
"logps/chosen": -916.9654541015625,
"logps/rejected": -1060.645751953125,
"loss": 0.42,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.5597081184387207,
"rewards/margins": 1.2362518310546875,
"rewards/rejected": -4.795959949493408,
"step": 1237
},
{
"epoch": 0.8085558004735081,
"grad_norm": 14.457346050782789,
"learning_rate": 1.6142302468002227e-08,
"logits/chosen": -1.4068070650100708,
"logits/rejected": -1.399592638015747,
"logps/chosen": -771.3148193359375,
"logps/rejected": -843.7901611328125,
"loss": 0.5132,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3284435272216797,
"rewards/margins": 0.8365193605422974,
"rewards/rejected": -4.164963245391846,
"step": 1238
},
{
"epoch": 0.8092089150134705,
"grad_norm": 37.50589751809788,
"learning_rate": 1.603640337842377e-08,
"logits/chosen": -1.5218846797943115,
"logits/rejected": -1.5328527688980103,
"logps/chosen": -876.4766235351562,
"logps/rejected": -1026.4345703125,
"loss": 0.4891,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1716041564941406,
"rewards/margins": 0.8838239312171936,
"rewards/rejected": -4.0554280281066895,
"step": 1239
},
{
"epoch": 0.8098620295534329,
"grad_norm": 39.42416181441758,
"learning_rate": 1.5930811202098737e-08,
"logits/chosen": -1.451581597328186,
"logits/rejected": -1.4357130527496338,
"logps/chosen": -736.4873046875,
"logps/rejected": -854.6323852539062,
"loss": 0.5309,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.231121778488159,
"rewards/margins": 0.7173824310302734,
"rewards/rejected": -3.9485039710998535,
"step": 1240
},
{
"epoch": 0.8105151440933954,
"grad_norm": 15.457594936787636,
"learning_rate": 1.5825526488648268e-08,
"logits/chosen": -1.5039798021316528,
"logits/rejected": -1.4966380596160889,
"logps/chosen": -838.3843994140625,
"logps/rejected": -906.8178100585938,
"loss": 0.5552,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.452669620513916,
"rewards/margins": 0.6774091124534607,
"rewards/rejected": -4.1300787925720215,
"step": 1241
},
{
"epoch": 0.8111682586333578,
"grad_norm": 30.649527114852525,
"learning_rate": 1.572054978609306e-08,
"logits/chosen": -1.4918582439422607,
"logits/rejected": -1.454628586769104,
"logps/chosen": -912.1357421875,
"logps/rejected": -949.4915161132812,
"loss": 0.5541,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.9472150802612305,
"rewards/margins": 0.8809834122657776,
"rewards/rejected": -4.828197956085205,
"step": 1242
},
{
"epoch": 0.8118213731733203,
"grad_norm": 15.819803195866424,
"learning_rate": 1.5615881640850653e-08,
"logits/chosen": -1.4457621574401855,
"logits/rejected": -1.3891899585723877,
"logps/chosen": -806.92626953125,
"logps/rejected": -855.1336669921875,
"loss": 0.4604,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.198312282562256,
"rewards/margins": 1.005974531173706,
"rewards/rejected": -4.204287052154541,
"step": 1243
},
{
"epoch": 0.8124744877132827,
"grad_norm": 20.662834289028424,
"learning_rate": 1.551152259773245e-08,
"logits/chosen": -1.4745066165924072,
"logits/rejected": -1.4477360248565674,
"logps/chosen": -888.340576171875,
"logps/rejected": -906.5769653320312,
"loss": 0.5284,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.483821392059326,
"rewards/margins": 0.7100880146026611,
"rewards/rejected": -4.193909168243408,
"step": 1244
},
{
"epoch": 0.8131276022532452,
"grad_norm": 40.934726173262696,
"learning_rate": 1.5407473199940978e-08,
"logits/chosen": -1.5310475826263428,
"logits/rejected": -1.5288856029510498,
"logps/chosen": -761.62451171875,
"logps/rejected": -932.880126953125,
"loss": 0.4466,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.145474433898926,
"rewards/margins": 1.0549681186676025,
"rewards/rejected": -4.200442790985107,
"step": 1245
},
{
"epoch": 0.8137807167932076,
"grad_norm": 59.614420768209214,
"learning_rate": 1.5303733989066992e-08,
"logits/chosen": -1.5001609325408936,
"logits/rejected": -1.4439678192138672,
"logps/chosen": -862.7713623046875,
"logps/rejected": -916.8041381835938,
"loss": 0.5735,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4643964767456055,
"rewards/margins": 0.6968106627464294,
"rewards/rejected": -4.16120719909668,
"step": 1246
},
{
"epoch": 0.81443383133317,
"grad_norm": 99.7769624540917,
"learning_rate": 1.5200305505086678e-08,
"logits/chosen": -1.451883316040039,
"logits/rejected": -1.4290595054626465,
"logps/chosen": -822.1856689453125,
"logps/rejected": -901.4767456054688,
"loss": 0.547,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.326678514480591,
"rewards/margins": 0.7827971577644348,
"rewards/rejected": -4.109475612640381,
"step": 1247
},
{
"epoch": 0.8150869458731325,
"grad_norm": 35.074610634798226,
"learning_rate": 1.509718828635887e-08,
"logits/chosen": -1.4157919883728027,
"logits/rejected": -1.437688946723938,
"logps/chosen": -747.04638671875,
"logps/rejected": -875.550537109375,
"loss": 0.4635,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.111088991165161,
"rewards/margins": 0.8460332155227661,
"rewards/rejected": -3.9571218490600586,
"step": 1248
},
{
"epoch": 0.815740060413095,
"grad_norm": 31.080617936177795,
"learning_rate": 1.4994382869622212e-08,
"logits/chosen": -1.5038199424743652,
"logits/rejected": -1.5081907510757446,
"logps/chosen": -842.7953491210938,
"logps/rejected": -893.8030395507812,
"loss": 0.5325,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.6291537284851074,
"rewards/margins": 0.5055683851242065,
"rewards/rejected": -4.134722709655762,
"step": 1249
},
{
"epoch": 0.8163931749530574,
"grad_norm": 72.9364075543982,
"learning_rate": 1.4891889789992385e-08,
"logits/chosen": -1.4778282642364502,
"logits/rejected": -1.4801677465438843,
"logps/chosen": -810.5230712890625,
"logps/rejected": -921.8294067382812,
"loss": 0.504,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.3633217811584473,
"rewards/margins": 1.1871894598007202,
"rewards/rejected": -4.550510883331299,
"step": 1250
},
{
"epoch": 0.8170462894930198,
"grad_norm": 25.634012106471616,
"learning_rate": 1.4789709580959304e-08,
"logits/chosen": -1.4490480422973633,
"logits/rejected": -1.4597145318984985,
"logps/chosen": -787.5851440429688,
"logps/rejected": -899.1139526367188,
"loss": 0.5296,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.343380928039551,
"rewards/margins": 0.7966138124465942,
"rewards/rejected": -4.139995098114014,
"step": 1251
},
{
"epoch": 0.8176994040329822,
"grad_norm": 13.077479982264203,
"learning_rate": 1.4687842774384366e-08,
"logits/chosen": -1.4068443775177002,
"logits/rejected": -1.4268113374710083,
"logps/chosen": -732.1669921875,
"logps/rejected": -751.791259765625,
"loss": 0.5786,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2558491230010986,
"rewards/margins": 0.2685055732727051,
"rewards/rejected": -3.5243544578552246,
"step": 1252
},
{
"epoch": 0.8183525185729448,
"grad_norm": 96.37113725713898,
"learning_rate": 1.4586289900497672e-08,
"logits/chosen": -1.5462439060211182,
"logits/rejected": -1.519241452217102,
"logps/chosen": -875.8187866210938,
"logps/rejected": -946.7884521484375,
"loss": 0.5182,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.811178207397461,
"rewards/margins": 1.0562658309936523,
"rewards/rejected": -4.867443561553955,
"step": 1253
},
{
"epoch": 0.8190056331129072,
"grad_norm": 47.14222768313448,
"learning_rate": 1.4485051487895208e-08,
"logits/chosen": -1.6328545808792114,
"logits/rejected": -1.5723670721054077,
"logps/chosen": -878.3783569335938,
"logps/rejected": -1000.5723876953125,
"loss": 0.4553,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.133354902267456,
"rewards/margins": 1.4496155977249146,
"rewards/rejected": -4.58297061920166,
"step": 1254
},
{
"epoch": 0.8196587476528696,
"grad_norm": 77.13494040128981,
"learning_rate": 1.4384128063536215e-08,
"logits/chosen": -1.5092742443084717,
"logits/rejected": -1.4484039545059204,
"logps/chosen": -795.540283203125,
"logps/rejected": -922.9327392578125,
"loss": 0.5052,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.0639872550964355,
"rewards/margins": 0.9547647833824158,
"rewards/rejected": -4.018751621246338,
"step": 1255
},
{
"epoch": 0.820311862192832,
"grad_norm": 27.794467326470485,
"learning_rate": 1.4283520152740358e-08,
"logits/chosen": -1.527742624282837,
"logits/rejected": -1.550565481185913,
"logps/chosen": -901.008056640625,
"logps/rejected": -977.6214599609375,
"loss": 0.443,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6517744064331055,
"rewards/margins": 0.9718442559242249,
"rewards/rejected": -4.6236186027526855,
"step": 1256
},
{
"epoch": 0.8209649767327946,
"grad_norm": 48.23537245171485,
"learning_rate": 1.4183228279184986e-08,
"logits/chosen": -1.4986786842346191,
"logits/rejected": -1.505009651184082,
"logps/chosen": -780.6826171875,
"logps/rejected": -851.0025024414062,
"loss": 0.4116,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.2454466819763184,
"rewards/margins": 0.8366904258728027,
"rewards/rejected": -4.082137107849121,
"step": 1257
},
{
"epoch": 0.821618091272757,
"grad_norm": 25.124612077309788,
"learning_rate": 1.4083252964902476e-08,
"logits/chosen": -1.4318965673446655,
"logits/rejected": -1.4011112451553345,
"logps/chosen": -767.1181640625,
"logps/rejected": -834.265380859375,
"loss": 0.4421,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2802295684814453,
"rewards/margins": 0.9027811884880066,
"rewards/rejected": -4.183011054992676,
"step": 1258
},
{
"epoch": 0.8222712058127194,
"grad_norm": 28.873103251905075,
"learning_rate": 1.3983594730277437e-08,
"logits/chosen": -1.504234790802002,
"logits/rejected": -1.4638817310333252,
"logps/chosen": -840.501220703125,
"logps/rejected": -915.8950805664062,
"loss": 0.5238,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.346611976623535,
"rewards/margins": 1.1296738386154175,
"rewards/rejected": -4.4762864112854,
"step": 1259
},
{
"epoch": 0.8229243203526818,
"grad_norm": 92.65936407387743,
"learning_rate": 1.388425409404406e-08,
"logits/chosen": -1.3982869386672974,
"logits/rejected": -1.4056661128997803,
"logps/chosen": -882.9879150390625,
"logps/rejected": -926.9945068359375,
"loss": 0.5427,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.3799350261688232,
"rewards/margins": 0.531534731388092,
"rewards/rejected": -3.9114699363708496,
"step": 1260
},
{
"epoch": 0.8235774348926443,
"grad_norm": 21.267825506734145,
"learning_rate": 1.378523157328338e-08,
"logits/chosen": -1.463804006576538,
"logits/rejected": -1.4482476711273193,
"logps/chosen": -843.3424682617188,
"logps/rejected": -922.429931640625,
"loss": 0.4699,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.516786813735962,
"rewards/margins": 0.4834230840206146,
"rewards/rejected": -4.000209808349609,
"step": 1261
},
{
"epoch": 0.8242305494326068,
"grad_norm": 44.39704770318735,
"learning_rate": 1.3686527683420598e-08,
"logits/chosen": -1.5227603912353516,
"logits/rejected": -1.4986686706542969,
"logps/chosen": -880.2915649414062,
"logps/rejected": -956.4801025390625,
"loss": 0.4738,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.749066114425659,
"rewards/margins": 0.984779417514801,
"rewards/rejected": -4.7338457107543945,
"step": 1262
},
{
"epoch": 0.8248836639725692,
"grad_norm": 40.800361223632876,
"learning_rate": 1.3588142938222421e-08,
"logits/chosen": -1.5330525636672974,
"logits/rejected": -1.5526435375213623,
"logps/chosen": -796.693603515625,
"logps/rejected": -926.6709594726562,
"loss": 0.4648,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.9938881397247314,
"rewards/margins": 1.054705262184143,
"rewards/rejected": -4.048593521118164,
"step": 1263
},
{
"epoch": 0.8255367785125316,
"grad_norm": 16.08173399476035,
"learning_rate": 1.3490077849794333e-08,
"logits/chosen": -1.5841691493988037,
"logits/rejected": -1.5560128688812256,
"logps/chosen": -929.83056640625,
"logps/rejected": -988.2870483398438,
"loss": 0.4453,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.5736300945281982,
"rewards/margins": 0.8916316032409668,
"rewards/rejected": -4.465262413024902,
"step": 1264
},
{
"epoch": 0.8261898930524941,
"grad_norm": 18.753303938177545,
"learning_rate": 1.3392332928577994e-08,
"logits/chosen": -1.5194913148880005,
"logits/rejected": -1.525078535079956,
"logps/chosen": -871.6900024414062,
"logps/rejected": -965.0679931640625,
"loss": 0.4307,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.447103977203369,
"rewards/margins": 0.7160016298294067,
"rewards/rejected": -4.163105487823486,
"step": 1265
},
{
"epoch": 0.8268430075924565,
"grad_norm": 37.23704455866974,
"learning_rate": 1.3294908683348535e-08,
"logits/chosen": -1.4927794933319092,
"logits/rejected": -1.459194302558899,
"logps/chosen": -876.373046875,
"logps/rejected": -885.762939453125,
"loss": 0.5455,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.8211159706115723,
"rewards/margins": 0.5798470973968506,
"rewards/rejected": -4.400962829589844,
"step": 1266
},
{
"epoch": 0.827496122132419,
"grad_norm": 41.88538792702506,
"learning_rate": 1.3197805621211925e-08,
"logits/chosen": -1.4792529344558716,
"logits/rejected": -1.4594306945800781,
"logps/chosen": -812.187744140625,
"logps/rejected": -929.7012939453125,
"loss": 0.475,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.358401298522949,
"rewards/margins": 0.9242434501647949,
"rewards/rejected": -4.282644748687744,
"step": 1267
},
{
"epoch": 0.8281492366723814,
"grad_norm": 143.83646829408002,
"learning_rate": 1.3101024247602339e-08,
"logits/chosen": -1.3687011003494263,
"logits/rejected": -1.416597843170166,
"logps/chosen": -873.91796875,
"logps/rejected": -925.439453125,
"loss": 0.501,
"rewards/accuracies": 0.5625,
"rewards/chosen": -3.801042318344116,
"rewards/margins": 0.3380916118621826,
"rewards/rejected": -4.139133930206299,
"step": 1268
},
{
"epoch": 0.8288023512123439,
"grad_norm": 12.16391760621639,
"learning_rate": 1.3004565066279519e-08,
"logits/chosen": -1.5240532159805298,
"logits/rejected": -1.5018377304077148,
"logps/chosen": -785.153076171875,
"logps/rejected": -848.844970703125,
"loss": 0.5177,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3579812049865723,
"rewards/margins": 0.8355432748794556,
"rewards/rejected": -4.193524360656738,
"step": 1269
},
{
"epoch": 0.8294554657523063,
"grad_norm": 12.97712151020145,
"learning_rate": 1.2908428579326159e-08,
"logits/chosen": -1.483147144317627,
"logits/rejected": -1.440002202987671,
"logps/chosen": -855.696044921875,
"logps/rejected": -937.5626831054688,
"loss": 0.513,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.419983148574829,
"rewards/margins": 1.2121243476867676,
"rewards/rejected": -4.632107734680176,
"step": 1270
},
{
"epoch": 0.8301085802922687,
"grad_norm": 61.10324755914619,
"learning_rate": 1.2812615287145276e-08,
"logits/chosen": -1.4333446025848389,
"logits/rejected": -1.4426469802856445,
"logps/chosen": -779.0205078125,
"logps/rejected": -882.3414306640625,
"loss": 0.5547,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.4126522541046143,
"rewards/margins": 0.8153400421142578,
"rewards/rejected": -4.227993011474609,
"step": 1271
},
{
"epoch": 0.8307616948322312,
"grad_norm": 49.65538467582004,
"learning_rate": 1.2717125688457627e-08,
"logits/chosen": -1.4278748035430908,
"logits/rejected": -1.4869357347488403,
"logps/chosen": -847.4019165039062,
"logps/rejected": -1110.864501953125,
"loss": 0.5348,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.525916576385498,
"rewards/margins": 1.2023837566375732,
"rewards/rejected": -4.728300094604492,
"step": 1272
},
{
"epoch": 0.8314148093721937,
"grad_norm": 21.25568560854086,
"learning_rate": 1.2621960280299093e-08,
"logits/chosen": -1.498213529586792,
"logits/rejected": -1.5190093517303467,
"logps/chosen": -889.1369018554688,
"logps/rejected": -935.4756469726562,
"loss": 0.5172,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.5845720767974854,
"rewards/margins": 1.0147137641906738,
"rewards/rejected": -4.599286079406738,
"step": 1273
},
{
"epoch": 0.8320679239121561,
"grad_norm": 20.186165187940464,
"learning_rate": 1.252711955801811e-08,
"logits/chosen": -1.4972944259643555,
"logits/rejected": -1.515435814857483,
"logps/chosen": -775.0640869140625,
"logps/rejected": -830.6438598632812,
"loss": 0.4944,
"rewards/accuracies": 0.71875,
"rewards/chosen": -2.7457919120788574,
"rewards/margins": 0.6729423999786377,
"rewards/rejected": -3.418734073638916,
"step": 1274
},
{
"epoch": 0.8327210384521185,
"grad_norm": 80.31553260847888,
"learning_rate": 1.2432604015273082e-08,
"logits/chosen": -1.5685982704162598,
"logits/rejected": -1.5332074165344238,
"logps/chosen": -850.53076171875,
"logps/rejected": -1026.168701171875,
"loss": 0.4492,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.52174973487854,
"rewards/margins": 1.3801482915878296,
"rewards/rejected": -4.901898384094238,
"step": 1275
},
{
"epoch": 0.833374152992081,
"grad_norm": 44.67438081658325,
"learning_rate": 1.2338414144029779e-08,
"logits/chosen": -1.4650599956512451,
"logits/rejected": -1.5019781589508057,
"logps/chosen": -831.616455078125,
"logps/rejected": -865.5159912109375,
"loss": 0.5538,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3090555667877197,
"rewards/margins": 0.5248807668685913,
"rewards/rejected": -3.8339362144470215,
"step": 1276
},
{
"epoch": 0.8340272675320435,
"grad_norm": 39.396696943484436,
"learning_rate": 1.2244550434558842e-08,
"logits/chosen": -1.461850643157959,
"logits/rejected": -1.453994631767273,
"logps/chosen": -843.3770751953125,
"logps/rejected": -1038.049072265625,
"loss": 0.5168,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4972283840179443,
"rewards/margins": 1.6629681587219238,
"rewards/rejected": -5.160196781158447,
"step": 1277
},
{
"epoch": 0.8346803820720059,
"grad_norm": 65.0379831567537,
"learning_rate": 1.2151013375433202e-08,
"logits/chosen": -1.553591012954712,
"logits/rejected": -1.507093906402588,
"logps/chosen": -870.8557739257812,
"logps/rejected": -965.6985473632812,
"loss": 0.4459,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.6408274173736572,
"rewards/margins": 1.1605440378189087,
"rewards/rejected": -4.801371097564697,
"step": 1278
},
{
"epoch": 0.8353334966119683,
"grad_norm": 30.676217173011974,
"learning_rate": 1.2057803453525502e-08,
"logits/chosen": -1.468326449394226,
"logits/rejected": -1.4737112522125244,
"logps/chosen": -782.1486206054688,
"logps/rejected": -842.2639770507812,
"loss": 0.5042,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.191077709197998,
"rewards/margins": 0.6720733046531677,
"rewards/rejected": -3.8631508350372314,
"step": 1279
},
{
"epoch": 0.8359866111519307,
"grad_norm": 107.16302969733803,
"learning_rate": 1.1964921154005631e-08,
"logits/chosen": -1.4318448305130005,
"logits/rejected": -1.4438923597335815,
"logps/chosen": -870.4271850585938,
"logps/rejected": -939.97900390625,
"loss": 0.5041,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3240041732788086,
"rewards/margins": 0.9650790095329285,
"rewards/rejected": -4.289083003997803,
"step": 1280
},
{
"epoch": 0.8366397256918933,
"grad_norm": 37.456288698932234,
"learning_rate": 1.187236696033812e-08,
"logits/chosen": -1.5122158527374268,
"logits/rejected": -1.454342007637024,
"logps/chosen": -791.8545532226562,
"logps/rejected": -874.0135498046875,
"loss": 0.4348,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.786623239517212,
"rewards/margins": 0.9403093457221985,
"rewards/rejected": -3.7269325256347656,
"step": 1281
},
{
"epoch": 0.8372928402318557,
"grad_norm": 32.26978309244302,
"learning_rate": 1.1780141354279698e-08,
"logits/chosen": -1.4852337837219238,
"logits/rejected": -1.4611694812774658,
"logps/chosen": -899.8671875,
"logps/rejected": -902.857666015625,
"loss": 0.4584,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.5996108055114746,
"rewards/margins": 0.4853794276714325,
"rewards/rejected": -4.08499002456665,
"step": 1282
},
{
"epoch": 0.8379459547718181,
"grad_norm": 67.20483755586449,
"learning_rate": 1.1688244815876735e-08,
"logits/chosen": -1.4753847122192383,
"logits/rejected": -1.495114803314209,
"logps/chosen": -775.127685546875,
"logps/rejected": -854.59423828125,
"loss": 0.4918,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.08735728263855,
"rewards/margins": 0.8931238055229187,
"rewards/rejected": -3.9804811477661133,
"step": 1283
},
{
"epoch": 0.8385990693117805,
"grad_norm": 31.10069114339872,
"learning_rate": 1.1596677823462769e-08,
"logits/chosen": -1.4978846311569214,
"logits/rejected": -1.5116081237792969,
"logps/chosen": -835.6173095703125,
"logps/rejected": -892.9727783203125,
"loss": 0.4633,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4220669269561768,
"rewards/margins": 1.0128189325332642,
"rewards/rejected": -4.4348859786987305,
"step": 1284
},
{
"epoch": 0.839252183851743,
"grad_norm": 95.67340364123025,
"learning_rate": 1.1505440853655996e-08,
"logits/chosen": -1.580156922340393,
"logits/rejected": -1.564608097076416,
"logps/chosen": -828.8040771484375,
"logps/rejected": -858.2422485351562,
"loss": 0.5397,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2406132221221924,
"rewards/margins": 0.5403940677642822,
"rewards/rejected": -3.7810075283050537,
"step": 1285
},
{
"epoch": 0.8399052983917055,
"grad_norm": 20.610401240104963,
"learning_rate": 1.1414534381356817e-08,
"logits/chosen": -1.457862377166748,
"logits/rejected": -1.4158681631088257,
"logps/chosen": -854.0062255859375,
"logps/rejected": -980.3665771484375,
"loss": 0.5507,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.427743911743164,
"rewards/margins": 1.1576130390167236,
"rewards/rejected": -4.585356712341309,
"step": 1286
},
{
"epoch": 0.8405584129316679,
"grad_norm": 15.387698511738025,
"learning_rate": 1.132395887974536e-08,
"logits/chosen": -1.512953758239746,
"logits/rejected": -1.4666166305541992,
"logps/chosen": -873.6098022460938,
"logps/rejected": -931.2805786132812,
"loss": 0.5066,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.396209478378296,
"rewards/margins": 0.8550810813903809,
"rewards/rejected": -4.251290798187256,
"step": 1287
},
{
"epoch": 0.8412115274716303,
"grad_norm": 17.455285669432005,
"learning_rate": 1.123371482027895e-08,
"logits/chosen": -1.4253538846969604,
"logits/rejected": -1.4188988208770752,
"logps/chosen": -823.5361328125,
"logps/rejected": -855.1928100585938,
"loss": 0.4807,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0535998344421387,
"rewards/margins": 1.1148144006729126,
"rewards/rejected": -4.16841459274292,
"step": 1288
},
{
"epoch": 0.8418646420115928,
"grad_norm": 95.33465590024296,
"learning_rate": 1.1143802672689772e-08,
"logits/chosen": -1.5128263235092163,
"logits/rejected": -1.5237712860107422,
"logps/chosen": -804.3637084960938,
"logps/rejected": -936.9696044921875,
"loss": 0.4619,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.063973903656006,
"rewards/margins": 0.989101231098175,
"rewards/rejected": -4.053074836730957,
"step": 1289
},
{
"epoch": 0.8425177565515553,
"grad_norm": 18.893079688911747,
"learning_rate": 1.1054222904982346e-08,
"logits/chosen": -1.4855834245681763,
"logits/rejected": -1.4457110166549683,
"logps/chosen": -862.0806274414062,
"logps/rejected": -923.0704956054688,
"loss": 0.4819,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3044419288635254,
"rewards/margins": 1.0387816429138184,
"rewards/rejected": -4.343223571777344,
"step": 1290
},
{
"epoch": 0.8431708710915177,
"grad_norm": 36.3373091887223,
"learning_rate": 1.0964975983431116e-08,
"logits/chosen": -1.5048320293426514,
"logits/rejected": -1.5244179964065552,
"logps/chosen": -848.87548828125,
"logps/rejected": -917.09619140625,
"loss": 0.4807,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.414567708969116,
"rewards/margins": 0.6657711267471313,
"rewards/rejected": -4.080338954925537,
"step": 1291
},
{
"epoch": 0.8438239856314801,
"grad_norm": 41.15768005703854,
"learning_rate": 1.0876062372578e-08,
"logits/chosen": -1.4884114265441895,
"logits/rejected": -1.495098352432251,
"logps/chosen": -854.0869750976562,
"logps/rejected": -878.1396484375,
"loss": 0.5553,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.5928542613983154,
"rewards/margins": 0.4925929307937622,
"rewards/rejected": -4.085447311401367,
"step": 1292
},
{
"epoch": 0.8444771001714426,
"grad_norm": 36.42635614805045,
"learning_rate": 1.0787482535230022e-08,
"logits/chosen": -1.4687424898147583,
"logits/rejected": -1.4143062829971313,
"logps/chosen": -813.6422729492188,
"logps/rejected": -855.9901123046875,
"loss": 0.516,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0875415802001953,
"rewards/margins": 0.5753628015518188,
"rewards/rejected": -3.6629042625427246,
"step": 1293
},
{
"epoch": 0.845130214711405,
"grad_norm": 29.828612800376433,
"learning_rate": 1.0699236932456835e-08,
"logits/chosen": -1.6260933876037598,
"logits/rejected": -1.5973844528198242,
"logps/chosen": -855.070068359375,
"logps/rejected": -896.7317504882812,
"loss": 0.5569,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.386779546737671,
"rewards/margins": 0.6484910249710083,
"rewards/rejected": -4.035270690917969,
"step": 1294
},
{
"epoch": 0.8457833292513675,
"grad_norm": 42.20579198570261,
"learning_rate": 1.0611326023588388e-08,
"logits/chosen": -1.4956504106521606,
"logits/rejected": -1.4156677722930908,
"logps/chosen": -786.1782836914062,
"logps/rejected": -816.1480102539062,
"loss": 0.563,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.0679595470428467,
"rewards/margins": 0.5685796141624451,
"rewards/rejected": -3.6365387439727783,
"step": 1295
},
{
"epoch": 0.8464364437913299,
"grad_norm": 13.728489619296084,
"learning_rate": 1.0523750266212483e-08,
"logits/chosen": -1.5465428829193115,
"logits/rejected": -1.500390887260437,
"logps/chosen": -882.738525390625,
"logps/rejected": -909.25634765625,
"loss": 0.5011,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.744199752807617,
"rewards/margins": 0.5797795653343201,
"rewards/rejected": -4.323978900909424,
"step": 1296
},
{
"epoch": 0.8470895583312924,
"grad_norm": 16.665414724980934,
"learning_rate": 1.0436510116172425e-08,
"logits/chosen": -1.4301414489746094,
"logits/rejected": -1.4165202379226685,
"logps/chosen": -852.8707275390625,
"logps/rejected": -919.6453857421875,
"loss": 0.5169,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.4321398735046387,
"rewards/margins": 0.7153012156486511,
"rewards/rejected": -4.1474409103393555,
"step": 1297
},
{
"epoch": 0.8477426728712548,
"grad_norm": 51.89482440292027,
"learning_rate": 1.0349606027564633e-08,
"logits/chosen": -1.4815446138381958,
"logits/rejected": -1.4437528848648071,
"logps/chosen": -808.208984375,
"logps/rejected": -885.4605712890625,
"loss": 0.4939,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.99233078956604,
"rewards/margins": 0.8016055822372437,
"rewards/rejected": -3.7939364910125732,
"step": 1298
},
{
"epoch": 0.8483957874112172,
"grad_norm": 47.61880788832745,
"learning_rate": 1.0263038452736292e-08,
"logits/chosen": -1.4789763689041138,
"logits/rejected": -1.4989715814590454,
"logps/chosen": -751.89453125,
"logps/rejected": -971.8363647460938,
"loss": 0.4495,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.0481152534484863,
"rewards/margins": 1.4892245531082153,
"rewards/rejected": -4.53734016418457,
"step": 1299
},
{
"epoch": 0.8490489019511797,
"grad_norm": 12.835741715040568,
"learning_rate": 1.0176807842282977e-08,
"logits/chosen": -1.4237325191497803,
"logits/rejected": -1.4008545875549316,
"logps/chosen": -822.1392211914062,
"logps/rejected": -825.3770751953125,
"loss": 0.5182,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1958327293395996,
"rewards/margins": 0.45696282386779785,
"rewards/rejected": -3.6527957916259766,
"step": 1300
},
{
"epoch": 0.8490489019511797,
"eval_logits/chosen": -1.4787728786468506,
"eval_logits/rejected": -1.4600926637649536,
"eval_logps/chosen": -842.8944702148438,
"eval_logps/rejected": -926.5396118164062,
"eval_loss": 0.49515822529792786,
"eval_rewards/accuracies": 0.7599999904632568,
"eval_rewards/chosen": -3.3254919052124023,
"eval_rewards/margins": 0.9175096750259399,
"eval_rewards/rejected": -4.243001461029053,
"eval_runtime": 296.4556,
"eval_samples_per_second": 13.493,
"eval_steps_per_second": 0.843,
"step": 1300
},
{
"epoch": 0.8497020164911422,
"grad_norm": 66.58487284404727,
"learning_rate": 1.009091464504633e-08,
"logits/chosen": -1.4258389472961426,
"logits/rejected": -1.3811410665512085,
"logps/chosen": -866.520751953125,
"logps/rejected": -932.500244140625,
"loss": 0.5058,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.8197102546691895,
"rewards/margins": 0.9209631085395813,
"rewards/rejected": -4.740673542022705,
"step": 1301
},
{
"epoch": 0.8503551310311046,
"grad_norm": 15.745580341420917,
"learning_rate": 1.0005359308111702e-08,
"logits/chosen": -1.535689353942871,
"logits/rejected": -1.5087151527404785,
"logps/chosen": -806.9986572265625,
"logps/rejected": -910.8426513671875,
"loss": 0.4717,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.248060941696167,
"rewards/margins": 1.305046558380127,
"rewards/rejected": -4.553107738494873,
"step": 1302
},
{
"epoch": 0.851008245571067,
"grad_norm": 62.332572039035284,
"learning_rate": 9.920142276805852e-09,
"logits/chosen": -1.465049386024475,
"logits/rejected": -1.4151432514190674,
"logps/chosen": -927.4882202148438,
"logps/rejected": -980.1682739257812,
"loss": 0.5446,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.472818374633789,
"rewards/margins": 0.7706753611564636,
"rewards/rejected": -4.243494033813477,
"step": 1303
},
{
"epoch": 0.8516613601110294,
"grad_norm": 33.565372393934965,
"learning_rate": 9.835263994694587e-09,
"logits/chosen": -1.46824049949646,
"logits/rejected": -1.4706170558929443,
"logps/chosen": -960.3832397460938,
"logps/rejected": -1062.023681640625,
"loss": 0.4433,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.9557857513427734,
"rewards/margins": 0.9899990558624268,
"rewards/rejected": -4.945785045623779,
"step": 1304
},
{
"epoch": 0.852314474650992,
"grad_norm": 25.81469581412769,
"learning_rate": 9.750724903580503e-09,
"logits/chosen": -1.5085800886154175,
"logits/rejected": -1.5068165063858032,
"logps/chosen": -845.734130859375,
"logps/rejected": -906.9862060546875,
"loss": 0.4746,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.1564383506774902,
"rewards/margins": 0.7602788209915161,
"rewards/rejected": -3.916717052459717,
"step": 1305
},
{
"epoch": 0.8529675891909544,
"grad_norm": 17.54341908103785,
"learning_rate": 9.666525443500666e-09,
"logits/chosen": -1.5109413862228394,
"logits/rejected": -1.397308349609375,
"logps/chosen": -848.3529052734375,
"logps/rejected": -978.5413818359375,
"loss": 0.5116,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.651090145111084,
"rewards/margins": 1.1326894760131836,
"rewards/rejected": -4.783779144287109,
"step": 1306
},
{
"epoch": 0.8536207037309168,
"grad_norm": 17.25055458777116,
"learning_rate": 9.582666052724305e-09,
"logits/chosen": -1.5393823385238647,
"logits/rejected": -1.5182414054870605,
"logps/chosen": -882.0748901367188,
"logps/rejected": -946.6273803710938,
"loss": 0.5355,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4100325107574463,
"rewards/margins": 0.5562119483947754,
"rewards/rejected": -3.9662444591522217,
"step": 1307
},
{
"epoch": 0.8542738182708792,
"grad_norm": 16.192143577334363,
"learning_rate": 9.499147167750541e-09,
"logits/chosen": -1.425965428352356,
"logits/rejected": -1.4162306785583496,
"logps/chosen": -811.1448364257812,
"logps/rejected": -926.71435546875,
"loss": 0.4845,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.059163808822632,
"rewards/margins": 0.9071324467658997,
"rewards/rejected": -3.9662961959838867,
"step": 1308
},
{
"epoch": 0.8549269328108418,
"grad_norm": 74.71761901774033,
"learning_rate": 9.415969223306133e-09,
"logits/chosen": -1.4104845523834229,
"logits/rejected": -1.4388418197631836,
"logps/chosen": -902.40380859375,
"logps/rejected": -968.486083984375,
"loss": 0.5375,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.3599321842193604,
"rewards/margins": 0.6312100291252136,
"rewards/rejected": -3.9911422729492188,
"step": 1309
},
{
"epoch": 0.8555800473508042,
"grad_norm": 33.85232737585008,
"learning_rate": 9.333132652343193e-09,
"logits/chosen": -1.5950632095336914,
"logits/rejected": -1.5360954999923706,
"logps/chosen": -822.3123779296875,
"logps/rejected": -889.439697265625,
"loss": 0.4481,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.34323787689209,
"rewards/margins": 1.0119768381118774,
"rewards/rejected": -4.355214595794678,
"step": 1310
},
{
"epoch": 0.8562331618907666,
"grad_norm": 64.61699538785297,
"learning_rate": 9.250637886036913e-09,
"logits/chosen": -1.4856970310211182,
"logits/rejected": -1.4961824417114258,
"logps/chosen": -820.3836059570312,
"logps/rejected": -874.966796875,
"loss": 0.4821,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.9959676265716553,
"rewards/margins": 0.6477092504501343,
"rewards/rejected": -3.6436767578125,
"step": 1311
},
{
"epoch": 0.856886276430729,
"grad_norm": 49.59060889243429,
"learning_rate": 9.16848535378339e-09,
"logits/chosen": -1.4910832643508911,
"logits/rejected": -1.4679546356201172,
"logps/chosen": -820.6103515625,
"logps/rejected": -934.1729125976562,
"loss": 0.4813,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2271547317504883,
"rewards/margins": 1.107942819595337,
"rewards/rejected": -4.335097312927246,
"step": 1312
},
{
"epoch": 0.8575393909706915,
"grad_norm": 67.84816938025705,
"learning_rate": 9.086675483197323e-09,
"logits/chosen": -1.4484561681747437,
"logits/rejected": -1.424304485321045,
"logps/chosen": -858.7589111328125,
"logps/rejected": -946.8106079101562,
"loss": 0.4282,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.125239849090576,
"rewards/margins": 0.7515519857406616,
"rewards/rejected": -3.876791477203369,
"step": 1313
},
{
"epoch": 0.858192505510654,
"grad_norm": 25.662991507947012,
"learning_rate": 9.005208700109817e-09,
"logits/chosen": -1.4877333641052246,
"logits/rejected": -1.4774301052093506,
"logps/chosen": -765.2407836914062,
"logps/rejected": -906.0118408203125,
"loss": 0.4055,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.1628265380859375,
"rewards/margins": 1.3297454118728638,
"rewards/rejected": -4.492571830749512,
"step": 1314
},
{
"epoch": 0.8588456200506164,
"grad_norm": 127.58031473226492,
"learning_rate": 8.924085428566163e-09,
"logits/chosen": -1.4710652828216553,
"logits/rejected": -1.4731792211532593,
"logps/chosen": -753.9927978515625,
"logps/rejected": -920.06005859375,
"loss": 0.5066,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.049560070037842,
"rewards/margins": 0.948535680770874,
"rewards/rejected": -3.998095750808716,
"step": 1315
},
{
"epoch": 0.8594987345905788,
"grad_norm": 45.1109159351009,
"learning_rate": 8.843306090823632e-09,
"logits/chosen": -1.45258367061615,
"logits/rejected": -1.4658076763153076,
"logps/chosen": -828.3963012695312,
"logps/rejected": -949.1826171875,
"loss": 0.4856,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.294034481048584,
"rewards/margins": 0.9175934791564941,
"rewards/rejected": -4.211627960205078,
"step": 1316
},
{
"epoch": 0.8601518491305413,
"grad_norm": 29.986901265876828,
"learning_rate": 8.762871107349267e-09,
"logits/chosen": -1.511345386505127,
"logits/rejected": -1.4521244764328003,
"logps/chosen": -778.9940795898438,
"logps/rejected": -893.1011962890625,
"loss": 0.5278,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.323246479034424,
"rewards/margins": 1.0293022394180298,
"rewards/rejected": -4.352548599243164,
"step": 1317
},
{
"epoch": 0.8608049636705037,
"grad_norm": 11.35851659183948,
"learning_rate": 8.682780896817716e-09,
"logits/chosen": -1.3951612710952759,
"logits/rejected": -1.4687902927398682,
"logps/chosen": -848.2562866210938,
"logps/rejected": -946.0987548828125,
"loss": 0.5036,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4302592277526855,
"rewards/margins": 0.6691716909408569,
"rewards/rejected": -4.099431037902832,
"step": 1318
},
{
"epoch": 0.8614580782104662,
"grad_norm": 11.967451850970171,
"learning_rate": 8.603035876109013e-09,
"logits/chosen": -1.5124762058258057,
"logits/rejected": -1.4754977226257324,
"logps/chosen": -814.0156860351562,
"logps/rejected": -796.652587890625,
"loss": 0.4813,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.291874647140503,
"rewards/margins": 0.5844862461090088,
"rewards/rejected": -3.8763608932495117,
"step": 1319
},
{
"epoch": 0.8621111927504286,
"grad_norm": 120.18322878609834,
"learning_rate": 8.523636460306463e-09,
"logits/chosen": -1.4285402297973633,
"logits/rejected": -1.4294164180755615,
"logps/chosen": -782.0220336914062,
"logps/rejected": -898.1871948242188,
"loss": 0.5255,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4391157627105713,
"rewards/margins": 0.8806977868080139,
"rewards/rejected": -4.3198137283325195,
"step": 1320
},
{
"epoch": 0.862764307290391,
"grad_norm": 48.61379884717116,
"learning_rate": 8.444583062694439e-09,
"logits/chosen": -1.3637633323669434,
"logits/rejected": -1.3878041505813599,
"logps/chosen": -757.3119506835938,
"logps/rejected": -860.0999145507812,
"loss": 0.5276,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.0247325897216797,
"rewards/margins": 0.7837187051773071,
"rewards/rejected": -3.8084514141082764,
"step": 1321
},
{
"epoch": 0.8634174218303535,
"grad_norm": 11.712512451580766,
"learning_rate": 8.365876094756228e-09,
"logits/chosen": -1.5599374771118164,
"logits/rejected": -1.5587773323059082,
"logps/chosen": -911.1922607421875,
"logps/rejected": -981.4324340820312,
"loss": 0.5147,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.283015251159668,
"rewards/margins": 0.8019892573356628,
"rewards/rejected": -4.085004806518555,
"step": 1322
},
{
"epoch": 0.864070536370316,
"grad_norm": 131.68861126525354,
"learning_rate": 8.287515966171928e-09,
"logits/chosen": -1.5320029258728027,
"logits/rejected": -1.516080379486084,
"logps/chosen": -924.517822265625,
"logps/rejected": -953.4563598632812,
"loss": 0.5502,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6483347415924072,
"rewards/margins": 0.7624077796936035,
"rewards/rejected": -4.41074275970459,
"step": 1323
},
{
"epoch": 0.8647236509102784,
"grad_norm": 24.356458817864755,
"learning_rate": 8.209503084816285e-09,
"logits/chosen": -1.546379566192627,
"logits/rejected": -1.4577432870864868,
"logps/chosen": -800.8156127929688,
"logps/rejected": -880.213134765625,
"loss": 0.4574,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0747878551483154,
"rewards/margins": 1.0365558862686157,
"rewards/rejected": -4.1113433837890625,
"step": 1324
},
{
"epoch": 0.8653767654502408,
"grad_norm": 33.56562789954909,
"learning_rate": 8.131837856756585e-09,
"logits/chosen": -1.4879792928695679,
"logits/rejected": -1.4731531143188477,
"logps/chosen": -843.258056640625,
"logps/rejected": -948.2938232421875,
"loss": 0.5075,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.485196828842163,
"rewards/margins": 1.1023627519607544,
"rewards/rejected": -4.587559223175049,
"step": 1325
},
{
"epoch": 0.8660298799902033,
"grad_norm": 60.746333244964944,
"learning_rate": 8.054520686250512e-09,
"logits/chosen": -1.4594736099243164,
"logits/rejected": -1.4537848234176636,
"logps/chosen": -705.6865844726562,
"logps/rejected": -833.6557006835938,
"loss": 0.4751,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.902322769165039,
"rewards/margins": 0.7870765924453735,
"rewards/rejected": -3.689399242401123,
"step": 1326
},
{
"epoch": 0.8666829945301657,
"grad_norm": 27.41375334113206,
"learning_rate": 7.977551975744088e-09,
"logits/chosen": -1.5716170072555542,
"logits/rejected": -1.5631660223007202,
"logps/chosen": -939.8978271484375,
"logps/rejected": -984.3545532226562,
"loss": 0.5208,
"rewards/accuracies": 0.6875,
"rewards/chosen": -4.000885963439941,
"rewards/margins": 0.5963870286941528,
"rewards/rejected": -4.597273349761963,
"step": 1327
},
{
"epoch": 0.8673361090701281,
"grad_norm": 16.93864295624459,
"learning_rate": 7.900932125869545e-09,
"logits/chosen": -1.4590840339660645,
"logits/rejected": -1.4381649494171143,
"logps/chosen": -835.8486938476562,
"logps/rejected": -1007.2196655273438,
"loss": 0.5122,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.375314712524414,
"rewards/margins": 1.4108258485794067,
"rewards/rejected": -4.786140441894531,
"step": 1328
},
{
"epoch": 0.8679892236100906,
"grad_norm": 28.716003483649484,
"learning_rate": 7.824661535443247e-09,
"logits/chosen": -1.6202168464660645,
"logits/rejected": -1.5609275102615356,
"logps/chosen": -914.7964477539062,
"logps/rejected": -1054.38525390625,
"loss": 0.5092,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.545833110809326,
"rewards/margins": 1.0506269931793213,
"rewards/rejected": -4.596460342407227,
"step": 1329
},
{
"epoch": 0.8686423381500531,
"grad_norm": 41.01865127499047,
"learning_rate": 7.748740601463622e-09,
"logits/chosen": -1.48884117603302,
"logits/rejected": -1.5338010787963867,
"logps/chosen": -731.7664794921875,
"logps/rejected": -812.95458984375,
"loss": 0.4394,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.1326169967651367,
"rewards/margins": 0.877723217010498,
"rewards/rejected": -4.010340213775635,
"step": 1330
},
{
"epoch": 0.8692954526900155,
"grad_norm": 64.11269025670589,
"learning_rate": 7.673169719109091e-09,
"logits/chosen": -1.4994860887527466,
"logits/rejected": -1.4728403091430664,
"logps/chosen": -873.9459228515625,
"logps/rejected": -884.364501953125,
"loss": 0.5138,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.492875337600708,
"rewards/margins": 0.4393343925476074,
"rewards/rejected": -3.9322092533111572,
"step": 1331
},
{
"epoch": 0.8699485672299779,
"grad_norm": 38.41559411716,
"learning_rate": 7.597949281736019e-09,
"logits/chosen": -1.5339343547821045,
"logits/rejected": -1.4958739280700684,
"logps/chosen": -808.7354125976562,
"logps/rejected": -932.888916015625,
"loss": 0.5561,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.077152967453003,
"rewards/margins": 1.2154319286346436,
"rewards/rejected": -4.2925848960876465,
"step": 1332
},
{
"epoch": 0.8706016817699403,
"grad_norm": 12.205155892018915,
"learning_rate": 7.523079680876613e-09,
"logits/chosen": -1.416054368019104,
"logits/rejected": -1.377394676208496,
"logps/chosen": -824.678955078125,
"logps/rejected": -965.9581909179688,
"loss": 0.4893,
"rewards/accuracies": 0.78125,
"rewards/chosen": -4.192705154418945,
"rewards/margins": 0.9551449418067932,
"rewards/rejected": -5.147850036621094,
"step": 1333
},
{
"epoch": 0.8712547963099029,
"grad_norm": 57.99000568713857,
"learning_rate": 7.448561306236989e-09,
"logits/chosen": -1.4369691610336304,
"logits/rejected": -1.4728755950927734,
"logps/chosen": -964.21435546875,
"logps/rejected": -1086.426513671875,
"loss": 0.4682,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6197071075439453,
"rewards/margins": 1.294512391090393,
"rewards/rejected": -4.914219379425049,
"step": 1334
},
{
"epoch": 0.8719079108498653,
"grad_norm": 20.044699015378878,
"learning_rate": 7.374394545695062e-09,
"logits/chosen": -1.4569401741027832,
"logits/rejected": -1.4874026775360107,
"logps/chosen": -895.0442504882812,
"logps/rejected": -896.1160888671875,
"loss": 0.4917,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.06990385055542,
"rewards/margins": 0.44757652282714844,
"rewards/rejected": -3.5174806118011475,
"step": 1335
},
{
"epoch": 0.8725610253898277,
"grad_norm": 37.40096323479501,
"learning_rate": 7.300579785298516e-09,
"logits/chosen": -1.5550789833068848,
"logits/rejected": -1.561784029006958,
"logps/chosen": -865.9757690429688,
"logps/rejected": -1000.7755126953125,
"loss": 0.461,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.3865833282470703,
"rewards/margins": 0.9827724695205688,
"rewards/rejected": -4.36935567855835,
"step": 1336
},
{
"epoch": 0.8732141399297901,
"grad_norm": 10.447772657085043,
"learning_rate": 7.227117409262912e-09,
"logits/chosen": -1.477044701576233,
"logits/rejected": -1.4742741584777832,
"logps/chosen": -804.1036987304688,
"logps/rejected": -945.7398681640625,
"loss": 0.4617,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0770020484924316,
"rewards/margins": 1.022882342338562,
"rewards/rejected": -4.099884033203125,
"step": 1337
},
{
"epoch": 0.8738672544697527,
"grad_norm": 16.999701561801512,
"learning_rate": 7.154007799969517e-09,
"logits/chosen": -1.5495110750198364,
"logits/rejected": -1.5202927589416504,
"logps/chosen": -835.4287109375,
"logps/rejected": -846.221435546875,
"loss": 0.4887,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.041454315185547,
"rewards/margins": 0.6531744003295898,
"rewards/rejected": -3.6946287155151367,
"step": 1338
},
{
"epoch": 0.8745203690097151,
"grad_norm": 93.85006471838479,
"learning_rate": 7.081251337963442e-09,
"logits/chosen": -1.581192970275879,
"logits/rejected": -1.5974817276000977,
"logps/chosen": -931.1429443359375,
"logps/rejected": -1022.624755859375,
"loss": 0.4658,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.4839534759521484,
"rewards/margins": 1.0064407587051392,
"rewards/rejected": -4.490394115447998,
"step": 1339
},
{
"epoch": 0.8751734835496775,
"grad_norm": 57.219568967053746,
"learning_rate": 7.008848401951622e-09,
"logits/chosen": -1.449042558670044,
"logits/rejected": -1.4050977230072021,
"logps/chosen": -781.509521484375,
"logps/rejected": -854.7266845703125,
"loss": 0.5102,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.965616226196289,
"rewards/margins": 0.8889732956886292,
"rewards/rejected": -3.8545899391174316,
"step": 1340
},
{
"epoch": 0.8758265980896399,
"grad_norm": 73.1771564404403,
"learning_rate": 6.9367993688008195e-09,
"logits/chosen": -1.4334369897842407,
"logits/rejected": -1.4349690675735474,
"logps/chosen": -838.9798583984375,
"logps/rejected": -894.887939453125,
"loss": 0.4708,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.614448070526123,
"rewards/margins": 0.6810629367828369,
"rewards/rejected": -4.295510768890381,
"step": 1341
},
{
"epoch": 0.8764797126296024,
"grad_norm": 88.67111576311824,
"learning_rate": 6.865104613535719e-09,
"logits/chosen": -1.5720555782318115,
"logits/rejected": -1.556181788444519,
"logps/chosen": -1001.600830078125,
"logps/rejected": -1061.5048828125,
"loss": 0.5376,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.814486503601074,
"rewards/margins": 0.7989178895950317,
"rewards/rejected": -4.613404273986816,
"step": 1342
},
{
"epoch": 0.8771328271695649,
"grad_norm": 107.66602305396016,
"learning_rate": 6.7937645093369076e-09,
"logits/chosen": -1.5547233819961548,
"logits/rejected": -1.5719074010849,
"logps/chosen": -969.04296875,
"logps/rejected": -1082.557373046875,
"loss": 0.4579,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.629171371459961,
"rewards/margins": 1.1535873413085938,
"rewards/rejected": -4.782759189605713,
"step": 1343
},
{
"epoch": 0.8777859417095273,
"grad_norm": 12.704339641522967,
"learning_rate": 6.722779427539007e-09,
"logits/chosen": -1.5757780075073242,
"logits/rejected": -1.5685677528381348,
"logps/chosen": -892.2096557617188,
"logps/rejected": -1070.7259521484375,
"loss": 0.5011,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5382652282714844,
"rewards/margins": 1.3167110681533813,
"rewards/rejected": -4.854976654052734,
"step": 1344
},
{
"epoch": 0.8784390562494897,
"grad_norm": 74.90456047914228,
"learning_rate": 6.6521497376286425e-09,
"logits/chosen": -1.4474490880966187,
"logits/rejected": -1.393038272857666,
"logps/chosen": -863.8945922851562,
"logps/rejected": -985.6024169921875,
"loss": 0.525,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4975996017456055,
"rewards/margins": 1.0218359231948853,
"rewards/rejected": -4.519434928894043,
"step": 1345
},
{
"epoch": 0.8790921707894522,
"grad_norm": 16.34267655213826,
"learning_rate": 6.581875807242643e-09,
"logits/chosen": -1.4592047929763794,
"logits/rejected": -1.4397943019866943,
"logps/chosen": -822.5889892578125,
"logps/rejected": -898.7081909179688,
"loss": 0.4558,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.987348794937134,
"rewards/margins": 0.8248506784439087,
"rewards/rejected": -3.812199115753174,
"step": 1346
},
{
"epoch": 0.8797452853294147,
"grad_norm": 16.0782479464605,
"learning_rate": 6.51195800216601e-09,
"logits/chosen": -1.5585129261016846,
"logits/rejected": -1.5897541046142578,
"logps/chosen": -826.855224609375,
"logps/rejected": -982.5360717773438,
"loss": 0.4817,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.9745357036590576,
"rewards/margins": 0.9419158697128296,
"rewards/rejected": -3.9164514541625977,
"step": 1347
},
{
"epoch": 0.8803983998693771,
"grad_norm": 15.944369178445228,
"learning_rate": 6.442396686330104e-09,
"logits/chosen": -1.5734515190124512,
"logits/rejected": -1.5492303371429443,
"logps/chosen": -877.200927734375,
"logps/rejected": -935.7286987304688,
"loss": 0.4639,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.669853925704956,
"rewards/margins": 0.9167975187301636,
"rewards/rejected": -4.58665132522583,
"step": 1348
},
{
"epoch": 0.8810515144093395,
"grad_norm": 12.702778495003988,
"learning_rate": 6.373192221810694e-09,
"logits/chosen": -1.5214407444000244,
"logits/rejected": -1.5045905113220215,
"logps/chosen": -798.3970947265625,
"logps/rejected": -847.7827758789062,
"loss": 0.4624,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3049206733703613,
"rewards/margins": 0.854720950126648,
"rewards/rejected": -4.159641742706299,
"step": 1349
},
{
"epoch": 0.881704628949302,
"grad_norm": 100.6971315662948,
"learning_rate": 6.304344968826094e-09,
"logits/chosen": -1.4444749355316162,
"logits/rejected": -1.461186170578003,
"logps/chosen": -823.1656494140625,
"logps/rejected": -854.8773193359375,
"loss": 0.4764,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.2248125076293945,
"rewards/margins": 0.8384161591529846,
"rewards/rejected": -4.063228607177734,
"step": 1350
},
{
"epoch": 0.8823577434892644,
"grad_norm": 28.966983505349088,
"learning_rate": 6.235855285735289e-09,
"logits/chosen": -1.373712182044983,
"logits/rejected": -1.3527984619140625,
"logps/chosen": -833.9160766601562,
"logps/rejected": -975.32080078125,
"loss": 0.4452,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4645469188690186,
"rewards/margins": 1.1589038372039795,
"rewards/rejected": -4.623451232910156,
"step": 1351
},
{
"epoch": 0.8830108580292269,
"grad_norm": 11.232677637593673,
"learning_rate": 6.167723529036051e-09,
"logits/chosen": -1.4311108589172363,
"logits/rejected": -1.4204224348068237,
"logps/chosen": -745.5303955078125,
"logps/rejected": -826.6415405273438,
"loss": 0.556,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9358832836151123,
"rewards/margins": 0.5993090271949768,
"rewards/rejected": -3.5351924896240234,
"step": 1352
},
{
"epoch": 0.8836639725691893,
"grad_norm": 52.83410951930048,
"learning_rate": 6.099950053363109e-09,
"logits/chosen": -1.5700377225875854,
"logits/rejected": -1.5207599401474,
"logps/chosen": -897.0535888671875,
"logps/rejected": -950.018798828125,
"loss": 0.4828,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4924488067626953,
"rewards/margins": 0.8256914019584656,
"rewards/rejected": -4.318140506744385,
"step": 1353
},
{
"epoch": 0.8843170871091518,
"grad_norm": 48.439495319236386,
"learning_rate": 6.032535211486303e-09,
"logits/chosen": -1.5001033544540405,
"logits/rejected": -1.4997605085372925,
"logps/chosen": -765.0709838867188,
"logps/rejected": -830.47900390625,
"loss": 0.4423,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.0950000286102295,
"rewards/margins": 0.8747947216033936,
"rewards/rejected": -3.969794273376465,
"step": 1354
},
{
"epoch": 0.8849702016491142,
"grad_norm": 56.98461717362285,
"learning_rate": 5.965479354308739e-09,
"logits/chosen": -1.4557162523269653,
"logits/rejected": -1.4349863529205322,
"logps/chosen": -910.533935546875,
"logps/rejected": -992.22998046875,
"loss": 0.5308,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4078073501586914,
"rewards/margins": 0.8652389049530029,
"rewards/rejected": -4.273046016693115,
"step": 1355
},
{
"epoch": 0.8856233161890766,
"grad_norm": 24.54186584954266,
"learning_rate": 5.898782830864909e-09,
"logits/chosen": -1.51798415184021,
"logits/rejected": -1.564520001411438,
"logps/chosen": -840.780029296875,
"logps/rejected": -975.0224609375,
"loss": 0.4677,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.289569139480591,
"rewards/margins": 1.0400049686431885,
"rewards/rejected": -4.329574108123779,
"step": 1356
},
{
"epoch": 0.886276430729039,
"grad_norm": 104.63221608787276,
"learning_rate": 5.832445988318996e-09,
"logits/chosen": -1.552783489227295,
"logits/rejected": -1.462197184562683,
"logps/chosen": -885.2568969726562,
"logps/rejected": -968.8541259765625,
"loss": 0.5129,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.7370312213897705,
"rewards/margins": 1.4136296510696411,
"rewards/rejected": -5.150660514831543,
"step": 1357
},
{
"epoch": 0.8869295452690016,
"grad_norm": 45.185665578640275,
"learning_rate": 5.766469171962943e-09,
"logits/chosen": -1.5153913497924805,
"logits/rejected": -1.4714546203613281,
"logps/chosen": -927.0983276367188,
"logps/rejected": -1037.369140625,
"loss": 0.4428,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4304184913635254,
"rewards/margins": 0.805862307548523,
"rewards/rejected": -4.23628044128418,
"step": 1358
},
{
"epoch": 0.887582659808964,
"grad_norm": 79.59548680251643,
"learning_rate": 5.7008527252147525e-09,
"logits/chosen": -1.5614638328552246,
"logits/rejected": -1.5693732500076294,
"logps/chosen": -825.9854736328125,
"logps/rejected": -924.3582763671875,
"loss": 0.4688,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.293821096420288,
"rewards/margins": 1.0692007541656494,
"rewards/rejected": -4.363021373748779,
"step": 1359
},
{
"epoch": 0.8882357743489264,
"grad_norm": 17.890553860709666,
"learning_rate": 5.635596989616628e-09,
"logits/chosen": -1.3857258558273315,
"logits/rejected": -1.3223706483840942,
"logps/chosen": -842.4962768554688,
"logps/rejected": -924.7880859375,
"loss": 0.5206,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.7687184810638428,
"rewards/margins": 0.9322965145111084,
"rewards/rejected": -4.701014518737793,
"step": 1360
},
{
"epoch": 0.8888888888888888,
"grad_norm": 86.3211811567686,
"learning_rate": 5.570702304833225e-09,
"logits/chosen": -1.4879322052001953,
"logits/rejected": -1.5010827779769897,
"logps/chosen": -828.9422607421875,
"logps/rejected": -911.2158203125,
"loss": 0.482,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.6344053745269775,
"rewards/margins": 0.5890753865242004,
"rewards/rejected": -4.223480701446533,
"step": 1361
},
{
"epoch": 0.8895420034288514,
"grad_norm": 44.45040636623173,
"learning_rate": 5.5061690086498995e-09,
"logits/chosen": -1.60628342628479,
"logits/rejected": -1.5504629611968994,
"logps/chosen": -918.0870971679688,
"logps/rejected": -938.4238891601562,
"loss": 0.5785,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.70293927192688,
"rewards/margins": 0.504538357257843,
"rewards/rejected": -4.207477569580078,
"step": 1362
},
{
"epoch": 0.8901951179688138,
"grad_norm": 59.14938077433194,
"learning_rate": 5.441997436970908e-09,
"logits/chosen": -1.5273542404174805,
"logits/rejected": -1.4696133136749268,
"logps/chosen": -905.05615234375,
"logps/rejected": -1024.7708740234375,
"loss": 0.5164,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6890876293182373,
"rewards/margins": 0.9259529709815979,
"rewards/rejected": -4.6150407791137695,
"step": 1363
},
{
"epoch": 0.8908482325087762,
"grad_norm": 20.847094913779035,
"learning_rate": 5.3781879238177175e-09,
"logits/chosen": -1.4799872636795044,
"logits/rejected": -1.4763946533203125,
"logps/chosen": -818.2520751953125,
"logps/rejected": -900.87353515625,
"loss": 0.5318,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.162236213684082,
"rewards/margins": 0.6138532757759094,
"rewards/rejected": -3.776089668273926,
"step": 1364
},
{
"epoch": 0.8915013470487386,
"grad_norm": 42.086051676513534,
"learning_rate": 5.314740801327189e-09,
"logits/chosen": -1.5711402893066406,
"logits/rejected": -1.5064854621887207,
"logps/chosen": -800.728515625,
"logps/rejected": -949.0936889648438,
"loss": 0.4662,
"rewards/accuracies": 0.96875,
"rewards/chosen": -3.1598963737487793,
"rewards/margins": 1.2613778114318848,
"rewards/rejected": -4.421274662017822,
"step": 1365
},
{
"epoch": 0.8921544615887012,
"grad_norm": 36.43558502369589,
"learning_rate": 5.251656399749948e-09,
"logits/chosen": -1.5465190410614014,
"logits/rejected": -1.5165586471557617,
"logps/chosen": -931.0474853515625,
"logps/rejected": -1152.05126953125,
"loss": 0.4131,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.7468013763427734,
"rewards/margins": 1.6344844102859497,
"rewards/rejected": -5.381285667419434,
"step": 1366
},
{
"epoch": 0.8928075761286636,
"grad_norm": 18.023822032353266,
"learning_rate": 5.1889350474485425e-09,
"logits/chosen": -1.4559390544891357,
"logits/rejected": -1.4189265966415405,
"logps/chosen": -787.0001831054688,
"logps/rejected": -834.5303344726562,
"loss": 0.4982,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.818913698196411,
"rewards/margins": 0.7834856510162354,
"rewards/rejected": -3.6023998260498047,
"step": 1367
},
{
"epoch": 0.893460690668626,
"grad_norm": 51.00256674463102,
"learning_rate": 5.126577070895851e-09,
"logits/chosen": -1.4489341974258423,
"logits/rejected": -1.4130151271820068,
"logps/chosen": -933.874267578125,
"logps/rejected": -1004.5009765625,
"loss": 0.5661,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.800851345062256,
"rewards/margins": 1.1616989374160767,
"rewards/rejected": -4.962550163269043,
"step": 1368
},
{
"epoch": 0.8941138052085884,
"grad_norm": 28.042878667934062,
"learning_rate": 5.064582794673322e-09,
"logits/chosen": -1.5686874389648438,
"logits/rejected": -1.5284960269927979,
"logps/chosen": -913.930908203125,
"logps/rejected": -1030.9652099609375,
"loss": 0.5008,
"rewards/accuracies": 0.71875,
"rewards/chosen": -4.0387115478515625,
"rewards/margins": 0.8935470581054688,
"rewards/rejected": -4.9322590827941895,
"step": 1369
},
{
"epoch": 0.8947669197485509,
"grad_norm": 10.695407778573784,
"learning_rate": 5.002952541469296e-09,
"logits/chosen": -1.4896258115768433,
"logits/rejected": -1.4801172018051147,
"logps/chosen": -879.490966796875,
"logps/rejected": -959.708251953125,
"loss": 0.4915,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.568885564804077,
"rewards/margins": 0.942472517490387,
"rewards/rejected": -4.51135778427124,
"step": 1370
},
{
"epoch": 0.8954200342885134,
"grad_norm": 31.529008960795142,
"learning_rate": 4.941686632077316e-09,
"logits/chosen": -1.4891539812088013,
"logits/rejected": -1.4347100257873535,
"logps/chosen": -856.7740478515625,
"logps/rejected": -881.070068359375,
"loss": 0.4438,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.407636880874634,
"rewards/margins": 0.6649507284164429,
"rewards/rejected": -4.072587490081787,
"step": 1371
},
{
"epoch": 0.8960731488284758,
"grad_norm": 63.6933646160359,
"learning_rate": 4.880785385394481e-09,
"logits/chosen": -1.5834522247314453,
"logits/rejected": -1.5488102436065674,
"logps/chosen": -841.972900390625,
"logps/rejected": -880.1702880859375,
"loss": 0.457,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.852036714553833,
"rewards/margins": 1.1045963764190674,
"rewards/rejected": -3.956632614135742,
"step": 1372
},
{
"epoch": 0.8967262633684382,
"grad_norm": 69.78064834856058,
"learning_rate": 4.820249118419753e-09,
"logits/chosen": -1.4307173490524292,
"logits/rejected": -1.4592258930206299,
"logps/chosen": -751.7630615234375,
"logps/rejected": -990.3554077148438,
"loss": 0.4792,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.6149396896362305,
"rewards/margins": 1.3603920936584473,
"rewards/rejected": -3.9753315448760986,
"step": 1373
},
{
"epoch": 0.8973793779084007,
"grad_norm": 10.250335849217539,
"learning_rate": 4.760078146252369e-09,
"logits/chosen": -1.5606201887130737,
"logits/rejected": -1.5576801300048828,
"logps/chosen": -913.7922973632812,
"logps/rejected": -964.5801391601562,
"loss": 0.4753,
"rewards/accuracies": 0.75,
"rewards/chosen": -4.296699523925781,
"rewards/margins": 0.7379961013793945,
"rewards/rejected": -5.034695625305176,
"step": 1374
},
{
"epoch": 0.8980324924483631,
"grad_norm": 20.39345965692167,
"learning_rate": 4.7002727820901145e-09,
"logits/chosen": -1.45150887966156,
"logits/rejected": -1.4612077474594116,
"logps/chosen": -808.03466796875,
"logps/rejected": -928.4208984375,
"loss": 0.4667,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.283160448074341,
"rewards/margins": 1.0168074369430542,
"rewards/rejected": -4.2999677658081055,
"step": 1375
},
{
"epoch": 0.8986856069883256,
"grad_norm": 10.800412600192082,
"learning_rate": 4.640833337227754e-09,
"logits/chosen": -1.5039968490600586,
"logits/rejected": -1.4786089658737183,
"logps/chosen": -886.8804931640625,
"logps/rejected": -959.9457397460938,
"loss": 0.4385,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.3790817260742188,
"rewards/margins": 1.2774152755737305,
"rewards/rejected": -4.656497001647949,
"step": 1376
},
{
"epoch": 0.899338721528288,
"grad_norm": 103.40843027693559,
"learning_rate": 4.581760121055392e-09,
"logits/chosen": -1.5100741386413574,
"logits/rejected": -1.435150146484375,
"logps/chosen": -883.2802124023438,
"logps/rejected": -883.0489501953125,
"loss": 0.54,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.2741122245788574,
"rewards/margins": 0.6780447959899902,
"rewards/rejected": -3.9521570205688477,
"step": 1377
},
{
"epoch": 0.8999918360682505,
"grad_norm": 48.575124614734456,
"learning_rate": 4.523053441056876e-09,
"logits/chosen": -1.5189893245697021,
"logits/rejected": -1.5679491758346558,
"logps/chosen": -803.8499755859375,
"logps/rejected": -894.542236328125,
"loss": 0.4704,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.50295352935791,
"rewards/margins": 0.8048915266990662,
"rewards/rejected": -4.307845592498779,
"step": 1378
},
{
"epoch": 0.9006449506082129,
"grad_norm": 10.423098002326993,
"learning_rate": 4.4647136028081536e-09,
"logits/chosen": -1.5686522722244263,
"logits/rejected": -1.5366755723953247,
"logps/chosen": -841.1636352539062,
"logps/rejected": -963.812255859375,
"loss": 0.4655,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.380739688873291,
"rewards/margins": 1.0111637115478516,
"rewards/rejected": -4.391903400421143,
"step": 1379
},
{
"epoch": 0.9012980651481753,
"grad_norm": 103.35704441058242,
"learning_rate": 4.4067409099757505e-09,
"logits/chosen": -1.4515259265899658,
"logits/rejected": -1.421863317489624,
"logps/chosen": -816.9014282226562,
"logps/rejected": -830.080810546875,
"loss": 0.5235,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.6674067974090576,
"rewards/margins": 0.3336777687072754,
"rewards/rejected": -4.001084327697754,
"step": 1380
},
{
"epoch": 0.9019511796881378,
"grad_norm": 17.484965241800026,
"learning_rate": 4.349135664315137e-09,
"logits/chosen": -1.5792274475097656,
"logits/rejected": -1.5765178203582764,
"logps/chosen": -947.952392578125,
"logps/rejected": -1038.84326171875,
"loss": 0.4834,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.7614831924438477,
"rewards/margins": 0.8391762971878052,
"rewards/rejected": -4.6006598472595215,
"step": 1381
},
{
"epoch": 0.9026042942281003,
"grad_norm": 29.173220260017832,
"learning_rate": 4.291898165669155e-09,
"logits/chosen": -1.5122828483581543,
"logits/rejected": -1.5348204374313354,
"logps/chosen": -920.7978515625,
"logps/rejected": -1017.3441162109375,
"loss": 0.446,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.638453960418701,
"rewards/margins": 0.9234171509742737,
"rewards/rejected": -4.56187105178833,
"step": 1382
},
{
"epoch": 0.9032574087680627,
"grad_norm": 20.850486826793972,
"learning_rate": 4.235028711966512e-09,
"logits/chosen": -1.5382293462753296,
"logits/rejected": -1.5173406600952148,
"logps/chosen": -957.9508056640625,
"logps/rejected": -1045.537353515625,
"loss": 0.5194,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.8855299949645996,
"rewards/margins": 1.027876615524292,
"rewards/rejected": -4.9134063720703125,
"step": 1383
},
{
"epoch": 0.9039105233080251,
"grad_norm": 34.38670135334981,
"learning_rate": 4.178527599220164e-09,
"logits/chosen": -1.420579195022583,
"logits/rejected": -1.3815526962280273,
"logps/chosen": -813.62255859375,
"logps/rejected": -969.841552734375,
"loss": 0.5321,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.1666247844696045,
"rewards/margins": 1.1781651973724365,
"rewards/rejected": -4.344789981842041,
"step": 1384
},
{
"epoch": 0.9045636378479875,
"grad_norm": 68.615502539619,
"learning_rate": 4.122395121525807e-09,
"logits/chosen": -1.4724527597427368,
"logits/rejected": -1.4684042930603027,
"logps/chosen": -836.29638671875,
"logps/rejected": -971.5530395507812,
"loss": 0.5204,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.4262635707855225,
"rewards/margins": 0.8687800765037537,
"rewards/rejected": -4.2950439453125,
"step": 1385
},
{
"epoch": 0.9052167523879501,
"grad_norm": 50.58217070416385,
"learning_rate": 4.0666315710603585e-09,
"logits/chosen": -1.4584506750106812,
"logits/rejected": -1.438408613204956,
"logps/chosen": -815.32373046875,
"logps/rejected": -983.1445922851562,
"loss": 0.4398,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.262063503265381,
"rewards/margins": 1.1712214946746826,
"rewards/rejected": -4.433284759521484,
"step": 1386
},
{
"epoch": 0.9058698669279125,
"grad_norm": 37.43119604963389,
"learning_rate": 4.011237238080412e-09,
"logits/chosen": -1.47329843044281,
"logits/rejected": -1.5162948369979858,
"logps/chosen": -834.7694091796875,
"logps/rejected": -891.7135620117188,
"loss": 0.4209,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.565274238586426,
"rewards/margins": 0.6577186584472656,
"rewards/rejected": -4.222992897033691,
"step": 1387
},
{
"epoch": 0.9065229814678749,
"grad_norm": 20.4350578675936,
"learning_rate": 3.956212410920731e-09,
"logits/chosen": -1.5497958660125732,
"logits/rejected": -1.5459709167480469,
"logps/chosen": -957.7774047851562,
"logps/rejected": -1025.171142578125,
"loss": 0.4836,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.694035053253174,
"rewards/margins": 1.2108312845230103,
"rewards/rejected": -4.9048662185668945,
"step": 1388
},
{
"epoch": 0.9071760960078373,
"grad_norm": 124.44180200304133,
"learning_rate": 3.90155737599277e-09,
"logits/chosen": -1.532940149307251,
"logits/rejected": -1.521911859512329,
"logps/chosen": -958.362060546875,
"logps/rejected": -1038.1849365234375,
"loss": 0.4972,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.649043321609497,
"rewards/margins": 1.0646413564682007,
"rewards/rejected": -4.713685035705566,
"step": 1389
},
{
"epoch": 0.9078292105477999,
"grad_norm": 21.63423361819083,
"learning_rate": 3.847272417783129e-09,
"logits/chosen": -1.5962800979614258,
"logits/rejected": -1.5604323148727417,
"logps/chosen": -924.5335083007812,
"logps/rejected": -991.5548095703125,
"loss": 0.4995,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.826894760131836,
"rewards/margins": 0.9313570261001587,
"rewards/rejected": -4.758251190185547,
"step": 1390
},
{
"epoch": 0.9084823250877623,
"grad_norm": 33.36489378101839,
"learning_rate": 3.793357818852141e-09,
"logits/chosen": -1.5311145782470703,
"logits/rejected": -1.5007251501083374,
"logps/chosen": -857.7115478515625,
"logps/rejected": -936.7415161132812,
"loss": 0.5142,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.454068660736084,
"rewards/margins": 1.1231262683868408,
"rewards/rejected": -4.577195167541504,
"step": 1391
},
{
"epoch": 0.9091354396277247,
"grad_norm": 23.390479098008697,
"learning_rate": 3.739813859832383e-09,
"logits/chosen": -1.5484957695007324,
"logits/rejected": -1.5225454568862915,
"logps/chosen": -945.0482788085938,
"logps/rejected": -1007.5425415039062,
"loss": 0.5321,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.9478251934051514,
"rewards/margins": 0.5874274969100952,
"rewards/rejected": -4.535252571105957,
"step": 1392
},
{
"epoch": 0.9097885541676871,
"grad_norm": 27.975290638594313,
"learning_rate": 3.686640819427164e-09,
"logits/chosen": -1.4847687482833862,
"logits/rejected": -1.4688597917556763,
"logps/chosen": -885.1260986328125,
"logps/rejected": -946.1005859375,
"loss": 0.5318,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.731746196746826,
"rewards/margins": 0.6578457355499268,
"rewards/rejected": -4.389591693878174,
"step": 1393
},
{
"epoch": 0.9104416687076496,
"grad_norm": 11.396992947468927,
"learning_rate": 3.633838974409148e-09,
"logits/chosen": -1.4794695377349854,
"logits/rejected": -1.458986759185791,
"logps/chosen": -792.629150390625,
"logps/rejected": -829.4083251953125,
"loss": 0.5528,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.2658491134643555,
"rewards/margins": 0.6557501554489136,
"rewards/rejected": -3.9215991497039795,
"step": 1394
},
{
"epoch": 0.9110947832476121,
"grad_norm": 51.767743124843925,
"learning_rate": 3.5814085996188516e-09,
"logits/chosen": -1.5730397701263428,
"logits/rejected": -1.5797572135925293,
"logps/chosen": -915.05859375,
"logps/rejected": -1105.7381591796875,
"loss": 0.4246,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.560077667236328,
"rewards/margins": 1.278112769126892,
"rewards/rejected": -4.83819055557251,
"step": 1395
},
{
"epoch": 0.9117478977875745,
"grad_norm": 45.396998730278426,
"learning_rate": 3.529349967963263e-09,
"logits/chosen": -1.4627869129180908,
"logits/rejected": -1.473419427871704,
"logps/chosen": -947.2510986328125,
"logps/rejected": -1030.341064453125,
"loss": 0.4942,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.280148506164551,
"rewards/margins": 0.8479899764060974,
"rewards/rejected": -4.128138065338135,
"step": 1396
},
{
"epoch": 0.9124010123275369,
"grad_norm": 15.614778720558112,
"learning_rate": 3.477663350414378e-09,
"logits/chosen": -1.5414352416992188,
"logits/rejected": -1.560785174369812,
"logps/chosen": -816.3881225585938,
"logps/rejected": -960.979736328125,
"loss": 0.4485,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.140190601348877,
"rewards/margins": 1.348671555519104,
"rewards/rejected": -4.488862037658691,
"step": 1397
},
{
"epoch": 0.9130541268674994,
"grad_norm": 73.82027329116843,
"learning_rate": 3.426349016007815e-09,
"logits/chosen": -1.4851999282836914,
"logits/rejected": -1.4966939687728882,
"logps/chosen": -769.3817749023438,
"logps/rejected": -939.4908447265625,
"loss": 0.4833,
"rewards/accuracies": 0.8125,
"rewards/chosen": -2.9074413776397705,
"rewards/margins": 1.105202555656433,
"rewards/rejected": -4.012643337249756,
"step": 1398
},
{
"epoch": 0.9137072414074618,
"grad_norm": 42.06804571860955,
"learning_rate": 3.3754072318414346e-09,
"logits/chosen": -1.4935294389724731,
"logits/rejected": -1.4656096696853638,
"logps/chosen": -913.9616088867188,
"logps/rejected": -963.2996826171875,
"loss": 0.5134,
"rewards/accuracies": 0.84375,
"rewards/chosen": -4.082716464996338,
"rewards/margins": 1.030174970626831,
"rewards/rejected": -5.112891674041748,
"step": 1399
},
{
"epoch": 0.9143603559474243,
"grad_norm": 52.8878293286426,
"learning_rate": 3.3248382630738813e-09,
"logits/chosen": -1.5851317644119263,
"logits/rejected": -1.5376235246658325,
"logps/chosen": -955.0020141601562,
"logps/rejected": -968.5732421875,
"loss": 0.5165,
"rewards/accuracies": 0.65625,
"rewards/chosen": -4.052890777587891,
"rewards/margins": 0.612097442150116,
"rewards/rejected": -4.664988040924072,
"step": 1400
},
{
"epoch": 0.9143603559474243,
"eval_logits/chosen": -1.479893684387207,
"eval_logits/rejected": -1.4609832763671875,
"eval_logps/chosen": -843.42822265625,
"eval_logps/rejected": -927.4912719726562,
"eval_loss": 0.4942656457424164,
"eval_rewards/accuracies": 0.7599999904632568,
"eval_rewards/chosen": -3.330828905105591,
"eval_rewards/margins": 0.9216902256011963,
"eval_rewards/rejected": -4.252519130706787,
"eval_runtime": 300.2025,
"eval_samples_per_second": 13.324,
"eval_steps_per_second": 0.833,
"step": 1400
},
{
"epoch": 0.9150134704873867,
"grad_norm": 43.39561827227798,
"learning_rate": 3.2746423729232945e-09,
"logits/chosen": -1.4578529596328735,
"logits/rejected": -1.4696025848388672,
"logps/chosen": -826.8597412109375,
"logps/rejected": -900.2799682617188,
"loss": 0.4753,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.285310983657837,
"rewards/margins": 0.8875605463981628,
"rewards/rejected": -4.172871112823486,
"step": 1401
},
{
"epoch": 0.9156665850273492,
"grad_norm": 12.907015811194501,
"learning_rate": 3.224819822665842e-09,
"logits/chosen": -1.4285743236541748,
"logits/rejected": -1.4410088062286377,
"logps/chosen": -723.3681640625,
"logps/rejected": -917.68701171875,
"loss": 0.4513,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.843869209289551,
"rewards/margins": 1.6680275201797485,
"rewards/rejected": -4.511897087097168,
"step": 1402
},
{
"epoch": 0.9163196995673116,
"grad_norm": 14.776408723830784,
"learning_rate": 3.1753708716344364e-09,
"logits/chosen": -1.5420118570327759,
"logits/rejected": -1.4822347164154053,
"logps/chosen": -804.4373779296875,
"logps/rejected": -927.0358276367188,
"loss": 0.4767,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.7062926292419434,
"rewards/margins": 0.9591672420501709,
"rewards/rejected": -4.665459632873535,
"step": 1403
},
{
"epoch": 0.916972814107274,
"grad_norm": 110.87083222991181,
"learning_rate": 3.1262957772173637e-09,
"logits/chosen": -1.4802658557891846,
"logits/rejected": -1.4908185005187988,
"logps/chosen": -890.1760864257812,
"logps/rejected": -956.2567138671875,
"loss": 0.4505,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.3808512687683105,
"rewards/margins": 0.817065417766571,
"rewards/rejected": -4.1979169845581055,
"step": 1404
},
{
"epoch": 0.9176259286472365,
"grad_norm": 61.96498893829653,
"learning_rate": 3.0775947948569162e-09,
"logits/chosen": -1.4553033113479614,
"logits/rejected": -1.469690203666687,
"logps/chosen": -904.5153198242188,
"logps/rejected": -954.6373901367188,
"loss": 0.4687,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.5189156532287598,
"rewards/margins": 0.6943820118904114,
"rewards/rejected": -4.2132978439331055,
"step": 1405
},
{
"epoch": 0.918279043187199,
"grad_norm": 67.39643533256593,
"learning_rate": 3.0292681780481027e-09,
"logits/chosen": -1.587602972984314,
"logits/rejected": -1.5586885213851929,
"logps/chosen": -860.71923828125,
"logps/rejected": -963.667724609375,
"loss": 0.4579,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.3475418090820312,
"rewards/margins": 0.9892367124557495,
"rewards/rejected": -4.336778163909912,
"step": 1406
},
{
"epoch": 0.9189321577271614,
"grad_norm": 21.82848056849987,
"learning_rate": 2.981316178337298e-09,
"logits/chosen": -1.4134702682495117,
"logits/rejected": -1.4143006801605225,
"logps/chosen": -781.74853515625,
"logps/rejected": -905.278564453125,
"loss": 0.4413,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.0943939685821533,
"rewards/margins": 1.127016305923462,
"rewards/rejected": -4.221409797668457,
"step": 1407
},
{
"epoch": 0.9195852722671238,
"grad_norm": 50.76090513680746,
"learning_rate": 2.933739045320946e-09,
"logits/chosen": -1.5723644495010376,
"logits/rejected": -1.4760441780090332,
"logps/chosen": -810.4500122070312,
"logps/rejected": -890.5816040039062,
"loss": 0.4265,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.274219512939453,
"rewards/margins": 1.089596152305603,
"rewards/rejected": -4.363815784454346,
"step": 1408
},
{
"epoch": 0.9202383868070863,
"grad_norm": 35.65848530691593,
"learning_rate": 2.886537026644259e-09,
"logits/chosen": -1.4684292078018188,
"logits/rejected": -1.492079257965088,
"logps/chosen": -921.974853515625,
"logps/rejected": -1068.1021728515625,
"loss": 0.4578,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.9635848999023438,
"rewards/margins": 1.2408589124679565,
"rewards/rejected": -5.20444393157959,
"step": 1409
},
{
"epoch": 0.9208915013470488,
"grad_norm": 13.02579519124362,
"learning_rate": 2.8397103679999535e-09,
"logits/chosen": -1.4359222650527954,
"logits/rejected": -1.4592533111572266,
"logps/chosen": -927.1942749023438,
"logps/rejected": -1130.98779296875,
"loss": 0.5307,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3110556602478027,
"rewards/margins": 1.212038278579712,
"rewards/rejected": -4.5230937004089355,
"step": 1410
},
{
"epoch": 0.9215446158870112,
"grad_norm": 71.1836961214141,
"learning_rate": 2.7932593131269085e-09,
"logits/chosen": -1.506251573562622,
"logits/rejected": -1.4403364658355713,
"logps/chosen": -857.0916748046875,
"logps/rejected": -886.8201293945312,
"loss": 0.5038,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.356765031814575,
"rewards/margins": 0.6197909712791443,
"rewards/rejected": -3.9765563011169434,
"step": 1411
},
{
"epoch": 0.9221977304269736,
"grad_norm": 13.552404267546486,
"learning_rate": 2.747184103808975e-09,
"logits/chosen": -1.531203269958496,
"logits/rejected": -1.497283935546875,
"logps/chosen": -899.0496215820312,
"logps/rejected": -952.4412841796875,
"loss": 0.4867,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.493135452270508,
"rewards/margins": 0.9362788796424866,
"rewards/rejected": -4.429414749145508,
"step": 1412
},
{
"epoch": 0.922850844966936,
"grad_norm": 83.47511406736928,
"learning_rate": 2.7014849798736526e-09,
"logits/chosen": -1.4521548748016357,
"logits/rejected": -1.4682101011276245,
"logps/chosen": -852.2609252929688,
"logps/rejected": -949.7713012695312,
"loss": 0.4316,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.3543503284454346,
"rewards/margins": 1.411728858947754,
"rewards/rejected": -4.766078948974609,
"step": 1413
},
{
"epoch": 0.9235039595068986,
"grad_norm": 20.341630193627967,
"learning_rate": 2.6561621791908654e-09,
"logits/chosen": -1.53240168094635,
"logits/rejected": -1.5243420600891113,
"logps/chosen": -909.6138916015625,
"logps/rejected": -951.72900390625,
"loss": 0.4342,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.3367176055908203,
"rewards/margins": 0.8963368535041809,
"rewards/rejected": -4.233054161071777,
"step": 1414
},
{
"epoch": 0.924157074046861,
"grad_norm": 41.165407344672616,
"learning_rate": 2.6112159376717456e-09,
"logits/chosen": -1.4888275861740112,
"logits/rejected": -1.4489132165908813,
"logps/chosen": -770.2021484375,
"logps/rejected": -767.581787109375,
"loss": 0.5119,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1834659576416016,
"rewards/margins": 0.5162588357925415,
"rewards/rejected": -3.6997246742248535,
"step": 1415
},
{
"epoch": 0.9248101885868234,
"grad_norm": 49.5990277647984,
"learning_rate": 2.5666464892673768e-09,
"logits/chosen": -1.4363664388656616,
"logits/rejected": -1.4629590511322021,
"logps/chosen": -779.8847045898438,
"logps/rejected": -841.2040405273438,
"loss": 0.4964,
"rewards/accuracies": 0.75,
"rewards/chosen": -2.621016502380371,
"rewards/margins": 0.6873450875282288,
"rewards/rejected": -3.308361530303955,
"step": 1416
},
{
"epoch": 0.9254633031267858,
"grad_norm": 80.94852955816721,
"learning_rate": 2.5224540659675692e-09,
"logits/chosen": -1.5635671615600586,
"logits/rejected": -1.5525977611541748,
"logps/chosen": -759.9940185546875,
"logps/rejected": -936.1673583984375,
"loss": 0.4763,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.061068058013916,
"rewards/margins": 1.1235350370407104,
"rewards/rejected": -4.184603214263916,
"step": 1417
},
{
"epoch": 0.9261164176667483,
"grad_norm": 42.64229099155968,
"learning_rate": 2.4786388977997034e-09,
"logits/chosen": -1.5770373344421387,
"logits/rejected": -1.4894176721572876,
"logps/chosen": -840.8875122070312,
"logps/rejected": -915.895263671875,
"loss": 0.5005,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.297794818878174,
"rewards/margins": 0.8870930671691895,
"rewards/rejected": -4.184887886047363,
"step": 1418
},
{
"epoch": 0.9267695322067108,
"grad_norm": 95.47316675894163,
"learning_rate": 2.435201212827456e-09,
"logits/chosen": -1.5091971158981323,
"logits/rejected": -1.4532575607299805,
"logps/chosen": -874.2222900390625,
"logps/rejected": -985.8672485351562,
"loss": 0.4799,
"rewards/accuracies": 0.5625,
"rewards/chosen": -3.3658037185668945,
"rewards/margins": 0.7014881372451782,
"rewards/rejected": -4.067291736602783,
"step": 1419
},
{
"epoch": 0.9274226467466732,
"grad_norm": 22.60308983441237,
"learning_rate": 2.3921412371496834e-09,
"logits/chosen": -1.481071949005127,
"logits/rejected": -1.403544545173645,
"logps/chosen": -835.0723266601562,
"logps/rejected": -833.1524047851562,
"loss": 0.5539,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.328864812850952,
"rewards/margins": 0.49991360306739807,
"rewards/rejected": -3.8287787437438965,
"step": 1420
},
{
"epoch": 0.9280757612866356,
"grad_norm": 22.71188518146946,
"learning_rate": 2.349459194899198e-09,
"logits/chosen": -1.4905717372894287,
"logits/rejected": -1.4012264013290405,
"logps/chosen": -864.1593017578125,
"logps/rejected": -974.7191162109375,
"loss": 0.4975,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.465050220489502,
"rewards/margins": 0.9912995100021362,
"rewards/rejected": -4.4563493728637695,
"step": 1421
},
{
"epoch": 0.9287288758265981,
"grad_norm": 38.38087407026408,
"learning_rate": 2.307155308241643e-09,
"logits/chosen": -1.51276695728302,
"logits/rejected": -1.5155949592590332,
"logps/chosen": -840.83056640625,
"logps/rejected": -936.4772338867188,
"loss": 0.4841,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4179623126983643,
"rewards/margins": 1.073073387145996,
"rewards/rejected": -4.4910359382629395,
"step": 1422
},
{
"epoch": 0.9293819903665606,
"grad_norm": 31.60230447821753,
"learning_rate": 2.2652297973742963e-09,
"logits/chosen": -1.5305612087249756,
"logits/rejected": -1.509539008140564,
"logps/chosen": -770.224609375,
"logps/rejected": -883.810302734375,
"loss": 0.5138,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.264803409576416,
"rewards/margins": 1.1701805591583252,
"rewards/rejected": -4.434983730316162,
"step": 1423
},
{
"epoch": 0.930035104906523,
"grad_norm": 77.90811255353113,
"learning_rate": 2.2236828805249184e-09,
"logits/chosen": -1.5154668092727661,
"logits/rejected": -1.516597867012024,
"logps/chosen": -771.7640380859375,
"logps/rejected": -854.826904296875,
"loss": 0.5166,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.9361448287963867,
"rewards/margins": 0.7485728859901428,
"rewards/rejected": -3.6847176551818848,
"step": 1424
},
{
"epoch": 0.9306882194464854,
"grad_norm": 37.785751655050305,
"learning_rate": 2.1825147739506805e-09,
"logits/chosen": -1.4908900260925293,
"logits/rejected": -1.5062459707260132,
"logps/chosen": -884.8120727539062,
"logps/rejected": -929.1033935546875,
"loss": 0.562,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.515050172805786,
"rewards/margins": 0.5184265375137329,
"rewards/rejected": -4.03347635269165,
"step": 1425
},
{
"epoch": 0.9313413339864479,
"grad_norm": 15.861485447329448,
"learning_rate": 2.141725691936963e-09,
"logits/chosen": -1.4910266399383545,
"logits/rejected": -1.4862860441207886,
"logps/chosen": -826.7039794921875,
"logps/rejected": -980.1998291015625,
"loss": 0.486,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.137228012084961,
"rewards/margins": 1.146528720855713,
"rewards/rejected": -4.283757209777832,
"step": 1426
},
{
"epoch": 0.9319944485264103,
"grad_norm": 28.488624408307338,
"learning_rate": 2.1013158467963004e-09,
"logits/chosen": -1.6420342922210693,
"logits/rejected": -1.6163609027862549,
"logps/chosen": -924.2805786132812,
"logps/rejected": -944.080810546875,
"loss": 0.5047,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.657701015472412,
"rewards/margins": 0.6087216138839722,
"rewards/rejected": -4.266422748565674,
"step": 1427
},
{
"epoch": 0.9326475630663728,
"grad_norm": 10.967689522386022,
"learning_rate": 2.0612854488672227e-09,
"logits/chosen": -1.4555834531784058,
"logits/rejected": -1.4464821815490723,
"logps/chosen": -761.0975952148438,
"logps/rejected": -887.48046875,
"loss": 0.448,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.0129241943359375,
"rewards/margins": 1.1502059698104858,
"rewards/rejected": -4.163130760192871,
"step": 1428
},
{
"epoch": 0.9333006776063352,
"grad_norm": 14.555394411736843,
"learning_rate": 2.0216347065132144e-09,
"logits/chosen": -1.4503599405288696,
"logits/rejected": -1.418792724609375,
"logps/chosen": -837.8484497070312,
"logps/rejected": -937.6596069335938,
"loss": 0.5286,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.435154914855957,
"rewards/margins": 0.9483442902565002,
"rewards/rejected": -4.3834991455078125,
"step": 1429
},
{
"epoch": 0.9339537921462977,
"grad_norm": 65.87984391899451,
"learning_rate": 1.982363826121583e-09,
"logits/chosen": -1.5405863523483276,
"logits/rejected": -1.545666217803955,
"logps/chosen": -946.0952758789062,
"logps/rejected": -1030.0819091796875,
"loss": 0.477,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5466394424438477,
"rewards/margins": 0.8139285445213318,
"rewards/rejected": -4.360568046569824,
"step": 1430
},
{
"epoch": 0.9346069066862601,
"grad_norm": 18.205479221571412,
"learning_rate": 1.943473012102409e-09,
"logits/chosen": -1.531268835067749,
"logits/rejected": -1.5460495948791504,
"logps/chosen": -898.9242553710938,
"logps/rejected": -946.4170532226562,
"loss": 0.54,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.561772346496582,
"rewards/margins": 0.8415082693099976,
"rewards/rejected": -4.403279781341553,
"step": 1431
},
{
"epoch": 0.9352600212262225,
"grad_norm": 22.739713888659217,
"learning_rate": 1.9049624668874886e-09,
"logits/chosen": -1.4721064567565918,
"logits/rejected": -1.457444667816162,
"logps/chosen": -727.5782470703125,
"logps/rejected": -881.88720703125,
"loss": 0.432,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.979851007461548,
"rewards/margins": 1.0425056219100952,
"rewards/rejected": -4.0223565101623535,
"step": 1432
},
{
"epoch": 0.935913135766185,
"grad_norm": 23.34819005960656,
"learning_rate": 1.866832390929243e-09,
"logits/chosen": -1.4648447036743164,
"logits/rejected": -1.4732989072799683,
"logps/chosen": -815.2117309570312,
"logps/rejected": -898.7406005859375,
"loss": 0.541,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4008235931396484,
"rewards/margins": 0.6835126280784607,
"rewards/rejected": -4.084336280822754,
"step": 1433
},
{
"epoch": 0.9365662503061475,
"grad_norm": 63.46180789960193,
"learning_rate": 1.8290829826997367e-09,
"logits/chosen": -1.559678554534912,
"logits/rejected": -1.5437560081481934,
"logps/chosen": -902.6143188476562,
"logps/rejected": -933.3184204101562,
"loss": 0.5227,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.4241020679473877,
"rewards/margins": 0.513259768486023,
"rewards/rejected": -3.9373619556427,
"step": 1434
},
{
"epoch": 0.9372193648461099,
"grad_norm": 34.479724315346566,
"learning_rate": 1.7917144386895926e-09,
"logits/chosen": -1.5433356761932373,
"logits/rejected": -1.4930343627929688,
"logps/chosen": -839.4082641601562,
"logps/rejected": -884.831298828125,
"loss": 0.5361,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.436253547668457,
"rewards/margins": 0.7238208651542664,
"rewards/rejected": -4.160074234008789,
"step": 1435
},
{
"epoch": 0.9378724793860723,
"grad_norm": 15.898555518439984,
"learning_rate": 1.7547269534069626e-09,
"logits/chosen": -1.4363042116165161,
"logits/rejected": -1.426371455192566,
"logps/chosen": -834.6913452148438,
"logps/rejected": -999.5419921875,
"loss": 0.4245,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.229515552520752,
"rewards/margins": 1.1570442914962769,
"rewards/rejected": -4.386559963226318,
"step": 1436
},
{
"epoch": 0.9385255939260347,
"grad_norm": 24.923739035573195,
"learning_rate": 1.7181207193765756e-09,
"logits/chosen": -1.5305100679397583,
"logits/rejected": -1.5034910440444946,
"logps/chosen": -884.4599609375,
"logps/rejected": -919.95947265625,
"loss": 0.4775,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.1505179405212402,
"rewards/margins": 0.9173092842102051,
"rewards/rejected": -4.067827224731445,
"step": 1437
},
{
"epoch": 0.9391787084659973,
"grad_norm": 22.943505162467492,
"learning_rate": 1.681895927138674e-09,
"logits/chosen": -1.532247543334961,
"logits/rejected": -1.5096689462661743,
"logps/chosen": -860.7628173828125,
"logps/rejected": -1032.169189453125,
"loss": 0.4488,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.4541678428649902,
"rewards/margins": 1.4435040950775146,
"rewards/rejected": -4.897672176361084,
"step": 1438
},
{
"epoch": 0.9398318230059597,
"grad_norm": 43.92276660865622,
"learning_rate": 1.646052765248046e-09,
"logits/chosen": -1.5007368326187134,
"logits/rejected": -1.539098858833313,
"logps/chosen": -893.9174194335938,
"logps/rejected": -938.9563598632812,
"loss": 0.5518,
"rewards/accuracies": 0.6875,
"rewards/chosen": -4.333798885345459,
"rewards/margins": 0.5584127902984619,
"rewards/rejected": -4.8922119140625,
"step": 1439
},
{
"epoch": 0.9404849375459221,
"grad_norm": 34.468585588337255,
"learning_rate": 1.6105914202730608e-09,
"logits/chosen": -1.4849194288253784,
"logits/rejected": -1.4576518535614014,
"logps/chosen": -796.8408813476562,
"logps/rejected": -966.2708129882812,
"loss": 0.4925,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.009169578552246,
"rewards/margins": 1.379059076309204,
"rewards/rejected": -4.388228893280029,
"step": 1440
},
{
"epoch": 0.9411380520858845,
"grad_norm": 40.61462323177837,
"learning_rate": 1.5755120767946607e-09,
"logits/chosen": -1.4398796558380127,
"logits/rejected": -1.3817028999328613,
"logps/chosen": -766.0177612304688,
"logps/rejected": -804.5883178710938,
"loss": 0.4724,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.0304973125457764,
"rewards/margins": 0.6275922656059265,
"rewards/rejected": -3.6580896377563477,
"step": 1441
},
{
"epoch": 0.941791166625847,
"grad_norm": 10.887896926258218,
"learning_rate": 1.5408149174054446e-09,
"logits/chosen": -1.4392205476760864,
"logits/rejected": -1.4316151142120361,
"logps/chosen": -780.572509765625,
"logps/rejected": -929.8662109375,
"loss": 0.4883,
"rewards/accuracies": 0.875,
"rewards/chosen": -2.8742194175720215,
"rewards/margins": 1.0152424573898315,
"rewards/rejected": -3.8894619941711426,
"step": 1442
},
{
"epoch": 0.9424442811658095,
"grad_norm": 11.226257970745792,
"learning_rate": 1.506500122708662e-09,
"logits/chosen": -1.4671475887298584,
"logits/rejected": -1.4885404109954834,
"logps/chosen": -872.2683715820312,
"logps/rejected": -914.68115234375,
"loss": 0.5137,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.3775594234466553,
"rewards/margins": 0.5714359283447266,
"rewards/rejected": -3.9489948749542236,
"step": 1443
},
{
"epoch": 0.9430973957057719,
"grad_norm": 75.39569486082955,
"learning_rate": 1.4725678713173207e-09,
"logits/chosen": -1.5088709592819214,
"logits/rejected": -1.502449870109558,
"logps/chosen": -910.4465942382812,
"logps/rejected": -961.0289306640625,
"loss": 0.527,
"rewards/accuracies": 0.59375,
"rewards/chosen": -3.666200637817383,
"rewards/margins": 0.5112284421920776,
"rewards/rejected": -4.177428722381592,
"step": 1444
},
{
"epoch": 0.9437505102457343,
"grad_norm": 12.36325740105416,
"learning_rate": 1.4390183398532457e-09,
"logits/chosen": -1.5747441053390503,
"logits/rejected": -1.5668046474456787,
"logps/chosen": -940.868408203125,
"logps/rejected": -1135.773681640625,
"loss": 0.4886,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.498914957046509,
"rewards/margins": 1.4520134925842285,
"rewards/rejected": -4.950928688049316,
"step": 1445
},
{
"epoch": 0.9444036247856968,
"grad_norm": 53.08556208724623,
"learning_rate": 1.405851702946148e-09,
"logits/chosen": -1.4660958051681519,
"logits/rejected": -1.4496102333068848,
"logps/chosen": -885.9743041992188,
"logps/rejected": -923.9781494140625,
"loss": 0.5484,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4384634494781494,
"rewards/margins": 0.7487789392471313,
"rewards/rejected": -4.18724250793457,
"step": 1446
},
{
"epoch": 0.9450567393256593,
"grad_norm": 10.230443690320527,
"learning_rate": 1.3730681332327242e-09,
"logits/chosen": -1.409921646118164,
"logits/rejected": -1.4098153114318848,
"logps/chosen": -854.5587158203125,
"logps/rejected": -979.5097045898438,
"loss": 0.4908,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.432528018951416,
"rewards/margins": 1.0839788913726807,
"rewards/rejected": -4.516506671905518,
"step": 1447
},
{
"epoch": 0.9457098538656217,
"grad_norm": 51.690528713979994,
"learning_rate": 1.3406678013557492e-09,
"logits/chosen": -1.4445513486862183,
"logits/rejected": -1.489593267440796,
"logps/chosen": -868.4315185546875,
"logps/rejected": -968.2102661132812,
"loss": 0.431,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.0471696853637695,
"rewards/margins": 0.9745462536811829,
"rewards/rejected": -4.021716117858887,
"step": 1448
},
{
"epoch": 0.9463629684055841,
"grad_norm": 66.01296376492611,
"learning_rate": 1.3086508759631936e-09,
"logits/chosen": -1.609882116317749,
"logits/rejected": -1.605045199394226,
"logps/chosen": -855.3104248046875,
"logps/rejected": -889.3307495117188,
"loss": 0.543,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.2741992473602295,
"rewards/margins": 0.46837118268013,
"rewards/rejected": -3.742570161819458,
"step": 1449
},
{
"epoch": 0.9470160829455466,
"grad_norm": 69.83844230333594,
"learning_rate": 1.2770175237073661e-09,
"logits/chosen": -1.5078953504562378,
"logits/rejected": -1.4897831678390503,
"logps/chosen": -891.066650390625,
"logps/rejected": -890.9603271484375,
"loss": 0.5112,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.392768383026123,
"rewards/margins": 0.624529242515564,
"rewards/rejected": -4.017297267913818,
"step": 1450
},
{
"epoch": 0.947669197485509,
"grad_norm": 96.35598691320328,
"learning_rate": 1.2457679092440054e-09,
"logits/chosen": -1.5438551902770996,
"logits/rejected": -1.5170389413833618,
"logps/chosen": -747.1302490234375,
"logps/rejected": -902.2919311523438,
"loss": 0.4724,
"rewards/accuracies": 0.90625,
"rewards/chosen": -2.6579556465148926,
"rewards/margins": 1.488484501838684,
"rewards/rejected": -4.146440505981445,
"step": 1451
},
{
"epoch": 0.9483223120254715,
"grad_norm": 10.598167900496241,
"learning_rate": 1.2149021952314654e-09,
"logits/chosen": -1.587843418121338,
"logits/rejected": -1.5127121210098267,
"logps/chosen": -762.05615234375,
"logps/rejected": -796.598388671875,
"loss": 0.5009,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.141162157058716,
"rewards/margins": 0.5582625865936279,
"rewards/rejected": -3.699424982070923,
"step": 1452
},
{
"epoch": 0.9489754265654339,
"grad_norm": 37.95658977591617,
"learning_rate": 1.1844205423298142e-09,
"logits/chosen": -1.5430374145507812,
"logits/rejected": -1.494556188583374,
"logps/chosen": -900.53173828125,
"logps/rejected": -893.8412475585938,
"loss": 0.4861,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4113199710845947,
"rewards/margins": 0.7247394323348999,
"rewards/rejected": -4.136059284210205,
"step": 1453
},
{
"epoch": 0.9496285411053964,
"grad_norm": 37.85534269742698,
"learning_rate": 1.15432310920007e-09,
"logits/chosen": -1.5332300662994385,
"logits/rejected": -1.5049644708633423,
"logps/chosen": -879.8773803710938,
"logps/rejected": -943.8563842773438,
"loss": 0.4722,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.507415771484375,
"rewards/margins": 0.8981890678405762,
"rewards/rejected": -4.405604839324951,
"step": 1454
},
{
"epoch": 0.9502816556453588,
"grad_norm": 13.249961940362768,
"learning_rate": 1.1246100525033165e-09,
"logits/chosen": -1.51246976852417,
"logits/rejected": -1.4882327318191528,
"logps/chosen": -900.8251342773438,
"logps/rejected": -959.567626953125,
"loss": 0.4925,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6211113929748535,
"rewards/margins": 0.7765340805053711,
"rewards/rejected": -4.397645473480225,
"step": 1455
},
{
"epoch": 0.9509347701853212,
"grad_norm": 78.0554509757586,
"learning_rate": 1.0952815268999049e-09,
"logits/chosen": -1.4893440008163452,
"logits/rejected": -1.458596110343933,
"logps/chosen": -775.9324951171875,
"logps/rejected": -817.9571533203125,
"loss": 0.4805,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.140364646911621,
"rewards/margins": 0.7064178586006165,
"rewards/rejected": -3.846782684326172,
"step": 1456
},
{
"epoch": 0.9515878847252837,
"grad_norm": 112.04084469499804,
"learning_rate": 1.0663376850486628e-09,
"logits/chosen": -1.5067682266235352,
"logits/rejected": -1.438143253326416,
"logps/chosen": -828.5684814453125,
"logps/rejected": -927.2904052734375,
"loss": 0.5531,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.495786666870117,
"rewards/margins": 0.701140820980072,
"rewards/rejected": -4.196927547454834,
"step": 1457
},
{
"epoch": 0.9522409992652462,
"grad_norm": 39.94282013087609,
"learning_rate": 1.0377786776060854e-09,
"logits/chosen": -1.5286598205566406,
"logits/rejected": -1.4984127283096313,
"logps/chosen": -771.9295654296875,
"logps/rejected": -820.18310546875,
"loss": 0.4538,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.990509033203125,
"rewards/margins": 1.0369949340820312,
"rewards/rejected": -4.027503967285156,
"step": 1458
},
{
"epoch": 0.9528941138052086,
"grad_norm": 93.88679149047655,
"learning_rate": 1.0096046532255376e-09,
"logits/chosen": -1.6264631748199463,
"logits/rejected": -1.5820178985595703,
"logps/chosen": -873.897705078125,
"logps/rejected": -970.2108154296875,
"loss": 0.5254,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.5638842582702637,
"rewards/margins": 0.973100483417511,
"rewards/rejected": -4.536985397338867,
"step": 1459
},
{
"epoch": 0.953547228345171,
"grad_norm": 19.02700301018552,
"learning_rate": 9.818157585565284e-10,
"logits/chosen": -1.4352535009384155,
"logits/rejected": -1.3992193937301636,
"logps/chosen": -827.4281616210938,
"logps/rejected": -891.892822265625,
"loss": 0.49,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.2847142219543457,
"rewards/margins": 0.696830153465271,
"rewards/rejected": -3.9815444946289062,
"step": 1460
},
{
"epoch": 0.9542003428851334,
"grad_norm": 12.308757033298024,
"learning_rate": 9.544121382438875e-10,
"logits/chosen": -1.5002676248550415,
"logits/rejected": -1.46719491481781,
"logps/chosen": -895.8308715820312,
"logps/rejected": -926.5982666015625,
"loss": 0.4543,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5023105144500732,
"rewards/margins": 0.7259926795959473,
"rewards/rejected": -4.2283034324646,
"step": 1461
},
{
"epoch": 0.954853457425096,
"grad_norm": 18.453922423435763,
"learning_rate": 9.273939349270565e-10,
"logits/chosen": -1.5010855197906494,
"logits/rejected": -1.3877067565917969,
"logps/chosen": -768.5657348632812,
"logps/rejected": -797.06396484375,
"loss": 0.5329,
"rewards/accuracies": 0.6875,
"rewards/chosen": -2.9298336505889893,
"rewards/margins": 0.8176116943359375,
"rewards/rejected": -3.7474451065063477,
"step": 1462
},
{
"epoch": 0.9555065719650584,
"grad_norm": 51.301955382401765,
"learning_rate": 9.00761289239324e-10,
"logits/chosen": -1.5010892152786255,
"logits/rejected": -1.4880714416503906,
"logps/chosen": -826.286865234375,
"logps/rejected": -911.9449462890625,
"loss": 0.4457,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.053290843963623,
"rewards/margins": 0.8817964792251587,
"rewards/rejected": -3.9350876808166504,
"step": 1463
},
{
"epoch": 0.9561596865050208,
"grad_norm": 25.443543945201426,
"learning_rate": 8.74514339807117e-10,
"logits/chosen": -1.4686068296432495,
"logits/rejected": -1.4487457275390625,
"logps/chosen": -860.71484375,
"logps/rejected": -960.4424438476562,
"loss": 0.5242,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.8519248962402344,
"rewards/margins": 1.1036741733551025,
"rewards/rejected": -4.955598831176758,
"step": 1464
},
{
"epoch": 0.9568128010449832,
"grad_norm": 47.38366581435745,
"learning_rate": 8.48653223249235e-10,
"logits/chosen": -1.4724491834640503,
"logits/rejected": -1.4838087558746338,
"logps/chosen": -881.0520629882812,
"logps/rejected": -929.6651611328125,
"loss": 0.4467,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.963942289352417,
"rewards/margins": 0.7111070156097412,
"rewards/rejected": -4.675049304962158,
"step": 1465
},
{
"epoch": 0.9574659155849458,
"grad_norm": 22.495822791256995,
"learning_rate": 8.23178074176184e-10,
"logits/chosen": -1.5290329456329346,
"logits/rejected": -1.5657262802124023,
"logps/chosen": -871.360107421875,
"logps/rejected": -918.6497802734375,
"loss": 0.4997,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.4585037231445312,
"rewards/margins": 0.9077770113945007,
"rewards/rejected": -4.3662800788879395,
"step": 1466
},
{
"epoch": 0.9581190301249082,
"grad_norm": 56.19391082571053,
"learning_rate": 7.980890251894606e-10,
"logits/chosen": -1.5316890478134155,
"logits/rejected": -1.5544103384017944,
"logps/chosen": -880.1229858398438,
"logps/rejected": -1003.7587890625,
"loss": 0.4847,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.2340664863586426,
"rewards/margins": 1.333704948425293,
"rewards/rejected": -4.567770957946777,
"step": 1467
},
{
"epoch": 0.9587721446648706,
"grad_norm": 19.370534224099423,
"learning_rate": 7.733862068808522e-10,
"logits/chosen": -1.5931593179702759,
"logits/rejected": -1.595098614692688,
"logps/chosen": -917.0612182617188,
"logps/rejected": -948.163818359375,
"loss": 0.5525,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.9437098503112793,
"rewards/margins": 0.40092504024505615,
"rewards/rejected": -4.344634532928467,
"step": 1468
},
{
"epoch": 0.959425259204833,
"grad_norm": 51.95592093037611,
"learning_rate": 7.490697478317709e-10,
"logits/chosen": -1.4602024555206299,
"logits/rejected": -1.472777247428894,
"logps/chosen": -886.6076049804688,
"logps/rejected": -977.1819458007812,
"loss": 0.4909,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.663419246673584,
"rewards/margins": 0.9881623983383179,
"rewards/rejected": -4.651581764221191,
"step": 1469
},
{
"epoch": 0.9600783737447955,
"grad_norm": 33.610490066674124,
"learning_rate": 7.251397746125709e-10,
"logits/chosen": -1.5568439960479736,
"logits/rejected": -1.548944354057312,
"logps/chosen": -813.5231323242188,
"logps/rejected": -983.3826904296875,
"loss": 0.5045,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4104156494140625,
"rewards/margins": 1.39304518699646,
"rewards/rejected": -4.803461074829102,
"step": 1470
},
{
"epoch": 0.960731488284758,
"grad_norm": 32.807228872430244,
"learning_rate": 7.01596411781899e-10,
"logits/chosen": -1.5096122026443481,
"logits/rejected": -1.489628791809082,
"logps/chosen": -898.908447265625,
"logps/rejected": -961.82958984375,
"loss": 0.4871,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.5367379188537598,
"rewards/margins": 1.0929946899414062,
"rewards/rejected": -4.629732608795166,
"step": 1471
},
{
"epoch": 0.9613846028247204,
"grad_norm": 12.404429113630796,
"learning_rate": 6.784397818860532e-10,
"logits/chosen": -1.4223408699035645,
"logits/rejected": -1.4489169120788574,
"logps/chosen": -749.7882690429688,
"logps/rejected": -879.7198486328125,
"loss": 0.478,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.0195603370666504,
"rewards/margins": 1.003127098083496,
"rewards/rejected": -4.022687911987305,
"step": 1472
},
{
"epoch": 0.9620377173646828,
"grad_norm": 25.259975174041397,
"learning_rate": 6.556700054583253e-10,
"logits/chosen": -1.4970064163208008,
"logits/rejected": -1.4991717338562012,
"logps/chosen": -767.5467529296875,
"logps/rejected": -903.8729248046875,
"loss": 0.5629,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0890581607818604,
"rewards/margins": 0.8758949041366577,
"rewards/rejected": -3.9649529457092285,
"step": 1473
},
{
"epoch": 0.9626908319046453,
"grad_norm": 56.02069174460643,
"learning_rate": 6.332872010183843e-10,
"logits/chosen": -1.5272411108016968,
"logits/rejected": -1.53700590133667,
"logps/chosen": -874.013427734375,
"logps/rejected": -928.57373046875,
"loss": 0.4722,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1264383792877197,
"rewards/margins": 0.8865586519241333,
"rewards/rejected": -4.012997150421143,
"step": 1474
},
{
"epoch": 0.9633439464446077,
"grad_norm": 12.235784698068322,
"learning_rate": 6.112914850716771e-10,
"logits/chosen": -1.4167723655700684,
"logits/rejected": -1.4353446960449219,
"logps/chosen": -847.5712890625,
"logps/rejected": -927.8067626953125,
"loss": 0.4272,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.4825477600097656,
"rewards/margins": 0.6743806004524231,
"rewards/rejected": -4.156928539276123,
"step": 1475
},
{
"epoch": 0.9639970609845702,
"grad_norm": 53.93753110085255,
"learning_rate": 5.896829721087709e-10,
"logits/chosen": -1.4274340867996216,
"logits/rejected": -1.419229507446289,
"logps/chosen": -805.4702758789062,
"logps/rejected": -839.0209350585938,
"loss": 0.4951,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.2356109619140625,
"rewards/margins": 0.7077401280403137,
"rewards/rejected": -3.9433510303497314,
"step": 1476
},
{
"epoch": 0.9646501755245326,
"grad_norm": 14.406009332648742,
"learning_rate": 5.684617746048198e-10,
"logits/chosen": -1.5064961910247803,
"logits/rejected": -1.4822742938995361,
"logps/chosen": -826.1743774414062,
"logps/rejected": -864.1240234375,
"loss": 0.4387,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.075737476348877,
"rewards/margins": 0.7606046199798584,
"rewards/rejected": -3.8363420963287354,
"step": 1477
},
{
"epoch": 0.9653032900644951,
"grad_norm": 38.670095331691506,
"learning_rate": 5.476280030189406e-10,
"logits/chosen": -1.5506258010864258,
"logits/rejected": -1.5424368381500244,
"logps/chosen": -908.591064453125,
"logps/rejected": -1038.738525390625,
"loss": 0.5371,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.491032838821411,
"rewards/margins": 1.2838757038116455,
"rewards/rejected": -4.774909019470215,
"step": 1478
},
{
"epoch": 0.9659564046044575,
"grad_norm": 25.905815918474016,
"learning_rate": 5.271817657936467e-10,
"logits/chosen": -1.4018712043762207,
"logits/rejected": -1.3280322551727295,
"logps/chosen": -733.0325317382812,
"logps/rejected": -867.226806640625,
"loss": 0.4825,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.2157974243164062,
"rewards/margins": 0.9778196811676025,
"rewards/rejected": -4.19361686706543,
"step": 1479
},
{
"epoch": 0.96660951914442,
"grad_norm": 21.34054547875047,
"learning_rate": 5.071231693542732e-10,
"logits/chosen": -1.5019912719726562,
"logits/rejected": -1.4946105480194092,
"logps/chosen": -867.0049438476562,
"logps/rejected": -912.6876220703125,
"loss": 0.4603,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.469493865966797,
"rewards/margins": 0.9318346977233887,
"rewards/rejected": -4.4013285636901855,
"step": 1480
},
{
"epoch": 0.9672626336843824,
"grad_norm": 53.27235117213442,
"learning_rate": 4.874523181084611e-10,
"logits/chosen": -1.520746111869812,
"logits/rejected": -1.5099103450775146,
"logps/chosen": -950.77880859375,
"logps/rejected": -1004.3994750976562,
"loss": 0.4787,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.89591908454895,
"rewards/margins": 0.5920171141624451,
"rewards/rejected": -4.487936019897461,
"step": 1481
},
{
"epoch": 0.9679157482243449,
"grad_norm": 28.111966813993863,
"learning_rate": 4.681693144455656e-10,
"logits/chosen": -1.5613229274749756,
"logits/rejected": -1.4852428436279297,
"logps/chosen": -881.259765625,
"logps/rejected": -977.8818969726562,
"loss": 0.5095,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.6972122192382812,
"rewards/margins": 0.9360721111297607,
"rewards/rejected": -4.633284568786621,
"step": 1482
},
{
"epoch": 0.9685688627643073,
"grad_norm": 54.78566289216356,
"learning_rate": 4.4927425873614867e-10,
"logits/chosen": -1.5294121503829956,
"logits/rejected": -1.4791772365570068,
"logps/chosen": -838.6341552734375,
"logps/rejected": -868.4017333984375,
"loss": 0.4892,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.21793794631958,
"rewards/margins": 0.8229467272758484,
"rewards/rejected": -4.040884971618652,
"step": 1483
},
{
"epoch": 0.9692219773042697,
"grad_norm": 18.897385067422782,
"learning_rate": 4.30767249331454e-10,
"logits/chosen": -1.5810683965682983,
"logits/rejected": -1.5569019317626953,
"logps/chosen": -889.7955322265625,
"logps/rejected": -911.5818481445312,
"loss": 0.504,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.5015814304351807,
"rewards/margins": 0.5960092544555664,
"rewards/rejected": -4.097590923309326,
"step": 1484
},
{
"epoch": 0.9698750918442322,
"grad_norm": 11.548963014295419,
"learning_rate": 4.1264838256289126e-10,
"logits/chosen": -1.5425231456756592,
"logits/rejected": -1.5341880321502686,
"logps/chosen": -866.2848510742188,
"logps/rejected": -1056.494384765625,
"loss": 0.4792,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4112586975097656,
"rewards/margins": 1.2582924365997314,
"rewards/rejected": -4.669550895690918,
"step": 1485
},
{
"epoch": 0.9705282063841947,
"grad_norm": 45.7932753999572,
"learning_rate": 3.9491775274153595e-10,
"logits/chosen": -1.4852381944656372,
"logits/rejected": -1.4904857873916626,
"logps/chosen": -817.4542846679688,
"logps/rejected": -949.4779663085938,
"loss": 0.466,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.714642286300659,
"rewards/margins": 0.9591977000236511,
"rewards/rejected": -4.673839569091797,
"step": 1486
},
{
"epoch": 0.9711813209241571,
"grad_norm": 21.73552603940438,
"learning_rate": 3.7757545215764686e-10,
"logits/chosen": -1.483793020248413,
"logits/rejected": -1.5184385776519775,
"logps/chosen": -764.6434936523438,
"logps/rejected": -1080.6500244140625,
"loss": 0.487,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.000844717025757,
"rewards/margins": 2.042339563369751,
"rewards/rejected": -5.043184280395508,
"step": 1487
},
{
"epoch": 0.9718344354641195,
"grad_norm": 30.8551493809766,
"learning_rate": 3.606215710801663e-10,
"logits/chosen": -1.5189553499221802,
"logits/rejected": -1.555790662765503,
"logps/chosen": -820.7545166015625,
"logps/rejected": -871.0463256835938,
"loss": 0.5234,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.4103848934173584,
"rewards/margins": 0.8298892974853516,
"rewards/rejected": -4.240273475646973,
"step": 1488
},
{
"epoch": 0.9724875500040819,
"grad_norm": 47.632353902452365,
"learning_rate": 3.440561977562789e-10,
"logits/chosen": -1.4855304956436157,
"logits/rejected": -1.4552634954452515,
"logps/chosen": -765.5675048828125,
"logps/rejected": -875.1714477539062,
"loss": 0.4912,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.0855441093444824,
"rewards/margins": 0.9937740564346313,
"rewards/rejected": -4.079318046569824,
"step": 1489
},
{
"epoch": 0.9731406645440445,
"grad_norm": 59.60831943859033,
"learning_rate": 3.278794184109118e-10,
"logits/chosen": -1.5224008560180664,
"logits/rejected": -1.5374361276626587,
"logps/chosen": -921.4551391601562,
"logps/rejected": -966.5579223632812,
"loss": 0.4683,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.87973690032959,
"rewards/margins": 0.5896239280700684,
"rewards/rejected": -4.469360828399658,
"step": 1490
},
{
"epoch": 0.9737937790840069,
"grad_norm": 40.26002383284324,
"learning_rate": 3.1209131724633517e-10,
"logits/chosen": -1.5184483528137207,
"logits/rejected": -1.5036489963531494,
"logps/chosen": -814.7634887695312,
"logps/rejected": -889.21923828125,
"loss": 0.5193,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4046013355255127,
"rewards/margins": 0.740337610244751,
"rewards/rejected": -4.144938945770264,
"step": 1491
},
{
"epoch": 0.9744468936239693,
"grad_norm": 34.218594481949225,
"learning_rate": 2.9669197644168755e-10,
"logits/chosen": -1.5170872211456299,
"logits/rejected": -1.5203460454940796,
"logps/chosen": -835.5225219726562,
"logps/rejected": -900.4053955078125,
"loss": 0.4703,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.076322078704834,
"rewards/margins": 0.8840709328651428,
"rewards/rejected": -3.9603934288024902,
"step": 1492
},
{
"epoch": 0.9751000081639317,
"grad_norm": 10.70959520154608,
"learning_rate": 2.8168147615254265e-10,
"logits/chosen": -1.4827511310577393,
"logits/rejected": -1.463191270828247,
"logps/chosen": -825.8433837890625,
"logps/rejected": -895.0802612304688,
"loss": 0.433,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.2821357250213623,
"rewards/margins": 0.6120182871818542,
"rewards/rejected": -3.8941545486450195,
"step": 1493
},
{
"epoch": 0.9757531227038942,
"grad_norm": 14.011776786991527,
"learning_rate": 2.6705989451054343e-10,
"logits/chosen": -1.3761827945709229,
"logits/rejected": -1.3667436838150024,
"logps/chosen": -856.0826416015625,
"logps/rejected": -964.3638305664062,
"loss": 0.527,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.3973817825317383,
"rewards/margins": 0.8620023727416992,
"rewards/rejected": -4.2593841552734375,
"step": 1494
},
{
"epoch": 0.9764062372438567,
"grad_norm": 56.388774885579274,
"learning_rate": 2.528273076229187e-10,
"logits/chosen": -1.4972938299179077,
"logits/rejected": -1.4492802619934082,
"logps/chosen": -858.5281982421875,
"logps/rejected": -913.3388671875,
"loss": 0.5209,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.4028518199920654,
"rewards/margins": 0.6677621006965637,
"rewards/rejected": -4.070613861083984,
"step": 1495
},
{
"epoch": 0.9770593517838191,
"grad_norm": 14.034593283767544,
"learning_rate": 2.389837895721586e-10,
"logits/chosen": -1.5359089374542236,
"logits/rejected": -1.5055701732635498,
"logps/chosen": -792.88037109375,
"logps/rejected": -861.2803955078125,
"loss": 0.4586,
"rewards/accuracies": 0.65625,
"rewards/chosen": -3.479644775390625,
"rewards/margins": 0.4910427927970886,
"rewards/rejected": -3.9706876277923584,
"step": 1496
},
{
"epoch": 0.9777124663237815,
"grad_norm": 57.03647156257448,
"learning_rate": 2.255294124155982e-10,
"logits/chosen": -1.5697842836380005,
"logits/rejected": -1.507396936416626,
"logps/chosen": -896.1490478515625,
"logps/rejected": -974.0924682617188,
"loss": 0.5092,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.4873480796813965,
"rewards/margins": 1.329317331314087,
"rewards/rejected": -4.8166656494140625,
"step": 1497
},
{
"epoch": 0.978365580863744,
"grad_norm": 46.79910545953451,
"learning_rate": 2.124642461850179e-10,
"logits/chosen": -1.5201869010925293,
"logits/rejected": -1.4976688623428345,
"logps/chosen": -853.4832153320312,
"logps/rejected": -928.146728515625,
"loss": 0.4592,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.6100432872772217,
"rewards/margins": 0.6564720273017883,
"rewards/rejected": -4.266514778137207,
"step": 1498
},
{
"epoch": 0.9790186954037065,
"grad_norm": 23.03773550817736,
"learning_rate": 1.997883588863436e-10,
"logits/chosen": -1.5425734519958496,
"logits/rejected": -1.5184190273284912,
"logps/chosen": -913.212890625,
"logps/rejected": -1053.3978271484375,
"loss": 0.4411,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.4705982208251953,
"rewards/margins": 0.9973933696746826,
"rewards/rejected": -4.467991352081299,
"step": 1499
},
{
"epoch": 0.9796718099436689,
"grad_norm": 81.36896110575177,
"learning_rate": 1.875018164992137e-10,
"logits/chosen": -1.534759283065796,
"logits/rejected": -1.4785943031311035,
"logps/chosen": -828.40234375,
"logps/rejected": -832.9805908203125,
"loss": 0.5192,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.459455728530884,
"rewards/margins": 0.5882536172866821,
"rewards/rejected": -4.0477094650268555,
"step": 1500
},
{
"epoch": 0.9796718099436689,
"eval_logits/chosen": -1.4782989025115967,
"eval_logits/rejected": -1.4591182470321655,
"eval_logps/chosen": -844.1143798828125,
"eval_logps/rejected": -928.2655029296875,
"eval_loss": 0.49420246481895447,
"eval_rewards/accuracies": 0.7620000243186951,
"eval_rewards/chosen": -3.33768892288208,
"eval_rewards/margins": 0.9225709438323975,
"eval_rewards/rejected": -4.260260105133057,
"eval_runtime": 296.4379,
"eval_samples_per_second": 13.494,
"eval_steps_per_second": 0.843,
"step": 1500
},
{
"epoch": 0.9803249244836313,
"grad_norm": 26.21653471442728,
"learning_rate": 1.7560468297669606e-10,
"logits/chosen": -1.5513707399368286,
"logits/rejected": -1.5569634437561035,
"logps/chosen": -814.3359375,
"logps/rejected": -854.103759765625,
"loss": 0.4669,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.062058925628662,
"rewards/margins": 0.5742703676223755,
"rewards/rejected": -3.6363296508789062,
"step": 1501
},
{
"epoch": 0.9809780390235938,
"grad_norm": 12.418015543507641,
"learning_rate": 1.640970202449382e-10,
"logits/chosen": -1.4305357933044434,
"logits/rejected": -1.4416850805282593,
"logps/chosen": -852.3341064453125,
"logps/rejected": -901.50537109375,
"loss": 0.4805,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.029092311859131,
"rewards/margins": 0.7821911573410034,
"rewards/rejected": -3.811283588409424,
"step": 1502
},
{
"epoch": 0.9816311535635562,
"grad_norm": 35.91459788005618,
"learning_rate": 1.52978888202826e-10,
"logits/chosen": -1.497631311416626,
"logits/rejected": -1.5434260368347168,
"logps/chosen": -787.7675170898438,
"logps/rejected": -954.6630249023438,
"loss": 0.4877,
"rewards/accuracies": 0.90625,
"rewards/chosen": -3.3874967098236084,
"rewards/margins": 1.1716786623001099,
"rewards/rejected": -4.55917501449585,
"step": 1503
},
{
"epoch": 0.9822842681035187,
"grad_norm": 15.545739037319393,
"learning_rate": 1.4225034472169217e-10,
"logits/chosen": -1.479112982749939,
"logits/rejected": -1.4773443937301636,
"logps/chosen": -777.0709228515625,
"logps/rejected": -973.9321899414062,
"loss": 0.4295,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.061960458755493,
"rewards/margins": 1.352473258972168,
"rewards/rejected": -4.414434432983398,
"step": 1504
},
{
"epoch": 0.9829373826434811,
"grad_norm": 10.049630463778348,
"learning_rate": 1.3191144564502488e-10,
"logits/chosen": -1.5108180046081543,
"logits/rejected": -1.474786400794983,
"logps/chosen": -910.9762573242188,
"logps/rejected": -990.8765869140625,
"loss": 0.5196,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.5121772289276123,
"rewards/margins": 0.993119478225708,
"rewards/rejected": -4.5052971839904785,
"step": 1505
},
{
"epoch": 0.9835904971834436,
"grad_norm": 31.720266818268886,
"learning_rate": 1.2196224478814297e-10,
"logits/chosen": -1.4539670944213867,
"logits/rejected": -1.4328685998916626,
"logps/chosen": -865.9363403320312,
"logps/rejected": -964.106201171875,
"loss": 0.483,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.5068650245666504,
"rewards/margins": 0.9375267028808594,
"rewards/rejected": -4.44439172744751,
"step": 1506
},
{
"epoch": 0.984243611723406,
"grad_norm": 45.556241819502084,
"learning_rate": 1.1240279393793795e-10,
"logits/chosen": -1.5094014406204224,
"logits/rejected": -1.4748339653015137,
"logps/chosen": -926.0755004882812,
"logps/rejected": -976.0172729492188,
"loss": 0.4813,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.7970991134643555,
"rewards/margins": 0.9570468068122864,
"rewards/rejected": -4.754145622253418,
"step": 1507
},
{
"epoch": 0.9848967262633684,
"grad_norm": 40.10867343247857,
"learning_rate": 1.0323314285260731e-10,
"logits/chosen": -1.54648756980896,
"logits/rejected": -1.495444893836975,
"logps/chosen": -783.347900390625,
"logps/rejected": -871.718505859375,
"loss": 0.4933,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.104883909225464,
"rewards/margins": 1.1070107221603394,
"rewards/rejected": -4.211894512176514,
"step": 1508
},
{
"epoch": 0.9855498408033309,
"grad_norm": 32.74695628540036,
"learning_rate": 9.445333926139665e-11,
"logits/chosen": -1.5436725616455078,
"logits/rejected": -1.5149645805358887,
"logps/chosen": -904.5325317382812,
"logps/rejected": -958.0336303710938,
"loss": 0.4634,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.427320957183838,
"rewards/margins": 0.8428292274475098,
"rewards/rejected": -4.270150661468506,
"step": 1509
},
{
"epoch": 0.9862029553432933,
"grad_norm": 50.87038917169162,
"learning_rate": 8.606342886432472e-11,
"logits/chosen": -1.5588243007659912,
"logits/rejected": -1.532243251800537,
"logps/chosen": -860.1405029296875,
"logps/rejected": -904.2451171875,
"loss": 0.4499,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0775833129882812,
"rewards/margins": 0.7431819438934326,
"rewards/rejected": -3.8207650184631348,
"step": 1510
},
{
"epoch": 0.9868560698832558,
"grad_norm": 29.345234968333465,
"learning_rate": 7.806345533197534e-11,
"logits/chosen": -1.5143674612045288,
"logits/rejected": -1.483849287033081,
"logps/chosen": -851.5826416015625,
"logps/rejected": -954.4253540039062,
"loss": 0.5093,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.5056495666503906,
"rewards/margins": 0.8994309902191162,
"rewards/rejected": -4.405080795288086,
"step": 1511
},
{
"epoch": 0.9875091844232182,
"grad_norm": 22.39474777978432,
"learning_rate": 7.045346030526423e-11,
"logits/chosen": -1.5250229835510254,
"logits/rejected": -1.4954551458358765,
"logps/chosen": -931.7249145507812,
"logps/rejected": -1065.670166015625,
"loss": 0.4391,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.5478012561798096,
"rewards/margins": 1.2276500463485718,
"rewards/rejected": -4.77545166015625,
"step": 1512
},
{
"epoch": 0.9881622989631806,
"grad_norm": 68.548727751151,
"learning_rate": 6.323348339521418e-11,
"logits/chosen": -1.5496234893798828,
"logits/rejected": -1.5093997716903687,
"logps/chosen": -938.7208251953125,
"logps/rejected": -940.5839233398438,
"loss": 0.5334,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.8004541397094727,
"rewards/margins": 0.6491421461105347,
"rewards/rejected": -4.449596405029297,
"step": 1513
},
{
"epoch": 0.9888154135031431,
"grad_norm": 15.003945051271025,
"learning_rate": 5.640356218274689e-11,
"logits/chosen": -1.4270832538604736,
"logits/rejected": -1.4271320104599,
"logps/chosen": -785.04052734375,
"logps/rejected": -928.6375732421875,
"loss": 0.5223,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.0938949584960938,
"rewards/margins": 0.9844279885292053,
"rewards/rejected": -4.078322887420654,
"step": 1514
},
{
"epoch": 0.9894685280431056,
"grad_norm": 52.68067104199576,
"learning_rate": 4.996373221849981e-11,
"logits/chosen": -1.4737398624420166,
"logits/rejected": -1.478607177734375,
"logps/chosen": -859.705810546875,
"logps/rejected": -980.18701171875,
"loss": 0.4386,
"rewards/accuracies": 0.875,
"rewards/chosen": -3.2958946228027344,
"rewards/margins": 1.2142889499664307,
"rewards/rejected": -4.510182857513428,
"step": 1515
},
{
"epoch": 0.990121642583068,
"grad_norm": 41.105753672966415,
"learning_rate": 4.391402702263458e-11,
"logits/chosen": -1.5442856550216675,
"logits/rejected": -1.5355677604675293,
"logps/chosen": -766.7518920898438,
"logps/rejected": -807.2686157226562,
"loss": 0.4765,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.052950620651245,
"rewards/margins": 0.6595571041107178,
"rewards/rejected": -3.712507724761963,
"step": 1516
},
{
"epoch": 0.9907747571230304,
"grad_norm": 68.18799629848323,
"learning_rate": 3.82544780846622e-11,
"logits/chosen": -1.4942461252212524,
"logits/rejected": -1.4979506731033325,
"logps/chosen": -900.58935546875,
"logps/rejected": -1016.7969970703125,
"loss": 0.4941,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.9103312492370605,
"rewards/margins": 0.9849377870559692,
"rewards/rejected": -4.89526891708374,
"step": 1517
},
{
"epoch": 0.9914278716629928,
"grad_norm": 20.36699675427797,
"learning_rate": 3.2985114863276484e-11,
"logits/chosen": -1.4821314811706543,
"logits/rejected": -1.4812616109848022,
"logps/chosen": -811.2876586914062,
"logps/rejected": -998.925048828125,
"loss": 0.5106,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.557675838470459,
"rewards/margins": 1.2705349922180176,
"rewards/rejected": -4.828211307525635,
"step": 1518
},
{
"epoch": 0.9920809862029554,
"grad_norm": 58.88999010926919,
"learning_rate": 2.810596478619587e-11,
"logits/chosen": -1.5028977394104004,
"logits/rejected": -1.4703741073608398,
"logps/chosen": -901.3570556640625,
"logps/rejected": -913.9461059570312,
"loss": 0.5199,
"rewards/accuracies": 0.625,
"rewards/chosen": -3.426830291748047,
"rewards/margins": 0.6236191391944885,
"rewards/rejected": -4.050449371337891,
"step": 1519
},
{
"epoch": 0.9927341007429178,
"grad_norm": 34.40888479605642,
"learning_rate": 2.3617053250046815e-11,
"logits/chosen": -1.5230963230133057,
"logits/rejected": -1.4947023391723633,
"logps/chosen": -770.20849609375,
"logps/rejected": -841.6630859375,
"loss": 0.4742,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.0242300033569336,
"rewards/margins": 0.8307550549507141,
"rewards/rejected": -3.854984760284424,
"step": 1520
},
{
"epoch": 0.9933872152828802,
"grad_norm": 22.3379376658494,
"learning_rate": 1.951840362018897e-11,
"logits/chosen": -1.5373893976211548,
"logits/rejected": -1.5419368743896484,
"logps/chosen": -797.622314453125,
"logps/rejected": -868.178955078125,
"loss": 0.4944,
"rewards/accuracies": 0.71875,
"rewards/chosen": -3.065598249435425,
"rewards/margins": 0.9640765190124512,
"rewards/rejected": -4.029674530029297,
"step": 1521
},
{
"epoch": 0.9940403298228426,
"grad_norm": 37.37852136541512,
"learning_rate": 1.5810037230648554e-11,
"logits/chosen": -1.4990514516830444,
"logits/rejected": -1.498454213142395,
"logps/chosen": -753.627685546875,
"logps/rejected": -842.1553344726562,
"loss": 0.4797,
"rewards/accuracies": 0.84375,
"rewards/chosen": -2.89300537109375,
"rewards/margins": 0.8533806204795837,
"rewards/rejected": -3.7463865280151367,
"step": 1522
},
{
"epoch": 0.9946934443628052,
"grad_norm": 35.34109312088976,
"learning_rate": 1.2491973383951803e-11,
"logits/chosen": -1.433802843093872,
"logits/rejected": -1.4264992475509644,
"logps/chosen": -843.60986328125,
"logps/rejected": -899.9165649414062,
"loss": 0.5295,
"rewards/accuracies": 0.6875,
"rewards/chosen": -3.5381946563720703,
"rewards/margins": 0.6915282011032104,
"rewards/rejected": -4.229722499847412,
"step": 1523
},
{
"epoch": 0.9953465589027676,
"grad_norm": 25.436468352463912,
"learning_rate": 9.564229351050056e-12,
"logits/chosen": -1.4901500940322876,
"logits/rejected": -1.4961150884628296,
"logps/chosen": -803.0394287109375,
"logps/rejected": -855.164794921875,
"loss": 0.5505,
"rewards/accuracies": 0.78125,
"rewards/chosen": -2.957271099090576,
"rewards/margins": 0.79988694190979,
"rewards/rejected": -3.757158041000366,
"step": 1524
},
{
"epoch": 0.99599967344273,
"grad_norm": 10.73432201222575,
"learning_rate": 7.0268203712448015e-12,
"logits/chosen": -1.5934062004089355,
"logits/rejected": -1.544229507446289,
"logps/chosen": -856.8956298828125,
"logps/rejected": -1123.9691162109375,
"loss": 0.4846,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.3569083213806152,
"rewards/margins": 2.04354190826416,
"rewards/rejected": -5.400450706481934,
"step": 1525
},
{
"epoch": 0.9966527879826924,
"grad_norm": 87.38096016034348,
"learning_rate": 4.879759652079429e-12,
"logits/chosen": -1.467991828918457,
"logits/rejected": -1.427109718322754,
"logps/chosen": -805.0746459960938,
"logps/rejected": -918.366455078125,
"loss": 0.5265,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.3931772708892822,
"rewards/margins": 0.822685956954956,
"rewards/rejected": -4.215863227844238,
"step": 1526
},
{
"epoch": 0.9973059025226549,
"grad_norm": 24.30638520435916,
"learning_rate": 3.123058369280951e-12,
"logits/chosen": -1.4472224712371826,
"logits/rejected": -1.4677808284759521,
"logps/chosen": -832.0319213867188,
"logps/rejected": -963.2366333007812,
"loss": 0.5076,
"rewards/accuracies": 0.8125,
"rewards/chosen": -3.1619575023651123,
"rewards/margins": 0.8776848912239075,
"rewards/rejected": -4.039642333984375,
"step": 1527
},
{
"epoch": 0.9979590170626174,
"grad_norm": 17.20772991475794,
"learning_rate": 1.756725666710035e-12,
"logits/chosen": -1.4970341920852661,
"logits/rejected": -1.4706482887268066,
"logps/chosen": -862.6663818359375,
"logps/rejected": -1025.07177734375,
"loss": 0.4429,
"rewards/accuracies": 0.78125,
"rewards/chosen": -3.1549482345581055,
"rewards/margins": 1.3957924842834473,
"rewards/rejected": -4.550740718841553,
"step": 1528
},
{
"epoch": 0.9986121316025798,
"grad_norm": 17.972242666702652,
"learning_rate": 7.80768656319375e-13,
"logits/chosen": -1.409681797027588,
"logits/rejected": -1.3731991052627563,
"logps/chosen": -763.9820556640625,
"logps/rejected": -802.5194702148438,
"loss": 0.4251,
"rewards/accuracies": 0.75,
"rewards/chosen": -3.1270015239715576,
"rewards/margins": 0.6430358290672302,
"rewards/rejected": -3.7700371742248535,
"step": 1529
},
{
"epoch": 0.9992652461425422,
"grad_norm": 30.620744387516318,
"learning_rate": 1.9519241807874897e-13,
"logits/chosen": -1.5083037614822388,
"logits/rejected": -1.4722343683242798,
"logps/chosen": -971.5647583007812,
"logps/rejected": -995.6734619140625,
"loss": 0.4784,
"rewards/accuracies": 0.71875,
"rewards/chosen": -4.024622440338135,
"rewards/margins": 0.6361601948738098,
"rewards/rejected": -4.660782814025879,
"step": 1530
},
{
"epoch": 0.9999183606825047,
"grad_norm": 25.960604541088838,
"learning_rate": 0.0,
"logits/chosen": -1.506800651550293,
"logits/rejected": -1.5056391954421997,
"logps/chosen": -892.1127319335938,
"logps/rejected": -1093.7352294921875,
"loss": 0.4507,
"rewards/accuracies": 0.84375,
"rewards/chosen": -3.275479793548584,
"rewards/margins": 1.1585028171539307,
"rewards/rejected": -4.433982849121094,
"step": 1531
},
{
"epoch": 0.9999183606825047,
"step": 1531,
"total_flos": 0.0,
"train_loss": 0.5524635882372952,
"train_runtime": 40329.8241,
"train_samples_per_second": 4.859,
"train_steps_per_second": 0.038
}
],
"logging_steps": 1,
"max_steps": 1531,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 500,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 2,
"trial_name": null,
"trial_params": null
}
Reference in New Issue View Git Blame Copy Permalink