YuchenLi01/genv3pair1NoGT_1.5B_cdpo_ebs32_lr1e-06_beta0.1_epoch16.0_42
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity
Files
742c3e40b827c42ef15da08514c62b32c4bc0f5b
genv3pair1NoGT_1.5B_cdpo_eb…/trainer_state.json
ModelHub XC 742c3e40b8 初始化项目，由ModelHub XC社区提供模型 
Model: YuchenLi01/genv3pair1NoGT_1.5B_cdpo_ebs32_lr1e-06_beta0.1_epoch16.0_42
Source: Original Platform
2026-05-25 23:10:23 +08:00

45291 lines
1.5 MiB
Raw Blame History

{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 16.0,
"eval_steps": 20,
"global_step": 2864,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.00558659217877095,
"grad_norm": 28.603565649816808,
"learning_rate": 3.4843205574912892e-09,
"logits/chosen": -2.4004337787628174,
"logits/rejected": -2.5142080783843994,
"logps/chosen": -60.584632873535156,
"logps/rejected": -42.787071228027344,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.0111731843575419,
"grad_norm": 29.276010564093625,
"learning_rate": 6.9686411149825785e-09,
"logits/chosen": -2.4949634075164795,
"logits/rejected": -2.4507484436035156,
"logps/chosen": -19.39887046813965,
"logps/rejected": -25.19920539855957,
"loss": 0.6931,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 2
},
{
"epoch": 0.01675977653631285,
"grad_norm": 28.859494722690478,
"learning_rate": 1.0452961672473868e-08,
"logits/chosen": -2.2804834842681885,
"logits/rejected": -2.27510666847229,
"logps/chosen": -27.924236297607422,
"logps/rejected": -39.671104431152344,
"loss": 0.689,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.009498548693954945,
"rewards/margins": 0.009498548693954945,
"rewards/rejected": 0.0,
"step": 3
},
{
"epoch": 0.0223463687150838,
"grad_norm": 32.50888450037989,
"learning_rate": 1.3937282229965157e-08,
"logits/chosen": -2.056849718093872,
"logits/rejected": -2.039930820465088,
"logps/chosen": -50.10342788696289,
"logps/rejected": -44.76918029785156,
"loss": 0.6933,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.02802419662475586,
"rewards/margins": 0.02802419662475586,
"rewards/rejected": 0.0,
"step": 4
},
{
"epoch": 0.027932960893854747,
"grad_norm": 28.857154004570457,
"learning_rate": 1.7421602787456446e-08,
"logits/chosen": -2.272435426712036,
"logits/rejected": -2.301142454147339,
"logps/chosen": -38.176116943359375,
"logps/rejected": -50.210296630859375,
"loss": 0.6965,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.011784981936216354,
"rewards/margins": -0.011784981936216354,
"rewards/rejected": 0.0,
"step": 5
},
{
"epoch": 0.0335195530726257,
"grad_norm": 26.415733391767393,
"learning_rate": 2.0905923344947736e-08,
"logits/chosen": -2.793938159942627,
"logits/rejected": -2.4558026790618896,
"logps/chosen": -18.068748474121094,
"logps/rejected": -63.842552185058594,
"loss": 0.6959,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.01914348639547825,
"rewards/margins": 0.01914348639547825,
"rewards/rejected": 0.0,
"step": 6
},
{
"epoch": 0.03910614525139665,
"grad_norm": 30.429564005701334,
"learning_rate": 2.4390243902439023e-08,
"logits/chosen": -2.6528186798095703,
"logits/rejected": -2.5782175064086914,
"logps/chosen": -18.585830688476562,
"logps/rejected": -21.081623077392578,
"loss": 0.6907,
"rewards/accuracies": 0.0,
"rewards/chosen": -0.010954786092042923,
"rewards/margins": -0.010954786092042923,
"rewards/rejected": 0.0,
"step": 7
},
{
"epoch": 0.0446927374301676,
"grad_norm": 31.759150028352973,
"learning_rate": 2.7874564459930314e-08,
"logits/chosen": -2.546023368835449,
"logits/rejected": -2.566352128982544,
"logps/chosen": -20.15776824951172,
"logps/rejected": -29.43256950378418,
"loss": 0.6885,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.008025838062167168,
"rewards/margins": 0.008025838062167168,
"rewards/rejected": 0.0,
"step": 8
},
{
"epoch": 0.05027932960893855,
"grad_norm": 26.754315287337864,
"learning_rate": 3.13588850174216e-08,
"logits/chosen": -2.5980911254882812,
"logits/rejected": -2.507265567779541,
"logps/chosen": -28.086103439331055,
"logps/rejected": -36.260292053222656,
"loss": 0.6946,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.005537462420761585,
"rewards/margins": -0.005537462420761585,
"rewards/rejected": 0.0,
"step": 9
},
{
"epoch": 0.055865921787709494,
"grad_norm": 27.21359249234941,
"learning_rate": 3.484320557491289e-08,
"logits/chosen": -2.3490583896636963,
"logits/rejected": -2.2292792797088623,
"logps/chosen": -53.409996032714844,
"logps/rejected": -55.07990264892578,
"loss": 0.693,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.013922978192567825,
"rewards/margins": -0.013922978192567825,
"rewards/rejected": 0.0,
"step": 10
},
{
"epoch": 0.061452513966480445,
"grad_norm": 28.43356949422403,
"learning_rate": 3.832752613240418e-08,
"logits/chosen": -2.4062752723693848,
"logits/rejected": -2.2725396156311035,
"logps/chosen": -18.990951538085938,
"logps/rejected": -27.86153793334961,
"loss": 0.6949,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.00549395103007555,
"rewards/margins": 0.00549395103007555,
"rewards/rejected": 0.0,
"step": 11
},
{
"epoch": 0.0670391061452514,
"grad_norm": 26.03024884762362,
"learning_rate": 4.181184668989547e-08,
"logits/chosen": -2.7887275218963623,
"logits/rejected": -2.4660379886627197,
"logps/chosen": -15.426248550415039,
"logps/rejected": -43.07337188720703,
"loss": 0.6963,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.004401779733598232,
"rewards/margins": -0.004401779733598232,
"rewards/rejected": 0.0,
"step": 12
},
{
"epoch": 0.07262569832402235,
"grad_norm": 29.15535304010801,
"learning_rate": 4.529616724738676e-08,
"logits/chosen": -2.372344493865967,
"logits/rejected": -2.0893101692199707,
"logps/chosen": -33.84700393676758,
"logps/rejected": -52.90555953979492,
"loss": 0.6907,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.025128602981567383,
"rewards/margins": 0.025128602981567383,
"rewards/rejected": 0.0,
"step": 13
},
{
"epoch": 0.0782122905027933,
"grad_norm": 29.514245218176903,
"learning_rate": 4.878048780487805e-08,
"logits/chosen": -2.565272331237793,
"logits/rejected": -2.392062187194824,
"logps/chosen": -18.883819580078125,
"logps/rejected": -48.590614318847656,
"loss": 0.6956,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.01730327680706978,
"rewards/margins": 0.01730327680706978,
"rewards/rejected": 0.0,
"step": 14
},
{
"epoch": 0.08379888268156424,
"grad_norm": 30.166208831944306,
"learning_rate": 5.2264808362369334e-08,
"logits/chosen": -2.1422746181488037,
"logits/rejected": -2.060894012451172,
"logps/chosen": -42.46976852416992,
"logps/rejected": -50.680320739746094,
"loss": 0.6963,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0040832520462572575,
"rewards/margins": -0.0040832520462572575,
"rewards/rejected": 0.0,
"step": 15
},
{
"epoch": 0.0893854748603352,
"grad_norm": 27.010877795355334,
"learning_rate": 5.574912891986063e-08,
"logits/chosen": -2.048851728439331,
"logits/rejected": -2.078244686126709,
"logps/chosen": -32.976806640625,
"logps/rejected": -39.53431701660156,
"loss": 0.6963,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.01168208196759224,
"rewards/margins": -0.01168208196759224,
"rewards/rejected": 0.0,
"step": 16
},
{
"epoch": 0.09497206703910614,
"grad_norm": 31.303112360150603,
"learning_rate": 5.9233449477351915e-08,
"logits/chosen": -2.792558193206787,
"logits/rejected": -2.1472291946411133,
"logps/chosen": -24.31830596923828,
"logps/rejected": -71.05191040039062,
"loss": 0.6949,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.008655214682221413,
"rewards/margins": 0.008655214682221413,
"rewards/rejected": 0.0,
"step": 17
},
{
"epoch": 0.1005586592178771,
"grad_norm": 30.389870821612302,
"learning_rate": 6.27177700348432e-08,
"logits/chosen": -2.4866530895233154,
"logits/rejected": -2.375359296798706,
"logps/chosen": -37.48762512207031,
"logps/rejected": -38.95714569091797,
"loss": 0.6976,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.0314456932246685,
"rewards/margins": -0.0314456932246685,
"rewards/rejected": 0.0,
"step": 18
},
{
"epoch": 0.10614525139664804,
"grad_norm": 29.563734417452384,
"learning_rate": 6.620209059233449e-08,
"logits/chosen": -2.5353310108184814,
"logits/rejected": -2.5976669788360596,
"logps/chosen": -24.29471778869629,
"logps/rejected": -25.59542465209961,
"loss": 0.694,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.017024923115968704,
"rewards/margins": 0.017024923115968704,
"rewards/rejected": 0.0,
"step": 19
},
{
"epoch": 0.11173184357541899,
"grad_norm": 28.023653019850883,
"learning_rate": 6.968641114982578e-08,
"logits/chosen": -2.481630563735962,
"logits/rejected": -2.3800740242004395,
"logps/chosen": -43.142112731933594,
"logps/rejected": -45.48561096191406,
"loss": 0.6912,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.030663300305604935,
"rewards/margins": -0.030663300305604935,
"rewards/rejected": 0.0,
"step": 20
},
{
"epoch": 0.11173184357541899,
"eval_logits/chosen": -2.378596305847168,
"eval_logits/rejected": -2.231410264968872,
"eval_logps/chosen": -30.04729652404785,
"eval_logps/rejected": -41.63738250732422,
"eval_loss": 0.6928394436836243,
"eval_rewards/accuracies": 0.5,
"eval_rewards/chosen": 0.002274421276524663,
"eval_rewards/margins": 0.002274421276524663,
"eval_rewards/rejected": 0.0,
"eval_runtime": 36.2639,
"eval_samples_per_second": 8.548,
"eval_steps_per_second": 0.276,
"step": 20
},
{
"epoch": 0.11731843575418995,
"grad_norm": 28.713464267987312,
"learning_rate": 7.317073170731706e-08,
"logits/chosen": -2.1179280281066895,
"logits/rejected": -1.9675930738449097,
"logps/chosen": -39.224002838134766,
"logps/rejected": -41.40558624267578,
"loss": 0.6925,
"rewards/accuracies": 0.5,
"rewards/chosen": 9.074225090444088e-05,
"rewards/margins": 9.074225090444088e-05,
"rewards/rejected": 0.0,
"step": 21
},
{
"epoch": 0.12290502793296089,
"grad_norm": 29.03402105884427,
"learning_rate": 7.665505226480836e-08,
"logits/chosen": -2.4489192962646484,
"logits/rejected": -2.542243242263794,
"logps/chosen": -33.14081954956055,
"logps/rejected": -38.48543167114258,
"loss": 0.6947,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.018711712211370468,
"rewards/margins": -0.018711712211370468,
"rewards/rejected": 0.0,
"step": 22
},
{
"epoch": 0.12849162011173185,
"grad_norm": 33.33325134064739,
"learning_rate": 8.013937282229964e-08,
"logits/chosen": -2.0170164108276367,
"logits/rejected": -2.1128463745117188,
"logps/chosen": -98.69719696044922,
"logps/rejected": -58.296714782714844,
"loss": 0.6915,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.018199540674686432,
"rewards/margins": -0.018199540674686432,
"rewards/rejected": 0.0,
"step": 23
},
{
"epoch": 0.1340782122905028,
"grad_norm": 27.061983809004143,
"learning_rate": 8.362369337979095e-08,
"logits/chosen": -2.227400302886963,
"logits/rejected": -2.2094178199768066,
"logps/chosen": -33.232845306396484,
"logps/rejected": -34.66838836669922,
"loss": 0.6937,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.0005850789602845907,
"rewards/margins": 0.0005850789602845907,
"rewards/rejected": 0.0,
"step": 24
},
{
"epoch": 0.13966480446927373,
"grad_norm": 30.631315818668092,
"learning_rate": 8.710801393728223e-08,
"logits/chosen": -2.4377002716064453,
"logits/rejected": -2.3588991165161133,
"logps/chosen": -21.7973575592041,
"logps/rejected": -40.57929611206055,
"loss": 0.6949,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.02112269401550293,
"rewards/margins": 0.02112269401550293,
"rewards/rejected": 0.0,
"step": 25
},
{
"epoch": 0.1452513966480447,
"grad_norm": 29.378197798170234,
"learning_rate": 9.059233449477352e-08,
"logits/chosen": -2.628924608230591,
"logits/rejected": -2.482461452484131,
"logps/chosen": -23.92325210571289,
"logps/rejected": -44.77006530761719,
"loss": 0.6928,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.004933738149702549,
"rewards/margins": 0.004933738149702549,
"rewards/rejected": 0.0,
"step": 26
},
{
"epoch": 0.15083798882681565,
"grad_norm": 29.535605005914892,
"learning_rate": 9.407665505226481e-08,
"logits/chosen": -2.4817450046539307,
"logits/rejected": -2.485926866531372,
"logps/chosen": -59.53948211669922,
"logps/rejected": -56.88829040527344,
"loss": 0.6901,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.02714724652469158,
"rewards/margins": 0.02714724652469158,
"rewards/rejected": 0.0,
"step": 27
},
{
"epoch": 0.1564245810055866,
"grad_norm": 27.59267993779762,
"learning_rate": 9.75609756097561e-08,
"logits/chosen": -2.586944580078125,
"logits/rejected": -2.4643876552581787,
"logps/chosen": -31.718379974365234,
"logps/rejected": -36.230384826660156,
"loss": 0.6909,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.027640867978334427,
"rewards/margins": -0.027640867978334427,
"rewards/rejected": 0.0,
"step": 28
},
{
"epoch": 0.16201117318435754,
"grad_norm": 29.667583017088273,
"learning_rate": 1.0104529616724739e-07,
"logits/chosen": -2.30635404586792,
"logits/rejected": -2.212883710861206,
"logps/chosen": -34.53228759765625,
"logps/rejected": -37.26000213623047,
"loss": 0.6898,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.021092604845762253,
"rewards/margins": 0.021092604845762253,
"rewards/rejected": 0.0,
"step": 29
},
{
"epoch": 0.16759776536312848,
"grad_norm": 29.98028101754352,
"learning_rate": 1.0452961672473867e-07,
"logits/chosen": -2.4045355319976807,
"logits/rejected": -2.4857683181762695,
"logps/chosen": -34.87571716308594,
"logps/rejected": -52.67020034790039,
"loss": 0.689,
"rewards/accuracies": 0.25,
"rewards/chosen": -0.011652231216430664,
"rewards/margins": -0.011652231216430664,
"rewards/rejected": 0.0,
"step": 30
},
{
"epoch": 0.17318435754189945,
"grad_norm": 30.079938013679776,
"learning_rate": 1.0801393728222996e-07,
"logits/chosen": -2.4215786457061768,
"logits/rejected": -2.1780693531036377,
"logps/chosen": -20.111202239990234,
"logps/rejected": -49.6775016784668,
"loss": 0.693,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.0008932588389143348,
"rewards/margins": -0.0008932588389143348,
"rewards/rejected": 0.0,
"step": 31
},
{
"epoch": 0.1787709497206704,
"grad_norm": 29.0354625721042,
"learning_rate": 1.1149825783972126e-07,
"logits/chosen": -2.314462423324585,
"logits/rejected": -2.171380043029785,
"logps/chosen": -28.89542007446289,
"logps/rejected": -44.24406433105469,
"loss": 0.6917,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.028475143015384674,
"rewards/margins": 0.028475143015384674,
"rewards/rejected": 0.0,
"step": 32
},
{
"epoch": 0.18435754189944134,
"grad_norm": 28.82319834972392,
"learning_rate": 1.1498257839721254e-07,
"logits/chosen": -2.473206043243408,
"logits/rejected": -2.5393941402435303,
"logps/chosen": -28.577171325683594,
"logps/rejected": -34.01260757446289,
"loss": 0.687,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.00867919996380806,
"rewards/margins": 0.00867919996380806,
"rewards/rejected": 0.0,
"step": 33
},
{
"epoch": 0.18994413407821228,
"grad_norm": 26.99801083582208,
"learning_rate": 1.1846689895470383e-07,
"logits/chosen": -2.5871434211730957,
"logits/rejected": -2.204688310623169,
"logps/chosen": -24.449928283691406,
"logps/rejected": -66.65013122558594,
"loss": 0.6854,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.022472620010375977,
"rewards/margins": 0.022472620010375977,
"rewards/rejected": 0.0,
"step": 34
},
{
"epoch": 0.19553072625698323,
"grad_norm": 28.371452721285568,
"learning_rate": 1.219512195121951e-07,
"logits/chosen": -2.438382148742676,
"logits/rejected": -2.376980781555176,
"logps/chosen": -20.550439834594727,
"logps/rejected": -30.104656219482422,
"loss": 0.6828,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.009609270840883255,
"rewards/margins": 0.009609270840883255,
"rewards/rejected": 0.0,
"step": 35
},
{
"epoch": 0.2011173184357542,
"grad_norm": 27.7150904468452,
"learning_rate": 1.254355400696864e-07,
"logits/chosen": -2.1353628635406494,
"logits/rejected": -2.0030341148376465,
"logps/chosen": -52.302974700927734,
"logps/rejected": -63.30662536621094,
"loss": 0.6824,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.006544208619743586,
"rewards/margins": 0.006544208619743586,
"rewards/rejected": 0.0,
"step": 36
},
{
"epoch": 0.20670391061452514,
"grad_norm": 28.68285689295653,
"learning_rate": 1.289198606271777e-07,
"logits/chosen": -2.4385170936584473,
"logits/rejected": -2.4843199253082275,
"logps/chosen": -30.546653747558594,
"logps/rejected": -37.64432907104492,
"loss": 0.6874,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.010574293322861195,
"rewards/margins": 0.010574293322861195,
"rewards/rejected": 0.0,
"step": 37
},
{
"epoch": 0.2122905027932961,
"grad_norm": 25.934450282985118,
"learning_rate": 1.3240418118466898e-07,
"logits/chosen": -2.391115665435791,
"logits/rejected": -2.1517765522003174,
"logps/chosen": -36.59046936035156,
"logps/rejected": -54.95623016357422,
"loss": 0.682,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.036093570291996,
"rewards/margins": 0.036093570291996,
"rewards/rejected": 0.0,
"step": 38
},
{
"epoch": 0.21787709497206703,
"grad_norm": 29.575773366216787,
"learning_rate": 1.3588850174216026e-07,
"logits/chosen": -2.0830588340759277,
"logits/rejected": -2.0630950927734375,
"logps/chosen": -51.38600158691406,
"logps/rejected": -51.52277374267578,
"loss": 0.6753,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.047041941434144974,
"rewards/margins": 0.047041941434144974,
"rewards/rejected": 0.0,
"step": 39
},
{
"epoch": 0.22346368715083798,
"grad_norm": 27.986584327852498,
"learning_rate": 1.3937282229965157e-07,
"logits/chosen": -2.1863017082214355,
"logits/rejected": -2.054046630859375,
"logps/chosen": -36.82038879394531,
"logps/rejected": -28.519256591796875,
"loss": 0.6775,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.05012211948633194,
"rewards/margins": 0.05012211948633194,
"rewards/rejected": 0.0,
"step": 40
},
{
"epoch": 0.22346368715083798,
"eval_logits/chosen": -2.383514881134033,
"eval_logits/rejected": -2.2417750358581543,
"eval_logps/chosen": -29.78684425354004,
"eval_logps/rejected": -41.39722442626953,
"eval_loss": 0.6780919432640076,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.028319690376520157,
"eval_rewards/margins": 0.028319690376520157,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.55,
"eval_samples_per_second": 9.24,
"eval_steps_per_second": 0.298,
"step": 40
},
{
"epoch": 0.22905027932960895,
"grad_norm": 27.886526350000842,
"learning_rate": 1.4285714285714285e-07,
"logits/chosen": -2.4471418857574463,
"logits/rejected": -2.276179552078247,
"logps/chosen": -38.71142578125,
"logps/rejected": -44.0977783203125,
"loss": 0.675,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.045044660568237305,
"rewards/margins": 0.045044660568237305,
"rewards/rejected": 0.0,
"step": 41
},
{
"epoch": 0.2346368715083799,
"grad_norm": 28.19236917478761,
"learning_rate": 1.4634146341463413e-07,
"logits/chosen": -2.6012725830078125,
"logits/rejected": -2.3238584995269775,
"logps/chosen": -21.031005859375,
"logps/rejected": -43.29925537109375,
"loss": 0.6763,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.051784299314022064,
"rewards/margins": 0.051784299314022064,
"rewards/rejected": 0.0,
"step": 42
},
{
"epoch": 0.24022346368715083,
"grad_norm": 27.00095263618661,
"learning_rate": 1.498257839721254e-07,
"logits/chosen": -2.532641887664795,
"logits/rejected": -2.5259850025177,
"logps/chosen": -46.74769592285156,
"logps/rejected": -31.371822357177734,
"loss": 0.6785,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.03024010732769966,
"rewards/margins": 0.03024010732769966,
"rewards/rejected": 0.0,
"step": 43
},
{
"epoch": 0.24581005586592178,
"grad_norm": 27.735960230898932,
"learning_rate": 1.5331010452961672e-07,
"logits/chosen": -2.554621934890747,
"logits/rejected": -2.3309123516082764,
"logps/chosen": -29.774852752685547,
"logps/rejected": -50.249359130859375,
"loss": 0.6735,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.015763472765684128,
"rewards/margins": 0.015763472765684128,
"rewards/rejected": 0.0,
"step": 44
},
{
"epoch": 0.25139664804469275,
"grad_norm": 25.760071781763443,
"learning_rate": 1.56794425087108e-07,
"logits/chosen": -2.679527521133423,
"logits/rejected": -2.3180019855499268,
"logps/chosen": -29.519779205322266,
"logps/rejected": -51.563804626464844,
"loss": 0.6833,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.019712496548891068,
"rewards/margins": 0.019712496548891068,
"rewards/rejected": 0.0,
"step": 45
},
{
"epoch": 0.2569832402234637,
"grad_norm": 27.542371432519047,
"learning_rate": 1.6027874564459928e-07,
"logits/chosen": -1.9994746446609497,
"logits/rejected": -2.045717239379883,
"logps/chosen": -67.6978530883789,
"logps/rejected": -36.14336395263672,
"loss": 0.67,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.0912964791059494,
"rewards/margins": 0.0912964791059494,
"rewards/rejected": 0.0,
"step": 46
},
{
"epoch": 0.26256983240223464,
"grad_norm": 26.745703810276925,
"learning_rate": 1.6376306620209058e-07,
"logits/chosen": -2.550858736038208,
"logits/rejected": -2.3014934062957764,
"logps/chosen": -26.110837936401367,
"logps/rejected": -42.080474853515625,
"loss": 0.6742,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.017519522458314896,
"rewards/margins": 0.017519522458314896,
"rewards/rejected": 0.0,
"step": 47
},
{
"epoch": 0.2681564245810056,
"grad_norm": 26.81973377757222,
"learning_rate": 1.672473867595819e-07,
"logits/chosen": -2.5700764656066895,
"logits/rejected": -2.4172418117523193,
"logps/chosen": -29.44329071044922,
"logps/rejected": -34.067237854003906,
"loss": 0.6669,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.0359685905277729,
"rewards/margins": 0.0359685905277729,
"rewards/rejected": 0.0,
"step": 48
},
{
"epoch": 0.2737430167597765,
"grad_norm": 27.593766502688794,
"learning_rate": 1.7073170731707317e-07,
"logits/chosen": -2.271967887878418,
"logits/rejected": -2.3903441429138184,
"logps/chosen": -56.53013610839844,
"logps/rejected": -41.234901428222656,
"loss": 0.6624,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.06545233726501465,
"rewards/margins": 0.06545233726501465,
"rewards/rejected": 0.0,
"step": 49
},
{
"epoch": 0.27932960893854747,
"grad_norm": 23.1080148340995,
"learning_rate": 1.7421602787456445e-07,
"logits/chosen": -2.7415006160736084,
"logits/rejected": -2.479442834854126,
"logps/chosen": -22.101638793945312,
"logps/rejected": -32.386566162109375,
"loss": 0.6633,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.07915637642145157,
"rewards/margins": 0.07915637642145157,
"rewards/rejected": 0.0,
"step": 50
},
{
"epoch": 0.2849162011173184,
"grad_norm": 26.221650919673365,
"learning_rate": 1.7770034843205576e-07,
"logits/chosen": -2.509009599685669,
"logits/rejected": -1.923651933670044,
"logps/chosen": -25.04236602783203,
"logps/rejected": -69.1952133178711,
"loss": 0.6521,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.09428238868713379,
"rewards/margins": 0.09428238868713379,
"rewards/rejected": 0.0,
"step": 51
},
{
"epoch": 0.2905027932960894,
"grad_norm": 24.728560083374884,
"learning_rate": 1.8118466898954704e-07,
"logits/chosen": -2.480639696121216,
"logits/rejected": -2.4123823642730713,
"logps/chosen": -23.67394256591797,
"logps/rejected": -28.476255416870117,
"loss": 0.6529,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.08266758918762207,
"rewards/margins": 0.08266758918762207,
"rewards/rejected": 0.0,
"step": 52
},
{
"epoch": 0.29608938547486036,
"grad_norm": 24.21861955215401,
"learning_rate": 1.8466898954703832e-07,
"logits/chosen": -2.817721128463745,
"logits/rejected": -2.583847761154175,
"logps/chosen": -28.653152465820312,
"logps/rejected": -53.04334259033203,
"loss": 0.6493,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.12885236740112305,
"rewards/margins": 0.12885236740112305,
"rewards/rejected": 0.0,
"step": 53
},
{
"epoch": 0.3016759776536313,
"grad_norm": 27.927454062977144,
"learning_rate": 1.8815331010452963e-07,
"logits/chosen": -2.7345516681671143,
"logits/rejected": -2.321146011352539,
"logps/chosen": -23.42495346069336,
"logps/rejected": -48.8504524230957,
"loss": 0.6421,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.08252206444740295,
"rewards/margins": 0.08252206444740295,
"rewards/rejected": 0.0,
"step": 54
},
{
"epoch": 0.30726256983240224,
"grad_norm": 26.878483310239424,
"learning_rate": 1.916376306620209e-07,
"logits/chosen": -2.289569616317749,
"logits/rejected": -2.479778289794922,
"logps/chosen": -55.436302185058594,
"logps/rejected": -37.89727020263672,
"loss": 0.6379,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.15553730726242065,
"rewards/margins": 0.15553730726242065,
"rewards/rejected": 0.0,
"step": 55
},
{
"epoch": 0.3128491620111732,
"grad_norm": 26.65427216585537,
"learning_rate": 1.951219512195122e-07,
"logits/chosen": -2.139517068862915,
"logits/rejected": -2.0566482543945312,
"logps/chosen": -48.204933166503906,
"logps/rejected": -44.711669921875,
"loss": 0.6426,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1732596606016159,
"rewards/margins": 0.1732596606016159,
"rewards/rejected": 0.0,
"step": 56
},
{
"epoch": 0.31843575418994413,
"grad_norm": 24.87856002674328,
"learning_rate": 1.9860627177700347e-07,
"logits/chosen": -2.4646410942077637,
"logits/rejected": -2.167778491973877,
"logps/chosen": -24.157302856445312,
"logps/rejected": -48.043636322021484,
"loss": 0.6335,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1372363567352295,
"rewards/margins": 0.1372363567352295,
"rewards/rejected": 0.0,
"step": 57
},
{
"epoch": 0.3240223463687151,
"grad_norm": 24.632709992488646,
"learning_rate": 2.0209059233449478e-07,
"logits/chosen": -2.522116184234619,
"logits/rejected": -2.328913927078247,
"logps/chosen": -35.4097900390625,
"logps/rejected": -47.49260711669922,
"loss": 0.64,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.11406522244215012,
"rewards/margins": 0.11406522244215012,
"rewards/rejected": 0.0,
"step": 58
},
{
"epoch": 0.329608938547486,
"grad_norm": 26.555932940372706,
"learning_rate": 2.0557491289198606e-07,
"logits/chosen": -2.312364339828491,
"logits/rejected": -2.3537187576293945,
"logps/chosen": -58.54545593261719,
"logps/rejected": -63.535945892333984,
"loss": 0.6279,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.10076670348644257,
"rewards/margins": 0.10076670348644257,
"rewards/rejected": 0.0,
"step": 59
},
{
"epoch": 0.33519553072625696,
"grad_norm": 23.362895588849486,
"learning_rate": 2.0905923344947734e-07,
"logits/chosen": -2.3553497791290283,
"logits/rejected": -2.3863284587860107,
"logps/chosen": -27.702287673950195,
"logps/rejected": -33.98295211791992,
"loss": 0.6352,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.09926161170005798,
"rewards/margins": 0.09926161170005798,
"rewards/rejected": 0.0,
"step": 60
},
{
"epoch": 0.33519553072625696,
"eval_logits/chosen": -2.450570821762085,
"eval_logits/rejected": -2.313969850540161,
"eval_logps/chosen": -28.935455322265625,
"eval_logps/rejected": -40.358070373535156,
"eval_loss": 0.633419930934906,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 0.11345873773097992,
"eval_rewards/margins": 0.11345873773097992,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.9021,
"eval_samples_per_second": 9.144,
"eval_steps_per_second": 0.295,
"step": 60
},
{
"epoch": 0.3407821229050279,
"grad_norm": 23.278035097694193,
"learning_rate": 2.1254355400696864e-07,
"logits/chosen": -2.9201314449310303,
"logits/rejected": -2.522038221359253,
"logps/chosen": -22.127304077148438,
"logps/rejected": -34.22399139404297,
"loss": 0.6292,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1686132699251175,
"rewards/margins": 0.1686132699251175,
"rewards/rejected": 0.0,
"step": 61
},
{
"epoch": 0.3463687150837989,
"grad_norm": 24.907424656036728,
"learning_rate": 2.1602787456445992e-07,
"logits/chosen": -2.494243621826172,
"logits/rejected": -2.46530818939209,
"logps/chosen": -25.39417839050293,
"logps/rejected": -51.274375915527344,
"loss": 0.6286,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.12252993881702423,
"rewards/margins": 0.12252993881702423,
"rewards/rejected": 0.0,
"step": 62
},
{
"epoch": 0.35195530726256985,
"grad_norm": 23.084074118066138,
"learning_rate": 2.195121951219512e-07,
"logits/chosen": -2.5782864093780518,
"logits/rejected": -2.543471574783325,
"logps/chosen": -19.811384201049805,
"logps/rejected": -32.101165771484375,
"loss": 0.6353,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1308431625366211,
"rewards/margins": 0.1308431625366211,
"rewards/rejected": 0.0,
"step": 63
},
{
"epoch": 0.3575418994413408,
"grad_norm": 22.982956154022173,
"learning_rate": 2.229965156794425e-07,
"logits/chosen": -2.518967390060425,
"logits/rejected": -2.424222230911255,
"logps/chosen": -21.03902816772461,
"logps/rejected": -33.82353973388672,
"loss": 0.6248,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.11641013622283936,
"rewards/margins": 0.11641013622283936,
"rewards/rejected": 0.0,
"step": 64
},
{
"epoch": 0.36312849162011174,
"grad_norm": 22.587929792055263,
"learning_rate": 2.264808362369338e-07,
"logits/chosen": -2.4381473064422607,
"logits/rejected": -2.238905191421509,
"logps/chosen": -23.96572494506836,
"logps/rejected": -44.4195671081543,
"loss": 0.6319,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.10577362030744553,
"rewards/margins": 0.10577362030744553,
"rewards/rejected": 0.0,
"step": 65
},
{
"epoch": 0.3687150837988827,
"grad_norm": 24.445932086362088,
"learning_rate": 2.2996515679442507e-07,
"logits/chosen": -2.3218841552734375,
"logits/rejected": -1.4848906993865967,
"logps/chosen": -41.72569274902344,
"logps/rejected": -65.5929183959961,
"loss": 0.6254,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.17065949738025665,
"rewards/margins": 0.17065949738025665,
"rewards/rejected": 0.0,
"step": 66
},
{
"epoch": 0.3743016759776536,
"grad_norm": 23.886191675716017,
"learning_rate": 2.3344947735191635e-07,
"logits/chosen": -2.294743299484253,
"logits/rejected": -2.3369648456573486,
"logps/chosen": -36.33357620239258,
"logps/rejected": -31.636873245239258,
"loss": 0.6213,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.11421214044094086,
"rewards/margins": 0.11421214044094086,
"rewards/rejected": 0.0,
"step": 67
},
{
"epoch": 0.37988826815642457,
"grad_norm": 23.951577576185752,
"learning_rate": 2.3693379790940766e-07,
"logits/chosen": -2.117687463760376,
"logits/rejected": -2.0123000144958496,
"logps/chosen": -51.87607955932617,
"logps/rejected": -49.09307098388672,
"loss": 0.6109,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.21755558252334595,
"rewards/margins": 0.21755558252334595,
"rewards/rejected": 0.0,
"step": 68
},
{
"epoch": 0.3854748603351955,
"grad_norm": 22.7328773168266,
"learning_rate": 2.4041811846689894e-07,
"logits/chosen": -2.36024808883667,
"logits/rejected": -2.2724578380584717,
"logps/chosen": -22.848896026611328,
"logps/rejected": -43.429954528808594,
"loss": 0.6047,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.18426723778247833,
"rewards/margins": 0.18426723778247833,
"rewards/rejected": 0.0,
"step": 69
},
{
"epoch": 0.39106145251396646,
"grad_norm": 22.123233307547665,
"learning_rate": 2.439024390243902e-07,
"logits/chosen": -2.68394136428833,
"logits/rejected": -2.5219666957855225,
"logps/chosen": -25.577144622802734,
"logps/rejected": -36.436180114746094,
"loss": 0.5927,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.18651428818702698,
"rewards/margins": 0.18651428818702698,
"rewards/rejected": 0.0,
"step": 70
},
{
"epoch": 0.39664804469273746,
"grad_norm": 20.74545972228492,
"learning_rate": 2.473867595818815e-07,
"logits/chosen": -2.5408196449279785,
"logits/rejected": -2.420463800430298,
"logps/chosen": -24.5079288482666,
"logps/rejected": -39.61929702758789,
"loss": 0.5961,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.17850160598754883,
"rewards/margins": 0.17850160598754883,
"rewards/rejected": 0.0,
"step": 71
},
{
"epoch": 0.4022346368715084,
"grad_norm": 22.381070649511372,
"learning_rate": 2.508710801393728e-07,
"logits/chosen": -2.9514803886413574,
"logits/rejected": -2.8442459106445312,
"logps/chosen": -19.637069702148438,
"logps/rejected": -32.112022399902344,
"loss": 0.5775,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.24960695207118988,
"rewards/margins": 0.24960695207118988,
"rewards/rejected": 0.0,
"step": 72
},
{
"epoch": 0.40782122905027934,
"grad_norm": 19.79207116570799,
"learning_rate": 2.543554006968641e-07,
"logits/chosen": -2.584350109100342,
"logits/rejected": -2.4062235355377197,
"logps/chosen": -13.208507537841797,
"logps/rejected": -39.940975189208984,
"loss": 0.5829,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.1439269483089447,
"rewards/margins": 0.1439269483089447,
"rewards/rejected": 0.0,
"step": 73
},
{
"epoch": 0.4134078212290503,
"grad_norm": 17.54968620695586,
"learning_rate": 2.578397212543554e-07,
"logits/chosen": -2.2153613567352295,
"logits/rejected": -2.2630043029785156,
"logps/chosen": -33.766082763671875,
"logps/rejected": -40.740177154541016,
"loss": 0.5786,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.20816154778003693,
"rewards/margins": 0.20816154778003693,
"rewards/rejected": 0.0,
"step": 74
},
{
"epoch": 0.41899441340782123,
"grad_norm": 18.12325013292998,
"learning_rate": 2.613240418118467e-07,
"logits/chosen": -2.317336082458496,
"logits/rejected": -2.2485673427581787,
"logps/chosen": -31.588523864746094,
"logps/rejected": -34.93479537963867,
"loss": 0.5801,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2886280119419098,
"rewards/margins": 0.2886280119419098,
"rewards/rejected": 0.0,
"step": 75
},
{
"epoch": 0.4245810055865922,
"grad_norm": 17.81810199398241,
"learning_rate": 2.6480836236933796e-07,
"logits/chosen": -2.9728236198425293,
"logits/rejected": -2.7282276153564453,
"logps/chosen": -21.17477035522461,
"logps/rejected": -34.339134216308594,
"loss": 0.5673,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.34478840231895447,
"rewards/margins": 0.34478840231895447,
"rewards/rejected": 0.0,
"step": 76
},
{
"epoch": 0.4301675977653631,
"grad_norm": 18.57560846576237,
"learning_rate": 2.682926829268293e-07,
"logits/chosen": -2.258829355239868,
"logits/rejected": -2.3700168132781982,
"logps/chosen": -40.9664192199707,
"logps/rejected": -43.83318328857422,
"loss": 0.5462,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.47076869010925293,
"rewards/margins": 0.47076869010925293,
"rewards/rejected": 0.0,
"step": 77
},
{
"epoch": 0.43575418994413406,
"grad_norm": 20.06020648655684,
"learning_rate": 2.717770034843205e-07,
"logits/chosen": -2.5632386207580566,
"logits/rejected": -2.5703723430633545,
"logps/chosen": -29.538589477539062,
"logps/rejected": -38.035972595214844,
"loss": 0.5599,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.33969616889953613,
"rewards/margins": 0.33969616889953613,
"rewards/rejected": 0.0,
"step": 78
},
{
"epoch": 0.441340782122905,
"grad_norm": 17.522456596520726,
"learning_rate": 2.7526132404181185e-07,
"logits/chosen": -2.571640729904175,
"logits/rejected": -2.5123398303985596,
"logps/chosen": -23.64766502380371,
"logps/rejected": -38.227935791015625,
"loss": 0.5603,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.30658215284347534,
"rewards/margins": 0.30658215284347534,
"rewards/rejected": 0.0,
"step": 79
},
{
"epoch": 0.44692737430167595,
"grad_norm": 17.609953811055686,
"learning_rate": 2.7874564459930313e-07,
"logits/chosen": -2.55928897857666,
"logits/rejected": -2.609794855117798,
"logps/chosen": -29.678600311279297,
"logps/rejected": -34.138824462890625,
"loss": 0.5527,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2765759229660034,
"rewards/margins": 0.2765759229660034,
"rewards/rejected": 0.0,
"step": 80
},
{
"epoch": 0.44692737430167595,
"eval_logits/chosen": -2.5961174964904785,
"eval_logits/rejected": -2.481539011001587,
"eval_logps/chosen": -26.89170265197754,
"eval_logps/rejected": -38.134918212890625,
"eval_loss": 0.5455653071403503,
"eval_rewards/accuracies": 1.0,
"eval_rewards/chosen": 0.3178338408470154,
"eval_rewards/margins": 0.3178338408470154,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5365,
"eval_samples_per_second": 9.244,
"eval_steps_per_second": 0.298,
"step": 80
},
{
"epoch": 0.45251396648044695,
"grad_norm": 17.243619602321587,
"learning_rate": 2.822299651567944e-07,
"logits/chosen": -2.766360282897949,
"logits/rejected": -2.7481331825256348,
"logps/chosen": -34.953697204589844,
"logps/rejected": -44.08012008666992,
"loss": 0.5556,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.386406809091568,
"rewards/margins": 0.386406809091568,
"rewards/rejected": 0.0,
"step": 81
},
{
"epoch": 0.4581005586592179,
"grad_norm": 18.12598041761261,
"learning_rate": 2.857142857142857e-07,
"logits/chosen": -2.5122439861297607,
"logits/rejected": -2.2476606369018555,
"logps/chosen": -38.5035285949707,
"logps/rejected": -58.14997482299805,
"loss": 0.5337,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6213464736938477,
"rewards/margins": 0.6213464736938477,
"rewards/rejected": 0.0,
"step": 82
},
{
"epoch": 0.46368715083798884,
"grad_norm": 18.26018052160107,
"learning_rate": 2.8919860627177703e-07,
"logits/chosen": -3.008633613586426,
"logits/rejected": -2.7243096828460693,
"logps/chosen": -23.37417984008789,
"logps/rejected": -25.58895492553711,
"loss": 0.5336,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.45575347542762756,
"rewards/margins": 0.45575347542762756,
"rewards/rejected": 0.0,
"step": 83
},
{
"epoch": 0.4692737430167598,
"grad_norm": 17.95021613887715,
"learning_rate": 2.9268292682926825e-07,
"logits/chosen": -2.8492836952209473,
"logits/rejected": -2.724342107772827,
"logps/chosen": -22.531551361083984,
"logps/rejected": -83.26819610595703,
"loss": 0.5373,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.4083956778049469,
"rewards/margins": 0.4083956778049469,
"rewards/rejected": 0.0,
"step": 84
},
{
"epoch": 0.4748603351955307,
"grad_norm": 19.55540877246056,
"learning_rate": 2.961672473867596e-07,
"logits/chosen": -2.547548770904541,
"logits/rejected": -2.4949235916137695,
"logps/chosen": -39.774715423583984,
"logps/rejected": -45.983741760253906,
"loss": 0.519,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.38765764236450195,
"rewards/margins": 0.38765764236450195,
"rewards/rejected": 0.0,
"step": 85
},
{
"epoch": 0.48044692737430167,
"grad_norm": 17.44629190598895,
"learning_rate": 2.996515679442508e-07,
"logits/chosen": -2.9293198585510254,
"logits/rejected": -2.554246187210083,
"logps/chosen": -15.574745178222656,
"logps/rejected": -54.118019104003906,
"loss": 0.5206,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.3024306893348694,
"rewards/margins": 0.3024306893348694,
"rewards/rejected": 0.0,
"step": 86
},
{
"epoch": 0.4860335195530726,
"grad_norm": 17.640304319217552,
"learning_rate": 3.0313588850174215e-07,
"logits/chosen": -2.509176015853882,
"logits/rejected": -2.6400609016418457,
"logps/chosen": -26.335580825805664,
"logps/rejected": -23.034029006958008,
"loss": 0.5128,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.2293638437986374,
"rewards/margins": 0.2293638437986374,
"rewards/rejected": 0.0,
"step": 87
},
{
"epoch": 0.49162011173184356,
"grad_norm": 17.076730441032804,
"learning_rate": 3.0662020905923343e-07,
"logits/chosen": -2.748363733291626,
"logits/rejected": -2.727149248123169,
"logps/chosen": -39.85935974121094,
"logps/rejected": -41.72587203979492,
"loss": 0.5257,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.3664461374282837,
"rewards/margins": 0.3664461374282837,
"rewards/rejected": 0.0,
"step": 88
},
{
"epoch": 0.4972067039106145,
"grad_norm": 16.7596222624691,
"learning_rate": 3.101045296167247e-07,
"logits/chosen": -2.7122080326080322,
"logits/rejected": -2.6718027591705322,
"logps/chosen": -21.423099517822266,
"logps/rejected": -39.17164611816406,
"loss": 0.5254,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.3437221646308899,
"rewards/margins": 0.3437221646308899,
"rewards/rejected": 0.0,
"step": 89
},
{
"epoch": 0.5027932960893855,
"grad_norm": 16.370624963433137,
"learning_rate": 3.13588850174216e-07,
"logits/chosen": -2.7640321254730225,
"logits/rejected": -2.6971404552459717,
"logps/chosen": -27.607778549194336,
"logps/rejected": -29.204383850097656,
"loss": 0.5203,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.34216684103012085,
"rewards/margins": 0.34216684103012085,
"rewards/rejected": 0.0,
"step": 90
},
{
"epoch": 0.5083798882681564,
"grad_norm": 16.501196308487856,
"learning_rate": 3.170731707317073e-07,
"logits/chosen": -2.977828025817871,
"logits/rejected": -2.7632179260253906,
"logps/chosen": -25.994243621826172,
"logps/rejected": -50.56151580810547,
"loss": 0.5309,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.41055914759635925,
"rewards/margins": 0.41055914759635925,
"rewards/rejected": 0.0,
"step": 91
},
{
"epoch": 0.5139664804469274,
"grad_norm": 16.668279912747,
"learning_rate": 3.2055749128919855e-07,
"logits/chosen": -2.616750955581665,
"logits/rejected": -2.6325173377990723,
"logps/chosen": -27.63555908203125,
"logps/rejected": -26.495452880859375,
"loss": 0.5155,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.44802695512771606,
"rewards/margins": 0.44802695512771606,
"rewards/rejected": 0.0,
"step": 92
},
{
"epoch": 0.5195530726256983,
"grad_norm": 16.585791801396,
"learning_rate": 3.240418118466899e-07,
"logits/chosen": -2.691551446914673,
"logits/rejected": -2.4979159832000732,
"logps/chosen": -26.484764099121094,
"logps/rejected": -34.90871810913086,
"loss": 0.5193,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.4211455285549164,
"rewards/margins": 0.4211455285549164,
"rewards/rejected": 0.0,
"step": 93
},
{
"epoch": 0.5251396648044693,
"grad_norm": 15.508507664030144,
"learning_rate": 3.2752613240418117e-07,
"logits/chosen": -2.3586292266845703,
"logits/rejected": -2.530395746231079,
"logps/chosen": -26.607080459594727,
"logps/rejected": -29.511329650878906,
"loss": 0.5221,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.381470650434494,
"rewards/margins": 0.381470650434494,
"rewards/rejected": 0.0,
"step": 94
},
{
"epoch": 0.5307262569832403,
"grad_norm": 14.73280514705423,
"learning_rate": 3.3101045296167245e-07,
"logits/chosen": -2.9987075328826904,
"logits/rejected": -2.319053888320923,
"logps/chosen": -15.4478178024292,
"logps/rejected": -50.605716705322266,
"loss": 0.5068,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.46311700344085693,
"rewards/margins": 0.46311700344085693,
"rewards/rejected": 0.0,
"step": 95
},
{
"epoch": 0.5363128491620112,
"grad_norm": 15.464939165348971,
"learning_rate": 3.344947735191638e-07,
"logits/chosen": -3.0241072177886963,
"logits/rejected": -2.82523775100708,
"logps/chosen": -15.733403205871582,
"logps/rejected": -23.26158332824707,
"loss": 0.4811,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.3380926251411438,
"rewards/margins": 0.3380926251411438,
"rewards/rejected": 0.0,
"step": 96
},
{
"epoch": 0.5418994413407822,
"grad_norm": 15.07425710254498,
"learning_rate": 3.3797909407665506e-07,
"logits/chosen": -2.829176425933838,
"logits/rejected": -2.7734498977661133,
"logps/chosen": -26.00864028930664,
"logps/rejected": -61.76011276245117,
"loss": 0.4938,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.47422951459884644,
"rewards/margins": 0.47422951459884644,
"rewards/rejected": 0.0,
"step": 97
},
{
"epoch": 0.547486033519553,
"grad_norm": 15.292202698963973,
"learning_rate": 3.4146341463414634e-07,
"logits/chosen": -3.094583749771118,
"logits/rejected": -2.988278388977051,
"logps/chosen": -15.285747528076172,
"logps/rejected": -23.24346923828125,
"loss": 0.4904,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.3607971668243408,
"rewards/margins": 0.3607971668243408,
"rewards/rejected": 0.0,
"step": 98
},
{
"epoch": 0.553072625698324,
"grad_norm": 14.280871491578443,
"learning_rate": 3.449477351916376e-07,
"logits/chosen": -3.0072247982025146,
"logits/rejected": -2.7647087574005127,
"logps/chosen": -17.79582405090332,
"logps/rejected": -30.290449142456055,
"loss": 0.4646,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.43808504939079285,
"rewards/margins": 0.43808504939079285,
"rewards/rejected": 0.0,
"step": 99
},
{
"epoch": 0.5586592178770949,
"grad_norm": 16.045439012707615,
"learning_rate": 3.484320557491289e-07,
"logits/chosen": -2.7985477447509766,
"logits/rejected": -2.6149985790252686,
"logps/chosen": -41.18121337890625,
"logps/rejected": -41.32271194458008,
"loss": 0.4155,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6721265316009521,
"rewards/margins": 0.6721265316009521,
"rewards/rejected": 0.0,
"step": 100
},
{
"epoch": 0.5586592178770949,
"eval_logits/chosen": -2.7876787185668945,
"eval_logits/rejected": -2.70434308052063,
"eval_logps/chosen": -24.622821807861328,
"eval_logps/rejected": -35.58250045776367,
"eval_loss": 0.4617948532104492,
"eval_rewards/accuracies": 1.0,
"eval_rewards/chosen": 0.5447221994400024,
"eval_rewards/margins": 0.5447221994400024,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5634,
"eval_samples_per_second": 9.236,
"eval_steps_per_second": 0.298,
"step": 100
},
{
"epoch": 0.5642458100558659,
"grad_norm": 12.573137728140779,
"learning_rate": 3.519163763066202e-07,
"logits/chosen": -2.688979387283325,
"logits/rejected": -2.658362865447998,
"logps/chosen": -16.801599502563477,
"logps/rejected": -17.431743621826172,
"loss": 0.4702,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.346937894821167,
"rewards/margins": 0.346937894821167,
"rewards/rejected": 0.0,
"step": 101
},
{
"epoch": 0.5698324022346368,
"grad_norm": 13.331911966929777,
"learning_rate": 3.554006968641115e-07,
"logits/chosen": -2.7960867881774902,
"logits/rejected": -2.776906728744507,
"logps/chosen": -37.72350311279297,
"logps/rejected": -41.60508728027344,
"loss": 0.4448,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7360150814056396,
"rewards/margins": 0.7360150814056396,
"rewards/rejected": 0.0,
"step": 102
},
{
"epoch": 0.5754189944134078,
"grad_norm": 13.730862772517419,
"learning_rate": 3.5888501742160274e-07,
"logits/chosen": -3.160386562347412,
"logits/rejected": -3.0882933139801025,
"logps/chosen": -14.691383361816406,
"logps/rejected": -32.46586608886719,
"loss": 0.4682,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.44620683789253235,
"rewards/margins": 0.44620683789253235,
"rewards/rejected": 0.0,
"step": 103
},
{
"epoch": 0.5810055865921788,
"grad_norm": 12.958763580555999,
"learning_rate": 3.623693379790941e-07,
"logits/chosen": -2.806156873703003,
"logits/rejected": -2.7647111415863037,
"logps/chosen": -26.239046096801758,
"logps/rejected": -27.40913963317871,
"loss": 0.4521,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6825430393218994,
"rewards/margins": 0.6825430393218994,
"rewards/rejected": 0.0,
"step": 104
},
{
"epoch": 0.5865921787709497,
"grad_norm": 13.360824364314807,
"learning_rate": 3.6585365853658536e-07,
"logits/chosen": -2.783945083618164,
"logits/rejected": -2.7649834156036377,
"logps/chosen": -23.880428314208984,
"logps/rejected": -24.65614128112793,
"loss": 0.4493,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5743331909179688,
"rewards/margins": 0.5743331909179688,
"rewards/rejected": 0.0,
"step": 105
},
{
"epoch": 0.5921787709497207,
"grad_norm": 13.104227770358209,
"learning_rate": 3.6933797909407664e-07,
"logits/chosen": -3.279587507247925,
"logits/rejected": -3.1188621520996094,
"logps/chosen": -18.71685791015625,
"logps/rejected": -33.31108474731445,
"loss": 0.437,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6923595070838928,
"rewards/margins": 0.6923595070838928,
"rewards/rejected": 0.0,
"step": 106
},
{
"epoch": 0.5977653631284916,
"grad_norm": 14.148746367362307,
"learning_rate": 3.728222996515679e-07,
"logits/chosen": -3.1686055660247803,
"logits/rejected": -3.0254077911376953,
"logps/chosen": -20.358896255493164,
"logps/rejected": -36.1294059753418,
"loss": 0.4331,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6534121036529541,
"rewards/margins": 0.6534121036529541,
"rewards/rejected": 0.0,
"step": 107
},
{
"epoch": 0.6033519553072626,
"grad_norm": 11.985725221178399,
"learning_rate": 3.7630662020905925e-07,
"logits/chosen": -2.710590124130249,
"logits/rejected": -2.8439455032348633,
"logps/chosen": -40.90412139892578,
"logps/rejected": -32.95746994018555,
"loss": 0.4134,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.029143214225769,
"rewards/margins": 1.029143214225769,
"rewards/rejected": 0.0,
"step": 108
},
{
"epoch": 0.6089385474860335,
"grad_norm": 13.248918349041723,
"learning_rate": 3.797909407665505e-07,
"logits/chosen": -2.8310110569000244,
"logits/rejected": -2.5916903018951416,
"logps/chosen": -28.208520889282227,
"logps/rejected": -52.245033264160156,
"loss": 0.4174,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.562522292137146,
"rewards/margins": 0.562522292137146,
"rewards/rejected": 0.0,
"step": 109
},
{
"epoch": 0.6145251396648045,
"grad_norm": 12.5998803430947,
"learning_rate": 3.832752613240418e-07,
"logits/chosen": -3.032209634780884,
"logits/rejected": -3.0512664318084717,
"logps/chosen": -19.49001693725586,
"logps/rejected": -72.57357788085938,
"loss": 0.4215,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7006881237030029,
"rewards/margins": 0.7006881237030029,
"rewards/rejected": 0.0,
"step": 110
},
{
"epoch": 0.6201117318435754,
"grad_norm": 13.65187382845731,
"learning_rate": 3.867595818815331e-07,
"logits/chosen": -2.8330180644989014,
"logits/rejected": -2.093581199645996,
"logps/chosen": -25.720317840576172,
"logps/rejected": -62.531890869140625,
"loss": 0.431,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6129947900772095,
"rewards/margins": 0.6129947900772095,
"rewards/rejected": 0.0,
"step": 111
},
{
"epoch": 0.6256983240223464,
"grad_norm": 13.466950507028253,
"learning_rate": 3.902439024390244e-07,
"logits/chosen": -3.234790325164795,
"logits/rejected": -3.256101369857788,
"logps/chosen": -21.19486427307129,
"logps/rejected": -40.322139739990234,
"loss": 0.451,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.36941981315612793,
"rewards/margins": 0.36941981315612793,
"rewards/rejected": 0.0,
"step": 112
},
{
"epoch": 0.6312849162011173,
"grad_norm": 13.653645461327002,
"learning_rate": 3.9372822299651566e-07,
"logits/chosen": -3.116182804107666,
"logits/rejected": -2.931743860244751,
"logps/chosen": -37.29005813598633,
"logps/rejected": -53.360008239746094,
"loss": 0.45,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6139600872993469,
"rewards/margins": 0.6139600872993469,
"rewards/rejected": 0.0,
"step": 113
},
{
"epoch": 0.6368715083798883,
"grad_norm": 12.366194224921594,
"learning_rate": 3.9721254355400694e-07,
"logits/chosen": -3.1295180320739746,
"logits/rejected": -3.1512725353240967,
"logps/chosen": -22.38964080810547,
"logps/rejected": -30.870153427124023,
"loss": 0.4218,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8451684713363647,
"rewards/margins": 0.8451684713363647,
"rewards/rejected": 0.0,
"step": 114
},
{
"epoch": 0.6424581005586593,
"grad_norm": 12.820270690808119,
"learning_rate": 4.006968641114982e-07,
"logits/chosen": -2.5849902629852295,
"logits/rejected": -2.5268383026123047,
"logps/chosen": -39.322906494140625,
"logps/rejected": -31.824663162231445,
"loss": 0.435,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6132646799087524,
"rewards/margins": 0.6132646799087524,
"rewards/rejected": 0.0,
"step": 115
},
{
"epoch": 0.6480446927374302,
"grad_norm": 12.544434650205087,
"learning_rate": 4.0418118466898955e-07,
"logits/chosen": -3.1694047451019287,
"logits/rejected": -3.014833688735962,
"logps/chosen": -32.03730010986328,
"logps/rejected": -36.730979919433594,
"loss": 0.407,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9823371171951294,
"rewards/margins": 0.9823371171951294,
"rewards/rejected": 0.0,
"step": 116
},
{
"epoch": 0.6536312849162011,
"grad_norm": 12.963061338666149,
"learning_rate": 4.076655052264808e-07,
"logits/chosen": -3.153256893157959,
"logits/rejected": -3.0860118865966797,
"logps/chosen": -28.097692489624023,
"logps/rejected": -47.145748138427734,
"loss": 0.3859,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9587994813919067,
"rewards/margins": 0.9587994813919067,
"rewards/rejected": 0.0,
"step": 117
},
{
"epoch": 0.659217877094972,
"grad_norm": 13.019617331540443,
"learning_rate": 4.111498257839721e-07,
"logits/chosen": -3.1117141246795654,
"logits/rejected": -2.981959581375122,
"logps/chosen": -29.46117401123047,
"logps/rejected": -54.425132751464844,
"loss": 0.3843,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9660426378250122,
"rewards/margins": 0.9660426378250122,
"rewards/rejected": 0.0,
"step": 118
},
{
"epoch": 0.664804469273743,
"grad_norm": 12.09790313144843,
"learning_rate": 4.146341463414634e-07,
"logits/chosen": -3.2568607330322266,
"logits/rejected": -3.243248701095581,
"logps/chosen": -35.24617004394531,
"logps/rejected": -28.864641189575195,
"loss": 0.4081,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5788088440895081,
"rewards/margins": 0.5788088440895081,
"rewards/rejected": 0.0,
"step": 119
},
{
"epoch": 0.6703910614525139,
"grad_norm": 12.299789893471367,
"learning_rate": 4.1811846689895467e-07,
"logits/chosen": -3.4779090881347656,
"logits/rejected": -3.5990796089172363,
"logps/chosen": -27.420997619628906,
"logps/rejected": -29.504467010498047,
"loss": 0.4148,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7487576603889465,
"rewards/margins": 0.7487576603889465,
"rewards/rejected": 0.0,
"step": 120
},
{
"epoch": 0.6703910614525139,
"eval_logits/chosen": -2.985457420349121,
"eval_logits/rejected": -2.9356820583343506,
"eval_logps/chosen": -22.773021697998047,
"eval_logps/rejected": -33.96015548706055,
"eval_loss": 0.41244789958000183,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 0.7297021746635437,
"eval_rewards/margins": 0.7297021746635437,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5893,
"eval_samples_per_second": 9.229,
"eval_steps_per_second": 0.298,
"step": 120
},
{
"epoch": 0.6759776536312849,
"grad_norm": 11.918901658967625,
"learning_rate": 4.2160278745644595e-07,
"logits/chosen": -3.361621618270874,
"logits/rejected": -3.4122440814971924,
"logps/chosen": -24.095653533935547,
"logps/rejected": -39.066993713378906,
"loss": 0.4248,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7933073043823242,
"rewards/margins": 0.7933073043823242,
"rewards/rejected": 0.0,
"step": 121
},
{
"epoch": 0.6815642458100558,
"grad_norm": 12.09693560581,
"learning_rate": 4.250871080139373e-07,
"logits/chosen": -3.1030969619750977,
"logits/rejected": -3.078477382659912,
"logps/chosen": -31.073591232299805,
"logps/rejected": -18.581279754638672,
"loss": 0.4445,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.20919391512870789,
"rewards/margins": 0.20919391512870789,
"rewards/rejected": 0.0,
"step": 122
},
{
"epoch": 0.6871508379888268,
"grad_norm": 13.230174521061821,
"learning_rate": 4.285714285714285e-07,
"logits/chosen": -2.724034547805786,
"logits/rejected": -2.712747573852539,
"logps/chosen": -38.76531982421875,
"logps/rejected": -40.227027893066406,
"loss": 0.4477,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.081432819366455,
"rewards/margins": 1.081432819366455,
"rewards/rejected": 0.0,
"step": 123
},
{
"epoch": 0.6927374301675978,
"grad_norm": 11.029542893807358,
"learning_rate": 4.3205574912891985e-07,
"logits/chosen": -3.2009103298187256,
"logits/rejected": -3.1295859813690186,
"logps/chosen": -17.012771606445312,
"logps/rejected": -38.768104553222656,
"loss": 0.3632,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8287367820739746,
"rewards/margins": 0.8287367820739746,
"rewards/rejected": 0.0,
"step": 124
},
{
"epoch": 0.6983240223463687,
"grad_norm": 12.92541684214979,
"learning_rate": 4.3554006968641113e-07,
"logits/chosen": -3.3594982624053955,
"logits/rejected": -3.184117317199707,
"logps/chosen": -15.785029411315918,
"logps/rejected": -44.183937072753906,
"loss": 0.4085,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5601776242256165,
"rewards/margins": 0.5601776242256165,
"rewards/rejected": 0.0,
"step": 125
},
{
"epoch": 0.7039106145251397,
"grad_norm": 12.418840197192134,
"learning_rate": 4.390243902439024e-07,
"logits/chosen": -3.018588066101074,
"logits/rejected": -2.9847147464752197,
"logps/chosen": -16.75999641418457,
"logps/rejected": -46.10063934326172,
"loss": 0.3922,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6556053161621094,
"rewards/margins": 0.6556053161621094,
"rewards/rejected": 0.0,
"step": 126
},
{
"epoch": 0.7094972067039106,
"grad_norm": 10.889083863099927,
"learning_rate": 4.425087108013937e-07,
"logits/chosen": -3.1690027713775635,
"logits/rejected": -3.025820016860962,
"logps/chosen": -35.54032516479492,
"logps/rejected": -34.36103820800781,
"loss": 0.3746,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0566309690475464,
"rewards/margins": 1.0566309690475464,
"rewards/rejected": 0.0,
"step": 127
},
{
"epoch": 0.7150837988826816,
"grad_norm": 12.906782926741812,
"learning_rate": 4.45993031358885e-07,
"logits/chosen": -3.336862087249756,
"logits/rejected": -3.1122004985809326,
"logps/chosen": -16.735675811767578,
"logps/rejected": -49.08568572998047,
"loss": 0.4002,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7246649861335754,
"rewards/margins": 0.7246649861335754,
"rewards/rejected": 0.0,
"step": 128
},
{
"epoch": 0.7206703910614525,
"grad_norm": 11.359331301126964,
"learning_rate": 4.494773519163763e-07,
"logits/chosen": -3.3350188732147217,
"logits/rejected": -3.3323018550872803,
"logps/chosen": -20.424598693847656,
"logps/rejected": -26.444061279296875,
"loss": 0.4034,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9043171405792236,
"rewards/margins": 0.9043171405792236,
"rewards/rejected": 0.0,
"step": 129
},
{
"epoch": 0.7262569832402235,
"grad_norm": 11.530762525487518,
"learning_rate": 4.529616724738676e-07,
"logits/chosen": -3.4406161308288574,
"logits/rejected": -3.4307963848114014,
"logps/chosen": -30.73476791381836,
"logps/rejected": -31.74262809753418,
"loss": 0.385,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6756500005722046,
"rewards/margins": 0.6756500005722046,
"rewards/rejected": 0.0,
"step": 130
},
{
"epoch": 0.7318435754189944,
"grad_norm": 15.939552333565983,
"learning_rate": 4.5644599303135886e-07,
"logits/chosen": -3.0016088485717773,
"logits/rejected": -3.0341885089874268,
"logps/chosen": -19.028980255126953,
"logps/rejected": -35.876129150390625,
"loss": 0.4286,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7295345067977905,
"rewards/margins": 0.7295345067977905,
"rewards/rejected": 0.0,
"step": 131
},
{
"epoch": 0.7374301675977654,
"grad_norm": 12.180173279607684,
"learning_rate": 4.5993031358885015e-07,
"logits/chosen": -3.136824369430542,
"logits/rejected": -2.9175052642822266,
"logps/chosen": -22.140581130981445,
"logps/rejected": -34.302146911621094,
"loss": 0.4175,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5968088507652283,
"rewards/margins": 0.5968088507652283,
"rewards/rejected": 0.0,
"step": 132
},
{
"epoch": 0.7430167597765364,
"grad_norm": 13.82002718598598,
"learning_rate": 4.634146341463415e-07,
"logits/chosen": -2.8966777324676514,
"logits/rejected": -2.788705587387085,
"logps/chosen": -30.218425750732422,
"logps/rejected": -40.37165832519531,
"loss": 0.4183,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0006399154663086,
"rewards/margins": 1.0006399154663086,
"rewards/rejected": 0.0,
"step": 133
},
{
"epoch": 0.7486033519553073,
"grad_norm": 11.279996388030757,
"learning_rate": 4.668989547038327e-07,
"logits/chosen": -3.263176441192627,
"logits/rejected": -3.253405809402466,
"logps/chosen": -22.389724731445312,
"logps/rejected": -28.712936401367188,
"loss": 0.3617,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8534319996833801,
"rewards/margins": 0.8534319996833801,
"rewards/rejected": 0.0,
"step": 134
},
{
"epoch": 0.7541899441340782,
"grad_norm": 13.31664654210755,
"learning_rate": 4.7038327526132404e-07,
"logits/chosen": -3.196460008621216,
"logits/rejected": -3.055830478668213,
"logps/chosen": -13.232107162475586,
"logps/rejected": -26.673044204711914,
"loss": 0.3906,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5391906499862671,
"rewards/margins": 0.5391906499862671,
"rewards/rejected": 0.0,
"step": 135
},
{
"epoch": 0.7597765363128491,
"grad_norm": 13.644525356031862,
"learning_rate": 4.738675958188153e-07,
"logits/chosen": -3.086418867111206,
"logits/rejected": -3.050128221511841,
"logps/chosen": -63.31258773803711,
"logps/rejected": -45.431800842285156,
"loss": 0.4112,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6270837783813477,
"rewards/margins": 0.6270837783813477,
"rewards/rejected": 0.0,
"step": 136
},
{
"epoch": 0.7653631284916201,
"grad_norm": 11.75120264957957,
"learning_rate": 4.773519163763066e-07,
"logits/chosen": -3.3567633628845215,
"logits/rejected": -3.200810432434082,
"logps/chosen": -23.672664642333984,
"logps/rejected": -31.51905059814453,
"loss": 0.4161,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5618863105773926,
"rewards/margins": 0.5618863105773926,
"rewards/rejected": 0.0,
"step": 137
},
{
"epoch": 0.770949720670391,
"grad_norm": 12.836460989737148,
"learning_rate": 4.808362369337979e-07,
"logits/chosen": -3.219993829727173,
"logits/rejected": -3.0484225749969482,
"logps/chosen": -51.30001449584961,
"logps/rejected": -43.791709899902344,
"loss": 0.3754,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.059951901435852,
"rewards/margins": 1.059951901435852,
"rewards/rejected": 0.0,
"step": 138
},
{
"epoch": 0.776536312849162,
"grad_norm": 13.634756786440422,
"learning_rate": 4.843205574912892e-07,
"logits/chosen": -2.9743027687072754,
"logits/rejected": -2.8742432594299316,
"logps/chosen": -31.341638565063477,
"logps/rejected": -30.591285705566406,
"loss": 0.3963,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0887293815612793,
"rewards/margins": 1.0887293815612793,
"rewards/rejected": 0.0,
"step": 139
},
{
"epoch": 0.7821229050279329,
"grad_norm": 13.257090900452434,
"learning_rate": 4.878048780487804e-07,
"logits/chosen": -3.090188503265381,
"logits/rejected": -2.824709415435791,
"logps/chosen": -37.55943298339844,
"logps/rejected": -23.54281234741211,
"loss": 0.4048,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.8553612232208252,
"rewards/margins": 0.8553612232208252,
"rewards/rejected": 0.0,
"step": 140
},
{
"epoch": 0.7821229050279329,
"eval_logits/chosen": -3.085087299346924,
"eval_logits/rejected": -3.0553345680236816,
"eval_logps/chosen": -21.801071166992188,
"eval_logps/rejected": -33.04508972167969,
"eval_loss": 0.3879704773426056,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 0.8268970251083374,
"eval_rewards/margins": 0.8268970251083374,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6022,
"eval_samples_per_second": 9.226,
"eval_steps_per_second": 0.298,
"step": 140
},
{
"epoch": 0.7877094972067039,
"grad_norm": 13.170388084647799,
"learning_rate": 4.912891986062717e-07,
"logits/chosen": -3.280094861984253,
"logits/rejected": -3.313344955444336,
"logps/chosen": -16.471755981445312,
"logps/rejected": -47.39679718017578,
"loss": 0.4417,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6201029419898987,
"rewards/margins": 0.6201029419898987,
"rewards/rejected": 0.0,
"step": 141
},
{
"epoch": 0.7932960893854749,
"grad_norm": 12.471190495621682,
"learning_rate": 4.94773519163763e-07,
"logits/chosen": -3.4723424911499023,
"logits/rejected": -3.2591540813446045,
"logps/chosen": -12.425056457519531,
"logps/rejected": -45.91838836669922,
"loss": 0.3801,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.765860378742218,
"rewards/margins": 0.765860378742218,
"rewards/rejected": 0.0,
"step": 142
},
{
"epoch": 0.7988826815642458,
"grad_norm": 11.753324312801528,
"learning_rate": 4.982578397212544e-07,
"logits/chosen": -3.0344104766845703,
"logits/rejected": -3.033134698867798,
"logps/chosen": -27.046661376953125,
"logps/rejected": -30.028705596923828,
"loss": 0.352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2494921684265137,
"rewards/margins": 1.2494921684265137,
"rewards/rejected": 0.0,
"step": 143
},
{
"epoch": 0.8044692737430168,
"grad_norm": 12.753673627655559,
"learning_rate": 5.017421602787456e-07,
"logits/chosen": -3.360502004623413,
"logits/rejected": -3.2027170658111572,
"logps/chosen": -18.972614288330078,
"logps/rejected": -27.654964447021484,
"loss": 0.3963,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8680293560028076,
"rewards/margins": 0.8680293560028076,
"rewards/rejected": 0.0,
"step": 144
},
{
"epoch": 0.8100558659217877,
"grad_norm": 14.327923210822185,
"learning_rate": 5.052264808362368e-07,
"logits/chosen": -3.0389859676361084,
"logits/rejected": -3.176422595977783,
"logps/chosen": -27.511253356933594,
"logps/rejected": -49.270713806152344,
"loss": 0.3575,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1916924715042114,
"rewards/margins": 1.1916924715042114,
"rewards/rejected": 0.0,
"step": 145
},
{
"epoch": 0.8156424581005587,
"grad_norm": 11.956878046826485,
"learning_rate": 5.087108013937282e-07,
"logits/chosen": -3.2901556491851807,
"logits/rejected": -3.1521785259246826,
"logps/chosen": -21.122501373291016,
"logps/rejected": -32.07401657104492,
"loss": 0.3995,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6621261835098267,
"rewards/margins": 0.6621261835098267,
"rewards/rejected": 0.0,
"step": 146
},
{
"epoch": 0.8212290502793296,
"grad_norm": 13.441070658644444,
"learning_rate": 5.121951219512195e-07,
"logits/chosen": -3.182860851287842,
"logits/rejected": -3.0047945976257324,
"logps/chosen": -25.162214279174805,
"logps/rejected": -86.56019592285156,
"loss": 0.3656,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7732621431350708,
"rewards/margins": 0.7732621431350708,
"rewards/rejected": 0.0,
"step": 147
},
{
"epoch": 0.8268156424581006,
"grad_norm": 12.43524346477179,
"learning_rate": 5.156794425087108e-07,
"logits/chosen": -3.2569799423217773,
"logits/rejected": -2.9275951385498047,
"logps/chosen": -18.716459274291992,
"logps/rejected": -23.8824462890625,
"loss": 0.3594,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7202714681625366,
"rewards/margins": 0.7202714681625366,
"rewards/rejected": 0.0,
"step": 148
},
{
"epoch": 0.8324022346368715,
"grad_norm": 10.290562092635898,
"learning_rate": 5.191637630662021e-07,
"logits/chosen": -3.4133076667785645,
"logits/rejected": -3.531622886657715,
"logps/chosen": -18.537792205810547,
"logps/rejected": -52.4014892578125,
"loss": 0.3551,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7665916681289673,
"rewards/margins": 0.7665916681289673,
"rewards/rejected": 0.0,
"step": 149
},
{
"epoch": 0.8379888268156425,
"grad_norm": 12.698613392610968,
"learning_rate": 5.226480836236934e-07,
"logits/chosen": -3.232588529586792,
"logits/rejected": -3.436432361602783,
"logps/chosen": -30.869352340698242,
"logps/rejected": -27.366065979003906,
"loss": 0.3824,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0566002130508423,
"rewards/margins": 1.0566002130508423,
"rewards/rejected": 0.0,
"step": 150
},
{
"epoch": 0.8435754189944135,
"grad_norm": 10.300001343681256,
"learning_rate": 5.261324041811846e-07,
"logits/chosen": -3.4442994594573975,
"logits/rejected": -3.2811100482940674,
"logps/chosen": -20.960033416748047,
"logps/rejected": -44.81601333618164,
"loss": 0.3153,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3567603826522827,
"rewards/margins": 1.3567603826522827,
"rewards/rejected": 0.0,
"step": 151
},
{
"epoch": 0.8491620111731844,
"grad_norm": 10.426058542643384,
"learning_rate": 5.296167247386759e-07,
"logits/chosen": -3.2275266647338867,
"logits/rejected": -2.9595117568969727,
"logps/chosen": -23.1798095703125,
"logps/rejected": -30.18008804321289,
"loss": 0.319,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.275942325592041,
"rewards/margins": 1.275942325592041,
"rewards/rejected": 0.0,
"step": 152
},
{
"epoch": 0.8547486033519553,
"grad_norm": 14.089341146153586,
"learning_rate": 5.331010452961672e-07,
"logits/chosen": -3.4573311805725098,
"logits/rejected": -3.4678773880004883,
"logps/chosen": -28.501304626464844,
"logps/rejected": -15.408483505249023,
"loss": 0.4136,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.659926176071167,
"rewards/margins": 0.659926176071167,
"rewards/rejected": 0.0,
"step": 153
},
{
"epoch": 0.8603351955307262,
"grad_norm": 13.935787499891406,
"learning_rate": 5.365853658536586e-07,
"logits/chosen": -3.2246856689453125,
"logits/rejected": -3.205880880355835,
"logps/chosen": -10.82982063293457,
"logps/rejected": -18.569210052490234,
"loss": 0.3684,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7554144263267517,
"rewards/margins": 0.7554144263267517,
"rewards/rejected": 0.0,
"step": 154
},
{
"epoch": 0.8659217877094972,
"grad_norm": 11.205255809753025,
"learning_rate": 5.400696864111498e-07,
"logits/chosen": -3.4092469215393066,
"logits/rejected": -3.27697491645813,
"logps/chosen": -19.24517250061035,
"logps/rejected": -27.456302642822266,
"loss": 0.3581,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9654081463813782,
"rewards/margins": 0.9654081463813782,
"rewards/rejected": 0.0,
"step": 155
},
{
"epoch": 0.8715083798882681,
"grad_norm": 13.326177301871766,
"learning_rate": 5.43554006968641e-07,
"logits/chosen": -3.5050294399261475,
"logits/rejected": -3.224290609359741,
"logps/chosen": -14.973371505737305,
"logps/rejected": -39.803672790527344,
"loss": 0.414,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8258927464485168,
"rewards/margins": 0.8258927464485168,
"rewards/rejected": 0.0,
"step": 156
},
{
"epoch": 0.8770949720670391,
"grad_norm": 13.809466960643286,
"learning_rate": 5.470383275261323e-07,
"logits/chosen": -3.009963035583496,
"logits/rejected": -2.6967086791992188,
"logps/chosen": -26.160964965820312,
"logps/rejected": -51.31365966796875,
"loss": 0.3614,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.279192566871643,
"rewards/margins": 1.279192566871643,
"rewards/rejected": 0.0,
"step": 157
},
{
"epoch": 0.88268156424581,
"grad_norm": 14.909423246602014,
"learning_rate": 5.505226480836237e-07,
"logits/chosen": -3.4372055530548096,
"logits/rejected": -3.270108938217163,
"logps/chosen": -13.904682159423828,
"logps/rejected": -23.854835510253906,
"loss": 0.3709,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.46371743083000183,
"rewards/margins": 0.46371743083000183,
"rewards/rejected": 0.0,
"step": 158
},
{
"epoch": 0.888268156424581,
"grad_norm": 11.144765174986764,
"learning_rate": 5.54006968641115e-07,
"logits/chosen": -3.4928090572357178,
"logits/rejected": -3.3432681560516357,
"logps/chosen": -12.412818908691406,
"logps/rejected": -23.834762573242188,
"loss": 0.3319,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7286128997802734,
"rewards/margins": 0.7286128997802734,
"rewards/rejected": 0.0,
"step": 159
},
{
"epoch": 0.8938547486033519,
"grad_norm": 13.863594250814781,
"learning_rate": 5.574912891986063e-07,
"logits/chosen": -3.4417412281036377,
"logits/rejected": -3.0007429122924805,
"logps/chosen": -31.18167495727539,
"logps/rejected": -49.418609619140625,
"loss": 0.367,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.020141363143921,
"rewards/margins": 1.020141363143921,
"rewards/rejected": 0.0,
"step": 160
},
{
"epoch": 0.8938547486033519,
"eval_logits/chosen": -3.1731648445129395,
"eval_logits/rejected": -3.1657137870788574,
"eval_logps/chosen": -20.766986846923828,
"eval_logps/rejected": -32.35370635986328,
"eval_loss": 0.3678908944129944,
"eval_rewards/accuracies": 1.0,
"eval_rewards/chosen": 0.9303056001663208,
"eval_rewards/margins": 0.9303056001663208,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.697,
"eval_samples_per_second": 9.2,
"eval_steps_per_second": 0.297,
"step": 160
},
{
"epoch": 0.8994413407821229,
"grad_norm": 10.547321075543701,
"learning_rate": 5.609756097560975e-07,
"logits/chosen": -3.3858096599578857,
"logits/rejected": -3.1765737533569336,
"logps/chosen": -18.24430274963379,
"logps/rejected": -43.009735107421875,
"loss": 0.3706,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0767275094985962,
"rewards/margins": 1.0767275094985962,
"rewards/rejected": 0.0,
"step": 161
},
{
"epoch": 0.9050279329608939,
"grad_norm": 13.31104092875125,
"learning_rate": 5.644599303135888e-07,
"logits/chosen": -3.298539876937866,
"logits/rejected": -2.8914856910705566,
"logps/chosen": -17.463916778564453,
"logps/rejected": -20.251060485839844,
"loss": 0.3565,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.924486517906189,
"rewards/margins": 0.924486517906189,
"rewards/rejected": 0.0,
"step": 162
},
{
"epoch": 0.9106145251396648,
"grad_norm": 12.189074268400924,
"learning_rate": 5.679442508710801e-07,
"logits/chosen": -3.2942914962768555,
"logits/rejected": -3.2542693614959717,
"logps/chosen": -21.734283447265625,
"logps/rejected": -25.956695556640625,
"loss": 0.364,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8711822032928467,
"rewards/margins": 0.8711822032928467,
"rewards/rejected": 0.0,
"step": 163
},
{
"epoch": 0.9162011173184358,
"grad_norm": 10.739512007740988,
"learning_rate": 5.714285714285714e-07,
"logits/chosen": -3.2355618476867676,
"logits/rejected": -3.083498239517212,
"logps/chosen": -21.765317916870117,
"logps/rejected": -21.472946166992188,
"loss": 0.3473,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1917572021484375,
"rewards/margins": 1.1917572021484375,
"rewards/rejected": 0.0,
"step": 164
},
{
"epoch": 0.9217877094972067,
"grad_norm": 11.746475390473849,
"learning_rate": 5.749128919860628e-07,
"logits/chosen": -3.3734073638916016,
"logits/rejected": -3.3708062171936035,
"logps/chosen": -37.94599151611328,
"logps/rejected": -33.02565383911133,
"loss": 0.3433,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.279693603515625,
"rewards/margins": 1.279693603515625,
"rewards/rejected": 0.0,
"step": 165
},
{
"epoch": 0.9273743016759777,
"grad_norm": 11.414057455521112,
"learning_rate": 5.783972125435541e-07,
"logits/chosen": -3.4064152240753174,
"logits/rejected": -3.3076775074005127,
"logps/chosen": -8.735708236694336,
"logps/rejected": -26.29998016357422,
"loss": 0.377,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.45992982387542725,
"rewards/margins": 0.45992982387542725,
"rewards/rejected": 0.0,
"step": 166
},
{
"epoch": 0.9329608938547486,
"grad_norm": 10.752674294933225,
"learning_rate": 5.818815331010452e-07,
"logits/chosen": -3.4760870933532715,
"logits/rejected": -3.190138101577759,
"logps/chosen": -19.67440414428711,
"logps/rejected": -27.942874908447266,
"loss": 0.3322,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.020792007446289,
"rewards/margins": 1.020792007446289,
"rewards/rejected": 0.0,
"step": 167
},
{
"epoch": 0.9385474860335196,
"grad_norm": 20.059327017371768,
"learning_rate": 5.853658536585365e-07,
"logits/chosen": -3.1289055347442627,
"logits/rejected": -3.1791739463806152,
"logps/chosen": -14.388964653015137,
"logps/rejected": -24.04684829711914,
"loss": 0.458,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6628751754760742,
"rewards/margins": 0.6628751754760742,
"rewards/rejected": 0.0,
"step": 168
},
{
"epoch": 0.9441340782122905,
"grad_norm": 19.40231327977891,
"learning_rate": 5.888501742160279e-07,
"logits/chosen": -3.560133934020996,
"logits/rejected": -3.58990216255188,
"logps/chosen": -19.82522201538086,
"logps/rejected": -28.30991554260254,
"loss": 0.4138,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.792276918888092,
"rewards/margins": 0.792276918888092,
"rewards/rejected": 0.0,
"step": 169
},
{
"epoch": 0.9497206703910615,
"grad_norm": 12.42947365843963,
"learning_rate": 5.923344947735192e-07,
"logits/chosen": -3.2126545906066895,
"logits/rejected": -3.3703675270080566,
"logps/chosen": -29.75718116760254,
"logps/rejected": -31.64392852783203,
"loss": 0.3579,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7188339829444885,
"rewards/margins": 0.7188339829444885,
"rewards/rejected": 0.0,
"step": 170
},
{
"epoch": 0.9553072625698324,
"grad_norm": 18.202343205238847,
"learning_rate": 5.958188153310105e-07,
"logits/chosen": -3.717379570007324,
"logits/rejected": -3.5793988704681396,
"logps/chosen": -20.444347381591797,
"logps/rejected": -33.93192672729492,
"loss": 0.3897,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9968618750572205,
"rewards/margins": 0.9968618750572205,
"rewards/rejected": 0.0,
"step": 171
},
{
"epoch": 0.9608938547486033,
"grad_norm": 12.832841852369326,
"learning_rate": 5.993031358885016e-07,
"logits/chosen": -3.4382286071777344,
"logits/rejected": -3.317105770111084,
"logps/chosen": -18.771831512451172,
"logps/rejected": -34.331298828125,
"loss": 0.3651,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8643215894699097,
"rewards/margins": 0.8643215894699097,
"rewards/rejected": 0.0,
"step": 172
},
{
"epoch": 0.9664804469273743,
"grad_norm": 11.857640826249451,
"learning_rate": 6.02787456445993e-07,
"logits/chosen": -3.6196374893188477,
"logits/rejected": -3.3753185272216797,
"logps/chosen": -19.114072799682617,
"logps/rejected": -29.907459259033203,
"loss": 0.3557,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9212259650230408,
"rewards/margins": 0.9212259650230408,
"rewards/rejected": 0.0,
"step": 173
},
{
"epoch": 0.9720670391061452,
"grad_norm": 11.028463714669611,
"learning_rate": 6.062717770034843e-07,
"logits/chosen": -3.2672202587127686,
"logits/rejected": -3.3163845539093018,
"logps/chosen": -23.178640365600586,
"logps/rejected": -55.991661071777344,
"loss": 0.3474,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8489004969596863,
"rewards/margins": 0.8489004969596863,
"rewards/rejected": 0.0,
"step": 174
},
{
"epoch": 0.9776536312849162,
"grad_norm": 12.810488081859203,
"learning_rate": 6.097560975609756e-07,
"logits/chosen": -3.150639057159424,
"logits/rejected": -3.3473706245422363,
"logps/chosen": -55.005409240722656,
"logps/rejected": -28.745824813842773,
"loss": 0.3478,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4712138175964355,
"rewards/margins": 1.4712138175964355,
"rewards/rejected": 0.0,
"step": 175
},
{
"epoch": 0.9832402234636871,
"grad_norm": 13.180414956676238,
"learning_rate": 6.132404181184669e-07,
"logits/chosen": -3.620823621749878,
"logits/rejected": -3.6447391510009766,
"logps/chosen": -10.457490921020508,
"logps/rejected": -29.693206787109375,
"loss": 0.3509,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6984259486198425,
"rewards/margins": 0.6984259486198425,
"rewards/rejected": 0.0,
"step": 176
},
{
"epoch": 0.9888268156424581,
"grad_norm": 22.420225225210046,
"learning_rate": 6.167247386759582e-07,
"logits/chosen": -3.5637428760528564,
"logits/rejected": -3.59216046333313,
"logps/chosen": -16.807857513427734,
"logps/rejected": -19.454391479492188,
"loss": 0.4543,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7911770343780518,
"rewards/margins": 0.7911770343780518,
"rewards/rejected": 0.0,
"step": 177
},
{
"epoch": 0.994413407821229,
"grad_norm": 11.80586886342996,
"learning_rate": 6.202090592334494e-07,
"logits/chosen": -3.3847668170928955,
"logits/rejected": -3.4201109409332275,
"logps/chosen": -14.308084487915039,
"logps/rejected": -50.307579040527344,
"loss": 0.3606,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.090038776397705,
"rewards/margins": 1.090038776397705,
"rewards/rejected": 0.0,
"step": 178
},
{
"epoch": 1.0,
"grad_norm": 16.952966215261824,
"learning_rate": 6.236933797909407e-07,
"logits/chosen": -3.4724905490875244,
"logits/rejected": -3.308558702468872,
"logps/chosen": -27.244325637817383,
"logps/rejected": -37.93553161621094,
"loss": 0.3722,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1800241470336914,
"rewards/margins": 1.1800241470336914,
"rewards/rejected": 0.0,
"step": 179
},
{
"epoch": 1.005586592178771,
"grad_norm": 9.105872976664521,
"learning_rate": 6.27177700348432e-07,
"logits/chosen": -3.6326866149902344,
"logits/rejected": -3.5780582427978516,
"logps/chosen": -12.944976806640625,
"logps/rejected": -18.96871566772461,
"loss": 0.3302,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0247759819030762,
"rewards/margins": 1.0247759819030762,
"rewards/rejected": 0.0,
"step": 180
},
{
"epoch": 1.005586592178771,
"eval_logits/chosen": -3.2065823078155518,
"eval_logits/rejected": -3.2086079120635986,
"eval_logps/chosen": -20.46588897705078,
"eval_logps/rejected": -31.99078369140625,
"eval_loss": 0.36149224638938904,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 0.960415244102478,
"eval_rewards/margins": 0.960415244102478,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6009,
"eval_samples_per_second": 9.226,
"eval_steps_per_second": 0.298,
"step": 180
},
{
"epoch": 1.011173184357542,
"grad_norm": 13.19416012523754,
"learning_rate": 6.306620209059234e-07,
"logits/chosen": -3.2777421474456787,
"logits/rejected": -3.3989064693450928,
"logps/chosen": -14.293594360351562,
"logps/rejected": -17.12541389465332,
"loss": 0.3666,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5113266706466675,
"rewards/margins": 0.5113266706466675,
"rewards/rejected": 0.0,
"step": 181
},
{
"epoch": 1.0167597765363128,
"grad_norm": 10.887028146578567,
"learning_rate": 6.341463414634146e-07,
"logits/chosen": -3.1348888874053955,
"logits/rejected": -3.224384307861328,
"logps/chosen": -21.554080963134766,
"logps/rejected": -20.96063232421875,
"loss": 0.3146,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9010570645332336,
"rewards/margins": 0.9010570645332336,
"rewards/rejected": 0.0,
"step": 182
},
{
"epoch": 1.0223463687150838,
"grad_norm": 8.963970647137387,
"learning_rate": 6.376306620209058e-07,
"logits/chosen": -3.125688076019287,
"logits/rejected": -3.066281795501709,
"logps/chosen": -17.731704711914062,
"logps/rejected": -16.59673309326172,
"loss": 0.2865,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9751240611076355,
"rewards/margins": 0.9751240611076355,
"rewards/rejected": 0.0,
"step": 183
},
{
"epoch": 1.0279329608938548,
"grad_norm": 10.288390190064316,
"learning_rate": 6.411149825783971e-07,
"logits/chosen": -3.3919193744659424,
"logits/rejected": -3.0586678981781006,
"logps/chosen": -16.1546688079834,
"logps/rejected": -39.81340026855469,
"loss": 0.3408,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.247828722000122,
"rewards/margins": 1.247828722000122,
"rewards/rejected": 0.0,
"step": 184
},
{
"epoch": 1.0335195530726258,
"grad_norm": 11.02350587817682,
"learning_rate": 6.445993031358885e-07,
"logits/chosen": -3.785944938659668,
"logits/rejected": -3.5418670177459717,
"logps/chosen": -16.335508346557617,
"logps/rejected": -20.58266830444336,
"loss": 0.3442,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0386725664138794,
"rewards/margins": 1.0386725664138794,
"rewards/rejected": 0.0,
"step": 185
},
{
"epoch": 1.0391061452513966,
"grad_norm": 16.289447666313198,
"learning_rate": 6.480836236933798e-07,
"logits/chosen": -3.668846845626831,
"logits/rejected": -3.4984891414642334,
"logps/chosen": -25.417245864868164,
"logps/rejected": -39.81985092163086,
"loss": 0.4073,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1094753742218018,
"rewards/margins": 1.1094753742218018,
"rewards/rejected": 0.0,
"step": 186
},
{
"epoch": 1.0446927374301676,
"grad_norm": 15.073431374332136,
"learning_rate": 6.51567944250871e-07,
"logits/chosen": -3.4357309341430664,
"logits/rejected": -3.447826623916626,
"logps/chosen": -21.84659194946289,
"logps/rejected": -23.085220336914062,
"loss": 0.3643,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3426122665405273,
"rewards/margins": 1.3426122665405273,
"rewards/rejected": 0.0,
"step": 187
},
{
"epoch": 1.0502793296089385,
"grad_norm": 10.116314298618896,
"learning_rate": 6.550522648083623e-07,
"logits/chosen": -3.488616704940796,
"logits/rejected": -3.547907590866089,
"logps/chosen": -20.659770965576172,
"logps/rejected": -44.12837219238281,
"loss": 0.3277,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8681479096412659,
"rewards/margins": 0.8681479096412659,
"rewards/rejected": 0.0,
"step": 188
},
{
"epoch": 1.0558659217877095,
"grad_norm": 10.04057046229293,
"learning_rate": 6.585365853658536e-07,
"logits/chosen": -3.3828887939453125,
"logits/rejected": -3.5210986137390137,
"logps/chosen": -15.001754760742188,
"logps/rejected": -18.628276824951172,
"loss": 0.3095,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7176334857940674,
"rewards/margins": 0.7176334857940674,
"rewards/rejected": 0.0,
"step": 189
},
{
"epoch": 1.0614525139664805,
"grad_norm": 15.007274677524912,
"learning_rate": 6.620209059233449e-07,
"logits/chosen": -3.0460798740386963,
"logits/rejected": -3.2319722175598145,
"logps/chosen": -32.32198715209961,
"logps/rejected": -35.442230224609375,
"loss": 0.379,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6987197399139404,
"rewards/margins": 0.6987197399139404,
"rewards/rejected": 0.0,
"step": 190
},
{
"epoch": 1.0670391061452513,
"grad_norm": 9.10362018167136,
"learning_rate": 6.655052264808362e-07,
"logits/chosen": -3.4198787212371826,
"logits/rejected": -3.2804646492004395,
"logps/chosen": -10.96152114868164,
"logps/rejected": -31.262535095214844,
"loss": 0.3236,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6560508012771606,
"rewards/margins": 0.6560508012771606,
"rewards/rejected": 0.0,
"step": 191
},
{
"epoch": 1.0726256983240223,
"grad_norm": 9.590605833890953,
"learning_rate": 6.689895470383276e-07,
"logits/chosen": -3.66292142868042,
"logits/rejected": -3.4462697505950928,
"logps/chosen": -16.839336395263672,
"logps/rejected": -34.59306716918945,
"loss": 0.3152,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2217531204223633,
"rewards/margins": 1.2217531204223633,
"rewards/rejected": 0.0,
"step": 192
},
{
"epoch": 1.0782122905027933,
"grad_norm": 14.136300690926081,
"learning_rate": 6.724738675958188e-07,
"logits/chosen": -3.2333831787109375,
"logits/rejected": -3.228888750076294,
"logps/chosen": -20.69516372680664,
"logps/rejected": -31.679576873779297,
"loss": 0.3615,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.5942341685295105,
"rewards/margins": 0.5942341685295105,
"rewards/rejected": 0.0,
"step": 193
},
{
"epoch": 1.0837988826815643,
"grad_norm": 9.73848287694558,
"learning_rate": 6.759581881533101e-07,
"logits/chosen": -2.731956720352173,
"logits/rejected": -2.8023910522460938,
"logps/chosen": -15.538613319396973,
"logps/rejected": -35.99656677246094,
"loss": 0.3398,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0924439430236816,
"rewards/margins": 1.0924439430236816,
"rewards/rejected": 0.0,
"step": 194
},
{
"epoch": 1.089385474860335,
"grad_norm": 13.504403269737955,
"learning_rate": 6.794425087108013e-07,
"logits/chosen": -3.1872870922088623,
"logits/rejected": -3.1580259799957275,
"logps/chosen": -48.09290313720703,
"logps/rejected": -29.916027069091797,
"loss": 0.3314,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0472760200500488,
"rewards/margins": 1.0472760200500488,
"rewards/rejected": 0.0,
"step": 195
},
{
"epoch": 1.094972067039106,
"grad_norm": 9.013293702308625,
"learning_rate": 6.829268292682927e-07,
"logits/chosen": -3.5854601860046387,
"logits/rejected": -3.320793867111206,
"logps/chosen": -15.177647590637207,
"logps/rejected": -46.74102783203125,
"loss": 0.3142,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8823977112770081,
"rewards/margins": 0.8823977112770081,
"rewards/rejected": 0.0,
"step": 196
},
{
"epoch": 1.100558659217877,
"grad_norm": 9.944518309167927,
"learning_rate": 6.86411149825784e-07,
"logits/chosen": -3.352811574935913,
"logits/rejected": -3.3195114135742188,
"logps/chosen": -14.030511856079102,
"logps/rejected": -51.092716217041016,
"loss": 0.3007,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9310657978057861,
"rewards/margins": 0.9310657978057861,
"rewards/rejected": 0.0,
"step": 197
},
{
"epoch": 1.106145251396648,
"grad_norm": 9.619422555624382,
"learning_rate": 6.898954703832752e-07,
"logits/chosen": -3.5843019485473633,
"logits/rejected": -3.1107637882232666,
"logps/chosen": -5.203390121459961,
"logps/rejected": -27.75282096862793,
"loss": 0.3516,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6825684905052185,
"rewards/margins": 0.6825684905052185,
"rewards/rejected": 0.0,
"step": 198
},
{
"epoch": 1.111731843575419,
"grad_norm": 12.838451492103111,
"learning_rate": 6.933797909407665e-07,
"logits/chosen": -3.3617866039276123,
"logits/rejected": -3.123973846435547,
"logps/chosen": -53.03567886352539,
"logps/rejected": -36.53765106201172,
"loss": 0.3275,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8718262314796448,
"rewards/margins": 0.8718262314796448,
"rewards/rejected": 0.0,
"step": 199
},
{
"epoch": 1.1173184357541899,
"grad_norm": 10.727712294253553,
"learning_rate": 6.968641114982578e-07,
"logits/chosen": -2.0662670135498047,
"logits/rejected": -2.08182692527771,
"logps/chosen": -52.00320816040039,
"logps/rejected": -73.17282104492188,
"loss": 0.3375,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.8044001460075378,
"rewards/margins": 0.8044001460075378,
"rewards/rejected": 0.0,
"step": 200
},
{
"epoch": 1.1173184357541899,
"eval_logits/chosen": -3.246267795562744,
"eval_logits/rejected": -3.2582263946533203,
"eval_logps/chosen": -20.083738327026367,
"eval_logps/rejected": -31.633167266845703,
"eval_loss": 0.3558894991874695,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 0.9986303448677063,
"eval_rewards/margins": 0.9986303448677063,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5976,
"eval_samples_per_second": 9.227,
"eval_steps_per_second": 0.298,
"step": 200
},
{
"epoch": 1.1229050279329609,
"grad_norm": 10.59758427732295,
"learning_rate": 7.003484320557491e-07,
"logits/chosen": -3.505729913711548,
"logits/rejected": -3.1870007514953613,
"logps/chosen": -10.572619438171387,
"logps/rejected": -52.28015899658203,
"loss": 0.3449,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7204795479774475,
"rewards/margins": 0.7204795479774475,
"rewards/rejected": 0.0,
"step": 201
},
{
"epoch": 1.1284916201117319,
"grad_norm": 15.850633721679678,
"learning_rate": 7.038327526132404e-07,
"logits/chosen": -3.2468059062957764,
"logits/rejected": -3.3311328887939453,
"logps/chosen": -19.04383659362793,
"logps/rejected": -15.986388206481934,
"loss": 0.3689,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9080885052680969,
"rewards/margins": 0.9080885052680969,
"rewards/rejected": 0.0,
"step": 202
},
{
"epoch": 1.1340782122905029,
"grad_norm": 12.773768896781231,
"learning_rate": 7.073170731707316e-07,
"logits/chosen": -3.3263590335845947,
"logits/rejected": -3.251199722290039,
"logps/chosen": -32.89419937133789,
"logps/rejected": -29.640453338623047,
"loss": 0.3193,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.5567541122436523,
"rewards/margins": 0.5567541122436523,
"rewards/rejected": 0.0,
"step": 203
},
{
"epoch": 1.1396648044692737,
"grad_norm": 10.169395063409917,
"learning_rate": 7.10801393728223e-07,
"logits/chosen": -3.357675075531006,
"logits/rejected": -3.479375123977661,
"logps/chosen": -18.883747100830078,
"logps/rejected": -26.06609344482422,
"loss": 0.2824,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0959761142730713,
"rewards/margins": 1.0959761142730713,
"rewards/rejected": 0.0,
"step": 204
},
{
"epoch": 1.1452513966480447,
"grad_norm": 11.535769216754757,
"learning_rate": 7.142857142857143e-07,
"logits/chosen": -3.1873953342437744,
"logits/rejected": -3.218331813812256,
"logps/chosen": -43.96869659423828,
"logps/rejected": -32.669837951660156,
"loss": 0.3341,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4571388959884644,
"rewards/margins": 1.4571388959884644,
"rewards/rejected": 0.0,
"step": 205
},
{
"epoch": 1.1508379888268156,
"grad_norm": 12.745600341451343,
"learning_rate": 7.177700348432055e-07,
"logits/chosen": -3.6761059761047363,
"logits/rejected": -3.639869213104248,
"logps/chosen": -19.122093200683594,
"logps/rejected": -23.711383819580078,
"loss": 0.3439,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8524357080459595,
"rewards/margins": 0.8524357080459595,
"rewards/rejected": 0.0,
"step": 206
},
{
"epoch": 1.1564245810055866,
"grad_norm": 9.378824141835013,
"learning_rate": 7.212543554006968e-07,
"logits/chosen": -3.5651609897613525,
"logits/rejected": -3.3586459159851074,
"logps/chosen": -21.362688064575195,
"logps/rejected": -30.987037658691406,
"loss": 0.3087,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2608428001403809,
"rewards/margins": 1.2608428001403809,
"rewards/rejected": 0.0,
"step": 207
},
{
"epoch": 1.1620111731843576,
"grad_norm": 14.298346495114684,
"learning_rate": 7.247386759581882e-07,
"logits/chosen": -3.445509910583496,
"logits/rejected": -3.559532403945923,
"logps/chosen": -15.627222061157227,
"logps/rejected": -48.640472412109375,
"loss": 0.3454,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.47943437099456787,
"rewards/margins": 0.47943437099456787,
"rewards/rejected": 0.0,
"step": 208
},
{
"epoch": 1.1675977653631284,
"grad_norm": 17.065773568044825,
"learning_rate": 7.282229965156794e-07,
"logits/chosen": -3.4207494258880615,
"logits/rejected": -3.4317336082458496,
"logps/chosen": -15.158013343811035,
"logps/rejected": -26.077003479003906,
"loss": 0.3768,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6948219537734985,
"rewards/margins": 0.6948219537734985,
"rewards/rejected": 0.0,
"step": 209
},
{
"epoch": 1.1731843575418994,
"grad_norm": 10.558484474657567,
"learning_rate": 7.317073170731707e-07,
"logits/chosen": -3.489943265914917,
"logits/rejected": -3.3792459964752197,
"logps/chosen": -13.803016662597656,
"logps/rejected": -21.387771606445312,
"loss": 0.2991,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9277787208557129,
"rewards/margins": 0.9277787208557129,
"rewards/rejected": 0.0,
"step": 210
},
{
"epoch": 1.1787709497206704,
"grad_norm": 14.878823416144341,
"learning_rate": 7.351916376306619e-07,
"logits/chosen": -3.1514551639556885,
"logits/rejected": -2.85518741607666,
"logps/chosen": -17.18071746826172,
"logps/rejected": -34.49565887451172,
"loss": 0.3657,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2792776823043823,
"rewards/margins": 1.2792776823043823,
"rewards/rejected": 0.0,
"step": 211
},
{
"epoch": 1.1843575418994414,
"grad_norm": 9.848084960506721,
"learning_rate": 7.386759581881533e-07,
"logits/chosen": -3.138712167739868,
"logits/rejected": -3.03584361076355,
"logps/chosen": -15.721365928649902,
"logps/rejected": -16.612977981567383,
"loss": 0.302,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1289442777633667,
"rewards/margins": 1.1289442777633667,
"rewards/rejected": 0.0,
"step": 212
},
{
"epoch": 1.1899441340782122,
"grad_norm": 10.805227371998999,
"learning_rate": 7.421602787456446e-07,
"logits/chosen": -3.6576991081237793,
"logits/rejected": -3.5775070190429688,
"logps/chosen": -22.21623992919922,
"logps/rejected": -49.09683609008789,
"loss": 0.315,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1914067268371582,
"rewards/margins": 1.1914067268371582,
"rewards/rejected": 0.0,
"step": 213
},
{
"epoch": 1.1955307262569832,
"grad_norm": 10.320243828582568,
"learning_rate": 7.456445993031358e-07,
"logits/chosen": -3.409869909286499,
"logits/rejected": -3.401746988296509,
"logps/chosen": -26.289350509643555,
"logps/rejected": -28.945384979248047,
"loss": 0.3352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3901445865631104,
"rewards/margins": 1.3901445865631104,
"rewards/rejected": 0.0,
"step": 214
},
{
"epoch": 1.2011173184357542,
"grad_norm": 14.18558383515029,
"learning_rate": 7.491289198606271e-07,
"logits/chosen": -3.62727427482605,
"logits/rejected": -3.4659011363983154,
"logps/chosen": -36.72397994995117,
"logps/rejected": -21.145355224609375,
"loss": 0.321,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7905348539352417,
"rewards/margins": 0.7905348539352417,
"rewards/rejected": 0.0,
"step": 215
},
{
"epoch": 1.2067039106145252,
"grad_norm": 10.488275048991248,
"learning_rate": 7.526132404181185e-07,
"logits/chosen": -3.6146132946014404,
"logits/rejected": -3.3445773124694824,
"logps/chosen": -22.47472381591797,
"logps/rejected": -28.731056213378906,
"loss": 0.2963,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4101334810256958,
"rewards/margins": 1.4101334810256958,
"rewards/rejected": 0.0,
"step": 216
},
{
"epoch": 1.2122905027932962,
"grad_norm": 8.704684794484079,
"learning_rate": 7.560975609756097e-07,
"logits/chosen": -3.5806429386138916,
"logits/rejected": -3.7555086612701416,
"logps/chosen": -16.60782814025879,
"logps/rejected": -26.24691390991211,
"loss": 0.2961,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0293834209442139,
"rewards/margins": 1.0293834209442139,
"rewards/rejected": 0.0,
"step": 217
},
{
"epoch": 1.217877094972067,
"grad_norm": 9.698408791817338,
"learning_rate": 7.59581881533101e-07,
"logits/chosen": -3.216071605682373,
"logits/rejected": -3.0813839435577393,
"logps/chosen": -23.708539962768555,
"logps/rejected": -29.808490753173828,
"loss": 0.3132,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2182135581970215,
"rewards/margins": 1.2182135581970215,
"rewards/rejected": 0.0,
"step": 218
},
{
"epoch": 1.223463687150838,
"grad_norm": 9.811614676428968,
"learning_rate": 7.630662020905922e-07,
"logits/chosen": -3.2354507446289062,
"logits/rejected": -3.1731200218200684,
"logps/chosen": -28.39469337463379,
"logps/rejected": -34.51610565185547,
"loss": 0.2879,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0406968593597412,
"rewards/margins": 1.0406968593597412,
"rewards/rejected": 0.0,
"step": 219
},
{
"epoch": 1.229050279329609,
"grad_norm": 11.451301511630755,
"learning_rate": 7.665505226480836e-07,
"logits/chosen": -3.4079675674438477,
"logits/rejected": -3.562708616256714,
"logps/chosen": -50.24363708496094,
"logps/rejected": -36.007076263427734,
"loss": 0.3172,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8625876903533936,
"rewards/margins": 0.8625876903533936,
"rewards/rejected": 0.0,
"step": 220
},
{
"epoch": 1.229050279329609,
"eval_logits/chosen": -3.2746100425720215,
"eval_logits/rejected": -3.2954063415527344,
"eval_logps/chosen": -19.88897705078125,
"eval_logps/rejected": -31.52643394470215,
"eval_loss": 0.35166430473327637,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0181065797805786,
"eval_rewards/margins": 1.0181065797805786,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5481,
"eval_samples_per_second": 9.24,
"eval_steps_per_second": 0.298,
"step": 220
},
{
"epoch": 1.23463687150838,
"grad_norm": 9.326879656223309,
"learning_rate": 7.700348432055749e-07,
"logits/chosen": -3.5331428050994873,
"logits/rejected": -3.4841206073760986,
"logps/chosen": -15.98175048828125,
"logps/rejected": -30.6879940032959,
"loss": 0.3262,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2041575908660889,
"rewards/margins": 1.2041575908660889,
"rewards/rejected": 0.0,
"step": 221
},
{
"epoch": 1.2402234636871508,
"grad_norm": 12.3079719112879,
"learning_rate": 7.735191637630662e-07,
"logits/chosen": -3.518021583557129,
"logits/rejected": -3.4510629177093506,
"logps/chosen": -27.463146209716797,
"logps/rejected": -17.00018310546875,
"loss": 0.3618,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.3559202551841736,
"rewards/margins": 0.3559202551841736,
"rewards/rejected": 0.0,
"step": 222
},
{
"epoch": 1.2458100558659218,
"grad_norm": 11.308218355800076,
"learning_rate": 7.770034843205574e-07,
"logits/chosen": -3.3539228439331055,
"logits/rejected": -3.5053908824920654,
"logps/chosen": -9.660523414611816,
"logps/rejected": -34.22919464111328,
"loss": 0.3076,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7283318638801575,
"rewards/margins": 0.7283318638801575,
"rewards/rejected": 0.0,
"step": 223
},
{
"epoch": 1.2513966480446927,
"grad_norm": 15.695720858566277,
"learning_rate": 7.804878048780488e-07,
"logits/chosen": -3.4702861309051514,
"logits/rejected": -3.6778104305267334,
"logps/chosen": -17.774089813232422,
"logps/rejected": -22.740568161010742,
"loss": 0.3975,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1346460580825806,
"rewards/margins": 1.1346460580825806,
"rewards/rejected": 0.0,
"step": 224
},
{
"epoch": 1.2569832402234637,
"grad_norm": 9.243920919200582,
"learning_rate": 7.8397212543554e-07,
"logits/chosen": -3.1038174629211426,
"logits/rejected": -3.1146140098571777,
"logps/chosen": -45.7878303527832,
"logps/rejected": -30.96743392944336,
"loss": 0.2978,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9360488057136536,
"rewards/margins": 0.9360488057136536,
"rewards/rejected": 0.0,
"step": 225
},
{
"epoch": 1.2625698324022347,
"grad_norm": 11.13129466714581,
"learning_rate": 7.874564459930313e-07,
"logits/chosen": -3.5532631874084473,
"logits/rejected": -3.4976348876953125,
"logps/chosen": -30.162281036376953,
"logps/rejected": -29.01420783996582,
"loss": 0.3294,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7946181893348694,
"rewards/margins": 0.7946181893348694,
"rewards/rejected": 0.0,
"step": 226
},
{
"epoch": 1.2681564245810055,
"grad_norm": 10.721423149528743,
"learning_rate": 7.909407665505227e-07,
"logits/chosen": -3.399590253829956,
"logits/rejected": -3.3550968170166016,
"logps/chosen": -28.896617889404297,
"logps/rejected": -39.47148895263672,
"loss": 0.3103,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5518330335617065,
"rewards/margins": 1.5518330335617065,
"rewards/rejected": 0.0,
"step": 227
},
{
"epoch": 1.2737430167597765,
"grad_norm": 10.794643046108307,
"learning_rate": 7.944250871080139e-07,
"logits/chosen": -3.1023850440979004,
"logits/rejected": -3.070626974105835,
"logps/chosen": -26.629072189331055,
"logps/rejected": -31.59149742126465,
"loss": 0.308,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7542858719825745,
"rewards/margins": 0.7542858719825745,
"rewards/rejected": 0.0,
"step": 228
},
{
"epoch": 1.2793296089385475,
"grad_norm": 9.710935244235399,
"learning_rate": 7.979094076655052e-07,
"logits/chosen": -3.3577988147735596,
"logits/rejected": -3.3791284561157227,
"logps/chosen": -34.757232666015625,
"logps/rejected": -45.71214294433594,
"loss": 0.2734,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.129239559173584,
"rewards/margins": 1.129239559173584,
"rewards/rejected": 0.0,
"step": 229
},
{
"epoch": 1.2849162011173183,
"grad_norm": 10.214850798494435,
"learning_rate": 8.013937282229964e-07,
"logits/chosen": -3.270902633666992,
"logits/rejected": -3.060398578643799,
"logps/chosen": -14.818272590637207,
"logps/rejected": -36.052574157714844,
"loss": 0.3036,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0607706308364868,
"rewards/margins": 1.0607706308364868,
"rewards/rejected": 0.0,
"step": 230
},
{
"epoch": 1.2905027932960893,
"grad_norm": 15.427190116461567,
"learning_rate": 8.048780487804878e-07,
"logits/chosen": -3.6726341247558594,
"logits/rejected": -3.629931926727295,
"logps/chosen": -15.026057243347168,
"logps/rejected": -25.69697380065918,
"loss": 0.3639,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7771012783050537,
"rewards/margins": 0.7771012783050537,
"rewards/rejected": 0.0,
"step": 231
},
{
"epoch": 1.2960893854748603,
"grad_norm": 16.71952562743431,
"learning_rate": 8.083623693379791e-07,
"logits/chosen": -3.426694631576538,
"logits/rejected": -3.055060863494873,
"logps/chosen": -23.176193237304688,
"logps/rejected": -103.94121551513672,
"loss": 0.3559,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7330259680747986,
"rewards/margins": 0.7330259680747986,
"rewards/rejected": 0.0,
"step": 232
},
{
"epoch": 1.3016759776536313,
"grad_norm": 13.460670617842231,
"learning_rate": 8.118466898954704e-07,
"logits/chosen": -2.560324192047119,
"logits/rejected": -2.674717903137207,
"logps/chosen": -43.27853775024414,
"logps/rejected": -37.60259246826172,
"loss": 0.3469,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.79660964012146,
"rewards/margins": 1.79660964012146,
"rewards/rejected": 0.0,
"step": 233
},
{
"epoch": 1.3072625698324023,
"grad_norm": 9.312146717749025,
"learning_rate": 8.153310104529616e-07,
"logits/chosen": -3.848695755004883,
"logits/rejected": -3.8059310913085938,
"logps/chosen": -16.219894409179688,
"logps/rejected": -28.943775177001953,
"loss": 0.2984,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3555338382720947,
"rewards/margins": 1.3555338382720947,
"rewards/rejected": 0.0,
"step": 234
},
{
"epoch": 1.3128491620111733,
"grad_norm": 14.641951211411621,
"learning_rate": 8.188153310104529e-07,
"logits/chosen": -2.8487823009490967,
"logits/rejected": -3.1932733058929443,
"logps/chosen": -52.92893981933594,
"logps/rejected": -19.277240753173828,
"loss": 0.383,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8276066780090332,
"rewards/margins": 0.8276066780090332,
"rewards/rejected": 0.0,
"step": 235
},
{
"epoch": 1.318435754189944,
"grad_norm": 8.445490926381995,
"learning_rate": 8.222996515679442e-07,
"logits/chosen": -3.67144775390625,
"logits/rejected": -3.691436529159546,
"logps/chosen": -31.82388687133789,
"logps/rejected": -26.39704132080078,
"loss": 0.2616,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5748435258865356,
"rewards/margins": 1.5748435258865356,
"rewards/rejected": 0.0,
"step": 236
},
{
"epoch": 1.324022346368715,
"grad_norm": 9.763384731492138,
"learning_rate": 8.257839721254355e-07,
"logits/chosen": -3.3791232109069824,
"logits/rejected": -3.5130560398101807,
"logps/chosen": -10.164708137512207,
"logps/rejected": -49.38196563720703,
"loss": 0.2883,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7769471406936646,
"rewards/margins": 0.7769471406936646,
"rewards/rejected": 0.0,
"step": 237
},
{
"epoch": 1.329608938547486,
"grad_norm": 15.350243675651864,
"learning_rate": 8.292682926829268e-07,
"logits/chosen": -3.604556083679199,
"logits/rejected": -3.3559699058532715,
"logps/chosen": -32.71151351928711,
"logps/rejected": -22.76406478881836,
"loss": 0.3143,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1828868389129639,
"rewards/margins": 1.1828868389129639,
"rewards/rejected": 0.0,
"step": 238
},
{
"epoch": 1.3351955307262569,
"grad_norm": 11.556880663671324,
"learning_rate": 8.327526132404182e-07,
"logits/chosen": -3.514572858810425,
"logits/rejected": -3.5072760581970215,
"logps/chosen": -12.79554557800293,
"logps/rejected": -48.96214294433594,
"loss": 0.3218,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2840112447738647,
"rewards/margins": 1.2840112447738647,
"rewards/rejected": 0.0,
"step": 239
},
{
"epoch": 1.3407821229050279,
"grad_norm": 11.486460847498583,
"learning_rate": 8.362369337979093e-07,
"logits/chosen": -3.382394552230835,
"logits/rejected": -3.1789932250976562,
"logps/chosen": -19.617023468017578,
"logps/rejected": -38.0474853515625,
"loss": 0.3177,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8296595215797424,
"rewards/margins": 0.8296595215797424,
"rewards/rejected": 0.0,
"step": 240
},
{
"epoch": 1.3407821229050279,
"eval_logits/chosen": -3.2941277027130127,
"eval_logits/rejected": -3.319910764694214,
"eval_logps/chosen": -19.708602905273438,
"eval_logps/rejected": -31.385814666748047,
"eval_loss": 0.34877967834472656,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0361440181732178,
"eval_rewards/margins": 1.0361440181732178,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5929,
"eval_samples_per_second": 9.228,
"eval_steps_per_second": 0.298,
"step": 240
},
{
"epoch": 1.3463687150837989,
"grad_norm": 10.085361199189645,
"learning_rate": 8.397212543554006e-07,
"logits/chosen": -3.6434545516967773,
"logits/rejected": -3.575937032699585,
"logps/chosen": -16.716463088989258,
"logps/rejected": -34.58618927001953,
"loss": 0.3216,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2842942476272583,
"rewards/margins": 1.2842942476272583,
"rewards/rejected": 0.0,
"step": 241
},
{
"epoch": 1.3519553072625698,
"grad_norm": 13.10261943018632,
"learning_rate": 8.432055749128919e-07,
"logits/chosen": -2.625068426132202,
"logits/rejected": -2.675497055053711,
"logps/chosen": -23.205856323242188,
"logps/rejected": -25.172626495361328,
"loss": 0.343,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9943975210189819,
"rewards/margins": 0.9943975210189819,
"rewards/rejected": 0.0,
"step": 242
},
{
"epoch": 1.3575418994413408,
"grad_norm": 9.178930009945624,
"learning_rate": 8.466898954703833e-07,
"logits/chosen": -3.7004106044769287,
"logits/rejected": -3.8466243743896484,
"logps/chosen": -11.391777992248535,
"logps/rejected": -27.08580780029297,
"loss": 0.2945,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8786574006080627,
"rewards/margins": 0.8786574006080627,
"rewards/rejected": 0.0,
"step": 243
},
{
"epoch": 1.3631284916201118,
"grad_norm": 12.984989957600904,
"learning_rate": 8.501742160278746e-07,
"logits/chosen": -3.639615297317505,
"logits/rejected": -3.54256272315979,
"logps/chosen": -18.723716735839844,
"logps/rejected": -28.6210880279541,
"loss": 0.3179,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9797879457473755,
"rewards/margins": 0.9797879457473755,
"rewards/rejected": 0.0,
"step": 244
},
{
"epoch": 1.3687150837988826,
"grad_norm": 10.364666344721781,
"learning_rate": 8.536585365853657e-07,
"logits/chosen": -3.396254062652588,
"logits/rejected": -3.1957669258117676,
"logps/chosen": -17.962038040161133,
"logps/rejected": -25.107196807861328,
"loss": 0.2727,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0272819995880127,
"rewards/margins": 1.0272819995880127,
"rewards/rejected": 0.0,
"step": 245
},
{
"epoch": 1.3743016759776536,
"grad_norm": 9.651011599680983,
"learning_rate": 8.57142857142857e-07,
"logits/chosen": -3.681082248687744,
"logits/rejected": -3.4493260383605957,
"logps/chosen": -24.472599029541016,
"logps/rejected": -26.830467224121094,
"loss": 0.2953,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5167380571365356,
"rewards/margins": 1.5167380571365356,
"rewards/rejected": 0.0,
"step": 246
},
{
"epoch": 1.3798882681564246,
"grad_norm": 10.838107337723281,
"learning_rate": 8.606271777003484e-07,
"logits/chosen": -3.1495206356048584,
"logits/rejected": -3.2228331565856934,
"logps/chosen": -12.363944053649902,
"logps/rejected": -28.788822174072266,
"loss": 0.346,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.5605226159095764,
"rewards/margins": 0.5605226159095764,
"rewards/rejected": 0.0,
"step": 247
},
{
"epoch": 1.3854748603351954,
"grad_norm": 11.827297529444412,
"learning_rate": 8.641114982578397e-07,
"logits/chosen": -3.316638231277466,
"logits/rejected": -3.3069708347320557,
"logps/chosen": -18.166549682617188,
"logps/rejected": -42.66516876220703,
"loss": 0.2912,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2724140882492065,
"rewards/margins": 1.2724140882492065,
"rewards/rejected": 0.0,
"step": 248
},
{
"epoch": 1.3910614525139664,
"grad_norm": 11.01082188378406,
"learning_rate": 8.67595818815331e-07,
"logits/chosen": -3.594625949859619,
"logits/rejected": -3.4195151329040527,
"logps/chosen": -17.03807830810547,
"logps/rejected": -22.575963973999023,
"loss": 0.2864,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2216452360153198,
"rewards/margins": 1.2216452360153198,
"rewards/rejected": 0.0,
"step": 249
},
{
"epoch": 1.3966480446927374,
"grad_norm": 11.611478895373097,
"learning_rate": 8.710801393728223e-07,
"logits/chosen": -3.7455551624298096,
"logits/rejected": -3.686239719390869,
"logps/chosen": -11.092876434326172,
"logps/rejected": -62.6893310546875,
"loss": 0.2938,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8907960653305054,
"rewards/margins": 0.8907960653305054,
"rewards/rejected": 0.0,
"step": 250
},
{
"epoch": 1.4022346368715084,
"grad_norm": 12.031609061258749,
"learning_rate": 8.745644599303135e-07,
"logits/chosen": -3.03202486038208,
"logits/rejected": -3.108268976211548,
"logps/chosen": -51.005760192871094,
"logps/rejected": -26.438575744628906,
"loss": 0.3098,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.7945677042007446,
"rewards/margins": 0.7945677042007446,
"rewards/rejected": 0.0,
"step": 251
},
{
"epoch": 1.4078212290502794,
"grad_norm": 8.882604544624783,
"learning_rate": 8.780487804878048e-07,
"logits/chosen": -3.4373087882995605,
"logits/rejected": -3.1588633060455322,
"logps/chosen": -30.948307037353516,
"logps/rejected": -26.18224334716797,
"loss": 0.2956,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5699577331542969,
"rewards/margins": 1.5699577331542969,
"rewards/rejected": 0.0,
"step": 252
},
{
"epoch": 1.4134078212290504,
"grad_norm": 11.334729057967122,
"learning_rate": 8.815331010452961e-07,
"logits/chosen": -3.963742256164551,
"logits/rejected": -3.785430908203125,
"logps/chosen": -18.78699493408203,
"logps/rejected": -36.91880798339844,
"loss": 0.3319,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2210891246795654,
"rewards/margins": 1.2210891246795654,
"rewards/rejected": 0.0,
"step": 253
},
{
"epoch": 1.4189944134078212,
"grad_norm": 12.317771739780161,
"learning_rate": 8.850174216027874e-07,
"logits/chosen": -3.338820219039917,
"logits/rejected": -3.345203399658203,
"logps/chosen": -11.049318313598633,
"logps/rejected": -27.54586410522461,
"loss": 0.3411,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1689317226409912,
"rewards/margins": 1.1689317226409912,
"rewards/rejected": 0.0,
"step": 254
},
{
"epoch": 1.4245810055865922,
"grad_norm": 12.878542430550699,
"learning_rate": 8.885017421602788e-07,
"logits/chosen": -3.6580207347869873,
"logits/rejected": -3.5710203647613525,
"logps/chosen": -15.82120418548584,
"logps/rejected": -26.688724517822266,
"loss": 0.3185,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1673306226730347,
"rewards/margins": 1.1673306226730347,
"rewards/rejected": 0.0,
"step": 255
},
{
"epoch": 1.4301675977653632,
"grad_norm": 13.135079685475205,
"learning_rate": 8.9198606271777e-07,
"logits/chosen": -3.5650031566619873,
"logits/rejected": -3.6054866313934326,
"logps/chosen": -10.30579662322998,
"logps/rejected": -19.772520065307617,
"loss": 0.3444,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7571910619735718,
"rewards/margins": 0.7571910619735718,
"rewards/rejected": 0.0,
"step": 256
},
{
"epoch": 1.435754189944134,
"grad_norm": 13.055299569071433,
"learning_rate": 8.954703832752612e-07,
"logits/chosen": -3.363001823425293,
"logits/rejected": -3.2151830196380615,
"logps/chosen": -21.344371795654297,
"logps/rejected": -34.520687103271484,
"loss": 0.3164,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5332088470458984,
"rewards/margins": 1.5332088470458984,
"rewards/rejected": 0.0,
"step": 257
},
{
"epoch": 1.441340782122905,
"grad_norm": 13.31330440079475,
"learning_rate": 8.989547038327526e-07,
"logits/chosen": -3.4222359657287598,
"logits/rejected": -3.3897311687469482,
"logps/chosen": -26.0482120513916,
"logps/rejected": -31.58816146850586,
"loss": 0.3375,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5745426416397095,
"rewards/margins": 1.5745426416397095,
"rewards/rejected": 0.0,
"step": 258
},
{
"epoch": 1.446927374301676,
"grad_norm": 9.644841054252685,
"learning_rate": 9.024390243902439e-07,
"logits/chosen": -3.189476728439331,
"logits/rejected": -3.0403568744659424,
"logps/chosen": -17.522367477416992,
"logps/rejected": -40.978904724121094,
"loss": 0.2895,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9727367162704468,
"rewards/margins": 0.9727367162704468,
"rewards/rejected": 0.0,
"step": 259
},
{
"epoch": 1.452513966480447,
"grad_norm": 9.595550125483722,
"learning_rate": 9.059233449477352e-07,
"logits/chosen": -3.685537815093994,
"logits/rejected": -3.7384848594665527,
"logps/chosen": -11.41524887084961,
"logps/rejected": -23.859867095947266,
"loss": 0.318,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.869933545589447,
"rewards/margins": 0.869933545589447,
"rewards/rejected": 0.0,
"step": 260
},
{
"epoch": 1.452513966480447,
"eval_logits/chosen": -3.3066470623016357,
"eval_logits/rejected": -3.338177442550659,
"eval_logps/chosen": -19.498775482177734,
"eval_logps/rejected": -31.355274200439453,
"eval_loss": 0.34678158164024353,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.057126522064209,
"eval_rewards/margins": 1.057126522064209,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6534,
"eval_samples_per_second": 9.212,
"eval_steps_per_second": 0.297,
"step": 260
},
{
"epoch": 1.458100558659218,
"grad_norm": 12.575604168259234,
"learning_rate": 9.094076655052264e-07,
"logits/chosen": -3.4521327018737793,
"logits/rejected": -3.5857722759246826,
"logps/chosen": -23.312461853027344,
"logps/rejected": -23.504331588745117,
"loss": 0.2838,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1274609565734863,
"rewards/margins": 1.1274609565734863,
"rewards/rejected": 0.0,
"step": 261
},
{
"epoch": 1.463687150837989,
"grad_norm": 13.736170560933992,
"learning_rate": 9.128919860627177e-07,
"logits/chosen": -3.2900683879852295,
"logits/rejected": -3.16727614402771,
"logps/chosen": -33.037010192871094,
"logps/rejected": -41.841793060302734,
"loss": 0.346,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.271679401397705,
"rewards/margins": 1.271679401397705,
"rewards/rejected": 0.0,
"step": 262
},
{
"epoch": 1.4692737430167597,
"grad_norm": 10.469962806407239,
"learning_rate": 9.16376306620209e-07,
"logits/chosen": -3.060082197189331,
"logits/rejected": -3.1097497940063477,
"logps/chosen": -14.170034408569336,
"logps/rejected": -15.353635787963867,
"loss": 0.3308,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7865402698516846,
"rewards/margins": 0.7865402698516846,
"rewards/rejected": 0.0,
"step": 263
},
{
"epoch": 1.4748603351955307,
"grad_norm": 11.331562925216264,
"learning_rate": 9.198606271777003e-07,
"logits/chosen": -3.761474370956421,
"logits/rejected": -3.6347577571868896,
"logps/chosen": -20.385032653808594,
"logps/rejected": -31.475448608398438,
"loss": 0.2778,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1574748754501343,
"rewards/margins": 1.1574748754501343,
"rewards/rejected": 0.0,
"step": 264
},
{
"epoch": 1.4804469273743017,
"grad_norm": 13.562600056540433,
"learning_rate": 9.233449477351916e-07,
"logits/chosen": -3.46614408493042,
"logits/rejected": -3.1842451095581055,
"logps/chosen": -15.610736846923828,
"logps/rejected": -65.3194351196289,
"loss": 0.3205,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.061760663986206,
"rewards/margins": 1.061760663986206,
"rewards/rejected": 0.0,
"step": 265
},
{
"epoch": 1.4860335195530725,
"grad_norm": 10.88847151339513,
"learning_rate": 9.26829268292683e-07,
"logits/chosen": -3.5326571464538574,
"logits/rejected": -3.43210506439209,
"logps/chosen": -17.096965789794922,
"logps/rejected": -25.843481063842773,
"loss": 0.3111,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9644085168838501,
"rewards/margins": 0.9644085168838501,
"rewards/rejected": 0.0,
"step": 266
},
{
"epoch": 1.4916201117318435,
"grad_norm": 9.930526188424318,
"learning_rate": 9.303135888501742e-07,
"logits/chosen": -3.2434909343719482,
"logits/rejected": -3.5659306049346924,
"logps/chosen": -36.433223724365234,
"logps/rejected": -46.459388732910156,
"loss": 0.292,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1708192825317383,
"rewards/margins": 1.1708192825317383,
"rewards/rejected": 0.0,
"step": 267
},
{
"epoch": 1.4972067039106145,
"grad_norm": 12.444217260653138,
"learning_rate": 9.337979094076654e-07,
"logits/chosen": -3.221174955368042,
"logits/rejected": -2.937802314758301,
"logps/chosen": -33.17811584472656,
"logps/rejected": -35.00051498413086,
"loss": 0.3233,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7740553021430969,
"rewards/margins": 0.7740553021430969,
"rewards/rejected": 0.0,
"step": 268
},
{
"epoch": 1.5027932960893855,
"grad_norm": 10.716801635731775,
"learning_rate": 9.372822299651567e-07,
"logits/chosen": -3.3607850074768066,
"logits/rejected": -3.2163217067718506,
"logps/chosen": -16.583881378173828,
"logps/rejected": -48.29698944091797,
"loss": 0.3303,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3584688901901245,
"rewards/margins": 1.3584688901901245,
"rewards/rejected": 0.0,
"step": 269
},
{
"epoch": 1.5083798882681565,
"grad_norm": 15.041626311606112,
"learning_rate": 9.407665505226481e-07,
"logits/chosen": -3.3310461044311523,
"logits/rejected": -3.4619643688201904,
"logps/chosen": -38.412445068359375,
"logps/rejected": -26.61621856689453,
"loss": 0.3557,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6411152482032776,
"rewards/margins": 0.6411152482032776,
"rewards/rejected": 0.0,
"step": 270
},
{
"epoch": 1.5139664804469275,
"grad_norm": 10.425063879304092,
"learning_rate": 9.442508710801394e-07,
"logits/chosen": -3.748817205429077,
"logits/rejected": -3.6500179767608643,
"logps/chosen": -21.386062622070312,
"logps/rejected": -22.98751449584961,
"loss": 0.2863,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0100945234298706,
"rewards/margins": 1.0100945234298706,
"rewards/rejected": 0.0,
"step": 271
},
{
"epoch": 1.5195530726256983,
"grad_norm": 12.140949409242403,
"learning_rate": 9.477351916376306e-07,
"logits/chosen": -3.5599870681762695,
"logits/rejected": -2.9221935272216797,
"logps/chosen": -35.05908203125,
"logps/rejected": -48.583229064941406,
"loss": 0.3589,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5782874226570129,
"rewards/margins": 0.5782874226570129,
"rewards/rejected": 0.0,
"step": 272
},
{
"epoch": 1.5251396648044693,
"grad_norm": 16.849180066616874,
"learning_rate": 9.512195121951218e-07,
"logits/chosen": -3.553140163421631,
"logits/rejected": -3.5460431575775146,
"logps/chosen": -17.00033950805664,
"logps/rejected": -20.65375518798828,
"loss": 0.3222,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0098237991333008,
"rewards/margins": 1.0098237991333008,
"rewards/rejected": 0.0,
"step": 273
},
{
"epoch": 1.5307262569832403,
"grad_norm": 10.37525339082753,
"learning_rate": 9.547038327526132e-07,
"logits/chosen": -3.491828441619873,
"logits/rejected": -2.971925973892212,
"logps/chosen": -19.172990798950195,
"logps/rejected": -57.93187713623047,
"loss": 0.2899,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0543417930603027,
"rewards/margins": 1.0543417930603027,
"rewards/rejected": 0.0,
"step": 274
},
{
"epoch": 1.536312849162011,
"grad_norm": 15.071930004216172,
"learning_rate": 9.581881533101046e-07,
"logits/chosen": -3.063854932785034,
"logits/rejected": -3.2093632221221924,
"logps/chosen": -36.64848327636719,
"logps/rejected": -28.312206268310547,
"loss": 0.3545,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0443449020385742,
"rewards/margins": 1.0443449020385742,
"rewards/rejected": 0.0,
"step": 275
},
{
"epoch": 1.541899441340782,
"grad_norm": 12.085634495974382,
"learning_rate": 9.616724738675958e-07,
"logits/chosen": -3.3405849933624268,
"logits/rejected": -3.4201269149780273,
"logps/chosen": -27.971378326416016,
"logps/rejected": -28.627212524414062,
"loss": 0.3322,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.19189453125,
"rewards/margins": 1.19189453125,
"rewards/rejected": 0.0,
"step": 276
},
{
"epoch": 1.547486033519553,
"grad_norm": 10.207854538997713,
"learning_rate": 9.65156794425087e-07,
"logits/chosen": -3.3752753734588623,
"logits/rejected": -3.546302556991577,
"logps/chosen": -38.091705322265625,
"logps/rejected": -34.98358917236328,
"loss": 0.3002,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0685861110687256,
"rewards/margins": 2.0685861110687256,
"rewards/rejected": 0.0,
"step": 277
},
{
"epoch": 1.553072625698324,
"grad_norm": 11.616172239405316,
"learning_rate": 9.686411149825783e-07,
"logits/chosen": -3.7080671787261963,
"logits/rejected": -3.5104024410247803,
"logps/chosen": -15.084697723388672,
"logps/rejected": -25.442108154296875,
"loss": 0.3144,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0713567733764648,
"rewards/margins": 1.0713567733764648,
"rewards/rejected": 0.0,
"step": 278
},
{
"epoch": 1.558659217877095,
"grad_norm": 13.22103411673907,
"learning_rate": 9.721254355400697e-07,
"logits/chosen": -3.4512689113616943,
"logits/rejected": -3.094844341278076,
"logps/chosen": -26.031299591064453,
"logps/rejected": -23.88300323486328,
"loss": 0.3545,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0946345329284668,
"rewards/margins": 1.0946345329284668,
"rewards/rejected": 0.0,
"step": 279
},
{
"epoch": 1.564245810055866,
"grad_norm": 10.027545681733605,
"learning_rate": 9.756097560975609e-07,
"logits/chosen": -3.328197479248047,
"logits/rejected": -3.5146186351776123,
"logps/chosen": -39.08555603027344,
"logps/rejected": -29.00870132446289,
"loss": 0.3004,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8183404207229614,
"rewards/margins": 0.8183404207229614,
"rewards/rejected": 0.0,
"step": 280
},
{
"epoch": 1.564245810055866,
"eval_logits/chosen": -3.3012683391571045,
"eval_logits/rejected": -3.3296456336975098,
"eval_logps/chosen": -19.454092025756836,
"eval_logps/rejected": -31.271724700927734,
"eval_loss": 0.34393545985221863,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0615949630737305,
"eval_rewards/margins": 1.0615949630737305,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6036,
"eval_samples_per_second": 9.225,
"eval_steps_per_second": 0.298,
"step": 280
},
{
"epoch": 1.5698324022346368,
"grad_norm": 11.965310318453831,
"learning_rate": 9.790940766550523e-07,
"logits/chosen": -3.3429605960845947,
"logits/rejected": -3.2995493412017822,
"logps/chosen": -22.374677658081055,
"logps/rejected": -16.40281105041504,
"loss": 0.3241,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.042400598526001,
"rewards/margins": 1.042400598526001,
"rewards/rejected": 0.0,
"step": 281
},
{
"epoch": 1.5754189944134078,
"grad_norm": 10.352211488442286,
"learning_rate": 9.825783972125434e-07,
"logits/chosen": -3.5352978706359863,
"logits/rejected": -3.5306782722473145,
"logps/chosen": -12.284505844116211,
"logps/rejected": -26.704856872558594,
"loss": 0.3093,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.870249330997467,
"rewards/margins": 0.870249330997467,
"rewards/rejected": 0.0,
"step": 282
},
{
"epoch": 1.5810055865921788,
"grad_norm": 12.032630843979712,
"learning_rate": 9.860627177700348e-07,
"logits/chosen": -3.429222583770752,
"logits/rejected": -3.3236958980560303,
"logps/chosen": -27.717411041259766,
"logps/rejected": -25.07077980041504,
"loss": 0.3017,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2042782306671143,
"rewards/margins": 1.2042782306671143,
"rewards/rejected": 0.0,
"step": 283
},
{
"epoch": 1.5865921787709496,
"grad_norm": 9.779522151659501,
"learning_rate": 9.89547038327526e-07,
"logits/chosen": -3.578108310699463,
"logits/rejected": -3.4658658504486084,
"logps/chosen": -13.309062004089355,
"logps/rejected": -26.286941528320312,
"loss": 0.3008,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9821389317512512,
"rewards/margins": 0.9821389317512512,
"rewards/rejected": 0.0,
"step": 284
},
{
"epoch": 1.5921787709497206,
"grad_norm": 10.263148728342822,
"learning_rate": 9.930313588850174e-07,
"logits/chosen": -3.5624876022338867,
"logits/rejected": -3.529696464538574,
"logps/chosen": -19.440580368041992,
"logps/rejected": -21.190061569213867,
"loss": 0.2961,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6123175621032715,
"rewards/margins": 0.6123175621032715,
"rewards/rejected": 0.0,
"step": 285
},
{
"epoch": 1.5977653631284916,
"grad_norm": 10.342644978214786,
"learning_rate": 9.965156794425088e-07,
"logits/chosen": -3.0572822093963623,
"logits/rejected": -3.336202383041382,
"logps/chosen": -17.684846878051758,
"logps/rejected": -34.97415542602539,
"loss": 0.2706,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2538070678710938,
"rewards/margins": 1.2538070678710938,
"rewards/rejected": 0.0,
"step": 286
},
{
"epoch": 1.6033519553072626,
"grad_norm": 17.455815237252825,
"learning_rate": 1e-06,
"logits/chosen": -3.422065496444702,
"logits/rejected": -3.199143409729004,
"logps/chosen": -12.241714477539062,
"logps/rejected": -20.09499740600586,
"loss": 0.3422,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9664725065231323,
"rewards/margins": 0.9664725065231323,
"rewards/rejected": 0.0,
"step": 287
},
{
"epoch": 1.6089385474860336,
"grad_norm": 10.920072184573593,
"learning_rate": 9.999996284554774e-07,
"logits/chosen": -3.2156920433044434,
"logits/rejected": -3.287858009338379,
"logps/chosen": -14.97732162475586,
"logps/rejected": -24.15703773498535,
"loss": 0.288,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1677476167678833,
"rewards/margins": 1.1677476167678833,
"rewards/rejected": 0.0,
"step": 288
},
{
"epoch": 1.6145251396648046,
"grad_norm": 10.17580058388229,
"learning_rate": 9.999985138224618e-07,
"logits/chosen": -3.4886186122894287,
"logits/rejected": -3.316105604171753,
"logps/chosen": -13.622641563415527,
"logps/rejected": -46.91883087158203,
"loss": 0.3019,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9123961925506592,
"rewards/margins": 0.9123961925506592,
"rewards/rejected": 0.0,
"step": 289
},
{
"epoch": 1.6201117318435754,
"grad_norm": 9.326145815880142,
"learning_rate": 9.999966561026099e-07,
"logits/chosen": -3.2829484939575195,
"logits/rejected": -3.075709104537964,
"logps/chosen": -10.445840835571289,
"logps/rejected": -40.42874526977539,
"loss": 0.3039,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8486101627349854,
"rewards/margins": 0.8486101627349854,
"rewards/rejected": 0.0,
"step": 290
},
{
"epoch": 1.6256983240223464,
"grad_norm": 9.776873356124641,
"learning_rate": 9.999940552986825e-07,
"logits/chosen": -3.469417095184326,
"logits/rejected": -3.279420852661133,
"logps/chosen": -17.53714370727539,
"logps/rejected": -23.5706787109375,
"loss": 0.2718,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2589737176895142,
"rewards/margins": 1.2589737176895142,
"rewards/rejected": 0.0,
"step": 291
},
{
"epoch": 1.6312849162011172,
"grad_norm": 10.429075276268662,
"learning_rate": 9.99990711414545e-07,
"logits/chosen": -3.7084476947784424,
"logits/rejected": -3.5754408836364746,
"logps/chosen": -24.693119049072266,
"logps/rejected": -33.41971969604492,
"loss": 0.3242,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.529041051864624,
"rewards/margins": 1.529041051864624,
"rewards/rejected": 0.0,
"step": 292
},
{
"epoch": 1.6368715083798882,
"grad_norm": 9.720924585440683,
"learning_rate": 9.999866244551666e-07,
"logits/chosen": -3.4124324321746826,
"logits/rejected": -3.210049867630005,
"logps/chosen": -11.158379554748535,
"logps/rejected": -43.5897331237793,
"loss": 0.3196,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0147404670715332,
"rewards/margins": 1.0147404670715332,
"rewards/rejected": 0.0,
"step": 293
},
{
"epoch": 1.6424581005586592,
"grad_norm": 9.944702206569138,
"learning_rate": 9.999817944266215e-07,
"logits/chosen": -3.3606770038604736,
"logits/rejected": -3.4656686782836914,
"logps/chosen": -6.251188278198242,
"logps/rejected": -21.617168426513672,
"loss": 0.3217,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5969514846801758,
"rewards/margins": 0.5969514846801758,
"rewards/rejected": 0.0,
"step": 294
},
{
"epoch": 1.6480446927374302,
"grad_norm": 10.385353254236556,
"learning_rate": 9.999762213360883e-07,
"logits/chosen": -3.466520071029663,
"logits/rejected": -3.5797107219696045,
"logps/chosen": -25.80699920654297,
"logps/rejected": -19.83127212524414,
"loss": 0.3194,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3000495433807373,
"rewards/margins": 1.3000495433807373,
"rewards/rejected": 0.0,
"step": 295
},
{
"epoch": 1.6536312849162011,
"grad_norm": 8.773867276405428,
"learning_rate": 9.99969905191849e-07,
"logits/chosen": -3.557112693786621,
"logits/rejected": -3.573246955871582,
"logps/chosen": -20.78133773803711,
"logps/rejected": -34.07285690307617,
"loss": 0.2889,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1277602910995483,
"rewards/margins": 1.1277602910995483,
"rewards/rejected": 0.0,
"step": 296
},
{
"epoch": 1.6592178770949721,
"grad_norm": 9.821752375679727,
"learning_rate": 9.99962846003291e-07,
"logits/chosen": -3.7407591342926025,
"logits/rejected": -3.767958879470825,
"logps/chosen": -14.965896606445312,
"logps/rejected": -20.833538055419922,
"loss": 0.2999,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9994887113571167,
"rewards/margins": 0.9994887113571167,
"rewards/rejected": 0.0,
"step": 297
},
{
"epoch": 1.6648044692737431,
"grad_norm": 10.928183640862303,
"learning_rate": 9.999550437809053e-07,
"logits/chosen": -3.2713358402252197,
"logits/rejected": -3.4095914363861084,
"logps/chosen": -27.807035446166992,
"logps/rejected": -38.68109130859375,
"loss": 0.2967,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2415056228637695,
"rewards/margins": 1.2415056228637695,
"rewards/rejected": 0.0,
"step": 298
},
{
"epoch": 1.670391061452514,
"grad_norm": 17.270546512795296,
"learning_rate": 9.999464985362873e-07,
"logits/chosen": -3.3679652214050293,
"logits/rejected": -3.5425937175750732,
"logps/chosen": -61.781837463378906,
"logps/rejected": -28.911365509033203,
"loss": 0.3695,
"rewards/accuracies": 0.5,
"rewards/chosen": 0.8019064664840698,
"rewards/margins": 0.8019064664840698,
"rewards/rejected": 0.0,
"step": 299
},
{
"epoch": 1.675977653631285,
"grad_norm": 14.185589116130966,
"learning_rate": 9.999372102821368e-07,
"logits/chosen": -3.5178616046905518,
"logits/rejected": -3.4500045776367188,
"logps/chosen": -11.943248748779297,
"logps/rejected": -31.703353881835938,
"loss": 0.3579,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0013020038604736,
"rewards/margins": 1.0013020038604736,
"rewards/rejected": 0.0,
"step": 300
},
{
"epoch": 1.675977653631285,
"eval_logits/chosen": -3.322420597076416,
"eval_logits/rejected": -3.361907958984375,
"eval_logps/chosen": -19.29962158203125,
"eval_logps/rejected": -31.117549896240234,
"eval_loss": 0.344208300113678,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0770422220230103,
"eval_rewards/margins": 1.0770422220230103,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6372,
"eval_samples_per_second": 9.216,
"eval_steps_per_second": 0.297,
"step": 300
},
{
"epoch": 1.6815642458100557,
"grad_norm": 9.939112746890036,
"learning_rate": 9.999271790322582e-07,
"logits/chosen": -3.5513951778411865,
"logits/rejected": -3.587817668914795,
"logps/chosen": -10.393442153930664,
"logps/rejected": -32.21281814575195,
"loss": 0.3038,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.921258807182312,
"rewards/margins": 0.921258807182312,
"rewards/rejected": 0.0,
"step": 301
},
{
"epoch": 1.6871508379888267,
"grad_norm": 15.239563770884459,
"learning_rate": 9.99916404801559e-07,
"logits/chosen": -3.6404824256896973,
"logits/rejected": -3.5286920070648193,
"logps/chosen": -32.23038864135742,
"logps/rejected": -39.896217346191406,
"loss": 0.3328,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8375921249389648,
"rewards/margins": 0.8375921249389648,
"rewards/rejected": 0.0,
"step": 302
},
{
"epoch": 1.6927374301675977,
"grad_norm": 10.592627695108975,
"learning_rate": 9.999048876060525e-07,
"logits/chosen": -3.9261133670806885,
"logits/rejected": -3.867961883544922,
"logps/chosen": -10.186397552490234,
"logps/rejected": -26.875429153442383,
"loss": 0.3275,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.019852638244629,
"rewards/margins": 1.019852638244629,
"rewards/rejected": 0.0,
"step": 303
},
{
"epoch": 1.6983240223463687,
"grad_norm": 11.76517910635558,
"learning_rate": 9.998926274628546e-07,
"logits/chosen": -3.2354116439819336,
"logits/rejected": -3.0572993755340576,
"logps/chosen": -36.67839050292969,
"logps/rejected": -22.477191925048828,
"loss": 0.2708,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.6770723462104797,
"rewards/margins": 0.6770723462104797,
"rewards/rejected": 0.0,
"step": 304
},
{
"epoch": 1.7039106145251397,
"grad_norm": 14.159098515470435,
"learning_rate": 9.99879624390186e-07,
"logits/chosen": -3.5911238193511963,
"logits/rejected": -3.755124092102051,
"logps/chosen": -21.760345458984375,
"logps/rejected": -15.202546119689941,
"loss": 0.3113,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0169107913970947,
"rewards/margins": 1.0169107913970947,
"rewards/rejected": 0.0,
"step": 305
},
{
"epoch": 1.7094972067039107,
"grad_norm": 9.736802850531827,
"learning_rate": 9.998658784073723e-07,
"logits/chosen": -3.1625123023986816,
"logits/rejected": -3.1821582317352295,
"logps/chosen": -21.424030303955078,
"logps/rejected": -20.46610450744629,
"loss": 0.2899,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1583060026168823,
"rewards/margins": 1.1583060026168823,
"rewards/rejected": 0.0,
"step": 306
},
{
"epoch": 1.7150837988826817,
"grad_norm": 11.009203336538388,
"learning_rate": 9.998513895348417e-07,
"logits/chosen": -3.2195072174072266,
"logits/rejected": -3.4642629623413086,
"logps/chosen": -17.911853790283203,
"logps/rejected": -15.646801948547363,
"loss": 0.3135,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1663609743118286,
"rewards/margins": 1.1663609743118286,
"rewards/rejected": 0.0,
"step": 307
},
{
"epoch": 1.7206703910614525,
"grad_norm": 10.53450471248475,
"learning_rate": 9.998361577941278e-07,
"logits/chosen": -3.488490104675293,
"logits/rejected": -3.322944164276123,
"logps/chosen": -16.006378173828125,
"logps/rejected": -28.02416229248047,
"loss": 0.3442,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.024048089981079,
"rewards/margins": 1.024048089981079,
"rewards/rejected": 0.0,
"step": 308
},
{
"epoch": 1.7262569832402235,
"grad_norm": 12.582435776032845,
"learning_rate": 9.998201832078674e-07,
"logits/chosen": -3.3339123725891113,
"logits/rejected": -3.316779851913452,
"logps/chosen": -28.325908660888672,
"logps/rejected": -16.219600677490234,
"loss": 0.3268,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9893666505813599,
"rewards/margins": 0.9893666505813599,
"rewards/rejected": 0.0,
"step": 309
},
{
"epoch": 1.7318435754189943,
"grad_norm": 13.41209684100597,
"learning_rate": 9.998034657998019e-07,
"logits/chosen": -3.640216588973999,
"logits/rejected": -3.4110727310180664,
"logps/chosen": -11.483031272888184,
"logps/rejected": -40.59260559082031,
"loss": 0.3189,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0720748901367188,
"rewards/margins": 1.0720748901367188,
"rewards/rejected": 0.0,
"step": 310
},
{
"epoch": 1.7374301675977653,
"grad_norm": 15.205084198489166,
"learning_rate": 9.997860055947756e-07,
"logits/chosen": -3.6733248233795166,
"logits/rejected": -3.6724252700805664,
"logps/chosen": -7.7731428146362305,
"logps/rejected": -32.35521697998047,
"loss": 0.3746,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5072684288024902,
"rewards/margins": 0.5072684288024902,
"rewards/rejected": 0.0,
"step": 311
},
{
"epoch": 1.7430167597765363,
"grad_norm": 11.414239251919467,
"learning_rate": 9.997678026187383e-07,
"logits/chosen": -3.6234049797058105,
"logits/rejected": -3.676602363586426,
"logps/chosen": -12.703027725219727,
"logps/rejected": -30.089244842529297,
"loss": 0.3159,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7911865711212158,
"rewards/margins": 0.7911865711212158,
"rewards/rejected": 0.0,
"step": 312
},
{
"epoch": 1.7486033519553073,
"grad_norm": 12.823829809057761,
"learning_rate": 9.997488568987425e-07,
"logits/chosen": -3.5169265270233154,
"logits/rejected": -3.595710039138794,
"logps/chosen": -11.737189292907715,
"logps/rejected": -31.50156021118164,
"loss": 0.3476,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8435858488082886,
"rewards/margins": 0.8435858488082886,
"rewards/rejected": 0.0,
"step": 313
},
{
"epoch": 1.7541899441340782,
"grad_norm": 10.401381588926625,
"learning_rate": 9.997291684629449e-07,
"logits/chosen": -3.3408241271972656,
"logits/rejected": -3.465125560760498,
"logps/chosen": -18.41136932373047,
"logps/rejected": -20.647062301635742,
"loss": 0.3035,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3910760879516602,
"rewards/margins": 1.3910760879516602,
"rewards/rejected": 0.0,
"step": 314
},
{
"epoch": 1.7597765363128492,
"grad_norm": 8.50936106344969,
"learning_rate": 9.99708737340606e-07,
"logits/chosen": -3.2572875022888184,
"logits/rejected": -3.19663143157959,
"logps/chosen": -24.815534591674805,
"logps/rejected": -24.796138763427734,
"loss": 0.2827,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.203381061553955,
"rewards/margins": 1.203381061553955,
"rewards/rejected": 0.0,
"step": 315
},
{
"epoch": 1.7653631284916202,
"grad_norm": 15.15456006027298,
"learning_rate": 9.996875635620902e-07,
"logits/chosen": -3.3994522094726562,
"logits/rejected": -3.299717903137207,
"logps/chosen": -14.031728744506836,
"logps/rejected": -47.74146270751953,
"loss": 0.3482,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1319087743759155,
"rewards/margins": 1.1319087743759155,
"rewards/rejected": 0.0,
"step": 316
},
{
"epoch": 1.770949720670391,
"grad_norm": 13.875581331692954,
"learning_rate": 9.996656471588656e-07,
"logits/chosen": -3.031449556350708,
"logits/rejected": -3.1041312217712402,
"logps/chosen": -19.398300170898438,
"logps/rejected": -33.0574951171875,
"loss": 0.3268,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2168288230895996,
"rewards/margins": 1.2168288230895996,
"rewards/rejected": 0.0,
"step": 317
},
{
"epoch": 1.776536312849162,
"grad_norm": 13.74240813346542,
"learning_rate": 9.996429881635036e-07,
"logits/chosen": -3.656426191329956,
"logits/rejected": -3.317042112350464,
"logps/chosen": -10.931537628173828,
"logps/rejected": -45.407440185546875,
"loss": 0.3717,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8999665975570679,
"rewards/margins": 0.8999665975570679,
"rewards/rejected": 0.0,
"step": 318
},
{
"epoch": 1.7821229050279328,
"grad_norm": 16.88114307609148,
"learning_rate": 9.996195866096797e-07,
"logits/chosen": -3.3901474475860596,
"logits/rejected": -3.2347769737243652,
"logps/chosen": -20.043804168701172,
"logps/rejected": -23.260557174682617,
"loss": 0.3657,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1257836818695068,
"rewards/margins": 1.1257836818695068,
"rewards/rejected": 0.0,
"step": 319
},
{
"epoch": 1.7877094972067038,
"grad_norm": 19.952904936930338,
"learning_rate": 9.995954425321727e-07,
"logits/chosen": -3.3588874340057373,
"logits/rejected": -3.2260642051696777,
"logps/chosen": -66.18074798583984,
"logps/rejected": -35.820491790771484,
"loss": 0.3457,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.2065391540527344,
"rewards/margins": 1.2065391540527344,
"rewards/rejected": 0.0,
"step": 320
},
{
"epoch": 1.7877094972067038,
"eval_logits/chosen": -3.3043503761291504,
"eval_logits/rejected": -3.3413078784942627,
"eval_logps/chosen": -19.3726749420166,
"eval_logps/rejected": -31.22524642944336,
"eval_loss": 0.34021639823913574,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0697367191314697,
"eval_rewards/margins": 1.0697367191314697,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6357,
"eval_samples_per_second": 9.216,
"eval_steps_per_second": 0.297,
"step": 320
},
{
"epoch": 1.7932960893854748,
"grad_norm": 9.301428863056444,
"learning_rate": 9.99570555966865e-07,
"logits/chosen": -3.5000479221343994,
"logits/rejected": -3.395080089569092,
"logps/chosen": -30.758852005004883,
"logps/rejected": -30.442750930786133,
"loss": 0.2747,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9963459968566895,
"rewards/margins": 1.9963459968566895,
"rewards/rejected": 0.0,
"step": 321
},
{
"epoch": 1.7988826815642458,
"grad_norm": 9.171701997345275,
"learning_rate": 9.995449269507422e-07,
"logits/chosen": -3.5941829681396484,
"logits/rejected": -3.524878978729248,
"logps/chosen": -9.524094581604004,
"logps/rejected": -32.728294372558594,
"loss": 0.2647,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7229211330413818,
"rewards/margins": 0.7229211330413818,
"rewards/rejected": 0.0,
"step": 322
},
{
"epoch": 1.8044692737430168,
"grad_norm": 10.69476295700697,
"learning_rate": 9.995185555218943e-07,
"logits/chosen": -3.359325408935547,
"logits/rejected": -3.411824941635132,
"logps/chosen": -14.76585578918457,
"logps/rejected": -23.467166900634766,
"loss": 0.3463,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.5161473155021667,
"rewards/margins": 0.5161473155021667,
"rewards/rejected": 0.0,
"step": 323
},
{
"epoch": 1.8100558659217878,
"grad_norm": 11.118347896847748,
"learning_rate": 9.994914417195133e-07,
"logits/chosen": -3.2842142581939697,
"logits/rejected": -3.1278810501098633,
"logps/chosen": -9.60197639465332,
"logps/rejected": -22.244657516479492,
"loss": 0.3278,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7514290809631348,
"rewards/margins": 0.7514290809631348,
"rewards/rejected": 0.0,
"step": 324
},
{
"epoch": 1.8156424581005588,
"grad_norm": 13.029770617319564,
"learning_rate": 9.994635855838955e-07,
"logits/chosen": -3.752204179763794,
"logits/rejected": -3.814516305923462,
"logps/chosen": -18.200443267822266,
"logps/rejected": -22.95499038696289,
"loss": 0.3628,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9537428617477417,
"rewards/margins": 0.9537428617477417,
"rewards/rejected": 0.0,
"step": 325
},
{
"epoch": 1.8212290502793296,
"grad_norm": 12.513334642173808,
"learning_rate": 9.994349871564399e-07,
"logits/chosen": -3.5632996559143066,
"logits/rejected": -3.520127296447754,
"logps/chosen": -26.24734115600586,
"logps/rejected": -28.493831634521484,
"loss": 0.2948,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1466038227081299,
"rewards/margins": 1.1466038227081299,
"rewards/rejected": 0.0,
"step": 326
},
{
"epoch": 1.8268156424581006,
"grad_norm": 9.798945596568291,
"learning_rate": 9.994056464796488e-07,
"logits/chosen": -2.756512403488159,
"logits/rejected": -2.7556052207946777,
"logps/chosen": -42.9817008972168,
"logps/rejected": -46.558284759521484,
"loss": 0.3303,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8298018574714661,
"rewards/margins": 0.8298018574714661,
"rewards/rejected": 0.0,
"step": 327
},
{
"epoch": 1.8324022346368714,
"grad_norm": 10.918392572339316,
"learning_rate": 9.993755635971276e-07,
"logits/chosen": -3.643589496612549,
"logits/rejected": -3.6738765239715576,
"logps/chosen": -17.016036987304688,
"logps/rejected": -27.112552642822266,
"loss": 0.3252,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4312465190887451,
"rewards/margins": 1.4312465190887451,
"rewards/rejected": 0.0,
"step": 328
},
{
"epoch": 1.8379888268156424,
"grad_norm": 13.706813583359736,
"learning_rate": 9.993447385535851e-07,
"logits/chosen": -2.72222638130188,
"logits/rejected": -2.739816427230835,
"logps/chosen": -28.710716247558594,
"logps/rejected": -27.147722244262695,
"loss": 0.3485,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7921134233474731,
"rewards/margins": 0.7921134233474731,
"rewards/rejected": 0.0,
"step": 329
},
{
"epoch": 1.8435754189944134,
"grad_norm": 13.114944075004717,
"learning_rate": 9.99313171394833e-07,
"logits/chosen": -3.390263795852661,
"logits/rejected": -3.270251750946045,
"logps/chosen": -27.65669059753418,
"logps/rejected": -27.90283203125,
"loss": 0.3475,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8979809284210205,
"rewards/margins": 1.8979809284210205,
"rewards/rejected": 0.0,
"step": 330
},
{
"epoch": 1.8491620111731844,
"grad_norm": 9.133700091484517,
"learning_rate": 9.992808621677849e-07,
"logits/chosen": -3.3948614597320557,
"logits/rejected": -3.347947597503662,
"logps/chosen": -17.9841251373291,
"logps/rejected": -37.400943756103516,
"loss": 0.2728,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.332104206085205,
"rewards/margins": 1.332104206085205,
"rewards/rejected": 0.0,
"step": 331
},
{
"epoch": 1.8547486033519553,
"grad_norm": 9.317769026143214,
"learning_rate": 9.992478109204587e-07,
"logits/chosen": -3.4927287101745605,
"logits/rejected": -3.7286641597747803,
"logps/chosen": -14.411958694458008,
"logps/rejected": -28.702346801757812,
"loss": 0.2789,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1069204807281494,
"rewards/margins": 1.1069204807281494,
"rewards/rejected": 0.0,
"step": 332
},
{
"epoch": 1.8603351955307263,
"grad_norm": 14.201362831463452,
"learning_rate": 9.992140177019743e-07,
"logits/chosen": -3.479283571243286,
"logits/rejected": -3.3960371017456055,
"logps/chosen": -11.420783996582031,
"logps/rejected": -22.475345611572266,
"loss": 0.3369,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8035995364189148,
"rewards/margins": 0.8035995364189148,
"rewards/rejected": 0.0,
"step": 333
},
{
"epoch": 1.8659217877094973,
"grad_norm": 11.175373020780594,
"learning_rate": 9.991794825625546e-07,
"logits/chosen": -3.3228561878204346,
"logits/rejected": -3.2982242107391357,
"logps/chosen": -5.908030986785889,
"logps/rejected": -56.5552978515625,
"loss": 0.3385,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6184911727905273,
"rewards/margins": 0.6184911727905273,
"rewards/rejected": 0.0,
"step": 334
},
{
"epoch": 1.8715083798882681,
"grad_norm": 13.277307948180802,
"learning_rate": 9.991442055535246e-07,
"logits/chosen": -3.829953670501709,
"logits/rejected": -3.4723265171051025,
"logps/chosen": -14.793371200561523,
"logps/rejected": -20.083637237548828,
"loss": 0.3148,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3210265636444092,
"rewards/margins": 1.3210265636444092,
"rewards/rejected": 0.0,
"step": 335
},
{
"epoch": 1.8770949720670391,
"grad_norm": 9.406315729563916,
"learning_rate": 9.991081867273126e-07,
"logits/chosen": -3.4278557300567627,
"logits/rejected": -3.540053606033325,
"logps/chosen": -18.8541259765625,
"logps/rejected": -33.45713806152344,
"loss": 0.2721,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3648288249969482,
"rewards/margins": 1.3648288249969482,
"rewards/rejected": 0.0,
"step": 336
},
{
"epoch": 1.88268156424581,
"grad_norm": 9.818598859075479,
"learning_rate": 9.990714261374489e-07,
"logits/chosen": -3.2425389289855957,
"logits/rejected": -3.331890821456909,
"logps/chosen": -10.86574935913086,
"logps/rejected": -31.01609230041504,
"loss": 0.3132,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.6972233057022095,
"rewards/margins": 0.6972233057022095,
"rewards/rejected": 0.0,
"step": 337
},
{
"epoch": 1.888268156424581,
"grad_norm": 10.079056695811483,
"learning_rate": 9.99033923838566e-07,
"logits/chosen": -3.457608222961426,
"logits/rejected": -3.2599005699157715,
"logps/chosen": -45.99235916137695,
"logps/rejected": -35.339759826660156,
"loss": 0.2789,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6016557216644287,
"rewards/margins": 1.6016557216644287,
"rewards/rejected": 0.0,
"step": 338
},
{
"epoch": 1.893854748603352,
"grad_norm": 11.410716488878608,
"learning_rate": 9.989956798863992e-07,
"logits/chosen": -3.503572940826416,
"logits/rejected": -3.4884138107299805,
"logps/chosen": -19.33344268798828,
"logps/rejected": -22.809539794921875,
"loss": 0.3277,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0641868114471436,
"rewards/margins": 1.0641868114471436,
"rewards/rejected": 0.0,
"step": 339
},
{
"epoch": 1.899441340782123,
"grad_norm": 12.560637580640856,
"learning_rate": 9.989566943377859e-07,
"logits/chosen": -3.2320213317871094,
"logits/rejected": -3.377939462661743,
"logps/chosen": -48.088279724121094,
"logps/rejected": -21.49866485595703,
"loss": 0.3362,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3534762859344482,
"rewards/margins": 1.3534762859344482,
"rewards/rejected": 0.0,
"step": 340
},
{
"epoch": 1.899441340782123,
"eval_logits/chosen": -3.3132081031799316,
"eval_logits/rejected": -3.350616931915283,
"eval_logps/chosen": -19.246543884277344,
"eval_logps/rejected": -31.018817901611328,
"eval_loss": 0.3389597237110138,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.0823500156402588,
"eval_rewards/margins": 1.0823500156402588,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5746,
"eval_samples_per_second": 9.233,
"eval_steps_per_second": 0.298,
"step": 340
},
{
"epoch": 1.905027932960894,
"grad_norm": 9.892599407239048,
"learning_rate": 9.989169672506653e-07,
"logits/chosen": -3.5700721740722656,
"logits/rejected": -3.4472672939300537,
"logps/chosen": -15.808403015136719,
"logps/rejected": -32.856849670410156,
"loss": 0.2814,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3196511268615723,
"rewards/margins": 1.3196511268615723,
"rewards/rejected": 0.0,
"step": 341
},
{
"epoch": 1.910614525139665,
"grad_norm": 13.061729648562062,
"learning_rate": 9.988764986840794e-07,
"logits/chosen": -3.519749402999878,
"logits/rejected": -3.3913304805755615,
"logps/chosen": -18.575572967529297,
"logps/rejected": -45.02961349487305,
"loss": 0.3381,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1368541717529297,
"rewards/margins": 1.1368541717529297,
"rewards/rejected": 0.0,
"step": 342
},
{
"epoch": 1.916201117318436,
"grad_norm": 14.711655580356501,
"learning_rate": 9.988352886981713e-07,
"logits/chosen": -3.5021438598632812,
"logits/rejected": -3.3737618923187256,
"logps/chosen": -9.830192565917969,
"logps/rejected": -30.507808685302734,
"loss": 0.3278,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9768810868263245,
"rewards/margins": 0.9768810868263245,
"rewards/rejected": 0.0,
"step": 343
},
{
"epoch": 1.9217877094972067,
"grad_norm": 12.603322659212049,
"learning_rate": 9.987933373541866e-07,
"logits/chosen": -3.649998903274536,
"logits/rejected": -3.4158873558044434,
"logps/chosen": -47.723350524902344,
"logps/rejected": -29.496013641357422,
"loss": 0.308,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.104366660118103,
"rewards/margins": 1.104366660118103,
"rewards/rejected": 0.0,
"step": 344
},
{
"epoch": 1.9273743016759777,
"grad_norm": 9.37885249451381,
"learning_rate": 9.98750644714472e-07,
"logits/chosen": -3.400846242904663,
"logits/rejected": -3.4327080249786377,
"logps/chosen": -12.303414344787598,
"logps/rejected": -16.289201736450195,
"loss": 0.2985,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.5962328314781189,
"rewards/margins": 0.5962328314781189,
"rewards/rejected": 0.0,
"step": 345
},
{
"epoch": 1.9329608938547485,
"grad_norm": 11.10219966687012,
"learning_rate": 9.987072108424771e-07,
"logits/chosen": -3.700981378555298,
"logits/rejected": -3.167149543762207,
"logps/chosen": -13.309455871582031,
"logps/rejected": -43.61916732788086,
"loss": 0.3056,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9612671732902527,
"rewards/margins": 0.9612671732902527,
"rewards/rejected": 0.0,
"step": 346
},
{
"epoch": 1.9385474860335195,
"grad_norm": 14.401670529551904,
"learning_rate": 9.986630358027519e-07,
"logits/chosen": -3.465524435043335,
"logits/rejected": -3.315683126449585,
"logps/chosen": -14.349411964416504,
"logps/rejected": -32.112144470214844,
"loss": 0.3166,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.256392478942871,
"rewards/margins": 1.256392478942871,
"rewards/rejected": 0.0,
"step": 347
},
{
"epoch": 1.9441340782122905,
"grad_norm": 11.261316217053585,
"learning_rate": 9.986181196609485e-07,
"logits/chosen": -3.340606689453125,
"logits/rejected": -3.401167154312134,
"logps/chosen": -34.154457092285156,
"logps/rejected": -33.26017761230469,
"loss": 0.3193,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.9464106559753418,
"rewards/margins": 0.9464106559753418,
"rewards/rejected": 0.0,
"step": 348
},
{
"epoch": 1.9497206703910615,
"grad_norm": 8.624436204570914,
"learning_rate": 9.9857246248382e-07,
"logits/chosen": -3.3766283988952637,
"logits/rejected": -3.463714599609375,
"logps/chosen": -8.690055847167969,
"logps/rejected": -27.917278289794922,
"loss": 0.2987,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7766602039337158,
"rewards/margins": 0.7766602039337158,
"rewards/rejected": 0.0,
"step": 349
},
{
"epoch": 1.9553072625698324,
"grad_norm": 14.03088083668422,
"learning_rate": 9.985260643392214e-07,
"logits/chosen": -3.6260368824005127,
"logits/rejected": -3.6323251724243164,
"logps/chosen": -27.713470458984375,
"logps/rejected": -29.578903198242188,
"loss": 0.2988,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.267095685005188,
"rewards/margins": 1.267095685005188,
"rewards/rejected": 0.0,
"step": 350
},
{
"epoch": 1.9608938547486034,
"grad_norm": 12.16822330191893,
"learning_rate": 9.984789252961087e-07,
"logits/chosen": -3.0514800548553467,
"logits/rejected": -2.936555862426758,
"logps/chosen": -56.688716888427734,
"logps/rejected": -36.15301513671875,
"loss": 0.319,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2467468976974487,
"rewards/margins": 1.2467468976974487,
"rewards/rejected": 0.0,
"step": 351
},
{
"epoch": 1.9664804469273744,
"grad_norm": 10.1811700058371,
"learning_rate": 9.984310454245387e-07,
"logits/chosen": -3.7579891681671143,
"logits/rejected": -3.6203415393829346,
"logps/chosen": -17.171201705932617,
"logps/rejected": -24.970687866210938,
"loss": 0.2861,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7694160342216492,
"rewards/margins": 0.7694160342216492,
"rewards/rejected": 0.0,
"step": 352
},
{
"epoch": 1.9720670391061452,
"grad_norm": 8.383463260037288,
"learning_rate": 9.983824247956697e-07,
"logits/chosen": -3.534579277038574,
"logits/rejected": -3.261944055557251,
"logps/chosen": -21.65915298461914,
"logps/rejected": -61.01673126220703,
"loss": 0.2666,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.189016342163086,
"rewards/margins": 1.189016342163086,
"rewards/rejected": 0.0,
"step": 353
},
{
"epoch": 1.9776536312849162,
"grad_norm": 10.72433684140575,
"learning_rate": 9.983330634817604e-07,
"logits/chosen": -3.474168539047241,
"logits/rejected": -3.3596277236938477,
"logps/chosen": -29.677217483520508,
"logps/rejected": -22.099822998046875,
"loss": 0.2935,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5245972871780396,
"rewards/margins": 1.5245972871780396,
"rewards/rejected": 0.0,
"step": 354
},
{
"epoch": 1.983240223463687,
"grad_norm": 9.773463439448527,
"learning_rate": 9.982829615561703e-07,
"logits/chosen": -3.77665638923645,
"logits/rejected": -3.8121116161346436,
"logps/chosen": -13.98667049407959,
"logps/rejected": -18.671918869018555,
"loss": 0.3192,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8396562337875366,
"rewards/margins": 0.8396562337875366,
"rewards/rejected": 0.0,
"step": 355
},
{
"epoch": 1.988826815642458,
"grad_norm": 14.832142493277997,
"learning_rate": 9.9823211909336e-07,
"logits/chosen": -3.797262668609619,
"logits/rejected": -3.7703962326049805,
"logps/chosen": -13.690839767456055,
"logps/rejected": -22.22066879272461,
"loss": 0.304,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1787787675857544,
"rewards/margins": 1.1787787675857544,
"rewards/rejected": 0.0,
"step": 356
},
{
"epoch": 1.994413407821229,
"grad_norm": 9.654778339478744,
"learning_rate": 9.981805361688905e-07,
"logits/chosen": -3.5876450538635254,
"logits/rejected": -3.2162351608276367,
"logps/chosen": -23.21221351623535,
"logps/rejected": -22.66413688659668,
"loss": 0.2689,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9285216331481934,
"rewards/margins": 1.9285216331481934,
"rewards/rejected": 0.0,
"step": 357
},
{
"epoch": 2.0,
"grad_norm": 10.368296764057684,
"learning_rate": 9.98128212859423e-07,
"logits/chosen": -3.411379814147949,
"logits/rejected": -3.3156909942626953,
"logps/chosen": -26.243820190429688,
"logps/rejected": -26.946678161621094,
"loss": 0.2911,
"rewards/accuracies": 0.75,
"rewards/chosen": 0.3683784008026123,
"rewards/margins": 0.3683784008026123,
"rewards/rejected": 0.0,
"step": 358
},
{
"epoch": 2.005586592178771,
"grad_norm": 6.853559733334418,
"learning_rate": 9.980751492427196e-07,
"logits/chosen": -2.940213680267334,
"logits/rejected": -3.107919692993164,
"logps/chosen": -23.83705711364746,
"logps/rejected": -21.831621170043945,
"loss": 0.2572,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8738245964050293,
"rewards/margins": 1.8738245964050293,
"rewards/rejected": 0.0,
"step": 359
},
{
"epoch": 2.011173184357542,
"grad_norm": 7.230543916216328,
"learning_rate": 9.98021345397642e-07,
"logits/chosen": -3.674678087234497,
"logits/rejected": -3.4114761352539062,
"logps/chosen": -7.945418834686279,
"logps/rejected": -19.77797508239746,
"loss": 0.2611,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9623333215713501,
"rewards/margins": 0.9623333215713501,
"rewards/rejected": 0.0,
"step": 360
},
{
"epoch": 2.011173184357542,
"eval_logits/chosen": -3.352512836456299,
"eval_logits/rejected": -3.3942763805389404,
"eval_logps/chosen": -19.009695053100586,
"eval_logps/rejected": -31.048065185546875,
"eval_loss": 0.3390383720397949,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.106034517288208,
"eval_rewards/margins": 1.106034517288208,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5901,
"eval_samples_per_second": 9.229,
"eval_steps_per_second": 0.298,
"step": 360
},
{
"epoch": 2.016759776536313,
"grad_norm": 7.137277220971277,
"learning_rate": 9.979668014041524e-07,
"logits/chosen": -3.667314052581787,
"logits/rejected": -3.686204195022583,
"logps/chosen": -13.76287841796875,
"logps/rejected": -24.61775779724121,
"loss": 0.2442,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6051509380340576,
"rewards/margins": 1.6051509380340576,
"rewards/rejected": 0.0,
"step": 361
},
{
"epoch": 2.022346368715084,
"grad_norm": 6.284282002970191,
"learning_rate": 9.979115173433128e-07,
"logits/chosen": -3.2760801315307617,
"logits/rejected": -3.306630849838257,
"logps/chosen": -31.155853271484375,
"logps/rejected": -43.76316833496094,
"loss": 0.2428,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9641227722167969,
"rewards/margins": 1.9641227722167969,
"rewards/rejected": 0.0,
"step": 362
},
{
"epoch": 2.0279329608938546,
"grad_norm": 7.587844952549412,
"learning_rate": 9.978554932972853e-07,
"logits/chosen": -3.5016777515411377,
"logits/rejected": -3.313396692276001,
"logps/chosen": -25.92389488220215,
"logps/rejected": -32.96814727783203,
"loss": 0.2582,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5138459205627441,
"rewards/margins": 1.5138459205627441,
"rewards/rejected": 0.0,
"step": 363
},
{
"epoch": 2.0335195530726256,
"grad_norm": 7.315707530774128,
"learning_rate": 9.977987293493316e-07,
"logits/chosen": -3.5630156993865967,
"logits/rejected": -3.6592371463775635,
"logps/chosen": -24.435527801513672,
"logps/rejected": -34.446014404296875,
"loss": 0.2369,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.220048189163208,
"rewards/margins": 1.220048189163208,
"rewards/rejected": 0.0,
"step": 364
},
{
"epoch": 2.0391061452513966,
"grad_norm": 6.938745469268472,
"learning_rate": 9.97741225583813e-07,
"logits/chosen": -3.7066171169281006,
"logits/rejected": -3.533735752105713,
"logps/chosen": -12.29153060913086,
"logps/rejected": -30.40150260925293,
"loss": 0.2508,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2298760414123535,
"rewards/margins": 1.2298760414123535,
"rewards/rejected": 0.0,
"step": 365
},
{
"epoch": 2.0446927374301676,
"grad_norm": 6.890346502664088,
"learning_rate": 9.976829820861904e-07,
"logits/chosen": -3.3776321411132812,
"logits/rejected": -3.2434563636779785,
"logps/chosen": -16.57495880126953,
"logps/rejected": -17.069612503051758,
"loss": 0.228,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5914559364318848,
"rewards/margins": 1.5914559364318848,
"rewards/rejected": 0.0,
"step": 366
},
{
"epoch": 2.0502793296089385,
"grad_norm": 8.398925349958642,
"learning_rate": 9.976239989430238e-07,
"logits/chosen": -3.385850667953491,
"logits/rejected": -3.657202959060669,
"logps/chosen": -10.436373710632324,
"logps/rejected": -21.774452209472656,
"loss": 0.2498,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.079164743423462,
"rewards/margins": 1.079164743423462,
"rewards/rejected": 0.0,
"step": 367
},
{
"epoch": 2.0558659217877095,
"grad_norm": 6.474590748831068,
"learning_rate": 9.975642762419727e-07,
"logits/chosen": -3.501469373703003,
"logits/rejected": -3.5434398651123047,
"logps/chosen": -10.090056419372559,
"logps/rejected": -40.99167251586914,
"loss": 0.2454,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0617396831512451,
"rewards/margins": 1.0617396831512451,
"rewards/rejected": 0.0,
"step": 368
},
{
"epoch": 2.0614525139664805,
"grad_norm": 7.754974370689558,
"learning_rate": 9.975038140717958e-07,
"logits/chosen": -3.7316436767578125,
"logits/rejected": -3.6640937328338623,
"logps/chosen": -9.623273849487305,
"logps/rejected": -28.749025344848633,
"loss": 0.2771,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1402578353881836,
"rewards/margins": 1.1402578353881836,
"rewards/rejected": 0.0,
"step": 369
},
{
"epoch": 2.0670391061452515,
"grad_norm": 7.824714594571133,
"learning_rate": 9.974426125223506e-07,
"logits/chosen": -3.76188063621521,
"logits/rejected": -3.493408441543579,
"logps/chosen": -8.70077896118164,
"logps/rejected": -20.87495231628418,
"loss": 0.2659,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1030755043029785,
"rewards/margins": 1.1030755043029785,
"rewards/rejected": 0.0,
"step": 370
},
{
"epoch": 2.0726256983240225,
"grad_norm": 8.03964632747558,
"learning_rate": 9.973806716845937e-07,
"logits/chosen": -3.290515184402466,
"logits/rejected": -3.3557488918304443,
"logps/chosen": -18.478015899658203,
"logps/rejected": -40.62440872192383,
"loss": 0.2351,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.738030195236206,
"rewards/margins": 1.738030195236206,
"rewards/rejected": 0.0,
"step": 371
},
{
"epoch": 2.078212290502793,
"grad_norm": 7.024462513908243,
"learning_rate": 9.973179916505799e-07,
"logits/chosen": -3.6067373752593994,
"logits/rejected": -3.547588586807251,
"logps/chosen": -62.18476867675781,
"logps/rejected": -46.74725341796875,
"loss": 0.2089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.101921558380127,
"rewards/margins": 2.101921558380127,
"rewards/rejected": 0.0,
"step": 372
},
{
"epoch": 2.083798882681564,
"grad_norm": 7.383675548720896,
"learning_rate": 9.972545725134628e-07,
"logits/chosen": -3.372720718383789,
"logits/rejected": -3.7493672370910645,
"logps/chosen": -13.201423645019531,
"logps/rejected": -40.3868408203125,
"loss": 0.2454,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3980975151062012,
"rewards/margins": 1.3980975151062012,
"rewards/rejected": 0.0,
"step": 373
},
{
"epoch": 2.089385474860335,
"grad_norm": 8.722114344066854,
"learning_rate": 9.97190414367495e-07,
"logits/chosen": -3.102618932723999,
"logits/rejected": -3.4442105293273926,
"logps/chosen": -38.9439697265625,
"logps/rejected": -23.601165771484375,
"loss": 0.2853,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5407699346542358,
"rewards/margins": 1.5407699346542358,
"rewards/rejected": 0.0,
"step": 374
},
{
"epoch": 2.094972067039106,
"grad_norm": 7.431900946832534,
"learning_rate": 9.971255173080266e-07,
"logits/chosen": -3.483156204223633,
"logits/rejected": -3.4012656211853027,
"logps/chosen": -13.505800247192383,
"logps/rejected": -27.328590393066406,
"loss": 0.2251,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.495887279510498,
"rewards/margins": 1.495887279510498,
"rewards/rejected": 0.0,
"step": 375
},
{
"epoch": 2.100558659217877,
"grad_norm": 7.470933693936137,
"learning_rate": 9.970598814315063e-07,
"logits/chosen": -3.3025078773498535,
"logits/rejected": -3.138798475265503,
"logps/chosen": -20.259449005126953,
"logps/rejected": -28.857093811035156,
"loss": 0.2419,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7965961694717407,
"rewards/margins": 1.7965961694717407,
"rewards/rejected": 0.0,
"step": 376
},
{
"epoch": 2.106145251396648,
"grad_norm": 9.27200140273602,
"learning_rate": 9.969935068354806e-07,
"logits/chosen": -3.6736648082733154,
"logits/rejected": -3.6531009674072266,
"logps/chosen": -10.967897415161133,
"logps/rejected": -24.558155059814453,
"loss": 0.2284,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3979521989822388,
"rewards/margins": 1.3979521989822388,
"rewards/rejected": 0.0,
"step": 377
},
{
"epoch": 2.111731843575419,
"grad_norm": 9.061565782826737,
"learning_rate": 9.969263936185942e-07,
"logits/chosen": -3.4653420448303223,
"logits/rejected": -3.643233299255371,
"logps/chosen": -42.82839584350586,
"logps/rejected": -40.906795501708984,
"loss": 0.2622,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3158607482910156,
"rewards/margins": 1.3158607482910156,
"rewards/rejected": 0.0,
"step": 378
},
{
"epoch": 2.11731843575419,
"grad_norm": 8.00290082777598,
"learning_rate": 9.96858541880589e-07,
"logits/chosen": -3.524867057800293,
"logits/rejected": -3.4138917922973633,
"logps/chosen": -15.306783676147461,
"logps/rejected": -24.926111221313477,
"loss": 0.2273,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2128949165344238,
"rewards/margins": 1.2128949165344238,
"rewards/rejected": 0.0,
"step": 379
},
{
"epoch": 2.122905027932961,
"grad_norm": 8.282038668199027,
"learning_rate": 9.96789951722305e-07,
"logits/chosen": -2.8289055824279785,
"logits/rejected": -2.834775686264038,
"logps/chosen": -15.831747055053711,
"logps/rejected": -30.430438995361328,
"loss": 0.2496,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0240150690078735,
"rewards/margins": 1.0240150690078735,
"rewards/rejected": 0.0,
"step": 380
},
{
"epoch": 2.122905027932961,
"eval_logits/chosen": -3.3850479125976562,
"eval_logits/rejected": -3.451110363006592,
"eval_logps/chosen": -19.02072525024414,
"eval_logps/rejected": -31.587453842163086,
"eval_loss": 0.3512902557849884,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.1049315929412842,
"eval_rewards/margins": 1.1049315929412842,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6034,
"eval_samples_per_second": 9.225,
"eval_steps_per_second": 0.298,
"step": 380
},
{
"epoch": 2.1284916201117317,
"grad_norm": 7.659119705802381,
"learning_rate": 9.967206232456792e-07,
"logits/chosen": -3.3204312324523926,
"logits/rejected": -3.4590940475463867,
"logps/chosen": -29.15018081665039,
"logps/rejected": -26.92755126953125,
"loss": 0.228,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5056679248809814,
"rewards/margins": 1.5056679248809814,
"rewards/rejected": 0.0,
"step": 381
},
{
"epoch": 2.1340782122905027,
"grad_norm": 7.341943731665606,
"learning_rate": 9.966505565537462e-07,
"logits/chosen": -3.463940382003784,
"logits/rejected": -3.5414977073669434,
"logps/chosen": -23.84471893310547,
"logps/rejected": -43.20133590698242,
"loss": 0.209,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0310187339782715,
"rewards/margins": 2.0310187339782715,
"rewards/rejected": 0.0,
"step": 382
},
{
"epoch": 2.1396648044692737,
"grad_norm": 7.65895099687802,
"learning_rate": 9.965797517506377e-07,
"logits/chosen": -3.5571227073669434,
"logits/rejected": -3.448068618774414,
"logps/chosen": -12.737523078918457,
"logps/rejected": -39.384822845458984,
"loss": 0.2409,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6982755661010742,
"rewards/margins": 1.6982755661010742,
"rewards/rejected": 0.0,
"step": 383
},
{
"epoch": 2.1452513966480447,
"grad_norm": 6.53867383915077,
"learning_rate": 9.965082089415818e-07,
"logits/chosen": -3.653449535369873,
"logits/rejected": -3.518336772918701,
"logps/chosen": -14.590206146240234,
"logps/rejected": -39.92870330810547,
"loss": 0.1985,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8198078870773315,
"rewards/margins": 1.8198078870773315,
"rewards/rejected": 0.0,
"step": 384
},
{
"epoch": 2.1508379888268156,
"grad_norm": 11.769028103839736,
"learning_rate": 9.964359282329042e-07,
"logits/chosen": -3.7708849906921387,
"logits/rejected": -3.714660406112671,
"logps/chosen": -5.746259689331055,
"logps/rejected": -25.73727798461914,
"loss": 0.2974,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9119479060173035,
"rewards/margins": 0.9119479060173035,
"rewards/rejected": 0.0,
"step": 385
},
{
"epoch": 2.1564245810055866,
"grad_norm": 7.368904952446889,
"learning_rate": 9.96362909732027e-07,
"logits/chosen": -3.477548837661743,
"logits/rejected": -3.552312135696411,
"logps/chosen": -23.343643188476562,
"logps/rejected": -42.16249084472656,
"loss": 0.2322,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6210466623306274,
"rewards/margins": 1.6210466623306274,
"rewards/rejected": 0.0,
"step": 386
},
{
"epoch": 2.1620111731843576,
"grad_norm": 7.806055517330862,
"learning_rate": 9.962891535474683e-07,
"logits/chosen": -3.7190170288085938,
"logits/rejected": -3.836012840270996,
"logps/chosen": -7.597694396972656,
"logps/rejected": -36.90325164794922,
"loss": 0.2476,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0124187469482422,
"rewards/margins": 1.0124187469482422,
"rewards/rejected": 0.0,
"step": 387
},
{
"epoch": 2.1675977653631286,
"grad_norm": 7.731071810124617,
"learning_rate": 9.962146597888433e-07,
"logits/chosen": -3.7043721675872803,
"logits/rejected": -3.74884033203125,
"logps/chosen": -37.62725830078125,
"logps/rejected": -27.989416122436523,
"loss": 0.2255,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7073400020599365,
"rewards/margins": 1.7073400020599365,
"rewards/rejected": 0.0,
"step": 388
},
{
"epoch": 2.1731843575418996,
"grad_norm": 7.002769887579688,
"learning_rate": 9.961394285668628e-07,
"logits/chosen": -3.4300410747528076,
"logits/rejected": -3.171128511428833,
"logps/chosen": -16.042146682739258,
"logps/rejected": -29.850797653198242,
"loss": 0.2075,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3781628608703613,
"rewards/margins": 1.3781628608703613,
"rewards/rejected": 0.0,
"step": 389
},
{
"epoch": 2.17877094972067,
"grad_norm": 8.783255120045123,
"learning_rate": 9.96063459993334e-07,
"logits/chosen": -3.6922378540039062,
"logits/rejected": -3.79630184173584,
"logps/chosen": -21.953340530395508,
"logps/rejected": -23.263206481933594,
"loss": 0.2502,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.887188196182251,
"rewards/margins": 1.887188196182251,
"rewards/rejected": 0.0,
"step": 390
},
{
"epoch": 2.184357541899441,
"grad_norm": 7.067604970449007,
"learning_rate": 9.959867541811595e-07,
"logits/chosen": -3.6089160442352295,
"logits/rejected": -3.683166980743408,
"logps/chosen": -20.209598541259766,
"logps/rejected": -29.633625030517578,
"loss": 0.204,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.683326244354248,
"rewards/margins": 1.683326244354248,
"rewards/rejected": 0.0,
"step": 391
},
{
"epoch": 2.189944134078212,
"grad_norm": 7.975251972964183,
"learning_rate": 9.959093112443377e-07,
"logits/chosen": -3.6021435260772705,
"logits/rejected": -3.445990562438965,
"logps/chosen": -24.468564987182617,
"logps/rejected": -21.645599365234375,
"loss": 0.2562,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0186338424682617,
"rewards/margins": 2.0186338424682617,
"rewards/rejected": 0.0,
"step": 392
},
{
"epoch": 2.195530726256983,
"grad_norm": 8.238477675873897,
"learning_rate": 9.958311312979631e-07,
"logits/chosen": -3.7424333095550537,
"logits/rejected": -3.7833523750305176,
"logps/chosen": -22.42348861694336,
"logps/rejected": -37.2231559753418,
"loss": 0.2484,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0180425643920898,
"rewards/margins": 1.0180425643920898,
"rewards/rejected": 0.0,
"step": 393
},
{
"epoch": 2.201117318435754,
"grad_norm": 7.069688666193769,
"learning_rate": 9.957522144582244e-07,
"logits/chosen": -3.653636932373047,
"logits/rejected": -3.844578266143799,
"logps/chosen": -49.03389358520508,
"logps/rejected": -65.41432189941406,
"loss": 0.1985,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2444148063659668,
"rewards/margins": 1.2444148063659668,
"rewards/rejected": 0.0,
"step": 394
},
{
"epoch": 2.206703910614525,
"grad_norm": 7.451897833919531,
"learning_rate": 9.956725608424065e-07,
"logits/chosen": -3.4446585178375244,
"logits/rejected": -3.402251958847046,
"logps/chosen": -21.064973831176758,
"logps/rejected": -19.122060775756836,
"loss": 0.2414,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3341941833496094,
"rewards/margins": 1.3341941833496094,
"rewards/rejected": 0.0,
"step": 395
},
{
"epoch": 2.212290502793296,
"grad_norm": 8.402157956861563,
"learning_rate": 9.955921705688887e-07,
"logits/chosen": -3.3678083419799805,
"logits/rejected": -3.352842330932617,
"logps/chosen": -50.160003662109375,
"logps/rejected": -31.446388244628906,
"loss": 0.2071,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7660503387451172,
"rewards/margins": 1.7660503387451172,
"rewards/rejected": 0.0,
"step": 396
},
{
"epoch": 2.217877094972067,
"grad_norm": 7.344066870799037,
"learning_rate": 9.955110437571454e-07,
"logits/chosen": -3.5992870330810547,
"logits/rejected": -3.6589908599853516,
"logps/chosen": -12.551340103149414,
"logps/rejected": -37.615997314453125,
"loss": 0.2539,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4545259475708008,
"rewards/margins": 1.4545259475708008,
"rewards/rejected": 0.0,
"step": 397
},
{
"epoch": 2.223463687150838,
"grad_norm": 7.220123935246985,
"learning_rate": 9.954291805277454e-07,
"logits/chosen": -3.710559606552124,
"logits/rejected": -3.6783299446105957,
"logps/chosen": -20.05228042602539,
"logps/rejected": -30.739959716796875,
"loss": 0.2515,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4118175506591797,
"rewards/margins": 1.4118175506591797,
"rewards/rejected": 0.0,
"step": 398
},
{
"epoch": 2.2290502793296088,
"grad_norm": 7.4890890379198884,
"learning_rate": 9.95346581002352e-07,
"logits/chosen": -3.7515759468078613,
"logits/rejected": -3.7081682682037354,
"logps/chosen": -14.216001510620117,
"logps/rejected": -45.78520202636719,
"loss": 0.241,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5549489259719849,
"rewards/margins": 1.5549489259719849,
"rewards/rejected": 0.0,
"step": 399
},
{
"epoch": 2.2346368715083798,
"grad_norm": 7.588464054751335,
"learning_rate": 9.952632453037227e-07,
"logits/chosen": -3.35744571685791,
"logits/rejected": -3.4066972732543945,
"logps/chosen": -11.357706069946289,
"logps/rejected": -35.88401412963867,
"loss": 0.2294,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3197273015975952,
"rewards/margins": 1.3197273015975952,
"rewards/rejected": 0.0,
"step": 400
},
{
"epoch": 2.2346368715083798,
"eval_logits/chosen": -3.406590223312378,
"eval_logits/rejected": -3.4720587730407715,
"eval_logps/chosen": -18.758052825927734,
"eval_logps/rejected": -31.479137420654297,
"eval_loss": 0.347927451133728,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.1311991214752197,
"eval_rewards/margins": 1.1311991214752197,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6225,
"eval_samples_per_second": 9.22,
"eval_steps_per_second": 0.297,
"step": 400
},
{
"epoch": 2.2402234636871508,
"grad_norm": 7.582352058613477,
"learning_rate": 9.951791735557093e-07,
"logits/chosen": -3.3883655071258545,
"logits/rejected": -3.6184258460998535,
"logps/chosen": -49.78010940551758,
"logps/rejected": -40.60430145263672,
"loss": 0.2156,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.381688117980957,
"rewards/margins": 1.381688117980957,
"rewards/rejected": 0.0,
"step": 401
},
{
"epoch": 2.2458100558659218,
"grad_norm": 7.350648779219589,
"learning_rate": 9.950943658832576e-07,
"logits/chosen": -3.5008554458618164,
"logits/rejected": -3.2886977195739746,
"logps/chosen": -10.4576997756958,
"logps/rejected": -28.83907699584961,
"loss": 0.2622,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4585580825805664,
"rewards/margins": 1.4585580825805664,
"rewards/rejected": 0.0,
"step": 402
},
{
"epoch": 2.2513966480446927,
"grad_norm": 6.75041204068673,
"learning_rate": 9.950088224124066e-07,
"logits/chosen": -3.7398993968963623,
"logits/rejected": -3.5482699871063232,
"logps/chosen": -10.035907745361328,
"logps/rejected": -38.832725524902344,
"loss": 0.2269,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1586952209472656,
"rewards/margins": 1.1586952209472656,
"rewards/rejected": 0.0,
"step": 403
},
{
"epoch": 2.2569832402234637,
"grad_norm": 7.8954947377312745,
"learning_rate": 9.949225432702892e-07,
"logits/chosen": -3.1041765213012695,
"logits/rejected": -3.3727102279663086,
"logps/chosen": -27.308000564575195,
"logps/rejected": -45.766178131103516,
"loss": 0.233,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3659512996673584,
"rewards/margins": 1.3659512996673584,
"rewards/rejected": 0.0,
"step": 404
},
{
"epoch": 2.2625698324022347,
"grad_norm": 7.76564918974186,
"learning_rate": 9.948355285851317e-07,
"logits/chosen": -3.777467966079712,
"logits/rejected": -3.669759750366211,
"logps/chosen": -19.442195892333984,
"logps/rejected": -21.301504135131836,
"loss": 0.233,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3987016677856445,
"rewards/margins": 1.3987016677856445,
"rewards/rejected": 0.0,
"step": 405
},
{
"epoch": 2.2681564245810057,
"grad_norm": 8.368115428995738,
"learning_rate": 9.947477784862535e-07,
"logits/chosen": -3.5681798458099365,
"logits/rejected": -3.5615100860595703,
"logps/chosen": -31.800512313842773,
"logps/rejected": -34.029930114746094,
"loss": 0.259,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5696308612823486,
"rewards/margins": 1.5696308612823486,
"rewards/rejected": 0.0,
"step": 406
},
{
"epoch": 2.2737430167597763,
"grad_norm": 7.02416768188933,
"learning_rate": 9.946592931040665e-07,
"logits/chosen": -3.4972879886627197,
"logits/rejected": -3.69376277923584,
"logps/chosen": -20.93242645263672,
"logps/rejected": -23.23131561279297,
"loss": 0.2577,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4333603382110596,
"rewards/margins": 1.4333603382110596,
"rewards/rejected": 0.0,
"step": 407
},
{
"epoch": 2.2793296089385473,
"grad_norm": 8.308033162842571,
"learning_rate": 9.945700725700759e-07,
"logits/chosen": -3.5655198097229004,
"logits/rejected": -3.555826187133789,
"logps/chosen": -11.822471618652344,
"logps/rejected": -23.9578800201416,
"loss": 0.2266,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2672007083892822,
"rewards/margins": 1.2672007083892822,
"rewards/rejected": 0.0,
"step": 408
},
{
"epoch": 2.2849162011173183,
"grad_norm": 7.178366503388651,
"learning_rate": 9.944801170168794e-07,
"logits/chosen": -3.5784828662872314,
"logits/rejected": -3.629476547241211,
"logps/chosen": -18.417068481445312,
"logps/rejected": -39.53338623046875,
"loss": 0.2267,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4985662698745728,
"rewards/margins": 1.4985662698745728,
"rewards/rejected": 0.0,
"step": 409
},
{
"epoch": 2.2905027932960893,
"grad_norm": 7.5812514192209415,
"learning_rate": 9.943894265781668e-07,
"logits/chosen": -2.8852853775024414,
"logits/rejected": -2.9510042667388916,
"logps/chosen": -25.948524475097656,
"logps/rejected": -56.017242431640625,
"loss": 0.2209,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.315206527709961,
"rewards/margins": 1.315206527709961,
"rewards/rejected": 0.0,
"step": 410
},
{
"epoch": 2.2960893854748603,
"grad_norm": 8.091361697251866,
"learning_rate": 9.942980013887206e-07,
"logits/chosen": -3.0868020057678223,
"logits/rejected": -3.217686891555786,
"logps/chosen": -14.222066879272461,
"logps/rejected": -55.355682373046875,
"loss": 0.2509,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.462952971458435,
"rewards/margins": 1.462952971458435,
"rewards/rejected": 0.0,
"step": 411
},
{
"epoch": 2.3016759776536313,
"grad_norm": 7.704261312067314,
"learning_rate": 9.942058415844147e-07,
"logits/chosen": -3.523170232772827,
"logits/rejected": -3.435476541519165,
"logps/chosen": -29.237510681152344,
"logps/rejected": -41.71282958984375,
"loss": 0.2056,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.677089214324951,
"rewards/margins": 2.677089214324951,
"rewards/rejected": 0.0,
"step": 412
},
{
"epoch": 2.3072625698324023,
"grad_norm": 7.968377469779117,
"learning_rate": 9.941129473022148e-07,
"logits/chosen": -3.4820969104766846,
"logits/rejected": -3.5996475219726562,
"logps/chosen": -10.383661270141602,
"logps/rejected": -46.434814453125,
"loss": 0.237,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.342440128326416,
"rewards/margins": 1.342440128326416,
"rewards/rejected": 0.0,
"step": 413
},
{
"epoch": 2.3128491620111733,
"grad_norm": 8.865221690856353,
"learning_rate": 9.940193186801786e-07,
"logits/chosen": -3.543195962905884,
"logits/rejected": -3.466761589050293,
"logps/chosen": -21.331546783447266,
"logps/rejected": -60.1800537109375,
"loss": 0.237,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.712676763534546,
"rewards/margins": 1.712676763534546,
"rewards/rejected": 0.0,
"step": 414
},
{
"epoch": 2.3184357541899443,
"grad_norm": 6.842146244984532,
"learning_rate": 9.939249558574548e-07,
"logits/chosen": -3.5901739597320557,
"logits/rejected": -3.483553647994995,
"logps/chosen": -12.781819343566895,
"logps/rejected": -24.080650329589844,
"loss": 0.2128,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1559152603149414,
"rewards/margins": 1.1559152603149414,
"rewards/rejected": 0.0,
"step": 415
},
{
"epoch": 2.3240223463687153,
"grad_norm": 7.230405389338254,
"learning_rate": 9.938298589742834e-07,
"logits/chosen": -3.8077878952026367,
"logits/rejected": -3.926487445831299,
"logps/chosen": -9.203804016113281,
"logps/rejected": -24.789108276367188,
"loss": 0.2251,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1044079065322876,
"rewards/margins": 1.1044079065322876,
"rewards/rejected": 0.0,
"step": 416
},
{
"epoch": 2.329608938547486,
"grad_norm": 8.61336686443517,
"learning_rate": 9.937340281719952e-07,
"logits/chosen": -3.538182497024536,
"logits/rejected": -3.41615629196167,
"logps/chosen": -15.180647850036621,
"logps/rejected": -33.629058837890625,
"loss": 0.2205,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.701636791229248,
"rewards/margins": 1.701636791229248,
"rewards/rejected": 0.0,
"step": 417
},
{
"epoch": 2.335195530726257,
"grad_norm": 10.429641400424977,
"learning_rate": 9.936374635930117e-07,
"logits/chosen": -3.1239843368530273,
"logits/rejected": -3.1988978385925293,
"logps/chosen": -32.015865325927734,
"logps/rejected": -20.780353546142578,
"loss": 0.2093,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3216593265533447,
"rewards/margins": 1.3216593265533447,
"rewards/rejected": 0.0,
"step": 418
},
{
"epoch": 2.340782122905028,
"grad_norm": 7.421014419343002,
"learning_rate": 9.935401653808457e-07,
"logits/chosen": -3.5012435913085938,
"logits/rejected": -3.516075849533081,
"logps/chosen": -22.774662017822266,
"logps/rejected": -39.66460418701172,
"loss": 0.2468,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.757613182067871,
"rewards/margins": 1.757613182067871,
"rewards/rejected": 0.0,
"step": 419
},
{
"epoch": 2.346368715083799,
"grad_norm": 7.576509851812004,
"learning_rate": 9.934421336800992e-07,
"logits/chosen": -3.8414852619171143,
"logits/rejected": -3.465494394302368,
"logps/chosen": -7.34761905670166,
"logps/rejected": -38.78349304199219,
"loss": 0.2657,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3509728908538818,
"rewards/margins": 1.3509728908538818,
"rewards/rejected": 0.0,
"step": 420
},
{
"epoch": 2.346368715083799,
"eval_logits/chosen": -3.4067509174346924,
"eval_logits/rejected": -3.4796643257141113,
"eval_logps/chosen": -18.939788818359375,
"eval_logps/rejected": -32.09042739868164,
"eval_loss": 0.3535655438899994,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.1130255460739136,
"eval_rewards/margins": 1.1130255460739136,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5488,
"eval_samples_per_second": 9.24,
"eval_steps_per_second": 0.298,
"step": 420
},
{
"epoch": 2.35195530726257,
"grad_norm": 8.850244258366178,
"learning_rate": 9.933433686364644e-07,
"logits/chosen": -3.617330312728882,
"logits/rejected": -3.3769824504852295,
"logps/chosen": -15.00255012512207,
"logps/rejected": -18.44194221496582,
"loss": 0.2446,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4632066488265991,
"rewards/margins": 1.4632066488265991,
"rewards/rejected": 0.0,
"step": 421
},
{
"epoch": 2.357541899441341,
"grad_norm": 7.3360611383001775,
"learning_rate": 9.932438703967246e-07,
"logits/chosen": -3.679149866104126,
"logits/rejected": -3.724022150039673,
"logps/chosen": -16.18329429626465,
"logps/rejected": -46.84700012207031,
"loss": 0.2214,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4130620956420898,
"rewards/margins": 1.4130620956420898,
"rewards/rejected": 0.0,
"step": 422
},
{
"epoch": 2.363128491620112,
"grad_norm": 8.520519128108717,
"learning_rate": 9.931436391087513e-07,
"logits/chosen": -3.628382444381714,
"logits/rejected": -3.237825632095337,
"logps/chosen": -6.118841171264648,
"logps/rejected": -23.551803588867188,
"loss": 0.2539,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0133485794067383,
"rewards/margins": 1.0133485794067383,
"rewards/rejected": 0.0,
"step": 423
},
{
"epoch": 2.368715083798883,
"grad_norm": 7.844051635317521,
"learning_rate": 9.930426749215063e-07,
"logits/chosen": -3.580740451812744,
"logits/rejected": -3.685987949371338,
"logps/chosen": -21.19434356689453,
"logps/rejected": -18.519508361816406,
"loss": 0.2257,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.82901132106781,
"rewards/margins": 1.82901132106781,
"rewards/rejected": 0.0,
"step": 424
},
{
"epoch": 2.3743016759776534,
"grad_norm": 10.937737527530368,
"learning_rate": 9.9294097798504e-07,
"logits/chosen": -3.3463058471679688,
"logits/rejected": -3.4205501079559326,
"logps/chosen": -20.50113296508789,
"logps/rejected": -44.7275390625,
"loss": 0.2862,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.96407949924469,
"rewards/margins": 1.96407949924469,
"rewards/rejected": 0.0,
"step": 425
},
{
"epoch": 2.3798882681564244,
"grad_norm": 9.66736665940558,
"learning_rate": 9.928385484504927e-07,
"logits/chosen": -3.729299306869507,
"logits/rejected": -3.614223003387451,
"logps/chosen": -11.2921142578125,
"logps/rejected": -28.89381980895996,
"loss": 0.2489,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2638719081878662,
"rewards/margins": 1.2638719081878662,
"rewards/rejected": 0.0,
"step": 426
},
{
"epoch": 2.3854748603351954,
"grad_norm": 6.717558676276133,
"learning_rate": 9.927353864700926e-07,
"logits/chosen": -3.8699488639831543,
"logits/rejected": -3.7441611289978027,
"logps/chosen": -8.718036651611328,
"logps/rejected": -17.121315002441406,
"loss": 0.2462,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3821943998336792,
"rewards/margins": 1.3821943998336792,
"rewards/rejected": 0.0,
"step": 427
},
{
"epoch": 2.3910614525139664,
"grad_norm": 7.9211753757705825,
"learning_rate": 9.926314921971568e-07,
"logits/chosen": -3.1198785305023193,
"logits/rejected": -3.305750608444214,
"logps/chosen": -20.42072296142578,
"logps/rejected": -23.660072326660156,
"loss": 0.23,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2869709730148315,
"rewards/margins": 1.2869709730148315,
"rewards/rejected": 0.0,
"step": 428
},
{
"epoch": 2.3966480446927374,
"grad_norm": 12.073967218598673,
"learning_rate": 9.925268657860906e-07,
"logits/chosen": -3.7187838554382324,
"logits/rejected": -3.7397525310516357,
"logps/chosen": -7.898566246032715,
"logps/rejected": -16.721696853637695,
"loss": 0.2562,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1372358798980713,
"rewards/margins": 1.1372358798980713,
"rewards/rejected": 0.0,
"step": 429
},
{
"epoch": 2.4022346368715084,
"grad_norm": 8.959279178556857,
"learning_rate": 9.924215073923878e-07,
"logits/chosen": -3.3832149505615234,
"logits/rejected": -3.1709272861480713,
"logps/chosen": -16.782611846923828,
"logps/rejected": -25.03054428100586,
"loss": 0.2458,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.549168586730957,
"rewards/margins": 1.549168586730957,
"rewards/rejected": 0.0,
"step": 430
},
{
"epoch": 2.4078212290502794,
"grad_norm": 10.542399892202047,
"learning_rate": 9.923154171726295e-07,
"logits/chosen": -3.636807441711426,
"logits/rejected": -2.7520697116851807,
"logps/chosen": -14.142077445983887,
"logps/rejected": -35.1505241394043,
"loss": 0.2751,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4068236351013184,
"rewards/margins": 1.4068236351013184,
"rewards/rejected": 0.0,
"step": 431
},
{
"epoch": 2.4134078212290504,
"grad_norm": 8.389676538659716,
"learning_rate": 9.922085952844844e-07,
"logits/chosen": -3.3838095664978027,
"logits/rejected": -3.5815417766571045,
"logps/chosen": -20.966859817504883,
"logps/rejected": -24.863521575927734,
"loss": 0.2465,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4586759805679321,
"rewards/margins": 1.4586759805679321,
"rewards/rejected": 0.0,
"step": 432
},
{
"epoch": 2.4189944134078214,
"grad_norm": 9.465693911271115,
"learning_rate": 9.921010418867095e-07,
"logits/chosen": -3.8997325897216797,
"logits/rejected": -3.8577191829681396,
"logps/chosen": -21.410032272338867,
"logps/rejected": -22.073474884033203,
"loss": 0.2564,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5003677606582642,
"rewards/margins": 1.5003677606582642,
"rewards/rejected": 0.0,
"step": 433
},
{
"epoch": 2.4245810055865924,
"grad_norm": 7.341157159790346,
"learning_rate": 9.91992757139148e-07,
"logits/chosen": -3.2469494342803955,
"logits/rejected": -3.2828922271728516,
"logps/chosen": -11.902131080627441,
"logps/rejected": -21.969844818115234,
"loss": 0.2129,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2229560613632202,
"rewards/margins": 1.2229560613632202,
"rewards/rejected": 0.0,
"step": 434
},
{
"epoch": 2.430167597765363,
"grad_norm": 6.575904750471763,
"learning_rate": 9.918837412027301e-07,
"logits/chosen": -3.707895278930664,
"logits/rejected": -3.5448644161224365,
"logps/chosen": -20.005281448364258,
"logps/rejected": -39.55828094482422,
"loss": 0.2249,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8787816762924194,
"rewards/margins": 1.8787816762924194,
"rewards/rejected": 0.0,
"step": 435
},
{
"epoch": 2.435754189944134,
"grad_norm": 7.2042726517935245,
"learning_rate": 9.917739942394732e-07,
"logits/chosen": -3.557511806488037,
"logits/rejected": -3.477961540222168,
"logps/chosen": -10.696720123291016,
"logps/rejected": -28.972217559814453,
"loss": 0.2361,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1771442890167236,
"rewards/margins": 1.1771442890167236,
"rewards/rejected": 0.0,
"step": 436
},
{
"epoch": 2.441340782122905,
"grad_norm": 7.603658907854762,
"learning_rate": 9.916635164124807e-07,
"logits/chosen": -3.7466676235198975,
"logits/rejected": -3.890836000442505,
"logps/chosen": -10.05591869354248,
"logps/rejected": -53.3323860168457,
"loss": 0.2175,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3430405855178833,
"rewards/margins": 1.3430405855178833,
"rewards/rejected": 0.0,
"step": 437
},
{
"epoch": 2.446927374301676,
"grad_norm": 7.134620867385865,
"learning_rate": 9.915523078859423e-07,
"logits/chosen": -3.170605421066284,
"logits/rejected": -3.387576103210449,
"logps/chosen": -25.830745697021484,
"logps/rejected": -18.52677345275879,
"loss": 0.235,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5619460344314575,
"rewards/margins": 1.5619460344314575,
"rewards/rejected": 0.0,
"step": 438
},
{
"epoch": 2.452513966480447,
"grad_norm": 7.558971744311051,
"learning_rate": 9.914403688251336e-07,
"logits/chosen": -3.418914794921875,
"logits/rejected": -3.3568410873413086,
"logps/chosen": -15.482650756835938,
"logps/rejected": -60.44960403442383,
"loss": 0.2175,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.548828363418579,
"rewards/margins": 1.548828363418579,
"rewards/rejected": 0.0,
"step": 439
},
{
"epoch": 2.458100558659218,
"grad_norm": 8.513656421826987,
"learning_rate": 9.913276993964162e-07,
"logits/chosen": -3.6556973457336426,
"logits/rejected": -3.377732992172241,
"logps/chosen": -33.013580322265625,
"logps/rejected": -30.232341766357422,
"loss": 0.2225,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6873899698257446,
"rewards/margins": 1.6873899698257446,
"rewards/rejected": 0.0,
"step": 440
},
{
"epoch": 2.458100558659218,
"eval_logits/chosen": -3.4010918140411377,
"eval_logits/rejected": -3.471395969390869,
"eval_logps/chosen": -18.80868148803711,
"eval_logps/rejected": -31.709808349609375,
"eval_loss": 0.3474169969558716,
"eval_rewards/accuracies": 0.9750000238418579,
"eval_rewards/chosen": 1.126136302947998,
"eval_rewards/margins": 1.126136302947998,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5932,
"eval_samples_per_second": 9.228,
"eval_steps_per_second": 0.298,
"step": 440
},
{
"epoch": 2.463687150837989,
"grad_norm": 8.45964119210809,
"learning_rate": 9.912142997672365e-07,
"logits/chosen": -3.669959306716919,
"logits/rejected": -3.6248857975006104,
"logps/chosen": -48.35924530029297,
"logps/rejected": -22.86141586303711,
"loss": 0.2135,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.139950752258301,
"rewards/margins": 2.139950752258301,
"rewards/rejected": 0.0,
"step": 441
},
{
"epoch": 2.46927374301676,
"grad_norm": 8.336543648116804,
"learning_rate": 9.911001701061272e-07,
"logits/chosen": -3.458721399307251,
"logits/rejected": -3.456511974334717,
"logps/chosen": -7.8299360275268555,
"logps/rejected": -15.928705215454102,
"loss": 0.2184,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.279940128326416,
"rewards/margins": 1.279940128326416,
"rewards/rejected": 0.0,
"step": 442
},
{
"epoch": 2.4748603351955305,
"grad_norm": 7.98398514794912,
"learning_rate": 9.909853105827049e-07,
"logits/chosen": -3.6968343257904053,
"logits/rejected": -3.656633138656616,
"logps/chosen": -11.460504531860352,
"logps/rejected": -29.274757385253906,
"loss": 0.2583,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0127944946289062,
"rewards/margins": 1.0127944946289062,
"rewards/rejected": 0.0,
"step": 443
},
{
"epoch": 2.4804469273743015,
"grad_norm": 6.471490355639204,
"learning_rate": 9.908697213676715e-07,
"logits/chosen": -3.431952476501465,
"logits/rejected": -3.3830695152282715,
"logps/chosen": -26.654708862304688,
"logps/rejected": -35.89697265625,
"loss": 0.2052,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6807575225830078,
"rewards/margins": 1.6807575225830078,
"rewards/rejected": 0.0,
"step": 444
},
{
"epoch": 2.4860335195530725,
"grad_norm": 8.775363392001863,
"learning_rate": 9.907534026328129e-07,
"logits/chosen": -3.72342848777771,
"logits/rejected": -3.810544729232788,
"logps/chosen": -8.127252578735352,
"logps/rejected": -72.51339721679688,
"loss": 0.2275,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1973881721496582,
"rewards/margins": 1.1973881721496582,
"rewards/rejected": 0.0,
"step": 445
},
{
"epoch": 2.4916201117318435,
"grad_norm": 9.159198455313787,
"learning_rate": 9.906363545509994e-07,
"logits/chosen": -3.7136473655700684,
"logits/rejected": -3.652353525161743,
"logps/chosen": -14.011004447937012,
"logps/rejected": -19.61683464050293,
"loss": 0.2467,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4680709838867188,
"rewards/margins": 1.4680709838867188,
"rewards/rejected": 0.0,
"step": 446
},
{
"epoch": 2.4972067039106145,
"grad_norm": 7.922207692774938,
"learning_rate": 9.905185772961856e-07,
"logits/chosen": -3.517378091812134,
"logits/rejected": -3.40279483795166,
"logps/chosen": -40.9967041015625,
"logps/rejected": -22.460494995117188,
"loss": 0.2311,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9493257999420166,
"rewards/margins": 1.9493257999420166,
"rewards/rejected": 0.0,
"step": 447
},
{
"epoch": 2.5027932960893855,
"grad_norm": 7.398737562546686,
"learning_rate": 9.904000710434096e-07,
"logits/chosen": -3.7379915714263916,
"logits/rejected": -3.6095502376556396,
"logps/chosen": -20.070119857788086,
"logps/rejected": -26.29829978942871,
"loss": 0.2253,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7872674465179443,
"rewards/margins": 1.7872674465179443,
"rewards/rejected": 0.0,
"step": 448
},
{
"epoch": 2.5083798882681565,
"grad_norm": 7.070392361472554,
"learning_rate": 9.902808359687924e-07,
"logits/chosen": -3.3179070949554443,
"logits/rejected": -3.313913583755493,
"logps/chosen": -11.521672248840332,
"logps/rejected": -53.297386169433594,
"loss": 0.2465,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6013566255569458,
"rewards/margins": 1.6013566255569458,
"rewards/rejected": 0.0,
"step": 449
},
{
"epoch": 2.5139664804469275,
"grad_norm": 6.400628795294673,
"learning_rate": 9.901608722495387e-07,
"logits/chosen": -3.8882439136505127,
"logits/rejected": -3.842254400253296,
"logps/chosen": -12.59760856628418,
"logps/rejected": -25.522594451904297,
"loss": 0.2314,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5315548181533813,
"rewards/margins": 1.5315548181533813,
"rewards/rejected": 0.0,
"step": 450
},
{
"epoch": 2.5195530726256985,
"grad_norm": 9.815768243812006,
"learning_rate": 9.90040180063936e-07,
"logits/chosen": -3.7810263633728027,
"logits/rejected": -3.626575231552124,
"logps/chosen": -7.732497215270996,
"logps/rejected": -29.06593132019043,
"loss": 0.2651,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.313602328300476,
"rewards/margins": 1.313602328300476,
"rewards/rejected": 0.0,
"step": 451
},
{
"epoch": 2.5251396648044695,
"grad_norm": 7.2065539188776375,
"learning_rate": 9.899187595913544e-07,
"logits/chosen": -3.0272104740142822,
"logits/rejected": -2.841782808303833,
"logps/chosen": -26.768041610717773,
"logps/rejected": -32.70100021362305,
"loss": 0.2298,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3412799835205078,
"rewards/margins": 1.3412799835205078,
"rewards/rejected": 0.0,
"step": 452
},
{
"epoch": 2.5307262569832405,
"grad_norm": 8.772903781327832,
"learning_rate": 9.89796611012246e-07,
"logits/chosen": -3.745570182800293,
"logits/rejected": -3.7247447967529297,
"logps/chosen": -12.117805480957031,
"logps/rejected": -28.30303192138672,
"loss": 0.2516,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.231385588645935,
"rewards/margins": 1.231385588645935,
"rewards/rejected": 0.0,
"step": 453
},
{
"epoch": 2.536312849162011,
"grad_norm": 8.16462910523264,
"learning_rate": 9.89673734508146e-07,
"logits/chosen": -3.556431531906128,
"logits/rejected": -3.6420626640319824,
"logps/chosen": -10.046062469482422,
"logps/rejected": -24.843278884887695,
"loss": 0.2247,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1439343690872192,
"rewards/margins": 1.1439343690872192,
"rewards/rejected": 0.0,
"step": 454
},
{
"epoch": 2.541899441340782,
"grad_norm": 8.260323686769588,
"learning_rate": 9.8955013026167e-07,
"logits/chosen": -3.498030424118042,
"logits/rejected": -3.5360107421875,
"logps/chosen": -13.82684326171875,
"logps/rejected": -24.774059295654297,
"loss": 0.2436,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1852829456329346,
"rewards/margins": 1.1852829456329346,
"rewards/rejected": 0.0,
"step": 455
},
{
"epoch": 2.547486033519553,
"grad_norm": 6.82526879346672,
"learning_rate": 9.894257984565167e-07,
"logits/chosen": -3.50545597076416,
"logits/rejected": -3.660874843597412,
"logps/chosen": -11.925575256347656,
"logps/rejected": -39.22270965576172,
"loss": 0.2215,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5857868194580078,
"rewards/margins": 1.5857868194580078,
"rewards/rejected": 0.0,
"step": 456
},
{
"epoch": 2.553072625698324,
"grad_norm": 7.818925205163151,
"learning_rate": 9.893007392774645e-07,
"logits/chosen": -3.6695072650909424,
"logits/rejected": -3.783515453338623,
"logps/chosen": -6.81137752532959,
"logps/rejected": -24.871543884277344,
"loss": 0.2204,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9589478969573975,
"rewards/margins": 0.9589478969573975,
"rewards/rejected": 0.0,
"step": 457
},
{
"epoch": 2.558659217877095,
"grad_norm": 7.808942118894329,
"learning_rate": 9.891749529103745e-07,
"logits/chosen": -3.488028049468994,
"logits/rejected": -3.680635929107666,
"logps/chosen": -7.478118896484375,
"logps/rejected": -46.97710418701172,
"loss": 0.2551,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.7159961462020874,
"rewards/margins": 0.7159961462020874,
"rewards/rejected": 0.0,
"step": 458
},
{
"epoch": 2.564245810055866,
"grad_norm": 7.385416058515253,
"learning_rate": 9.890484395421869e-07,
"logits/chosen": -3.5422539710998535,
"logits/rejected": -3.697225570678711,
"logps/chosen": -7.821656703948975,
"logps/rejected": -36.82724380493164,
"loss": 0.2341,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1111254692077637,
"rewards/margins": 1.1111254692077637,
"rewards/rejected": 0.0,
"step": 459
},
{
"epoch": 2.5698324022346366,
"grad_norm": 9.10891011608351,
"learning_rate": 9.889211993609234e-07,
"logits/chosen": -3.4917874336242676,
"logits/rejected": -3.5584025382995605,
"logps/chosen": -29.07617950439453,
"logps/rejected": -73.38496398925781,
"loss": 0.2554,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4889638423919678,
"rewards/margins": 1.4889638423919678,
"rewards/rejected": 0.0,
"step": 460
},
{
"epoch": 2.5698324022346366,
"eval_logits/chosen": -3.4113235473632812,
"eval_logits/rejected": -3.4855563640594482,
"eval_logps/chosen": -18.93764877319336,
"eval_logps/rejected": -31.75335121154785,
"eval_loss": 0.3509919345378876,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.1132395267486572,
"eval_rewards/margins": 1.1132395267486572,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5433,
"eval_samples_per_second": 9.242,
"eval_steps_per_second": 0.298,
"step": 460
},
{
"epoch": 2.5754189944134076,
"grad_norm": 7.519226930117591,
"learning_rate": 9.887932325556858e-07,
"logits/chosen": -3.36040997505188,
"logits/rejected": -3.4190354347229004,
"logps/chosen": -41.65183639526367,
"logps/rejected": -25.77553367614746,
"loss": 0.2352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.757643699645996,
"rewards/margins": 1.757643699645996,
"rewards/rejected": 0.0,
"step": 461
},
{
"epoch": 2.5810055865921786,
"grad_norm": 8.008746265043428,
"learning_rate": 9.88664539316655e-07,
"logits/chosen": -3.5281214714050293,
"logits/rejected": -3.6912050247192383,
"logps/chosen": -19.70013427734375,
"logps/rejected": -29.371292114257812,
"loss": 0.2059,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6448612213134766,
"rewards/margins": 1.6448612213134766,
"rewards/rejected": 0.0,
"step": 462
},
{
"epoch": 2.5865921787709496,
"grad_norm": 7.121357396179553,
"learning_rate": 9.885351198350927e-07,
"logits/chosen": -3.2395377159118652,
"logits/rejected": -3.192934513092041,
"logps/chosen": -12.96796989440918,
"logps/rejected": -40.11272048950195,
"loss": 0.2265,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2172858715057373,
"rewards/margins": 1.2172858715057373,
"rewards/rejected": 0.0,
"step": 463
},
{
"epoch": 2.5921787709497206,
"grad_norm": 6.943555180366184,
"learning_rate": 9.884049743033388e-07,
"logits/chosen": -3.6315016746520996,
"logits/rejected": -3.5012624263763428,
"logps/chosen": -10.144268989562988,
"logps/rejected": -36.149906158447266,
"loss": 0.2218,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6167594194412231,
"rewards/margins": 1.6167594194412231,
"rewards/rejected": 0.0,
"step": 464
},
{
"epoch": 2.5977653631284916,
"grad_norm": 9.093457875219448,
"learning_rate": 9.882741029148127e-07,
"logits/chosen": -3.257049083709717,
"logits/rejected": -3.3369534015655518,
"logps/chosen": -61.227699279785156,
"logps/rejected": -29.78754425048828,
"loss": 0.2268,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8105043172836304,
"rewards/margins": 1.8105043172836304,
"rewards/rejected": 0.0,
"step": 465
},
{
"epoch": 2.6033519553072626,
"grad_norm": 9.024743638308934,
"learning_rate": 9.88142505864013e-07,
"logits/chosen": -3.7688000202178955,
"logits/rejected": -3.7803730964660645,
"logps/chosen": -16.94661521911621,
"logps/rejected": -28.79592514038086,
"loss": 0.2684,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6676511764526367,
"rewards/margins": 1.6676511764526367,
"rewards/rejected": 0.0,
"step": 466
},
{
"epoch": 2.6089385474860336,
"grad_norm": 8.921806265731254,
"learning_rate": 9.880101833465162e-07,
"logits/chosen": -3.630502700805664,
"logits/rejected": -3.7189152240753174,
"logps/chosen": -8.833230018615723,
"logps/rejected": -24.829486846923828,
"loss": 0.2445,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2556135654449463,
"rewards/margins": 1.2556135654449463,
"rewards/rejected": 0.0,
"step": 467
},
{
"epoch": 2.6145251396648046,
"grad_norm": 7.116479813143785,
"learning_rate": 9.87877135558977e-07,
"logits/chosen": -3.725851535797119,
"logits/rejected": -3.6491265296936035,
"logps/chosen": -25.265533447265625,
"logps/rejected": -33.25012969970703,
"loss": 0.1997,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8423246145248413,
"rewards/margins": 1.8423246145248413,
"rewards/rejected": 0.0,
"step": 468
},
{
"epoch": 2.6201117318435756,
"grad_norm": 8.115055828693515,
"learning_rate": 9.877433626991283e-07,
"logits/chosen": -3.5986835956573486,
"logits/rejected": -3.6389102935791016,
"logps/chosen": -12.681318283081055,
"logps/rejected": -43.688018798828125,
"loss": 0.2724,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4984073638916016,
"rewards/margins": 1.4984073638916016,
"rewards/rejected": 0.0,
"step": 469
},
{
"epoch": 2.6256983240223466,
"grad_norm": 7.880393166220747,
"learning_rate": 9.876088649657802e-07,
"logits/chosen": -4.028558731079102,
"logits/rejected": -3.991490602493286,
"logps/chosen": -11.951882362365723,
"logps/rejected": -16.695070266723633,
"loss": 0.2507,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3863277435302734,
"rewards/margins": 1.3863277435302734,
"rewards/rejected": 0.0,
"step": 470
},
{
"epoch": 2.631284916201117,
"grad_norm": 8.075539955799195,
"learning_rate": 9.874736425588204e-07,
"logits/chosen": -3.785693407058716,
"logits/rejected": -3.8653554916381836,
"logps/chosen": -13.701693534851074,
"logps/rejected": -27.03780746459961,
"loss": 0.2314,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8585710525512695,
"rewards/margins": 1.8585710525512695,
"rewards/rejected": 0.0,
"step": 471
},
{
"epoch": 2.636871508379888,
"grad_norm": 8.101379256216118,
"learning_rate": 9.873376956792136e-07,
"logits/chosen": -3.4866418838500977,
"logits/rejected": -3.533543825149536,
"logps/chosen": -11.929169654846191,
"logps/rejected": -19.965227127075195,
"loss": 0.2392,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2067784070968628,
"rewards/margins": 1.2067784070968628,
"rewards/rejected": 0.0,
"step": 472
},
{
"epoch": 2.642458100558659,
"grad_norm": 7.236270883694466,
"learning_rate": 9.872010245290008e-07,
"logits/chosen": -3.7483625411987305,
"logits/rejected": -3.351423978805542,
"logps/chosen": -16.332185745239258,
"logps/rejected": -20.266708374023438,
"loss": 0.1975,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5048670768737793,
"rewards/margins": 1.5048670768737793,
"rewards/rejected": 0.0,
"step": 473
},
{
"epoch": 2.64804469273743,
"grad_norm": 9.621756868632152,
"learning_rate": 9.870636293112998e-07,
"logits/chosen": -3.9652161598205566,
"logits/rejected": -3.7689316272735596,
"logps/chosen": -17.28591537475586,
"logps/rejected": -25.530614852905273,
"loss": 0.2742,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6187388896942139,
"rewards/margins": 1.6187388896942139,
"rewards/rejected": 0.0,
"step": 474
},
{
"epoch": 2.653631284916201,
"grad_norm": 7.0671066458309975,
"learning_rate": 9.869255102303043e-07,
"logits/chosen": -3.672330617904663,
"logits/rejected": -3.6384193897247314,
"logps/chosen": -20.17595100402832,
"logps/rejected": -28.395429611206055,
"loss": 0.2522,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3444428443908691,
"rewards/margins": 1.3444428443908691,
"rewards/rejected": 0.0,
"step": 475
},
{
"epoch": 2.659217877094972,
"grad_norm": 8.07970111958636,
"learning_rate": 9.86786667491284e-07,
"logits/chosen": -3.524941921234131,
"logits/rejected": -3.5708658695220947,
"logps/chosen": -38.21711349487305,
"logps/rejected": -21.77924919128418,
"loss": 0.2274,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5858030319213867,
"rewards/margins": 1.5858030319213867,
"rewards/rejected": 0.0,
"step": 476
},
{
"epoch": 2.664804469273743,
"grad_norm": 7.784828158908748,
"learning_rate": 9.866471013005837e-07,
"logits/chosen": -3.4317426681518555,
"logits/rejected": -3.399907112121582,
"logps/chosen": -29.516658782958984,
"logps/rejected": -28.625930786132812,
"loss": 0.2304,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.044358253479004,
"rewards/margins": 2.044358253479004,
"rewards/rejected": 0.0,
"step": 477
},
{
"epoch": 2.6703910614525137,
"grad_norm": 8.152705669502572,
"learning_rate": 9.865068118656239e-07,
"logits/chosen": -3.7083640098571777,
"logits/rejected": -3.397965669631958,
"logps/chosen": -12.964552879333496,
"logps/rejected": -51.40794372558594,
"loss": 0.2329,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4145393371582031,
"rewards/margins": 1.4145393371582031,
"rewards/rejected": 0.0,
"step": 478
},
{
"epoch": 2.6759776536312847,
"grad_norm": 8.268904824398946,
"learning_rate": 9.863657993948996e-07,
"logits/chosen": -3.613013982772827,
"logits/rejected": -3.5054068565368652,
"logps/chosen": -5.858638763427734,
"logps/rejected": -18.040542602539062,
"loss": 0.2504,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9335199594497681,
"rewards/margins": 0.9335199594497681,
"rewards/rejected": 0.0,
"step": 479
},
{
"epoch": 2.6815642458100557,
"grad_norm": 7.9052242520574705,
"learning_rate": 9.862240640979801e-07,
"logits/chosen": -3.3727173805236816,
"logits/rejected": -3.4428517818450928,
"logps/chosen": -17.80929946899414,
"logps/rejected": -37.69042205810547,
"loss": 0.2282,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.604065179824829,
"rewards/margins": 1.604065179824829,
"rewards/rejected": 0.0,
"step": 480
},
{
"epoch": 2.6815642458100557,
"eval_logits/chosen": -3.4260826110839844,
"eval_logits/rejected": -3.4980411529541016,
"eval_logps/chosen": -18.827129364013672,
"eval_logps/rejected": -31.69999122619629,
"eval_loss": 0.34813663363456726,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.124291181564331,
"eval_rewards/margins": 1.124291181564331,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5466,
"eval_samples_per_second": 9.241,
"eval_steps_per_second": 0.298,
"step": 480
},
{
"epoch": 2.6871508379888267,
"grad_norm": 7.902342051974376,
"learning_rate": 9.860816061855097e-07,
"logits/chosen": -3.8566508293151855,
"logits/rejected": -3.9841768741607666,
"logps/chosen": -15.78173828125,
"logps/rejected": -19.682714462280273,
"loss": 0.2259,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2836341857910156,
"rewards/margins": 1.2836341857910156,
"rewards/rejected": 0.0,
"step": 481
},
{
"epoch": 2.6927374301675977,
"grad_norm": 8.128367088301074,
"learning_rate": 9.859384258692063e-07,
"logits/chosen": -3.5985052585601807,
"logits/rejected": -3.4468796253204346,
"logps/chosen": -13.781343460083008,
"logps/rejected": -39.90230178833008,
"loss": 0.2144,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5477707386016846,
"rewards/margins": 1.5477707386016846,
"rewards/rejected": 0.0,
"step": 482
},
{
"epoch": 2.6983240223463687,
"grad_norm": 7.945758910520379,
"learning_rate": 9.85794523361861e-07,
"logits/chosen": -3.4471285343170166,
"logits/rejected": -3.401249647140503,
"logps/chosen": -29.81804656982422,
"logps/rejected": -40.78144836425781,
"loss": 0.2225,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9413402080535889,
"rewards/margins": 1.9413402080535889,
"rewards/rejected": 0.0,
"step": 483
},
{
"epoch": 2.7039106145251397,
"grad_norm": 7.482279160486083,
"learning_rate": 9.856498988773388e-07,
"logits/chosen": -3.764979600906372,
"logits/rejected": -3.8511509895324707,
"logps/chosen": -6.9386138916015625,
"logps/rejected": -45.165382385253906,
"loss": 0.2231,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9394733309745789,
"rewards/margins": 0.9394733309745789,
"rewards/rejected": 0.0,
"step": 484
},
{
"epoch": 2.7094972067039107,
"grad_norm": 9.236290688906081,
"learning_rate": 9.855045526305772e-07,
"logits/chosen": -3.6754257678985596,
"logits/rejected": -3.456483840942383,
"logps/chosen": -9.34485149383545,
"logps/rejected": -17.740493774414062,
"loss": 0.2672,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4246737957000732,
"rewards/margins": 1.4246737957000732,
"rewards/rejected": 0.0,
"step": 485
},
{
"epoch": 2.7150837988826817,
"grad_norm": 8.512651763195937,
"learning_rate": 9.85358484837587e-07,
"logits/chosen": -3.905971050262451,
"logits/rejected": -3.810239791870117,
"logps/chosen": -10.29494571685791,
"logps/rejected": -13.215850830078125,
"loss": 0.2433,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.171828269958496,
"rewards/margins": 1.171828269958496,
"rewards/rejected": 0.0,
"step": 486
},
{
"epoch": 2.7206703910614527,
"grad_norm": 7.14735012300244,
"learning_rate": 9.852116957154505e-07,
"logits/chosen": -3.829719066619873,
"logits/rejected": -3.6599392890930176,
"logps/chosen": -14.203679084777832,
"logps/rejected": -27.11825180053711,
"loss": 0.2368,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.285489797592163,
"rewards/margins": 1.285489797592163,
"rewards/rejected": 0.0,
"step": 487
},
{
"epoch": 2.7262569832402237,
"grad_norm": 8.321189811146263,
"learning_rate": 9.850641854823227e-07,
"logits/chosen": -3.6217041015625,
"logits/rejected": -3.5015790462493896,
"logps/chosen": -19.579452514648438,
"logps/rejected": -30.540557861328125,
"loss": 0.2378,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5553691387176514,
"rewards/margins": 1.5553691387176514,
"rewards/rejected": 0.0,
"step": 488
},
{
"epoch": 2.7318435754189943,
"grad_norm": 8.77169091111931,
"learning_rate": 9.8491595435743e-07,
"logits/chosen": -3.7543747425079346,
"logits/rejected": -3.7363133430480957,
"logps/chosen": -11.50475788116455,
"logps/rejected": -21.293922424316406,
"loss": 0.2359,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4948424100875854,
"rewards/margins": 1.4948424100875854,
"rewards/rejected": 0.0,
"step": 489
},
{
"epoch": 2.7374301675977653,
"grad_norm": 7.473231553816813,
"learning_rate": 9.847670025610706e-07,
"logits/chosen": -3.4229557514190674,
"logits/rejected": -3.723675489425659,
"logps/chosen": -11.39542293548584,
"logps/rejected": -31.53961753845215,
"loss": 0.1943,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.520459771156311,
"rewards/margins": 1.520459771156311,
"rewards/rejected": 0.0,
"step": 490
},
{
"epoch": 2.7430167597765363,
"grad_norm": 7.167214974189555,
"learning_rate": 9.84617330314613e-07,
"logits/chosen": -3.6125621795654297,
"logits/rejected": -3.4417295455932617,
"logps/chosen": -9.274349212646484,
"logps/rejected": -23.995391845703125,
"loss": 0.2352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3021409511566162,
"rewards/margins": 1.3021409511566162,
"rewards/rejected": 0.0,
"step": 491
},
{
"epoch": 2.7486033519553073,
"grad_norm": 6.673448358635546,
"learning_rate": 9.84466937840497e-07,
"logits/chosen": -3.8617923259735107,
"logits/rejected": -3.861264944076538,
"logps/chosen": -12.381843566894531,
"logps/rejected": -40.608482360839844,
"loss": 0.2046,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6555776596069336,
"rewards/margins": 1.6555776596069336,
"rewards/rejected": 0.0,
"step": 492
},
{
"epoch": 2.7541899441340782,
"grad_norm": 6.906722025226586,
"learning_rate": 9.843158253622323e-07,
"logits/chosen": -3.7957656383514404,
"logits/rejected": -3.8069732189178467,
"logps/chosen": -11.624693870544434,
"logps/rejected": -24.539522171020508,
"loss": 0.2147,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3188285827636719,
"rewards/margins": 1.3188285827636719,
"rewards/rejected": 0.0,
"step": 493
},
{
"epoch": 2.7597765363128492,
"grad_norm": 7.284973496779757,
"learning_rate": 9.841639931043993e-07,
"logits/chosen": -3.497393846511841,
"logits/rejected": -3.6471757888793945,
"logps/chosen": -6.985918045043945,
"logps/rejected": -34.42020034790039,
"loss": 0.2501,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.8936736583709717,
"rewards/margins": 0.8936736583709717,
"rewards/rejected": 0.0,
"step": 494
},
{
"epoch": 2.7653631284916202,
"grad_norm": 9.564494734379064,
"learning_rate": 9.840114412926477e-07,
"logits/chosen": -3.9058496952056885,
"logits/rejected": -4.01422643661499,
"logps/chosen": -9.518606185913086,
"logps/rejected": -35.89985275268555,
"loss": 0.226,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1451034545898438,
"rewards/margins": 1.1451034545898438,
"rewards/rejected": 0.0,
"step": 495
},
{
"epoch": 2.770949720670391,
"grad_norm": 10.982171123662225,
"learning_rate": 9.838581701536966e-07,
"logits/chosen": -3.650115728378296,
"logits/rejected": -3.7084121704101562,
"logps/chosen": -20.638866424560547,
"logps/rejected": -29.952651977539062,
"loss": 0.2463,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4969639778137207,
"rewards/margins": 2.4969639778137207,
"rewards/rejected": 0.0,
"step": 496
},
{
"epoch": 2.776536312849162,
"grad_norm": 10.535507629208146,
"learning_rate": 9.83704179915334e-07,
"logits/chosen": -3.655679225921631,
"logits/rejected": -3.5606353282928467,
"logps/chosen": -17.765445709228516,
"logps/rejected": -28.804914474487305,
"loss": 0.2594,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9974253177642822,
"rewards/margins": 0.9974253177642822,
"rewards/rejected": 0.0,
"step": 497
},
{
"epoch": 2.782122905027933,
"grad_norm": 7.847733143629526,
"learning_rate": 9.835494708064174e-07,
"logits/chosen": -3.6984760761260986,
"logits/rejected": -3.708005666732788,
"logps/chosen": -8.091940879821777,
"logps/rejected": -45.89476013183594,
"loss": 0.2171,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3567581176757812,
"rewards/margins": 1.3567581176757812,
"rewards/rejected": 0.0,
"step": 498
},
{
"epoch": 2.787709497206704,
"grad_norm": 6.614344245198836,
"learning_rate": 9.833940430568716e-07,
"logits/chosen": -3.3254106044769287,
"logits/rejected": -3.47277569770813,
"logps/chosen": -18.11175537109375,
"logps/rejected": -30.400558471679688,
"loss": 0.1786,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.502833366394043,
"rewards/margins": 1.502833366394043,
"rewards/rejected": 0.0,
"step": 499
},
{
"epoch": 2.793296089385475,
"grad_norm": 7.448538904626172,
"learning_rate": 9.8323789689769e-07,
"logits/chosen": -3.6233277320861816,
"logits/rejected": -3.7375636100769043,
"logps/chosen": -20.592002868652344,
"logps/rejected": -21.011043548583984,
"loss": 0.2073,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9914714097976685,
"rewards/margins": 1.9914714097976685,
"rewards/rejected": 0.0,
"step": 500
},
{
"epoch": 2.793296089385475,
"eval_logits/chosen": -3.4131484031677246,
"eval_logits/rejected": -3.4875328540802,
"eval_logps/chosen": -18.787128448486328,
"eval_logps/rejected": -31.88308334350586,
"eval_loss": 0.3499985337257385,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.128291368484497,
"eval_rewards/margins": 1.128291368484497,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6147,
"eval_samples_per_second": 9.222,
"eval_steps_per_second": 0.297,
"step": 500
},
{
"epoch": 2.798882681564246,
"grad_norm": 7.006173844335316,
"learning_rate": 9.830810325609336e-07,
"logits/chosen": -3.4901397228240967,
"logits/rejected": -3.6741926670074463,
"logps/chosen": -8.288501739501953,
"logps/rejected": -35.86272048950195,
"loss": 0.2157,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1337344646453857,
"rewards/margins": 1.1337344646453857,
"rewards/rejected": 0.0,
"step": 501
},
{
"epoch": 2.804469273743017,
"grad_norm": 6.873407950602945,
"learning_rate": 9.829234502797308e-07,
"logits/chosen": -3.667271137237549,
"logits/rejected": -3.5253448486328125,
"logps/chosen": -8.964853286743164,
"logps/rejected": -29.030786514282227,
"loss": 0.2287,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5016932487487793,
"rewards/margins": 1.5016932487487793,
"rewards/rejected": 0.0,
"step": 502
},
{
"epoch": 2.810055865921788,
"grad_norm": 8.405649575131,
"learning_rate": 9.82765150288277e-07,
"logits/chosen": -3.6840107440948486,
"logits/rejected": -3.177286386489868,
"logps/chosen": -10.823659896850586,
"logps/rejected": -64.39732360839844,
"loss": 0.232,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.265246868133545,
"rewards/margins": 1.265246868133545,
"rewards/rejected": 0.0,
"step": 503
},
{
"epoch": 2.815642458100559,
"grad_norm": 7.593265729493165,
"learning_rate": 9.82606132821834e-07,
"logits/chosen": -3.047468423843384,
"logits/rejected": -3.244701385498047,
"logps/chosen": -47.37431716918945,
"logps/rejected": -28.011714935302734,
"loss": 0.2279,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6064038276672363,
"rewards/margins": 1.6064038276672363,
"rewards/rejected": 0.0,
"step": 504
},
{
"epoch": 2.82122905027933,
"grad_norm": 6.952402357724541,
"learning_rate": 9.824463981167303e-07,
"logits/chosen": -3.7553091049194336,
"logits/rejected": -3.633820056915283,
"logps/chosen": -8.730056762695312,
"logps/rejected": -17.908893585205078,
"loss": 0.2106,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2194769382476807,
"rewards/margins": 1.2194769382476807,
"rewards/rejected": 0.0,
"step": 505
},
{
"epoch": 2.826815642458101,
"grad_norm": 6.967895770582853,
"learning_rate": 9.8228594641036e-07,
"logits/chosen": -3.747936248779297,
"logits/rejected": -3.849527359008789,
"logps/chosen": -10.058979034423828,
"logps/rejected": -21.122617721557617,
"loss": 0.2191,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4362872838974,
"rewards/margins": 1.4362872838974,
"rewards/rejected": 0.0,
"step": 506
},
{
"epoch": 2.8324022346368714,
"grad_norm": 10.638736519793175,
"learning_rate": 9.82124777941183e-07,
"logits/chosen": -3.6542043685913086,
"logits/rejected": -3.6405322551727295,
"logps/chosen": -13.411842346191406,
"logps/rejected": -28.5947265625,
"loss": 0.2546,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1960914134979248,
"rewards/margins": 1.1960914134979248,
"rewards/rejected": 0.0,
"step": 507
},
{
"epoch": 2.8379888268156424,
"grad_norm": 6.862882851062471,
"learning_rate": 9.81962892948724e-07,
"logits/chosen": -3.875190258026123,
"logits/rejected": -3.661947250366211,
"logps/chosen": -18.41823387145996,
"logps/rejected": -29.832141876220703,
"loss": 0.2041,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8571171760559082,
"rewards/margins": 1.8571171760559082,
"rewards/rejected": 0.0,
"step": 508
},
{
"epoch": 2.8435754189944134,
"grad_norm": 8.190086694414964,
"learning_rate": 9.818002916735733e-07,
"logits/chosen": -4.057163715362549,
"logits/rejected": -3.9159181118011475,
"logps/chosen": -11.774065017700195,
"logps/rejected": -21.041200637817383,
"loss": 0.2129,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1493196487426758,
"rewards/margins": 1.1493196487426758,
"rewards/rejected": 0.0,
"step": 509
},
{
"epoch": 2.8491620111731844,
"grad_norm": 8.668457802964342,
"learning_rate": 9.816369743573851e-07,
"logits/chosen": -3.3303356170654297,
"logits/rejected": -3.314500093460083,
"logps/chosen": -42.74287414550781,
"logps/rejected": -19.058147430419922,
"loss": 0.2275,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6401937007904053,
"rewards/margins": 1.6401937007904053,
"rewards/rejected": 0.0,
"step": 510
},
{
"epoch": 2.8547486033519553,
"grad_norm": 7.721684645719408,
"learning_rate": 9.814729412428785e-07,
"logits/chosen": -3.3857314586639404,
"logits/rejected": -3.330643892288208,
"logps/chosen": -20.835432052612305,
"logps/rejected": -30.11691665649414,
"loss": 0.1992,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.359020471572876,
"rewards/margins": 1.359020471572876,
"rewards/rejected": 0.0,
"step": 511
},
{
"epoch": 2.8603351955307263,
"grad_norm": 9.145606388010972,
"learning_rate": 9.813081925738353e-07,
"logits/chosen": -3.6362464427948,
"logits/rejected": -3.593132972717285,
"logps/chosen": -23.955276489257812,
"logps/rejected": -30.369144439697266,
"loss": 0.22,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6455130577087402,
"rewards/margins": 1.6455130577087402,
"rewards/rejected": 0.0,
"step": 512
},
{
"epoch": 2.8659217877094973,
"grad_norm": 9.01464953544514,
"learning_rate": 9.811427285951016e-07,
"logits/chosen": -3.5415940284729004,
"logits/rejected": -3.3448758125305176,
"logps/chosen": -24.666587829589844,
"logps/rejected": -30.633447647094727,
"loss": 0.2234,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7907155752182007,
"rewards/margins": 1.7907155752182007,
"rewards/rejected": 0.0,
"step": 513
},
{
"epoch": 2.871508379888268,
"grad_norm": 6.905190973282463,
"learning_rate": 9.809765495525865e-07,
"logits/chosen": -3.708414316177368,
"logits/rejected": -3.320551872253418,
"logps/chosen": -22.223108291625977,
"logps/rejected": -56.73927688598633,
"loss": 0.2047,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0079169273376465,
"rewards/margins": 2.0079169273376465,
"rewards/rejected": 0.0,
"step": 514
},
{
"epoch": 2.877094972067039,
"grad_norm": 7.27438558863278,
"learning_rate": 9.808096556932615e-07,
"logits/chosen": -3.25369930267334,
"logits/rejected": -3.4669570922851562,
"logps/chosen": -15.14928913116455,
"logps/rejected": -34.39002227783203,
"loss": 0.2087,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2067991495132446,
"rewards/margins": 1.2067991495132446,
"rewards/rejected": 0.0,
"step": 515
},
{
"epoch": 2.88268156424581,
"grad_norm": 8.695252550756638,
"learning_rate": 9.806420472651606e-07,
"logits/chosen": -3.400707960128784,
"logits/rejected": -3.3857288360595703,
"logps/chosen": -19.411767959594727,
"logps/rejected": -33.77351760864258,
"loss": 0.2396,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2417821884155273,
"rewards/margins": 2.2417821884155273,
"rewards/rejected": 0.0,
"step": 516
},
{
"epoch": 2.888268156424581,
"grad_norm": 6.249609765410309,
"learning_rate": 9.8047372451738e-07,
"logits/chosen": -3.5469932556152344,
"logits/rejected": -3.7076032161712646,
"logps/chosen": -34.2308349609375,
"logps/rejected": -22.720203399658203,
"loss": 0.1958,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6457676887512207,
"rewards/margins": 1.6457676887512207,
"rewards/rejected": 0.0,
"step": 517
},
{
"epoch": 2.893854748603352,
"grad_norm": 8.155276171497226,
"learning_rate": 9.803046877000769e-07,
"logits/chosen": -3.419424533843994,
"logits/rejected": -3.2604241371154785,
"logps/chosen": -50.358097076416016,
"logps/rejected": -29.022581100463867,
"loss": 0.2053,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7308439016342163,
"rewards/margins": 1.7308439016342163,
"rewards/rejected": 0.0,
"step": 518
},
{
"epoch": 2.899441340782123,
"grad_norm": 9.309994435464231,
"learning_rate": 9.801349370644702e-07,
"logits/chosen": -3.366185426712036,
"logits/rejected": -3.3842945098876953,
"logps/chosen": -5.38421630859375,
"logps/rejected": -48.389190673828125,
"loss": 0.2357,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.9194574356079102,
"rewards/margins": 0.9194574356079102,
"rewards/rejected": 0.0,
"step": 519
},
{
"epoch": 2.905027932960894,
"grad_norm": 9.390118081596956,
"learning_rate": 9.799644728628399e-07,
"logits/chosen": -3.5186169147491455,
"logits/rejected": -3.3356292247772217,
"logps/chosen": -15.109439849853516,
"logps/rejected": -19.943649291992188,
"loss": 0.2345,
"rewards/accuracies": 1.0,
"rewards/chosen": 0.694863498210907,
"rewards/margins": 0.694863498210907,
"rewards/rejected": 0.0,
"step": 520
},
{
"epoch": 2.905027932960894,
"eval_logits/chosen": -3.4156765937805176,
"eval_logits/rejected": -3.4941775798797607,
"eval_logps/chosen": -18.871572494506836,
"eval_logps/rejected": -31.991893768310547,
"eval_loss": 0.35296905040740967,
"eval_rewards/accuracies": 0.949999988079071,
"eval_rewards/chosen": 1.1198468208312988,
"eval_rewards/margins": 1.1198468208312988,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.604,
"eval_samples_per_second": 9.225,
"eval_steps_per_second": 0.298,
"step": 520
},
{
"epoch": 2.910614525139665,
"grad_norm": 9.195833038191644,
"learning_rate": 9.797932953485255e-07,
"logits/chosen": -3.8919308185577393,
"logits/rejected": -3.7082579135894775,
"logps/chosen": -18.82931900024414,
"logps/rejected": -58.64323425292969,
"loss": 0.2005,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8860766887664795,
"rewards/margins": 1.8860766887664795,
"rewards/rejected": 0.0,
"step": 521
},
{
"epoch": 2.916201117318436,
"grad_norm": 11.431428415303868,
"learning_rate": 9.796214047759282e-07,
"logits/chosen": -3.6113293170928955,
"logits/rejected": -3.451960325241089,
"logps/chosen": -31.122095108032227,
"logps/rejected": -16.40538215637207,
"loss": 0.232,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5730338096618652,
"rewards/margins": 1.5730338096618652,
"rewards/rejected": 0.0,
"step": 522
},
{
"epoch": 2.921787709497207,
"grad_norm": 7.757319062171428,
"learning_rate": 9.79448801400507e-07,
"logits/chosen": -3.6258773803710938,
"logits/rejected": -3.2235469818115234,
"logps/chosen": -16.786638259887695,
"logps/rejected": -26.439407348632812,
"loss": 0.2112,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4600039720535278,
"rewards/margins": 1.4600039720535278,
"rewards/rejected": 0.0,
"step": 523
},
{
"epoch": 2.927374301675978,
"grad_norm": 8.896054053489028,
"learning_rate": 9.792754854787822e-07,
"logits/chosen": -3.507004499435425,
"logits/rejected": -3.4773612022399902,
"logps/chosen": -23.045026779174805,
"logps/rejected": -33.94160079956055,
"loss": 0.2574,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8514593839645386,
"rewards/margins": 1.8514593839645386,
"rewards/rejected": 0.0,
"step": 524
},
{
"epoch": 2.9329608938547485,
"grad_norm": 7.173060291129353,
"learning_rate": 9.791014572683315e-07,
"logits/chosen": -3.7352986335754395,
"logits/rejected": -3.105494499206543,
"logps/chosen": -13.770273208618164,
"logps/rejected": -40.190067291259766,
"loss": 0.2375,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5704180002212524,
"rewards/margins": 1.5704180002212524,
"rewards/rejected": 0.0,
"step": 525
},
{
"epoch": 2.9385474860335195,
"grad_norm": 7.558496228552007,
"learning_rate": 9.78926717027792e-07,
"logits/chosen": -3.6851797103881836,
"logits/rejected": -3.4479734897613525,
"logps/chosen": -16.109195709228516,
"logps/rejected": -34.50910186767578,
"loss": 0.2203,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.467513918876648,
"rewards/margins": 1.467513918876648,
"rewards/rejected": 0.0,
"step": 526
},
{
"epoch": 2.9441340782122905,
"grad_norm": 7.028257607923507,
"learning_rate": 9.78751265016859e-07,
"logits/chosen": -3.689628839492798,
"logits/rejected": -3.438250780105591,
"logps/chosen": -26.680370330810547,
"logps/rejected": -34.16841125488281,
"loss": 0.1758,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5826314687728882,
"rewards/margins": 1.5826314687728882,
"rewards/rejected": 0.0,
"step": 527
},
{
"epoch": 2.9497206703910615,
"grad_norm": 7.452248237330729,
"learning_rate": 9.785751014962851e-07,
"logits/chosen": -3.894042730331421,
"logits/rejected": -3.8289926052093506,
"logps/chosen": -10.17105770111084,
"logps/rejected": -21.5932559967041,
"loss": 0.1969,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5189507007598877,
"rewards/margins": 1.5189507007598877,
"rewards/rejected": 0.0,
"step": 528
},
{
"epoch": 2.9553072625698324,
"grad_norm": 7.606588657029743,
"learning_rate": 9.783982267278808e-07,
"logits/chosen": -3.688990831375122,
"logits/rejected": -3.6647708415985107,
"logps/chosen": -24.94097900390625,
"logps/rejected": -26.211376190185547,
"loss": 0.2151,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5228779315948486,
"rewards/margins": 1.5228779315948486,
"rewards/rejected": 0.0,
"step": 529
},
{
"epoch": 2.9608938547486034,
"grad_norm": 7.300424464888586,
"learning_rate": 9.782206409745134e-07,
"logits/chosen": -3.7147138118743896,
"logits/rejected": -3.255930185317993,
"logps/chosen": -8.188423156738281,
"logps/rejected": -33.88630294799805,
"loss": 0.2095,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.0383009910583496,
"rewards/margins": 1.0383009910583496,
"rewards/rejected": 0.0,
"step": 530
},
{
"epoch": 2.9664804469273744,
"grad_norm": 7.532588767917452,
"learning_rate": 9.780423445001072e-07,
"logits/chosen": -3.4041967391967773,
"logits/rejected": -3.420712947845459,
"logps/chosen": -43.992828369140625,
"logps/rejected": -31.57377052307129,
"loss": 0.2052,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2596793174743652,
"rewards/margins": 2.2596793174743652,
"rewards/rejected": 0.0,
"step": 531
},
{
"epoch": 2.972067039106145,
"grad_norm": 9.639239919983133,
"learning_rate": 9.778633375696424e-07,
"logits/chosen": -3.5303702354431152,
"logits/rejected": -3.5252037048339844,
"logps/chosen": -17.531352996826172,
"logps/rejected": -27.08810043334961,
"loss": 0.2215,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5886486768722534,
"rewards/margins": 1.5886486768722534,
"rewards/rejected": 0.0,
"step": 532
},
{
"epoch": 2.977653631284916,
"grad_norm": 7.825710810550395,
"learning_rate": 9.776836204491552e-07,
"logits/chosen": -3.514402151107788,
"logits/rejected": -3.4888627529144287,
"logps/chosen": -28.722347259521484,
"logps/rejected": -41.736698150634766,
"loss": 0.2249,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1408169269561768,
"rewards/margins": 1.1408169269561768,
"rewards/rejected": 0.0,
"step": 533
},
{
"epoch": 2.983240223463687,
"grad_norm": 7.344748276963433,
"learning_rate": 9.775031934057373e-07,
"logits/chosen": -3.348198652267456,
"logits/rejected": -3.621284246444702,
"logps/chosen": -15.124707221984863,
"logps/rejected": -28.8204345703125,
"loss": 0.2101,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4255895614624023,
"rewards/margins": 1.4255895614624023,
"rewards/rejected": 0.0,
"step": 534
},
{
"epoch": 2.988826815642458,
"grad_norm": 6.686215734213654,
"learning_rate": 9.77322056707535e-07,
"logits/chosen": -3.156296968460083,
"logits/rejected": -3.096890449523926,
"logps/chosen": -22.622802734375,
"logps/rejected": -37.47029113769531,
"loss": 0.1905,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8413368463516235,
"rewards/margins": 1.8413368463516235,
"rewards/rejected": 0.0,
"step": 535
},
{
"epoch": 2.994413407821229,
"grad_norm": 8.883809445571819,
"learning_rate": 9.771402106237503e-07,
"logits/chosen": -3.5562469959259033,
"logits/rejected": -3.5949959754943848,
"logps/chosen": -14.242838859558105,
"logps/rejected": -36.978660583496094,
"loss": 0.2287,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5399631261825562,
"rewards/margins": 1.5399631261825562,
"rewards/rejected": 0.0,
"step": 536
},
{
"epoch": 3.0,
"grad_norm": 5.951618447991793,
"learning_rate": 9.769576554246383e-07,
"logits/chosen": -3.464947462081909,
"logits/rejected": -3.4362552165985107,
"logps/chosen": -14.894698143005371,
"logps/rejected": -22.000164031982422,
"loss": 0.2003,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3502216339111328,
"rewards/margins": 1.3502216339111328,
"rewards/rejected": 0.0,
"step": 537
},
{
"epoch": 3.005586592178771,
"grad_norm": 4.245095056572733,
"learning_rate": 9.767743913815092e-07,
"logits/chosen": -3.527196168899536,
"logits/rejected": -3.772918701171875,
"logps/chosen": -9.334918975830078,
"logps/rejected": -30.373565673828125,
"loss": 0.183,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5863807201385498,
"rewards/margins": 1.5863807201385498,
"rewards/rejected": 0.0,
"step": 538
},
{
"epoch": 3.011173184357542,
"grad_norm": 5.1840999250382245,
"learning_rate": 9.765904187667254e-07,
"logits/chosen": -3.642043113708496,
"logits/rejected": -3.252462863922119,
"logps/chosen": -6.1675238609313965,
"logps/rejected": -28.275943756103516,
"loss": 0.2001,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4355202913284302,
"rewards/margins": 1.4355202913284302,
"rewards/rejected": 0.0,
"step": 539
},
{
"epoch": 3.016759776536313,
"grad_norm": 4.85681935133445,
"learning_rate": 9.764057378537034e-07,
"logits/chosen": -3.1955578327178955,
"logits/rejected": -3.1691908836364746,
"logps/chosen": -43.302391052246094,
"logps/rejected": -48.15077590942383,
"loss": 0.1987,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.09118390083313,
"rewards/margins": 2.09118390083313,
"rewards/rejected": 0.0,
"step": 540
},
{
"epoch": 3.016759776536313,
"eval_logits/chosen": -3.426398754119873,
"eval_logits/rejected": -3.50640869140625,
"eval_logps/chosen": -18.674394607543945,
"eval_logps/rejected": -32.07451629638672,
"eval_loss": 0.3496454656124115,
"eval_rewards/accuracies": 0.925000011920929,
"eval_rewards/chosen": 1.1395649909973145,
"eval_rewards/margins": 1.1395649909973145,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5941,
"eval_samples_per_second": 9.228,
"eval_steps_per_second": 0.298,
"step": 540
},
{
"epoch": 3.022346368715084,
"grad_norm": 4.687407332922764,
"learning_rate": 9.762203489169115e-07,
"logits/chosen": -3.765223264694214,
"logits/rejected": -3.6322054862976074,
"logps/chosen": -10.266185760498047,
"logps/rejected": -21.843515396118164,
"loss": 0.182,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8560717105865479,
"rewards/margins": 1.8560717105865479,
"rewards/rejected": 0.0,
"step": 541
},
{
"epoch": 3.0279329608938546,
"grad_norm": 4.518162104408857,
"learning_rate": 9.76034252231871e-07,
"logits/chosen": -3.720940589904785,
"logits/rejected": -3.8452420234680176,
"logps/chosen": -19.095359802246094,
"logps/rejected": -28.30400276184082,
"loss": 0.1657,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2671868801116943,
"rewards/margins": 2.2671868801116943,
"rewards/rejected": 0.0,
"step": 542
},
{
"epoch": 3.0335195530726256,
"grad_norm": 5.579688457027353,
"learning_rate": 9.758474480751547e-07,
"logits/chosen": -3.8501555919647217,
"logits/rejected": -3.6848278045654297,
"logps/chosen": -13.278109550476074,
"logps/rejected": -39.80680465698242,
"loss": 0.1818,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9774212837219238,
"rewards/margins": 1.9774212837219238,
"rewards/rejected": 0.0,
"step": 543
},
{
"epoch": 3.0391061452513966,
"grad_norm": 4.411804230080304,
"learning_rate": 9.756599367243868e-07,
"logits/chosen": -3.7062551975250244,
"logits/rejected": -3.4251039028167725,
"logps/chosen": -15.456059455871582,
"logps/rejected": -33.30834197998047,
"loss": 0.1826,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.043917179107666,
"rewards/margins": 2.043917179107666,
"rewards/rejected": 0.0,
"step": 544
},
{
"epoch": 3.0446927374301676,
"grad_norm": 4.009334068053905,
"learning_rate": 9.754717184582422e-07,
"logits/chosen": -2.9957168102264404,
"logits/rejected": -2.999037981033325,
"logps/chosen": -33.969444274902344,
"logps/rejected": -48.57637023925781,
"loss": 0.1799,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5263586044311523,
"rewards/margins": 2.5263586044311523,
"rewards/rejected": 0.0,
"step": 545
},
{
"epoch": 3.0502793296089385,
"grad_norm": 4.7000551597518125,
"learning_rate": 9.752827935564474e-07,
"logits/chosen": -3.714012622833252,
"logits/rejected": -3.739675521850586,
"logps/chosen": -32.73251724243164,
"logps/rejected": -31.016817092895508,
"loss": 0.1843,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6957581043243408,
"rewards/margins": 1.6957581043243408,
"rewards/rejected": 0.0,
"step": 546
},
{
"epoch": 3.0558659217877095,
"grad_norm": 5.036818270344218,
"learning_rate": 9.75093162299778e-07,
"logits/chosen": -3.67802095413208,
"logits/rejected": -3.584538459777832,
"logps/chosen": -31.94599151611328,
"logps/rejected": -31.759218215942383,
"loss": 0.1651,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.74308443069458,
"rewards/margins": 2.74308443069458,
"rewards/rejected": 0.0,
"step": 547
},
{
"epoch": 3.0614525139664805,
"grad_norm": 4.3369665011037375,
"learning_rate": 9.7490282497006e-07,
"logits/chosen": -3.5875184535980225,
"logits/rejected": -3.5783565044403076,
"logps/chosen": -15.221088409423828,
"logps/rejected": -18.263507843017578,
"loss": 0.1478,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2133846282958984,
"rewards/margins": 2.2133846282958984,
"rewards/rejected": 0.0,
"step": 548
},
{
"epoch": 3.0670391061452515,
"grad_norm": 4.64409448625801,
"learning_rate": 9.747117818501687e-07,
"logits/chosen": -3.67057204246521,
"logits/rejected": -3.667259931564331,
"logps/chosen": -5.813396453857422,
"logps/rejected": -16.091121673583984,
"loss": 0.1681,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5230515003204346,
"rewards/margins": 1.5230515003204346,
"rewards/rejected": 0.0,
"step": 549
},
{
"epoch": 3.0726256983240225,
"grad_norm": 5.647539059115093,
"learning_rate": 9.745200332240278e-07,
"logits/chosen": -3.5769731998443604,
"logits/rejected": -3.6055614948272705,
"logps/chosen": -15.150192260742188,
"logps/rejected": -37.067344665527344,
"loss": 0.1799,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.007539749145508,
"rewards/margins": 2.007539749145508,
"rewards/rejected": 0.0,
"step": 550
},
{
"epoch": 3.078212290502793,
"grad_norm": 5.14672413289712,
"learning_rate": 9.743275793766102e-07,
"logits/chosen": -3.8731751441955566,
"logits/rejected": -3.747708320617676,
"logps/chosen": -26.233776092529297,
"logps/rejected": -35.369693756103516,
"loss": 0.1869,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7416244745254517,
"rewards/margins": 1.7416244745254517,
"rewards/rejected": 0.0,
"step": 551
},
{
"epoch": 3.083798882681564,
"grad_norm": 4.849462299874129,
"learning_rate": 9.741344205939364e-07,
"logits/chosen": -3.7174947261810303,
"logits/rejected": -3.245631217956543,
"logps/chosen": -7.081581115722656,
"logps/rejected": -37.488609313964844,
"loss": 0.1541,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5629022121429443,
"rewards/margins": 1.5629022121429443,
"rewards/rejected": 0.0,
"step": 552
},
{
"epoch": 3.089385474860335,
"grad_norm": 4.209133371240973,
"learning_rate": 9.739405571630752e-07,
"logits/chosen": -3.8818559646606445,
"logits/rejected": -3.9326014518737793,
"logps/chosen": -13.727590560913086,
"logps/rejected": -22.858854293823242,
"loss": 0.1753,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7776994705200195,
"rewards/margins": 1.7776994705200195,
"rewards/rejected": 0.0,
"step": 553
},
{
"epoch": 3.094972067039106,
"grad_norm": 4.999596680859089,
"learning_rate": 9.737459893721415e-07,
"logits/chosen": -3.8012993335723877,
"logits/rejected": -3.6896021366119385,
"logps/chosen": -3.1425883769989014,
"logps/rejected": -53.1058464050293,
"loss": 0.1771,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3161832094192505,
"rewards/margins": 1.3161832094192505,
"rewards/rejected": 0.0,
"step": 554
},
{
"epoch": 3.100558659217877,
"grad_norm": 4.611992849402734,
"learning_rate": 9.735507175102983e-07,
"logits/chosen": -3.735285520553589,
"logits/rejected": -3.5629961490631104,
"logps/chosen": -7.803858757019043,
"logps/rejected": -42.674530029296875,
"loss": 0.1612,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3933154344558716,
"rewards/margins": 1.3933154344558716,
"rewards/rejected": 0.0,
"step": 555
},
{
"epoch": 3.106145251396648,
"grad_norm": 4.925793425741922,
"learning_rate": 9.733547418677538e-07,
"logits/chosen": -3.6361725330352783,
"logits/rejected": -3.8588008880615234,
"logps/chosen": -10.72397232055664,
"logps/rejected": -59.879058837890625,
"loss": 0.1996,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8188023567199707,
"rewards/margins": 1.8188023567199707,
"rewards/rejected": 0.0,
"step": 556
},
{
"epoch": 3.111731843575419,
"grad_norm": 5.475225802231006,
"learning_rate": 9.731580627357635e-07,
"logits/chosen": -3.7604262828826904,
"logits/rejected": -3.319699287414551,
"logps/chosen": -7.3406982421875,
"logps/rejected": -58.49966049194336,
"loss": 0.2007,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6027517318725586,
"rewards/margins": 1.6027517318725586,
"rewards/rejected": 0.0,
"step": 557
},
{
"epoch": 3.11731843575419,
"grad_norm": 4.809392527499902,
"learning_rate": 9.729606804066268e-07,
"logits/chosen": -3.924943208694458,
"logits/rejected": -3.829216241836548,
"logps/chosen": -7.529721260070801,
"logps/rejected": -54.85078811645508,
"loss": 0.1983,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4421954154968262,
"rewards/margins": 1.4421954154968262,
"rewards/rejected": 0.0,
"step": 558
},
{
"epoch": 3.122905027932961,
"grad_norm": 4.345413799217623,
"learning_rate": 9.727625951736894e-07,
"logits/chosen": -3.5260565280914307,
"logits/rejected": -3.7493531703948975,
"logps/chosen": -3.8053598403930664,
"logps/rejected": -39.421329498291016,
"loss": 0.1672,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2054493427276611,
"rewards/margins": 1.2054493427276611,
"rewards/rejected": 0.0,
"step": 559
},
{
"epoch": 3.1284916201117317,
"grad_norm": 5.22143163482338,
"learning_rate": 9.725638073313412e-07,
"logits/chosen": -2.7598049640655518,
"logits/rejected": -2.6982693672180176,
"logps/chosen": -28.23615264892578,
"logps/rejected": -41.56056213378906,
"loss": 0.1838,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8444175720214844,
"rewards/margins": 1.8444175720214844,
"rewards/rejected": 0.0,
"step": 560
},
{
"epoch": 3.1284916201117317,
"eval_logits/chosen": -3.452139377593994,
"eval_logits/rejected": -3.551227569580078,
"eval_logps/chosen": -19.524335861206055,
"eval_logps/rejected": -34.0481071472168,
"eval_loss": 0.38711562752723694,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.0545704364776611,
"eval_rewards/margins": 1.0545704364776611,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.6012,
"eval_samples_per_second": 9.226,
"eval_steps_per_second": 0.298,
"step": 560
},
{
"epoch": 3.1340782122905027,
"grad_norm": 4.742808432195938,
"learning_rate": 9.723643171750161e-07,
"logits/chosen": -3.5465304851531982,
"logits/rejected": -3.5002918243408203,
"logps/chosen": -3.8645851612091064,
"logps/rejected": -37.22589111328125,
"loss": 0.1957,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2577381134033203,
"rewards/margins": 1.2577381134033203,
"rewards/rejected": 0.0,
"step": 561
},
{
"epoch": 3.1396648044692737,
"grad_norm": 4.531387635169225,
"learning_rate": 9.721641250011926e-07,
"logits/chosen": -3.4351985454559326,
"logits/rejected": -3.562939167022705,
"logps/chosen": -25.417177200317383,
"logps/rejected": -27.687837600708008,
"loss": 0.1714,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8707070350646973,
"rewards/margins": 1.8707070350646973,
"rewards/rejected": 0.0,
"step": 562
},
{
"epoch": 3.1452513966480447,
"grad_norm": 5.785023128231763,
"learning_rate": 9.719632311073912e-07,
"logits/chosen": -3.410985231399536,
"logits/rejected": -3.544358730316162,
"logps/chosen": -7.332213401794434,
"logps/rejected": -29.132205963134766,
"loss": 0.1919,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.501019835472107,
"rewards/margins": 1.501019835472107,
"rewards/rejected": 0.0,
"step": 563
},
{
"epoch": 3.1508379888268156,
"grad_norm": 4.816389939776029,
"learning_rate": 9.717616357921764e-07,
"logits/chosen": -3.499091386795044,
"logits/rejected": -3.3932666778564453,
"logps/chosen": -9.977988243103027,
"logps/rejected": -31.942049026489258,
"loss": 0.2038,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4507946968078613,
"rewards/margins": 1.4507946968078613,
"rewards/rejected": 0.0,
"step": 564
},
{
"epoch": 3.1564245810055866,
"grad_norm": 5.080691874623269,
"learning_rate": 9.715593393551545e-07,
"logits/chosen": -3.698702096939087,
"logits/rejected": -3.818716287612915,
"logps/chosen": -6.350429534912109,
"logps/rejected": -32.38847732543945,
"loss": 0.1626,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4038186073303223,
"rewards/margins": 1.4038186073303223,
"rewards/rejected": 0.0,
"step": 565
},
{
"epoch": 3.1620111731843576,
"grad_norm": 4.900675475627213,
"learning_rate": 9.713563420969744e-07,
"logits/chosen": -3.4896557331085205,
"logits/rejected": -3.595815420150757,
"logps/chosen": -7.538320541381836,
"logps/rejected": -51.425331115722656,
"loss": 0.1624,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9275681972503662,
"rewards/margins": 1.9275681972503662,
"rewards/rejected": 0.0,
"step": 566
},
{
"epoch": 3.1675977653631286,
"grad_norm": 4.772537916253512,
"learning_rate": 9.711526443193258e-07,
"logits/chosen": -3.5935449600219727,
"logits/rejected": -3.7333123683929443,
"logps/chosen": -14.802519798278809,
"logps/rejected": -22.667762756347656,
"loss": 0.1506,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9541294574737549,
"rewards/margins": 1.9541294574737549,
"rewards/rejected": 0.0,
"step": 567
},
{
"epoch": 3.1731843575418996,
"grad_norm": 4.410642601758276,
"learning_rate": 9.709482463249402e-07,
"logits/chosen": -3.8915655612945557,
"logits/rejected": -3.8851208686828613,
"logps/chosen": -4.589975357055664,
"logps/rejected": -21.963491439819336,
"loss": 0.1891,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3541920185089111,
"rewards/margins": 1.3541920185089111,
"rewards/rejected": 0.0,
"step": 568
},
{
"epoch": 3.17877094972067,
"grad_norm": 5.001923513173423,
"learning_rate": 9.707431484175893e-07,
"logits/chosen": -3.3935484886169434,
"logits/rejected": -3.537262439727783,
"logps/chosen": -25.665592193603516,
"logps/rejected": -27.040699005126953,
"loss": 0.1691,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2004714012145996,
"rewards/margins": 2.2004714012145996,
"rewards/rejected": 0.0,
"step": 569
},
{
"epoch": 3.184357541899441,
"grad_norm": 6.779016548131147,
"learning_rate": 9.705373509020848e-07,
"logits/chosen": -3.615057945251465,
"logits/rejected": -3.2200875282287598,
"logps/chosen": -6.432480335235596,
"logps/rejected": -58.13328170776367,
"loss": 0.1804,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4290056228637695,
"rewards/margins": 1.4290056228637695,
"rewards/rejected": 0.0,
"step": 570
},
{
"epoch": 3.189944134078212,
"grad_norm": 4.736026679285262,
"learning_rate": 9.70330854084279e-07,
"logits/chosen": -3.712716817855835,
"logits/rejected": -3.572997570037842,
"logps/chosen": -23.108036041259766,
"logps/rejected": -22.306825637817383,
"loss": 0.1618,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.165341854095459,
"rewards/margins": 2.165341854095459,
"rewards/rejected": 0.0,
"step": 571
},
{
"epoch": 3.195530726256983,
"grad_norm": 5.626284254757533,
"learning_rate": 9.701236582710627e-07,
"logits/chosen": -3.353090763092041,
"logits/rejected": -3.665696859359741,
"logps/chosen": -27.17815399169922,
"logps/rejected": -34.68605041503906,
"loss": 0.1495,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3534014225006104,
"rewards/margins": 2.3534014225006104,
"rewards/rejected": 0.0,
"step": 572
},
{
"epoch": 3.201117318435754,
"grad_norm": 4.870352035036208,
"learning_rate": 9.699157637703656e-07,
"logits/chosen": -3.9493651390075684,
"logits/rejected": -3.839723825454712,
"logps/chosen": -7.73994255065918,
"logps/rejected": -40.916465759277344,
"loss": 0.149,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5562421083450317,
"rewards/margins": 1.5562421083450317,
"rewards/rejected": 0.0,
"step": 573
},
{
"epoch": 3.206703910614525,
"grad_norm": 5.277463053572278,
"learning_rate": 9.697071708911563e-07,
"logits/chosen": -3.6510212421417236,
"logits/rejected": -3.637693166732788,
"logps/chosen": -8.033740997314453,
"logps/rejected": -30.34661293029785,
"loss": 0.1869,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5631598234176636,
"rewards/margins": 1.5631598234176636,
"rewards/rejected": 0.0,
"step": 574
},
{
"epoch": 3.212290502793296,
"grad_norm": 5.264094144924965,
"learning_rate": 9.694978799434408e-07,
"logits/chosen": -3.5529937744140625,
"logits/rejected": -3.5506813526153564,
"logps/chosen": -4.875071048736572,
"logps/rejected": -36.16624450683594,
"loss": 0.1843,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1266826391220093,
"rewards/margins": 1.1266826391220093,
"rewards/rejected": 0.0,
"step": 575
},
{
"epoch": 3.217877094972067,
"grad_norm": 5.485669842712066,
"learning_rate": 9.692878912382625e-07,
"logits/chosen": -3.9466843605041504,
"logits/rejected": -3.9598653316497803,
"logps/chosen": -6.40760612487793,
"logps/rejected": -22.46565055847168,
"loss": 0.1884,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5205612182617188,
"rewards/margins": 1.5205612182617188,
"rewards/rejected": 0.0,
"step": 576
},
{
"epoch": 3.223463687150838,
"grad_norm": 4.98802706808277,
"learning_rate": 9.690772050877024e-07,
"logits/chosen": -3.943087339401245,
"logits/rejected": -3.983419418334961,
"logps/chosen": -44.90244674682617,
"logps/rejected": -22.856143951416016,
"loss": 0.1743,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.329369306564331,
"rewards/margins": 2.329369306564331,
"rewards/rejected": 0.0,
"step": 577
},
{
"epoch": 3.2290502793296088,
"grad_norm": 4.842990265689181,
"learning_rate": 9.688658218048773e-07,
"logits/chosen": -3.3331680297851562,
"logits/rejected": -3.435561418533325,
"logps/chosen": -51.9761848449707,
"logps/rejected": -41.16301727294922,
"loss": 0.1442,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0154778957366943,
"rewards/margins": 2.0154778957366943,
"rewards/rejected": 0.0,
"step": 578
},
{
"epoch": 3.2346368715083798,
"grad_norm": 5.312142250461408,
"learning_rate": 9.686537417039407e-07,
"logits/chosen": -3.582887887954712,
"logits/rejected": -3.7412376403808594,
"logps/chosen": -11.86709213256836,
"logps/rejected": -42.311580657958984,
"loss": 0.1803,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5465190410614014,
"rewards/margins": 1.5465190410614014,
"rewards/rejected": 0.0,
"step": 579
},
{
"epoch": 3.2402234636871508,
"grad_norm": 9.410288475259977,
"learning_rate": 9.684409651000811e-07,
"logits/chosen": -3.9253406524658203,
"logits/rejected": -3.460865020751953,
"logps/chosen": -10.171557426452637,
"logps/rejected": -44.976715087890625,
"loss": 0.1988,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9322619438171387,
"rewards/margins": 1.9322619438171387,
"rewards/rejected": 0.0,
"step": 580
},
{
"epoch": 3.2402234636871508,
"eval_logits/chosen": -3.462258815765381,
"eval_logits/rejected": -3.5622501373291016,
"eval_logps/chosen": -19.60980224609375,
"eval_logps/rejected": -33.709877014160156,
"eval_loss": 0.3881343901157379,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.0460238456726074,
"eval_rewards/margins": 1.0460238456726074,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.583,
"eval_samples_per_second": 9.231,
"eval_steps_per_second": 0.298,
"step": 580
},
{
"epoch": 3.2458100558659218,
"grad_norm": 6.161358425089739,
"learning_rate": 9.682274923095229e-07,
"logits/chosen": -3.1356165409088135,
"logits/rejected": -3.395218849182129,
"logps/chosen": -31.436750411987305,
"logps/rejected": -31.18406867980957,
"loss": 0.2037,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7328619956970215,
"rewards/margins": 2.7328619956970215,
"rewards/rejected": 0.0,
"step": 581
},
{
"epoch": 3.2513966480446927,
"grad_norm": 4.578745281068452,
"learning_rate": 9.68013323649524e-07,
"logits/chosen": -3.405984878540039,
"logits/rejected": -3.7583210468292236,
"logps/chosen": -7.036988258361816,
"logps/rejected": -24.276416778564453,
"loss": 0.1941,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6983870267868042,
"rewards/margins": 1.6983870267868042,
"rewards/rejected": 0.0,
"step": 582
},
{
"epoch": 3.2569832402234637,
"grad_norm": 5.9174722341795825,
"learning_rate": 9.677984594383778e-07,
"logits/chosen": -3.7339975833892822,
"logits/rejected": -3.9361507892608643,
"logps/chosen": -17.876081466674805,
"logps/rejected": -24.841102600097656,
"loss": 0.1399,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7446861267089844,
"rewards/margins": 1.7446861267089844,
"rewards/rejected": 0.0,
"step": 583
},
{
"epoch": 3.2625698324022347,
"grad_norm": 6.793699025911994,
"learning_rate": 9.675828999954104e-07,
"logits/chosen": -3.9633615016937256,
"logits/rejected": -3.6193535327911377,
"logps/chosen": -9.287819862365723,
"logps/rejected": -38.18955993652344,
"loss": 0.1919,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.739070177078247,
"rewards/margins": 1.739070177078247,
"rewards/rejected": 0.0,
"step": 584
},
{
"epoch": 3.2681564245810057,
"grad_norm": 5.075795557710363,
"learning_rate": 9.673666456409817e-07,
"logits/chosen": -3.2759780883789062,
"logits/rejected": -3.0620763301849365,
"logps/chosen": -22.58893585205078,
"logps/rejected": -75.91329956054688,
"loss": 0.1507,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.643172264099121,
"rewards/margins": 1.643172264099121,
"rewards/rejected": 0.0,
"step": 585
},
{
"epoch": 3.2737430167597763,
"grad_norm": 4.83784737013629,
"learning_rate": 9.671496966964842e-07,
"logits/chosen": -3.522770881652832,
"logits/rejected": -3.5017261505126953,
"logps/chosen": -24.941234588623047,
"logps/rejected": -32.05644989013672,
"loss": 0.1712,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4505114555358887,
"rewards/margins": 2.4505114555358887,
"rewards/rejected": 0.0,
"step": 586
},
{
"epoch": 3.2793296089385473,
"grad_norm": 4.962956416097082,
"learning_rate": 9.669320534843425e-07,
"logits/chosen": -3.6959798336029053,
"logits/rejected": -3.7863824367523193,
"logps/chosen": -23.056482315063477,
"logps/rejected": -43.51755142211914,
"loss": 0.1876,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9795132875442505,
"rewards/margins": 1.9795132875442505,
"rewards/rejected": 0.0,
"step": 587
},
{
"epoch": 3.2849162011173183,
"grad_norm": 5.535105657873138,
"learning_rate": 9.667137163280133e-07,
"logits/chosen": -3.588747024536133,
"logits/rejected": -3.5322866439819336,
"logps/chosen": -24.421964645385742,
"logps/rejected": -36.126529693603516,
"loss": 0.1672,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2174808979034424,
"rewards/margins": 2.2174808979034424,
"rewards/rejected": 0.0,
"step": 588
},
{
"epoch": 3.2905027932960893,
"grad_norm": 4.949491208974845,
"learning_rate": 9.664946855519843e-07,
"logits/chosen": -3.6174607276916504,
"logits/rejected": -3.545384407043457,
"logps/chosen": -10.511805534362793,
"logps/rejected": -33.92623519897461,
"loss": 0.1705,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7961182594299316,
"rewards/margins": 1.7961182594299316,
"rewards/rejected": 0.0,
"step": 589
},
{
"epoch": 3.2960893854748603,
"grad_norm": 4.832052588159931,
"learning_rate": 9.662749614817746e-07,
"logits/chosen": -3.5949666500091553,
"logits/rejected": -3.5336391925811768,
"logps/chosen": -16.682287216186523,
"logps/rejected": -39.20673370361328,
"loss": 0.1686,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7638062238693237,
"rewards/margins": 1.7638062238693237,
"rewards/rejected": 0.0,
"step": 590
},
{
"epoch": 3.3016759776536313,
"grad_norm": 4.374631165657219,
"learning_rate": 9.660545444439332e-07,
"logits/chosen": -3.721179485321045,
"logits/rejected": -3.7360987663269043,
"logps/chosen": -8.91193962097168,
"logps/rejected": -36.664329528808594,
"loss": 0.1618,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7523739337921143,
"rewards/margins": 1.7523739337921143,
"rewards/rejected": 0.0,
"step": 591
},
{
"epoch": 3.3072625698324023,
"grad_norm": 6.306868741111923,
"learning_rate": 9.658334347660389e-07,
"logits/chosen": -2.937652349472046,
"logits/rejected": -3.0750551223754883,
"logps/chosen": -18.957584381103516,
"logps/rejected": -55.87245178222656,
"loss": 0.1598,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6617270708084106,
"rewards/margins": 1.6617270708084106,
"rewards/rejected": 0.0,
"step": 592
},
{
"epoch": 3.3128491620111733,
"grad_norm": 6.186761405552854,
"learning_rate": 9.656116327767002e-07,
"logits/chosen": -3.8097431659698486,
"logits/rejected": -3.948094129562378,
"logps/chosen": -19.085521697998047,
"logps/rejected": -32.111228942871094,
"loss": 0.1701,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1641557216644287,
"rewards/margins": 2.1641557216644287,
"rewards/rejected": 0.0,
"step": 593
},
{
"epoch": 3.3184357541899443,
"grad_norm": 4.4902103816695,
"learning_rate": 9.65389138805554e-07,
"logits/chosen": -3.5979137420654297,
"logits/rejected": -3.61662220954895,
"logps/chosen": -7.313446998596191,
"logps/rejected": -33.970970153808594,
"loss": 0.1669,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4582792520523071,
"rewards/margins": 1.4582792520523071,
"rewards/rejected": 0.0,
"step": 594
},
{
"epoch": 3.3240223463687153,
"grad_norm": 5.527789456110965,
"learning_rate": 9.651659531832665e-07,
"logits/chosen": -3.4975526332855225,
"logits/rejected": -3.5394434928894043,
"logps/chosen": -26.84719467163086,
"logps/rejected": -37.31050109863281,
"loss": 0.1295,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6903393268585205,
"rewards/margins": 2.6903393268585205,
"rewards/rejected": 0.0,
"step": 595
},
{
"epoch": 3.329608938547486,
"grad_norm": 6.680730860628917,
"learning_rate": 9.64942076241531e-07,
"logits/chosen": -3.551564931869507,
"logits/rejected": -3.630307674407959,
"logps/chosen": -13.276336669921875,
"logps/rejected": -29.829763412475586,
"loss": 0.1894,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.529152750968933,
"rewards/margins": 1.529152750968933,
"rewards/rejected": 0.0,
"step": 596
},
{
"epoch": 3.335195530726257,
"grad_norm": 4.386712935890993,
"learning_rate": 9.647175083130687e-07,
"logits/chosen": -3.4997851848602295,
"logits/rejected": -3.688692331314087,
"logps/chosen": -3.3455991744995117,
"logps/rejected": -41.719268798828125,
"loss": 0.1659,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3428454399108887,
"rewards/margins": 1.3428454399108887,
"rewards/rejected": 0.0,
"step": 597
},
{
"epoch": 3.340782122905028,
"grad_norm": 5.648244605274734,
"learning_rate": 9.644922497316272e-07,
"logits/chosen": -3.492734670639038,
"logits/rejected": -3.720151662826538,
"logps/chosen": -14.767166137695312,
"logps/rejected": -77.58329010009766,
"loss": 0.1865,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.538488507270813,
"rewards/margins": 1.538488507270813,
"rewards/rejected": 0.0,
"step": 598
},
{
"epoch": 3.346368715083799,
"grad_norm": 4.483493491856391,
"learning_rate": 9.642663008319811e-07,
"logits/chosen": -3.735584020614624,
"logits/rejected": -3.4455156326293945,
"logps/chosen": -8.765111923217773,
"logps/rejected": -49.826454162597656,
"loss": 0.163,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.848058819770813,
"rewards/margins": 1.848058819770813,
"rewards/rejected": 0.0,
"step": 599
},
{
"epoch": 3.35195530726257,
"grad_norm": 4.781236911208581,
"learning_rate": 9.640396619499307e-07,
"logits/chosen": -3.654815435409546,
"logits/rejected": -3.8434436321258545,
"logps/chosen": -5.785635948181152,
"logps/rejected": -44.263404846191406,
"loss": 0.1733,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4081251621246338,
"rewards/margins": 1.4081251621246338,
"rewards/rejected": 0.0,
"step": 600
},
{
"epoch": 3.35195530726257,
"eval_logits/chosen": -3.4567484855651855,
"eval_logits/rejected": -3.5548903942108154,
"eval_logps/chosen": -19.334123611450195,
"eval_logps/rejected": -33.267757415771484,
"eval_loss": 0.3792971074581146,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.073591947555542,
"eval_rewards/margins": 1.073591947555542,
"eval_rewards/rejected": 0.0,
"eval_runtime": 33.5681,
"eval_samples_per_second": 9.235,
"eval_steps_per_second": 0.298,
"step": 600
},
{
"epoch": 3.357541899441341,
"grad_norm": 5.051227255910184,
"learning_rate": 9.638123334223016e-07,
"logits/chosen": -3.534510850906372,
"logits/rejected": -3.752387523651123,
"logps/chosen": -5.093735218048096,
"logps/rejected": -64.19776153564453,
"loss": 0.1873,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4543613195419312,
"rewards/margins": 1.4543613195419312,
"rewards/rejected": 0.0,
"step": 601
},
{
"epoch": 3.363128491620112,
"grad_norm": 4.166054699411632,
"learning_rate": 9.635843155869448e-07,
"logits/chosen": -3.5230698585510254,
"logits/rejected": -3.564009428024292,
"logps/chosen": -26.481098175048828,
"logps/rejected": -31.756576538085938,
"loss": 0.1649,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0945353507995605,
"rewards/margins": 2.0945353507995605,
"rewards/rejected": 0.0,
"step": 602
},
{
"epoch": 3.368715083798883,
"grad_norm": 4.527167153730143,
"learning_rate": 9.63355608782735e-07,
"logits/chosen": -3.5326149463653564,
"logits/rejected": -3.4358036518096924,
"logps/chosen": -25.61269760131836,
"logps/rejected": -17.74207305908203,
"loss": 0.1714,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.192561149597168,
"rewards/margins": 2.192561149597168,
"rewards/rejected": 0.0,
"step": 603
},
{
"epoch": 3.3743016759776534,
"grad_norm": 4.987759145273387,
"learning_rate": 9.631262133495714e-07,
"logits/chosen": -3.794609308242798,
"logits/rejected": -3.7175488471984863,
"logps/chosen": -8.738481521606445,
"logps/rejected": -51.309539794921875,
"loss": 0.1394,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7903088331222534,
"rewards/margins": 1.7903088331222534,
"rewards/rejected": 0.0,
"step": 604
},
{
"epoch": 3.3798882681564244,
"grad_norm": 5.759532896560646,
"learning_rate": 9.628961296283764e-07,
"logits/chosen": -3.6752233505249023,
"logits/rejected": -3.772005558013916,
"logps/chosen": -5.829462051391602,
"logps/rejected": -27.75981903076172,
"loss": 0.1769,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6819231510162354,
"rewards/margins": 1.6819231510162354,
"rewards/rejected": 0.0,
"step": 605
},
{
"epoch": 3.3854748603351954,
"grad_norm": 4.743147267197351,
"learning_rate": 9.626653579610958e-07,
"logits/chosen": -3.6155364513397217,
"logits/rejected": -3.660079002380371,
"logps/chosen": -13.428018569946289,
"logps/rejected": -47.90098571777344,
"loss": 0.1733,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.271134614944458,
"rewards/margins": 2.271134614944458,
"rewards/rejected": 0.0,
"step": 606
},
{
"epoch": 3.3910614525139664,
"grad_norm": 4.898652037325377,
"learning_rate": 9.624338986906967e-07,
"logits/chosen": -3.5013554096221924,
"logits/rejected": -3.5532264709472656,
"logps/chosen": -14.624397277832031,
"logps/rejected": -30.702373504638672,
"loss": 0.1768,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7548508644104004,
"rewards/margins": 1.7548508644104004,
"rewards/rejected": 0.0,
"step": 607
},
{
"epoch": 3.3966480446927374,
"grad_norm": 6.47308077269089,
"learning_rate": 9.622017521611697e-07,
"logits/chosen": -3.6072585582733154,
"logits/rejected": -3.499119997024536,
"logps/chosen": -12.317621231079102,
"logps/rejected": -28.305444717407227,
"loss": 0.1748,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.382249355316162,
"rewards/margins": 2.382249355316162,
"rewards/rejected": 0.0,
"step": 608
},
{
"epoch": 3.4022346368715084,
"grad_norm": 4.6331298377841,
"learning_rate": 9.619689187175248e-07,
"logits/chosen": -3.594680070877075,
"logits/rejected": -3.6533143520355225,
"logps/chosen": -10.815997123718262,
"logps/rejected": -52.948158264160156,
"loss": 0.1669,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.53365159034729,
"rewards/margins": 1.53365159034729,
"rewards/rejected": 0.0,
"step": 609
},
{
"epoch": 3.4078212290502794,
"grad_norm": 4.438532761181371,
"learning_rate": 9.61735398705795e-07,
"logits/chosen": -3.5103824138641357,
"logits/rejected": -3.302504062652588,
"logps/chosen": -31.74606704711914,
"logps/rejected": -17.0972957611084,
"loss": 0.1491,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.507812023162842,
"rewards/margins": 2.507812023162842,
"rewards/rejected": 0.0,
"step": 610
},
{
"epoch": 3.4134078212290504,
"grad_norm": 5.218958127801242,
"learning_rate": 9.615011924730318e-07,
"logits/chosen": -3.8670122623443604,
"logits/rejected": -3.679332733154297,
"logps/chosen": -14.15910530090332,
"logps/rejected": -21.305110931396484,
"loss": 0.1687,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9713423252105713,
"rewards/margins": 1.9713423252105713,
"rewards/rejected": 0.0,
"step": 611
},
{
"epoch": 3.4189944134078214,
"grad_norm": 4.811857730384431,
"learning_rate": 9.612663003673077e-07,
"logits/chosen": -3.7115883827209473,
"logits/rejected": -3.7226953506469727,
"logps/chosen": -9.147367477416992,
"logps/rejected": -20.090017318725586,
"loss": 0.1743,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.535884141921997,
"rewards/margins": 1.535884141921997,
"rewards/rejected": 0.0,
"step": 612
},
{
"epoch": 3.4245810055865924,
"grad_norm": 5.1760160227896135,
"learning_rate": 9.610307227377145e-07,
"logits/chosen": -4.140496253967285,
"logits/rejected": -3.9840750694274902,
"logps/chosen": -37.68970489501953,
"logps/rejected": -21.389732360839844,
"loss": 0.1513,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4988536834716797,
"rewards/margins": 2.4988536834716797,
"rewards/rejected": 0.0,
"step": 613
},
{
"epoch": 3.430167597765363,
"grad_norm": 6.542052114074927,
"learning_rate": 9.60794459934362e-07,
"logits/chosen": -3.9749577045440674,
"logits/rejected": -3.6949622631073,
"logps/chosen": -7.262472152709961,
"logps/rejected": -27.838661193847656,
"loss": 0.1995,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7984397411346436,
"rewards/margins": 1.7984397411346436,
"rewards/rejected": 0.0,
"step": 614
},
{
"epoch": 3.435754189944134,
"grad_norm": 4.723249123539818,
"learning_rate": 9.60557512308379e-07,
"logits/chosen": -3.4244918823242188,
"logits/rejected": -3.4502954483032227,
"logps/chosen": -16.041728973388672,
"logps/rejected": -34.97552490234375,
"loss": 0.1455,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9372085332870483,
"rewards/margins": 1.9372085332870483,
"rewards/rejected": 0.0,
"step": 615
},
{
"epoch": 3.441340782122905,
"grad_norm": 6.321918155936006,
"learning_rate": 9.60319880211912e-07,
"logits/chosen": -3.6923134326934814,
"logits/rejected": -3.7828528881073,
"logps/chosen": -8.835647583007812,
"logps/rejected": -19.127933502197266,
"loss": 0.1878,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.688408374786377,
"rewards/margins": 1.688408374786377,
"rewards/rejected": 0.0,
"step": 616
},
{
"epoch": 3.446927374301676,
"grad_norm": 5.15241985028114,
"learning_rate": 9.600815639981248e-07,
"logits/chosen": -3.5907249450683594,
"logits/rejected": -3.7290914058685303,
"logps/chosen": -20.6009578704834,
"logps/rejected": -23.31732177734375,
"loss": 0.1511,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.027575969696045,
"rewards/margins": 2.027575969696045,
"rewards/rejected": 0.0,
"step": 617
},
{
"epoch": 3.452513966480447,
"grad_norm": 6.098796287127406,
"learning_rate": 9.598425640211971e-07,
"logits/chosen": -3.561833620071411,
"logits/rejected": -3.593796730041504,
"logps/chosen": -14.328054428100586,
"logps/rejected": -27.98879623413086,
"loss": 0.1893,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8051799535751343,
"rewards/margins": 1.8051799535751343,
"rewards/rejected": 0.0,
"step": 618
},
{
"epoch": 3.458100558659218,
"grad_norm": 6.881151069002195,
"learning_rate": 9.596028806363258e-07,
"logits/chosen": -3.9076590538024902,
"logits/rejected": -3.656128406524658,
"logps/chosen": -24.803630828857422,
"logps/rejected": -40.879093170166016,
"loss": 0.1708,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7125649452209473,
"rewards/margins": 1.7125649452209473,
"rewards/rejected": 0.0,
"step": 619
},
{
"epoch": 3.463687150837989,
"grad_norm": 4.453490548371726,
"learning_rate": 9.593625141997232e-07,
"logits/chosen": -3.6640028953552246,
"logits/rejected": -3.509269952774048,
"logps/chosen": -9.31806755065918,
"logps/rejected": -26.860980987548828,
"loss": 0.1543,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7607637643814087,
"rewards/margins": 1.7607637643814087,
"rewards/rejected": 0.0,
"step": 620
},
{
"epoch": 3.463687150837989,
"eval_logits/chosen": -3.4551804065704346,
"eval_logits/rejected": -3.559019088745117,
"eval_logps/chosen": -19.769975662231445,
"eval_logps/rejected": -34.009849548339844,
"eval_loss": 0.3931514024734497,
"eval_rewards/accuracies": 0.925000011920929,
"eval_rewards/chosen": 1.0300068855285645,
"eval_rewards/margins": 1.0300068855285645,
"eval_rewards/rejected": 0.0,
"eval_runtime": 34.8509,
"eval_samples_per_second": 8.895,
"eval_steps_per_second": 0.287,
"step": 620
},
{
"epoch": 3.46927374301676,
"grad_norm": 4.69081008723961,
"learning_rate": 9.591214650686165e-07,
"logits/chosen": -3.429396867752075,
"logits/rejected": -3.533478260040283,
"logps/chosen": -9.281721115112305,
"logps/rejected": -41.42967224121094,
"loss": 0.1535,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.78550124168396,
"rewards/margins": 1.78550124168396,
"rewards/rejected": 0.0,
"step": 621
},
{
"epoch": 3.4748603351955305,
"grad_norm": 5.138819012890861,
"learning_rate": 9.588797336012476e-07,
"logits/chosen": -3.8199877738952637,
"logits/rejected": -3.800265073776245,
"logps/chosen": -8.140040397644043,
"logps/rejected": -24.334957122802734,
"loss": 0.1587,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5762226581573486,
"rewards/margins": 1.5762226581573486,
"rewards/rejected": 0.0,
"step": 622
},
{
"epoch": 3.4804469273743015,
"grad_norm": 5.094867464072331,
"learning_rate": 9.586373201568727e-07,
"logits/chosen": -3.436248540878296,
"logits/rejected": -3.528372049331665,
"logps/chosen": -6.533449172973633,
"logps/rejected": -46.51861572265625,
"loss": 0.178,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7865111827850342,
"rewards/margins": 1.7865111827850342,
"rewards/rejected": 0.0,
"step": 623
},
{
"epoch": 3.4860335195530725,
"grad_norm": 7.0326792611564315,
"learning_rate": 9.58394225095761e-07,
"logits/chosen": -3.6675281524658203,
"logits/rejected": -3.461420774459839,
"logps/chosen": -12.002155303955078,
"logps/rejected": -31.428363800048828,
"loss": 0.1891,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4570937156677246,
"rewards/margins": 1.4570937156677246,
"rewards/rejected": 0.0,
"step": 624
},
{
"epoch": 3.4916201117318435,
"grad_norm": 5.30470515311673,
"learning_rate": 9.581504487791954e-07,
"logits/chosen": -3.8229575157165527,
"logits/rejected": -3.6838696002960205,
"logps/chosen": -5.357357501983643,
"logps/rejected": -30.263629913330078,
"loss": 0.186,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4722356796264648,
"rewards/margins": 1.4722356796264648,
"rewards/rejected": 0.0,
"step": 625
},
{
"epoch": 3.4972067039106145,
"grad_norm": 5.262133806438183,
"learning_rate": 9.579059915694706e-07,
"logits/chosen": -3.353064775466919,
"logits/rejected": -3.4130282402038574,
"logps/chosen": -17.963783264160156,
"logps/rejected": -30.49808120727539,
"loss": 0.2108,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7664779424667358,
"rewards/margins": 1.7664779424667358,
"rewards/rejected": 0.0,
"step": 626
},
{
"epoch": 3.5027932960893855,
"grad_norm": 5.004012675634213,
"learning_rate": 9.57660853829894e-07,
"logits/chosen": -3.4588680267333984,
"logits/rejected": -3.574172258377075,
"logps/chosen": -6.091219902038574,
"logps/rejected": -29.69709014892578,
"loss": 0.1659,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4152452945709229,
"rewards/margins": 1.4152452945709229,
"rewards/rejected": 0.0,
"step": 627
},
{
"epoch": 3.5083798882681565,
"grad_norm": 4.626292696994778,
"learning_rate": 9.574150359247835e-07,
"logits/chosen": -3.640641689300537,
"logits/rejected": -3.775499105453491,
"logps/chosen": -15.644683837890625,
"logps/rejected": -39.21204376220703,
"loss": 0.1422,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2473812103271484,
"rewards/margins": 2.2473812103271484,
"rewards/rejected": 0.0,
"step": 628
},
{
"epoch": 3.5139664804469275,
"grad_norm": 7.093816370215724,
"learning_rate": 9.571685382194685e-07,
"logits/chosen": -3.4563698768615723,
"logits/rejected": -3.6638269424438477,
"logps/chosen": -5.070566654205322,
"logps/rejected": -36.20479965209961,
"loss": 0.196,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.20818293094635,
"rewards/margins": 1.20818293094635,
"rewards/rejected": 0.0,
"step": 629
},
{
"epoch": 3.5195530726256985,
"grad_norm": 5.626556316151276,
"learning_rate": 9.569213610802883e-07,
"logits/chosen": -3.2967770099639893,
"logits/rejected": -3.477936029434204,
"logps/chosen": -20.65081024169922,
"logps/rejected": -46.64710235595703,
"loss": 0.1941,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5222539901733398,
"rewards/margins": 1.5222539901733398,
"rewards/rejected": 0.0,
"step": 630
},
{
"epoch": 3.5251396648044695,
"grad_norm": 5.238010505375191,
"learning_rate": 9.566735048745925e-07,
"logits/chosen": -3.8448498249053955,
"logits/rejected": -3.834787130355835,
"logps/chosen": -12.810737609863281,
"logps/rejected": -23.271757125854492,
"loss": 0.1689,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8085863590240479,
"rewards/margins": 1.8085863590240479,
"rewards/rejected": 0.0,
"step": 631
},
{
"epoch": 3.5307262569832405,
"grad_norm": 5.361118050321566,
"learning_rate": 9.564249699707393e-07,
"logits/chosen": -3.449556350708008,
"logits/rejected": -3.5498440265655518,
"logps/chosen": -14.951969146728516,
"logps/rejected": -39.278709411621094,
"loss": 0.1504,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.020573616027832,
"rewards/margins": 2.020573616027832,
"rewards/rejected": 0.0,
"step": 632
},
{
"epoch": 3.536312849162011,
"grad_norm": 5.868396347645879,
"learning_rate": 9.561757567380958e-07,
"logits/chosen": -3.4344379901885986,
"logits/rejected": -3.4854111671447754,
"logps/chosen": -10.592303276062012,
"logps/rejected": -27.1976375579834,
"loss": 0.1792,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5092116594314575,
"rewards/margins": 1.5092116594314575,
"rewards/rejected": 0.0,
"step": 633
},
{
"epoch": 3.541899441340782,
"grad_norm": 5.212421228813011,
"learning_rate": 9.559258655470375e-07,
"logits/chosen": -3.630856513977051,
"logits/rejected": -3.720335006713867,
"logps/chosen": -11.691306114196777,
"logps/rejected": -57.59962463378906,
"loss": 0.1744,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.189746379852295,
"rewards/margins": 2.189746379852295,
"rewards/rejected": 0.0,
"step": 634
},
{
"epoch": 3.547486033519553,
"grad_norm": 5.150740438504914,
"learning_rate": 9.556752967689469e-07,
"logits/chosen": -3.617985725402832,
"logits/rejected": -3.5494420528411865,
"logps/chosen": -11.184094429016113,
"logps/rejected": -33.076351165771484,
"loss": 0.1861,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.091069221496582,
"rewards/margins": 2.091069221496582,
"rewards/rejected": 0.0,
"step": 635
},
{
"epoch": 3.553072625698324,
"grad_norm": 5.440410797672074,
"learning_rate": 9.55424050776214e-07,
"logits/chosen": -3.30224347114563,
"logits/rejected": -3.57208514213562,
"logps/chosen": -10.038273811340332,
"logps/rejected": -29.3889217376709,
"loss": 0.1916,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6328983306884766,
"rewards/margins": 1.6328983306884766,
"rewards/rejected": 0.0,
"step": 636
},
{
"epoch": 3.558659217877095,
"grad_norm": 4.65723960855201,
"learning_rate": 9.55172127942235e-07,
"logits/chosen": -3.5845320224761963,
"logits/rejected": -3.3195221424102783,
"logps/chosen": -12.603439331054688,
"logps/rejected": -73.64794921875,
"loss": 0.1546,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.90388822555542,
"rewards/margins": 1.90388822555542,
"rewards/rejected": 0.0,
"step": 637
},
{
"epoch": 3.564245810055866,
"grad_norm": 6.528446821475113,
"learning_rate": 9.549195286414122e-07,
"logits/chosen": -3.185835123062134,
"logits/rejected": -3.4180290699005127,
"logps/chosen": -31.241548538208008,
"logps/rejected": -42.23626708984375,
"loss": 0.1527,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1054329872131348,
"rewards/margins": 2.1054329872131348,
"rewards/rejected": 0.0,
"step": 638
},
{
"epoch": 3.5698324022346366,
"grad_norm": 4.677698484388385,
"learning_rate": 9.54666253249153e-07,
"logits/chosen": -3.7723662853240967,
"logits/rejected": -3.773625373840332,
"logps/chosen": -3.9417715072631836,
"logps/rejected": -16.878204345703125,
"loss": 0.1987,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1821473836898804,
"rewards/margins": 1.1821473836898804,
"rewards/rejected": 0.0,
"step": 639
},
{
"epoch": 3.5754189944134076,
"grad_norm": 5.117589561037088,
"learning_rate": 9.5441230214187e-07,
"logits/chosen": -3.399118661880493,
"logits/rejected": -3.243917942047119,
"logps/chosen": -18.99847412109375,
"logps/rejected": -34.145355224609375,
"loss": 0.17,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8375072479248047,
"rewards/margins": 1.8375072479248047,
"rewards/rejected": 0.0,
"step": 640
},
{
"epoch": 3.5754189944134076,
"eval_logits/chosen": -3.4527626037597656,
"eval_logits/rejected": -3.556689739227295,
"eval_logps/chosen": -19.967891693115234,
"eval_logps/rejected": -34.24546432495117,
"eval_loss": 0.40128594636917114,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.0102150440216064,
"eval_rewards/margins": 1.0102150440216064,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7062,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 640
},
{
"epoch": 3.5810055865921786,
"grad_norm": 5.133411055097609,
"learning_rate": 9.541576756969796e-07,
"logits/chosen": -3.5344247817993164,
"logits/rejected": -3.62461256980896,
"logps/chosen": -23.3912296295166,
"logps/rejected": -42.15888595581055,
"loss": 0.1522,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3202648162841797,
"rewards/margins": 2.3202648162841797,
"rewards/rejected": 0.0,
"step": 641
},
{
"epoch": 3.5865921787709496,
"grad_norm": 4.570418247843243,
"learning_rate": 9.539023742929021e-07,
"logits/chosen": -3.8138322830200195,
"logits/rejected": -3.665846586227417,
"logps/chosen": -9.992279052734375,
"logps/rejected": -33.50912857055664,
"loss": 0.1555,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0003204345703125,
"rewards/margins": 2.0003204345703125,
"rewards/rejected": 0.0,
"step": 642
},
{
"epoch": 3.5921787709497206,
"grad_norm": 7.227085737302612,
"learning_rate": 9.536463983090606e-07,
"logits/chosen": -3.5766735076904297,
"logits/rejected": -3.831467390060425,
"logps/chosen": -8.467517852783203,
"logps/rejected": -32.18593978881836,
"loss": 0.1831,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8903911113739014,
"rewards/margins": 1.8903911113739014,
"rewards/rejected": 0.0,
"step": 643
},
{
"epoch": 3.5977653631284916,
"grad_norm": 5.682147119312251,
"learning_rate": 9.533897481258815e-07,
"logits/chosen": -3.520840883255005,
"logits/rejected": -3.321063280105591,
"logps/chosen": -13.972931861877441,
"logps/rejected": -66.25763702392578,
"loss": 0.1799,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5537865161895752,
"rewards/margins": 1.5537865161895752,
"rewards/rejected": 0.0,
"step": 644
},
{
"epoch": 3.6033519553072626,
"grad_norm": 5.696592487103718,
"learning_rate": 9.531324241247922e-07,
"logits/chosen": -3.572374105453491,
"logits/rejected": -3.5924270153045654,
"logps/chosen": -24.20823097229004,
"logps/rejected": -24.128681182861328,
"loss": 0.1591,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.120861053466797,
"rewards/margins": 2.120861053466797,
"rewards/rejected": 0.0,
"step": 645
},
{
"epoch": 3.6089385474860336,
"grad_norm": 6.012466721685493,
"learning_rate": 9.528744266882223e-07,
"logits/chosen": -3.735198497772217,
"logits/rejected": -3.734330892562866,
"logps/chosen": -26.19061851501465,
"logps/rejected": -27.667203903198242,
"loss": 0.1793,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7986745834350586,
"rewards/margins": 2.7986745834350586,
"rewards/rejected": 0.0,
"step": 646
},
{
"epoch": 3.6145251396648046,
"grad_norm": 4.644339200519325,
"learning_rate": 9.526157561996017e-07,
"logits/chosen": -3.922447443008423,
"logits/rejected": -3.8625240325927734,
"logps/chosen": -17.55810546875,
"logps/rejected": -28.501117706298828,
"loss": 0.1665,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.302061080932617,
"rewards/margins": 2.302061080932617,
"rewards/rejected": 0.0,
"step": 647
},
{
"epoch": 3.6201117318435756,
"grad_norm": 4.455673398979726,
"learning_rate": 9.523564130433609e-07,
"logits/chosen": -3.790313959121704,
"logits/rejected": -3.7689945697784424,
"logps/chosen": -4.420231819152832,
"logps/rejected": -21.001861572265625,
"loss": 0.1689,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5885013341903687,
"rewards/margins": 1.5885013341903687,
"rewards/rejected": 0.0,
"step": 648
},
{
"epoch": 3.6256983240223466,
"grad_norm": 5.067755669359323,
"learning_rate": 9.520963976049301e-07,
"logits/chosen": -3.3586041927337646,
"logits/rejected": -3.1449713706970215,
"logps/chosen": -25.404495239257812,
"logps/rejected": -51.243751525878906,
"loss": 0.1795,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6585184335708618,
"rewards/margins": 1.6585184335708618,
"rewards/rejected": 0.0,
"step": 649
},
{
"epoch": 3.631284916201117,
"grad_norm": 5.296324090772392,
"learning_rate": 9.518357102707386e-07,
"logits/chosen": -3.651949882507324,
"logits/rejected": -3.735245943069458,
"logps/chosen": -4.745665550231934,
"logps/rejected": -27.808101654052734,
"loss": 0.1833,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3110535144805908,
"rewards/margins": 1.3110535144805908,
"rewards/rejected": 0.0,
"step": 650
},
{
"epoch": 3.636871508379888,
"grad_norm": 4.365839520605308,
"learning_rate": 9.51574351428214e-07,
"logits/chosen": -3.827695608139038,
"logits/rejected": -3.716557502746582,
"logps/chosen": -11.460929870605469,
"logps/rejected": -32.29026794433594,
"loss": 0.2001,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.666351079940796,
"rewards/margins": 1.666351079940796,
"rewards/rejected": 0.0,
"step": 651
},
{
"epoch": 3.642458100558659,
"grad_norm": 4.479324786427868,
"learning_rate": 9.513123214657821e-07,
"logits/chosen": -3.68721866607666,
"logits/rejected": -3.325680732727051,
"logps/chosen": -9.46074390411377,
"logps/rejected": -41.34431457519531,
"loss": 0.1789,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.501688838005066,
"rewards/margins": 1.501688838005066,
"rewards/rejected": 0.0,
"step": 652
},
{
"epoch": 3.64804469273743,
"grad_norm": 5.149847701080681,
"learning_rate": 9.510496207728662e-07,
"logits/chosen": -3.6695034503936768,
"logits/rejected": -3.5877413749694824,
"logps/chosen": -15.89958381652832,
"logps/rejected": -28.580312728881836,
"loss": 0.1903,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0160269737243652,
"rewards/margins": 2.0160269737243652,
"rewards/rejected": 0.0,
"step": 653
},
{
"epoch": 3.653631284916201,
"grad_norm": 5.109958931687544,
"learning_rate": 9.507862497398864e-07,
"logits/chosen": -4.000102996826172,
"logits/rejected": -3.9663286209106445,
"logps/chosen": -7.500527381896973,
"logps/rejected": -24.804828643798828,
"loss": 0.1609,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7702608108520508,
"rewards/margins": 1.7702608108520508,
"rewards/rejected": 0.0,
"step": 654
},
{
"epoch": 3.659217877094972,
"grad_norm": 4.656953780482441,
"learning_rate": 9.505222087582588e-07,
"logits/chosen": -3.6987736225128174,
"logits/rejected": -3.8206393718719482,
"logps/chosen": -10.570422172546387,
"logps/rejected": -27.170337677001953,
"loss": 0.1954,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9592132568359375,
"rewards/margins": 1.9592132568359375,
"rewards/rejected": 0.0,
"step": 655
},
{
"epoch": 3.664804469273743,
"grad_norm": 4.713062170943476,
"learning_rate": 9.502574982203954e-07,
"logits/chosen": -3.754716157913208,
"logits/rejected": -3.68346905708313,
"logps/chosen": -13.092695236206055,
"logps/rejected": -32.32396697998047,
"loss": 0.1857,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1699705123901367,
"rewards/margins": 2.1699705123901367,
"rewards/rejected": 0.0,
"step": 656
},
{
"epoch": 3.6703910614525137,
"grad_norm": 4.215207701061461,
"learning_rate": 9.499921185197032e-07,
"logits/chosen": -3.3362271785736084,
"logits/rejected": -3.278992176055908,
"logps/chosen": -4.92309045791626,
"logps/rejected": -34.32732009887695,
"loss": 0.1664,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4023206233978271,
"rewards/margins": 1.4023206233978271,
"rewards/rejected": 0.0,
"step": 657
},
{
"epoch": 3.6759776536312847,
"grad_norm": 5.333585474707274,
"learning_rate": 9.497260700505835e-07,
"logits/chosen": -3.3382411003112793,
"logits/rejected": -3.0705416202545166,
"logps/chosen": -14.22011947631836,
"logps/rejected": -36.958946228027344,
"loss": 0.2132,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.247107982635498,
"rewards/margins": 1.247107982635498,
"rewards/rejected": 0.0,
"step": 658
},
{
"epoch": 3.6815642458100557,
"grad_norm": 5.676048291789358,
"learning_rate": 9.494593532084321e-07,
"logits/chosen": -3.778850793838501,
"logits/rejected": -3.78900408744812,
"logps/chosen": -7.684193134307861,
"logps/rejected": -40.97362518310547,
"loss": 0.1799,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6917026042938232,
"rewards/margins": 1.6917026042938232,
"rewards/rejected": 0.0,
"step": 659
},
{
"epoch": 3.6871508379888267,
"grad_norm": 4.901932214135667,
"learning_rate": 9.491919683896373e-07,
"logits/chosen": -3.4505538940429688,
"logits/rejected": -3.496281147003174,
"logps/chosen": -4.362544059753418,
"logps/rejected": -25.02591323852539,
"loss": 0.1905,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1565850973129272,
"rewards/margins": 1.1565850973129272,
"rewards/rejected": 0.0,
"step": 660
},
{
"epoch": 3.6871508379888267,
"eval_logits/chosen": -3.4496493339538574,
"eval_logits/rejected": -3.5553760528564453,
"eval_logps/chosen": -20.094425201416016,
"eval_logps/rejected": -34.56037139892578,
"eval_loss": 0.39944276213645935,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9975619316101074,
"eval_rewards/margins": 0.9975619316101074,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.736,
"eval_samples_per_second": 9.47,
"eval_steps_per_second": 0.305,
"step": 660
},
{
"epoch": 3.6927374301675977,
"grad_norm": 7.205953665972467,
"learning_rate": 9.48923915991581e-07,
"logits/chosen": -3.631009101867676,
"logits/rejected": -3.6369123458862305,
"logps/chosen": -34.10837936401367,
"logps/rejected": -23.194347381591797,
"loss": 0.1604,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9607927799224854,
"rewards/margins": 1.9607927799224854,
"rewards/rejected": 0.0,
"step": 661
},
{
"epoch": 3.6983240223463687,
"grad_norm": 5.019279966582101,
"learning_rate": 9.486551964126362e-07,
"logits/chosen": -3.6323862075805664,
"logits/rejected": -3.6116278171539307,
"logps/chosen": -5.30668830871582,
"logps/rejected": -29.413761138916016,
"loss": 0.1511,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2488832473754883,
"rewards/margins": 1.2488832473754883,
"rewards/rejected": 0.0,
"step": 662
},
{
"epoch": 3.7039106145251397,
"grad_norm": 4.732609891488001,
"learning_rate": 9.483858100521686e-07,
"logits/chosen": -3.6550815105438232,
"logits/rejected": -3.189901113510132,
"logps/chosen": -8.990190505981445,
"logps/rejected": -46.08880615234375,
"loss": 0.1565,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7714471817016602,
"rewards/margins": 1.7714471817016602,
"rewards/rejected": 0.0,
"step": 663
},
{
"epoch": 3.7094972067039107,
"grad_norm": 5.120186565591483,
"learning_rate": 9.481157573105343e-07,
"logits/chosen": -3.8208818435668945,
"logits/rejected": -3.5488696098327637,
"logps/chosen": -6.346719741821289,
"logps/rejected": -25.36515998840332,
"loss": 0.1615,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6430108547210693,
"rewards/margins": 1.6430108547210693,
"rewards/rejected": 0.0,
"step": 664
},
{
"epoch": 3.7150837988826817,
"grad_norm": 4.466537483585655,
"learning_rate": 9.478450385890793e-07,
"logits/chosen": -3.73468279838562,
"logits/rejected": -3.436859369277954,
"logps/chosen": -14.974721908569336,
"logps/rejected": -31.110218048095703,
"loss": 0.1636,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5450804233551025,
"rewards/margins": 2.5450804233551025,
"rewards/rejected": 0.0,
"step": 665
},
{
"epoch": 3.7206703910614527,
"grad_norm": 5.622882110084028,
"learning_rate": 9.475736542901402e-07,
"logits/chosen": -3.625474691390991,
"logits/rejected": -3.738278865814209,
"logps/chosen": -10.976929664611816,
"logps/rejected": -31.13936996459961,
"loss": 0.1662,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.916973352432251,
"rewards/margins": 1.916973352432251,
"rewards/rejected": 0.0,
"step": 666
},
{
"epoch": 3.7262569832402237,
"grad_norm": 6.155213830518234,
"learning_rate": 9.473016048170424e-07,
"logits/chosen": -3.7011091709136963,
"logits/rejected": -3.6307482719421387,
"logps/chosen": -9.683084487915039,
"logps/rejected": -50.2158088684082,
"loss": 0.2093,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3847639560699463,
"rewards/margins": 1.3847639560699463,
"rewards/rejected": 0.0,
"step": 667
},
{
"epoch": 3.7318435754189943,
"grad_norm": 5.543091406250431,
"learning_rate": 9.470288905740997e-07,
"logits/chosen": -3.4428887367248535,
"logits/rejected": -3.432431936264038,
"logps/chosen": -23.892230987548828,
"logps/rejected": -20.082015991210938,
"loss": 0.1712,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1780343055725098,
"rewards/margins": 2.1780343055725098,
"rewards/rejected": 0.0,
"step": 668
},
{
"epoch": 3.7374301675977653,
"grad_norm": 8.02166683785944,
"learning_rate": 9.467555119666142e-07,
"logits/chosen": -3.8772687911987305,
"logits/rejected": -3.8502283096313477,
"logps/chosen": -5.98099422454834,
"logps/rejected": -44.68482208251953,
"loss": 0.1743,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5072976350784302,
"rewards/margins": 1.5072976350784302,
"rewards/rejected": 0.0,
"step": 669
},
{
"epoch": 3.7430167597765363,
"grad_norm": 6.802613644354898,
"learning_rate": 9.464814694008751e-07,
"logits/chosen": -3.6652045249938965,
"logits/rejected": -3.6151578426361084,
"logps/chosen": -13.424704551696777,
"logps/rejected": -20.253692626953125,
"loss": 0.2114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2190957069396973,
"rewards/margins": 2.2190957069396973,
"rewards/rejected": 0.0,
"step": 670
},
{
"epoch": 3.7486033519553073,
"grad_norm": 4.92290955136957,
"learning_rate": 9.462067632841584e-07,
"logits/chosen": -3.666156053543091,
"logits/rejected": -3.7569448947906494,
"logps/chosen": -7.920692443847656,
"logps/rejected": -29.915081024169922,
"loss": 0.1734,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3987126350402832,
"rewards/margins": 1.3987126350402832,
"rewards/rejected": 0.0,
"step": 671
},
{
"epoch": 3.7541899441340782,
"grad_norm": 6.164093722805633,
"learning_rate": 9.459313940247265e-07,
"logits/chosen": -3.394481897354126,
"logits/rejected": -3.5197486877441406,
"logps/chosen": -29.63011360168457,
"logps/rejected": -15.040891647338867,
"loss": 0.1581,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.591725468635559,
"rewards/margins": 1.591725468635559,
"rewards/rejected": 0.0,
"step": 672
},
{
"epoch": 3.7597765363128492,
"grad_norm": 4.5768595129802225,
"learning_rate": 9.456553620318268e-07,
"logits/chosen": -3.614335298538208,
"logits/rejected": -3.7072131633758545,
"logps/chosen": -7.686119079589844,
"logps/rejected": -48.3739128112793,
"loss": 0.1537,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6674258708953857,
"rewards/margins": 1.6674258708953857,
"rewards/rejected": 0.0,
"step": 673
},
{
"epoch": 3.7653631284916202,
"grad_norm": 5.756451394662651,
"learning_rate": 9.453786677156925e-07,
"logits/chosen": -3.025435209274292,
"logits/rejected": -2.9831600189208984,
"logps/chosen": -32.42550277709961,
"logps/rejected": -43.135040283203125,
"loss": 0.1592,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9021589756011963,
"rewards/margins": 1.9021589756011963,
"rewards/rejected": 0.0,
"step": 674
},
{
"epoch": 3.770949720670391,
"grad_norm": 5.462064085578757,
"learning_rate": 9.451013114875403e-07,
"logits/chosen": -3.7452549934387207,
"logits/rejected": -3.701018810272217,
"logps/chosen": -23.871150970458984,
"logps/rejected": -49.515480041503906,
"loss": 0.1527,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.078068733215332,
"rewards/margins": 2.078068733215332,
"rewards/rejected": 0.0,
"step": 675
},
{
"epoch": 3.776536312849162,
"grad_norm": 5.547590022778659,
"learning_rate": 9.448232937595709e-07,
"logits/chosen": -3.4784772396087646,
"logits/rejected": -3.3884263038635254,
"logps/chosen": -25.33246612548828,
"logps/rejected": -26.745288848876953,
"loss": 0.1654,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1457791328430176,
"rewards/margins": 2.1457791328430176,
"rewards/rejected": 0.0,
"step": 676
},
{
"epoch": 3.782122905027933,
"grad_norm": 6.291665828229705,
"learning_rate": 9.445446149449686e-07,
"logits/chosen": -3.669466018676758,
"logits/rejected": -3.5027315616607666,
"logps/chosen": -8.840606689453125,
"logps/rejected": -26.029987335205078,
"loss": 0.1681,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6782493591308594,
"rewards/margins": 1.6782493591308594,
"rewards/rejected": 0.0,
"step": 677
},
{
"epoch": 3.787709497206704,
"grad_norm": 5.7516916496717645,
"learning_rate": 9.442652754578991e-07,
"logits/chosen": -3.5253586769104004,
"logits/rejected": -3.5481929779052734,
"logps/chosen": -38.82970428466797,
"logps/rejected": -32.559940338134766,
"loss": 0.1614,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7083141803741455,
"rewards/margins": 1.7083141803741455,
"rewards/rejected": 0.0,
"step": 678
},
{
"epoch": 3.793296089385475,
"grad_norm": 5.676763386544318,
"learning_rate": 9.439852757135109e-07,
"logits/chosen": -3.507185935974121,
"logits/rejected": -3.475709915161133,
"logps/chosen": -24.707077026367188,
"logps/rejected": -32.674659729003906,
"loss": 0.1582,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5474183559417725,
"rewards/margins": 2.5474183559417725,
"rewards/rejected": 0.0,
"step": 679
},
{
"epoch": 3.798882681564246,
"grad_norm": 6.652992280245279,
"learning_rate": 9.437046161279338e-07,
"logits/chosen": -3.5408294200897217,
"logits/rejected": -3.46547794342041,
"logps/chosen": -4.129613876342773,
"logps/rejected": -43.75081253051758,
"loss": 0.2188,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.249681830406189,
"rewards/margins": 1.249681830406189,
"rewards/rejected": 0.0,
"step": 680
},
{
"epoch": 3.798882681564246,
"eval_logits/chosen": -3.46787691116333,
"eval_logits/rejected": -3.571978807449341,
"eval_logps/chosen": -19.941822052001953,
"eval_logps/rejected": -34.716636657714844,
"eval_loss": 0.3982676863670349,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.012821912765503,
"eval_rewards/margins": 1.012821912765503,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6974,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 680
},
{
"epoch": 3.804469273743017,
"grad_norm": 6.225168750349631,
"learning_rate": 9.434232971182776e-07,
"logits/chosen": -2.9479496479034424,
"logits/rejected": -2.898322582244873,
"logps/chosen": -31.89510154724121,
"logps/rejected": -45.454803466796875,
"loss": 0.1695,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8217171430587769,
"rewards/margins": 1.8217171430587769,
"rewards/rejected": 0.0,
"step": 681
},
{
"epoch": 3.810055865921788,
"grad_norm": 8.990530591590625,
"learning_rate": 9.431413191026324e-07,
"logits/chosen": -3.5850820541381836,
"logits/rejected": -3.771740674972534,
"logps/chosen": -10.17590045928955,
"logps/rejected": -54.2763671875,
"loss": 0.1855,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7644659280776978,
"rewards/margins": 1.7644659280776978,
"rewards/rejected": 0.0,
"step": 682
},
{
"epoch": 3.815642458100559,
"grad_norm": 5.372783330517193,
"learning_rate": 9.42858682500068e-07,
"logits/chosen": -3.7883987426757812,
"logits/rejected": -3.7369813919067383,
"logps/chosen": -10.788247108459473,
"logps/rejected": -33.16901397705078,
"loss": 0.1667,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7213985919952393,
"rewards/margins": 1.7213985919952393,
"rewards/rejected": 0.0,
"step": 683
},
{
"epoch": 3.82122905027933,
"grad_norm": 5.640058295563442,
"learning_rate": 9.425753877306325e-07,
"logits/chosen": -3.765312671661377,
"logits/rejected": -3.813652753829956,
"logps/chosen": -21.540424346923828,
"logps/rejected": -34.084686279296875,
"loss": 0.1856,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9886975288391113,
"rewards/margins": 1.9886975288391113,
"rewards/rejected": 0.0,
"step": 684
},
{
"epoch": 3.826815642458101,
"grad_norm": 4.9159808943093335,
"learning_rate": 9.422914352153523e-07,
"logits/chosen": -3.9867115020751953,
"logits/rejected": -3.6477715969085693,
"logps/chosen": -8.451070785522461,
"logps/rejected": -68.41416931152344,
"loss": 0.1519,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7793676853179932,
"rewards/margins": 1.7793676853179932,
"rewards/rejected": 0.0,
"step": 685
},
{
"epoch": 3.8324022346368714,
"grad_norm": 5.463507510300138,
"learning_rate": 9.420068253762318e-07,
"logits/chosen": -3.3537416458129883,
"logits/rejected": -3.5590171813964844,
"logps/chosen": -43.074703216552734,
"logps/rejected": -40.1202507019043,
"loss": 0.1659,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.508090019226074,
"rewards/margins": 2.508090019226074,
"rewards/rejected": 0.0,
"step": 686
},
{
"epoch": 3.8379888268156424,
"grad_norm": 5.973513320758181,
"learning_rate": 9.417215586362517e-07,
"logits/chosen": -3.8050286769866943,
"logits/rejected": -3.7647716999053955,
"logps/chosen": -13.847637176513672,
"logps/rejected": -22.573652267456055,
"loss": 0.2194,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.968073844909668,
"rewards/margins": 1.968073844909668,
"rewards/rejected": 0.0,
"step": 687
},
{
"epoch": 3.8435754189944134,
"grad_norm": 5.376646776434792,
"learning_rate": 9.414356354193691e-07,
"logits/chosen": -3.6765923500061035,
"logits/rejected": -3.552823305130005,
"logps/chosen": -6.152892112731934,
"logps/rejected": -32.36994171142578,
"loss": 0.1739,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7726759910583496,
"rewards/margins": 1.7726759910583496,
"rewards/rejected": 0.0,
"step": 688
},
{
"epoch": 3.8491620111731844,
"grad_norm": 7.975341527703978,
"learning_rate": 9.41149056150517e-07,
"logits/chosen": -3.5974161624908447,
"logits/rejected": -3.5041980743408203,
"logps/chosen": -5.884551048278809,
"logps/rejected": -22.4666748046875,
"loss": 0.2047,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4673993587493896,
"rewards/margins": 1.4673993587493896,
"rewards/rejected": 0.0,
"step": 689
},
{
"epoch": 3.8547486033519553,
"grad_norm": 5.3843626924306145,
"learning_rate": 9.40861821255603e-07,
"logits/chosen": -3.120640277862549,
"logits/rejected": -3.1205742359161377,
"logps/chosen": -16.056499481201172,
"logps/rejected": -31.95566177368164,
"loss": 0.1774,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2459497451782227,
"rewards/margins": 2.2459497451782227,
"rewards/rejected": 0.0,
"step": 690
},
{
"epoch": 3.8603351955307263,
"grad_norm": 5.970205138837227,
"learning_rate": 9.405739311615093e-07,
"logits/chosen": -3.544402837753296,
"logits/rejected": -3.556819438934326,
"logps/chosen": -8.344869613647461,
"logps/rejected": -27.311933517456055,
"loss": 0.1976,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.778423547744751,
"rewards/margins": 1.778423547744751,
"rewards/rejected": 0.0,
"step": 691
},
{
"epoch": 3.8659217877094973,
"grad_norm": 5.514562150326068,
"learning_rate": 9.402853862960922e-07,
"logits/chosen": -3.2727842330932617,
"logits/rejected": -3.3912999629974365,
"logps/chosen": -16.866605758666992,
"logps/rejected": -35.99789810180664,
"loss": 0.1617,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2165558338165283,
"rewards/margins": 2.2165558338165283,
"rewards/rejected": 0.0,
"step": 692
},
{
"epoch": 3.871508379888268,
"grad_norm": 5.08182089806298,
"learning_rate": 9.399961870881803e-07,
"logits/chosen": -3.807812213897705,
"logits/rejected": -3.3577778339385986,
"logps/chosen": -11.646402359008789,
"logps/rejected": -44.892578125,
"loss": 0.1617,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.149644136428833,
"rewards/margins": 2.149644136428833,
"rewards/rejected": 0.0,
"step": 693
},
{
"epoch": 3.877094972067039,
"grad_norm": 5.277433404269871,
"learning_rate": 9.397063339675755e-07,
"logits/chosen": -3.4058542251586914,
"logits/rejected": -3.51452898979187,
"logps/chosen": -16.585853576660156,
"logps/rejected": -25.68677520751953,
"loss": 0.1634,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.123507261276245,
"rewards/margins": 2.123507261276245,
"rewards/rejected": 0.0,
"step": 694
},
{
"epoch": 3.88268156424581,
"grad_norm": 5.577486384805503,
"learning_rate": 9.39415827365051e-07,
"logits/chosen": -3.2718281745910645,
"logits/rejected": -3.2168197631835938,
"logps/chosen": -39.55937957763672,
"logps/rejected": -34.023231506347656,
"loss": 0.18,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.823746681213379,
"rewards/margins": 1.823746681213379,
"rewards/rejected": 0.0,
"step": 695
},
{
"epoch": 3.888268156424581,
"grad_norm": 6.045947313979074,
"learning_rate": 9.391246677123514e-07,
"logits/chosen": -3.422668695449829,
"logits/rejected": -3.5968868732452393,
"logps/chosen": -20.796070098876953,
"logps/rejected": -29.368831634521484,
"loss": 0.191,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6497948169708252,
"rewards/margins": 1.6497948169708252,
"rewards/rejected": 0.0,
"step": 696
},
{
"epoch": 3.893854748603352,
"grad_norm": 7.031171292255311,
"learning_rate": 9.388328554421916e-07,
"logits/chosen": -3.605628728866577,
"logits/rejected": -3.9246270656585693,
"logps/chosen": -7.370364189147949,
"logps/rejected": -28.591331481933594,
"loss": 0.1916,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4823625087738037,
"rewards/margins": 1.4823625087738037,
"rewards/rejected": 0.0,
"step": 697
},
{
"epoch": 3.899441340782123,
"grad_norm": 5.939034876154746,
"learning_rate": 9.385403909882568e-07,
"logits/chosen": -3.5976762771606445,
"logits/rejected": -3.3776772022247314,
"logps/chosen": -47.60350036621094,
"logps/rejected": -34.28583526611328,
"loss": 0.1746,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.683462619781494,
"rewards/margins": 2.683462619781494,
"rewards/rejected": 0.0,
"step": 698
},
{
"epoch": 3.905027932960894,
"grad_norm": 5.466067710144353,
"learning_rate": 9.382472747852013e-07,
"logits/chosen": -3.800515651702881,
"logits/rejected": -3.8023061752319336,
"logps/chosen": -8.329952239990234,
"logps/rejected": -18.521623611450195,
"loss": 0.1745,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.788851261138916,
"rewards/margins": 1.788851261138916,
"rewards/rejected": 0.0,
"step": 699
},
{
"epoch": 3.910614525139665,
"grad_norm": 5.108787830059381,
"learning_rate": 9.379535072686479e-07,
"logits/chosen": -4.01161003112793,
"logits/rejected": -3.925262689590454,
"logps/chosen": -13.309797286987305,
"logps/rejected": -23.609291076660156,
"loss": 0.1544,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0578761100769043,
"rewards/margins": 2.0578761100769043,
"rewards/rejected": 0.0,
"step": 700
},
{
"epoch": 3.910614525139665,
"eval_logits/chosen": -3.4569759368896484,
"eval_logits/rejected": -3.5611720085144043,
"eval_logps/chosen": -19.63014793395996,
"eval_logps/rejected": -34.3280029296875,
"eval_loss": 0.3954744040966034,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.0439895391464233,
"eval_rewards/margins": 1.0439895391464233,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7175,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 700
},
{
"epoch": 3.916201117318436,
"grad_norm": 6.2322334272570235,
"learning_rate": 9.376590888751874e-07,
"logits/chosen": -3.695904016494751,
"logits/rejected": -3.675644636154175,
"logps/chosen": -6.758902549743652,
"logps/rejected": -61.0106086730957,
"loss": 0.1784,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6046574115753174,
"rewards/margins": 1.6046574115753174,
"rewards/rejected": 0.0,
"step": 701
},
{
"epoch": 3.921787709497207,
"grad_norm": 4.9212756999100735,
"learning_rate": 9.373640200423781e-07,
"logits/chosen": -3.793055295944214,
"logits/rejected": -3.6425044536590576,
"logps/chosen": -13.498632431030273,
"logps/rejected": -17.117267608642578,
"loss": 0.1791,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3323898315429688,
"rewards/margins": 2.3323898315429688,
"rewards/rejected": 0.0,
"step": 702
},
{
"epoch": 3.927374301675978,
"grad_norm": 8.869097302222126,
"learning_rate": 9.370683012087446e-07,
"logits/chosen": -3.450406551361084,
"logits/rejected": -3.642115831375122,
"logps/chosen": -34.950592041015625,
"logps/rejected": -32.08348846435547,
"loss": 0.1862,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3248953819274902,
"rewards/margins": 2.3248953819274902,
"rewards/rejected": 0.0,
"step": 703
},
{
"epoch": 3.9329608938547485,
"grad_norm": 5.4792668465797965,
"learning_rate": 9.36771932813778e-07,
"logits/chosen": -3.65859317779541,
"logits/rejected": -3.6708462238311768,
"logps/chosen": -9.614830017089844,
"logps/rejected": -31.10177993774414,
"loss": 0.1733,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6424806118011475,
"rewards/margins": 1.6424806118011475,
"rewards/rejected": 0.0,
"step": 704
},
{
"epoch": 3.9385474860335195,
"grad_norm": 5.624524800100163,
"learning_rate": 9.364749152979343e-07,
"logits/chosen": -3.3821287155151367,
"logits/rejected": -3.666095733642578,
"logps/chosen": -19.01564598083496,
"logps/rejected": -27.14424705505371,
"loss": 0.1645,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2867445945739746,
"rewards/margins": 2.2867445945739746,
"rewards/rejected": 0.0,
"step": 705
},
{
"epoch": 3.9441340782122905,
"grad_norm": 5.402923801872257,
"learning_rate": 9.361772491026346e-07,
"logits/chosen": -3.9651875495910645,
"logits/rejected": -4.0007171630859375,
"logps/chosen": -9.126187324523926,
"logps/rejected": -20.278427124023438,
"loss": 0.1801,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7250056266784668,
"rewards/margins": 1.7250056266784668,
"rewards/rejected": 0.0,
"step": 706
},
{
"epoch": 3.9497206703910615,
"grad_norm": 4.557502412492466,
"learning_rate": 9.358789346702638e-07,
"logits/chosen": -3.9285929203033447,
"logits/rejected": -3.7418060302734375,
"logps/chosen": -31.137500762939453,
"logps/rejected": -34.382118225097656,
"loss": 0.1544,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.618072509765625,
"rewards/margins": 2.618072509765625,
"rewards/rejected": 0.0,
"step": 707
},
{
"epoch": 3.9553072625698324,
"grad_norm": 5.425198650099388,
"learning_rate": 9.355799724441703e-07,
"logits/chosen": -3.633681297302246,
"logits/rejected": -3.6181023120880127,
"logps/chosen": -20.004682540893555,
"logps/rejected": -15.314538955688477,
"loss": 0.1568,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.233163356781006,
"rewards/margins": 2.233163356781006,
"rewards/rejected": 0.0,
"step": 708
},
{
"epoch": 3.9608938547486034,
"grad_norm": 7.821883107455096,
"learning_rate": 9.352803628686652e-07,
"logits/chosen": -3.8305516242980957,
"logits/rejected": -3.6730797290802,
"logps/chosen": -11.294788360595703,
"logps/rejected": -34.06853103637695,
"loss": 0.1885,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.016578435897827,
"rewards/margins": 2.016578435897827,
"rewards/rejected": 0.0,
"step": 709
},
{
"epoch": 3.9664804469273744,
"grad_norm": 6.80654800888443,
"learning_rate": 9.349801063890217e-07,
"logits/chosen": -3.607896089553833,
"logits/rejected": -3.580332040786743,
"logps/chosen": -7.952505111694336,
"logps/rejected": -50.59882354736328,
"loss": 0.1772,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.695425033569336,
"rewards/margins": 1.695425033569336,
"rewards/rejected": 0.0,
"step": 710
},
{
"epoch": 3.972067039106145,
"grad_norm": 5.817496764436292,
"learning_rate": 9.346792034514745e-07,
"logits/chosen": -3.566187858581543,
"logits/rejected": -3.4624316692352295,
"logps/chosen": -12.475229263305664,
"logps/rejected": -24.530345916748047,
"loss": 0.1821,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.693335771560669,
"rewards/margins": 1.693335771560669,
"rewards/rejected": 0.0,
"step": 711
},
{
"epoch": 3.977653631284916,
"grad_norm": 5.012839322578175,
"learning_rate": 9.343776545032187e-07,
"logits/chosen": -3.9504566192626953,
"logits/rejected": -3.617793560028076,
"logps/chosen": -14.308114051818848,
"logps/rejected": -35.20501708984375,
"loss": 0.16,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9731078147888184,
"rewards/margins": 1.9731078147888184,
"rewards/rejected": 0.0,
"step": 712
},
{
"epoch": 3.983240223463687,
"grad_norm": 5.075294614685221,
"learning_rate": 9.340754599924098e-07,
"logits/chosen": -3.4559686183929443,
"logits/rejected": -3.4902491569519043,
"logps/chosen": -10.13359260559082,
"logps/rejected": -24.145774841308594,
"loss": 0.156,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5532363653182983,
"rewards/margins": 1.5532363653182983,
"rewards/rejected": 0.0,
"step": 713
},
{
"epoch": 3.988826815642458,
"grad_norm": 6.10877500712863,
"learning_rate": 9.337726203681628e-07,
"logits/chosen": -3.6558167934417725,
"logits/rejected": -3.7465085983276367,
"logps/chosen": -9.029916763305664,
"logps/rejected": -35.61157989501953,
"loss": 0.1821,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6430715322494507,
"rewards/margins": 1.6430715322494507,
"rewards/rejected": 0.0,
"step": 714
},
{
"epoch": 3.994413407821229,
"grad_norm": 5.645512806316481,
"learning_rate": 9.334691360805515e-07,
"logits/chosen": -3.5487918853759766,
"logits/rejected": -3.520137310028076,
"logps/chosen": -22.947519302368164,
"logps/rejected": -24.66504669189453,
"loss": 0.1675,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2151451110839844,
"rewards/margins": 2.2151451110839844,
"rewards/rejected": 0.0,
"step": 715
},
{
"epoch": 4.0,
"grad_norm": 3.3119225096540705,
"learning_rate": 9.33165007580607e-07,
"logits/chosen": -3.453325033187866,
"logits/rejected": -3.3989417552948,
"logps/chosen": -10.77746868133545,
"logps/rejected": -19.99879264831543,
"loss": 0.1574,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.782940149307251,
"rewards/margins": 1.782940149307251,
"rewards/rejected": 0.0,
"step": 716
},
{
"epoch": 4.005586592178771,
"grad_norm": 3.1066199059131994,
"learning_rate": 9.32860235320319e-07,
"logits/chosen": -3.86627197265625,
"logits/rejected": -3.867809534072876,
"logps/chosen": -7.358992099761963,
"logps/rejected": -28.105648040771484,
"loss": 0.1327,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9243531227111816,
"rewards/margins": 1.9243531227111816,
"rewards/rejected": 0.0,
"step": 717
},
{
"epoch": 4.011173184357542,
"grad_norm": 3.139164996115159,
"learning_rate": 9.325548197526329e-07,
"logits/chosen": -3.582491874694824,
"logits/rejected": -3.5838065147399902,
"logps/chosen": -12.756607055664062,
"logps/rejected": -31.588409423828125,
"loss": 0.1393,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5849655866622925,
"rewards/margins": 1.5849655866622925,
"rewards/rejected": 0.0,
"step": 718
},
{
"epoch": 4.016759776536313,
"grad_norm": 2.8827433677868246,
"learning_rate": 9.322487613314511e-07,
"logits/chosen": -3.928121566772461,
"logits/rejected": -3.5920283794403076,
"logps/chosen": -5.253528594970703,
"logps/rejected": -29.994462966918945,
"loss": 0.1629,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.885958194732666,
"rewards/margins": 1.885958194732666,
"rewards/rejected": 0.0,
"step": 719
},
{
"epoch": 4.022346368715084,
"grad_norm": 4.710942819453221,
"learning_rate": 9.319420605116306e-07,
"logits/chosen": -3.8790347576141357,
"logits/rejected": -3.5468173027038574,
"logps/chosen": -5.437746524810791,
"logps/rejected": -17.863117218017578,
"loss": 0.1485,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.861025333404541,
"rewards/margins": 1.861025333404541,
"rewards/rejected": 0.0,
"step": 720
},
{
"epoch": 4.022346368715084,
"eval_logits/chosen": -3.4488277435302734,
"eval_logits/rejected": -3.556626796722412,
"eval_logps/chosen": -19.6168270111084,
"eval_logps/rejected": -34.507118225097656,
"eval_loss": 0.39809659123420715,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 1.0453212261199951,
"eval_rewards/margins": 1.0453212261199951,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7389,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 720
},
{
"epoch": 4.027932960893855,
"grad_norm": 2.705266372688293,
"learning_rate": 9.316347177489834e-07,
"logits/chosen": -3.5220789909362793,
"logits/rejected": -3.2003581523895264,
"logps/chosen": -10.75131607055664,
"logps/rejected": -18.33860969543457,
"loss": 0.1392,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5228328704833984,
"rewards/margins": 2.5228328704833984,
"rewards/rejected": 0.0,
"step": 721
},
{
"epoch": 4.033519553072626,
"grad_norm": 2.790758874166294,
"learning_rate": 9.313267335002758e-07,
"logits/chosen": -3.6527023315429688,
"logits/rejected": -3.6668760776519775,
"logps/chosen": -13.89764404296875,
"logps/rejected": -54.01905822753906,
"loss": 0.1253,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4421184062957764,
"rewards/margins": 2.4421184062957764,
"rewards/rejected": 0.0,
"step": 722
},
{
"epoch": 4.039106145251397,
"grad_norm": 3.0446453914168727,
"learning_rate": 9.310181082232272e-07,
"logits/chosen": -3.152895927429199,
"logits/rejected": -2.9274024963378906,
"logps/chosen": -18.17841339111328,
"logps/rejected": -21.236652374267578,
"loss": 0.137,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.428865432739258,
"rewards/margins": 2.428865432739258,
"rewards/rejected": 0.0,
"step": 723
},
{
"epoch": 4.044692737430168,
"grad_norm": 3.4419644926394546,
"learning_rate": 9.307088423765095e-07,
"logits/chosen": -3.4919638633728027,
"logits/rejected": -3.390697956085205,
"logps/chosen": -6.289459705352783,
"logps/rejected": -33.2501335144043,
"loss": 0.1375,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4905641078948975,
"rewards/margins": 1.4905641078948975,
"rewards/rejected": 0.0,
"step": 724
},
{
"epoch": 4.050279329608939,
"grad_norm": 3.115348890510477,
"learning_rate": 9.303989364197468e-07,
"logits/chosen": -3.5298280715942383,
"logits/rejected": -3.65049409866333,
"logps/chosen": -8.518899917602539,
"logps/rejected": -64.3823013305664,
"loss": 0.1311,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.480935573577881,
"rewards/margins": 2.480935573577881,
"rewards/rejected": 0.0,
"step": 725
},
{
"epoch": 4.055865921787709,
"grad_norm": 3.074773937376554,
"learning_rate": 9.300883908135151e-07,
"logits/chosen": -3.6087746620178223,
"logits/rejected": -3.4459707736968994,
"logps/chosen": -11.586292266845703,
"logps/rejected": -28.29360008239746,
"loss": 0.1524,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2884156703948975,
"rewards/margins": 2.2884156703948975,
"rewards/rejected": 0.0,
"step": 726
},
{
"epoch": 4.06145251396648,
"grad_norm": 3.724823833704586,
"learning_rate": 9.297772060193399e-07,
"logits/chosen": -3.662719488143921,
"logits/rejected": -3.6215732097625732,
"logps/chosen": -8.40018081665039,
"logps/rejected": -24.1854190826416,
"loss": 0.146,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9744935035705566,
"rewards/margins": 1.9744935035705566,
"rewards/rejected": 0.0,
"step": 727
},
{
"epoch": 4.067039106145251,
"grad_norm": 2.8456632938699813,
"learning_rate": 9.294653824996974e-07,
"logits/chosen": -3.7286922931671143,
"logits/rejected": -3.5332837104797363,
"logps/chosen": -3.7866902351379395,
"logps/rejected": -48.27838134765625,
"loss": 0.1576,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6437840461730957,
"rewards/margins": 1.6437840461730957,
"rewards/rejected": 0.0,
"step": 728
},
{
"epoch": 4.072625698324022,
"grad_norm": 3.3654213272524216,
"learning_rate": 9.291529207180131e-07,
"logits/chosen": -3.652053117752075,
"logits/rejected": -3.7424676418304443,
"logps/chosen": -4.034215927124023,
"logps/rejected": -38.23637390136719,
"loss": 0.1525,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.381847858428955,
"rewards/margins": 1.381847858428955,
"rewards/rejected": 0.0,
"step": 729
},
{
"epoch": 4.078212290502793,
"grad_norm": 2.8708320826782803,
"learning_rate": 9.288398211386606e-07,
"logits/chosen": -3.9891786575317383,
"logits/rejected": -3.8775477409362793,
"logps/chosen": -8.785106658935547,
"logps/rejected": -16.385774612426758,
"loss": 0.1338,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3422060012817383,
"rewards/margins": 2.3422060012817383,
"rewards/rejected": 0.0,
"step": 730
},
{
"epoch": 4.083798882681564,
"grad_norm": 2.882045035043276,
"learning_rate": 9.285260842269616e-07,
"logits/chosen": -3.5995490550994873,
"logits/rejected": -3.3151626586914062,
"logps/chosen": -6.383079528808594,
"logps/rejected": -46.785770416259766,
"loss": 0.1506,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5607208013534546,
"rewards/margins": 1.5607208013534546,
"rewards/rejected": 0.0,
"step": 731
},
{
"epoch": 4.089385474860335,
"grad_norm": 3.2785859506970882,
"learning_rate": 9.28211710449185e-07,
"logits/chosen": -3.473612070083618,
"logits/rejected": -3.66225266456604,
"logps/chosen": -2.542301654815674,
"logps/rejected": -73.18181610107422,
"loss": 0.1471,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4122735261917114,
"rewards/margins": 1.4122735261917114,
"rewards/rejected": 0.0,
"step": 732
},
{
"epoch": 4.094972067039106,
"grad_norm": 2.974385466628785,
"learning_rate": 9.278967002725464e-07,
"logits/chosen": -3.808043956756592,
"logits/rejected": -3.915055513381958,
"logps/chosen": -11.477127075195312,
"logps/rejected": -49.63106155395508,
"loss": 0.1653,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7113841772079468,
"rewards/margins": 1.7113841772079468,
"rewards/rejected": 0.0,
"step": 733
},
{
"epoch": 4.100558659217877,
"grad_norm": 2.511075806636554,
"learning_rate": 9.275810541652071e-07,
"logits/chosen": -3.7008209228515625,
"logits/rejected": -3.6690826416015625,
"logps/chosen": -5.514899253845215,
"logps/rejected": -39.429718017578125,
"loss": 0.115,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1070127487182617,
"rewards/margins": 2.1070127487182617,
"rewards/rejected": 0.0,
"step": 734
},
{
"epoch": 4.106145251396648,
"grad_norm": 3.211602338759111,
"learning_rate": 9.272647725962732e-07,
"logits/chosen": -3.5125110149383545,
"logits/rejected": -3.681480646133423,
"logps/chosen": -32.70913314819336,
"logps/rejected": -31.92466926574707,
"loss": 0.1502,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6375980377197266,
"rewards/margins": 2.6375980377197266,
"rewards/rejected": 0.0,
"step": 735
},
{
"epoch": 4.111731843575419,
"grad_norm": 3.522666740853587,
"learning_rate": 9.269478560357956e-07,
"logits/chosen": -3.8538997173309326,
"logits/rejected": -3.648585796356201,
"logps/chosen": -19.528947830200195,
"logps/rejected": -23.962234497070312,
"loss": 0.1666,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0669362545013428,
"rewards/margins": 2.0669362545013428,
"rewards/rejected": 0.0,
"step": 736
},
{
"epoch": 4.11731843575419,
"grad_norm": 3.204503869499785,
"learning_rate": 9.266303049547687e-07,
"logits/chosen": -3.47011137008667,
"logits/rejected": -3.659250259399414,
"logps/chosen": -10.742396354675293,
"logps/rejected": -27.028169631958008,
"loss": 0.1412,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4456121921539307,
"rewards/margins": 2.4456121921539307,
"rewards/rejected": 0.0,
"step": 737
},
{
"epoch": 4.122905027932961,
"grad_norm": 3.167416898284249,
"learning_rate": 9.263121198251298e-07,
"logits/chosen": -3.63665771484375,
"logits/rejected": -3.7554972171783447,
"logps/chosen": -8.60072135925293,
"logps/rejected": -38.334529876708984,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9187707901000977,
"rewards/margins": 1.9187707901000977,
"rewards/rejected": 0.0,
"step": 738
},
{
"epoch": 4.128491620111732,
"grad_norm": 6.0027105137760906,
"learning_rate": 9.25993301119759e-07,
"logits/chosen": -3.747448444366455,
"logits/rejected": -3.8216421604156494,
"logps/chosen": -6.165863990783691,
"logps/rejected": -30.87430191040039,
"loss": 0.1472,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2946873903274536,
"rewards/margins": 1.2946873903274536,
"rewards/rejected": 0.0,
"step": 739
},
{
"epoch": 4.134078212290503,
"grad_norm": 2.9984148756243285,
"learning_rate": 9.256738493124773e-07,
"logits/chosen": -3.3836264610290527,
"logits/rejected": -3.621556520462036,
"logps/chosen": -24.779489517211914,
"logps/rejected": -45.01148223876953,
"loss": 0.1312,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5112781524658203,
"rewards/margins": 2.5112781524658203,
"rewards/rejected": 0.0,
"step": 740
},
{
"epoch": 4.134078212290503,
"eval_logits/chosen": -3.4461867809295654,
"eval_logits/rejected": -3.5609054565429688,
"eval_logps/chosen": -20.51335906982422,
"eval_logps/rejected": -37.09926986694336,
"eval_loss": 0.45483899116516113,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9556684494018555,
"eval_rewards/margins": 0.9556684494018555,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7462,
"eval_samples_per_second": 9.467,
"eval_steps_per_second": 0.305,
"step": 740
},
{
"epoch": 4.139664804469274,
"grad_norm": 3.6893463435658416,
"learning_rate": 9.253537648780472e-07,
"logits/chosen": -3.7558329105377197,
"logits/rejected": -3.8279576301574707,
"logps/chosen": -14.115180969238281,
"logps/rejected": -22.908321380615234,
"loss": 0.1611,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6616134643554688,
"rewards/margins": 2.6616134643554688,
"rewards/rejected": 0.0,
"step": 741
},
{
"epoch": 4.145251396648045,
"grad_norm": 3.81926847554455,
"learning_rate": 9.250330482921711e-07,
"logits/chosen": -3.778927803039551,
"logits/rejected": -3.671673536300659,
"logps/chosen": -1.6724507808685303,
"logps/rejected": -75.70147705078125,
"loss": 0.1529,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1681830883026123,
"rewards/margins": 1.1681830883026123,
"rewards/rejected": 0.0,
"step": 742
},
{
"epoch": 4.150837988826815,
"grad_norm": 6.57804840798006,
"learning_rate": 9.24711700031491e-07,
"logits/chosen": -3.6145131587982178,
"logits/rejected": -3.5747292041778564,
"logps/chosen": -13.17055892944336,
"logps/rejected": -25.868175506591797,
"loss": 0.1455,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.250964403152466,
"rewards/margins": 2.250964403152466,
"rewards/rejected": 0.0,
"step": 743
},
{
"epoch": 4.156424581005586,
"grad_norm": 2.9823534594630368,
"learning_rate": 9.243897205735878e-07,
"logits/chosen": -3.810460090637207,
"logits/rejected": -3.8142824172973633,
"logps/chosen": -3.4041643142700195,
"logps/rejected": -65.01181030273438,
"loss": 0.1223,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5365550518035889,
"rewards/margins": 1.5365550518035889,
"rewards/rejected": 0.0,
"step": 744
},
{
"epoch": 4.162011173184357,
"grad_norm": 3.0972227791265206,
"learning_rate": 9.2406711039698e-07,
"logits/chosen": -3.4885849952697754,
"logits/rejected": -3.400820255279541,
"logps/chosen": -3.5544004440307617,
"logps/rejected": -31.687374114990234,
"loss": 0.1193,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8177826404571533,
"rewards/margins": 1.8177826404571533,
"rewards/rejected": 0.0,
"step": 745
},
{
"epoch": 4.167597765363128,
"grad_norm": 3.0874643852325176,
"learning_rate": 9.237438699811238e-07,
"logits/chosen": -3.5004031658172607,
"logits/rejected": -3.6086864471435547,
"logps/chosen": -8.36198616027832,
"logps/rejected": -24.751819610595703,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.266913414001465,
"rewards/margins": 2.266913414001465,
"rewards/rejected": 0.0,
"step": 746
},
{
"epoch": 4.173184357541899,
"grad_norm": 3.213656743187189,
"learning_rate": 9.234199998064124e-07,
"logits/chosen": -3.846193313598633,
"logits/rejected": -3.6449530124664307,
"logps/chosen": -8.9894380569458,
"logps/rejected": -40.17134094238281,
"loss": 0.1473,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7603962421417236,
"rewards/margins": 1.7603962421417236,
"rewards/rejected": 0.0,
"step": 747
},
{
"epoch": 4.17877094972067,
"grad_norm": 3.039480742921061,
"learning_rate": 9.230955003541742e-07,
"logits/chosen": -3.5335752964019775,
"logits/rejected": -3.739990711212158,
"logps/chosen": -2.6747734546661377,
"logps/rejected": -38.45859909057617,
"loss": 0.1374,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.272268295288086,
"rewards/margins": 1.272268295288086,
"rewards/rejected": 0.0,
"step": 748
},
{
"epoch": 4.184357541899441,
"grad_norm": 3.1356275674508804,
"learning_rate": 9.227703721066733e-07,
"logits/chosen": -3.3352243900299072,
"logits/rejected": -3.300250768661499,
"logps/chosen": -21.90550994873047,
"logps/rejected": -33.50829315185547,
"loss": 0.1298,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1247329711914062,
"rewards/margins": 2.1247329711914062,
"rewards/rejected": 0.0,
"step": 749
},
{
"epoch": 4.189944134078212,
"grad_norm": 3.3401965954263444,
"learning_rate": 9.224446155471082e-07,
"logits/chosen": -3.4573476314544678,
"logits/rejected": -3.706101179122925,
"logps/chosen": -42.64149856567383,
"logps/rejected": -27.08307456970215,
"loss": 0.1482,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8673996925354004,
"rewards/margins": 2.8673996925354004,
"rewards/rejected": 0.0,
"step": 750
},
{
"epoch": 4.195530726256983,
"grad_norm": 3.8568903726565766,
"learning_rate": 9.221182311596111e-07,
"logits/chosen": -3.56864070892334,
"logits/rejected": -3.4958670139312744,
"logps/chosen": -19.352327346801758,
"logps/rejected": -35.47402572631836,
"loss": 0.1408,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6086671352386475,
"rewards/margins": 2.6086671352386475,
"rewards/rejected": 0.0,
"step": 751
},
{
"epoch": 4.201117318435754,
"grad_norm": 4.538585470300417,
"learning_rate": 9.217912194292473e-07,
"logits/chosen": -3.2495787143707275,
"logits/rejected": -3.3900630474090576,
"logps/chosen": -3.3145623207092285,
"logps/rejected": -51.09520721435547,
"loss": 0.1361,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7654829025268555,
"rewards/margins": 1.7654829025268555,
"rewards/rejected": 0.0,
"step": 752
},
{
"epoch": 4.206703910614525,
"grad_norm": 5.442019841494312,
"learning_rate": 9.214635808420146e-07,
"logits/chosen": -3.792428731918335,
"logits/rejected": -3.8434760570526123,
"logps/chosen": -12.232793807983398,
"logps/rejected": -32.03156661987305,
"loss": 0.15,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4639711380004883,
"rewards/margins": 2.4639711380004883,
"rewards/rejected": 0.0,
"step": 753
},
{
"epoch": 4.212290502793296,
"grad_norm": 3.116250654254641,
"learning_rate": 9.211353158848421e-07,
"logits/chosen": -3.395803689956665,
"logits/rejected": -3.3356313705444336,
"logps/chosen": -11.544515609741211,
"logps/rejected": -31.22180938720703,
"loss": 0.1504,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2545347213745117,
"rewards/margins": 2.2545347213745117,
"rewards/rejected": 0.0,
"step": 754
},
{
"epoch": 4.217877094972067,
"grad_norm": 3.321881698001037,
"learning_rate": 9.208064250455902e-07,
"logits/chosen": -3.676140069961548,
"logits/rejected": -3.8750996589660645,
"logps/chosen": -4.331568717956543,
"logps/rejected": -27.3017578125,
"loss": 0.1454,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8052383661270142,
"rewards/margins": 1.8052383661270142,
"rewards/rejected": 0.0,
"step": 755
},
{
"epoch": 4.223463687150838,
"grad_norm": 3.4889153207537933,
"learning_rate": 9.204769088130491e-07,
"logits/chosen": -3.7074637413024902,
"logits/rejected": -3.801008939743042,
"logps/chosen": -9.631816864013672,
"logps/rejected": -45.95659637451172,
"loss": 0.1231,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.324375629425049,
"rewards/margins": 2.324375629425049,
"rewards/rejected": 0.0,
"step": 756
},
{
"epoch": 4.229050279329609,
"grad_norm": 3.147759262629917,
"learning_rate": 9.201467676769388e-07,
"logits/chosen": -3.383687973022461,
"logits/rejected": -3.453930139541626,
"logps/chosen": -31.3158016204834,
"logps/rejected": -25.914657592773438,
"loss": 0.1345,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4040439128875732,
"rewards/margins": 2.4040439128875732,
"rewards/rejected": 0.0,
"step": 757
},
{
"epoch": 4.23463687150838,
"grad_norm": 3.2858190977651427,
"learning_rate": 9.198160021279076e-07,
"logits/chosen": -3.4089393615722656,
"logits/rejected": -3.566396474838257,
"logps/chosen": -10.551408767700195,
"logps/rejected": -35.372467041015625,
"loss": 0.1546,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9044461250305176,
"rewards/margins": 1.9044461250305176,
"rewards/rejected": 0.0,
"step": 758
},
{
"epoch": 4.240223463687151,
"grad_norm": 3.293436579521988,
"learning_rate": 9.19484612657532e-07,
"logits/chosen": -4.007391929626465,
"logits/rejected": -3.845597505569458,
"logps/chosen": -3.860405683517456,
"logps/rejected": -29.047256469726562,
"loss": 0.1478,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.710767388343811,
"rewards/margins": 1.710767388343811,
"rewards/rejected": 0.0,
"step": 759
},
{
"epoch": 4.245810055865922,
"grad_norm": 3.6355080196955285,
"learning_rate": 9.19152599758316e-07,
"logits/chosen": -3.7784767150878906,
"logits/rejected": -3.559087038040161,
"logps/chosen": -7.721414566040039,
"logps/rejected": -39.091068267822266,
"loss": 0.137,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.056973457336426,
"rewards/margins": 2.056973457336426,
"rewards/rejected": 0.0,
"step": 760
},
{
"epoch": 4.245810055865922,
"eval_logits/chosen": -3.4631340503692627,
"eval_logits/rejected": -3.580005645751953,
"eval_logps/chosen": -20.73528480529785,
"eval_logps/rejected": -37.09739303588867,
"eval_loss": 0.4545563757419586,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9334754943847656,
"eval_rewards/margins": 0.9334754943847656,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6984,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 760
},
{
"epoch": 4.251396648044693,
"grad_norm": 4.944444500555194,
"learning_rate": 9.188199639236896e-07,
"logits/chosen": -3.581573963165283,
"logits/rejected": -3.8438568115234375,
"logps/chosen": -5.980813026428223,
"logps/rejected": -34.75967025756836,
"loss": 0.1302,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1940486431121826,
"rewards/margins": 2.1940486431121826,
"rewards/rejected": 0.0,
"step": 761
},
{
"epoch": 4.256983240223463,
"grad_norm": 3.0382683861843462,
"learning_rate": 9.184867056480092e-07,
"logits/chosen": -3.32403826713562,
"logits/rejected": -3.279866933822632,
"logps/chosen": -15.873481750488281,
"logps/rejected": -32.66768264770508,
"loss": 0.1418,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.380401134490967,
"rewards/margins": 2.380401134490967,
"rewards/rejected": 0.0,
"step": 762
},
{
"epoch": 4.262569832402234,
"grad_norm": 3.525863356895964,
"learning_rate": 9.181528254265558e-07,
"logits/chosen": -3.561990737915039,
"logits/rejected": -3.3689842224121094,
"logps/chosen": -14.201573371887207,
"logps/rejected": -36.46363067626953,
"loss": 0.1306,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.926465392112732,
"rewards/margins": 1.926465392112732,
"rewards/rejected": 0.0,
"step": 763
},
{
"epoch": 4.268156424581005,
"grad_norm": 3.4135660098550615,
"learning_rate": 9.178183237555348e-07,
"logits/chosen": -3.6096136569976807,
"logits/rejected": -3.447056770324707,
"logps/chosen": -4.089483261108398,
"logps/rejected": -35.93095397949219,
"loss": 0.1558,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7403192520141602,
"rewards/margins": 1.7403192520141602,
"rewards/rejected": 0.0,
"step": 764
},
{
"epoch": 4.273743016759776,
"grad_norm": 3.3652695327928948,
"learning_rate": 9.174832011320755e-07,
"logits/chosen": -3.407003164291382,
"logits/rejected": -3.7466068267822266,
"logps/chosen": -4.882349014282227,
"logps/rejected": -75.54251098632812,
"loss": 0.1344,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7903443574905396,
"rewards/margins": 1.7903443574905396,
"rewards/rejected": 0.0,
"step": 765
},
{
"epoch": 4.279329608938547,
"grad_norm": 3.6507183315670475,
"learning_rate": 9.171474580542295e-07,
"logits/chosen": -3.983333110809326,
"logits/rejected": -3.87296199798584,
"logps/chosen": -5.923532485961914,
"logps/rejected": -20.535736083984375,
"loss": 0.1247,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8490628004074097,
"rewards/margins": 1.8490628004074097,
"rewards/rejected": 0.0,
"step": 766
},
{
"epoch": 4.284916201117318,
"grad_norm": 3.592633095633263,
"learning_rate": 9.168110950209709e-07,
"logits/chosen": -3.6785390377044678,
"logits/rejected": -3.9248623847961426,
"logps/chosen": -20.322904586791992,
"logps/rejected": -33.33665466308594,
"loss": 0.1372,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.527787446975708,
"rewards/margins": 2.527787446975708,
"rewards/rejected": 0.0,
"step": 767
},
{
"epoch": 4.290502793296089,
"grad_norm": 3.8296725215002936,
"learning_rate": 9.164741125321951e-07,
"logits/chosen": -3.448434591293335,
"logits/rejected": -3.6976163387298584,
"logps/chosen": -6.484898567199707,
"logps/rejected": -96.32450103759766,
"loss": 0.1548,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.413029432296753,
"rewards/margins": 2.413029432296753,
"rewards/rejected": 0.0,
"step": 768
},
{
"epoch": 4.29608938547486,
"grad_norm": 3.7465486737345475,
"learning_rate": 9.16136511088718e-07,
"logits/chosen": -3.265479803085327,
"logits/rejected": -3.4205503463745117,
"logps/chosen": -9.510969161987305,
"logps/rejected": -42.40364074707031,
"loss": 0.1472,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9431638717651367,
"rewards/margins": 1.9431638717651367,
"rewards/rejected": 0.0,
"step": 769
},
{
"epoch": 4.301675977653631,
"grad_norm": 4.749079193531236,
"learning_rate": 9.157982911922758e-07,
"logits/chosen": -3.686668872833252,
"logits/rejected": -3.5783064365386963,
"logps/chosen": -11.030715942382812,
"logps/rejected": -29.405502319335938,
"loss": 0.1341,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.23626971244812,
"rewards/margins": 2.23626971244812,
"rewards/rejected": 0.0,
"step": 770
},
{
"epoch": 4.307262569832402,
"grad_norm": 3.234669849885449,
"learning_rate": 9.154594533455231e-07,
"logits/chosen": -3.8428173065185547,
"logits/rejected": -3.558997869491577,
"logps/chosen": -3.6590447425842285,
"logps/rejected": -25.97920799255371,
"loss": 0.1443,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.66550874710083,
"rewards/margins": 1.66550874710083,
"rewards/rejected": 0.0,
"step": 771
},
{
"epoch": 4.312849162011173,
"grad_norm": 3.2309803622647215,
"learning_rate": 9.151199980520335e-07,
"logits/chosen": -3.673877477645874,
"logits/rejected": -3.547835350036621,
"logps/chosen": -11.635046005249023,
"logps/rejected": -38.26511001586914,
"loss": 0.1473,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.192197561264038,
"rewards/margins": 2.192197561264038,
"rewards/rejected": 0.0,
"step": 772
},
{
"epoch": 4.318435754189944,
"grad_norm": 2.866293349271251,
"learning_rate": 9.147799258162981e-07,
"logits/chosen": -3.293854236602783,
"logits/rejected": -3.5211727619171143,
"logps/chosen": -8.397584915161133,
"logps/rejected": -41.11590576171875,
"loss": 0.1224,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8701279163360596,
"rewards/margins": 1.8701279163360596,
"rewards/rejected": 0.0,
"step": 773
},
{
"epoch": 4.324022346368715,
"grad_norm": 3.5849916471280334,
"learning_rate": 9.144392371437244e-07,
"logits/chosen": -3.6938962936401367,
"logits/rejected": -3.7141709327697754,
"logps/chosen": -4.073142051696777,
"logps/rejected": -26.94833755493164,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8245025873184204,
"rewards/margins": 1.8245025873184204,
"rewards/rejected": 0.0,
"step": 774
},
{
"epoch": 4.329608938547486,
"grad_norm": 2.7710639424681016,
"learning_rate": 9.140979325406369e-07,
"logits/chosen": -3.309889554977417,
"logits/rejected": -3.4348950386047363,
"logps/chosen": -4.105931758880615,
"logps/rejected": -58.37015914916992,
"loss": 0.1559,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.727717638015747,
"rewards/margins": 1.727717638015747,
"rewards/rejected": 0.0,
"step": 775
},
{
"epoch": 4.335195530726257,
"grad_norm": 4.0928340001123775,
"learning_rate": 9.137560125142748e-07,
"logits/chosen": -3.6457533836364746,
"logits/rejected": -3.8680081367492676,
"logps/chosen": -5.9027228355407715,
"logps/rejected": -48.547664642333984,
"loss": 0.1387,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5917106866836548,
"rewards/margins": 1.5917106866836548,
"rewards/rejected": 0.0,
"step": 776
},
{
"epoch": 4.340782122905028,
"grad_norm": 3.2766470621291197,
"learning_rate": 9.134134775727921e-07,
"logits/chosen": -3.70273756980896,
"logits/rejected": -3.5022263526916504,
"logps/chosen": -6.853073596954346,
"logps/rejected": -37.53110885620117,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.01503324508667,
"rewards/margins": 2.01503324508667,
"rewards/rejected": 0.0,
"step": 777
},
{
"epoch": 4.346368715083799,
"grad_norm": 4.640038749004769,
"learning_rate": 9.13070328225257e-07,
"logits/chosen": -3.47983455657959,
"logits/rejected": -3.3371739387512207,
"logps/chosen": -7.3134355545043945,
"logps/rejected": -32.42387390136719,
"loss": 0.1463,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5095090866088867,
"rewards/margins": 2.5095090866088867,
"rewards/rejected": 0.0,
"step": 778
},
{
"epoch": 4.351955307262569,
"grad_norm": 3.844517908712181,
"learning_rate": 9.127265649816503e-07,
"logits/chosen": -3.4332523345947266,
"logits/rejected": -3.6509153842926025,
"logps/chosen": -13.059277534484863,
"logps/rejected": -48.963497161865234,
"loss": 0.1535,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9094536304473877,
"rewards/margins": 1.9094536304473877,
"rewards/rejected": 0.0,
"step": 779
},
{
"epoch": 4.35754189944134,
"grad_norm": 2.9517341839947115,
"learning_rate": 9.123821883528653e-07,
"logits/chosen": -3.7786710262298584,
"logits/rejected": -3.6058902740478516,
"logps/chosen": -2.9812262058258057,
"logps/rejected": -54.710079193115234,
"loss": 0.1264,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6125730276107788,
"rewards/margins": 1.6125730276107788,
"rewards/rejected": 0.0,
"step": 780
},
{
"epoch": 4.35754189944134,
"eval_logits/chosen": -3.43599271774292,
"eval_logits/rejected": -3.5524775981903076,
"eval_logps/chosen": -20.9012393951416,
"eval_logps/rejected": -38.057952880859375,
"eval_loss": 0.4690379500389099,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9168804883956909,
"eval_rewards/margins": 0.9168804883956909,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7009,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 780
},
{
"epoch": 4.363128491620111,
"grad_norm": 5.343100546730451,
"learning_rate": 9.120371988507073e-07,
"logits/chosen": -3.424809217453003,
"logits/rejected": -3.5738799571990967,
"logps/chosen": -35.25627899169922,
"logps/rejected": -46.489288330078125,
"loss": 0.1704,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8060996532440186,
"rewards/margins": 1.8060996532440186,
"rewards/rejected": 0.0,
"step": 781
},
{
"epoch": 4.368715083798882,
"grad_norm": 3.10295127317863,
"learning_rate": 9.11691596987892e-07,
"logits/chosen": -3.82004976272583,
"logits/rejected": -3.744163751602173,
"logps/chosen": -18.236862182617188,
"logps/rejected": -40.948020935058594,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6354761123657227,
"rewards/margins": 2.6354761123657227,
"rewards/rejected": 0.0,
"step": 782
},
{
"epoch": 4.374301675977653,
"grad_norm": 6.635272543839055,
"learning_rate": 9.113453832780452e-07,
"logits/chosen": -3.6957621574401855,
"logits/rejected": -3.819979429244995,
"logps/chosen": -11.850873947143555,
"logps/rejected": -44.036354064941406,
"loss": 0.1585,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.082808494567871,
"rewards/margins": 2.082808494567871,
"rewards/rejected": 0.0,
"step": 783
},
{
"epoch": 4.379888268156424,
"grad_norm": 3.854798709356382,
"learning_rate": 9.109985582357024e-07,
"logits/chosen": -3.4215121269226074,
"logits/rejected": -3.3615832328796387,
"logps/chosen": -9.830392837524414,
"logps/rejected": -46.99778366088867,
"loss": 0.1185,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1329922676086426,
"rewards/margins": 2.1329922676086426,
"rewards/rejected": 0.0,
"step": 784
},
{
"epoch": 4.385474860335195,
"grad_norm": 3.787443142878194,
"learning_rate": 9.106511223763072e-07,
"logits/chosen": -3.693557024002075,
"logits/rejected": -3.639054298400879,
"logps/chosen": -12.23244857788086,
"logps/rejected": -44.668846130371094,
"loss": 0.1374,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.178089141845703,
"rewards/margins": 2.178089141845703,
"rewards/rejected": 0.0,
"step": 785
},
{
"epoch": 4.391061452513966,
"grad_norm": 8.74467544616392,
"learning_rate": 9.10303076216211e-07,
"logits/chosen": -3.5624606609344482,
"logits/rejected": -3.3727633953094482,
"logps/chosen": -10.703680038452148,
"logps/rejected": -34.77155685424805,
"loss": 0.1832,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7109159231185913,
"rewards/margins": 1.7109159231185913,
"rewards/rejected": 0.0,
"step": 786
},
{
"epoch": 4.396648044692737,
"grad_norm": 3.4825215321185357,
"learning_rate": 9.099544202726727e-07,
"logits/chosen": -3.783106803894043,
"logits/rejected": -3.583214282989502,
"logps/chosen": -23.693105697631836,
"logps/rejected": -20.57411003112793,
"loss": 0.136,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.845611333847046,
"rewards/margins": 2.845611333847046,
"rewards/rejected": 0.0,
"step": 787
},
{
"epoch": 4.402234636871508,
"grad_norm": 4.135132450259034,
"learning_rate": 9.096051550638571e-07,
"logits/chosen": -3.5946578979492188,
"logits/rejected": -3.6001033782958984,
"logps/chosen": -41.17884063720703,
"logps/rejected": -32.217872619628906,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4946250915527344,
"rewards/margins": 2.4946250915527344,
"rewards/rejected": 0.0,
"step": 788
},
{
"epoch": 4.407821229050279,
"grad_norm": 8.7635921333923,
"learning_rate": 9.092552811088341e-07,
"logits/chosen": -3.9743850231170654,
"logits/rejected": -3.7202308177948,
"logps/chosen": -10.13995361328125,
"logps/rejected": -32.595420837402344,
"loss": 0.199,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.177818775177002,
"rewards/margins": 2.177818775177002,
"rewards/rejected": 0.0,
"step": 789
},
{
"epoch": 4.41340782122905,
"grad_norm": 4.1633615918267655,
"learning_rate": 9.089047989275792e-07,
"logits/chosen": -3.6054015159606934,
"logits/rejected": -3.681760787963867,
"logps/chosen": -6.8949151039123535,
"logps/rejected": -27.745790481567383,
"loss": 0.1358,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0137648582458496,
"rewards/margins": 2.0137648582458496,
"rewards/rejected": 0.0,
"step": 790
},
{
"epoch": 4.418994413407821,
"grad_norm": 2.9385780581086913,
"learning_rate": 9.085537090409711e-07,
"logits/chosen": -3.7114686965942383,
"logits/rejected": -3.642517328262329,
"logps/chosen": -9.158451080322266,
"logps/rejected": -57.30244064331055,
"loss": 0.1191,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2197651863098145,
"rewards/margins": 2.2197651863098145,
"rewards/rejected": 0.0,
"step": 791
},
{
"epoch": 4.424581005586592,
"grad_norm": 4.421768930792704,
"learning_rate": 9.082020119707919e-07,
"logits/chosen": -3.851954698562622,
"logits/rejected": -3.7828562259674072,
"logps/chosen": -9.329696655273438,
"logps/rejected": -44.28007507324219,
"loss": 0.1303,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.448639392852783,
"rewards/margins": 2.448639392852783,
"rewards/rejected": 0.0,
"step": 792
},
{
"epoch": 4.430167597765363,
"grad_norm": 4.710272386727113,
"learning_rate": 9.078497082397262e-07,
"logits/chosen": -3.7823007106781006,
"logits/rejected": -3.5928187370300293,
"logps/chosen": -12.154468536376953,
"logps/rejected": -22.485620498657227,
"loss": 0.1354,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6742687225341797,
"rewards/margins": 2.6742687225341797,
"rewards/rejected": 0.0,
"step": 793
},
{
"epoch": 4.435754189944134,
"grad_norm": 3.4723079306981233,
"learning_rate": 9.074967983713599e-07,
"logits/chosen": -3.509960174560547,
"logits/rejected": -3.4457461833953857,
"logps/chosen": -21.281532287597656,
"logps/rejected": -21.520544052124023,
"loss": 0.1339,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3013064861297607,
"rewards/margins": 2.3013064861297607,
"rewards/rejected": 0.0,
"step": 794
},
{
"epoch": 4.441340782122905,
"grad_norm": 4.243171534513039,
"learning_rate": 9.071432828901799e-07,
"logits/chosen": -3.7948055267333984,
"logits/rejected": -3.835109233856201,
"logps/chosen": -5.634298324584961,
"logps/rejected": -28.5150089263916,
"loss": 0.1525,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.071786403656006,
"rewards/margins": 2.071786403656006,
"rewards/rejected": 0.0,
"step": 795
},
{
"epoch": 4.446927374301676,
"grad_norm": 4.408833849680875,
"learning_rate": 9.067891623215735e-07,
"logits/chosen": -3.2833240032196045,
"logits/rejected": -3.6118528842926025,
"logps/chosen": -38.4232177734375,
"logps/rejected": -52.00761795043945,
"loss": 0.1572,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1886892318725586,
"rewards/margins": 2.1886892318725586,
"rewards/rejected": 0.0,
"step": 796
},
{
"epoch": 4.452513966480447,
"grad_norm": 3.505128841617623,
"learning_rate": 9.064344371918267e-07,
"logits/chosen": -2.8041765689849854,
"logits/rejected": -2.882662057876587,
"logps/chosen": -13.393410682678223,
"logps/rejected": -55.70225143432617,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9918265342712402,
"rewards/margins": 1.9918265342712402,
"rewards/rejected": 0.0,
"step": 797
},
{
"epoch": 4.4581005586592175,
"grad_norm": 4.033312450932131,
"learning_rate": 9.06079108028124e-07,
"logits/chosen": -3.5890114307403564,
"logits/rejected": -3.583430051803589,
"logps/chosen": -15.296281814575195,
"logps/rejected": -51.344356536865234,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.810650587081909,
"rewards/margins": 2.810650587081909,
"rewards/rejected": 0.0,
"step": 798
},
{
"epoch": 4.4636871508379885,
"grad_norm": 3.3631755261112306,
"learning_rate": 9.057231753585483e-07,
"logits/chosen": -3.6827738285064697,
"logits/rejected": -3.0112385749816895,
"logps/chosen": -14.723047256469727,
"logps/rejected": -47.55908966064453,
"loss": 0.1522,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.315462589263916,
"rewards/margins": 2.315462589263916,
"rewards/rejected": 0.0,
"step": 799
},
{
"epoch": 4.4692737430167595,
"grad_norm": 5.462703338360426,
"learning_rate": 9.053666397120785e-07,
"logits/chosen": -3.467860221862793,
"logits/rejected": -3.461380958557129,
"logps/chosen": -5.588186264038086,
"logps/rejected": -35.241554260253906,
"loss": 0.1676,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9312646389007568,
"rewards/margins": 1.9312646389007568,
"rewards/rejected": 0.0,
"step": 800
},
{
"epoch": 4.4692737430167595,
"eval_logits/chosen": -3.439868927001953,
"eval_logits/rejected": -3.557157516479492,
"eval_logps/chosen": -20.577713012695312,
"eval_logps/rejected": -36.924678802490234,
"eval_loss": 0.4478033185005188,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9492327570915222,
"eval_rewards/margins": 0.9492327570915222,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6994,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 800
},
{
"epoch": 4.4748603351955305,
"grad_norm": 2.7917272173132552,
"learning_rate": 9.050095016185902e-07,
"logits/chosen": -3.5902280807495117,
"logits/rejected": -3.9027185440063477,
"logps/chosen": -15.269607543945312,
"logps/rejected": -32.67121124267578,
"loss": 0.1222,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.40834903717041,
"rewards/margins": 2.40834903717041,
"rewards/rejected": 0.0,
"step": 801
},
{
"epoch": 4.4804469273743015,
"grad_norm": 4.542744408179407,
"learning_rate": 9.046517616088544e-07,
"logits/chosen": -3.6178178787231445,
"logits/rejected": -3.6688997745513916,
"logps/chosen": -8.076221466064453,
"logps/rejected": -34.97129440307617,
"loss": 0.1415,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.155116558074951,
"rewards/margins": 2.155116558074951,
"rewards/rejected": 0.0,
"step": 802
},
{
"epoch": 4.4860335195530725,
"grad_norm": 3.7891416567916476,
"learning_rate": 9.042934202145362e-07,
"logits/chosen": -3.5725836753845215,
"logits/rejected": -3.2765321731567383,
"logps/chosen": -20.888517379760742,
"logps/rejected": -30.360881805419922,
"loss": 0.1447,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8129642009735107,
"rewards/margins": 2.8129642009735107,
"rewards/rejected": 0.0,
"step": 803
},
{
"epoch": 4.4916201117318435,
"grad_norm": 3.25904645570946,
"learning_rate": 9.039344779681949e-07,
"logits/chosen": -3.187351942062378,
"logits/rejected": -3.4457833766937256,
"logps/chosen": -4.032071590423584,
"logps/rejected": -53.12700271606445,
"loss": 0.1552,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8995722532272339,
"rewards/margins": 1.8995722532272339,
"rewards/rejected": 0.0,
"step": 804
},
{
"epoch": 4.4972067039106145,
"grad_norm": 2.635557739396476,
"learning_rate": 9.035749354032825e-07,
"logits/chosen": -3.5149707794189453,
"logits/rejected": -3.59647536277771,
"logps/chosen": -5.334439277648926,
"logps/rejected": -32.08082580566406,
"loss": 0.1218,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8468266725540161,
"rewards/margins": 1.8468266725540161,
"rewards/rejected": 0.0,
"step": 805
},
{
"epoch": 4.5027932960893855,
"grad_norm": 5.125379045597751,
"learning_rate": 9.032147930541434e-07,
"logits/chosen": -3.469554901123047,
"logits/rejected": -3.7115323543548584,
"logps/chosen": -8.944414138793945,
"logps/rejected": -29.938426971435547,
"loss": 0.1352,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.080336570739746,
"rewards/margins": 2.080336570739746,
"rewards/rejected": 0.0,
"step": 806
},
{
"epoch": 4.5083798882681565,
"grad_norm": 4.342979054517213,
"learning_rate": 9.028540514560131e-07,
"logits/chosen": -3.495372772216797,
"logits/rejected": -3.6461708545684814,
"logps/chosen": -6.056365489959717,
"logps/rejected": -36.85601043701172,
"loss": 0.1792,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.217763900756836,
"rewards/margins": 2.217763900756836,
"rewards/rejected": 0.0,
"step": 807
},
{
"epoch": 4.5139664804469275,
"grad_norm": 6.01102840188483,
"learning_rate": 9.02492711145018e-07,
"logits/chosen": -3.437669038772583,
"logits/rejected": -3.7074410915374756,
"logps/chosen": -77.30291748046875,
"logps/rejected": -25.93654441833496,
"loss": 0.152,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7220306396484375,
"rewards/margins": 2.7220306396484375,
"rewards/rejected": 0.0,
"step": 808
},
{
"epoch": 4.5195530726256985,
"grad_norm": 11.144390793167407,
"learning_rate": 9.021307726581742e-07,
"logits/chosen": -3.0836281776428223,
"logits/rejected": -3.0821447372436523,
"logps/chosen": -31.108070373535156,
"logps/rejected": -50.718505859375,
"loss": 0.1831,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1599106788635254,
"rewards/margins": 3.1599106788635254,
"rewards/rejected": 0.0,
"step": 809
},
{
"epoch": 4.5251396648044695,
"grad_norm": 6.333980529322715,
"learning_rate": 9.017682365333866e-07,
"logits/chosen": -3.517589569091797,
"logits/rejected": -3.588407278060913,
"logps/chosen": -7.623435974121094,
"logps/rejected": -25.48214340209961,
"loss": 0.134,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.081376552581787,
"rewards/margins": 2.081376552581787,
"rewards/rejected": 0.0,
"step": 810
},
{
"epoch": 4.5307262569832405,
"grad_norm": 3.1007002978569576,
"learning_rate": 9.014051033094483e-07,
"logits/chosen": -3.4597530364990234,
"logits/rejected": -3.6112377643585205,
"logps/chosen": -10.739227294921875,
"logps/rejected": -32.503944396972656,
"loss": 0.1334,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9882330894470215,
"rewards/margins": 1.9882330894470215,
"rewards/rejected": 0.0,
"step": 811
},
{
"epoch": 4.5363128491620115,
"grad_norm": 4.298191902823217,
"learning_rate": 9.010413735260404e-07,
"logits/chosen": -3.3846542835235596,
"logits/rejected": -3.2056400775909424,
"logps/chosen": -14.382534980773926,
"logps/rejected": -53.71867752075195,
"loss": 0.1465,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1941494941711426,
"rewards/margins": 2.1941494941711426,
"rewards/rejected": 0.0,
"step": 812
},
{
"epoch": 4.5418994413407825,
"grad_norm": 6.505601145909155,
"learning_rate": 9.006770477237298e-07,
"logits/chosen": -3.448770523071289,
"logits/rejected": -3.4331557750701904,
"logps/chosen": -4.308927536010742,
"logps/rejected": -28.490875244140625,
"loss": 0.1749,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4533464908599854,
"rewards/margins": 1.4533464908599854,
"rewards/rejected": 0.0,
"step": 813
},
{
"epoch": 4.547486033519553,
"grad_norm": 6.117768020643388,
"learning_rate": 9.003121264439696e-07,
"logits/chosen": -3.6977038383483887,
"logits/rejected": -3.5017526149749756,
"logps/chosen": -7.282244682312012,
"logps/rejected": -50.60511779785156,
"loss": 0.1413,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1659576892852783,
"rewards/margins": 1.1659576892852783,
"rewards/rejected": 0.0,
"step": 814
},
{
"epoch": 4.553072625698324,
"grad_norm": 3.6236896664606175,
"learning_rate": 8.999466102290978e-07,
"logits/chosen": -3.7826175689697266,
"logits/rejected": -3.6877007484436035,
"logps/chosen": -8.810380935668945,
"logps/rejected": -30.6740779876709,
"loss": 0.1511,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8914289474487305,
"rewards/margins": 1.8914289474487305,
"rewards/rejected": 0.0,
"step": 815
},
{
"epoch": 4.558659217877095,
"grad_norm": 6.724955984642115,
"learning_rate": 8.995804996223366e-07,
"logits/chosen": -3.9157938957214355,
"logits/rejected": -3.9207956790924072,
"logps/chosen": -15.186198234558105,
"logps/rejected": -35.508544921875,
"loss": 0.149,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9603586196899414,
"rewards/margins": 1.9603586196899414,
"rewards/rejected": 0.0,
"step": 816
},
{
"epoch": 4.564245810055866,
"grad_norm": 7.206472291417148,
"learning_rate": 8.992137951677916e-07,
"logits/chosen": -3.4340057373046875,
"logits/rejected": -3.5320987701416016,
"logps/chosen": -3.658921957015991,
"logps/rejected": -24.223121643066406,
"loss": 0.1704,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5331815481185913,
"rewards/margins": 1.5331815481185913,
"rewards/rejected": 0.0,
"step": 817
},
{
"epoch": 4.569832402234637,
"grad_norm": 3.2631195235899497,
"learning_rate": 8.988464974104509e-07,
"logits/chosen": -3.5969433784484863,
"logits/rejected": -3.9172513484954834,
"logps/chosen": -5.465658187866211,
"logps/rejected": -56.673118591308594,
"loss": 0.1523,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8671543598175049,
"rewards/margins": 1.8671543598175049,
"rewards/rejected": 0.0,
"step": 818
},
{
"epoch": 4.575418994413408,
"grad_norm": 3.32968828017197,
"learning_rate": 8.984786068961843e-07,
"logits/chosen": -3.925957202911377,
"logits/rejected": -3.8275678157806396,
"logps/chosen": -14.856788635253906,
"logps/rejected": -25.89576530456543,
"loss": 0.1508,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.517965793609619,
"rewards/margins": 2.517965793609619,
"rewards/rejected": 0.0,
"step": 819
},
{
"epoch": 4.581005586592179,
"grad_norm": 3.693836941985265,
"learning_rate": 8.981101241717427e-07,
"logits/chosen": -3.3613967895507812,
"logits/rejected": -3.5036234855651855,
"logps/chosen": -5.705307483673096,
"logps/rejected": -23.447486877441406,
"loss": 0.1669,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9145679473876953,
"rewards/margins": 1.9145679473876953,
"rewards/rejected": 0.0,
"step": 820
},
{
"epoch": 4.581005586592179,
"eval_logits/chosen": -3.43200945854187,
"eval_logits/rejected": -3.550126552581787,
"eval_logps/chosen": -20.60776138305664,
"eval_logps/rejected": -36.83216094970703,
"eval_loss": 0.45647016167640686,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9462281465530396,
"eval_rewards/margins": 0.9462281465530396,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7826,
"eval_samples_per_second": 9.456,
"eval_steps_per_second": 0.305,
"step": 820
},
{
"epoch": 4.58659217877095,
"grad_norm": 3.2726788992489713,
"learning_rate": 8.977410497847571e-07,
"logits/chosen": -3.7218239307403564,
"logits/rejected": -3.643237590789795,
"logps/chosen": -11.734451293945312,
"logps/rejected": -33.06830596923828,
"loss": 0.1304,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9959282875061035,
"rewards/margins": 1.9959282875061035,
"rewards/rejected": 0.0,
"step": 821
},
{
"epoch": 4.592178770949721,
"grad_norm": 3.8309460160415325,
"learning_rate": 8.973713842837378e-07,
"logits/chosen": -3.600090503692627,
"logits/rejected": -3.7199010848999023,
"logps/chosen": -11.833852767944336,
"logps/rejected": -28.740642547607422,
"loss": 0.1357,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4016051292419434,
"rewards/margins": 2.4016051292419434,
"rewards/rejected": 0.0,
"step": 822
},
{
"epoch": 4.597765363128492,
"grad_norm": 3.698424680136248,
"learning_rate": 8.970011282180733e-07,
"logits/chosen": -3.499113082885742,
"logits/rejected": -3.7541868686676025,
"logps/chosen": -24.649921417236328,
"logps/rejected": -37.322784423828125,
"loss": 0.1373,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4891157150268555,
"rewards/margins": 2.4891157150268555,
"rewards/rejected": 0.0,
"step": 823
},
{
"epoch": 4.603351955307263,
"grad_norm": 3.314445175410359,
"learning_rate": 8.966302821380302e-07,
"logits/chosen": -3.5853443145751953,
"logits/rejected": -3.616360902786255,
"logps/chosen": -7.119479656219482,
"logps/rejected": -29.054214477539062,
"loss": 0.1364,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3403353691101074,
"rewards/margins": 2.3403353691101074,
"rewards/rejected": 0.0,
"step": 824
},
{
"epoch": 4.608938547486034,
"grad_norm": 3.813873948837155,
"learning_rate": 8.96258846594752e-07,
"logits/chosen": -3.519484281539917,
"logits/rejected": -3.314337730407715,
"logps/chosen": -3.3653411865234375,
"logps/rejected": -64.49395751953125,
"loss": 0.1381,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6540875434875488,
"rewards/margins": 1.6540875434875488,
"rewards/rejected": 0.0,
"step": 825
},
{
"epoch": 4.614525139664805,
"grad_norm": 4.044003592559312,
"learning_rate": 8.95886822140258e-07,
"logits/chosen": -3.553804874420166,
"logits/rejected": -3.6048622131347656,
"logps/chosen": -11.220242500305176,
"logps/rejected": -12.231485366821289,
"loss": 0.1332,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2462263107299805,
"rewards/margins": 2.2462263107299805,
"rewards/rejected": 0.0,
"step": 826
},
{
"epoch": 4.620111731843576,
"grad_norm": 3.0775225868967766,
"learning_rate": 8.955142093274426e-07,
"logits/chosen": -3.465428352355957,
"logits/rejected": -3.665736675262451,
"logps/chosen": -5.22053337097168,
"logps/rejected": -22.805583953857422,
"loss": 0.1119,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.120401382446289,
"rewards/margins": 2.120401382446289,
"rewards/rejected": 0.0,
"step": 827
},
{
"epoch": 4.625698324022347,
"grad_norm": 3.165298406526679,
"learning_rate": 8.951410087100749e-07,
"logits/chosen": -3.749488353729248,
"logits/rejected": -3.714769124984741,
"logps/chosen": -2.7161521911621094,
"logps/rejected": -33.90278244018555,
"loss": 0.1542,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5537762641906738,
"rewards/margins": 1.5537762641906738,
"rewards/rejected": 0.0,
"step": 828
},
{
"epoch": 4.631284916201118,
"grad_norm": 7.899198610372127,
"learning_rate": 8.947672208427976e-07,
"logits/chosen": -3.750854730606079,
"logits/rejected": -3.6782498359680176,
"logps/chosen": -6.590742588043213,
"logps/rejected": -46.52386474609375,
"loss": 0.1743,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.104245662689209,
"rewards/margins": 2.104245662689209,
"rewards/rejected": 0.0,
"step": 829
},
{
"epoch": 4.636871508379889,
"grad_norm": 2.6030914754418926,
"learning_rate": 8.943928462811257e-07,
"logits/chosen": -3.698634386062622,
"logits/rejected": -3.8758480548858643,
"logps/chosen": -17.00765609741211,
"logps/rejected": -34.08386993408203,
"loss": 0.1394,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.54410457611084,
"rewards/margins": 2.54410457611084,
"rewards/rejected": 0.0,
"step": 830
},
{
"epoch": 4.64245810055866,
"grad_norm": 4.6086346471003985,
"learning_rate": 8.940178855814468e-07,
"logits/chosen": -3.7106263637542725,
"logits/rejected": -3.6128463745117188,
"logps/chosen": -14.202127456665039,
"logps/rejected": -31.858396530151367,
"loss": 0.126,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.933572769165039,
"rewards/margins": 2.933572769165039,
"rewards/rejected": 0.0,
"step": 831
},
{
"epoch": 4.648044692737431,
"grad_norm": 5.784086044638562,
"learning_rate": 8.936423393010193e-07,
"logits/chosen": -2.882251024246216,
"logits/rejected": -3.0542917251586914,
"logps/chosen": -28.222265243530273,
"logps/rejected": -54.81916427612305,
"loss": 0.1184,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.224271774291992,
"rewards/margins": 2.224271774291992,
"rewards/rejected": 0.0,
"step": 832
},
{
"epoch": 4.653631284916202,
"grad_norm": 3.5986533913630003,
"learning_rate": 8.932662079979717e-07,
"logits/chosen": -3.5953774452209473,
"logits/rejected": -3.6140482425689697,
"logps/chosen": -4.883005619049072,
"logps/rejected": -67.65734100341797,
"loss": 0.1584,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9462710618972778,
"rewards/margins": 1.9462710618972778,
"rewards/rejected": 0.0,
"step": 833
},
{
"epoch": 4.659217877094972,
"grad_norm": 5.502857848087159,
"learning_rate": 8.928894922313022e-07,
"logits/chosen": -3.6708383560180664,
"logits/rejected": -3.9550371170043945,
"logps/chosen": -24.624183654785156,
"logps/rejected": -26.975831985473633,
"loss": 0.1528,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.015517234802246,
"rewards/margins": 2.015517234802246,
"rewards/rejected": 0.0,
"step": 834
},
{
"epoch": 4.664804469273743,
"grad_norm": 4.464777725710083,
"learning_rate": 8.925121925608772e-07,
"logits/chosen": -3.9531307220458984,
"logits/rejected": -3.589585542678833,
"logps/chosen": -9.957625389099121,
"logps/rejected": -56.52036666870117,
"loss": 0.1909,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4638419151306152,
"rewards/margins": 2.4638419151306152,
"rewards/rejected": 0.0,
"step": 835
},
{
"epoch": 4.670391061452514,
"grad_norm": 4.323240389193277,
"learning_rate": 8.921343095474316e-07,
"logits/chosen": -3.4413089752197266,
"logits/rejected": -3.384950876235962,
"logps/chosen": -12.17386245727539,
"logps/rejected": -44.575828552246094,
"loss": 0.148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.343230724334717,
"rewards/margins": 2.343230724334717,
"rewards/rejected": 0.0,
"step": 836
},
{
"epoch": 4.675977653631285,
"grad_norm": 4.871331071988654,
"learning_rate": 8.917558437525667e-07,
"logits/chosen": -3.6823201179504395,
"logits/rejected": -3.872624397277832,
"logps/chosen": -3.832653045654297,
"logps/rejected": -62.0364875793457,
"loss": 0.1318,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7777758836746216,
"rewards/margins": 1.7777758836746216,
"rewards/rejected": 0.0,
"step": 837
},
{
"epoch": 4.681564245810056,
"grad_norm": 2.739482165182619,
"learning_rate": 8.913767957387501e-07,
"logits/chosen": -3.8687398433685303,
"logits/rejected": -3.897033452987671,
"logps/chosen": -8.018453598022461,
"logps/rejected": -42.729331970214844,
"loss": 0.1188,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.448753833770752,
"rewards/margins": 2.448753833770752,
"rewards/rejected": 0.0,
"step": 838
},
{
"epoch": 4.687150837988827,
"grad_norm": 3.183335515081523,
"learning_rate": 8.909971660693147e-07,
"logits/chosen": -3.34853196144104,
"logits/rejected": -3.4277381896972656,
"logps/chosen": -8.967146873474121,
"logps/rejected": -23.089962005615234,
"loss": 0.1311,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.13580060005188,
"rewards/margins": 2.13580060005188,
"rewards/rejected": 0.0,
"step": 839
},
{
"epoch": 4.692737430167598,
"grad_norm": 7.559023766067752,
"learning_rate": 8.906169553084577e-07,
"logits/chosen": -3.4968342781066895,
"logits/rejected": -3.479490280151367,
"logps/chosen": -4.84238338470459,
"logps/rejected": -29.289892196655273,
"loss": 0.2087,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.613553524017334,
"rewards/margins": 1.613553524017334,
"rewards/rejected": 0.0,
"step": 840
},
{
"epoch": 4.692737430167598,
"eval_logits/chosen": -3.4382870197296143,
"eval_logits/rejected": -3.557262897491455,
"eval_logps/chosen": -20.889551162719727,
"eval_logps/rejected": -37.64310073852539,
"eval_loss": 0.4653335511684418,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9180494546890259,
"eval_rewards/margins": 0.9180494546890259,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7538,
"eval_samples_per_second": 9.465,
"eval_steps_per_second": 0.305,
"step": 840
},
{
"epoch": 4.698324022346369,
"grad_norm": 4.164925855011585,
"learning_rate": 8.9023616402124e-07,
"logits/chosen": -3.6509761810302734,
"logits/rejected": -3.4124083518981934,
"logps/chosen": -13.185310363769531,
"logps/rejected": -19.683773040771484,
"loss": 0.1361,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4277474880218506,
"rewards/margins": 2.4277474880218506,
"rewards/rejected": 0.0,
"step": 841
},
{
"epoch": 4.70391061452514,
"grad_norm": 4.095634852224189,
"learning_rate": 8.898547927735853e-07,
"logits/chosen": -3.4244139194488525,
"logits/rejected": -3.3410980701446533,
"logps/chosen": -20.241806030273438,
"logps/rejected": -25.040424346923828,
"loss": 0.1308,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5954225063323975,
"rewards/margins": 2.5954225063323975,
"rewards/rejected": 0.0,
"step": 842
},
{
"epoch": 4.709497206703911,
"grad_norm": 3.8874672480164203,
"learning_rate": 8.894728421322792e-07,
"logits/chosen": -3.538426160812378,
"logits/rejected": -3.6762707233428955,
"logps/chosen": -34.24104690551758,
"logps/rejected": -28.45409393310547,
"loss": 0.1536,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.510376453399658,
"rewards/margins": 2.510376453399658,
"rewards/rejected": 0.0,
"step": 843
},
{
"epoch": 4.715083798882682,
"grad_norm": 5.641882413941325,
"learning_rate": 8.890903126649685e-07,
"logits/chosen": -3.534832000732422,
"logits/rejected": -3.3883702754974365,
"logps/chosen": -6.195542335510254,
"logps/rejected": -21.10651397705078,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.074922800064087,
"rewards/margins": 2.074922800064087,
"rewards/rejected": 0.0,
"step": 844
},
{
"epoch": 4.720670391061453,
"grad_norm": 3.225454777473204,
"learning_rate": 8.887072049401598e-07,
"logits/chosen": -3.827566385269165,
"logits/rejected": -3.703700304031372,
"logps/chosen": -6.3797197341918945,
"logps/rejected": -25.23931312561035,
"loss": 0.1498,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.050833225250244,
"rewards/margins": 2.050833225250244,
"rewards/rejected": 0.0,
"step": 845
},
{
"epoch": 4.726256983240224,
"grad_norm": 4.180459311375749,
"learning_rate": 8.883235195272196e-07,
"logits/chosen": -3.547757148742676,
"logits/rejected": -3.5946600437164307,
"logps/chosen": -7.756105899810791,
"logps/rejected": -29.495033264160156,
"loss": 0.1188,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3571243286132812,
"rewards/margins": 2.3571243286132812,
"rewards/rejected": 0.0,
"step": 846
},
{
"epoch": 4.731843575418995,
"grad_norm": 5.9734829206994995,
"learning_rate": 8.879392569963728e-07,
"logits/chosen": -3.4734814167022705,
"logits/rejected": -3.33512020111084,
"logps/chosen": -12.2142333984375,
"logps/rejected": -28.297042846679688,
"loss": 0.1844,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.387676954269409,
"rewards/margins": 2.387676954269409,
"rewards/rejected": 0.0,
"step": 847
},
{
"epoch": 4.737430167597766,
"grad_norm": 4.2661422590982,
"learning_rate": 8.87554417918702e-07,
"logits/chosen": -3.9001755714416504,
"logits/rejected": -3.9348902702331543,
"logps/chosen": -6.842373371124268,
"logps/rejected": -26.677532196044922,
"loss": 0.1663,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8079452514648438,
"rewards/margins": 1.8079452514648438,
"rewards/rejected": 0.0,
"step": 848
},
{
"epoch": 4.743016759776537,
"grad_norm": 5.05967023339862,
"learning_rate": 8.871690028661464e-07,
"logits/chosen": -3.479654550552368,
"logits/rejected": -3.278282403945923,
"logps/chosen": -20.254270553588867,
"logps/rejected": -81.40043640136719,
"loss": 0.1558,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.504744052886963,
"rewards/margins": 2.504744052886963,
"rewards/rejected": 0.0,
"step": 849
},
{
"epoch": 4.748603351955307,
"grad_norm": 3.7319728586212753,
"learning_rate": 8.867830124115014e-07,
"logits/chosen": -3.176271915435791,
"logits/rejected": -3.3628337383270264,
"logps/chosen": -16.692127227783203,
"logps/rejected": -42.161102294921875,
"loss": 0.1304,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4846765995025635,
"rewards/margins": 2.4846765995025635,
"rewards/rejected": 0.0,
"step": 850
},
{
"epoch": 4.754189944134078,
"grad_norm": 4.783110484472577,
"learning_rate": 8.863964471284178e-07,
"logits/chosen": -3.7601890563964844,
"logits/rejected": -3.7013044357299805,
"logps/chosen": -5.958396911621094,
"logps/rejected": -22.45055389404297,
"loss": 0.1333,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.179135322570801,
"rewards/margins": 2.179135322570801,
"rewards/rejected": 0.0,
"step": 851
},
{
"epoch": 4.759776536312849,
"grad_norm": 5.7151029669725935,
"learning_rate": 8.860093075914004e-07,
"logits/chosen": -3.2503085136413574,
"logits/rejected": -3.472109794616699,
"logps/chosen": -3.8963963985443115,
"logps/rejected": -80.78500366210938,
"loss": 0.1591,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7483668327331543,
"rewards/margins": 1.7483668327331543,
"rewards/rejected": 0.0,
"step": 852
},
{
"epoch": 4.76536312849162,
"grad_norm": 9.094997919284872,
"learning_rate": 8.856215943758074e-07,
"logits/chosen": -3.5655927658081055,
"logits/rejected": -3.1865780353546143,
"logps/chosen": -11.715600967407227,
"logps/rejected": -69.3724594116211,
"loss": 0.1441,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4587438106536865,
"rewards/margins": 2.4587438106536865,
"rewards/rejected": 0.0,
"step": 853
},
{
"epoch": 4.770949720670391,
"grad_norm": 5.112652606314636,
"learning_rate": 8.852333080578495e-07,
"logits/chosen": -3.502084255218506,
"logits/rejected": -3.748291492462158,
"logps/chosen": -9.790876388549805,
"logps/rejected": -19.313125610351562,
"loss": 0.1628,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8427542448043823,
"rewards/margins": 1.8427542448043823,
"rewards/rejected": 0.0,
"step": 854
},
{
"epoch": 4.776536312849162,
"grad_norm": 4.848025576669861,
"learning_rate": 8.848444492145899e-07,
"logits/chosen": -3.5827841758728027,
"logits/rejected": -3.7325146198272705,
"logps/chosen": -13.776187896728516,
"logps/rejected": -30.147903442382812,
"loss": 0.1605,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8059005737304688,
"rewards/margins": 2.8059005737304688,
"rewards/rejected": 0.0,
"step": 855
},
{
"epoch": 4.782122905027933,
"grad_norm": 4.2627260769307025,
"learning_rate": 8.844550184239414e-07,
"logits/chosen": -3.709784746170044,
"logits/rejected": -3.637117624282837,
"logps/chosen": -27.101978302001953,
"logps/rejected": -38.568138122558594,
"loss": 0.1394,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3738842010498047,
"rewards/margins": 2.3738842010498047,
"rewards/rejected": 0.0,
"step": 856
},
{
"epoch": 4.787709497206704,
"grad_norm": 3.2765455080161354,
"learning_rate": 8.840650162646679e-07,
"logits/chosen": -3.7055535316467285,
"logits/rejected": -3.7811439037323,
"logps/chosen": -6.742583274841309,
"logps/rejected": -61.49567413330078,
"loss": 0.1335,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0676863193511963,
"rewards/margins": 2.0676863193511963,
"rewards/rejected": 0.0,
"step": 857
},
{
"epoch": 4.793296089385475,
"grad_norm": 3.440546922947156,
"learning_rate": 8.836744433163821e-07,
"logits/chosen": -3.4895920753479004,
"logits/rejected": -3.6672732830047607,
"logps/chosen": -32.132694244384766,
"logps/rejected": -29.041641235351562,
"loss": 0.1212,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5951497554779053,
"rewards/margins": 2.5951497554779053,
"rewards/rejected": 0.0,
"step": 858
},
{
"epoch": 4.798882681564246,
"grad_norm": 6.333532675924077,
"learning_rate": 8.832833001595448e-07,
"logits/chosen": -3.485452175140381,
"logits/rejected": -3.5626187324523926,
"logps/chosen": -6.243794918060303,
"logps/rejected": -54.92760467529297,
"loss": 0.1666,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4682459831237793,
"rewards/margins": 1.4682459831237793,
"rewards/rejected": 0.0,
"step": 859
},
{
"epoch": 4.804469273743017,
"grad_norm": 9.321206136113556,
"learning_rate": 8.828915873754643e-07,
"logits/chosen": -3.8284683227539062,
"logits/rejected": -3.896667718887329,
"logps/chosen": -7.002790451049805,
"logps/rejected": -38.846466064453125,
"loss": 0.1917,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6483442783355713,
"rewards/margins": 1.6483442783355713,
"rewards/rejected": 0.0,
"step": 860
},
{
"epoch": 4.804469273743017,
"eval_logits/chosen": -3.4337158203125,
"eval_logits/rejected": -3.5523319244384766,
"eval_logps/chosen": -20.837451934814453,
"eval_logps/rejected": -37.41223907470703,
"eval_loss": 0.45792660117149353,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.9232591390609741,
"eval_rewards/margins": 0.9232591390609741,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7351,
"eval_samples_per_second": 9.47,
"eval_steps_per_second": 0.305,
"step": 860
},
{
"epoch": 4.810055865921788,
"grad_norm": 10.558126361014091,
"learning_rate": 8.82499305546296e-07,
"logits/chosen": -3.3911778926849365,
"logits/rejected": -3.5714023113250732,
"logps/chosen": -29.563304901123047,
"logps/rejected": -45.68463897705078,
"loss": 0.1503,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3393919467926025,
"rewards/margins": 2.3393919467926025,
"rewards/rejected": 0.0,
"step": 861
},
{
"epoch": 4.815642458100559,
"grad_norm": 3.6575243733620337,
"learning_rate": 8.821064552550399e-07,
"logits/chosen": -3.264042615890503,
"logits/rejected": -3.4649696350097656,
"logps/chosen": -7.36049222946167,
"logps/rejected": -49.74733352661133,
"loss": 0.1207,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.71518874168396,
"rewards/margins": 1.71518874168396,
"rewards/rejected": 0.0,
"step": 862
},
{
"epoch": 4.82122905027933,
"grad_norm": 5.312389583932255,
"learning_rate": 8.817130370855421e-07,
"logits/chosen": -3.665332794189453,
"logits/rejected": -3.640456199645996,
"logps/chosen": -7.155714988708496,
"logps/rejected": -18.648021697998047,
"loss": 0.1943,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.0625817775726318,
"rewards/margins": 1.0625817775726318,
"rewards/rejected": 0.0,
"step": 863
},
{
"epoch": 4.826815642458101,
"grad_norm": 4.271741315348231,
"learning_rate": 8.813190516224917e-07,
"logits/chosen": -3.786005973815918,
"logits/rejected": -3.8781280517578125,
"logps/chosen": -18.279701232910156,
"logps/rejected": -27.392444610595703,
"loss": 0.1428,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.885749340057373,
"rewards/margins": 2.885749340057373,
"rewards/rejected": 0.0,
"step": 864
},
{
"epoch": 4.832402234636872,
"grad_norm": 8.007426726489133,
"learning_rate": 8.809244994514216e-07,
"logits/chosen": -3.5738139152526855,
"logits/rejected": -3.54390811920166,
"logps/chosen": -9.149717330932617,
"logps/rejected": -60.3470458984375,
"loss": 0.1444,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.350733757019043,
"rewards/margins": 2.350733757019043,
"rewards/rejected": 0.0,
"step": 865
},
{
"epoch": 4.837988826815643,
"grad_norm": 3.501883519918708,
"learning_rate": 8.80529381158706e-07,
"logits/chosen": -3.549853563308716,
"logits/rejected": -3.669538736343384,
"logps/chosen": -5.122585773468018,
"logps/rejected": -53.249053955078125,
"loss": 0.1332,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3994872570037842,
"rewards/margins": 1.3994872570037842,
"rewards/rejected": 0.0,
"step": 866
},
{
"epoch": 4.843575418994414,
"grad_norm": 4.794695180692528,
"learning_rate": 8.801336973315617e-07,
"logits/chosen": -3.5816879272460938,
"logits/rejected": -3.6130690574645996,
"logps/chosen": -13.329113960266113,
"logps/rejected": -46.01040267944336,
"loss": 0.1446,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4738619327545166,
"rewards/margins": 2.4738619327545166,
"rewards/rejected": 0.0,
"step": 867
},
{
"epoch": 4.849162011173185,
"grad_norm": 3.4681825650643416,
"learning_rate": 8.79737448558045e-07,
"logits/chosen": -3.8831658363342285,
"logits/rejected": -3.7107584476470947,
"logps/chosen": -7.576117515563965,
"logps/rejected": -37.6939582824707,
"loss": 0.1578,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.096287250518799,
"rewards/margins": 2.096287250518799,
"rewards/rejected": 0.0,
"step": 868
},
{
"epoch": 4.854748603351956,
"grad_norm": 3.522115677977157,
"learning_rate": 8.793406354270522e-07,
"logits/chosen": -3.5656511783599854,
"logits/rejected": -3.636209487915039,
"logps/chosen": -9.377242088317871,
"logps/rejected": -19.11086654663086,
"loss": 0.1471,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1785616874694824,
"rewards/margins": 2.1785616874694824,
"rewards/rejected": 0.0,
"step": 869
},
{
"epoch": 4.860335195530726,
"grad_norm": 3.636940330890014,
"learning_rate": 8.789432585283182e-07,
"logits/chosen": -3.9074556827545166,
"logits/rejected": -3.8623340129852295,
"logps/chosen": -9.630167961120605,
"logps/rejected": -20.304290771484375,
"loss": 0.1159,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2927615642547607,
"rewards/margins": 2.2927615642547607,
"rewards/rejected": 0.0,
"step": 870
},
{
"epoch": 4.865921787709497,
"grad_norm": 8.786746912693621,
"learning_rate": 8.785453184524161e-07,
"logits/chosen": -3.4771811962127686,
"logits/rejected": -3.7050230503082275,
"logps/chosen": -5.9504594802856445,
"logps/rejected": -50.320777893066406,
"loss": 0.1872,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.0571459531784058,
"rewards/margins": 1.0571459531784058,
"rewards/rejected": 0.0,
"step": 871
},
{
"epoch": 4.871508379888268,
"grad_norm": 4.016442501619869,
"learning_rate": 8.781468157907554e-07,
"logits/chosen": -3.5626220703125,
"logits/rejected": -3.583447217941284,
"logps/chosen": -9.067211151123047,
"logps/rejected": -26.456647872924805,
"loss": 0.1499,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9865537881851196,
"rewards/margins": 1.9865537881851196,
"rewards/rejected": 0.0,
"step": 872
},
{
"epoch": 4.877094972067039,
"grad_norm": 4.53543096354546,
"learning_rate": 8.777477511355822e-07,
"logits/chosen": -3.3158986568450928,
"logits/rejected": -3.3363637924194336,
"logps/chosen": -6.712955951690674,
"logps/rejected": -32.11681365966797,
"loss": 0.148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.324026346206665,
"rewards/margins": 2.324026346206665,
"rewards/rejected": 0.0,
"step": 873
},
{
"epoch": 4.88268156424581,
"grad_norm": 3.3915553921064108,
"learning_rate": 8.773481250799776e-07,
"logits/chosen": -3.3401966094970703,
"logits/rejected": -3.218843936920166,
"logps/chosen": -5.150437831878662,
"logps/rejected": -48.85731506347656,
"loss": 0.1429,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6175756454467773,
"rewards/margins": 1.6175756454467773,
"rewards/rejected": 0.0,
"step": 874
},
{
"epoch": 4.888268156424581,
"grad_norm": 5.366800397444954,
"learning_rate": 8.76947938217857e-07,
"logits/chosen": -3.6672873497009277,
"logits/rejected": -3.4623639583587646,
"logps/chosen": -4.0337114334106445,
"logps/rejected": -21.869462966918945,
"loss": 0.1619,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2740421295166016,
"rewards/margins": 1.2740421295166016,
"rewards/rejected": 0.0,
"step": 875
},
{
"epoch": 4.893854748603352,
"grad_norm": 6.212216346123267,
"learning_rate": 8.765471911439697e-07,
"logits/chosen": -3.3895440101623535,
"logits/rejected": -3.4063079357147217,
"logps/chosen": -33.54104995727539,
"logps/rejected": -39.06547546386719,
"loss": 0.1685,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6400959491729736,
"rewards/margins": 3.6400959491729736,
"rewards/rejected": 0.0,
"step": 876
},
{
"epoch": 4.899441340782123,
"grad_norm": 8.414250727496093,
"learning_rate": 8.76145884453897e-07,
"logits/chosen": -3.5738282203674316,
"logits/rejected": -3.5223355293273926,
"logps/chosen": -12.802833557128906,
"logps/rejected": -20.496601104736328,
"loss": 0.2038,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.4248250722885132,
"rewards/margins": 1.4248250722885132,
"rewards/rejected": 0.0,
"step": 877
},
{
"epoch": 4.905027932960894,
"grad_norm": 6.795900358584056,
"learning_rate": 8.757440187440519e-07,
"logits/chosen": -3.712117910385132,
"logits/rejected": -3.529858350753784,
"logps/chosen": -8.133525848388672,
"logps/rejected": -24.60988998413086,
"loss": 0.1829,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.5177634954452515,
"rewards/margins": 1.5177634954452515,
"rewards/rejected": 0.0,
"step": 878
},
{
"epoch": 4.910614525139665,
"grad_norm": 4.620793169748076,
"learning_rate": 8.753415946116786e-07,
"logits/chosen": -3.4427971839904785,
"logits/rejected": -3.714064836502075,
"logps/chosen": -3.4842402935028076,
"logps/rejected": -55.89486312866211,
"loss": 0.1461,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6254570484161377,
"rewards/margins": 1.6254570484161377,
"rewards/rejected": 0.0,
"step": 879
},
{
"epoch": 4.916201117318436,
"grad_norm": 3.6278667379031915,
"learning_rate": 8.74938612654851e-07,
"logits/chosen": -3.660454511642456,
"logits/rejected": -3.769160509109497,
"logps/chosen": -7.781861305236816,
"logps/rejected": -52.43263244628906,
"loss": 0.1446,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0503783226013184,
"rewards/margins": 2.0503783226013184,
"rewards/rejected": 0.0,
"step": 880
},
{
"epoch": 4.916201117318436,
"eval_logits/chosen": -3.4445042610168457,
"eval_logits/rejected": -3.562300205230713,
"eval_logps/chosen": -21.081167221069336,
"eval_logps/rejected": -37.07769012451172,
"eval_loss": 0.45790335536003113,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.8988874554634094,
"eval_rewards/margins": 0.8988874554634094,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6974,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 880
},
{
"epoch": 4.921787709497207,
"grad_norm": 3.532124702646352,
"learning_rate": 8.745350734724721e-07,
"logits/chosen": -3.6082777976989746,
"logits/rejected": -3.5758960247039795,
"logps/chosen": -5.217126846313477,
"logps/rejected": -43.59714126586914,
"loss": 0.1326,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.950150728225708,
"rewards/margins": 1.950150728225708,
"rewards/rejected": 0.0,
"step": 881
},
{
"epoch": 4.927374301675978,
"grad_norm": 5.148709902345891,
"learning_rate": 8.741309776642732e-07,
"logits/chosen": -3.537998914718628,
"logits/rejected": -3.392167568206787,
"logps/chosen": -4.527282238006592,
"logps/rejected": -38.122215270996094,
"loss": 0.1304,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.907090663909912,
"rewards/margins": 1.907090663909912,
"rewards/rejected": 0.0,
"step": 882
},
{
"epoch": 4.932960893854749,
"grad_norm": 4.587751704336057,
"learning_rate": 8.73726325830812e-07,
"logits/chosen": -3.6477560997009277,
"logits/rejected": -3.5672593116760254,
"logps/chosen": -15.186811447143555,
"logps/rejected": -20.84063148498535,
"loss": 0.1278,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6742024421691895,
"rewards/margins": 2.6742024421691895,
"rewards/rejected": 0.0,
"step": 883
},
{
"epoch": 4.93854748603352,
"grad_norm": 3.195306283341391,
"learning_rate": 8.733211185734739e-07,
"logits/chosen": -3.3010315895080566,
"logits/rejected": -3.403163194656372,
"logps/chosen": -19.622634887695312,
"logps/rejected": -43.96053695678711,
"loss": 0.1401,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4762537479400635,
"rewards/margins": 2.4762537479400635,
"rewards/rejected": 0.0,
"step": 884
},
{
"epoch": 4.94413407821229,
"grad_norm": 4.803074260751064,
"learning_rate": 8.729153564944687e-07,
"logits/chosen": -3.221707582473755,
"logits/rejected": -3.2944374084472656,
"logps/chosen": -14.754175186157227,
"logps/rejected": -35.13188934326172,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9772963523864746,
"rewards/margins": 1.9772963523864746,
"rewards/rejected": 0.0,
"step": 885
},
{
"epoch": 4.949720670391061,
"grad_norm": 4.0082091117970915,
"learning_rate": 8.725090401968312e-07,
"logits/chosen": -3.5048813819885254,
"logits/rejected": -3.5432684421539307,
"logps/chosen": -11.013408660888672,
"logps/rejected": -22.639747619628906,
"loss": 0.1307,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.150505542755127,
"rewards/margins": 2.150505542755127,
"rewards/rejected": 0.0,
"step": 886
},
{
"epoch": 4.955307262569832,
"grad_norm": 3.0866883461040384,
"learning_rate": 8.721021702844196e-07,
"logits/chosen": -3.496249198913574,
"logits/rejected": -3.6273727416992188,
"logps/chosen": -6.44735050201416,
"logps/rejected": -57.3204345703125,
"loss": 0.1356,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3810064792633057,
"rewards/margins": 2.3810064792633057,
"rewards/rejected": 0.0,
"step": 887
},
{
"epoch": 4.960893854748603,
"grad_norm": 3.206021847477474,
"learning_rate": 8.716947473619153e-07,
"logits/chosen": -3.3977081775665283,
"logits/rejected": -3.6455678939819336,
"logps/chosen": -11.624605178833008,
"logps/rejected": -52.049560546875,
"loss": 0.1345,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2550697326660156,
"rewards/margins": 2.2550697326660156,
"rewards/rejected": 0.0,
"step": 888
},
{
"epoch": 4.966480446927374,
"grad_norm": 3.1866427838508047,
"learning_rate": 8.712867720348211e-07,
"logits/chosen": -3.441948890686035,
"logits/rejected": -3.4032344818115234,
"logps/chosen": -13.188152313232422,
"logps/rejected": -31.50457763671875,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3232598304748535,
"rewards/margins": 2.3232598304748535,
"rewards/rejected": 0.0,
"step": 889
},
{
"epoch": 4.972067039106145,
"grad_norm": 2.5771609283540378,
"learning_rate": 8.708782449094611e-07,
"logits/chosen": -3.64654803276062,
"logits/rejected": -3.7692649364471436,
"logps/chosen": -14.121740341186523,
"logps/rejected": -22.496761322021484,
"loss": 0.122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.407217502593994,
"rewards/margins": 2.407217502593994,
"rewards/rejected": 0.0,
"step": 890
},
{
"epoch": 4.977653631284916,
"grad_norm": 5.04488215485463,
"learning_rate": 8.704691665929795e-07,
"logits/chosen": -3.8497824668884277,
"logits/rejected": -3.7043254375457764,
"logps/chosen": -11.662101745605469,
"logps/rejected": -25.833295822143555,
"loss": 0.1808,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9730836153030396,
"rewards/margins": 1.9730836153030396,
"rewards/rejected": 0.0,
"step": 891
},
{
"epoch": 4.983240223463687,
"grad_norm": 3.658861469188171,
"learning_rate": 8.700595376933393e-07,
"logits/chosen": -3.8118834495544434,
"logits/rejected": -3.5197701454162598,
"logps/chosen": -13.414787292480469,
"logps/rejected": -56.63683319091797,
"loss": 0.1361,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0512075424194336,
"rewards/margins": 2.0512075424194336,
"rewards/rejected": 0.0,
"step": 892
},
{
"epoch": 4.988826815642458,
"grad_norm": 2.8903311625392654,
"learning_rate": 8.696493588193221e-07,
"logits/chosen": -3.328495979309082,
"logits/rejected": -3.4698615074157715,
"logps/chosen": -3.700801134109497,
"logps/rejected": -44.19495391845703,
"loss": 0.1304,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.874070405960083,
"rewards/margins": 1.874070405960083,
"rewards/rejected": 0.0,
"step": 893
},
{
"epoch": 4.994413407821229,
"grad_norm": 3.2934790346791663,
"learning_rate": 8.692386305805269e-07,
"logits/chosen": -3.1227805614471436,
"logits/rejected": -3.09611439704895,
"logps/chosen": -35.340484619140625,
"logps/rejected": -68.12325286865234,
"loss": 0.1361,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.268045425415039,
"rewards/margins": 2.268045425415039,
"rewards/rejected": 0.0,
"step": 894
},
{
"epoch": 5.0,
"grad_norm": 2.0998319711198743,
"learning_rate": 8.688273535873686e-07,
"logits/chosen": -3.8451521396636963,
"logits/rejected": -3.4695792198181152,
"logps/chosen": -9.727790832519531,
"logps/rejected": -28.658916473388672,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6223816871643066,
"rewards/margins": 2.6223816871643066,
"rewards/rejected": 0.0,
"step": 895
},
{
"epoch": 5.005586592178771,
"grad_norm": 6.127492385172778,
"learning_rate": 8.684155284510783e-07,
"logits/chosen": -3.7467594146728516,
"logits/rejected": -3.6886558532714844,
"logps/chosen": -12.076903343200684,
"logps/rejected": -22.749643325805664,
"loss": 0.144,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.371918201446533,
"rewards/margins": 2.371918201446533,
"rewards/rejected": 0.0,
"step": 896
},
{
"epoch": 5.011173184357542,
"grad_norm": 2.0759935000059735,
"learning_rate": 8.680031557837018e-07,
"logits/chosen": -3.2360999584198,
"logits/rejected": -3.2278568744659424,
"logps/chosen": -11.073640823364258,
"logps/rejected": -63.07777786254883,
"loss": 0.1214,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6215603351593018,
"rewards/margins": 2.6215603351593018,
"rewards/rejected": 0.0,
"step": 897
},
{
"epoch": 5.016759776536313,
"grad_norm": 7.738618865663303,
"learning_rate": 8.675902361980978e-07,
"logits/chosen": -3.3227431774139404,
"logits/rejected": -3.1236305236816406,
"logps/chosen": -4.390902042388916,
"logps/rejected": -19.178699493408203,
"loss": 0.1502,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.052748441696167,
"rewards/margins": 2.052748441696167,
"rewards/rejected": 0.0,
"step": 898
},
{
"epoch": 5.022346368715084,
"grad_norm": 2.6501499890783338,
"learning_rate": 8.671767703079386e-07,
"logits/chosen": -3.8698344230651855,
"logits/rejected": -3.894357204437256,
"logps/chosen": -21.294483184814453,
"logps/rejected": -44.14567184448242,
"loss": 0.1191,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.803455352783203,
"rewards/margins": 3.803455352783203,
"rewards/rejected": 0.0,
"step": 899
},
{
"epoch": 5.027932960893855,
"grad_norm": 2.386846191424922,
"learning_rate": 8.667627587277081e-07,
"logits/chosen": -3.7314257621765137,
"logits/rejected": -3.7088756561279297,
"logps/chosen": -3.20855450630188,
"logps/rejected": -48.41551971435547,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.193094491958618,
"rewards/margins": 2.193094491958618,
"rewards/rejected": 0.0,
"step": 900
},
{
"epoch": 5.027932960893855,
"eval_logits/chosen": -3.435518980026245,
"eval_logits/rejected": -3.5585150718688965,
"eval_logps/chosen": -21.40536117553711,
"eval_logps/rejected": -38.2454719543457,
"eval_loss": 0.4943534731864929,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.8664682507514954,
"eval_rewards/margins": 0.8664682507514954,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7129,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 900
},
{
"epoch": 5.033519553072626,
"grad_norm": 2.5779826951393097,
"learning_rate": 8.663482020727013e-07,
"logits/chosen": -3.718381643295288,
"logits/rejected": -3.7165474891662598,
"logps/chosen": -12.928625106811523,
"logps/rejected": -37.67301940917969,
"loss": 0.1181,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.815427780151367,
"rewards/margins": 2.815427780151367,
"rewards/rejected": 0.0,
"step": 901
},
{
"epoch": 5.039106145251397,
"grad_norm": 2.3829605902021043,
"learning_rate": 8.659331009590231e-07,
"logits/chosen": -3.8124420642852783,
"logits/rejected": -3.894714593887329,
"logps/chosen": -6.802081108093262,
"logps/rejected": -59.38578796386719,
"loss": 0.1232,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1188838481903076,
"rewards/margins": 2.1188838481903076,
"rewards/rejected": 0.0,
"step": 902
},
{
"epoch": 5.044692737430168,
"grad_norm": 3.295346521656783,
"learning_rate": 8.655174560035878e-07,
"logits/chosen": -3.3991622924804688,
"logits/rejected": -3.6705405712127686,
"logps/chosen": -19.94412612915039,
"logps/rejected": -27.22773551940918,
"loss": 0.1441,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.003748893737793,
"rewards/margins": 3.003748893737793,
"rewards/rejected": 0.0,
"step": 903
},
{
"epoch": 5.050279329608939,
"grad_norm": 2.4254931524029373,
"learning_rate": 8.651012678241179e-07,
"logits/chosen": -3.432058811187744,
"logits/rejected": -3.484321117401123,
"logps/chosen": -7.184386730194092,
"logps/rejected": -19.5164794921875,
"loss": 0.1367,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8795058727264404,
"rewards/margins": 1.8795058727264404,
"rewards/rejected": 0.0,
"step": 904
},
{
"epoch": 5.055865921787709,
"grad_norm": 5.762149383593757,
"learning_rate": 8.646845370391429e-07,
"logits/chosen": -3.544438362121582,
"logits/rejected": -3.552574872970581,
"logps/chosen": -3.3116674423217773,
"logps/rejected": -44.91862487792969,
"loss": 0.1599,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1848440170288086,
"rewards/margins": 2.1848440170288086,
"rewards/rejected": 0.0,
"step": 905
},
{
"epoch": 5.06145251396648,
"grad_norm": 2.238674137807379,
"learning_rate": 8.642672642679991e-07,
"logits/chosen": -2.8455119132995605,
"logits/rejected": -3.1013801097869873,
"logps/chosen": -37.575096130371094,
"logps/rejected": -67.72724914550781,
"loss": 0.1286,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.719628095626831,
"rewards/margins": 2.719628095626831,
"rewards/rejected": 0.0,
"step": 906
},
{
"epoch": 5.067039106145251,
"grad_norm": 3.1359154613767766,
"learning_rate": 8.638494501308282e-07,
"logits/chosen": -3.614899158477783,
"logits/rejected": -3.5207931995391846,
"logps/chosen": -2.8579413890838623,
"logps/rejected": -29.47831916809082,
"loss": 0.1145,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9729458093643188,
"rewards/margins": 1.9729458093643188,
"rewards/rejected": 0.0,
"step": 907
},
{
"epoch": 5.072625698324022,
"grad_norm": 2.9435008272268743,
"learning_rate": 8.634310952485764e-07,
"logits/chosen": -3.7244222164154053,
"logits/rejected": -3.8382468223571777,
"logps/chosen": -5.40566349029541,
"logps/rejected": -39.49150085449219,
"loss": 0.1289,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.449627161026001,
"rewards/margins": 2.449627161026001,
"rewards/rejected": 0.0,
"step": 908
},
{
"epoch": 5.078212290502793,
"grad_norm": 3.07483432464254,
"learning_rate": 8.630122002429934e-07,
"logits/chosen": -3.888763666152954,
"logits/rejected": -3.8629233837127686,
"logps/chosen": -4.398175239562988,
"logps/rejected": -40.81145095825195,
"loss": 0.1191,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1777243614196777,
"rewards/margins": 2.1777243614196777,
"rewards/rejected": 0.0,
"step": 909
},
{
"epoch": 5.083798882681564,
"grad_norm": 2.0242722462680076,
"learning_rate": 8.625927657366321e-07,
"logits/chosen": -3.602264165878296,
"logits/rejected": -3.629054307937622,
"logps/chosen": -7.341497421264648,
"logps/rejected": -29.58710289001465,
"loss": 0.1171,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.576740026473999,
"rewards/margins": 2.576740026473999,
"rewards/rejected": 0.0,
"step": 910
},
{
"epoch": 5.089385474860335,
"grad_norm": 4.349726842397969,
"learning_rate": 8.621727923528465e-07,
"logits/chosen": -3.780391216278076,
"logits/rejected": -3.8685684204101562,
"logps/chosen": -13.424625396728516,
"logps/rejected": -41.788330078125,
"loss": 0.1632,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.475114345550537,
"rewards/margins": 2.475114345550537,
"rewards/rejected": 0.0,
"step": 911
},
{
"epoch": 5.094972067039106,
"grad_norm": 2.1555804248891564,
"learning_rate": 8.61752280715792e-07,
"logits/chosen": -3.635103225708008,
"logits/rejected": -3.7968673706054688,
"logps/chosen": -4.576326370239258,
"logps/rejected": -49.913028717041016,
"loss": 0.0981,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.112658739089966,
"rewards/margins": 2.112658739089966,
"rewards/rejected": 0.0,
"step": 912
},
{
"epoch": 5.100558659217877,
"grad_norm": 3.6088824487821602,
"learning_rate": 8.613312314504239e-07,
"logits/chosen": -3.730335235595703,
"logits/rejected": -3.8086090087890625,
"logps/chosen": -3.0457799434661865,
"logps/rejected": -28.66973114013672,
"loss": 0.13,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5662554502487183,
"rewards/margins": 1.5662554502487183,
"rewards/rejected": 0.0,
"step": 913
},
{
"epoch": 5.106145251396648,
"grad_norm": 5.127206199475611,
"learning_rate": 8.609096451824962e-07,
"logits/chosen": -3.3814682960510254,
"logits/rejected": -3.2327253818511963,
"logps/chosen": -3.7657828330993652,
"logps/rejected": -39.985130310058594,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6491355895996094,
"rewards/margins": 1.6491355895996094,
"rewards/rejected": 0.0,
"step": 914
},
{
"epoch": 5.111731843575419,
"grad_norm": 2.9408725419231034,
"learning_rate": 8.604875225385612e-07,
"logits/chosen": -3.7518861293792725,
"logits/rejected": -3.776332139968872,
"logps/chosen": -10.089970588684082,
"logps/rejected": -22.130268096923828,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.302517890930176,
"rewards/margins": 2.302517890930176,
"rewards/rejected": 0.0,
"step": 915
},
{
"epoch": 5.11731843575419,
"grad_norm": 2.2874280726807887,
"learning_rate": 8.600648641459687e-07,
"logits/chosen": -3.748547077178955,
"logits/rejected": -3.8262057304382324,
"logps/chosen": -3.8874173164367676,
"logps/rejected": -28.511350631713867,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9695693254470825,
"rewards/margins": 1.9695693254470825,
"rewards/rejected": 0.0,
"step": 916
},
{
"epoch": 5.122905027932961,
"grad_norm": 3.3819493698529732,
"learning_rate": 8.596416706328638e-07,
"logits/chosen": -3.5149059295654297,
"logits/rejected": -3.633885622024536,
"logps/chosen": -5.141606330871582,
"logps/rejected": -32.707542419433594,
"loss": 0.1455,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6533509492874146,
"rewards/margins": 1.6533509492874146,
"rewards/rejected": 0.0,
"step": 917
},
{
"epoch": 5.128491620111732,
"grad_norm": 2.907818534904172,
"learning_rate": 8.592179426281878e-07,
"logits/chosen": -3.6833579540252686,
"logits/rejected": -3.493891477584839,
"logps/chosen": -6.513550758361816,
"logps/rejected": -42.60105895996094,
"loss": 0.1299,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.287080764770508,
"rewards/margins": 2.287080764770508,
"rewards/rejected": 0.0,
"step": 918
},
{
"epoch": 5.134078212290503,
"grad_norm": 2.2665989967634226,
"learning_rate": 8.587936807616756e-07,
"logits/chosen": -3.459026336669922,
"logits/rejected": -3.541001558303833,
"logps/chosen": -20.24863052368164,
"logps/rejected": -31.118236541748047,
"loss": 0.1317,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.420464038848877,
"rewards/margins": 2.420464038848877,
"rewards/rejected": 0.0,
"step": 919
},
{
"epoch": 5.139664804469274,
"grad_norm": 2.123177189518529,
"learning_rate": 8.583688856638563e-07,
"logits/chosen": -3.3693435192108154,
"logits/rejected": -3.469839572906494,
"logps/chosen": -8.63968276977539,
"logps/rejected": -47.11341857910156,
"loss": 0.1107,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0286364555358887,
"rewards/margins": 2.0286364555358887,
"rewards/rejected": 0.0,
"step": 920
},
{
"epoch": 5.139664804469274,
"eval_logits/chosen": -3.425973415374756,
"eval_logits/rejected": -3.551579236984253,
"eval_logps/chosen": -21.862384796142578,
"eval_logps/rejected": -39.4764289855957,
"eval_loss": 0.5151383876800537,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.8207657933235168,
"eval_rewards/margins": 0.8207657933235168,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7056,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 920
},
{
"epoch": 5.145251396648045,
"grad_norm": 2.6781754907340325,
"learning_rate": 8.579435579660509e-07,
"logits/chosen": -3.6018340587615967,
"logits/rejected": -3.6825060844421387,
"logps/chosen": -4.666955471038818,
"logps/rejected": -52.6751708984375,
"loss": 0.1357,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0365028381347656,
"rewards/margins": 2.0365028381347656,
"rewards/rejected": 0.0,
"step": 921
},
{
"epoch": 5.150837988826815,
"grad_norm": 2.302785965792041,
"learning_rate": 8.57517698300372e-07,
"logits/chosen": -3.6873726844787598,
"logits/rejected": -3.595874309539795,
"logps/chosen": -14.56936264038086,
"logps/rejected": -22.20042610168457,
"loss": 0.1046,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0235393047332764,
"rewards/margins": 3.0235393047332764,
"rewards/rejected": 0.0,
"step": 922
},
{
"epoch": 5.156424581005586,
"grad_norm": 3.426254757697152,
"learning_rate": 8.570913072997233e-07,
"logits/chosen": -3.8000307083129883,
"logits/rejected": -3.5224392414093018,
"logps/chosen": -6.43786096572876,
"logps/rejected": -31.06023406982422,
"loss": 0.1258,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4343795776367188,
"rewards/margins": 2.4343795776367188,
"rewards/rejected": 0.0,
"step": 923
},
{
"epoch": 5.162011173184357,
"grad_norm": 4.169084053169075,
"learning_rate": 8.566643855977971e-07,
"logits/chosen": -3.583242416381836,
"logits/rejected": -3.4318792819976807,
"logps/chosen": -9.185176849365234,
"logps/rejected": -17.2728271484375,
"loss": 0.1301,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.572232246398926,
"rewards/margins": 2.572232246398926,
"rewards/rejected": 0.0,
"step": 924
},
{
"epoch": 5.167597765363128,
"grad_norm": 2.1962278849430503,
"learning_rate": 8.562369338290756e-07,
"logits/chosen": -3.6337318420410156,
"logits/rejected": -3.725064754486084,
"logps/chosen": -5.347441673278809,
"logps/rejected": -33.45294952392578,
"loss": 0.1434,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.316202163696289,
"rewards/margins": 2.316202163696289,
"rewards/rejected": 0.0,
"step": 925
},
{
"epoch": 5.173184357541899,
"grad_norm": 4.294902947201775,
"learning_rate": 8.558089526288281e-07,
"logits/chosen": -3.2959165573120117,
"logits/rejected": -3.4752299785614014,
"logps/chosen": -13.800309181213379,
"logps/rejected": -29.787723541259766,
"loss": 0.1588,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.654491901397705,
"rewards/margins": 2.654491901397705,
"rewards/rejected": 0.0,
"step": 926
},
{
"epoch": 5.17877094972067,
"grad_norm": 3.2968934448645784,
"learning_rate": 8.55380442633111e-07,
"logits/chosen": -3.612819194793701,
"logits/rejected": -3.552001953125,
"logps/chosen": -26.93370819091797,
"logps/rejected": -24.36080551147461,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.425076961517334,
"rewards/margins": 3.425076961517334,
"rewards/rejected": 0.0,
"step": 927
},
{
"epoch": 5.184357541899441,
"grad_norm": 2.134847232261565,
"learning_rate": 8.549514044787663e-07,
"logits/chosen": -3.552821397781372,
"logits/rejected": -3.3308751583099365,
"logps/chosen": -6.113100051879883,
"logps/rejected": -52.883689880371094,
"loss": 0.1512,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0449371337890625,
"rewards/margins": 2.0449371337890625,
"rewards/rejected": 0.0,
"step": 928
},
{
"epoch": 5.189944134078212,
"grad_norm": 1.9644718559300236,
"learning_rate": 8.54521838803421e-07,
"logits/chosen": -3.4582157135009766,
"logits/rejected": -3.522745370864868,
"logps/chosen": -10.016764640808105,
"logps/rejected": -36.827789306640625,
"loss": 0.1209,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.741762161254883,
"rewards/margins": 2.741762161254883,
"rewards/rejected": 0.0,
"step": 929
},
{
"epoch": 5.195530726256983,
"grad_norm": 9.682737619449624,
"learning_rate": 8.540917462454864e-07,
"logits/chosen": -3.9171395301818848,
"logits/rejected": -3.842108964920044,
"logps/chosen": -5.504759311676025,
"logps/rejected": -33.843650817871094,
"loss": 0.1553,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1367480754852295,
"rewards/margins": 2.1367480754852295,
"rewards/rejected": 0.0,
"step": 930
},
{
"epoch": 5.201117318435754,
"grad_norm": 2.0908246070117453,
"learning_rate": 8.536611274441566e-07,
"logits/chosen": -3.6773064136505127,
"logits/rejected": -3.590070962905884,
"logps/chosen": -38.34971618652344,
"logps/rejected": -41.589012145996094,
"loss": 0.1193,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5228641033172607,
"rewards/margins": 3.5228641033172607,
"rewards/rejected": 0.0,
"step": 931
},
{
"epoch": 5.206703910614525,
"grad_norm": 2.873060339890528,
"learning_rate": 8.532299830394079e-07,
"logits/chosen": -3.4728243350982666,
"logits/rejected": -3.8036398887634277,
"logps/chosen": -5.717551231384277,
"logps/rejected": -60.32939910888672,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.168911933898926,
"rewards/margins": 2.168911933898926,
"rewards/rejected": 0.0,
"step": 932
},
{
"epoch": 5.212290502793296,
"grad_norm": 4.196625161083795,
"learning_rate": 8.527983136719976e-07,
"logits/chosen": -3.3160006999969482,
"logits/rejected": -3.38995623588562,
"logps/chosen": -22.3360595703125,
"logps/rejected": -26.047075271606445,
"loss": 0.1157,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3260152339935303,
"rewards/margins": 2.3260152339935303,
"rewards/rejected": 0.0,
"step": 933
},
{
"epoch": 5.217877094972067,
"grad_norm": 2.684892475589337,
"learning_rate": 8.523661199834631e-07,
"logits/chosen": -3.8219635486602783,
"logits/rejected": -3.9562270641326904,
"logps/chosen": -5.9781270027160645,
"logps/rejected": -39.06108093261719,
"loss": 0.109,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3151016235351562,
"rewards/margins": 2.3151016235351562,
"rewards/rejected": 0.0,
"step": 934
},
{
"epoch": 5.223463687150838,
"grad_norm": 3.361433215437716,
"learning_rate": 8.519334026161214e-07,
"logits/chosen": -3.4284021854400635,
"logits/rejected": -3.430989980697632,
"logps/chosen": -1.641413688659668,
"logps/rejected": -48.60231018066406,
"loss": 0.1605,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6644656658172607,
"rewards/margins": 1.6644656658172607,
"rewards/rejected": 0.0,
"step": 935
},
{
"epoch": 5.229050279329609,
"grad_norm": 2.448103045769202,
"learning_rate": 8.515001622130675e-07,
"logits/chosen": -3.7019660472869873,
"logits/rejected": -3.4524085521698,
"logps/chosen": -4.677493095397949,
"logps/rejected": -43.681678771972656,
"loss": 0.122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2800111770629883,
"rewards/margins": 2.2800111770629883,
"rewards/rejected": 0.0,
"step": 936
},
{
"epoch": 5.23463687150838,
"grad_norm": 3.3938076413380376,
"learning_rate": 8.510663994181738e-07,
"logits/chosen": -3.1296334266662598,
"logits/rejected": -3.234459400177002,
"logps/chosen": -25.453676223754883,
"logps/rejected": -38.42242431640625,
"loss": 0.1216,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0531082153320312,
"rewards/margins": 3.0531082153320312,
"rewards/rejected": 0.0,
"step": 937
},
{
"epoch": 5.240223463687151,
"grad_norm": 3.0708984028748247,
"learning_rate": 8.506321148760891e-07,
"logits/chosen": -3.6749613285064697,
"logits/rejected": -3.7708475589752197,
"logps/chosen": -3.0763132572174072,
"logps/rejected": -33.902042388916016,
"loss": 0.1323,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.474084734916687,
"rewards/margins": 1.474084734916687,
"rewards/rejected": 0.0,
"step": 938
},
{
"epoch": 5.245810055865922,
"grad_norm": 3.9396728681411206,
"learning_rate": 8.501973092322375e-07,
"logits/chosen": -3.4395763874053955,
"logits/rejected": -3.355232000350952,
"logps/chosen": -5.489841938018799,
"logps/rejected": -23.398866653442383,
"loss": 0.1297,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4068026542663574,
"rewards/margins": 2.4068026542663574,
"rewards/rejected": 0.0,
"step": 939
},
{
"epoch": 5.251396648044693,
"grad_norm": 3.2175744448504893,
"learning_rate": 8.497619831328178e-07,
"logits/chosen": -3.805591106414795,
"logits/rejected": -3.9668242931365967,
"logps/chosen": -10.686816215515137,
"logps/rejected": -28.034881591796875,
"loss": 0.1205,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7061309814453125,
"rewards/margins": 2.7061309814453125,
"rewards/rejected": 0.0,
"step": 940
},
{
"epoch": 5.251396648044693,
"eval_logits/chosen": -3.4313652515411377,
"eval_logits/rejected": -3.555899143218994,
"eval_logps/chosen": -21.915748596191406,
"eval_logps/rejected": -40.4402961730957,
"eval_loss": 0.5229962468147278,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.8154290914535522,
"eval_rewards/margins": 0.8154290914535522,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7408,
"eval_samples_per_second": 9.468,
"eval_steps_per_second": 0.305,
"step": 940
},
{
"epoch": 5.256983240223463,
"grad_norm": 2.1227916790899037,
"learning_rate": 8.493261372248017e-07,
"logits/chosen": -3.7516915798187256,
"logits/rejected": -3.5086164474487305,
"logps/chosen": -8.845268249511719,
"logps/rejected": -27.884584426879883,
"loss": 0.1345,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5056686401367188,
"rewards/margins": 2.5056686401367188,
"rewards/rejected": 0.0,
"step": 941
},
{
"epoch": 5.262569832402234,
"grad_norm": 2.86054465219148,
"learning_rate": 8.488897721559343e-07,
"logits/chosen": -3.6603825092315674,
"logits/rejected": -3.7184808254241943,
"logps/chosen": -5.2650980949401855,
"logps/rejected": -37.26551055908203,
"loss": 0.121,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9481420516967773,
"rewards/margins": 1.9481420516967773,
"rewards/rejected": 0.0,
"step": 942
},
{
"epoch": 5.268156424581005,
"grad_norm": 3.4441727530514172,
"learning_rate": 8.484528885747315e-07,
"logits/chosen": -3.4722037315368652,
"logits/rejected": -3.4777796268463135,
"logps/chosen": -3.823040008544922,
"logps/rejected": -30.969818115234375,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9734690189361572,
"rewards/margins": 1.9734690189361572,
"rewards/rejected": 0.0,
"step": 943
},
{
"epoch": 5.273743016759776,
"grad_norm": 5.2095703916482154,
"learning_rate": 8.480154871304801e-07,
"logits/chosen": -3.396315097808838,
"logits/rejected": -3.4404098987579346,
"logps/chosen": -11.265424728393555,
"logps/rejected": -14.863361358642578,
"loss": 0.1526,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.128246545791626,
"rewards/margins": 2.128246545791626,
"rewards/rejected": 0.0,
"step": 944
},
{
"epoch": 5.279329608938547,
"grad_norm": 3.3338926579251162,
"learning_rate": 8.475775684732369e-07,
"logits/chosen": -3.5682334899902344,
"logits/rejected": -3.294985771179199,
"logps/chosen": -1.143467903137207,
"logps/rejected": -56.27909469604492,
"loss": 0.1497,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2283358573913574,
"rewards/margins": 1.2283358573913574,
"rewards/rejected": 0.0,
"step": 945
},
{
"epoch": 5.284916201117318,
"grad_norm": 2.9230370638327585,
"learning_rate": 8.471391332538266e-07,
"logits/chosen": -3.6190085411071777,
"logits/rejected": -3.870126962661743,
"logps/chosen": -11.431694030761719,
"logps/rejected": -18.61604881286621,
"loss": 0.1224,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6102752685546875,
"rewards/margins": 2.6102752685546875,
"rewards/rejected": 0.0,
"step": 946
},
{
"epoch": 5.290502793296089,
"grad_norm": 3.2666822273478417,
"learning_rate": 8.467001821238421e-07,
"logits/chosen": -3.7518441677093506,
"logits/rejected": -3.6237411499023438,
"logps/chosen": -1.5083858966827393,
"logps/rejected": -82.01118469238281,
"loss": 0.1361,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.424487590789795,
"rewards/margins": 1.424487590789795,
"rewards/rejected": 0.0,
"step": 947
},
{
"epoch": 5.29608938547486,
"grad_norm": 2.402537784639173,
"learning_rate": 8.462607157356434e-07,
"logits/chosen": -3.1628594398498535,
"logits/rejected": -3.3524177074432373,
"logps/chosen": -4.144198894500732,
"logps/rejected": -45.98481750488281,
"loss": 0.1267,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9893189668655396,
"rewards/margins": 1.9893189668655396,
"rewards/rejected": 0.0,
"step": 948
},
{
"epoch": 5.301675977653631,
"grad_norm": 2.1287793565080166,
"learning_rate": 8.458207347423553e-07,
"logits/chosen": -3.2780921459198,
"logits/rejected": -3.2125978469848633,
"logps/chosen": -5.203969478607178,
"logps/rejected": -66.4188232421875,
"loss": 0.1184,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.149090528488159,
"rewards/margins": 2.149090528488159,
"rewards/rejected": 0.0,
"step": 949
},
{
"epoch": 5.307262569832402,
"grad_norm": 3.8601327518074102,
"learning_rate": 8.45380239797868e-07,
"logits/chosen": -3.5509846210479736,
"logits/rejected": -3.5342063903808594,
"logps/chosen": -10.181859970092773,
"logps/rejected": -30.597375869750977,
"loss": 0.1141,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5514869689941406,
"rewards/margins": 2.5514869689941406,
"rewards/rejected": 0.0,
"step": 950
},
{
"epoch": 5.312849162011173,
"grad_norm": 2.3130618426281937,
"learning_rate": 8.449392315568355e-07,
"logits/chosen": -3.836066484451294,
"logits/rejected": -3.620694637298584,
"logps/chosen": -8.613641738891602,
"logps/rejected": -36.907615661621094,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1679139137268066,
"rewards/margins": 2.1679139137268066,
"rewards/rejected": 0.0,
"step": 951
},
{
"epoch": 5.318435754189944,
"grad_norm": 2.0727404018073043,
"learning_rate": 8.444977106746746e-07,
"logits/chosen": -3.6299524307250977,
"logits/rejected": -3.7042176723480225,
"logps/chosen": -25.02996826171875,
"logps/rejected": -30.16958236694336,
"loss": 0.1168,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.044982433319092,
"rewards/margins": 3.044982433319092,
"rewards/rejected": 0.0,
"step": 952
},
{
"epoch": 5.324022346368715,
"grad_norm": 3.5282975987744707,
"learning_rate": 8.44055677807564e-07,
"logits/chosen": -3.3784050941467285,
"logits/rejected": -3.7101690769195557,
"logps/chosen": -6.744629859924316,
"logps/rejected": -48.71521759033203,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.211911201477051,
"rewards/margins": 2.211911201477051,
"rewards/rejected": 0.0,
"step": 953
},
{
"epoch": 5.329608938547486,
"grad_norm": 1.88849840259381,
"learning_rate": 8.436131336124431e-07,
"logits/chosen": -3.7207722663879395,
"logits/rejected": -3.75814151763916,
"logps/chosen": -2.950428009033203,
"logps/rejected": -43.67927551269531,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.291203022003174,
"rewards/margins": 2.291203022003174,
"rewards/rejected": 0.0,
"step": 954
},
{
"epoch": 5.335195530726257,
"grad_norm": 2.862908630145159,
"learning_rate": 8.431700787470116e-07,
"logits/chosen": -3.9300715923309326,
"logits/rejected": -3.9752509593963623,
"logps/chosen": -37.81331253051758,
"logps/rejected": -30.72120475769043,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.501981735229492,
"rewards/margins": 3.501981735229492,
"rewards/rejected": 0.0,
"step": 955
},
{
"epoch": 5.340782122905028,
"grad_norm": 3.7350135696278914,
"learning_rate": 8.427265138697279e-07,
"logits/chosen": -3.7004594802856445,
"logits/rejected": -3.6338562965393066,
"logps/chosen": -8.439620971679688,
"logps/rejected": -29.650474548339844,
"loss": 0.129,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3722708225250244,
"rewards/margins": 2.3722708225250244,
"rewards/rejected": 0.0,
"step": 956
},
{
"epoch": 5.346368715083799,
"grad_norm": 2.3180477658725267,
"learning_rate": 8.422824396398081e-07,
"logits/chosen": -3.6553523540496826,
"logits/rejected": -3.615277051925659,
"logps/chosen": -5.601487159729004,
"logps/rejected": -32.682823181152344,
"loss": 0.0878,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2934601306915283,
"rewards/margins": 2.2934601306915283,
"rewards/rejected": 0.0,
"step": 957
},
{
"epoch": 5.351955307262569,
"grad_norm": 2.624376384375354,
"learning_rate": 8.41837856717226e-07,
"logits/chosen": -3.8188042640686035,
"logits/rejected": -3.8369505405426025,
"logps/chosen": -14.519266128540039,
"logps/rejected": -36.771400451660156,
"loss": 0.1166,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2506394386291504,
"rewards/margins": 2.2506394386291504,
"rewards/rejected": 0.0,
"step": 958
},
{
"epoch": 5.35754189944134,
"grad_norm": 2.9579438881756746,
"learning_rate": 8.413927657627109e-07,
"logits/chosen": -3.6174941062927246,
"logits/rejected": -3.5712153911590576,
"logps/chosen": -3.6067914962768555,
"logps/rejected": -53.74940490722656,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3034090995788574,
"rewards/margins": 2.3034090995788574,
"rewards/rejected": 0.0,
"step": 959
},
{
"epoch": 5.363128491620111,
"grad_norm": 3.1962700511590163,
"learning_rate": 8.40947167437747e-07,
"logits/chosen": -3.660860061645508,
"logits/rejected": -3.572692632675171,
"logps/chosen": -9.876673698425293,
"logps/rejected": -41.77517318725586,
"loss": 0.1228,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6486129760742188,
"rewards/margins": 2.6486129760742188,
"rewards/rejected": 0.0,
"step": 960
},
{
"epoch": 5.363128491620111,
"eval_logits/chosen": -3.4246811866760254,
"eval_logits/rejected": -3.549966335296631,
"eval_logps/chosen": -21.75457763671875,
"eval_logps/rejected": -40.558349609375,
"eval_loss": 0.5337956547737122,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.8315466046333313,
"eval_rewards/margins": 0.8315466046333313,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6947,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 960
},
{
"epoch": 5.368715083798882,
"grad_norm": 4.816839452806974,
"learning_rate": 8.40501062404573e-07,
"logits/chosen": -3.4338021278381348,
"logits/rejected": -3.2634763717651367,
"logps/chosen": -15.61409854888916,
"logps/rejected": -34.017295837402344,
"loss": 0.1534,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.017667770385742,
"rewards/margins": 3.017667770385742,
"rewards/rejected": 0.0,
"step": 961
},
{
"epoch": 5.374301675977653,
"grad_norm": 2.482510011792642,
"learning_rate": 8.400544513261803e-07,
"logits/chosen": -3.8516783714294434,
"logits/rejected": -3.761363983154297,
"logps/chosen": -3.9727489948272705,
"logps/rejected": -32.69411849975586,
"loss": 0.108,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.200329303741455,
"rewards/margins": 2.200329303741455,
"rewards/rejected": 0.0,
"step": 962
},
{
"epoch": 5.379888268156424,
"grad_norm": 1.748209005735185,
"learning_rate": 8.396073348663125e-07,
"logits/chosen": -3.537409543991089,
"logits/rejected": -3.4711556434631348,
"logps/chosen": -6.442181587219238,
"logps/rejected": -26.657386779785156,
"loss": 0.1021,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5487351417541504,
"rewards/margins": 2.5487351417541504,
"rewards/rejected": 0.0,
"step": 963
},
{
"epoch": 5.385474860335195,
"grad_norm": 3.6730868297201518,
"learning_rate": 8.391597136894644e-07,
"logits/chosen": -3.5208287239074707,
"logits/rejected": -3.2740936279296875,
"logps/chosen": -11.94963550567627,
"logps/rejected": -44.65607833862305,
"loss": 0.1398,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4625320434570312,
"rewards/margins": 2.4625320434570312,
"rewards/rejected": 0.0,
"step": 964
},
{
"epoch": 5.391061452513966,
"grad_norm": 2.632398028082482,
"learning_rate": 8.387115884608807e-07,
"logits/chosen": -3.6703941822052,
"logits/rejected": -3.7788689136505127,
"logps/chosen": -1.2421746253967285,
"logps/rejected": -90.82178497314453,
"loss": 0.1521,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3596502542495728,
"rewards/margins": 1.3596502542495728,
"rewards/rejected": 0.0,
"step": 965
},
{
"epoch": 5.396648044692737,
"grad_norm": 3.7936712908039394,
"learning_rate": 8.382629598465554e-07,
"logits/chosen": -3.911432981491089,
"logits/rejected": -3.613752841949463,
"logps/chosen": -4.657186508178711,
"logps/rejected": -24.86302947998047,
"loss": 0.1416,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.330866813659668,
"rewards/margins": 2.330866813659668,
"rewards/rejected": 0.0,
"step": 966
},
{
"epoch": 5.402234636871508,
"grad_norm": 1.7917417081482783,
"learning_rate": 8.378138285132303e-07,
"logits/chosen": -3.687436819076538,
"logits/rejected": -3.62178373336792,
"logps/chosen": -1.7335727214813232,
"logps/rejected": -30.068944931030273,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4629037380218506,
"rewards/margins": 1.4629037380218506,
"rewards/rejected": 0.0,
"step": 967
},
{
"epoch": 5.407821229050279,
"grad_norm": 3.3786380351906637,
"learning_rate": 8.373641951283947e-07,
"logits/chosen": -3.4602854251861572,
"logits/rejected": -3.360727071762085,
"logps/chosen": -2.335566997528076,
"logps/rejected": -38.54754638671875,
"loss": 0.1297,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7878601551055908,
"rewards/margins": 1.7878601551055908,
"rewards/rejected": 0.0,
"step": 968
},
{
"epoch": 5.41340782122905,
"grad_norm": 3.062911125741669,
"learning_rate": 8.36914060360284e-07,
"logits/chosen": -3.9719436168670654,
"logits/rejected": -3.806168556213379,
"logps/chosen": -7.921912670135498,
"logps/rejected": -35.76070022583008,
"loss": 0.1531,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.394432544708252,
"rewards/margins": 2.394432544708252,
"rewards/rejected": 0.0,
"step": 969
},
{
"epoch": 5.418994413407821,
"grad_norm": 4.118220377183595,
"learning_rate": 8.364634248778783e-07,
"logits/chosen": -3.5636565685272217,
"logits/rejected": -3.6488091945648193,
"logps/chosen": -6.545294761657715,
"logps/rejected": -37.022560119628906,
"loss": 0.1299,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.072500705718994,
"rewards/margins": 2.072500705718994,
"rewards/rejected": 0.0,
"step": 970
},
{
"epoch": 5.424581005586592,
"grad_norm": 2.423093707243962,
"learning_rate": 8.360122893509025e-07,
"logits/chosen": -3.716005802154541,
"logits/rejected": -3.667078733444214,
"logps/chosen": -6.051375865936279,
"logps/rejected": -51.206546783447266,
"loss": 0.1303,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1058497428894043,
"rewards/margins": 2.1058497428894043,
"rewards/rejected": 0.0,
"step": 971
},
{
"epoch": 5.430167597765363,
"grad_norm": 2.469023269894222,
"learning_rate": 8.355606544498241e-07,
"logits/chosen": -3.6141245365142822,
"logits/rejected": -3.656458854675293,
"logps/chosen": -6.50161075592041,
"logps/rejected": -28.264385223388672,
"loss": 0.1372,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2572598457336426,
"rewards/margins": 2.2572598457336426,
"rewards/rejected": 0.0,
"step": 972
},
{
"epoch": 5.435754189944134,
"grad_norm": 2.870405773440076,
"learning_rate": 8.351085208458533e-07,
"logits/chosen": -3.440138816833496,
"logits/rejected": -3.4495396614074707,
"logps/chosen": -5.213408470153809,
"logps/rejected": -53.99237060546875,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.179966449737549,
"rewards/margins": 2.179966449737549,
"rewards/rejected": 0.0,
"step": 973
},
{
"epoch": 5.441340782122905,
"grad_norm": 4.779981011272239,
"learning_rate": 8.346558892109407e-07,
"logits/chosen": -3.6187150478363037,
"logits/rejected": -3.481257200241089,
"logps/chosen": -7.725311279296875,
"logps/rejected": -26.0552978515625,
"loss": 0.1366,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9844796657562256,
"rewards/margins": 1.9844796657562256,
"rewards/rejected": 0.0,
"step": 974
},
{
"epoch": 5.446927374301676,
"grad_norm": 3.662552416936445,
"learning_rate": 8.342027602177779e-07,
"logits/chosen": -3.3446497917175293,
"logits/rejected": -3.2567250728607178,
"logps/chosen": -18.352100372314453,
"logps/rejected": -29.235855102539062,
"loss": 0.1411,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.959120988845825,
"rewards/margins": 2.959120988845825,
"rewards/rejected": 0.0,
"step": 975
},
{
"epoch": 5.452513966480447,
"grad_norm": 2.457014682852703,
"learning_rate": 8.337491345397952e-07,
"logits/chosen": -3.4628405570983887,
"logits/rejected": -3.2928173542022705,
"logps/chosen": -11.77698040008545,
"logps/rejected": -48.920654296875,
"loss": 0.1138,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4243571758270264,
"rewards/margins": 2.4243571758270264,
"rewards/rejected": 0.0,
"step": 976
},
{
"epoch": 5.4581005586592175,
"grad_norm": 2.4380306326679055,
"learning_rate": 8.332950128511611e-07,
"logits/chosen": -3.9781696796417236,
"logits/rejected": -3.8392252922058105,
"logps/chosen": -3.0315113067626953,
"logps/rejected": -49.192378997802734,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9341998100280762,
"rewards/margins": 1.9341998100280762,
"rewards/rejected": 0.0,
"step": 977
},
{
"epoch": 5.4636871508379885,
"grad_norm": 4.0072399880982585,
"learning_rate": 8.328403958267813e-07,
"logits/chosen": -3.6522932052612305,
"logits/rejected": -3.7077629566192627,
"logps/chosen": -10.203707695007324,
"logps/rejected": -29.24556541442871,
"loss": 0.1183,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6389682292938232,
"rewards/margins": 2.6389682292938232,
"rewards/rejected": 0.0,
"step": 978
},
{
"epoch": 5.4692737430167595,
"grad_norm": 4.9940949419550416,
"learning_rate": 8.323852841422976e-07,
"logits/chosen": -3.3061370849609375,
"logits/rejected": -3.5059430599212646,
"logps/chosen": -1.0360535383224487,
"logps/rejected": -56.605987548828125,
"loss": 0.1476,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1801810264587402,
"rewards/margins": 1.1801810264587402,
"rewards/rejected": 0.0,
"step": 979
},
{
"epoch": 5.4748603351955305,
"grad_norm": 4.580311287796682,
"learning_rate": 8.319296784740872e-07,
"logits/chosen": -3.534660816192627,
"logits/rejected": -3.7138571739196777,
"logps/chosen": -11.394460678100586,
"logps/rejected": -25.93413543701172,
"loss": 0.1248,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7908012866973877,
"rewards/margins": 2.7908012866973877,
"rewards/rejected": 0.0,
"step": 980
},
{
"epoch": 5.4748603351955305,
"eval_logits/chosen": -3.4305739402770996,
"eval_logits/rejected": -3.554555892944336,
"eval_logps/chosen": -22.076717376708984,
"eval_logps/rejected": -40.22883987426758,
"eval_loss": 0.5252670049667358,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.7993324398994446,
"eval_rewards/margins": 0.7993324398994446,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7101,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 980
},
{
"epoch": 5.4804469273743015,
"grad_norm": 2.8556803617902577,
"learning_rate": 8.314735794992611e-07,
"logits/chosen": -3.6257777214050293,
"logits/rejected": -3.6955454349517822,
"logps/chosen": -28.178733825683594,
"logps/rejected": -27.137306213378906,
"loss": 0.1562,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.659531593322754,
"rewards/margins": 2.659531593322754,
"rewards/rejected": 0.0,
"step": 981
},
{
"epoch": 5.4860335195530725,
"grad_norm": 2.670914142698,
"learning_rate": 8.310169878956636e-07,
"logits/chosen": -3.495283365249634,
"logits/rejected": -3.2490310668945312,
"logps/chosen": -3.7130908966064453,
"logps/rejected": -30.080886840820312,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.114500045776367,
"rewards/margins": 2.114500045776367,
"rewards/rejected": 0.0,
"step": 982
},
{
"epoch": 5.4916201117318435,
"grad_norm": 10.020042671020262,
"learning_rate": 8.305599043418711e-07,
"logits/chosen": -3.626030206680298,
"logits/rejected": -3.4159045219421387,
"logps/chosen": -11.557868003845215,
"logps/rejected": -47.024723052978516,
"loss": 0.1472,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.241159439086914,
"rewards/margins": 2.241159439086914,
"rewards/rejected": 0.0,
"step": 983
},
{
"epoch": 5.4972067039106145,
"grad_norm": 6.178298743082472,
"learning_rate": 8.301023295171915e-07,
"logits/chosen": -3.4850172996520996,
"logits/rejected": -3.5257372856140137,
"logps/chosen": -10.838176727294922,
"logps/rejected": -26.925825119018555,
"loss": 0.132,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4204211235046387,
"rewards/margins": 2.4204211235046387,
"rewards/rejected": 0.0,
"step": 984
},
{
"epoch": 5.5027932960893855,
"grad_norm": 3.4672961029394913,
"learning_rate": 8.296442641016621e-07,
"logits/chosen": -3.311988592147827,
"logits/rejected": -3.3689091205596924,
"logps/chosen": -4.270512580871582,
"logps/rejected": -45.78513717651367,
"loss": 0.1356,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1297318935394287,
"rewards/margins": 2.1297318935394287,
"rewards/rejected": 0.0,
"step": 985
},
{
"epoch": 5.5083798882681565,
"grad_norm": 3.9497562228771117,
"learning_rate": 8.291857087760498e-07,
"logits/chosen": -3.668921947479248,
"logits/rejected": -3.7324578762054443,
"logps/chosen": -11.818612098693848,
"logps/rejected": -46.838600158691406,
"loss": 0.1385,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.626201629638672,
"rewards/margins": 2.626201629638672,
"rewards/rejected": 0.0,
"step": 986
},
{
"epoch": 5.5139664804469275,
"grad_norm": 4.330539574087278,
"learning_rate": 8.287266642218495e-07,
"logits/chosen": -3.6318037509918213,
"logits/rejected": -3.7037880420684814,
"logps/chosen": -7.9843573570251465,
"logps/rejected": -31.925270080566406,
"loss": 0.1409,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.511212110519409,
"rewards/margins": 2.511212110519409,
"rewards/rejected": 0.0,
"step": 987
},
{
"epoch": 5.5195530726256985,
"grad_norm": 3.2623845153838342,
"learning_rate": 8.282671311212832e-07,
"logits/chosen": -3.3751072883605957,
"logits/rejected": -3.4393019676208496,
"logps/chosen": -16.140539169311523,
"logps/rejected": -38.205322265625,
"loss": 0.1286,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.369124174118042,
"rewards/margins": 2.369124174118042,
"rewards/rejected": 0.0,
"step": 988
},
{
"epoch": 5.5251396648044695,
"grad_norm": 3.8885666395174705,
"learning_rate": 8.278071101572988e-07,
"logits/chosen": -3.4989941120147705,
"logits/rejected": -3.589448928833008,
"logps/chosen": -6.971655368804932,
"logps/rejected": -29.12643051147461,
"loss": 0.1435,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9628701210021973,
"rewards/margins": 1.9628701210021973,
"rewards/rejected": 0.0,
"step": 989
},
{
"epoch": 5.5307262569832405,
"grad_norm": 4.020234784130751,
"learning_rate": 8.273466020135694e-07,
"logits/chosen": -3.7029707431793213,
"logits/rejected": -3.753220558166504,
"logps/chosen": -8.310090065002441,
"logps/rejected": -27.959949493408203,
"loss": 0.1257,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.644866466522217,
"rewards/margins": 2.644866466522217,
"rewards/rejected": 0.0,
"step": 990
},
{
"epoch": 5.5363128491620115,
"grad_norm": 2.8017690945691913,
"learning_rate": 8.268856073744923e-07,
"logits/chosen": -3.3072383403778076,
"logits/rejected": -3.4897496700286865,
"logps/chosen": -13.173815727233887,
"logps/rejected": -89.54933166503906,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.632387161254883,
"rewards/margins": 2.632387161254883,
"rewards/rejected": 0.0,
"step": 991
},
{
"epoch": 5.5418994413407825,
"grad_norm": 5.324536947997927,
"learning_rate": 8.264241269251875e-07,
"logits/chosen": -3.8031108379364014,
"logits/rejected": -3.923901319503784,
"logps/chosen": -30.693422317504883,
"logps/rejected": -28.303131103515625,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6970062255859375,
"rewards/margins": 3.6970062255859375,
"rewards/rejected": 0.0,
"step": 992
},
{
"epoch": 5.547486033519553,
"grad_norm": 13.275659618085774,
"learning_rate": 8.259621613514969e-07,
"logits/chosen": -3.4624850749969482,
"logits/rejected": -3.402142286300659,
"logps/chosen": -9.430938720703125,
"logps/rejected": -51.62495422363281,
"loss": 0.1678,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3033833503723145,
"rewards/margins": 2.3033833503723145,
"rewards/rejected": 0.0,
"step": 993
},
{
"epoch": 5.553072625698324,
"grad_norm": 5.129162530894774,
"learning_rate": 8.254997113399842e-07,
"logits/chosen": -4.001574516296387,
"logits/rejected": -3.6959242820739746,
"logps/chosen": -9.59112548828125,
"logps/rejected": -36.154266357421875,
"loss": 0.1283,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5434792041778564,
"rewards/margins": 2.5434792041778564,
"rewards/rejected": 0.0,
"step": 994
},
{
"epoch": 5.558659217877095,
"grad_norm": 3.12863945151769,
"learning_rate": 8.250367775779319e-07,
"logits/chosen": -3.3619141578674316,
"logits/rejected": -3.4087536334991455,
"logps/chosen": -4.648746013641357,
"logps/rejected": -88.69186401367188,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.778944969177246,
"rewards/margins": 1.778944969177246,
"rewards/rejected": 0.0,
"step": 995
},
{
"epoch": 5.564245810055866,
"grad_norm": 4.191695662470572,
"learning_rate": 8.245733607533423e-07,
"logits/chosen": -3.499629259109497,
"logits/rejected": -3.460113525390625,
"logps/chosen": -14.256868362426758,
"logps/rejected": -28.99484634399414,
"loss": 0.1157,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6720776557922363,
"rewards/margins": 2.6720776557922363,
"rewards/rejected": 0.0,
"step": 996
},
{
"epoch": 5.569832402234637,
"grad_norm": 2.8226776540352794,
"learning_rate": 8.241094615549352e-07,
"logits/chosen": -3.471566915512085,
"logits/rejected": -3.5952682495117188,
"logps/chosen": -5.305325031280518,
"logps/rejected": -55.97410202026367,
"loss": 0.1006,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.113853931427002,
"rewards/margins": 2.113853931427002,
"rewards/rejected": 0.0,
"step": 997
},
{
"epoch": 5.575418994413408,
"grad_norm": 3.013186677958537,
"learning_rate": 8.236450806721475e-07,
"logits/chosen": -3.428640127182007,
"logits/rejected": -3.337449550628662,
"logps/chosen": -3.7617545127868652,
"logps/rejected": -34.300498962402344,
"loss": 0.1294,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9175052642822266,
"rewards/margins": 1.9175052642822266,
"rewards/rejected": 0.0,
"step": 998
},
{
"epoch": 5.581005586592179,
"grad_norm": 3.7475550308034378,
"learning_rate": 8.23180218795132e-07,
"logits/chosen": -3.6567182540893555,
"logits/rejected": -3.4699931144714355,
"logps/chosen": -8.725936889648438,
"logps/rejected": -26.261375427246094,
"loss": 0.1452,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1150784492492676,
"rewards/margins": 2.1150784492492676,
"rewards/rejected": 0.0,
"step": 999
},
{
"epoch": 5.58659217877095,
"grad_norm": 2.4820108158390095,
"learning_rate": 8.22714876614756e-07,
"logits/chosen": -3.543668508529663,
"logits/rejected": -3.7309658527374268,
"logps/chosen": -19.350299835205078,
"logps/rejected": -23.483177185058594,
"loss": 0.1221,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6947455406188965,
"rewards/margins": 2.6947455406188965,
"rewards/rejected": 0.0,
"step": 1000
},
{
"epoch": 5.58659217877095,
"eval_logits/chosen": -3.4348416328430176,
"eval_logits/rejected": -3.5552430152893066,
"eval_logps/chosen": -21.70493507385254,
"eval_logps/rejected": -39.91596221923828,
"eval_loss": 0.5152106285095215,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.8365108370780945,
"eval_rewards/margins": 0.8365108370780945,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7193,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 1000
},
{
"epoch": 5.592178770949721,
"grad_norm": 2.432680359777377,
"learning_rate": 8.222490548226011e-07,
"logits/chosen": -3.6886191368103027,
"logits/rejected": -3.5714361667633057,
"logps/chosen": -5.722451210021973,
"logps/rejected": -28.130802154541016,
"loss": 0.13,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3259849548339844,
"rewards/margins": 2.3259849548339844,
"rewards/rejected": 0.0,
"step": 1001
},
{
"epoch": 5.597765363128492,
"grad_norm": 2.2978400226359725,
"learning_rate": 8.217827541109612e-07,
"logits/chosen": -3.263062000274658,
"logits/rejected": -3.352325677871704,
"logps/chosen": -3.045107126235962,
"logps/rejected": -60.54975128173828,
"loss": 0.12,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.559848666191101,
"rewards/margins": 1.559848666191101,
"rewards/rejected": 0.0,
"step": 1002
},
{
"epoch": 5.603351955307263,
"grad_norm": 1.9418465558548306,
"learning_rate": 8.213159751728424e-07,
"logits/chosen": -3.3491270542144775,
"logits/rejected": -3.388939619064331,
"logps/chosen": -16.784282684326172,
"logps/rejected": -29.645170211791992,
"loss": 0.1022,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.189891815185547,
"rewards/margins": 3.189891815185547,
"rewards/rejected": 0.0,
"step": 1003
},
{
"epoch": 5.608938547486034,
"grad_norm": 4.783557383585265,
"learning_rate": 8.208487187019612e-07,
"logits/chosen": -3.413283348083496,
"logits/rejected": -3.386101245880127,
"logps/chosen": -1.149272084236145,
"logps/rejected": -29.626628875732422,
"loss": 0.1461,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.537308692932129,
"rewards/margins": 1.537308692932129,
"rewards/rejected": 0.0,
"step": 1004
},
{
"epoch": 5.614525139664805,
"grad_norm": 3.2186521172932294,
"learning_rate": 8.203809853927439e-07,
"logits/chosen": -3.5104498863220215,
"logits/rejected": -3.4505863189697266,
"logps/chosen": -15.869890213012695,
"logps/rejected": -57.78500747680664,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.779179334640503,
"rewards/margins": 2.779179334640503,
"rewards/rejected": 0.0,
"step": 1005
},
{
"epoch": 5.620111731843576,
"grad_norm": 6.515149987517597,
"learning_rate": 8.199127759403257e-07,
"logits/chosen": -3.5511887073516846,
"logits/rejected": -3.4592039585113525,
"logps/chosen": -7.230931282043457,
"logps/rejected": -35.0115966796875,
"loss": 0.1398,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.145414352416992,
"rewards/margins": 2.145414352416992,
"rewards/rejected": 0.0,
"step": 1006
},
{
"epoch": 5.625698324022347,
"grad_norm": 5.188676319462628,
"learning_rate": 8.194440910405492e-07,
"logits/chosen": -3.5458908081054688,
"logits/rejected": -3.509129285812378,
"logps/chosen": -5.889590263366699,
"logps/rejected": -36.376590728759766,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3784074783325195,
"rewards/margins": 2.3784074783325195,
"rewards/rejected": 0.0,
"step": 1007
},
{
"epoch": 5.631284916201118,
"grad_norm": 2.9165117914502177,
"learning_rate": 8.189749313899632e-07,
"logits/chosen": -3.3884472846984863,
"logits/rejected": -3.6048972606658936,
"logps/chosen": -3.1345629692077637,
"logps/rejected": -105.1725845336914,
"loss": 0.1442,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.012176990509033,
"rewards/margins": 2.012176990509033,
"rewards/rejected": 0.0,
"step": 1008
},
{
"epoch": 5.636871508379889,
"grad_norm": 4.732486447912168,
"learning_rate": 8.185052976858231e-07,
"logits/chosen": -3.626363754272461,
"logits/rejected": -3.478931188583374,
"logps/chosen": -5.381372451782227,
"logps/rejected": -42.6785888671875,
"loss": 0.1473,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6120800971984863,
"rewards/margins": 2.6120800971984863,
"rewards/rejected": 0.0,
"step": 1009
},
{
"epoch": 5.64245810055866,
"grad_norm": 2.897188146080436,
"learning_rate": 8.180351906260878e-07,
"logits/chosen": -3.9950497150421143,
"logits/rejected": -3.830092668533325,
"logps/chosen": -9.810124397277832,
"logps/rejected": -27.96949577331543,
"loss": 0.1298,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.570422649383545,
"rewards/margins": 2.570422649383545,
"rewards/rejected": 0.0,
"step": 1010
},
{
"epoch": 5.648044692737431,
"grad_norm": 2.2812328890419256,
"learning_rate": 8.175646109094204e-07,
"logits/chosen": -3.7176613807678223,
"logits/rejected": -3.571798801422119,
"logps/chosen": -15.496166229248047,
"logps/rejected": -27.049665451049805,
"loss": 0.1145,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.841841459274292,
"rewards/margins": 2.841841459274292,
"rewards/rejected": 0.0,
"step": 1011
},
{
"epoch": 5.653631284916202,
"grad_norm": 3.264848508506336,
"learning_rate": 8.17093559235186e-07,
"logits/chosen": -3.532919406890869,
"logits/rejected": -3.543309211730957,
"logps/chosen": -2.1750073432922363,
"logps/rejected": -64.8692855834961,
"loss": 0.1453,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9546127319335938,
"rewards/margins": 1.9546127319335938,
"rewards/rejected": 0.0,
"step": 1012
},
{
"epoch": 5.659217877094972,
"grad_norm": 5.876623350033097,
"learning_rate": 8.166220363034513e-07,
"logits/chosen": -3.6311070919036865,
"logits/rejected": -3.5080888271331787,
"logps/chosen": -10.898408889770508,
"logps/rejected": -21.208297729492188,
"loss": 0.1423,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5477521419525146,
"rewards/margins": 2.5477521419525146,
"rewards/rejected": 0.0,
"step": 1013
},
{
"epoch": 5.664804469273743,
"grad_norm": 2.544333298111979,
"learning_rate": 8.161500428149832e-07,
"logits/chosen": -3.4825284481048584,
"logits/rejected": -3.7296884059906006,
"logps/chosen": -4.902135848999023,
"logps/rejected": -39.969940185546875,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2294304370880127,
"rewards/margins": 2.2294304370880127,
"rewards/rejected": 0.0,
"step": 1014
},
{
"epoch": 5.670391061452514,
"grad_norm": 4.401093520263053,
"learning_rate": 8.156775794712483e-07,
"logits/chosen": -3.7538087368011475,
"logits/rejected": -3.8753857612609863,
"logps/chosen": -7.648348808288574,
"logps/rejected": -53.634117126464844,
"loss": 0.1273,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.640012264251709,
"rewards/margins": 2.640012264251709,
"rewards/rejected": 0.0,
"step": 1015
},
{
"epoch": 5.675977653631285,
"grad_norm": 8.629381180647874,
"learning_rate": 8.152046469744114e-07,
"logits/chosen": -3.2993342876434326,
"logits/rejected": -3.439915895462036,
"logps/chosen": -3.192535400390625,
"logps/rejected": -77.20162963867188,
"loss": 0.1339,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2494940757751465,
"rewards/margins": 2.2494940757751465,
"rewards/rejected": 0.0,
"step": 1016
},
{
"epoch": 5.681564245810056,
"grad_norm": 2.9299287157690532,
"learning_rate": 8.14731246027334e-07,
"logits/chosen": -3.7943758964538574,
"logits/rejected": -3.711491823196411,
"logps/chosen": -10.302391052246094,
"logps/rejected": -29.03504753112793,
"loss": 0.1226,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7932891845703125,
"rewards/margins": 2.7932891845703125,
"rewards/rejected": 0.0,
"step": 1017
},
{
"epoch": 5.687150837988827,
"grad_norm": 3.790138742496397,
"learning_rate": 8.142573773335744e-07,
"logits/chosen": -3.3959226608276367,
"logits/rejected": -3.3778514862060547,
"logps/chosen": -6.375226974487305,
"logps/rejected": -24.757173538208008,
"loss": 0.118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3533895015716553,
"rewards/margins": 2.3533895015716553,
"rewards/rejected": 0.0,
"step": 1018
},
{
"epoch": 5.692737430167598,
"grad_norm": 4.346472453742642,
"learning_rate": 8.137830415973861e-07,
"logits/chosen": -3.6288812160491943,
"logits/rejected": -3.6254985332489014,
"logps/chosen": -3.6216018199920654,
"logps/rejected": -40.97502899169922,
"loss": 0.1124,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5600640773773193,
"rewards/margins": 1.5600640773773193,
"rewards/rejected": 0.0,
"step": 1019
},
{
"epoch": 5.698324022346369,
"grad_norm": 1.5770828705824953,
"learning_rate": 8.133082395237162e-07,
"logits/chosen": -3.769218683242798,
"logits/rejected": -3.8052897453308105,
"logps/chosen": -2.879276752471924,
"logps/rejected": -27.094223022460938,
"loss": 0.113,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.373605489730835,
"rewards/margins": 2.373605489730835,
"rewards/rejected": 0.0,
"step": 1020
},
{
"epoch": 5.698324022346369,
"eval_logits/chosen": -3.4237537384033203,
"eval_logits/rejected": -3.548767566680908,
"eval_logps/chosen": -21.93404197692871,
"eval_logps/rejected": -40.45651626586914,
"eval_loss": 0.5296804308891296,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.8135999441146851,
"eval_rewards/margins": 0.8135999441146851,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7144,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 1020
},
{
"epoch": 5.70391061452514,
"grad_norm": 3.777357781410946,
"learning_rate": 8.128329718182052e-07,
"logits/chosen": -3.8146719932556152,
"logits/rejected": -3.7009477615356445,
"logps/chosen": -6.057042598724365,
"logps/rejected": -34.36450958251953,
"loss": 0.1312,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0107688903808594,
"rewards/margins": 2.0107688903808594,
"rewards/rejected": 0.0,
"step": 1021
},
{
"epoch": 5.709497206703911,
"grad_norm": 2.5088495568479985,
"learning_rate": 8.123572391871855e-07,
"logits/chosen": -3.6293792724609375,
"logits/rejected": -3.667412281036377,
"logps/chosen": -5.019689559936523,
"logps/rejected": -38.291656494140625,
"loss": 0.1445,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.206315517425537,
"rewards/margins": 2.206315517425537,
"rewards/rejected": 0.0,
"step": 1022
},
{
"epoch": 5.715083798882682,
"grad_norm": 4.798630862002979,
"learning_rate": 8.118810423376808e-07,
"logits/chosen": -3.55879545211792,
"logits/rejected": -3.4452855587005615,
"logps/chosen": -11.348285675048828,
"logps/rejected": -28.296905517578125,
"loss": 0.1337,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0398333072662354,
"rewards/margins": 3.0398333072662354,
"rewards/rejected": 0.0,
"step": 1023
},
{
"epoch": 5.720670391061453,
"grad_norm": 5.99365526397894,
"learning_rate": 8.11404381977404e-07,
"logits/chosen": -3.6069586277008057,
"logits/rejected": -3.7181835174560547,
"logps/chosen": -7.550654888153076,
"logps/rejected": -25.838010787963867,
"loss": 0.1261,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.031066417694092,
"rewards/margins": 2.031066417694092,
"rewards/rejected": 0.0,
"step": 1024
},
{
"epoch": 5.726256983240224,
"grad_norm": 2.65472959335182,
"learning_rate": 8.109272588147576e-07,
"logits/chosen": -3.678741931915283,
"logits/rejected": -3.7180609703063965,
"logps/chosen": -11.061220169067383,
"logps/rejected": -19.70529556274414,
"loss": 0.1223,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.749025344848633,
"rewards/margins": 2.749025344848633,
"rewards/rejected": 0.0,
"step": 1025
},
{
"epoch": 5.731843575418995,
"grad_norm": 3.7313982458131734,
"learning_rate": 8.104496735588316e-07,
"logits/chosen": -3.725174903869629,
"logits/rejected": -3.6457762718200684,
"logps/chosen": -11.355713844299316,
"logps/rejected": -36.744564056396484,
"loss": 0.128,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9542886018753052,
"rewards/margins": 1.9542886018753052,
"rewards/rejected": 0.0,
"step": 1026
},
{
"epoch": 5.737430167597766,
"grad_norm": 4.039599732016783,
"learning_rate": 8.099716269194025e-07,
"logits/chosen": -3.6347756385803223,
"logits/rejected": -4.101277828216553,
"logps/chosen": -4.197378635406494,
"logps/rejected": -55.23566436767578,
"loss": 0.1538,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1737470626831055,
"rewards/margins": 2.1737470626831055,
"rewards/rejected": 0.0,
"step": 1027
},
{
"epoch": 5.743016759776537,
"grad_norm": 9.835650662654556,
"learning_rate": 8.094931196069327e-07,
"logits/chosen": -3.9969725608825684,
"logits/rejected": -4.015323162078857,
"logps/chosen": -6.438171863555908,
"logps/rejected": -23.971437454223633,
"loss": 0.1465,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.403317451477051,
"rewards/margins": 2.403317451477051,
"rewards/rejected": 0.0,
"step": 1028
},
{
"epoch": 5.748603351955307,
"grad_norm": 3.742173466066059,
"learning_rate": 8.090141523325698e-07,
"logits/chosen": -3.7224981784820557,
"logits/rejected": -3.749600648880005,
"logps/chosen": -20.025390625,
"logps/rejected": -26.75469207763672,
"loss": 0.1077,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9519076347351074,
"rewards/margins": 2.9519076347351074,
"rewards/rejected": 0.0,
"step": 1029
},
{
"epoch": 5.754189944134078,
"grad_norm": 3.520843064313235,
"learning_rate": 8.085347258081439e-07,
"logits/chosen": -3.7439513206481934,
"logits/rejected": -3.7707877159118652,
"logps/chosen": -6.769908905029297,
"logps/rejected": -40.603904724121094,
"loss": 0.1516,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.562917947769165,
"rewards/margins": 2.562917947769165,
"rewards/rejected": 0.0,
"step": 1030
},
{
"epoch": 5.759776536312849,
"grad_norm": 2.988490256300306,
"learning_rate": 8.080548407461684e-07,
"logits/chosen": -3.547328472137451,
"logits/rejected": -3.535720109939575,
"logps/chosen": -10.439088821411133,
"logps/rejected": -41.473087310791016,
"loss": 0.1374,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3769493103027344,
"rewards/margins": 2.3769493103027344,
"rewards/rejected": 0.0,
"step": 1031
},
{
"epoch": 5.76536312849162,
"grad_norm": 3.8913613629080124,
"learning_rate": 8.07574497859838e-07,
"logits/chosen": -3.9097604751586914,
"logits/rejected": -3.8350741863250732,
"logps/chosen": -2.4992494583129883,
"logps/rejected": -51.154930114746094,
"loss": 0.1552,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5016214847564697,
"rewards/margins": 1.5016214847564697,
"rewards/rejected": 0.0,
"step": 1032
},
{
"epoch": 5.770949720670391,
"grad_norm": 3.2573820900391017,
"learning_rate": 8.070936978630279e-07,
"logits/chosen": -3.601395845413208,
"logits/rejected": -3.5447804927825928,
"logps/chosen": -10.644867897033691,
"logps/rejected": -23.28240203857422,
"loss": 0.1316,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.515011787414551,
"rewards/margins": 2.515011787414551,
"rewards/rejected": 0.0,
"step": 1033
},
{
"epoch": 5.776536312849162,
"grad_norm": 4.333177457770317,
"learning_rate": 8.066124414702921e-07,
"logits/chosen": -3.6848273277282715,
"logits/rejected": -3.734130620956421,
"logps/chosen": -9.491111755371094,
"logps/rejected": -26.304800033569336,
"loss": 0.1287,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5129756927490234,
"rewards/margins": 2.5129756927490234,
"rewards/rejected": 0.0,
"step": 1034
},
{
"epoch": 5.782122905027933,
"grad_norm": 2.348425439806568,
"learning_rate": 8.06130729396864e-07,
"logits/chosen": -3.655939817428589,
"logits/rejected": -3.356834888458252,
"logps/chosen": -4.262138366699219,
"logps/rejected": -49.86429977416992,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8290448188781738,
"rewards/margins": 1.8290448188781738,
"rewards/rejected": 0.0,
"step": 1035
},
{
"epoch": 5.787709497206704,
"grad_norm": 8.205788491058067,
"learning_rate": 8.056485623586529e-07,
"logits/chosen": -3.6774847507476807,
"logits/rejected": -3.6492533683776855,
"logps/chosen": -28.121702194213867,
"logps/rejected": -27.963851928710938,
"loss": 0.1575,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6612625122070312,
"rewards/margins": 2.6612625122070312,
"rewards/rejected": 0.0,
"step": 1036
},
{
"epoch": 5.793296089385475,
"grad_norm": 3.7346531198512865,
"learning_rate": 8.051659410722449e-07,
"logits/chosen": -3.5452606678009033,
"logits/rejected": -3.6970622539520264,
"logps/chosen": -8.915273666381836,
"logps/rejected": -38.14643096923828,
"loss": 0.1517,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.321495532989502,
"rewards/margins": 2.321495532989502,
"rewards/rejected": 0.0,
"step": 1037
},
{
"epoch": 5.798882681564246,
"grad_norm": 4.525769205141487,
"learning_rate": 8.046828662549016e-07,
"logits/chosen": -3.406954526901245,
"logits/rejected": -3.3492534160614014,
"logps/chosen": -19.924955368041992,
"logps/rejected": -37.408931732177734,
"loss": 0.1248,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.026550769805908,
"rewards/margins": 2.026550769805908,
"rewards/rejected": 0.0,
"step": 1038
},
{
"epoch": 5.804469273743017,
"grad_norm": 2.844572501441276,
"learning_rate": 8.04199338624558e-07,
"logits/chosen": -3.7834746837615967,
"logits/rejected": -3.807988405227661,
"logps/chosen": -6.659724235534668,
"logps/rejected": -28.40852165222168,
"loss": 0.1231,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8151044845581055,
"rewards/margins": 2.8151044845581055,
"rewards/rejected": 0.0,
"step": 1039
},
{
"epoch": 5.810055865921788,
"grad_norm": 7.162773196323615,
"learning_rate": 8.037153588998223e-07,
"logits/chosen": -3.5483334064483643,
"logits/rejected": -3.4280738830566406,
"logps/chosen": -22.878292083740234,
"logps/rejected": -43.18260192871094,
"loss": 0.1521,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.384970188140869,
"rewards/margins": 2.384970188140869,
"rewards/rejected": 0.0,
"step": 1040
},
{
"epoch": 5.810055865921788,
"eval_logits/chosen": -3.41727876663208,
"eval_logits/rejected": -3.543991804122925,
"eval_logps/chosen": -21.917339324951172,
"eval_logps/rejected": -40.18777847290039,
"eval_loss": 0.5309573411941528,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.8152703046798706,
"eval_rewards/margins": 0.8152703046798706,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7306,
"eval_samples_per_second": 9.471,
"eval_steps_per_second": 0.306,
"step": 1040
},
{
"epoch": 5.815642458100559,
"grad_norm": 5.937504038655379,
"learning_rate": 8.032309277999745e-07,
"logits/chosen": -3.1624813079833984,
"logits/rejected": -3.0980820655822754,
"logps/chosen": -14.842673301696777,
"logps/rejected": -29.77103614807129,
"loss": 0.1342,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7918972969055176,
"rewards/margins": 2.7918972969055176,
"rewards/rejected": 0.0,
"step": 1041
},
{
"epoch": 5.82122905027933,
"grad_norm": 2.2425943551312097,
"learning_rate": 8.027460460449655e-07,
"logits/chosen": -3.483254909515381,
"logits/rejected": -3.4363527297973633,
"logps/chosen": -11.010003089904785,
"logps/rejected": -24.420269012451172,
"loss": 0.1253,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.189192295074463,
"rewards/margins": 2.189192295074463,
"rewards/rejected": 0.0,
"step": 1042
},
{
"epoch": 5.826815642458101,
"grad_norm": 3.844927643246877,
"learning_rate": 8.022607143554159e-07,
"logits/chosen": -3.4990077018737793,
"logits/rejected": -3.4998536109924316,
"logps/chosen": -7.874691963195801,
"logps/rejected": -34.47044372558594,
"loss": 0.1196,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.21986985206604,
"rewards/margins": 2.21986985206604,
"rewards/rejected": 0.0,
"step": 1043
},
{
"epoch": 5.832402234636872,
"grad_norm": 4.981271074838592,
"learning_rate": 8.017749334526151e-07,
"logits/chosen": -3.855689764022827,
"logits/rejected": -3.902771234512329,
"logps/chosen": -10.003469467163086,
"logps/rejected": -64.29655456542969,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3023786544799805,
"rewards/margins": 2.3023786544799805,
"rewards/rejected": 0.0,
"step": 1044
},
{
"epoch": 5.837988826815643,
"grad_norm": 3.429561411893467,
"learning_rate": 8.012887040585201e-07,
"logits/chosen": -3.5651566982269287,
"logits/rejected": -3.5544581413269043,
"logps/chosen": -13.647613525390625,
"logps/rejected": -47.53219985961914,
"loss": 0.1406,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.001678466796875,
"rewards/margins": 2.001678466796875,
"rewards/rejected": 0.0,
"step": 1045
},
{
"epoch": 5.843575418994414,
"grad_norm": 8.603228515077392,
"learning_rate": 8.008020268957542e-07,
"logits/chosen": -3.314690351486206,
"logits/rejected": -3.366624355316162,
"logps/chosen": -9.321515083312988,
"logps/rejected": -22.403581619262695,
"loss": 0.1534,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4352869987487793,
"rewards/margins": 2.4352869987487793,
"rewards/rejected": 0.0,
"step": 1046
},
{
"epoch": 5.849162011173185,
"grad_norm": 3.6846850123231274,
"learning_rate": 8.003149026876064e-07,
"logits/chosen": -3.2948601245880127,
"logits/rejected": -3.2105228900909424,
"logps/chosen": -3.3570847511291504,
"logps/rejected": -51.15724563598633,
"loss": 0.1373,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.01554012298584,
"rewards/margins": 2.01554012298584,
"rewards/rejected": 0.0,
"step": 1047
},
{
"epoch": 5.854748603351956,
"grad_norm": 2.189238114286271,
"learning_rate": 7.998273321580301e-07,
"logits/chosen": -3.7571964263916016,
"logits/rejected": -3.6517891883850098,
"logps/chosen": -23.5264949798584,
"logps/rejected": -43.30960464477539,
"loss": 0.0983,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.620962142944336,
"rewards/margins": 3.620962142944336,
"rewards/rejected": 0.0,
"step": 1048
},
{
"epoch": 5.860335195530726,
"grad_norm": 5.025232426897415,
"learning_rate": 7.99339316031642e-07,
"logits/chosen": -3.4721052646636963,
"logits/rejected": -3.5792720317840576,
"logps/chosen": -16.34332847595215,
"logps/rejected": -51.911041259765625,
"loss": 0.1335,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.748723268508911,
"rewards/margins": 2.748723268508911,
"rewards/rejected": 0.0,
"step": 1049
},
{
"epoch": 5.865921787709497,
"grad_norm": 5.05088904377592,
"learning_rate": 7.988508550337206e-07,
"logits/chosen": -3.437150001525879,
"logits/rejected": -3.4027011394500732,
"logps/chosen": -12.300931930541992,
"logps/rejected": -26.506114959716797,
"loss": 0.1632,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8190996646881104,
"rewards/margins": 2.8190996646881104,
"rewards/rejected": 0.0,
"step": 1050
},
{
"epoch": 5.871508379888268,
"grad_norm": 3.384330851640664,
"learning_rate": 7.983619498902063e-07,
"logits/chosen": -3.9197616577148438,
"logits/rejected": -3.4385604858398438,
"logps/chosen": -8.68655014038086,
"logps/rejected": -55.796478271484375,
"loss": 0.1511,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1019668579101562,
"rewards/margins": 2.1019668579101562,
"rewards/rejected": 0.0,
"step": 1051
},
{
"epoch": 5.877094972067039,
"grad_norm": 4.992564809277405,
"learning_rate": 7.978726013276993e-07,
"logits/chosen": -3.516857862472534,
"logits/rejected": -3.541065216064453,
"logps/chosen": -2.68699049949646,
"logps/rejected": -36.906333923339844,
"loss": 0.1532,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6484850645065308,
"rewards/margins": 1.6484850645065308,
"rewards/rejected": 0.0,
"step": 1052
},
{
"epoch": 5.88268156424581,
"grad_norm": 5.892173992077341,
"learning_rate": 7.973828100734583e-07,
"logits/chosen": -3.618582248687744,
"logits/rejected": -3.6364614963531494,
"logps/chosen": -36.14175796508789,
"logps/rejected": -38.002197265625,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6985483169555664,
"rewards/margins": 2.6985483169555664,
"rewards/rejected": 0.0,
"step": 1053
},
{
"epoch": 5.888268156424581,
"grad_norm": 2.657512711385388,
"learning_rate": 7.968925768554005e-07,
"logits/chosen": -3.326869487762451,
"logits/rejected": -3.615558385848999,
"logps/chosen": -12.183609008789062,
"logps/rejected": -40.154624938964844,
"loss": 0.1214,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.736098051071167,
"rewards/margins": 2.736098051071167,
"rewards/rejected": 0.0,
"step": 1054
},
{
"epoch": 5.893854748603352,
"grad_norm": 2.7462352809129795,
"learning_rate": 7.964019024021e-07,
"logits/chosen": -3.8306450843811035,
"logits/rejected": -3.7286834716796875,
"logps/chosen": -6.44540548324585,
"logps/rejected": -84.51054382324219,
"loss": 0.1018,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.596806526184082,
"rewards/margins": 2.596806526184082,
"rewards/rejected": 0.0,
"step": 1055
},
{
"epoch": 5.899441340782123,
"grad_norm": 2.1605553755845204,
"learning_rate": 7.959107874427862e-07,
"logits/chosen": -3.3188657760620117,
"logits/rejected": -3.349465847015381,
"logps/chosen": -2.676693916320801,
"logps/rejected": -86.0643081665039,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8748719692230225,
"rewards/margins": 1.8748719692230225,
"rewards/rejected": 0.0,
"step": 1056
},
{
"epoch": 5.905027932960894,
"grad_norm": 6.44224485187991,
"learning_rate": 7.954192327073434e-07,
"logits/chosen": -3.76053524017334,
"logits/rejected": -3.7138631343841553,
"logps/chosen": -27.124252319335938,
"logps/rejected": -33.14948654174805,
"loss": 0.1293,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.01265025138855,
"rewards/margins": 3.01265025138855,
"rewards/rejected": 0.0,
"step": 1057
},
{
"epoch": 5.910614525139665,
"grad_norm": 5.831035365072501,
"learning_rate": 7.949272389263095e-07,
"logits/chosen": -3.683530807495117,
"logits/rejected": -3.5552704334259033,
"logps/chosen": -12.019757270812988,
"logps/rejected": -34.60222625732422,
"loss": 0.1243,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1793899536132812,
"rewards/margins": 3.1793899536132812,
"rewards/rejected": 0.0,
"step": 1058
},
{
"epoch": 5.916201117318436,
"grad_norm": 3.837026863586912,
"learning_rate": 7.944348068308749e-07,
"logits/chosen": -3.480384349822998,
"logits/rejected": -3.5684092044830322,
"logps/chosen": -3.4614875316619873,
"logps/rejected": -17.95250701904297,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.811203956604004,
"rewards/margins": 1.811203956604004,
"rewards/rejected": 0.0,
"step": 1059
},
{
"epoch": 5.921787709497207,
"grad_norm": 4.358596389031213,
"learning_rate": 7.939419371528813e-07,
"logits/chosen": -3.3316895961761475,
"logits/rejected": -3.5211856365203857,
"logps/chosen": -7.271438121795654,
"logps/rejected": -38.95193099975586,
"loss": 0.1454,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3479537963867188,
"rewards/margins": 2.3479537963867188,
"rewards/rejected": 0.0,
"step": 1060
},
{
"epoch": 5.921787709497207,
"eval_logits/chosen": -3.4203200340270996,
"eval_logits/rejected": -3.545351028442383,
"eval_logps/chosen": -21.673046112060547,
"eval_logps/rejected": -40.24152755737305,
"eval_loss": 0.5311393141746521,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.8396994471549988,
"eval_rewards/margins": 0.8396994471549988,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7014,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 1060
},
{
"epoch": 5.927374301675978,
"grad_norm": 6.649311192603633,
"learning_rate": 7.93448630624821e-07,
"logits/chosen": -3.718939781188965,
"logits/rejected": -3.7220516204833984,
"logps/chosen": -20.689510345458984,
"logps/rejected": -36.891448974609375,
"loss": 0.1423,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5494956970214844,
"rewards/margins": 2.5494956970214844,
"rewards/rejected": 0.0,
"step": 1061
},
{
"epoch": 5.932960893854749,
"grad_norm": 4.235624484730155,
"learning_rate": 7.929548879798352e-07,
"logits/chosen": -3.562255859375,
"logits/rejected": -3.632988691329956,
"logps/chosen": -23.885807037353516,
"logps/rejected": -48.20660400390625,
"loss": 0.1076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7936930656433105,
"rewards/margins": 3.7936930656433105,
"rewards/rejected": 0.0,
"step": 1062
},
{
"epoch": 5.93854748603352,
"grad_norm": 6.259998385431391,
"learning_rate": 7.924607099517134e-07,
"logits/chosen": -3.5328803062438965,
"logits/rejected": -3.5107877254486084,
"logps/chosen": -24.228809356689453,
"logps/rejected": -40.1533317565918,
"loss": 0.1482,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.37412166595459,
"rewards/margins": 2.37412166595459,
"rewards/rejected": 0.0,
"step": 1063
},
{
"epoch": 5.94413407821229,
"grad_norm": 2.4960125141521794,
"learning_rate": 7.919660972748923e-07,
"logits/chosen": -3.378631591796875,
"logits/rejected": -3.5496344566345215,
"logps/chosen": -2.1727137565612793,
"logps/rejected": -64.11479187011719,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8212121725082397,
"rewards/margins": 1.8212121725082397,
"rewards/rejected": 0.0,
"step": 1064
},
{
"epoch": 5.949720670391061,
"grad_norm": 2.855849965245941,
"learning_rate": 7.914710506844544e-07,
"logits/chosen": -3.1338109970092773,
"logits/rejected": -3.261589765548706,
"logps/chosen": -9.45151138305664,
"logps/rejected": -44.79124069213867,
"loss": 0.1463,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.204951763153076,
"rewards/margins": 2.204951763153076,
"rewards/rejected": 0.0,
"step": 1065
},
{
"epoch": 5.955307262569832,
"grad_norm": 2.7537517537632286,
"learning_rate": 7.909755709161268e-07,
"logits/chosen": -3.887150526046753,
"logits/rejected": -3.8906705379486084,
"logps/chosen": -2.40024995803833,
"logps/rejected": -54.04871368408203,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8678940534591675,
"rewards/margins": 1.8678940534591675,
"rewards/rejected": 0.0,
"step": 1066
},
{
"epoch": 5.960893854748603,
"grad_norm": 3.683780045334,
"learning_rate": 7.904796587062811e-07,
"logits/chosen": -3.3832907676696777,
"logits/rejected": -3.313124418258667,
"logps/chosen": -16.009614944458008,
"logps/rejected": -37.07920837402344,
"loss": 0.1333,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.613959789276123,
"rewards/margins": 2.613959789276123,
"rewards/rejected": 0.0,
"step": 1067
},
{
"epoch": 5.966480446927374,
"grad_norm": 3.518954282147113,
"learning_rate": 7.899833147919309e-07,
"logits/chosen": -2.8122658729553223,
"logits/rejected": -2.786893367767334,
"logps/chosen": -6.672562599182129,
"logps/rejected": -40.3980712890625,
"loss": 0.1257,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3488590717315674,
"rewards/margins": 2.3488590717315674,
"rewards/rejected": 0.0,
"step": 1068
},
{
"epoch": 5.972067039106145,
"grad_norm": 5.106324928377724,
"learning_rate": 7.894865399107317e-07,
"logits/chosen": -3.834169387817383,
"logits/rejected": -3.4647724628448486,
"logps/chosen": -18.958227157592773,
"logps/rejected": -43.90966796875,
"loss": 0.129,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4929468631744385,
"rewards/margins": 2.4929468631744385,
"rewards/rejected": 0.0,
"step": 1069
},
{
"epoch": 5.977653631284916,
"grad_norm": 8.920813378166514,
"learning_rate": 7.889893348009795e-07,
"logits/chosen": -3.744349956512451,
"logits/rejected": -3.489316940307617,
"logps/chosen": -17.75255012512207,
"logps/rejected": -43.721099853515625,
"loss": 0.152,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.822836399078369,
"rewards/margins": 2.822836399078369,
"rewards/rejected": 0.0,
"step": 1070
},
{
"epoch": 5.983240223463687,
"grad_norm": 13.605019074673555,
"learning_rate": 7.884917002016096e-07,
"logits/chosen": -3.423605442047119,
"logits/rejected": -3.5889503955841064,
"logps/chosen": -14.498847007751465,
"logps/rejected": -23.585159301757812,
"loss": 0.1945,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.329728126525879,
"rewards/margins": 1.329728126525879,
"rewards/rejected": 0.0,
"step": 1071
},
{
"epoch": 5.988826815642458,
"grad_norm": 3.1418183217119156,
"learning_rate": 7.879936368521956e-07,
"logits/chosen": -3.825878143310547,
"logits/rejected": -3.856353998184204,
"logps/chosen": -4.65446662902832,
"logps/rejected": -27.227596282958984,
"loss": 0.1321,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.210226058959961,
"rewards/margins": 2.210226058959961,
"rewards/rejected": 0.0,
"step": 1072
},
{
"epoch": 5.994413407821229,
"grad_norm": 6.986694525808208,
"learning_rate": 7.874951454929483e-07,
"logits/chosen": -3.3854620456695557,
"logits/rejected": -3.423590660095215,
"logps/chosen": -4.558568477630615,
"logps/rejected": -60.397037506103516,
"loss": 0.124,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3430333137512207,
"rewards/margins": 2.3430333137512207,
"rewards/rejected": 0.0,
"step": 1073
},
{
"epoch": 6.0,
"grad_norm": 2.334469710023271,
"learning_rate": 7.869962268647149e-07,
"logits/chosen": -3.527007818222046,
"logits/rejected": -3.4847359657287598,
"logps/chosen": -4.882283687591553,
"logps/rejected": -26.33693504333496,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9604823589324951,
"rewards/margins": 1.9604823589324951,
"rewards/rejected": 0.0,
"step": 1074
},
{
"epoch": 6.005586592178771,
"grad_norm": 2.4628652540684617,
"learning_rate": 7.864968817089771e-07,
"logits/chosen": -3.7268893718719482,
"logits/rejected": -3.7700512409210205,
"logps/chosen": -6.030098915100098,
"logps/rejected": -31.229232788085938,
"loss": 0.1234,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.246878147125244,
"rewards/margins": 2.246878147125244,
"rewards/rejected": 0.0,
"step": 1075
},
{
"epoch": 6.011173184357542,
"grad_norm": 2.202430159342966,
"learning_rate": 7.859971107678507e-07,
"logits/chosen": -3.4105281829833984,
"logits/rejected": -3.4312808513641357,
"logps/chosen": -8.663097381591797,
"logps/rejected": -19.917152404785156,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.780505895614624,
"rewards/margins": 2.780505895614624,
"rewards/rejected": 0.0,
"step": 1076
},
{
"epoch": 6.016759776536313,
"grad_norm": 2.712624777228195,
"learning_rate": 7.854969147840844e-07,
"logits/chosen": -3.333019733428955,
"logits/rejected": -3.460301399230957,
"logps/chosen": -39.58592224121094,
"logps/rejected": -66.58130645751953,
"loss": 0.1376,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9631738662719727,
"rewards/margins": 2.9631738662719727,
"rewards/rejected": 0.0,
"step": 1077
},
{
"epoch": 6.022346368715084,
"grad_norm": 3.5542963226582662,
"learning_rate": 7.849962945010586e-07,
"logits/chosen": -3.498375415802002,
"logits/rejected": -3.4425761699676514,
"logps/chosen": -21.79305648803711,
"logps/rejected": -28.10723304748535,
"loss": 0.1335,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7957966327667236,
"rewards/margins": 2.7957966327667236,
"rewards/rejected": 0.0,
"step": 1078
},
{
"epoch": 6.027932960893855,
"grad_norm": 1.490515196796577,
"learning_rate": 7.84495250662784e-07,
"logits/chosen": -3.419386625289917,
"logits/rejected": -3.4304134845733643,
"logps/chosen": -4.168264389038086,
"logps/rejected": -52.03262710571289,
"loss": 0.1092,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.539543390274048,
"rewards/margins": 2.539543390274048,
"rewards/rejected": 0.0,
"step": 1079
},
{
"epoch": 6.033519553072626,
"grad_norm": 2.0665741161577853,
"learning_rate": 7.839937840139012e-07,
"logits/chosen": -3.3358426094055176,
"logits/rejected": -3.5571582317352295,
"logps/chosen": -12.014814376831055,
"logps/rejected": -39.008941650390625,
"loss": 0.0924,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8855361938476562,
"rewards/margins": 2.8855361938476562,
"rewards/rejected": 0.0,
"step": 1080
},
{
"epoch": 6.033519553072626,
"eval_logits/chosen": -3.4154648780822754,
"eval_logits/rejected": -3.541161298751831,
"eval_logps/chosen": -22.391386032104492,
"eval_logps/rejected": -40.84318161010742,
"eval_loss": 0.5555992126464844,
"eval_rewards/accuracies": 0.8999999761581421,
"eval_rewards/chosen": 0.7678656578063965,
"eval_rewards/margins": 0.7678656578063965,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7461,
"eval_samples_per_second": 9.467,
"eval_steps_per_second": 0.305,
"step": 1080
},
{
"epoch": 6.039106145251397,
"grad_norm": 3.631732747504115,
"learning_rate": 7.834918952996787e-07,
"logits/chosen": -3.3401753902435303,
"logits/rejected": -3.1974246501922607,
"logps/chosen": -13.715397834777832,
"logps/rejected": -34.694549560546875,
"loss": 0.1262,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8850817680358887,
"rewards/margins": 2.8850817680358887,
"rewards/rejected": 0.0,
"step": 1081
},
{
"epoch": 6.044692737430168,
"grad_norm": 1.400512054416602,
"learning_rate": 7.829895852660129e-07,
"logits/chosen": -3.5798497200012207,
"logits/rejected": -3.602731466293335,
"logps/chosen": -9.231255531311035,
"logps/rejected": -17.62391471862793,
"loss": 0.1119,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3963284492492676,
"rewards/margins": 2.3963284492492676,
"rewards/rejected": 0.0,
"step": 1082
},
{
"epoch": 6.050279329608939,
"grad_norm": 3.3975025017852483,
"learning_rate": 7.824868546594254e-07,
"logits/chosen": -3.515927791595459,
"logits/rejected": -3.5182993412017822,
"logps/chosen": -10.096548080444336,
"logps/rejected": -19.822113037109375,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.589899778366089,
"rewards/margins": 2.589899778366089,
"rewards/rejected": 0.0,
"step": 1083
},
{
"epoch": 6.055865921787709,
"grad_norm": 2.0597535081699094,
"learning_rate": 7.819837042270639e-07,
"logits/chosen": -3.2524971961975098,
"logits/rejected": -3.3197362422943115,
"logps/chosen": -5.943603515625,
"logps/rejected": -34.231170654296875,
"loss": 0.13,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.120120048522949,
"rewards/margins": 2.120120048522949,
"rewards/rejected": 0.0,
"step": 1084
},
{
"epoch": 6.06145251396648,
"grad_norm": 1.3017971192466566,
"learning_rate": 7.814801347166992e-07,
"logits/chosen": -3.6389377117156982,
"logits/rejected": -3.7064850330352783,
"logps/chosen": -1.456046223640442,
"logps/rejected": -51.3591194152832,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.416407823562622,
"rewards/margins": 1.416407823562622,
"rewards/rejected": 0.0,
"step": 1085
},
{
"epoch": 6.067039106145251,
"grad_norm": 1.680492345553586,
"learning_rate": 7.809761468767256e-07,
"logits/chosen": -3.8416240215301514,
"logits/rejected": -3.710263729095459,
"logps/chosen": -17.1885986328125,
"logps/rejected": -58.12712097167969,
"loss": 0.1084,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.589221954345703,
"rewards/margins": 2.589221954345703,
"rewards/rejected": 0.0,
"step": 1086
},
{
"epoch": 6.072625698324022,
"grad_norm": 5.760094213348393,
"learning_rate": 7.804717414561585e-07,
"logits/chosen": -3.3121144771575928,
"logits/rejected": -3.3707149028778076,
"logps/chosen": -4.068519592285156,
"logps/rejected": -72.06900024414062,
"loss": 0.1257,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.890760898590088,
"rewards/margins": 1.890760898590088,
"rewards/rejected": 0.0,
"step": 1087
},
{
"epoch": 6.078212290502793,
"grad_norm": 5.42449667131543,
"learning_rate": 7.799669192046344e-07,
"logits/chosen": -3.530776023864746,
"logits/rejected": -3.664048433303833,
"logps/chosen": -5.868477821350098,
"logps/rejected": -43.213623046875,
"loss": 0.151,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.247279167175293,
"rewards/margins": 2.247279167175293,
"rewards/rejected": 0.0,
"step": 1088
},
{
"epoch": 6.083798882681564,
"grad_norm": 1.3059898216049628,
"learning_rate": 7.794616808724089e-07,
"logits/chosen": -3.378438711166382,
"logits/rejected": -3.362328052520752,
"logps/chosen": -13.31649112701416,
"logps/rejected": -52.14571762084961,
"loss": 0.1027,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.272395133972168,
"rewards/margins": 2.272395133972168,
"rewards/rejected": 0.0,
"step": 1089
},
{
"epoch": 6.089385474860335,
"grad_norm": 5.847080787887179,
"learning_rate": 7.789560272103563e-07,
"logits/chosen": -3.6775078773498535,
"logits/rejected": -3.790405511856079,
"logps/chosen": -6.6467742919921875,
"logps/rejected": -34.40592575073242,
"loss": 0.1252,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8223204612731934,
"rewards/margins": 2.8223204612731934,
"rewards/rejected": 0.0,
"step": 1090
},
{
"epoch": 6.094972067039106,
"grad_norm": 2.3429210829400433,
"learning_rate": 7.78449958969968e-07,
"logits/chosen": -3.4835174083709717,
"logits/rejected": -3.3803036212921143,
"logps/chosen": -3.2445549964904785,
"logps/rejected": -44.495182037353516,
"loss": 0.1124,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4009182453155518,
"rewards/margins": 2.4009182453155518,
"rewards/rejected": 0.0,
"step": 1091
},
{
"epoch": 6.100558659217877,
"grad_norm": 3.9284252763194507,
"learning_rate": 7.779434769033513e-07,
"logits/chosen": -3.2931106090545654,
"logits/rejected": -3.4320995807647705,
"logps/chosen": -6.39832067489624,
"logps/rejected": -28.484161376953125,
"loss": 0.1472,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5139878988265991,
"rewards/margins": 1.5139878988265991,
"rewards/rejected": 0.0,
"step": 1092
},
{
"epoch": 6.106145251396648,
"grad_norm": 2.075697482606786,
"learning_rate": 7.774365817632289e-07,
"logits/chosen": -3.5117132663726807,
"logits/rejected": -3.3721182346343994,
"logps/chosen": -5.439365386962891,
"logps/rejected": -24.08738136291504,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.383086681365967,
"rewards/margins": 2.383086681365967,
"rewards/rejected": 0.0,
"step": 1093
},
{
"epoch": 6.111731843575419,
"grad_norm": 3.798888734456071,
"learning_rate": 7.769292743029373e-07,
"logits/chosen": -3.698251724243164,
"logits/rejected": -3.4040896892547607,
"logps/chosen": -11.444732666015625,
"logps/rejected": -29.642263412475586,
"loss": 0.1101,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.803191661834717,
"rewards/margins": 2.803191661834717,
"rewards/rejected": 0.0,
"step": 1094
},
{
"epoch": 6.11731843575419,
"grad_norm": 1.4851485378801812,
"learning_rate": 7.764215552764258e-07,
"logits/chosen": -3.5323593616485596,
"logits/rejected": -3.5734283924102783,
"logps/chosen": -1.9695065021514893,
"logps/rejected": -35.47942352294922,
"loss": 0.1262,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.191638231277466,
"rewards/margins": 2.191638231277466,
"rewards/rejected": 0.0,
"step": 1095
},
{
"epoch": 6.122905027932961,
"grad_norm": 3.6155021859206116,
"learning_rate": 7.759134254382549e-07,
"logits/chosen": -3.7087323665618896,
"logits/rejected": -3.744468927383423,
"logps/chosen": -1.8839842081069946,
"logps/rejected": -15.816478729248047,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.832279920578003,
"rewards/margins": 1.832279920578003,
"rewards/rejected": 0.0,
"step": 1096
},
{
"epoch": 6.128491620111732,
"grad_norm": 1.505737569211408,
"learning_rate": 7.754048855435964e-07,
"logits/chosen": -3.5456957817077637,
"logits/rejected": -3.611483335494995,
"logps/chosen": -6.43549919128418,
"logps/rejected": -24.106260299682617,
"loss": 0.1291,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4427661895751953,
"rewards/margins": 2.4427661895751953,
"rewards/rejected": 0.0,
"step": 1097
},
{
"epoch": 6.134078212290503,
"grad_norm": 3.6197195471728216,
"learning_rate": 7.748959363482313e-07,
"logits/chosen": -3.683819532394409,
"logits/rejected": -3.587161064147949,
"logps/chosen": -5.716750621795654,
"logps/rejected": -20.286861419677734,
"loss": 0.1201,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9873279333114624,
"rewards/margins": 1.9873279333114624,
"rewards/rejected": 0.0,
"step": 1098
},
{
"epoch": 6.139664804469274,
"grad_norm": 2.5314161282485736,
"learning_rate": 7.743865786085481e-07,
"logits/chosen": -3.505862236022949,
"logits/rejected": -3.561208963394165,
"logps/chosen": -14.792437553405762,
"logps/rejected": -40.24578857421875,
"loss": 0.1318,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.70249605178833,
"rewards/margins": 2.70249605178833,
"rewards/rejected": 0.0,
"step": 1099
},
{
"epoch": 6.145251396648045,
"grad_norm": 1.6066077502153424,
"learning_rate": 7.738768130815438e-07,
"logits/chosen": -3.3270411491394043,
"logits/rejected": -3.530663251876831,
"logps/chosen": -4.3255462646484375,
"logps/rejected": -35.94929504394531,
"loss": 0.0872,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5646371841430664,
"rewards/margins": 2.5646371841430664,
"rewards/rejected": 0.0,
"step": 1100
},
{
"epoch": 6.145251396648045,
"eval_logits/chosen": -3.408015727996826,
"eval_logits/rejected": -3.5352530479431152,
"eval_logps/chosen": -23.45020866394043,
"eval_logps/rejected": -42.12792205810547,
"eval_loss": 0.5975612998008728,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.6619831919670105,
"eval_rewards/margins": 0.6619831919670105,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6794,
"eval_samples_per_second": 9.486,
"eval_steps_per_second": 0.306,
"step": 1100
},
{
"epoch": 6.150837988826815,
"grad_norm": 1.566825548015347,
"learning_rate": 7.733666405248203e-07,
"logits/chosen": -3.6134276390075684,
"logits/rejected": -3.5345585346221924,
"logps/chosen": -13.918472290039062,
"logps/rejected": -36.42315673828125,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.099599838256836,
"rewards/margins": 3.099599838256836,
"rewards/rejected": 0.0,
"step": 1101
},
{
"epoch": 6.156424581005586,
"grad_norm": 3.3667862061296234,
"learning_rate": 7.728560616965851e-07,
"logits/chosen": -3.281236171722412,
"logits/rejected": -3.4886529445648193,
"logps/chosen": -17.144258499145508,
"logps/rejected": -23.47547149658203,
"loss": 0.0909,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8349051475524902,
"rewards/margins": 2.8349051475524902,
"rewards/rejected": 0.0,
"step": 1102
},
{
"epoch": 6.162011173184357,
"grad_norm": 6.492421992577809,
"learning_rate": 7.723450773556493e-07,
"logits/chosen": -3.350792646408081,
"logits/rejected": -3.4875328540802,
"logps/chosen": -38.94091796875,
"logps/rejected": -45.85662841796875,
"loss": 0.1303,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9481279850006104,
"rewards/margins": 3.9481279850006104,
"rewards/rejected": 0.0,
"step": 1103
},
{
"epoch": 6.167597765363128,
"grad_norm": 2.360987681881043,
"learning_rate": 7.718336882614264e-07,
"logits/chosen": -3.499664545059204,
"logits/rejected": -3.5045888423919678,
"logps/chosen": -14.796363830566406,
"logps/rejected": -41.72339630126953,
"loss": 0.1211,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8475091457366943,
"rewards/margins": 2.8475091457366943,
"rewards/rejected": 0.0,
"step": 1104
},
{
"epoch": 6.173184357541899,
"grad_norm": 4.889157305192587,
"learning_rate": 7.71321895173932e-07,
"logits/chosen": -3.3693530559539795,
"logits/rejected": -3.3101730346679688,
"logps/chosen": -10.217439651489258,
"logps/rejected": -41.31106185913086,
"loss": 0.1324,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.839888095855713,
"rewards/margins": 2.839888095855713,
"rewards/rejected": 0.0,
"step": 1105
},
{
"epoch": 6.17877094972067,
"grad_norm": 1.4209093612701365,
"learning_rate": 7.708096988537815e-07,
"logits/chosen": -3.4527056217193604,
"logits/rejected": -3.456287145614624,
"logps/chosen": -1.576203465461731,
"logps/rejected": -31.562179565429688,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8774808645248413,
"rewards/margins": 1.8774808645248413,
"rewards/rejected": 0.0,
"step": 1106
},
{
"epoch": 6.184357541899441,
"grad_norm": 2.245537768734017,
"learning_rate": 7.702971000621898e-07,
"logits/chosen": -3.5811023712158203,
"logits/rejected": -3.478609561920166,
"logps/chosen": -20.62826156616211,
"logps/rejected": -53.61195373535156,
"loss": 0.1081,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.106804609298706,
"rewards/margins": 3.106804609298706,
"rewards/rejected": 0.0,
"step": 1107
},
{
"epoch": 6.189944134078212,
"grad_norm": 5.3723395536066265,
"learning_rate": 7.697840995609703e-07,
"logits/chosen": -3.7272963523864746,
"logits/rejected": -3.635612964630127,
"logps/chosen": -3.4792723655700684,
"logps/rejected": -27.30158805847168,
"loss": 0.1521,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.251101016998291,
"rewards/margins": 2.251101016998291,
"rewards/rejected": 0.0,
"step": 1108
},
{
"epoch": 6.195530726256983,
"grad_norm": 7.300222095177023,
"learning_rate": 7.692706981125328e-07,
"logits/chosen": -3.196547508239746,
"logits/rejected": -3.2908828258514404,
"logps/chosen": -3.8858413696289062,
"logps/rejected": -41.947906494140625,
"loss": 0.1611,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6939663887023926,
"rewards/margins": 1.6939663887023926,
"rewards/rejected": 0.0,
"step": 1109
},
{
"epoch": 6.201117318435754,
"grad_norm": 3.8000838724091492,
"learning_rate": 7.687568964798835e-07,
"logits/chosen": -3.5294885635375977,
"logits/rejected": -3.667855739593506,
"logps/chosen": -23.49508285522461,
"logps/rejected": -42.863075256347656,
"loss": 0.1323,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.693263053894043,
"rewards/margins": 3.693263053894043,
"rewards/rejected": 0.0,
"step": 1110
},
{
"epoch": 6.206703910614525,
"grad_norm": 1.4707324906429422,
"learning_rate": 7.682426954266231e-07,
"logits/chosen": -3.5440869331359863,
"logits/rejected": -3.706747531890869,
"logps/chosen": -10.006698608398438,
"logps/rejected": -21.6866455078125,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6204147338867188,
"rewards/margins": 2.6204147338867188,
"rewards/rejected": 0.0,
"step": 1111
},
{
"epoch": 6.212290502793296,
"grad_norm": 2.95793580906938,
"learning_rate": 7.677280957169459e-07,
"logits/chosen": -3.447383165359497,
"logits/rejected": -3.511056661605835,
"logps/chosen": -3.379901647567749,
"logps/rejected": -44.22206497192383,
"loss": 0.1525,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.633761167526245,
"rewards/margins": 2.633761167526245,
"rewards/rejected": 0.0,
"step": 1112
},
{
"epoch": 6.217877094972067,
"grad_norm": 3.9685239765315483,
"learning_rate": 7.672130981156386e-07,
"logits/chosen": -3.4409449100494385,
"logits/rejected": -3.4873509407043457,
"logps/chosen": -4.753771781921387,
"logps/rejected": -38.67449188232422,
"loss": 0.1307,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.859330654144287,
"rewards/margins": 1.859330654144287,
"rewards/rejected": 0.0,
"step": 1113
},
{
"epoch": 6.223463687150838,
"grad_norm": 5.311161386369822,
"learning_rate": 7.666977033880795e-07,
"logits/chosen": -3.696667432785034,
"logits/rejected": -3.5870654582977295,
"logps/chosen": -11.565668106079102,
"logps/rejected": -21.682395935058594,
"loss": 0.1037,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.505594491958618,
"rewards/margins": 2.505594491958618,
"rewards/rejected": 0.0,
"step": 1114
},
{
"epoch": 6.229050279329609,
"grad_norm": 2.649344486648601,
"learning_rate": 7.661819123002366e-07,
"logits/chosen": -3.3797621726989746,
"logits/rejected": -3.473595380783081,
"logps/chosen": -2.7563633918762207,
"logps/rejected": -90.32902526855469,
"loss": 0.1225,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.310032367706299,
"rewards/margins": 2.310032367706299,
"rewards/rejected": 0.0,
"step": 1115
},
{
"epoch": 6.23463687150838,
"grad_norm": 2.7494213517881807,
"learning_rate": 7.65665725618668e-07,
"logits/chosen": -3.3604629039764404,
"logits/rejected": -3.2099506855010986,
"logps/chosen": -22.696582794189453,
"logps/rejected": -47.31389617919922,
"loss": 0.1023,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.088898181915283,
"rewards/margins": 3.088898181915283,
"rewards/rejected": 0.0,
"step": 1116
},
{
"epoch": 6.240223463687151,
"grad_norm": 4.959981557959078,
"learning_rate": 7.651491441105187e-07,
"logits/chosen": -3.936720848083496,
"logits/rejected": -3.9750871658325195,
"logps/chosen": -2.2815120220184326,
"logps/rejected": -25.254283905029297,
"loss": 0.1183,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.22707200050354,
"rewards/margins": 2.22707200050354,
"rewards/rejected": 0.0,
"step": 1117
},
{
"epoch": 6.245810055865922,
"grad_norm": 4.277199276005058,
"learning_rate": 7.646321685435205e-07,
"logits/chosen": -3.290116548538208,
"logits/rejected": -3.431396245956421,
"logps/chosen": -2.931079864501953,
"logps/rejected": -33.018619537353516,
"loss": 0.1366,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4505295753479004,
"rewards/margins": 2.4505295753479004,
"rewards/rejected": 0.0,
"step": 1118
},
{
"epoch": 6.251396648044693,
"grad_norm": 1.613438108862121,
"learning_rate": 7.641147996859916e-07,
"logits/chosen": -3.8043417930603027,
"logits/rejected": -3.849745035171509,
"logps/chosen": -1.462568759918213,
"logps/rejected": -29.720266342163086,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.598609447479248,
"rewards/margins": 1.598609447479248,
"rewards/rejected": 0.0,
"step": 1119
},
{
"epoch": 6.256983240223463,
"grad_norm": 2.9443967723445663,
"learning_rate": 7.63597038306834e-07,
"logits/chosen": -3.5113418102264404,
"logits/rejected": -3.5904808044433594,
"logps/chosen": -1.8839757442474365,
"logps/rejected": -21.58673858642578,
"loss": 0.1171,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9584304094314575,
"rewards/margins": 1.9584304094314575,
"rewards/rejected": 0.0,
"step": 1120
},
{
"epoch": 6.256983240223463,
"eval_logits/chosen": -3.404155731201172,
"eval_logits/rejected": -3.5301384925842285,
"eval_logps/chosen": -23.150352478027344,
"eval_logps/rejected": -42.37028503417969,
"eval_loss": 0.586790144443512,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.6919690370559692,
"eval_rewards/margins": 0.6919690370559692,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7298,
"eval_samples_per_second": 9.471,
"eval_steps_per_second": 0.306,
"step": 1120
},
{
"epoch": 6.262569832402234,
"grad_norm": 2.1725677329370794,
"learning_rate": 7.630788851755334e-07,
"logits/chosen": -3.3949060440063477,
"logits/rejected": -3.3819336891174316,
"logps/chosen": -2.6754307746887207,
"logps/rejected": -24.352420806884766,
"loss": 0.1182,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.228742837905884,
"rewards/margins": 2.228742837905884,
"rewards/rejected": 0.0,
"step": 1121
},
{
"epoch": 6.268156424581005,
"grad_norm": 1.3327566934330697,
"learning_rate": 7.625603410621578e-07,
"logits/chosen": -3.430837392807007,
"logits/rejected": -3.391392946243286,
"logps/chosen": -7.514692306518555,
"logps/rejected": -31.07344627380371,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1787056922912598,
"rewards/margins": 2.1787056922912598,
"rewards/rejected": 0.0,
"step": 1122
},
{
"epoch": 6.273743016759776,
"grad_norm": 2.3769441085508225,
"learning_rate": 7.620414067373558e-07,
"logits/chosen": -3.2495040893554688,
"logits/rejected": -3.4531049728393555,
"logps/chosen": -11.662211418151855,
"logps/rejected": -50.61724090576172,
"loss": 0.1177,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.59584641456604,
"rewards/margins": 2.59584641456604,
"rewards/rejected": 0.0,
"step": 1123
},
{
"epoch": 6.279329608938547,
"grad_norm": 2.8236764416772027,
"learning_rate": 7.615220829723563e-07,
"logits/chosen": -3.635977268218994,
"logits/rejected": -3.314913749694824,
"logps/chosen": -2.3783607482910156,
"logps/rejected": -71.88471984863281,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5717906951904297,
"rewards/margins": 1.5717906951904297,
"rewards/rejected": 0.0,
"step": 1124
},
{
"epoch": 6.284916201117318,
"grad_norm": 5.067874331233562,
"learning_rate": 7.610023705389672e-07,
"logits/chosen": -3.6423583030700684,
"logits/rejected": -3.7084062099456787,
"logps/chosen": -2.9335761070251465,
"logps/rejected": -29.922504425048828,
"loss": 0.1198,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2978029251098633,
"rewards/margins": 2.2978029251098633,
"rewards/rejected": 0.0,
"step": 1125
},
{
"epoch": 6.290502793296089,
"grad_norm": 1.682862918959285,
"learning_rate": 7.604822702095733e-07,
"logits/chosen": -3.428931951522827,
"logits/rejected": -3.549999713897705,
"logps/chosen": -9.16401195526123,
"logps/rejected": -43.981082916259766,
"loss": 0.1285,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.355374813079834,
"rewards/margins": 2.355374813079834,
"rewards/rejected": 0.0,
"step": 1126
},
{
"epoch": 6.29608938547486,
"grad_norm": 1.5641957951116872,
"learning_rate": 7.599617827571366e-07,
"logits/chosen": -3.205566167831421,
"logits/rejected": -3.266538619995117,
"logps/chosen": -3.9005112648010254,
"logps/rejected": -35.63095474243164,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4671335220336914,
"rewards/margins": 2.4671335220336914,
"rewards/rejected": 0.0,
"step": 1127
},
{
"epoch": 6.301675977653631,
"grad_norm": 1.3903496328047276,
"learning_rate": 7.594409089551941e-07,
"logits/chosen": -3.446619987487793,
"logits/rejected": -3.578960657119751,
"logps/chosen": -3.6555418968200684,
"logps/rejected": -26.66632080078125,
"loss": 0.1282,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.413005828857422,
"rewards/margins": 2.413005828857422,
"rewards/rejected": 0.0,
"step": 1128
},
{
"epoch": 6.307262569832402,
"grad_norm": 4.070485445328956,
"learning_rate": 7.58919649577857e-07,
"logits/chosen": -3.253666639328003,
"logits/rejected": -3.2835707664489746,
"logps/chosen": -12.576066017150879,
"logps/rejected": -61.259986877441406,
"loss": 0.1205,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4137935638427734,
"rewards/margins": 2.4137935638427734,
"rewards/rejected": 0.0,
"step": 1129
},
{
"epoch": 6.312849162011173,
"grad_norm": 6.158340194259556,
"learning_rate": 7.583980053998094e-07,
"logits/chosen": -3.6586406230926514,
"logits/rejected": -3.458526372909546,
"logps/chosen": -9.13015365600586,
"logps/rejected": -44.24652099609375,
"loss": 0.119,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.887941360473633,
"rewards/margins": 2.887941360473633,
"rewards/rejected": 0.0,
"step": 1130
},
{
"epoch": 6.318435754189944,
"grad_norm": 3.9187148644017924,
"learning_rate": 7.578759771963077e-07,
"logits/chosen": -3.6340296268463135,
"logits/rejected": -3.5403378009796143,
"logps/chosen": -30.52096939086914,
"logps/rejected": -27.17243194580078,
"loss": 0.133,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3830909729003906,
"rewards/margins": 3.3830909729003906,
"rewards/rejected": 0.0,
"step": 1131
},
{
"epoch": 6.324022346368715,
"grad_norm": 2.388480448849047,
"learning_rate": 7.573535657431786e-07,
"logits/chosen": -3.5924248695373535,
"logits/rejected": -3.5247814655303955,
"logps/chosen": -8.00168514251709,
"logps/rejected": -55.73610305786133,
"loss": 0.124,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.651998519897461,
"rewards/margins": 2.651998519897461,
"rewards/rejected": 0.0,
"step": 1132
},
{
"epoch": 6.329608938547486,
"grad_norm": 2.735087380655357,
"learning_rate": 7.568307718168189e-07,
"logits/chosen": -3.3794960975646973,
"logits/rejected": -3.631844997406006,
"logps/chosen": -3.5662617683410645,
"logps/rejected": -48.561580657958984,
"loss": 0.1406,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1723389625549316,
"rewards/margins": 2.1723389625549316,
"rewards/rejected": 0.0,
"step": 1133
},
{
"epoch": 6.335195530726257,
"grad_norm": 3.0594035105757844,
"learning_rate": 7.563075961941929e-07,
"logits/chosen": -3.596266746520996,
"logits/rejected": -3.5486643314361572,
"logps/chosen": -12.418999671936035,
"logps/rejected": -45.26614761352539,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0499706268310547,
"rewards/margins": 3.0499706268310547,
"rewards/rejected": 0.0,
"step": 1134
},
{
"epoch": 6.340782122905028,
"grad_norm": 4.442309247909587,
"learning_rate": 7.557840396528333e-07,
"logits/chosen": -3.2268638610839844,
"logits/rejected": -3.0698015689849854,
"logps/chosen": -6.380509376525879,
"logps/rejected": -24.476642608642578,
"loss": 0.0994,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8292136192321777,
"rewards/margins": 2.8292136192321777,
"rewards/rejected": 0.0,
"step": 1135
},
{
"epoch": 6.346368715083799,
"grad_norm": 6.983921678944319,
"learning_rate": 7.55260102970838e-07,
"logits/chosen": -3.9538321495056152,
"logits/rejected": -3.7950656414031982,
"logps/chosen": -12.06083869934082,
"logps/rejected": -28.718433380126953,
"loss": 0.1032,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0041298866271973,
"rewards/margins": 3.0041298866271973,
"rewards/rejected": 0.0,
"step": 1136
},
{
"epoch": 6.351955307262569,
"grad_norm": 1.6742125231229434,
"learning_rate": 7.547357869268704e-07,
"logits/chosen": -3.7847020626068115,
"logits/rejected": -3.9054253101348877,
"logps/chosen": -3.6618285179138184,
"logps/rejected": -38.363895416259766,
"loss": 0.1174,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2039732933044434,
"rewards/margins": 2.2039732933044434,
"rewards/rejected": 0.0,
"step": 1137
},
{
"epoch": 6.35754189944134,
"grad_norm": 3.190302920370684,
"learning_rate": 7.542110923001576e-07,
"logits/chosen": -3.446341037750244,
"logits/rejected": -3.7196896076202393,
"logps/chosen": -5.214937686920166,
"logps/rejected": -48.24104309082031,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1412739753723145,
"rewards/margins": 2.1412739753723145,
"rewards/rejected": 0.0,
"step": 1138
},
{
"epoch": 6.363128491620111,
"grad_norm": 5.672963080835398,
"learning_rate": 7.53686019870489e-07,
"logits/chosen": -3.458178997039795,
"logits/rejected": -3.370027780532837,
"logps/chosen": -13.658634185791016,
"logps/rejected": -23.169994354248047,
"loss": 0.1501,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8471968173980713,
"rewards/margins": 2.8471968173980713,
"rewards/rejected": 0.0,
"step": 1139
},
{
"epoch": 6.368715083798882,
"grad_norm": 3.58288083102813,
"learning_rate": 7.531605704182158e-07,
"logits/chosen": -3.4480130672454834,
"logits/rejected": -3.4778430461883545,
"logps/chosen": -1.6349246501922607,
"logps/rejected": -40.36206817626953,
"loss": 0.1352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9295605421066284,
"rewards/margins": 1.9295605421066284,
"rewards/rejected": 0.0,
"step": 1140
},
{
"epoch": 6.368715083798882,
"eval_logits/chosen": -3.388373613357544,
"eval_logits/rejected": -3.5139546394348145,
"eval_logps/chosen": -23.564157485961914,
"eval_logps/rejected": -42.2648811340332,
"eval_loss": 0.5904839634895325,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/chosen": 0.6505883932113647,
"eval_rewards/margins": 0.6505883932113647,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.737,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 1140
},
{
"epoch": 6.374301675977653,
"grad_norm": 2.07573234783886,
"learning_rate": 7.526347447242496e-07,
"logits/chosen": -3.5074241161346436,
"logits/rejected": -3.385617733001709,
"logps/chosen": -21.858230590820312,
"logps/rejected": -38.5920524597168,
"loss": 0.1189,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0649476051330566,
"rewards/margins": 3.0649476051330566,
"rewards/rejected": 0.0,
"step": 1141
},
{
"epoch": 6.379888268156424,
"grad_norm": 1.9055519576325777,
"learning_rate": 7.521085435700609e-07,
"logits/chosen": -3.439046859741211,
"logits/rejected": -3.6108994483947754,
"logps/chosen": -4.852899551391602,
"logps/rejected": -29.23027992248535,
"loss": 0.1119,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1748762130737305,
"rewards/margins": 2.1748762130737305,
"rewards/rejected": 0.0,
"step": 1142
},
{
"epoch": 6.385474860335195,
"grad_norm": 7.2018815283622315,
"learning_rate": 7.515819677376786e-07,
"logits/chosen": -3.479985237121582,
"logits/rejected": -3.486938714981079,
"logps/chosen": -7.141788482666016,
"logps/rejected": -39.78456115722656,
"loss": 0.142,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5886242389678955,
"rewards/margins": 2.5886242389678955,
"rewards/rejected": 0.0,
"step": 1143
},
{
"epoch": 6.391061452513966,
"grad_norm": 4.999108636137947,
"learning_rate": 7.510550180096876e-07,
"logits/chosen": -3.5652968883514404,
"logits/rejected": -3.490980625152588,
"logps/chosen": -10.460119247436523,
"logps/rejected": -16.658226013183594,
"loss": 0.128,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4940853118896484,
"rewards/margins": 2.4940853118896484,
"rewards/rejected": 0.0,
"step": 1144
},
{
"epoch": 6.396648044692737,
"grad_norm": 2.5838107318157224,
"learning_rate": 7.505276951692296e-07,
"logits/chosen": -3.634608030319214,
"logits/rejected": -3.643927812576294,
"logps/chosen": -4.419398307800293,
"logps/rejected": -59.13410949707031,
"loss": 0.098,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5251169204711914,
"rewards/margins": 2.5251169204711914,
"rewards/rejected": 0.0,
"step": 1145
},
{
"epoch": 6.402234636871508,
"grad_norm": 3.3446696073869164,
"learning_rate": 7.5e-07,
"logits/chosen": -3.5030786991119385,
"logits/rejected": -3.564039468765259,
"logps/chosen": -9.80803108215332,
"logps/rejected": -40.293479919433594,
"loss": 0.1074,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9979138374328613,
"rewards/margins": 2.9979138374328613,
"rewards/rejected": 0.0,
"step": 1146
},
{
"epoch": 6.407821229050279,
"grad_norm": 1.9616353941615796,
"learning_rate": 7.494719332862478e-07,
"logits/chosen": -3.623610258102417,
"logits/rejected": -3.1363115310668945,
"logps/chosen": -1.4352056980133057,
"logps/rejected": -26.92434310913086,
"loss": 0.1204,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8442667722702026,
"rewards/margins": 1.8442667722702026,
"rewards/rejected": 0.0,
"step": 1147
},
{
"epoch": 6.41340782122905,
"grad_norm": 3.3440966999791693,
"learning_rate": 7.489434958127742e-07,
"logits/chosen": -3.5022785663604736,
"logits/rejected": -3.3467249870300293,
"logps/chosen": -13.047353744506836,
"logps/rejected": -58.06136703491211,
"loss": 0.1256,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.292217969894409,
"rewards/margins": 2.292217969894409,
"rewards/rejected": 0.0,
"step": 1148
},
{
"epoch": 6.418994413407821,
"grad_norm": 9.914571903712174,
"learning_rate": 7.484146883649315e-07,
"logits/chosen": -3.668579339981079,
"logits/rejected": -3.719177722930908,
"logps/chosen": -3.6147677898406982,
"logps/rejected": -43.331233978271484,
"loss": 0.1395,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.225457191467285,
"rewards/margins": 2.225457191467285,
"rewards/rejected": 0.0,
"step": 1149
},
{
"epoch": 6.424581005586592,
"grad_norm": 1.8114041552518896,
"learning_rate": 7.478855117286216e-07,
"logits/chosen": -3.4629340171813965,
"logits/rejected": -3.67208194732666,
"logps/chosen": -22.60823631286621,
"logps/rejected": -29.602725982666016,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.589573860168457,
"rewards/margins": 2.589573860168457,
"rewards/rejected": 0.0,
"step": 1150
},
{
"epoch": 6.430167597765363,
"grad_norm": 8.34282059434499,
"learning_rate": 7.473559666902953e-07,
"logits/chosen": -3.734013557434082,
"logits/rejected": -3.6599209308624268,
"logps/chosen": -1.3387436866760254,
"logps/rejected": -31.555988311767578,
"loss": 0.1494,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5991384983062744,
"rewards/margins": 1.5991384983062744,
"rewards/rejected": 0.0,
"step": 1151
},
{
"epoch": 6.435754189944134,
"grad_norm": 2.1156571176367254,
"learning_rate": 7.468260540369509e-07,
"logits/chosen": -3.589686632156372,
"logits/rejected": -3.5712993144989014,
"logps/chosen": -9.595553398132324,
"logps/rejected": -41.65563201904297,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.34926700592041,
"rewards/margins": 2.34926700592041,
"rewards/rejected": 0.0,
"step": 1152
},
{
"epoch": 6.441340782122905,
"grad_norm": 2.068700473887163,
"learning_rate": 7.462957745561327e-07,
"logits/chosen": -3.513237953186035,
"logits/rejected": -3.693328619003296,
"logps/chosen": -14.855696678161621,
"logps/rejected": -40.235191345214844,
"loss": 0.1264,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3135600090026855,
"rewards/margins": 3.3135600090026855,
"rewards/rejected": 0.0,
"step": 1153
},
{
"epoch": 6.446927374301676,
"grad_norm": 2.8361763613552813,
"learning_rate": 7.457651290359306e-07,
"logits/chosen": -3.4706835746765137,
"logits/rejected": -3.6301958560943604,
"logps/chosen": -31.08868408203125,
"logps/rejected": -78.93798065185547,
"loss": 0.1326,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.295527935028076,
"rewards/margins": 2.295527935028076,
"rewards/rejected": 0.0,
"step": 1154
},
{
"epoch": 6.452513966480447,
"grad_norm": 2.782776862415484,
"learning_rate": 7.452341182649787e-07,
"logits/chosen": -3.7000601291656494,
"logits/rejected": -3.6847100257873535,
"logps/chosen": -8.851727485656738,
"logps/rejected": -35.246849060058594,
"loss": 0.1126,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5256786346435547,
"rewards/margins": 2.5256786346435547,
"rewards/rejected": 0.0,
"step": 1155
},
{
"epoch": 6.4581005586592175,
"grad_norm": 5.337794330407838,
"learning_rate": 7.447027430324529e-07,
"logits/chosen": -3.444211006164551,
"logits/rejected": -3.5518124103546143,
"logps/chosen": -10.179937362670898,
"logps/rejected": -24.692378997802734,
"loss": 0.1521,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.374847888946533,
"rewards/margins": 2.374847888946533,
"rewards/rejected": 0.0,
"step": 1156
},
{
"epoch": 6.4636871508379885,
"grad_norm": 2.392044190739456,
"learning_rate": 7.441710041280719e-07,
"logits/chosen": -3.629194498062134,
"logits/rejected": -3.6020259857177734,
"logps/chosen": -1.9871127605438232,
"logps/rejected": -81.12548828125,
"loss": 0.1302,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.277148723602295,
"rewards/margins": 2.277148723602295,
"rewards/rejected": 0.0,
"step": 1157
},
{
"epoch": 6.4692737430167595,
"grad_norm": 1.7949174942430215,
"learning_rate": 7.436389023420942e-07,
"logits/chosen": -3.7679221630096436,
"logits/rejected": -3.652407646179199,
"logps/chosen": -5.68017578125,
"logps/rejected": -47.26246643066406,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.207443952560425,
"rewards/margins": 2.207443952560425,
"rewards/rejected": 0.0,
"step": 1158
},
{
"epoch": 6.4748603351955305,
"grad_norm": 2.546044260638092,
"learning_rate": 7.431064384653181e-07,
"logits/chosen": -3.419480085372925,
"logits/rejected": -3.672830104827881,
"logps/chosen": -3.2580134868621826,
"logps/rejected": -58.788848876953125,
"loss": 0.1396,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.158781051635742,
"rewards/margins": 2.158781051635742,
"rewards/rejected": 0.0,
"step": 1159
},
{
"epoch": 6.4804469273743015,
"grad_norm": 3.3398948644505384,
"learning_rate": 7.425736132890794e-07,
"logits/chosen": -3.4635164737701416,
"logits/rejected": -3.459397315979004,
"logps/chosen": -15.559146881103516,
"logps/rejected": -37.985374450683594,
"loss": 0.1121,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4702444076538086,
"rewards/margins": 2.4702444076538086,
"rewards/rejected": 0.0,
"step": 1160
},
{
"epoch": 6.4804469273743015,
"eval_logits/chosen": -3.3824734687805176,
"eval_logits/rejected": -3.5105621814727783,
"eval_logps/chosen": -22.876380920410156,
"eval_logps/rejected": -42.273921966552734,
"eval_loss": 0.5851587057113647,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/chosen": 0.7193660736083984,
"eval_rewards/margins": 0.7193660736083984,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6808,
"eval_samples_per_second": 9.486,
"eval_steps_per_second": 0.306,
"step": 1160
},
{
"epoch": 6.4860335195530725,
"grad_norm": 1.61387712350269,
"learning_rate": 7.420404276052516e-07,
"logits/chosen": -3.4014992713928223,
"logits/rejected": -3.5311710834503174,
"logps/chosen": -4.388230800628662,
"logps/rejected": -46.40802001953125,
"loss": 0.1188,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1586849689483643,
"rewards/margins": 2.1586849689483643,
"rewards/rejected": 0.0,
"step": 1161
},
{
"epoch": 6.4916201117318435,
"grad_norm": 1.800848057129923,
"learning_rate": 7.415068822062431e-07,
"logits/chosen": -3.586928367614746,
"logits/rejected": -3.7211384773254395,
"logps/chosen": -3.3956210613250732,
"logps/rejected": -46.08708190917969,
"loss": 0.1189,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.251286029815674,
"rewards/margins": 2.251286029815674,
"rewards/rejected": 0.0,
"step": 1162
},
{
"epoch": 6.4972067039106145,
"grad_norm": 4.4067031135950145,
"learning_rate": 7.409729778849976e-07,
"logits/chosen": -3.4703292846679688,
"logits/rejected": -3.460296869277954,
"logps/chosen": -5.283385276794434,
"logps/rejected": -55.7769889831543,
"loss": 0.1091,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5026423931121826,
"rewards/margins": 2.5026423931121826,
"rewards/rejected": 0.0,
"step": 1163
},
{
"epoch": 6.5027932960893855,
"grad_norm": 7.293108498783106,
"learning_rate": 7.40438715434992e-07,
"logits/chosen": -3.7694311141967773,
"logits/rejected": -3.788268804550171,
"logps/chosen": -1.3201477527618408,
"logps/rejected": -35.48125076293945,
"loss": 0.1356,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9290229082107544,
"rewards/margins": 1.9290229082107544,
"rewards/rejected": 0.0,
"step": 1164
},
{
"epoch": 6.5083798882681565,
"grad_norm": 1.9437543708834484,
"learning_rate": 7.399040956502357e-07,
"logits/chosen": -3.42350697517395,
"logits/rejected": -3.4623332023620605,
"logps/chosen": -6.216604709625244,
"logps/rejected": -44.93048095703125,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.248582601547241,
"rewards/margins": 2.248582601547241,
"rewards/rejected": 0.0,
"step": 1165
},
{
"epoch": 6.5139664804469275,
"grad_norm": 4.817821483946756,
"learning_rate": 7.393691193252685e-07,
"logits/chosen": -3.5924384593963623,
"logits/rejected": -3.4373843669891357,
"logps/chosen": -12.131484985351562,
"logps/rejected": -19.567914962768555,
"loss": 0.1526,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4102120399475098,
"rewards/margins": 2.4102120399475098,
"rewards/rejected": 0.0,
"step": 1166
},
{
"epoch": 6.5195530726256985,
"grad_norm": 4.3216972734564605,
"learning_rate": 7.388337872551606e-07,
"logits/chosen": -3.5208194255828857,
"logits/rejected": -3.3143310546875,
"logps/chosen": -31.798171997070312,
"logps/rejected": -21.487539291381836,
"loss": 0.1089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6758854389190674,
"rewards/margins": 2.6758854389190674,
"rewards/rejected": 0.0,
"step": 1167
},
{
"epoch": 6.5251396648044695,
"grad_norm": 2.638264524988585,
"learning_rate": 7.382981002355111e-07,
"logits/chosen": -3.5087356567382812,
"logits/rejected": -3.5776567459106445,
"logps/chosen": -25.8616943359375,
"logps/rejected": -48.079994201660156,
"loss": 0.1161,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.036346673965454,
"rewards/margins": 3.036346673965454,
"rewards/rejected": 0.0,
"step": 1168
},
{
"epoch": 6.5307262569832405,
"grad_norm": 1.879004053663191,
"learning_rate": 7.377620590624461e-07,
"logits/chosen": -3.4802727699279785,
"logits/rejected": -3.360588788986206,
"logps/chosen": -15.909881591796875,
"logps/rejected": -30.90283203125,
"loss": 0.1076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.166745662689209,
"rewards/margins": 3.166745662689209,
"rewards/rejected": 0.0,
"step": 1169
},
{
"epoch": 6.5363128491620115,
"grad_norm": 5.624151661105335,
"learning_rate": 7.37225664532618e-07,
"logits/chosen": -3.6656854152679443,
"logits/rejected": -3.3708481788635254,
"logps/chosen": -5.943014144897461,
"logps/rejected": -62.0150260925293,
"loss": 0.1463,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.5379159450531006,
"rewards/margins": 1.5379159450531006,
"rewards/rejected": 0.0,
"step": 1170
},
{
"epoch": 6.5418994413407825,
"grad_norm": 1.4296245046803395,
"learning_rate": 7.366889174432049e-07,
"logits/chosen": -3.0365090370178223,
"logits/rejected": -3.0734381675720215,
"logps/chosen": -15.989898681640625,
"logps/rejected": -30.509754180908203,
"loss": 0.0897,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0527782440185547,
"rewards/margins": 3.0527782440185547,
"rewards/rejected": 0.0,
"step": 1171
},
{
"epoch": 6.547486033519553,
"grad_norm": 3.3616999767523104,
"learning_rate": 7.361518185919086e-07,
"logits/chosen": -3.6535892486572266,
"logits/rejected": -3.5110924243927,
"logps/chosen": -5.396595001220703,
"logps/rejected": -33.7409782409668,
"loss": 0.1239,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0897598266601562,
"rewards/margins": 2.0897598266601562,
"rewards/rejected": 0.0,
"step": 1172
},
{
"epoch": 6.553072625698324,
"grad_norm": 2.334322615680733,
"learning_rate": 7.356143687769534e-07,
"logits/chosen": -3.3524158000946045,
"logits/rejected": -3.2852206230163574,
"logps/chosen": -18.785200119018555,
"logps/rejected": -63.185218811035156,
"loss": 0.1047,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.363424301147461,
"rewards/margins": 2.363424301147461,
"rewards/rejected": 0.0,
"step": 1173
},
{
"epoch": 6.558659217877095,
"grad_norm": 6.407386995876666,
"learning_rate": 7.350765687970856e-07,
"logits/chosen": -3.524543523788452,
"logits/rejected": -3.4610118865966797,
"logps/chosen": -1.330913782119751,
"logps/rejected": -51.27202606201172,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0210142135620117,
"rewards/margins": 2.0210142135620117,
"rewards/rejected": 0.0,
"step": 1174
},
{
"epoch": 6.564245810055866,
"grad_norm": 5.064980797784388,
"learning_rate": 7.345384194515718e-07,
"logits/chosen": -3.58046293258667,
"logits/rejected": -3.786489248275757,
"logps/chosen": -4.507716178894043,
"logps/rejected": -48.74477005004883,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.252613067626953,
"rewards/margins": 2.252613067626953,
"rewards/rejected": 0.0,
"step": 1175
},
{
"epoch": 6.569832402234637,
"grad_norm": 1.8492597615441708,
"learning_rate": 7.339999215401975e-07,
"logits/chosen": -3.372981548309326,
"logits/rejected": -3.4837727546691895,
"logps/chosen": -6.701407432556152,
"logps/rejected": -37.21700668334961,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7453620433807373,
"rewards/margins": 2.7453620433807373,
"rewards/rejected": 0.0,
"step": 1176
},
{
"epoch": 6.575418994413408,
"grad_norm": 6.3422045402981855,
"learning_rate": 7.334610758632668e-07,
"logits/chosen": -3.6905386447906494,
"logits/rejected": -3.767056703567505,
"logps/chosen": -31.702373504638672,
"logps/rejected": -46.73040771484375,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9266233444213867,
"rewards/margins": 2.9266233444213867,
"rewards/rejected": 0.0,
"step": 1177
},
{
"epoch": 6.581005586592179,
"grad_norm": 3.1790101358671246,
"learning_rate": 7.329218832216003e-07,
"logits/chosen": -3.395127773284912,
"logits/rejected": -3.559117555618286,
"logps/chosen": -5.663971900939941,
"logps/rejected": -48.885009765625,
"loss": 0.118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.305607318878174,
"rewards/margins": 2.305607318878174,
"rewards/rejected": 0.0,
"step": 1178
},
{
"epoch": 6.58659217877095,
"grad_norm": 7.017385742847004,
"learning_rate": 7.32382344416534e-07,
"logits/chosen": -3.26444411277771,
"logits/rejected": -3.290856122970581,
"logps/chosen": -6.893622875213623,
"logps/rejected": -60.45686721801758,
"loss": 0.1248,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7198753356933594,
"rewards/margins": 2.7198753356933594,
"rewards/rejected": 0.0,
"step": 1179
},
{
"epoch": 6.592178770949721,
"grad_norm": 3.2928264012315513,
"learning_rate": 7.318424602499192e-07,
"logits/chosen": -3.4912145137786865,
"logits/rejected": -3.486445426940918,
"logps/chosen": -8.337770462036133,
"logps/rejected": -35.21287155151367,
"loss": 0.1095,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.795140504837036,
"rewards/margins": 2.795140504837036,
"rewards/rejected": 0.0,
"step": 1180
},
{
"epoch": 6.592178770949721,
"eval_logits/chosen": -3.3981025218963623,
"eval_logits/rejected": -3.5221188068389893,
"eval_logps/chosen": -23.001008987426758,
"eval_logps/rejected": -42.521881103515625,
"eval_loss": 0.5904117822647095,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.706903338432312,
"eval_rewards/margins": 0.706903338432312,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7111,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 1180
},
{
"epoch": 6.597765363128492,
"grad_norm": 2.2059531916875192,
"learning_rate": 7.313022315241194e-07,
"logits/chosen": -3.7348954677581787,
"logits/rejected": -3.775949239730835,
"logps/chosen": -16.370988845825195,
"logps/rejected": -45.253597259521484,
"loss": 0.0886,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4794254302978516,
"rewards/margins": 2.4794254302978516,
"rewards/rejected": 0.0,
"step": 1181
},
{
"epoch": 6.603351955307263,
"grad_norm": 1.3834486669385506,
"learning_rate": 7.307616590420109e-07,
"logits/chosen": -3.329288959503174,
"logits/rejected": -3.528491497039795,
"logps/chosen": -5.240683555603027,
"logps/rejected": -26.84988784790039,
"loss": 0.1112,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5287373065948486,
"rewards/margins": 2.5287373065948486,
"rewards/rejected": 0.0,
"step": 1182
},
{
"epoch": 6.608938547486034,
"grad_norm": 4.938881978910298,
"learning_rate": 7.302207436069806e-07,
"logits/chosen": -3.7750566005706787,
"logits/rejected": -3.910111427307129,
"logps/chosen": -3.403721332550049,
"logps/rejected": -88.3388442993164,
"loss": 0.1272,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8856760263442993,
"rewards/margins": 1.8856760263442993,
"rewards/rejected": 0.0,
"step": 1183
},
{
"epoch": 6.614525139664805,
"grad_norm": 1.7965923799071406,
"learning_rate": 7.296794860229253e-07,
"logits/chosen": -3.5343782901763916,
"logits/rejected": -3.388319969177246,
"logps/chosen": -5.0793962478637695,
"logps/rejected": -32.3995361328125,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3262863159179688,
"rewards/margins": 2.3262863159179688,
"rewards/rejected": 0.0,
"step": 1184
},
{
"epoch": 6.620111731843576,
"grad_norm": 1.5554842365439236,
"learning_rate": 7.2913788709425e-07,
"logits/chosen": -3.6333446502685547,
"logits/rejected": -3.594719409942627,
"logps/chosen": -5.939208030700684,
"logps/rejected": -34.853302001953125,
"loss": 0.1046,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.663480281829834,
"rewards/margins": 2.663480281829834,
"rewards/rejected": 0.0,
"step": 1185
},
{
"epoch": 6.625698324022347,
"grad_norm": 1.7944320372729725,
"learning_rate": 7.285959476258673e-07,
"logits/chosen": -3.796753406524658,
"logits/rejected": -3.6380088329315186,
"logps/chosen": -12.204524993896484,
"logps/rejected": -24.707441329956055,
"loss": 0.0904,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.130892276763916,
"rewards/margins": 3.130892276763916,
"rewards/rejected": 0.0,
"step": 1186
},
{
"epoch": 6.631284916201118,
"grad_norm": 1.8467357194452632,
"learning_rate": 7.280536684231957e-07,
"logits/chosen": -3.501314163208008,
"logits/rejected": -3.2968432903289795,
"logps/chosen": -3.581840753555298,
"logps/rejected": -25.83201789855957,
"loss": 0.1271,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.264824628829956,
"rewards/margins": 2.264824628829956,
"rewards/rejected": 0.0,
"step": 1187
},
{
"epoch": 6.636871508379889,
"grad_norm": 3.1295977153882824,
"learning_rate": 7.275110502921588e-07,
"logits/chosen": -3.4816713333129883,
"logits/rejected": -3.6248412132263184,
"logps/chosen": -1.5265557765960693,
"logps/rejected": -31.821849822998047,
"loss": 0.1009,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9344596862792969,
"rewards/margins": 1.9344596862792969,
"rewards/rejected": 0.0,
"step": 1188
},
{
"epoch": 6.64245810055866,
"grad_norm": 2.7051695361680923,
"learning_rate": 7.269680940391835e-07,
"logits/chosen": -3.3442203998565674,
"logits/rejected": -3.4781646728515625,
"logps/chosen": -9.685230255126953,
"logps/rejected": -36.17442321777344,
"loss": 0.1061,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.676804304122925,
"rewards/margins": 2.676804304122925,
"rewards/rejected": 0.0,
"step": 1189
},
{
"epoch": 6.648044692737431,
"grad_norm": 6.929736209226811,
"learning_rate": 7.264248004711997e-07,
"logits/chosen": -3.505671501159668,
"logits/rejected": -3.411191701889038,
"logps/chosen": -8.421308517456055,
"logps/rejected": -42.89100646972656,
"loss": 0.1181,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6684720516204834,
"rewards/margins": 2.6684720516204834,
"rewards/rejected": 0.0,
"step": 1190
},
{
"epoch": 6.653631284916202,
"grad_norm": 3.004093237871956,
"learning_rate": 7.258811703956385e-07,
"logits/chosen": -3.009087324142456,
"logits/rejected": -2.9700117111206055,
"logps/chosen": -8.3806791305542,
"logps/rejected": -40.365577697753906,
"loss": 0.0992,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.768009901046753,
"rewards/margins": 2.768009901046753,
"rewards/rejected": 0.0,
"step": 1191
},
{
"epoch": 6.659217877094972,
"grad_norm": 1.8564056250718206,
"learning_rate": 7.253372046204307e-07,
"logits/chosen": -3.6647820472717285,
"logits/rejected": -3.746948003768921,
"logps/chosen": -1.0470430850982666,
"logps/rejected": -22.98351287841797,
"loss": 0.1428,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2635326385498047,
"rewards/margins": 1.2635326385498047,
"rewards/rejected": 0.0,
"step": 1192
},
{
"epoch": 6.664804469273743,
"grad_norm": 2.596685756261615,
"learning_rate": 7.247929039540064e-07,
"logits/chosen": -3.5840508937835693,
"logits/rejected": -3.5116047859191895,
"logps/chosen": -10.925397872924805,
"logps/rejected": -37.749664306640625,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.496830940246582,
"rewards/margins": 2.496830940246582,
"rewards/rejected": 0.0,
"step": 1193
},
{
"epoch": 6.670391061452514,
"grad_norm": 1.7800121822891553,
"learning_rate": 7.242482692052935e-07,
"logits/chosen": -3.11643385887146,
"logits/rejected": -3.096829652786255,
"logps/chosen": -9.554300308227539,
"logps/rejected": -43.78571319580078,
"loss": 0.0859,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.657076597213745,
"rewards/margins": 2.657076597213745,
"rewards/rejected": 0.0,
"step": 1194
},
{
"epoch": 6.675977653631285,
"grad_norm": 1.8899008327500304,
"learning_rate": 7.237033011837161e-07,
"logits/chosen": -3.393386125564575,
"logits/rejected": -3.544006586074829,
"logps/chosen": -5.600795745849609,
"logps/rejected": -52.57780456542969,
"loss": 0.1486,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2170066833496094,
"rewards/margins": 2.2170066833496094,
"rewards/rejected": 0.0,
"step": 1195
},
{
"epoch": 6.681564245810056,
"grad_norm": 4.892910432832869,
"learning_rate": 7.23158000699194e-07,
"logits/chosen": -3.600586175918579,
"logits/rejected": -3.6606061458587646,
"logps/chosen": -28.68573760986328,
"logps/rejected": -27.99110221862793,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1784684658050537,
"rewards/margins": 3.1784684658050537,
"rewards/rejected": 0.0,
"step": 1196
},
{
"epoch": 6.687150837988827,
"grad_norm": 4.778481457666094,
"learning_rate": 7.226123685621403e-07,
"logits/chosen": -3.7846198081970215,
"logits/rejected": -3.756338119506836,
"logps/chosen": -1.9063308238983154,
"logps/rejected": -26.37069320678711,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7744323015213013,
"rewards/margins": 1.7744323015213013,
"rewards/rejected": 0.0,
"step": 1197
},
{
"epoch": 6.692737430167598,
"grad_norm": 6.191543108954036,
"learning_rate": 7.22066405583462e-07,
"logits/chosen": -3.42878794670105,
"logits/rejected": -3.6990392208099365,
"logps/chosen": -14.416027069091797,
"logps/rejected": -42.47001266479492,
"loss": 0.1245,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.790872812271118,
"rewards/margins": 2.790872812271118,
"rewards/rejected": 0.0,
"step": 1198
},
{
"epoch": 6.698324022346369,
"grad_norm": 3.200562659039872,
"learning_rate": 7.21520112574557e-07,
"logits/chosen": -3.419428586959839,
"logits/rejected": -3.5295228958129883,
"logps/chosen": -1.1163108348846436,
"logps/rejected": -48.84260177612305,
"loss": 0.1344,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7757481336593628,
"rewards/margins": 1.7757481336593628,
"rewards/rejected": 0.0,
"step": 1199
},
{
"epoch": 6.70391061452514,
"grad_norm": 3.4983642115084916,
"learning_rate": 7.209734903473141e-07,
"logits/chosen": -3.2450597286224365,
"logits/rejected": -3.299987554550171,
"logps/chosen": -2.303595781326294,
"logps/rejected": -40.97272872924805,
"loss": 0.1029,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.18654203414917,
"rewards/margins": 2.18654203414917,
"rewards/rejected": 0.0,
"step": 1200
},
{
"epoch": 6.70391061452514,
"eval_logits/chosen": -3.398895740509033,
"eval_logits/rejected": -3.5260047912597656,
"eval_logps/chosen": -23.452024459838867,
"eval_logps/rejected": -43.03379821777344,
"eval_loss": 0.617557942867279,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/chosen": 0.6618017554283142,
"eval_rewards/margins": 0.6618017554283142,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7097,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 1200
},
{
"epoch": 6.709497206703911,
"grad_norm": 2.1236245304175703,
"learning_rate": 7.204265397141115e-07,
"logits/chosen": -3.3595433235168457,
"logits/rejected": -3.3600683212280273,
"logps/chosen": -3.071162223815918,
"logps/rejected": -47.003204345703125,
"loss": 0.1352,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.686194896697998,
"rewards/margins": 1.686194896697998,
"rewards/rejected": 0.0,
"step": 1201
},
{
"epoch": 6.715083798882682,
"grad_norm": 3.0521846397375367,
"learning_rate": 7.19879261487815e-07,
"logits/chosen": -3.675001859664917,
"logits/rejected": -3.700336456298828,
"logps/chosen": -5.996543884277344,
"logps/rejected": -38.44892883300781,
"loss": 0.1201,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.87506103515625,
"rewards/margins": 2.87506103515625,
"rewards/rejected": 0.0,
"step": 1202
},
{
"epoch": 6.720670391061453,
"grad_norm": 2.3043975156469174,
"learning_rate": 7.193316564817775e-07,
"logits/chosen": -3.619896173477173,
"logits/rejected": -3.757736921310425,
"logps/chosen": -4.08686637878418,
"logps/rejected": -42.31098937988281,
"loss": 0.1036,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.071007013320923,
"rewards/margins": 2.071007013320923,
"rewards/rejected": 0.0,
"step": 1203
},
{
"epoch": 6.726256983240224,
"grad_norm": 9.080938342219454,
"learning_rate": 7.187837255098378e-07,
"logits/chosen": -3.453091859817505,
"logits/rejected": -3.3350491523742676,
"logps/chosen": -11.954121589660645,
"logps/rejected": -69.71781921386719,
"loss": 0.1533,
"rewards/accuracies": 0.75,
"rewards/chosen": 1.75028395652771,
"rewards/margins": 1.75028395652771,
"rewards/rejected": 0.0,
"step": 1204
},
{
"epoch": 6.731843575418995,
"grad_norm": 6.330344122312587,
"learning_rate": 7.182354693863185e-07,
"logits/chosen": -3.3193860054016113,
"logits/rejected": -3.7034122943878174,
"logps/chosen": -1.3383597135543823,
"logps/rejected": -49.510345458984375,
"loss": 0.159,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6445035934448242,
"rewards/margins": 1.6445035934448242,
"rewards/rejected": 0.0,
"step": 1205
},
{
"epoch": 6.737430167597766,
"grad_norm": 4.194445576780564,
"learning_rate": 7.176868889260264e-07,
"logits/chosen": -3.772306442260742,
"logits/rejected": -3.756890058517456,
"logps/chosen": -2.6748602390289307,
"logps/rejected": -42.153343200683594,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1727726459503174,
"rewards/margins": 2.1727726459503174,
"rewards/rejected": 0.0,
"step": 1206
},
{
"epoch": 6.743016759776537,
"grad_norm": 2.1329505604911145,
"learning_rate": 7.171379849442492e-07,
"logits/chosen": -3.4764533042907715,
"logits/rejected": -3.428299903869629,
"logps/chosen": -26.964725494384766,
"logps/rejected": -50.2739143371582,
"loss": 0.1237,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7658278942108154,
"rewards/margins": 2.7658278942108154,
"rewards/rejected": 0.0,
"step": 1207
},
{
"epoch": 6.748603351955307,
"grad_norm": 2.042392428716392,
"learning_rate": 7.165887582567562e-07,
"logits/chosen": -3.3654351234436035,
"logits/rejected": -3.2815451622009277,
"logps/chosen": -10.788187026977539,
"logps/rejected": -42.93102264404297,
"loss": 0.1271,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.399768829345703,
"rewards/margins": 2.399768829345703,
"rewards/rejected": 0.0,
"step": 1208
},
{
"epoch": 6.754189944134078,
"grad_norm": 1.8204590297465313,
"learning_rate": 7.160392096797962e-07,
"logits/chosen": -3.4422526359558105,
"logits/rejected": -3.43233323097229,
"logps/chosen": -6.396111011505127,
"logps/rejected": -38.756385803222656,
"loss": 0.1214,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2732014656066895,
"rewards/margins": 2.2732014656066895,
"rewards/rejected": 0.0,
"step": 1209
},
{
"epoch": 6.759776536312849,
"grad_norm": 1.9915316249999613,
"learning_rate": 7.154893400300961e-07,
"logits/chosen": -3.3545596599578857,
"logits/rejected": -3.3272500038146973,
"logps/chosen": -15.552148818969727,
"logps/rejected": -38.22154235839844,
"loss": 0.1129,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.815664052963257,
"rewards/margins": 3.815664052963257,
"rewards/rejected": 0.0,
"step": 1210
},
{
"epoch": 6.76536312849162,
"grad_norm": 2.1818868474595057,
"learning_rate": 7.1493915012486e-07,
"logits/chosen": -3.586728572845459,
"logits/rejected": -3.6481289863586426,
"logps/chosen": -5.913375377655029,
"logps/rejected": -59.924766540527344,
"loss": 0.1095,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5586137771606445,
"rewards/margins": 2.5586137771606445,
"rewards/rejected": 0.0,
"step": 1211
},
{
"epoch": 6.770949720670391,
"grad_norm": 6.423805073766037,
"learning_rate": 7.143886407817685e-07,
"logits/chosen": -3.373922348022461,
"logits/rejected": -3.477799415588379,
"logps/chosen": -2.9880781173706055,
"logps/rejected": -27.999588012695312,
"loss": 0.1459,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9803999662399292,
"rewards/margins": 1.9803999662399292,
"rewards/rejected": 0.0,
"step": 1212
},
{
"epoch": 6.776536312849162,
"grad_norm": 6.5904341124871975,
"learning_rate": 7.138378128189762e-07,
"logits/chosen": -3.2786753177642822,
"logits/rejected": -3.3929131031036377,
"logps/chosen": -0.9103558659553528,
"logps/rejected": -32.21274185180664,
"loss": 0.1478,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2840886116027832,
"rewards/margins": 1.2840886116027832,
"rewards/rejected": 0.0,
"step": 1213
},
{
"epoch": 6.782122905027933,
"grad_norm": 2.7972847424818963,
"learning_rate": 7.132866670551116e-07,
"logits/chosen": -3.7327377796173096,
"logits/rejected": -3.74499249458313,
"logps/chosen": -13.38409423828125,
"logps/rejected": -38.56711196899414,
"loss": 0.1038,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2696335315704346,
"rewards/margins": 3.2696335315704346,
"rewards/rejected": 0.0,
"step": 1214
},
{
"epoch": 6.787709497206704,
"grad_norm": 1.3959306891227423,
"learning_rate": 7.127352043092754e-07,
"logits/chosen": -3.496448278427124,
"logits/rejected": -3.335744619369507,
"logps/chosen": -6.675405502319336,
"logps/rejected": -44.683937072753906,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6381330490112305,
"rewards/margins": 2.6381330490112305,
"rewards/rejected": 0.0,
"step": 1215
},
{
"epoch": 6.793296089385475,
"grad_norm": 2.259854727960776,
"learning_rate": 7.121834254010397e-07,
"logits/chosen": -3.683525323867798,
"logits/rejected": -3.503322124481201,
"logps/chosen": -2.7960424423217773,
"logps/rejected": -32.17485046386719,
"loss": 0.1358,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0533640384674072,
"rewards/margins": 2.0533640384674072,
"rewards/rejected": 0.0,
"step": 1216
},
{
"epoch": 6.798882681564246,
"grad_norm": 1.7479237032343986,
"learning_rate": 7.116313311504459e-07,
"logits/chosen": -3.8045952320098877,
"logits/rejected": -3.823420524597168,
"logps/chosen": -8.632901191711426,
"logps/rejected": -41.81629180908203,
"loss": 0.1257,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8498501777648926,
"rewards/margins": 2.8498501777648926,
"rewards/rejected": 0.0,
"step": 1217
},
{
"epoch": 6.804469273743017,
"grad_norm": 3.3165446290546354,
"learning_rate": 7.110789223780045e-07,
"logits/chosen": -3.540322780609131,
"logits/rejected": -3.6948049068450928,
"logps/chosen": -7.688474178314209,
"logps/rejected": -58.047393798828125,
"loss": 0.1218,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5741453170776367,
"rewards/margins": 2.5741453170776367,
"rewards/rejected": 0.0,
"step": 1218
},
{
"epoch": 6.810055865921788,
"grad_norm": 4.9105938149798485,
"learning_rate": 7.105261999046934e-07,
"logits/chosen": -3.63484787940979,
"logits/rejected": -3.6123716831207275,
"logps/chosen": -6.4083356857299805,
"logps/rejected": -45.80668640136719,
"loss": 0.1398,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0915567874908447,
"rewards/margins": 2.0915567874908447,
"rewards/rejected": 0.0,
"step": 1219
},
{
"epoch": 6.815642458100559,
"grad_norm": 2.493230433513645,
"learning_rate": 7.099731645519567e-07,
"logits/chosen": -3.586568832397461,
"logits/rejected": -3.7155306339263916,
"logps/chosen": -16.714839935302734,
"logps/rejected": -30.529937744140625,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.616159439086914,
"rewards/margins": 2.616159439086914,
"rewards/rejected": 0.0,
"step": 1220
},
{
"epoch": 6.815642458100559,
"eval_logits/chosen": -3.400486707687378,
"eval_logits/rejected": -3.5274319648742676,
"eval_logps/chosen": -23.52200698852539,
"eval_logps/rejected": -42.7762451171875,
"eval_loss": 0.6176553964614868,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/chosen": 0.6548033356666565,
"eval_rewards/margins": 0.6548033356666565,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7579,
"eval_samples_per_second": 9.463,
"eval_steps_per_second": 0.305,
"step": 1220
},
{
"epoch": 6.82122905027933,
"grad_norm": 1.8848191142706658,
"learning_rate": 7.094198171417031e-07,
"logits/chosen": -3.562079429626465,
"logits/rejected": -3.5121777057647705,
"logps/chosen": -1.9159338474273682,
"logps/rejected": -44.53816223144531,
"loss": 0.1196,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6635804176330566,
"rewards/margins": 1.6635804176330566,
"rewards/rejected": 0.0,
"step": 1221
},
{
"epoch": 6.826815642458101,
"grad_norm": 9.830095342409699,
"learning_rate": 7.088661584963058e-07,
"logits/chosen": -3.6274678707122803,
"logits/rejected": -3.4499831199645996,
"logps/chosen": -10.85792064666748,
"logps/rejected": -25.182781219482422,
"loss": 0.1478,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5888285636901855,
"rewards/margins": 2.5888285636901855,
"rewards/rejected": 0.0,
"step": 1222
},
{
"epoch": 6.832402234636872,
"grad_norm": 3.349762567016223,
"learning_rate": 7.083121894385999e-07,
"logits/chosen": -3.280095338821411,
"logits/rejected": -3.4245829582214355,
"logps/chosen": -11.596260070800781,
"logps/rejected": -51.15528869628906,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3517813682556152,
"rewards/margins": 2.3517813682556152,
"rewards/rejected": 0.0,
"step": 1223
},
{
"epoch": 6.837988826815643,
"grad_norm": 2.407717174388427,
"learning_rate": 7.077579107918821e-07,
"logits/chosen": -3.8237805366516113,
"logits/rejected": -3.719397783279419,
"logps/chosen": -5.843072414398193,
"logps/rejected": -27.59650421142578,
"loss": 0.1151,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6652581691741943,
"rewards/margins": 2.6652581691741943,
"rewards/rejected": 0.0,
"step": 1224
},
{
"epoch": 6.843575418994414,
"grad_norm": 2.360322438871608,
"learning_rate": 7.072033233799091e-07,
"logits/chosen": -3.5416440963745117,
"logits/rejected": -3.6952686309814453,
"logps/chosen": -2.770350694656372,
"logps/rejected": -73.64277648925781,
"loss": 0.1258,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.074777603149414,
"rewards/margins": 2.074777603149414,
"rewards/rejected": 0.0,
"step": 1225
},
{
"epoch": 6.849162011173185,
"grad_norm": 3.7667916496111733,
"learning_rate": 7.066484280268967e-07,
"logits/chosen": -3.4020073413848877,
"logits/rejected": -3.483330726623535,
"logps/chosen": -6.428084373474121,
"logps/rejected": -30.403133392333984,
"loss": 0.1671,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0210697650909424,
"rewards/margins": 3.0210697650909424,
"rewards/rejected": 0.0,
"step": 1226
},
{
"epoch": 6.854748603351956,
"grad_norm": 4.398224507010163,
"learning_rate": 7.060932255575182e-07,
"logits/chosen": -3.457984447479248,
"logits/rejected": -3.4139060974121094,
"logps/chosen": -2.470066785812378,
"logps/rejected": -100.52609252929688,
"loss": 0.1239,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1537039279937744,
"rewards/margins": 2.1537039279937744,
"rewards/rejected": 0.0,
"step": 1227
},
{
"epoch": 6.860335195530726,
"grad_norm": 2.487376116821704,
"learning_rate": 7.055377167969034e-07,
"logits/chosen": -3.5053632259368896,
"logits/rejected": -3.6033573150634766,
"logps/chosen": -2.783202648162842,
"logps/rejected": -30.8094425201416,
"loss": 0.1117,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.058837413787842,
"rewards/margins": 2.058837413787842,
"rewards/rejected": 0.0,
"step": 1228
},
{
"epoch": 6.865921787709497,
"grad_norm": 2.0118986319633723,
"learning_rate": 7.049819025706371e-07,
"logits/chosen": -3.568035840988159,
"logits/rejected": -3.5127785205841064,
"logps/chosen": -2.541534900665283,
"logps/rejected": -27.387203216552734,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.272038459777832,
"rewards/margins": 2.272038459777832,
"rewards/rejected": 0.0,
"step": 1229
},
{
"epoch": 6.871508379888268,
"grad_norm": 3.518051662240881,
"learning_rate": 7.044257837047583e-07,
"logits/chosen": -3.6467161178588867,
"logits/rejected": -3.5425987243652344,
"logps/chosen": -6.234301567077637,
"logps/rejected": -38.334869384765625,
"loss": 0.142,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4043684005737305,
"rewards/margins": 2.4043684005737305,
"rewards/rejected": 0.0,
"step": 1230
},
{
"epoch": 6.877094972067039,
"grad_norm": 4.108878555529441,
"learning_rate": 7.038693610257588e-07,
"logits/chosen": -3.5054585933685303,
"logits/rejected": -3.3625519275665283,
"logps/chosen": -9.478939056396484,
"logps/rejected": -25.657081604003906,
"loss": 0.1414,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4863359928131104,
"rewards/margins": 2.4863359928131104,
"rewards/rejected": 0.0,
"step": 1231
},
{
"epoch": 6.88268156424581,
"grad_norm": 8.374221840904603,
"learning_rate": 7.033126353605816e-07,
"logits/chosen": -3.3767518997192383,
"logits/rejected": -3.1736011505126953,
"logps/chosen": -9.914738655090332,
"logps/rejected": -28.2974853515625,
"loss": 0.1791,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1386775970458984,
"rewards/margins": 2.1386775970458984,
"rewards/rejected": 0.0,
"step": 1232
},
{
"epoch": 6.888268156424581,
"grad_norm": 1.7720574028531357,
"learning_rate": 7.027556075366202e-07,
"logits/chosen": -3.438734531402588,
"logits/rejected": -3.4485371112823486,
"logps/chosen": -3.397510051727295,
"logps/rejected": -41.679359436035156,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.425100803375244,
"rewards/margins": 2.425100803375244,
"rewards/rejected": 0.0,
"step": 1233
},
{
"epoch": 6.893854748603352,
"grad_norm": 4.468675600832932,
"learning_rate": 7.021982783817172e-07,
"logits/chosen": -3.4875001907348633,
"logits/rejected": -3.5468428134918213,
"logps/chosen": -9.436210632324219,
"logps/rejected": -37.21449279785156,
"loss": 0.1091,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.802164077758789,
"rewards/margins": 2.802164077758789,
"rewards/rejected": 0.0,
"step": 1234
},
{
"epoch": 6.899441340782123,
"grad_norm": 4.452848779091001,
"learning_rate": 7.01640648724163e-07,
"logits/chosen": -3.4259796142578125,
"logits/rejected": -3.4563775062561035,
"logps/chosen": -5.967182159423828,
"logps/rejected": -51.27204513549805,
"loss": 0.1195,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.487522602081299,
"rewards/margins": 2.487522602081299,
"rewards/rejected": 0.0,
"step": 1235
},
{
"epoch": 6.905027932960894,
"grad_norm": 3.8344758884608425,
"learning_rate": 7.010827193926946e-07,
"logits/chosen": -3.5665276050567627,
"logits/rejected": -3.6301960945129395,
"logps/chosen": -0.8891786336898804,
"logps/rejected": -61.749176025390625,
"loss": 0.1233,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3752551078796387,
"rewards/margins": 1.3752551078796387,
"rewards/rejected": 0.0,
"step": 1236
},
{
"epoch": 6.910614525139665,
"grad_norm": 1.4105069292029504,
"learning_rate": 7.005244912164941e-07,
"logits/chosen": -3.363919734954834,
"logits/rejected": -3.5228559970855713,
"logps/chosen": -6.403987884521484,
"logps/rejected": -29.543251037597656,
"loss": 0.1053,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6485252380371094,
"rewards/margins": 2.6485252380371094,
"rewards/rejected": 0.0,
"step": 1237
},
{
"epoch": 6.916201117318436,
"grad_norm": 1.567324172711957,
"learning_rate": 6.999659650251885e-07,
"logits/chosen": -3.5861716270446777,
"logits/rejected": -3.5285205841064453,
"logps/chosen": -19.173553466796875,
"logps/rejected": -30.03134536743164,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.41219162940979,
"rewards/margins": 3.41219162940979,
"rewards/rejected": 0.0,
"step": 1238
},
{
"epoch": 6.921787709497207,
"grad_norm": 2.9385896411756254,
"learning_rate": 6.994071416488467e-07,
"logits/chosen": -3.4926271438598633,
"logits/rejected": -3.8362624645233154,
"logps/chosen": -2.885002851486206,
"logps/rejected": -75.10549926757812,
"loss": 0.1026,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.034637212753296,
"rewards/margins": 2.034637212753296,
"rewards/rejected": 0.0,
"step": 1239
},
{
"epoch": 6.927374301675978,
"grad_norm": 1.8495862875255806,
"learning_rate": 6.988480219179801e-07,
"logits/chosen": -3.431464195251465,
"logits/rejected": -3.0147957801818848,
"logps/chosen": -1.5547078847885132,
"logps/rejected": -49.615867614746094,
"loss": 0.1516,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7282437086105347,
"rewards/margins": 1.7282437086105347,
"rewards/rejected": 0.0,
"step": 1240
},
{
"epoch": 6.927374301675978,
"eval_logits/chosen": -3.40470814704895,
"eval_logits/rejected": -3.530635118484497,
"eval_logps/chosen": -23.34773826599121,
"eval_logps/rejected": -43.31367874145508,
"eval_loss": 0.6201418042182922,
"eval_rewards/accuracies": 0.875,
"eval_rewards/chosen": 0.6722301840782166,
"eval_rewards/margins": 0.6722301840782166,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7094,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 1240
},
{
"epoch": 6.932960893854749,
"grad_norm": 3.022799239575308,
"learning_rate": 6.982886066635398e-07,
"logits/chosen": -3.509122371673584,
"logits/rejected": -3.611570358276367,
"logps/chosen": -5.77220344543457,
"logps/rejected": -30.937862396240234,
"loss": 0.0919,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.851273536682129,
"rewards/margins": 2.851273536682129,
"rewards/rejected": 0.0,
"step": 1241
},
{
"epoch": 6.93854748603352,
"grad_norm": 3.6542646780509718,
"learning_rate": 6.97728896716917e-07,
"logits/chosen": -3.1856911182403564,
"logits/rejected": -3.246323823928833,
"logps/chosen": -5.146172523498535,
"logps/rejected": -26.055557250976562,
"loss": 0.1565,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2226128578186035,
"rewards/margins": 2.2226128578186035,
"rewards/rejected": 0.0,
"step": 1242
},
{
"epoch": 6.94413407821229,
"grad_norm": 2.724818942837176,
"learning_rate": 6.9716889290994e-07,
"logits/chosen": -3.484645128250122,
"logits/rejected": -3.6872596740722656,
"logps/chosen": -15.662714004516602,
"logps/rejected": -20.226940155029297,
"loss": 0.0977,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2329084873199463,
"rewards/margins": 2.2329084873199463,
"rewards/rejected": 0.0,
"step": 1243
},
{
"epoch": 6.949720670391061,
"grad_norm": 3.233174886627538,
"learning_rate": 6.966085960748741e-07,
"logits/chosen": -3.3511745929718018,
"logits/rejected": -3.443159818649292,
"logps/chosen": -15.10945987701416,
"logps/rejected": -27.131855010986328,
"loss": 0.1387,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1416854858398438,
"rewards/margins": 3.1416854858398438,
"rewards/rejected": 0.0,
"step": 1244
},
{
"epoch": 6.955307262569832,
"grad_norm": 1.7873288884349265,
"learning_rate": 6.960480070444205e-07,
"logits/chosen": -3.6787734031677246,
"logits/rejected": -3.816436767578125,
"logps/chosen": -8.850296020507812,
"logps/rejected": -18.008380889892578,
"loss": 0.0784,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0034384727478027,
"rewards/margins": 3.0034384727478027,
"rewards/rejected": 0.0,
"step": 1245
},
{
"epoch": 6.960893854748603,
"grad_norm": 3.933611288564565,
"learning_rate": 6.954871266517142e-07,
"logits/chosen": -3.684723138809204,
"logits/rejected": -3.5727052688598633,
"logps/chosen": -9.609380722045898,
"logps/rejected": -21.106409072875977,
"loss": 0.1188,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8668670654296875,
"rewards/margins": 2.8668670654296875,
"rewards/rejected": 0.0,
"step": 1246
},
{
"epoch": 6.966480446927374,
"grad_norm": 3.188199242094565,
"learning_rate": 6.949259557303232e-07,
"logits/chosen": -3.592222213745117,
"logits/rejected": -3.5707173347473145,
"logps/chosen": -3.5327694416046143,
"logps/rejected": -23.995849609375,
"loss": 0.1362,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4314236640930176,
"rewards/margins": 2.4314236640930176,
"rewards/rejected": 0.0,
"step": 1247
},
{
"epoch": 6.972067039106145,
"grad_norm": 6.666790429932437,
"learning_rate": 6.943644951142477e-07,
"logits/chosen": -3.750788688659668,
"logits/rejected": -3.7600064277648926,
"logps/chosen": -25.545528411865234,
"logps/rejected": -38.38447570800781,
"loss": 0.1738,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.568408250808716,
"rewards/margins": 3.568408250808716,
"rewards/rejected": 0.0,
"step": 1248
},
{
"epoch": 6.977653631284916,
"grad_norm": 1.9302870448901726,
"learning_rate": 6.938027456379181e-07,
"logits/chosen": -3.6586012840270996,
"logits/rejected": -3.828281879425049,
"logps/chosen": -2.125006914138794,
"logps/rejected": -25.19756317138672,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2528858184814453,
"rewards/margins": 2.2528858184814453,
"rewards/rejected": 0.0,
"step": 1249
},
{
"epoch": 6.983240223463687,
"grad_norm": 4.1458219129676985,
"learning_rate": 6.93240708136194e-07,
"logits/chosen": -3.2185046672821045,
"logits/rejected": -3.186969518661499,
"logps/chosen": -46.381134033203125,
"logps/rejected": -36.031314849853516,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.10642409324646,
"rewards/margins": 2.10642409324646,
"rewards/rejected": 0.0,
"step": 1250
},
{
"epoch": 6.988826815642458,
"grad_norm": 2.0311511446524064,
"learning_rate": 6.926783834443633e-07,
"logits/chosen": -3.650008201599121,
"logits/rejected": -3.5961015224456787,
"logps/chosen": -2.199688196182251,
"logps/rejected": -33.29227828979492,
"loss": 0.1303,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8048226833343506,
"rewards/margins": 1.8048226833343506,
"rewards/rejected": 0.0,
"step": 1251
},
{
"epoch": 6.994413407821229,
"grad_norm": 5.875325680193311,
"learning_rate": 6.921157723981408e-07,
"logits/chosen": -3.4090051651000977,
"logits/rejected": -3.5231189727783203,
"logps/chosen": -37.33778762817383,
"logps/rejected": -23.965280532836914,
"loss": 0.127,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1387805938720703,
"rewards/margins": 2.1387805938720703,
"rewards/rejected": 0.0,
"step": 1252
},
{
"epoch": 7.0,
"grad_norm": 4.5952188096971085,
"learning_rate": 6.915528758336664e-07,
"logits/chosen": -3.5402402877807617,
"logits/rejected": -3.58508563041687,
"logps/chosen": -10.173696517944336,
"logps/rejected": -33.168235778808594,
"loss": 0.1321,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.598673105239868,
"rewards/margins": 2.598673105239868,
"rewards/rejected": 0.0,
"step": 1253
},
{
"epoch": 7.005586592178771,
"grad_norm": 1.9089672348865112,
"learning_rate": 6.909896945875049e-07,
"logits/chosen": -3.530308723449707,
"logits/rejected": -3.4146568775177,
"logps/chosen": -7.088906288146973,
"logps/rejected": -26.716922760009766,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2254104614257812,
"rewards/margins": 3.2254104614257812,
"rewards/rejected": 0.0,
"step": 1254
},
{
"epoch": 7.011173184357542,
"grad_norm": 1.9922181482129386,
"learning_rate": 6.904262294966438e-07,
"logits/chosen": -3.759143114089966,
"logits/rejected": -3.964033842086792,
"logps/chosen": -8.769649505615234,
"logps/rejected": -39.40819549560547,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.294891357421875,
"rewards/margins": 3.294891357421875,
"rewards/rejected": 0.0,
"step": 1255
},
{
"epoch": 7.016759776536313,
"grad_norm": 2.8377968284119115,
"learning_rate": 6.898624813984927e-07,
"logits/chosen": -3.4786267280578613,
"logits/rejected": -3.4019179344177246,
"logps/chosen": -3.114628791809082,
"logps/rejected": -68.92147827148438,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.311002254486084,
"rewards/margins": 2.311002254486084,
"rewards/rejected": 0.0,
"step": 1256
},
{
"epoch": 7.022346368715084,
"grad_norm": 1.6878198951718528,
"learning_rate": 6.892984511308813e-07,
"logits/chosen": -3.630856990814209,
"logits/rejected": -3.5413107872009277,
"logps/chosen": -4.833260536193848,
"logps/rejected": -46.88436508178711,
"loss": 0.1113,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4715824127197266,
"rewards/margins": 2.4715824127197266,
"rewards/rejected": 0.0,
"step": 1257
},
{
"epoch": 7.027932960893855,
"grad_norm": 1.3587412021660126,
"learning_rate": 6.887341395320596e-07,
"logits/chosen": -3.56107234954834,
"logits/rejected": -3.789989948272705,
"logps/chosen": -1.250314712524414,
"logps/rejected": -30.90006446838379,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7148796319961548,
"rewards/margins": 1.7148796319961548,
"rewards/rejected": 0.0,
"step": 1258
},
{
"epoch": 7.033519553072626,
"grad_norm": 1.6857703997299212,
"learning_rate": 6.881695474406947e-07,
"logits/chosen": -3.827329635620117,
"logits/rejected": -3.8502554893493652,
"logps/chosen": -2.4466514587402344,
"logps/rejected": -41.08592987060547,
"loss": 0.1031,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9332318305969238,
"rewards/margins": 1.9332318305969238,
"rewards/rejected": 0.0,
"step": 1259
},
{
"epoch": 7.039106145251397,
"grad_norm": 2.4773497393362045,
"learning_rate": 6.876046756958713e-07,
"logits/chosen": -3.564875841140747,
"logits/rejected": -3.6816511154174805,
"logps/chosen": -13.001611709594727,
"logps/rejected": -52.05155944824219,
"loss": 0.1004,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8277783393859863,
"rewards/margins": 2.8277783393859863,
"rewards/rejected": 0.0,
"step": 1260
},
{
"epoch": 7.039106145251397,
"eval_logits/chosen": -3.3980274200439453,
"eval_logits/rejected": -3.524296283721924,
"eval_logps/chosen": -23.279247283935547,
"eval_logps/rejected": -42.626617431640625,
"eval_loss": 0.6063249111175537,
"eval_rewards/accuracies": 0.8500000238418579,
"eval_rewards/chosen": 0.6790796518325806,
"eval_rewards/margins": 0.6790796518325806,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7375,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 1260
},
{
"epoch": 7.044692737430168,
"grad_norm": 3.0503634919952907,
"learning_rate": 6.870395251370893e-07,
"logits/chosen": -3.6887543201446533,
"logits/rejected": -3.7471749782562256,
"logps/chosen": -13.48740005493164,
"logps/rejected": -36.2447395324707,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6292214393615723,
"rewards/margins": 3.6292214393615723,
"rewards/rejected": 0.0,
"step": 1261
},
{
"epoch": 7.050279329608939,
"grad_norm": 2.355830206300344,
"learning_rate": 6.864740966042628e-07,
"logits/chosen": -3.3249711990356445,
"logits/rejected": -3.4159483909606934,
"logps/chosen": -2.5284647941589355,
"logps/rejected": -30.496246337890625,
"loss": 0.1099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2643685340881348,
"rewards/margins": 2.2643685340881348,
"rewards/rejected": 0.0,
"step": 1262
},
{
"epoch": 7.055865921787709,
"grad_norm": 2.7909867720166153,
"learning_rate": 6.859083909377196e-07,
"logits/chosen": -3.4230048656463623,
"logits/rejected": -3.4332618713378906,
"logps/chosen": -12.817394256591797,
"logps/rejected": -29.420516967773438,
"loss": 0.1351,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.819218873977661,
"rewards/margins": 2.819218873977661,
"rewards/rejected": 0.0,
"step": 1263
},
{
"epoch": 7.06145251396648,
"grad_norm": 1.1915542948733544,
"learning_rate": 6.85342408978199e-07,
"logits/chosen": -3.303455352783203,
"logits/rejected": -3.5971484184265137,
"logps/chosen": -7.716994762420654,
"logps/rejected": -54.277198791503906,
"loss": 0.1486,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4927916526794434,
"rewards/margins": 2.4927916526794434,
"rewards/rejected": 0.0,
"step": 1264
},
{
"epoch": 7.067039106145251,
"grad_norm": 4.254128587452519,
"learning_rate": 6.847761515668511e-07,
"logits/chosen": -3.419790506362915,
"logits/rejected": -3.2606422901153564,
"logps/chosen": -16.361888885498047,
"logps/rejected": -57.17011642456055,
"loss": 0.1224,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.781437397003174,
"rewards/margins": 2.781437397003174,
"rewards/rejected": 0.0,
"step": 1265
},
{
"epoch": 7.072625698324022,
"grad_norm": 1.572679292363798,
"learning_rate": 6.842096195452349e-07,
"logits/chosen": -3.235483169555664,
"logits/rejected": -3.463332414627075,
"logps/chosen": -2.7052762508392334,
"logps/rejected": -94.96967315673828,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9657814502716064,
"rewards/margins": 1.9657814502716064,
"rewards/rejected": 0.0,
"step": 1266
},
{
"epoch": 7.078212290502793,
"grad_norm": 3.991580921121,
"learning_rate": 6.836428137553183e-07,
"logits/chosen": -3.574601411819458,
"logits/rejected": -3.6130170822143555,
"logps/chosen": -7.246374607086182,
"logps/rejected": -52.97753143310547,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0662167072296143,
"rewards/margins": 3.0662167072296143,
"rewards/rejected": 0.0,
"step": 1267
},
{
"epoch": 7.083798882681564,
"grad_norm": 1.5920670141128423,
"learning_rate": 6.830757350394754e-07,
"logits/chosen": -3.4136204719543457,
"logits/rejected": -3.473283529281616,
"logps/chosen": -1.0013225078582764,
"logps/rejected": -64.36286926269531,
"loss": 0.1319,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.665197491645813,
"rewards/margins": 1.665197491645813,
"rewards/rejected": 0.0,
"step": 1268
},
{
"epoch": 7.089385474860335,
"grad_norm": 1.9658474063902145,
"learning_rate": 6.825083842404862e-07,
"logits/chosen": -3.531083583831787,
"logits/rejected": -3.6414847373962402,
"logps/chosen": -1.0360397100448608,
"logps/rejected": -97.47354888916016,
"loss": 0.1368,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8494411706924438,
"rewards/margins": 1.8494411706924438,
"rewards/rejected": 0.0,
"step": 1269
},
{
"epoch": 7.094972067039106,
"grad_norm": 1.8695660026719596,
"learning_rate": 6.819407622015352e-07,
"logits/chosen": -3.7749783992767334,
"logits/rejected": -3.5601696968078613,
"logps/chosen": -13.966944694519043,
"logps/rejected": -26.170475006103516,
"loss": 0.1157,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.500847578048706,
"rewards/margins": 3.500847578048706,
"rewards/rejected": 0.0,
"step": 1270
},
{
"epoch": 7.100558659217877,
"grad_norm": 2.098493826176811,
"learning_rate": 6.813728697662096e-07,
"logits/chosen": -3.7067666053771973,
"logits/rejected": -3.4574549198150635,
"logps/chosen": -6.075762748718262,
"logps/rejected": -60.33864974975586,
"loss": 0.0944,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.639904737472534,
"rewards/margins": 2.639904737472534,
"rewards/rejected": 0.0,
"step": 1271
},
{
"epoch": 7.106145251396648,
"grad_norm": 1.1897135710493125,
"learning_rate": 6.808047077784987e-07,
"logits/chosen": -3.3910703659057617,
"logits/rejected": -3.6197991371154785,
"logps/chosen": -8.341439247131348,
"logps/rejected": -38.314170837402344,
"loss": 0.0781,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6378109455108643,
"rewards/margins": 2.6378109455108643,
"rewards/rejected": 0.0,
"step": 1272
},
{
"epoch": 7.111731843575419,
"grad_norm": 1.4566550515248526,
"learning_rate": 6.802362770827926e-07,
"logits/chosen": -3.423917770385742,
"logits/rejected": -3.4773166179656982,
"logps/chosen": -9.888988494873047,
"logps/rejected": -38.88087844848633,
"loss": 0.1033,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1006269454956055,
"rewards/margins": 3.1006269454956055,
"rewards/rejected": 0.0,
"step": 1273
},
{
"epoch": 7.11731843575419,
"grad_norm": 1.5518900819244872,
"learning_rate": 6.796675785238803e-07,
"logits/chosen": -3.8480265140533447,
"logits/rejected": -3.7801334857940674,
"logps/chosen": -7.182609558105469,
"logps/rejected": -22.92597198486328,
"loss": 0.0888,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.425747871398926,
"rewards/margins": 3.425747871398926,
"rewards/rejected": 0.0,
"step": 1274
},
{
"epoch": 7.122905027932961,
"grad_norm": 1.0502692279830264,
"learning_rate": 6.790986129469494e-07,
"logits/chosen": -3.9019930362701416,
"logits/rejected": -3.716822624206543,
"logps/chosen": -3.439368963241577,
"logps/rejected": -28.22347640991211,
"loss": 0.0755,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6313304901123047,
"rewards/margins": 2.6313304901123047,
"rewards/rejected": 0.0,
"step": 1275
},
{
"epoch": 7.128491620111732,
"grad_norm": 1.6363874006250305,
"learning_rate": 6.785293811975838e-07,
"logits/chosen": -3.638106346130371,
"logits/rejected": -3.825218915939331,
"logps/chosen": -1.558204174041748,
"logps/rejected": -50.177581787109375,
"loss": 0.1165,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8915550708770752,
"rewards/margins": 1.8915550708770752,
"rewards/rejected": 0.0,
"step": 1276
},
{
"epoch": 7.134078212290503,
"grad_norm": 2.603208295074469,
"learning_rate": 6.779598841217635e-07,
"logits/chosen": -3.2676031589508057,
"logits/rejected": -3.6138453483581543,
"logps/chosen": -1.8926982879638672,
"logps/rejected": -48.11708068847656,
"loss": 0.1167,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1727561950683594,
"rewards/margins": 2.1727561950683594,
"rewards/rejected": 0.0,
"step": 1277
},
{
"epoch": 7.139664804469274,
"grad_norm": 2.3267088087405785,
"learning_rate": 6.773901225658625e-07,
"logits/chosen": -3.5336978435516357,
"logits/rejected": -3.351433038711548,
"logps/chosen": -2.522913694381714,
"logps/rejected": -54.36046600341797,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.253704786300659,
"rewards/margins": 2.253704786300659,
"rewards/rejected": 0.0,
"step": 1278
},
{
"epoch": 7.145251396648045,
"grad_norm": 1.0778863521725175,
"learning_rate": 6.768200973766478e-07,
"logits/chosen": -3.33052396774292,
"logits/rejected": -3.2629311084747314,
"logps/chosen": -4.489758491516113,
"logps/rejected": -36.76377868652344,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9570207595825195,
"rewards/margins": 2.9570207595825195,
"rewards/rejected": 0.0,
"step": 1279
},
{
"epoch": 7.150837988826815,
"grad_norm": 1.4641419644916576,
"learning_rate": 6.762498094012786e-07,
"logits/chosen": -3.2497832775115967,
"logits/rejected": -3.2622108459472656,
"logps/chosen": -5.682682991027832,
"logps/rejected": -47.43547821044922,
"loss": 0.1213,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.835364580154419,
"rewards/margins": 1.835364580154419,
"rewards/rejected": 0.0,
"step": 1280
},
{
"epoch": 7.150837988826815,
"eval_logits/chosen": -3.366528034210205,
"eval_logits/rejected": -3.4954731464385986,
"eval_logps/chosen": -25.02907943725586,
"eval_logps/rejected": -45.20484161376953,
"eval_loss": 0.688734769821167,
"eval_rewards/accuracies": 0.75,
"eval_rewards/chosen": 0.5040964484214783,
"eval_rewards/margins": 0.5040964484214783,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7258,
"eval_samples_per_second": 9.473,
"eval_steps_per_second": 0.306,
"step": 1280
},
{
"epoch": 7.156424581005586,
"grad_norm": 2.3060896288623574,
"learning_rate": 6.756792594873042e-07,
"logits/chosen": -3.3901827335357666,
"logits/rejected": -3.4088780879974365,
"logps/chosen": -6.201159477233887,
"logps/rejected": -30.820392608642578,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5701470375061035,
"rewards/margins": 2.5701470375061035,
"rewards/rejected": 0.0,
"step": 1281
},
{
"epoch": 7.162011173184357,
"grad_norm": 1.126589676343673,
"learning_rate": 6.751084484826635e-07,
"logits/chosen": -3.718583106994629,
"logits/rejected": -3.7990591526031494,
"logps/chosen": -2.9070825576782227,
"logps/rejected": -38.872406005859375,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5956931114196777,
"rewards/margins": 2.5956931114196777,
"rewards/rejected": 0.0,
"step": 1282
},
{
"epoch": 7.167597765363128,
"grad_norm": 1.243585930640092,
"learning_rate": 6.74537377235683e-07,
"logits/chosen": -3.553194761276245,
"logits/rejected": -3.4584012031555176,
"logps/chosen": -2.784400463104248,
"logps/rejected": -50.803123474121094,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5433735847473145,
"rewards/margins": 2.5433735847473145,
"rewards/rejected": 0.0,
"step": 1283
},
{
"epoch": 7.173184357541899,
"grad_norm": 1.0585329819655018,
"learning_rate": 6.739660465950769e-07,
"logits/chosen": -2.675751209259033,
"logits/rejected": -2.5353894233703613,
"logps/chosen": -16.220413208007812,
"logps/rejected": -52.3974609375,
"loss": 0.0922,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.149181842803955,
"rewards/margins": 2.149181842803955,
"rewards/rejected": 0.0,
"step": 1284
},
{
"epoch": 7.17877094972067,
"grad_norm": 2.700793634333128,
"learning_rate": 6.733944574099436e-07,
"logits/chosen": -3.5255300998687744,
"logits/rejected": -3.5196940898895264,
"logps/chosen": -12.33900260925293,
"logps/rejected": -53.60920715332031,
"loss": 0.1398,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0111923217773438,
"rewards/margins": 3.0111923217773438,
"rewards/rejected": 0.0,
"step": 1285
},
{
"epoch": 7.184357541899441,
"grad_norm": 1.0390366336173877,
"learning_rate": 6.728226105297667e-07,
"logits/chosen": -3.7337443828582764,
"logits/rejected": -3.573716402053833,
"logps/chosen": -1.5995138883590698,
"logps/rejected": -57.31018829345703,
"loss": 0.0987,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3661184310913086,
"rewards/margins": 2.3661184310913086,
"rewards/rejected": 0.0,
"step": 1286
},
{
"epoch": 7.189944134078212,
"grad_norm": 1.104951125795208,
"learning_rate": 6.722505068044127e-07,
"logits/chosen": -3.5012717247009277,
"logits/rejected": -3.511260747909546,
"logps/chosen": -18.973304748535156,
"logps/rejected": -33.1102409362793,
"loss": 0.0739,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6038103103637695,
"rewards/margins": 3.6038103103637695,
"rewards/rejected": 0.0,
"step": 1287
},
{
"epoch": 7.195530726256983,
"grad_norm": 1.6404524953515758,
"learning_rate": 6.716781470841295e-07,
"logits/chosen": -3.426778554916382,
"logits/rejected": -3.508471727371216,
"logps/chosen": -1.3954945802688599,
"logps/rejected": -61.31631088256836,
"loss": 0.119,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7865869998931885,
"rewards/margins": 1.7865869998931885,
"rewards/rejected": 0.0,
"step": 1288
},
{
"epoch": 7.201117318435754,
"grad_norm": 2.086553690039233,
"learning_rate": 6.711055322195455e-07,
"logits/chosen": -3.692363739013672,
"logits/rejected": -3.5762698650360107,
"logps/chosen": -7.974408149719238,
"logps/rejected": -28.232044219970703,
"loss": 0.0908,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3824398517608643,
"rewards/margins": 3.3824398517608643,
"rewards/rejected": 0.0,
"step": 1289
},
{
"epoch": 7.206703910614525,
"grad_norm": 1.3407543422176662,
"learning_rate": 6.705326630616683e-07,
"logits/chosen": -3.232694625854492,
"logits/rejected": -3.614248037338257,
"logps/chosen": -1.9202529191970825,
"logps/rejected": -54.0255126953125,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.092041254043579,
"rewards/margins": 2.092041254043579,
"rewards/rejected": 0.0,
"step": 1290
},
{
"epoch": 7.212290502793296,
"grad_norm": 2.7446037098899643,
"learning_rate": 6.699595404618836e-07,
"logits/chosen": -3.6303162574768066,
"logits/rejected": -3.593430519104004,
"logps/chosen": -12.793563842773438,
"logps/rejected": -23.735065460205078,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.336655616760254,
"rewards/margins": 2.336655616760254,
"rewards/rejected": 0.0,
"step": 1291
},
{
"epoch": 7.217877094972067,
"grad_norm": 1.1170534654840107,
"learning_rate": 6.693861652719536e-07,
"logits/chosen": -3.604806661605835,
"logits/rejected": -3.708782196044922,
"logps/chosen": -5.701390266418457,
"logps/rejected": -73.40631103515625,
"loss": 0.1009,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7073326110839844,
"rewards/margins": 2.7073326110839844,
"rewards/rejected": 0.0,
"step": 1292
},
{
"epoch": 7.223463687150838,
"grad_norm": 2.337765936858399,
"learning_rate": 6.68812538344016e-07,
"logits/chosen": -3.408153533935547,
"logits/rejected": -3.438262462615967,
"logps/chosen": -51.61593246459961,
"logps/rejected": -25.005937576293945,
"loss": 0.1092,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4474377632141113,
"rewards/margins": 3.4474377632141113,
"rewards/rejected": 0.0,
"step": 1293
},
{
"epoch": 7.229050279329609,
"grad_norm": 1.1792228953093058,
"learning_rate": 6.682386605305827e-07,
"logits/chosen": -3.4808995723724365,
"logits/rejected": -3.4360406398773193,
"logps/chosen": -3.751499652862549,
"logps/rejected": -42.26259994506836,
"loss": 0.1093,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.625836133956909,
"rewards/margins": 2.625836133956909,
"rewards/rejected": 0.0,
"step": 1294
},
{
"epoch": 7.23463687150838,
"grad_norm": 4.963918382196016,
"learning_rate": 6.676645326845379e-07,
"logits/chosen": -3.3212473392486572,
"logits/rejected": -3.2344558238983154,
"logps/chosen": -8.124931335449219,
"logps/rejected": -29.031097412109375,
"loss": 0.1292,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6535797119140625,
"rewards/margins": 2.6535797119140625,
"rewards/rejected": 0.0,
"step": 1295
},
{
"epoch": 7.240223463687151,
"grad_norm": 1.1186223144673042,
"learning_rate": 6.670901556591383e-07,
"logits/chosen": -3.446485996246338,
"logits/rejected": -3.4796955585479736,
"logps/chosen": -1.435394287109375,
"logps/rejected": -35.34324645996094,
"loss": 0.1226,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9041576385498047,
"rewards/margins": 1.9041576385498047,
"rewards/rejected": 0.0,
"step": 1296
},
{
"epoch": 7.245810055865922,
"grad_norm": 2.754067509491694,
"learning_rate": 6.665155303080103e-07,
"logits/chosen": -3.3536322116851807,
"logits/rejected": -3.560081720352173,
"logps/chosen": -10.537589073181152,
"logps/rejected": -37.63751220703125,
"loss": 0.093,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.356642723083496,
"rewards/margins": 3.356642723083496,
"rewards/rejected": 0.0,
"step": 1297
},
{
"epoch": 7.251396648044693,
"grad_norm": 1.9364333731940997,
"learning_rate": 6.659406574851494e-07,
"logits/chosen": -3.443791151046753,
"logits/rejected": -3.346338987350464,
"logps/chosen": -6.035096168518066,
"logps/rejected": -64.06090545654297,
"loss": 0.0988,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0059869289398193,
"rewards/margins": 3.0059869289398193,
"rewards/rejected": 0.0,
"step": 1298
},
{
"epoch": 7.256983240223463,
"grad_norm": 1.4585896778906036,
"learning_rate": 6.653655380449192e-07,
"logits/chosen": -2.924417018890381,
"logits/rejected": -2.9158060550689697,
"logps/chosen": -22.547714233398438,
"logps/rejected": -62.49207305908203,
"loss": 0.1306,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.475858211517334,
"rewards/margins": 2.475858211517334,
"rewards/rejected": 0.0,
"step": 1299
},
{
"epoch": 7.262569832402234,
"grad_norm": 2.772557178346172,
"learning_rate": 6.647901728420493e-07,
"logits/chosen": -3.595991373062134,
"logits/rejected": -3.5744409561157227,
"logps/chosen": -18.906320571899414,
"logps/rejected": -29.811229705810547,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.889451742172241,
"rewards/margins": 2.889451742172241,
"rewards/rejected": 0.0,
"step": 1300
},
{
"epoch": 7.262569832402234,
"eval_logits/chosen": -3.3801612854003906,
"eval_logits/rejected": -3.5086123943328857,
"eval_logps/chosen": -24.592073440551758,
"eval_logps/rejected": -44.9486083984375,
"eval_loss": 0.6780498027801514,
"eval_rewards/accuracies": 0.824999988079071,
"eval_rewards/chosen": 0.5477972626686096,
"eval_rewards/margins": 0.5477972626686096,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7423,
"eval_samples_per_second": 9.468,
"eval_steps_per_second": 0.305,
"step": 1300
},
{
"epoch": 7.268156424581005,
"grad_norm": 1.6144703090043173,
"learning_rate": 6.64214562731635e-07,
"logits/chosen": -3.5145280361175537,
"logits/rejected": -3.453125476837158,
"logps/chosen": -15.009201049804688,
"logps/rejected": -35.985870361328125,
"loss": 0.0935,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4529547691345215,
"rewards/margins": 3.4529547691345215,
"rewards/rejected": 0.0,
"step": 1301
},
{
"epoch": 7.273743016759776,
"grad_norm": 13.75040159306429,
"learning_rate": 6.636387085691355e-07,
"logits/chosen": -3.710878610610962,
"logits/rejected": -3.6133155822753906,
"logps/chosen": -19.006427764892578,
"logps/rejected": -27.593320846557617,
"loss": 0.1549,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8588337898254395,
"rewards/margins": 2.8588337898254395,
"rewards/rejected": 0.0,
"step": 1302
},
{
"epoch": 7.279329608938547,
"grad_norm": 1.5685706957219425,
"learning_rate": 6.630626112103727e-07,
"logits/chosen": -3.2809746265411377,
"logits/rejected": -3.403027057647705,
"logps/chosen": -5.856812477111816,
"logps/rejected": -106.83222198486328,
"loss": 0.0835,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7082905769348145,
"rewards/margins": 2.7082905769348145,
"rewards/rejected": 0.0,
"step": 1303
},
{
"epoch": 7.284916201117318,
"grad_norm": 5.498404105160891,
"learning_rate": 6.624862715115297e-07,
"logits/chosen": -3.476365566253662,
"logits/rejected": -3.4864091873168945,
"logps/chosen": -8.03408432006836,
"logps/rejected": -54.53045654296875,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.623234272003174,
"rewards/margins": 2.623234272003174,
"rewards/rejected": 0.0,
"step": 1304
},
{
"epoch": 7.290502793296089,
"grad_norm": 1.7987959234105906,
"learning_rate": 6.6190969032915e-07,
"logits/chosen": -3.5582940578460693,
"logits/rejected": -3.632675886154175,
"logps/chosen": -4.348865985870361,
"logps/rejected": -73.341552734375,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.675520181655884,
"rewards/margins": 2.675520181655884,
"rewards/rejected": 0.0,
"step": 1305
},
{
"epoch": 7.29608938547486,
"grad_norm": 7.064715714550168,
"learning_rate": 6.613328685201361e-07,
"logits/chosen": -3.6168596744537354,
"logits/rejected": -3.6552817821502686,
"logps/chosen": -9.071053504943848,
"logps/rejected": -21.970504760742188,
"loss": 0.1761,
"rewards/accuracies": 0.75,
"rewards/chosen": 2.4692320823669434,
"rewards/margins": 2.4692320823669434,
"rewards/rejected": 0.0,
"step": 1306
},
{
"epoch": 7.301675977653631,
"grad_norm": 1.3862521233978937,
"learning_rate": 6.607558069417476e-07,
"logits/chosen": -3.6154673099517822,
"logits/rejected": -3.7376348972320557,
"logps/chosen": -8.16813850402832,
"logps/rejected": -33.8210334777832,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.583134651184082,
"rewards/margins": 2.583134651184082,
"rewards/rejected": 0.0,
"step": 1307
},
{
"epoch": 7.307262569832402,
"grad_norm": 1.8055495272432047,
"learning_rate": 6.60178506451601e-07,
"logits/chosen": -3.5150320529937744,
"logits/rejected": -3.4990432262420654,
"logps/chosen": -1.468370795249939,
"logps/rejected": -52.503963470458984,
"loss": 0.14,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.68797767162323,
"rewards/margins": 1.68797767162323,
"rewards/rejected": 0.0,
"step": 1308
},
{
"epoch": 7.312849162011173,
"grad_norm": 2.222508934949117,
"learning_rate": 6.596009679076675e-07,
"logits/chosen": -3.292964458465576,
"logits/rejected": -3.464843988418579,
"logps/chosen": -11.528118133544922,
"logps/rejected": -59.465240478515625,
"loss": 0.0994,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.302183151245117,
"rewards/margins": 2.302183151245117,
"rewards/rejected": 0.0,
"step": 1309
},
{
"epoch": 7.318435754189944,
"grad_norm": 1.0094083066470343,
"learning_rate": 6.590231921682724e-07,
"logits/chosen": -3.560662269592285,
"logits/rejected": -3.674679756164551,
"logps/chosen": -9.870981216430664,
"logps/rejected": -36.52283477783203,
"loss": 0.099,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.251835346221924,
"rewards/margins": 3.251835346221924,
"rewards/rejected": 0.0,
"step": 1310
},
{
"epoch": 7.324022346368715,
"grad_norm": 1.2169747164098956,
"learning_rate": 6.584451800920932e-07,
"logits/chosen": -3.272944450378418,
"logits/rejected": -3.4275331497192383,
"logps/chosen": -0.7554080486297607,
"logps/rejected": -94.23004150390625,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7341724634170532,
"rewards/margins": 1.7341724634170532,
"rewards/rejected": 0.0,
"step": 1311
},
{
"epoch": 7.329608938547486,
"grad_norm": 6.70686787687767,
"learning_rate": 6.578669325381589e-07,
"logits/chosen": -3.729093551635742,
"logits/rejected": -3.688983678817749,
"logps/chosen": -12.128386497497559,
"logps/rejected": -46.44956970214844,
"loss": 0.1367,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1582024097442627,
"rewards/margins": 3.1582024097442627,
"rewards/rejected": 0.0,
"step": 1312
},
{
"epoch": 7.335195530726257,
"grad_norm": 1.0753440753555625,
"learning_rate": 6.572884503658482e-07,
"logits/chosen": -2.8142926692962646,
"logits/rejected": -2.823444128036499,
"logps/chosen": -69.76309967041016,
"logps/rejected": -50.779598236083984,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2833805084228516,
"rewards/margins": 3.2833805084228516,
"rewards/rejected": 0.0,
"step": 1313
},
{
"epoch": 7.340782122905028,
"grad_norm": 2.4490095582853115,
"learning_rate": 6.567097344348889e-07,
"logits/chosen": -3.2057392597198486,
"logits/rejected": -3.4715209007263184,
"logps/chosen": -3.2136125564575195,
"logps/rejected": -28.98649787902832,
"loss": 0.1164,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4721896648406982,
"rewards/margins": 2.4721896648406982,
"rewards/rejected": 0.0,
"step": 1314
},
{
"epoch": 7.346368715083799,
"grad_norm": 2.898508073567845,
"learning_rate": 6.561307856053555e-07,
"logits/chosen": -3.6992645263671875,
"logits/rejected": -3.738304615020752,
"logps/chosen": -8.727750778198242,
"logps/rejected": -50.29473114013672,
"loss": 0.082,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2454607486724854,
"rewards/margins": 3.2454607486724854,
"rewards/rejected": 0.0,
"step": 1315
},
{
"epoch": 7.351955307262569,
"grad_norm": 1.9034543468916336,
"learning_rate": 6.555516047376695e-07,
"logits/chosen": -3.5489768981933594,
"logits/rejected": -3.7010304927825928,
"logps/chosen": -4.087454795837402,
"logps/rejected": -40.70397186279297,
"loss": 0.1156,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.49857234954834,
"rewards/margins": 2.49857234954834,
"rewards/rejected": 0.0,
"step": 1316
},
{
"epoch": 7.35754189944134,
"grad_norm": 7.340890756786072,
"learning_rate": 6.549721926925965e-07,
"logits/chosen": -3.744727849960327,
"logits/rejected": -3.801370620727539,
"logps/chosen": -17.25729751586914,
"logps/rejected": -53.50926971435547,
"loss": 0.1597,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3661885261535645,
"rewards/margins": 2.3661885261535645,
"rewards/rejected": 0.0,
"step": 1317
},
{
"epoch": 7.363128491620111,
"grad_norm": 1.7321282384917218,
"learning_rate": 6.543925503312462e-07,
"logits/chosen": -3.7766597270965576,
"logits/rejected": -3.6987500190734863,
"logps/chosen": -3.585883617401123,
"logps/rejected": -24.525768280029297,
"loss": 0.0963,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4641125202178955,
"rewards/margins": 2.4641125202178955,
"rewards/rejected": 0.0,
"step": 1318
},
{
"epoch": 7.368715083798882,
"grad_norm": 1.643057365136865,
"learning_rate": 6.538126785150704e-07,
"logits/chosen": -3.5576860904693604,
"logits/rejected": -3.6237735748291016,
"logps/chosen": -11.430519104003906,
"logps/rejected": -24.14063262939453,
"loss": 0.1243,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.605529308319092,
"rewards/margins": 3.605529308319092,
"rewards/rejected": 0.0,
"step": 1319
},
{
"epoch": 7.374301675977653,
"grad_norm": 1.0075092519372641,
"learning_rate": 6.532325781058616e-07,
"logits/chosen": -3.627290964126587,
"logits/rejected": -3.6239960193634033,
"logps/chosen": -4.140728950500488,
"logps/rejected": -57.34563446044922,
"loss": 0.0938,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8437869548797607,
"rewards/margins": 2.8437869548797607,
"rewards/rejected": 0.0,
"step": 1320
},
{
"epoch": 7.374301675977653,
"eval_logits/chosen": -3.3619918823242188,
"eval_logits/rejected": -3.4917492866516113,
"eval_logps/chosen": -24.981060028076172,
"eval_logps/rejected": -44.87649917602539,
"eval_loss": 0.6837815046310425,
"eval_rewards/accuracies": 0.7749999761581421,
"eval_rewards/chosen": 0.5088983774185181,
"eval_rewards/margins": 0.5088983774185181,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7419,
"eval_samples_per_second": 9.468,
"eval_steps_per_second": 0.305,
"step": 1320
},
{
"epoch": 7.379888268156424,
"grad_norm": 1.933407383049931,
"learning_rate": 6.526522499657526e-07,
"logits/chosen": -3.636347770690918,
"logits/rejected": -3.7644951343536377,
"logps/chosen": -18.489547729492188,
"logps/rejected": -52.53203582763672,
"loss": 0.0957,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.291066884994507,
"rewards/margins": 3.291066884994507,
"rewards/rejected": 0.0,
"step": 1321
},
{
"epoch": 7.385474860335195,
"grad_norm": 2.543537622248895,
"learning_rate": 6.520716949572142e-07,
"logits/chosen": -3.3230035305023193,
"logits/rejected": -3.4805049896240234,
"logps/chosen": -9.104710578918457,
"logps/rejected": -39.75267791748047,
"loss": 0.1369,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.337641716003418,
"rewards/margins": 3.337641716003418,
"rewards/rejected": 0.0,
"step": 1322
},
{
"epoch": 7.391061452513966,
"grad_norm": 2.669030596207329,
"learning_rate": 6.514909139430548e-07,
"logits/chosen": -3.0569634437561035,
"logits/rejected": -3.1443471908569336,
"logps/chosen": -9.642778396606445,
"logps/rejected": -35.89179992675781,
"loss": 0.1183,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8764891624450684,
"rewards/margins": 2.8764891624450684,
"rewards/rejected": 0.0,
"step": 1323
},
{
"epoch": 7.396648044692737,
"grad_norm": 1.3889452322670262,
"learning_rate": 6.509099077864179e-07,
"logits/chosen": -3.413339853286743,
"logits/rejected": -3.6618871688842773,
"logps/chosen": -3.332961082458496,
"logps/rejected": -51.81547927856445,
"loss": 0.099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3606014251708984,
"rewards/margins": 2.3606014251708984,
"rewards/rejected": 0.0,
"step": 1324
},
{
"epoch": 7.402234636871508,
"grad_norm": 2.696801803963323,
"learning_rate": 6.503286773507828e-07,
"logits/chosen": -3.6457481384277344,
"logits/rejected": -3.445002317428589,
"logps/chosen": -7.384559631347656,
"logps/rejected": -45.16246032714844,
"loss": 0.1266,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4800567626953125,
"rewards/margins": 2.4800567626953125,
"rewards/rejected": 0.0,
"step": 1325
},
{
"epoch": 7.407821229050279,
"grad_norm": 1.676843188205714,
"learning_rate": 6.497472234999608e-07,
"logits/chosen": -3.5677103996276855,
"logits/rejected": -3.4524805545806885,
"logps/chosen": -12.379425048828125,
"logps/rejected": -22.92210578918457,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.429171562194824,
"rewards/margins": 3.429171562194824,
"rewards/rejected": 0.0,
"step": 1326
},
{
"epoch": 7.41340782122905,
"grad_norm": 3.349754471052375,
"learning_rate": 6.491655470980964e-07,
"logits/chosen": -3.7054364681243896,
"logits/rejected": -3.617354154586792,
"logps/chosen": -2.215886116027832,
"logps/rejected": -32.634761810302734,
"loss": 0.1018,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1901493072509766,
"rewards/margins": 2.1901493072509766,
"rewards/rejected": 0.0,
"step": 1327
},
{
"epoch": 7.418994413407821,
"grad_norm": 7.6664406022562614,
"learning_rate": 6.485836490096639e-07,
"logits/chosen": -3.5133492946624756,
"logits/rejected": -3.573883295059204,
"logps/chosen": -2.703808307647705,
"logps/rejected": -37.35500717163086,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.566051483154297,
"rewards/margins": 2.566051483154297,
"rewards/rejected": 0.0,
"step": 1328
},
{
"epoch": 7.424581005586592,
"grad_norm": 0.9687419745063361,
"learning_rate": 6.480015300994677e-07,
"logits/chosen": -2.770904064178467,
"logits/rejected": -2.7978460788726807,
"logps/chosen": -11.93392562866211,
"logps/rejected": -45.412567138671875,
"loss": 0.0909,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1353366374969482,
"rewards/margins": 3.1353366374969482,
"rewards/rejected": 0.0,
"step": 1329
},
{
"epoch": 7.430167597765363,
"grad_norm": 3.5022599679615394,
"learning_rate": 6.474191912326403e-07,
"logits/chosen": -3.3307547569274902,
"logits/rejected": -3.483093738555908,
"logps/chosen": -3.9055068492889404,
"logps/rejected": -44.35377502441406,
"loss": 0.1128,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.005783796310425,
"rewards/margins": 2.005783796310425,
"rewards/rejected": 0.0,
"step": 1330
},
{
"epoch": 7.435754189944134,
"grad_norm": 0.9502362216600457,
"learning_rate": 6.468366332746406e-07,
"logits/chosen": -3.3484418392181396,
"logits/rejected": -3.2991244792938232,
"logps/chosen": -3.163999080657959,
"logps/rejected": -51.34564208984375,
"loss": 0.0864,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.339700222015381,
"rewards/margins": 2.339700222015381,
"rewards/rejected": 0.0,
"step": 1331
},
{
"epoch": 7.441340782122905,
"grad_norm": 1.3659937131735682,
"learning_rate": 6.462538570912538e-07,
"logits/chosen": -3.2325432300567627,
"logits/rejected": -3.4618213176727295,
"logps/chosen": -3.2940423488616943,
"logps/rejected": -40.990081787109375,
"loss": 0.1101,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.100055694580078,
"rewards/margins": 3.100055694580078,
"rewards/rejected": 0.0,
"step": 1332
},
{
"epoch": 7.446927374301676,
"grad_norm": 1.3018900265061804,
"learning_rate": 6.456708635485889e-07,
"logits/chosen": -3.415616750717163,
"logits/rejected": -3.306875228881836,
"logps/chosen": -8.464548110961914,
"logps/rejected": -21.140100479125977,
"loss": 0.0953,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3580121994018555,
"rewards/margins": 2.3580121994018555,
"rewards/rejected": 0.0,
"step": 1333
},
{
"epoch": 7.452513966480447,
"grad_norm": 1.73838771227929,
"learning_rate": 6.450876535130781e-07,
"logits/chosen": -3.4838478565216064,
"logits/rejected": -3.6847469806671143,
"logps/chosen": -1.210036039352417,
"logps/rejected": -52.936614990234375,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.044281005859375,
"rewards/margins": 2.044281005859375,
"rewards/rejected": 0.0,
"step": 1334
},
{
"epoch": 7.4581005586592175,
"grad_norm": 1.3309924063638015,
"learning_rate": 6.445042278514758e-07,
"logits/chosen": -3.348163366317749,
"logits/rejected": -3.305093288421631,
"logps/chosen": -6.772055149078369,
"logps/rejected": -59.42049026489258,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2093491554260254,
"rewards/margins": 2.2093491554260254,
"rewards/rejected": 0.0,
"step": 1335
},
{
"epoch": 7.4636871508379885,
"grad_norm": 1.2217187287545177,
"learning_rate": 6.439205874308558e-07,
"logits/chosen": -3.406967878341675,
"logits/rejected": -3.4981319904327393,
"logps/chosen": -4.282097816467285,
"logps/rejected": -32.48981857299805,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5642776489257812,
"rewards/margins": 2.5642776489257812,
"rewards/rejected": 0.0,
"step": 1336
},
{
"epoch": 7.4692737430167595,
"grad_norm": 1.4537598880879743,
"learning_rate": 6.433367331186121e-07,
"logits/chosen": -3.6480681896209717,
"logits/rejected": -3.6677560806274414,
"logps/chosen": -2.5883102416992188,
"logps/rejected": -42.5275764465332,
"loss": 0.1105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.205169439315796,
"rewards/margins": 2.205169439315796,
"rewards/rejected": 0.0,
"step": 1337
},
{
"epoch": 7.4748603351955305,
"grad_norm": 3.323751503592095,
"learning_rate": 6.427526657824563e-07,
"logits/chosen": -3.100895404815674,
"logits/rejected": -3.221378803253174,
"logps/chosen": -9.436579704284668,
"logps/rejected": -37.93687438964844,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.607145071029663,
"rewards/margins": 2.607145071029663,
"rewards/rejected": 0.0,
"step": 1338
},
{
"epoch": 7.4804469273743015,
"grad_norm": 2.498765371403953,
"learning_rate": 6.421683862904162e-07,
"logits/chosen": -3.614877223968506,
"logits/rejected": -3.744763135910034,
"logps/chosen": -12.693595886230469,
"logps/rejected": -46.41218566894531,
"loss": 0.1353,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1861581802368164,
"rewards/margins": 3.1861581802368164,
"rewards/rejected": 0.0,
"step": 1339
},
{
"epoch": 7.4860335195530725,
"grad_norm": 3.5104236517711866,
"learning_rate": 6.415838955108352e-07,
"logits/chosen": -3.611034631729126,
"logits/rejected": -3.6447675228118896,
"logps/chosen": -3.8875608444213867,
"logps/rejected": -37.52974319458008,
"loss": 0.142,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5034303665161133,
"rewards/margins": 2.5034303665161133,
"rewards/rejected": 0.0,
"step": 1340
},
{
"epoch": 7.4860335195530725,
"eval_logits/chosen": -3.3717868328094482,
"eval_logits/rejected": -3.5007662773132324,
"eval_logps/chosen": -24.902873992919922,
"eval_logps/rejected": -45.34352493286133,
"eval_loss": 0.6902182102203369,
"eval_rewards/accuracies": 0.7749999761581421,
"eval_rewards/chosen": 0.5167169570922852,
"eval_rewards/margins": 0.5167169570922852,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7296,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 1340
},
{
"epoch": 7.4916201117318435,
"grad_norm": 1.9746337321102174,
"learning_rate": 6.409991943123708e-07,
"logits/chosen": -3.668294906616211,
"logits/rejected": -3.6877212524414062,
"logps/chosen": -9.687832832336426,
"logps/rejected": -51.59500503540039,
"loss": 0.1021,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5231974124908447,
"rewards/margins": 3.5231974124908447,
"rewards/rejected": 0.0,
"step": 1341
},
{
"epoch": 7.4972067039106145,
"grad_norm": 3.10042766484964,
"learning_rate": 6.40414283563993e-07,
"logits/chosen": -3.335176706314087,
"logits/rejected": -3.398043632507324,
"logps/chosen": -29.620380401611328,
"logps/rejected": -28.167808532714844,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.102391481399536,
"rewards/margins": 3.102391481399536,
"rewards/rejected": 0.0,
"step": 1342
},
{
"epoch": 7.5027932960893855,
"grad_norm": 1.8660695386987614,
"learning_rate": 6.398291641349835e-07,
"logits/chosen": -3.3379101753234863,
"logits/rejected": -3.1388955116271973,
"logps/chosen": -3.3202102184295654,
"logps/rejected": -45.3973388671875,
"loss": 0.1241,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2802844047546387,
"rewards/margins": 2.2802844047546387,
"rewards/rejected": 0.0,
"step": 1343
},
{
"epoch": 7.5083798882681565,
"grad_norm": 1.118053151884192,
"learning_rate": 6.392438368949338e-07,
"logits/chosen": -3.490548849105835,
"logits/rejected": -3.5045557022094727,
"logps/chosen": -17.330812454223633,
"logps/rejected": -46.646480560302734,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.551020622253418,
"rewards/margins": 2.551020622253418,
"rewards/rejected": 0.0,
"step": 1344
},
{
"epoch": 7.5139664804469275,
"grad_norm": 1.5846715669566198,
"learning_rate": 6.386583027137446e-07,
"logits/chosen": -3.4728899002075195,
"logits/rejected": -3.342719078063965,
"logps/chosen": -6.973177909851074,
"logps/rejected": -72.66995239257812,
"loss": 0.121,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5646984577178955,
"rewards/margins": 2.5646984577178955,
"rewards/rejected": 0.0,
"step": 1345
},
{
"epoch": 7.5195530726256985,
"grad_norm": 2.014230838623146,
"learning_rate": 6.380725624616239e-07,
"logits/chosen": -3.4623639583587646,
"logits/rejected": -3.6248762607574463,
"logps/chosen": -7.952892303466797,
"logps/rejected": -51.53924560546875,
"loss": 0.1505,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5686726570129395,
"rewards/margins": 2.5686726570129395,
"rewards/rejected": 0.0,
"step": 1346
},
{
"epoch": 7.5251396648044695,
"grad_norm": 1.158398029141674,
"learning_rate": 6.374866170090858e-07,
"logits/chosen": -3.652508020401001,
"logits/rejected": -3.475364923477173,
"logps/chosen": -6.89300012588501,
"logps/rejected": -44.19477844238281,
"loss": 0.1357,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.634492874145508,
"rewards/margins": 2.634492874145508,
"rewards/rejected": 0.0,
"step": 1347
},
{
"epoch": 7.5307262569832405,
"grad_norm": 1.2878840045831155,
"learning_rate": 6.3690046722695e-07,
"logits/chosen": -3.477654457092285,
"logits/rejected": -3.5552642345428467,
"logps/chosen": -9.743959426879883,
"logps/rejected": -57.691184997558594,
"loss": 0.1222,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.865281581878662,
"rewards/margins": 2.865281581878662,
"rewards/rejected": 0.0,
"step": 1348
},
{
"epoch": 7.5363128491620115,
"grad_norm": 2.260138463270068,
"learning_rate": 6.363141139863393e-07,
"logits/chosen": -3.523780584335327,
"logits/rejected": -3.8094823360443115,
"logps/chosen": -2.557177782058716,
"logps/rejected": -60.25283432006836,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2795915603637695,
"rewards/margins": 2.2795915603637695,
"rewards/rejected": 0.0,
"step": 1349
},
{
"epoch": 7.5418994413407825,
"grad_norm": 3.1029917301525454,
"learning_rate": 6.357275581586787e-07,
"logits/chosen": -3.578190565109253,
"logits/rejected": -3.648209571838379,
"logps/chosen": -5.742236137390137,
"logps/rejected": -31.89114761352539,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.640815496444702,
"rewards/margins": 2.640815496444702,
"rewards/rejected": 0.0,
"step": 1350
},
{
"epoch": 7.547486033519553,
"grad_norm": 1.8185286470035993,
"learning_rate": 6.351408006156952e-07,
"logits/chosen": -3.3877675533294678,
"logits/rejected": -3.6783602237701416,
"logps/chosen": -4.36691951751709,
"logps/rejected": -93.2117691040039,
"loss": 0.1141,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9463605880737305,
"rewards/margins": 1.9463605880737305,
"rewards/rejected": 0.0,
"step": 1351
},
{
"epoch": 7.553072625698324,
"grad_norm": 1.4962646551008194,
"learning_rate": 6.345538422294146e-07,
"logits/chosen": -3.5849568843841553,
"logits/rejected": -3.41233491897583,
"logps/chosen": -8.534076690673828,
"logps/rejected": -20.666913986206055,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.941847801208496,
"rewards/margins": 2.941847801208496,
"rewards/rejected": 0.0,
"step": 1352
},
{
"epoch": 7.558659217877095,
"grad_norm": 3.0771917352843436,
"learning_rate": 6.339666838721618e-07,
"logits/chosen": -3.6744654178619385,
"logits/rejected": -3.5337159633636475,
"logps/chosen": -6.005249500274658,
"logps/rejected": -44.79070281982422,
"loss": 0.1613,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5271992683410645,
"rewards/margins": 2.5271992683410645,
"rewards/rejected": 0.0,
"step": 1353
},
{
"epoch": 7.564245810055866,
"grad_norm": 1.4739657920602398,
"learning_rate": 6.333793264165585e-07,
"logits/chosen": -3.5482747554779053,
"logits/rejected": -3.4847331047058105,
"logps/chosen": -2.1421570777893066,
"logps/rejected": -73.30380249023438,
"loss": 0.0898,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2679672241210938,
"rewards/margins": 2.2679672241210938,
"rewards/rejected": 0.0,
"step": 1354
},
{
"epoch": 7.569832402234637,
"grad_norm": 3.218194655153971,
"learning_rate": 6.327917707355227e-07,
"logits/chosen": -3.604827642440796,
"logits/rejected": -3.4153332710266113,
"logps/chosen": -2.7689061164855957,
"logps/rejected": -67.11903381347656,
"loss": 0.1253,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2885615825653076,
"rewards/margins": 2.2885615825653076,
"rewards/rejected": 0.0,
"step": 1355
},
{
"epoch": 7.575418994413408,
"grad_norm": 2.9586776277300273,
"learning_rate": 6.322040177022666e-07,
"logits/chosen": -3.6942250728607178,
"logits/rejected": -3.5640459060668945,
"logps/chosen": -10.615409851074219,
"logps/rejected": -36.34138488769531,
"loss": 0.117,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.105916976928711,
"rewards/margins": 3.105916976928711,
"rewards/rejected": 0.0,
"step": 1356
},
{
"epoch": 7.581005586592179,
"grad_norm": 1.2298329754678976,
"learning_rate": 6.316160681902959e-07,
"logits/chosen": -3.546031951904297,
"logits/rejected": -3.579220771789551,
"logps/chosen": -2.216337203979492,
"logps/rejected": -45.56477355957031,
"loss": 0.107,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4204089641571045,
"rewards/margins": 2.4204089641571045,
"rewards/rejected": 0.0,
"step": 1357
},
{
"epoch": 7.58659217877095,
"grad_norm": 1.1688780139728336,
"learning_rate": 6.310279230734084e-07,
"logits/chosen": -3.5705759525299072,
"logits/rejected": -3.547123908996582,
"logps/chosen": -3.538571357727051,
"logps/rejected": -37.826637268066406,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.537139415740967,
"rewards/margins": 2.537139415740967,
"rewards/rejected": 0.0,
"step": 1358
},
{
"epoch": 7.592178770949721,
"grad_norm": 5.304963121125634,
"learning_rate": 6.304395832256925e-07,
"logits/chosen": -3.28761625289917,
"logits/rejected": -3.271043539047241,
"logps/chosen": -4.488574504852295,
"logps/rejected": -73.1676025390625,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.421415328979492,
"rewards/margins": 2.421415328979492,
"rewards/rejected": 0.0,
"step": 1359
},
{
"epoch": 7.597765363128492,
"grad_norm": 2.023221321031377,
"learning_rate": 6.298510495215258e-07,
"logits/chosen": -3.6465795040130615,
"logits/rejected": -3.650517225265503,
"logps/chosen": -5.583526611328125,
"logps/rejected": -19.741701126098633,
"loss": 0.127,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5471010208129883,
"rewards/margins": 2.5471010208129883,
"rewards/rejected": 0.0,
"step": 1360
},
{
"epoch": 7.597765363128492,
"eval_logits/chosen": -3.377352476119995,
"eval_logits/rejected": -3.5058913230895996,
"eval_logps/chosen": -24.613401412963867,
"eval_logps/rejected": -45.38679122924805,
"eval_loss": 0.6831362247467041,
"eval_rewards/accuracies": 0.7749999761581421,
"eval_rewards/chosen": 0.5456639528274536,
"eval_rewards/margins": 0.5456639528274536,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6911,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 1360
},
{
"epoch": 7.603351955307263,
"grad_norm": 1.8617615502700466,
"learning_rate": 6.292623228355742e-07,
"logits/chosen": -3.835742712020874,
"logits/rejected": -3.8247580528259277,
"logps/chosen": -17.866558074951172,
"logps/rejected": -32.8570442199707,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.352844715118408,
"rewards/margins": 3.352844715118408,
"rewards/rejected": 0.0,
"step": 1361
},
{
"epoch": 7.608938547486034,
"grad_norm": 1.1518435267353249,
"learning_rate": 6.286734040427908e-07,
"logits/chosen": -3.4593331813812256,
"logits/rejected": -3.492604970932007,
"logps/chosen": -7.050579071044922,
"logps/rejected": -48.49711608886719,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9449377059936523,
"rewards/margins": 2.9449377059936523,
"rewards/rejected": 0.0,
"step": 1362
},
{
"epoch": 7.614525139664805,
"grad_norm": 2.0091735988016413,
"learning_rate": 6.280842940184132e-07,
"logits/chosen": -3.2921576499938965,
"logits/rejected": -3.3532657623291016,
"logps/chosen": -4.33574104309082,
"logps/rejected": -46.30808639526367,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.816415309906006,
"rewards/margins": 2.816415309906006,
"rewards/rejected": 0.0,
"step": 1363
},
{
"epoch": 7.620111731843576,
"grad_norm": 6.099087823168098,
"learning_rate": 6.274949936379643e-07,
"logits/chosen": -3.1784050464630127,
"logits/rejected": -3.2182962894439697,
"logps/chosen": -18.77121353149414,
"logps/rejected": -33.73542022705078,
"loss": 0.1596,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.436699628829956,
"rewards/margins": 2.436699628829956,
"rewards/rejected": 0.0,
"step": 1364
},
{
"epoch": 7.625698324022347,
"grad_norm": 1.6373569972408615,
"learning_rate": 6.269055037772491e-07,
"logits/chosen": -3.4280288219451904,
"logits/rejected": -3.6520519256591797,
"logps/chosen": -1.1923744678497314,
"logps/rejected": -47.18501281738281,
"loss": 0.1218,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9932301044464111,
"rewards/margins": 1.9932301044464111,
"rewards/rejected": 0.0,
"step": 1365
},
{
"epoch": 7.631284916201118,
"grad_norm": 2.7189590731815687,
"learning_rate": 6.263158253123547e-07,
"logits/chosen": -3.5601494312286377,
"logits/rejected": -3.5052757263183594,
"logps/chosen": -9.23727035522461,
"logps/rejected": -65.1521224975586,
"loss": 0.1183,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8892292976379395,
"rewards/margins": 2.8892292976379395,
"rewards/rejected": 0.0,
"step": 1366
},
{
"epoch": 7.636871508379889,
"grad_norm": 5.333570060425901,
"learning_rate": 6.257259591196484e-07,
"logits/chosen": -3.7892396450042725,
"logits/rejected": -3.836857318878174,
"logps/chosen": -14.42813491821289,
"logps/rejected": -40.66142272949219,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.175755500793457,
"rewards/margins": 2.175755500793457,
"rewards/rejected": 0.0,
"step": 1367
},
{
"epoch": 7.64245810055866,
"grad_norm": 1.5204927438185927,
"learning_rate": 6.251359060757762e-07,
"logits/chosen": -3.2748348712921143,
"logits/rejected": -3.301457643508911,
"logps/chosen": -1.2259882688522339,
"logps/rejected": -36.65496063232422,
"loss": 0.1097,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.735575795173645,
"rewards/margins": 1.735575795173645,
"rewards/rejected": 0.0,
"step": 1368
},
{
"epoch": 7.648044692737431,
"grad_norm": 3.374061606176809,
"learning_rate": 6.245456670576621e-07,
"logits/chosen": -3.1197638511657715,
"logits/rejected": -3.1260502338409424,
"logps/chosen": -26.035179138183594,
"logps/rejected": -49.376625061035156,
"loss": 0.1317,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.248068332672119,
"rewards/margins": 2.248068332672119,
"rewards/rejected": 0.0,
"step": 1369
},
{
"epoch": 7.653631284916202,
"grad_norm": 7.988430551598308,
"learning_rate": 6.239552429425063e-07,
"logits/chosen": -3.3330259323120117,
"logits/rejected": -3.3822152614593506,
"logps/chosen": -5.702103614807129,
"logps/rejected": -44.26451873779297,
"loss": 0.1481,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.56434965133667,
"rewards/margins": 2.56434965133667,
"rewards/rejected": 0.0,
"step": 1370
},
{
"epoch": 7.659217877094972,
"grad_norm": 2.0178085734716706,
"learning_rate": 6.233646346077844e-07,
"logits/chosen": -3.6526939868927,
"logits/rejected": -3.6671154499053955,
"logps/chosen": -4.053013801574707,
"logps/rejected": -42.36505889892578,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2508602142333984,
"rewards/margins": 2.2508602142333984,
"rewards/rejected": 0.0,
"step": 1371
},
{
"epoch": 7.664804469273743,
"grad_norm": 4.93361389110381,
"learning_rate": 6.227738429312455e-07,
"logits/chosen": -3.4513962268829346,
"logits/rejected": -3.3486549854278564,
"logps/chosen": -3.926506996154785,
"logps/rejected": -51.46466064453125,
"loss": 0.1274,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7176238298416138,
"rewards/margins": 1.7176238298416138,
"rewards/rejected": 0.0,
"step": 1372
},
{
"epoch": 7.670391061452514,
"grad_norm": 1.2575238577960888,
"learning_rate": 6.221828687909108e-07,
"logits/chosen": -3.532562732696533,
"logits/rejected": -3.596668004989624,
"logps/chosen": -3.615190267562866,
"logps/rejected": -54.00471115112305,
"loss": 0.1243,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.430891990661621,
"rewards/margins": 2.430891990661621,
"rewards/rejected": 0.0,
"step": 1373
},
{
"epoch": 7.675977653631285,
"grad_norm": 2.0991029561289083,
"learning_rate": 6.215917130650739e-07,
"logits/chosen": -3.5558907985687256,
"logits/rejected": -3.493356704711914,
"logps/chosen": -2.2038702964782715,
"logps/rejected": -34.609466552734375,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.418152332305908,
"rewards/margins": 2.418152332305908,
"rewards/rejected": 0.0,
"step": 1374
},
{
"epoch": 7.681564245810056,
"grad_norm": 1.631611969715661,
"learning_rate": 6.210003766322969e-07,
"logits/chosen": -3.4739394187927246,
"logits/rejected": -3.3518495559692383,
"logps/chosen": -14.285957336425781,
"logps/rejected": -38.5153923034668,
"loss": 0.1243,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2559762001037598,
"rewards/margins": 3.2559762001037598,
"rewards/rejected": 0.0,
"step": 1375
},
{
"epoch": 7.687150837988827,
"grad_norm": 2.013490296619291,
"learning_rate": 6.204088603714111e-07,
"logits/chosen": -3.411038398742676,
"logits/rejected": -3.455512762069702,
"logps/chosen": -2.3266758918762207,
"logps/rejected": -44.296722412109375,
"loss": 0.1025,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3725152015686035,
"rewards/margins": 2.3725152015686035,
"rewards/rejected": 0.0,
"step": 1376
},
{
"epoch": 7.692737430167598,
"grad_norm": 2.404214334111452,
"learning_rate": 6.198171651615154e-07,
"logits/chosen": -3.2101945877075195,
"logits/rejected": -3.4877567291259766,
"logps/chosen": -2.0130977630615234,
"logps/rejected": -56.4400634765625,
"loss": 0.1071,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.49690580368042,
"rewards/margins": 1.49690580368042,
"rewards/rejected": 0.0,
"step": 1377
},
{
"epoch": 7.698324022346369,
"grad_norm": 12.290736935137124,
"learning_rate": 6.192252918819739e-07,
"logits/chosen": -3.4152767658233643,
"logits/rejected": -3.3613784313201904,
"logps/chosen": -26.862014770507812,
"logps/rejected": -40.26948547363281,
"loss": 0.1156,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6470947265625,
"rewards/margins": 3.6470947265625,
"rewards/rejected": 0.0,
"step": 1378
},
{
"epoch": 7.70391061452514,
"grad_norm": 2.6495774255832147,
"learning_rate": 6.186332414124159e-07,
"logits/chosen": -2.77083158493042,
"logits/rejected": -2.869678020477295,
"logps/chosen": -6.970669746398926,
"logps/rejected": -46.74264907836914,
"loss": 0.1035,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.762322425842285,
"rewards/margins": 2.762322425842285,
"rewards/rejected": 0.0,
"step": 1379
},
{
"epoch": 7.709497206703911,
"grad_norm": 1.446488271803958,
"learning_rate": 6.180410146327336e-07,
"logits/chosen": -3.3112292289733887,
"logits/rejected": -3.009565830230713,
"logps/chosen": -11.268660545349121,
"logps/rejected": -48.60431671142578,
"loss": 0.0918,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6027472019195557,
"rewards/margins": 2.6027472019195557,
"rewards/rejected": 0.0,
"step": 1380
},
{
"epoch": 7.709497206703911,
"eval_logits/chosen": -3.383289337158203,
"eval_logits/rejected": -3.508803129196167,
"eval_logps/chosen": -24.955305099487305,
"eval_logps/rejected": -45.27484893798828,
"eval_loss": 0.6841699481010437,
"eval_rewards/accuracies": 0.7749999761581421,
"eval_rewards/chosen": 0.5114735960960388,
"eval_rewards/margins": 0.5114735960960388,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6665,
"eval_samples_per_second": 9.49,
"eval_steps_per_second": 0.306,
"step": 1380
},
{
"epoch": 7.715083798882682,
"grad_norm": 2.171345951728274,
"learning_rate": 6.174486124230817e-07,
"logits/chosen": -3.494107961654663,
"logits/rejected": -3.584721803665161,
"logps/chosen": -1.2909730672836304,
"logps/rejected": -39.61445236206055,
"loss": 0.1355,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7321256399154663,
"rewards/margins": 1.7321256399154663,
"rewards/rejected": 0.0,
"step": 1381
},
{
"epoch": 7.720670391061453,
"grad_norm": 2.1896169795722726,
"learning_rate": 6.168560356638751e-07,
"logits/chosen": -3.6455185413360596,
"logits/rejected": -3.780693531036377,
"logps/chosen": -8.931732177734375,
"logps/rejected": -32.342994689941406,
"loss": 0.097,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.001253128051758,
"rewards/margins": 3.001253128051758,
"rewards/rejected": 0.0,
"step": 1382
},
{
"epoch": 7.726256983240224,
"grad_norm": 1.1964830984574129,
"learning_rate": 6.162632852357887e-07,
"logits/chosen": -3.446711540222168,
"logits/rejected": -3.3967432975769043,
"logps/chosen": -8.087589263916016,
"logps/rejected": -30.408592224121094,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4150784015655518,
"rewards/margins": 2.4150784015655518,
"rewards/rejected": 0.0,
"step": 1383
},
{
"epoch": 7.731843575418995,
"grad_norm": 1.4030991044792858,
"learning_rate": 6.156703620197552e-07,
"logits/chosen": -3.4570517539978027,
"logits/rejected": -3.558060884475708,
"logps/chosen": -4.256320953369141,
"logps/rejected": -38.700252532958984,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5645179748535156,
"rewards/margins": 2.5645179748535156,
"rewards/rejected": 0.0,
"step": 1384
},
{
"epoch": 7.737430167597766,
"grad_norm": 2.858857863046088,
"learning_rate": 6.150772668969639e-07,
"logits/chosen": -3.7470078468322754,
"logits/rejected": -3.738076686859131,
"logps/chosen": -14.466458320617676,
"logps/rejected": -39.010990142822266,
"loss": 0.1212,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.303770065307617,
"rewards/margins": 3.303770065307617,
"rewards/rejected": 0.0,
"step": 1385
},
{
"epoch": 7.743016759776537,
"grad_norm": 1.8436366580988155,
"learning_rate": 6.144840007488598e-07,
"logits/chosen": -3.9643008708953857,
"logits/rejected": -3.5621531009674072,
"logps/chosen": -3.323282480239868,
"logps/rejected": -33.563270568847656,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2505550384521484,
"rewards/margins": 2.2505550384521484,
"rewards/rejected": 0.0,
"step": 1386
},
{
"epoch": 7.748603351955307,
"grad_norm": 33.18464451137197,
"learning_rate": 6.13890564457142e-07,
"logits/chosen": -3.5103845596313477,
"logits/rejected": -3.5708584785461426,
"logps/chosen": -5.393838882446289,
"logps/rejected": -81.33026123046875,
"loss": 0.1234,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.862974166870117,
"rewards/margins": 2.862974166870117,
"rewards/rejected": 0.0,
"step": 1387
},
{
"epoch": 7.754189944134078,
"grad_norm": 2.6710716417733202,
"learning_rate": 6.132969589037629e-07,
"logits/chosen": -3.633074998855591,
"logits/rejected": -3.5671002864837646,
"logps/chosen": -2.210402727127075,
"logps/rejected": -37.03133010864258,
"loss": 0.1126,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8473315238952637,
"rewards/margins": 1.8473315238952637,
"rewards/rejected": 0.0,
"step": 1388
},
{
"epoch": 7.759776536312849,
"grad_norm": 1.9040518769852222,
"learning_rate": 6.127031849709256e-07,
"logits/chosen": -3.615497350692749,
"logits/rejected": -3.615487813949585,
"logps/chosen": -8.555794715881348,
"logps/rejected": -52.548377990722656,
"loss": 0.1161,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9727935791015625,
"rewards/margins": 2.9727935791015625,
"rewards/rejected": 0.0,
"step": 1389
},
{
"epoch": 7.76536312849162,
"grad_norm": 3.1192216970559175,
"learning_rate": 6.12109243541084e-07,
"logits/chosen": -3.539001703262329,
"logits/rejected": -3.648571491241455,
"logps/chosen": -1.8654552698135376,
"logps/rejected": -49.581146240234375,
"loss": 0.1286,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5140178203582764,
"rewards/margins": 2.5140178203582764,
"rewards/rejected": 0.0,
"step": 1390
},
{
"epoch": 7.770949720670391,
"grad_norm": 1.357172241819613,
"learning_rate": 6.115151354969413e-07,
"logits/chosen": -3.668170690536499,
"logits/rejected": -3.449744939804077,
"logps/chosen": -17.671070098876953,
"logps/rejected": -23.16139793395996,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8469767570495605,
"rewards/margins": 2.8469767570495605,
"rewards/rejected": 0.0,
"step": 1391
},
{
"epoch": 7.776536312849162,
"grad_norm": 1.929706858236694,
"learning_rate": 6.10920861721447e-07,
"logits/chosen": -3.673409938812256,
"logits/rejected": -3.611703872680664,
"logps/chosen": -1.868884563446045,
"logps/rejected": -102.60816955566406,
"loss": 0.1323,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1316471099853516,
"rewards/margins": 2.1316471099853516,
"rewards/rejected": 0.0,
"step": 1392
},
{
"epoch": 7.782122905027933,
"grad_norm": 4.120154268907915,
"learning_rate": 6.103264230977986e-07,
"logits/chosen": -3.464028835296631,
"logits/rejected": -3.5241379737854004,
"logps/chosen": -8.902050018310547,
"logps/rejected": -28.881071090698242,
"loss": 0.1233,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9710700511932373,
"rewards/margins": 2.9710700511932373,
"rewards/rejected": 0.0,
"step": 1393
},
{
"epoch": 7.787709497206704,
"grad_norm": 2.72460910417234,
"learning_rate": 6.097318205094373e-07,
"logits/chosen": -3.392451286315918,
"logits/rejected": -3.666982412338257,
"logps/chosen": -13.357746124267578,
"logps/rejected": -72.52420806884766,
"loss": 0.0985,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2183361053466797,
"rewards/margins": 2.2183361053466797,
"rewards/rejected": 0.0,
"step": 1394
},
{
"epoch": 7.793296089385475,
"grad_norm": 1.9192607441144842,
"learning_rate": 6.091370548400485e-07,
"logits/chosen": -3.409224033355713,
"logits/rejected": -3.3265554904937744,
"logps/chosen": -1.9528727531433105,
"logps/rejected": -83.61040496826172,
"loss": 0.1048,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0228054523468018,
"rewards/margins": 2.0228054523468018,
"rewards/rejected": 0.0,
"step": 1395
},
{
"epoch": 7.798882681564246,
"grad_norm": 1.867699872620725,
"learning_rate": 6.0854212697356e-07,
"logits/chosen": -3.6545894145965576,
"logits/rejected": -3.3519484996795654,
"logps/chosen": -7.546158790588379,
"logps/rejected": -30.187633514404297,
"loss": 0.116,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.773406982421875,
"rewards/margins": 2.773406982421875,
"rewards/rejected": 0.0,
"step": 1396
},
{
"epoch": 7.804469273743017,
"grad_norm": 2.2953090984405846,
"learning_rate": 6.079470377941406e-07,
"logits/chosen": -3.3122012615203857,
"logits/rejected": -3.3559436798095703,
"logps/chosen": -3.9074928760528564,
"logps/rejected": -51.002899169921875,
"loss": 0.1215,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5583930015563965,
"rewards/margins": 2.5583930015563965,
"rewards/rejected": 0.0,
"step": 1397
},
{
"epoch": 7.810055865921788,
"grad_norm": 1.544662917361502,
"learning_rate": 6.073517881861987e-07,
"logits/chosen": -3.674224615097046,
"logits/rejected": -3.7264983654022217,
"logps/chosen": -4.493802547454834,
"logps/rejected": -24.393909454345703,
"loss": 0.0847,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.688265562057495,
"rewards/margins": 2.688265562057495,
"rewards/rejected": 0.0,
"step": 1398
},
{
"epoch": 7.815642458100559,
"grad_norm": 1.439069885695365,
"learning_rate": 6.067563790343813e-07,
"logits/chosen": -3.4475038051605225,
"logits/rejected": -3.591827869415283,
"logps/chosen": -5.838848114013672,
"logps/rejected": -49.471961975097656,
"loss": 0.1321,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5757193565368652,
"rewards/margins": 2.5757193565368652,
"rewards/rejected": 0.0,
"step": 1399
},
{
"epoch": 7.82122905027933,
"grad_norm": 1.1246715943382999,
"learning_rate": 6.061608112235723e-07,
"logits/chosen": -3.374066114425659,
"logits/rejected": -3.497426986694336,
"logps/chosen": -1.7087607383728027,
"logps/rejected": -63.19783020019531,
"loss": 0.1178,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7700719833374023,
"rewards/margins": 1.7700719833374023,
"rewards/rejected": 0.0,
"step": 1400
},
{
"epoch": 7.82122905027933,
"eval_logits/chosen": -3.366156816482544,
"eval_logits/rejected": -3.4939181804656982,
"eval_logps/chosen": -25.586759567260742,
"eval_logps/rejected": -46.298309326171875,
"eval_loss": 0.7026585936546326,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.4483281075954437,
"eval_rewards/margins": 0.4483281075954437,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.698,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 1400
},
{
"epoch": 7.826815642458101,
"grad_norm": 2.260126420189014,
"learning_rate": 6.055650856388916e-07,
"logits/chosen": -3.5005438327789307,
"logits/rejected": -3.439100742340088,
"logps/chosen": -1.6743199825286865,
"logps/rejected": -31.64121437072754,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0029587745666504,
"rewards/margins": 2.0029587745666504,
"rewards/rejected": 0.0,
"step": 1401
},
{
"epoch": 7.832402234636872,
"grad_norm": 3.736988949502141,
"learning_rate": 6.049692031656934e-07,
"logits/chosen": -3.4197769165039062,
"logits/rejected": -3.3480851650238037,
"logps/chosen": -6.497530460357666,
"logps/rejected": -40.690635681152344,
"loss": 0.1166,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5752921104431152,
"rewards/margins": 2.5752921104431152,
"rewards/rejected": 0.0,
"step": 1402
},
{
"epoch": 7.837988826815643,
"grad_norm": 4.0719521366359785,
"learning_rate": 6.043731646895655e-07,
"logits/chosen": -3.29199481010437,
"logits/rejected": -3.2787442207336426,
"logps/chosen": -4.126980304718018,
"logps/rejected": -25.43942642211914,
"loss": 0.1306,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.086606025695801,
"rewards/margins": 2.086606025695801,
"rewards/rejected": 0.0,
"step": 1403
},
{
"epoch": 7.843575418994414,
"grad_norm": 2.2633841595929978,
"learning_rate": 6.037769710963271e-07,
"logits/chosen": -3.43090558052063,
"logits/rejected": -3.5747225284576416,
"logps/chosen": -4.280132293701172,
"logps/rejected": -33.68571472167969,
"loss": 0.1028,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.464203119277954,
"rewards/margins": 2.464203119277954,
"rewards/rejected": 0.0,
"step": 1404
},
{
"epoch": 7.849162011173185,
"grad_norm": 1.9929313553569437,
"learning_rate": 6.031806232720277e-07,
"logits/chosen": -3.09665846824646,
"logits/rejected": -3.1195924282073975,
"logps/chosen": -32.372947692871094,
"logps/rejected": -27.54912567138672,
"loss": 0.1052,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.689082622528076,
"rewards/margins": 3.689082622528076,
"rewards/rejected": 0.0,
"step": 1405
},
{
"epoch": 7.854748603351956,
"grad_norm": 1.611789802544927,
"learning_rate": 6.025841221029469e-07,
"logits/chosen": -3.410153865814209,
"logits/rejected": -3.562817335128784,
"logps/chosen": -3.033844470977783,
"logps/rejected": -54.31105041503906,
"loss": 0.0946,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2825844287872314,
"rewards/margins": 2.2825844287872314,
"rewards/rejected": 0.0,
"step": 1406
},
{
"epoch": 7.860335195530726,
"grad_norm": 6.442979724468341,
"learning_rate": 6.019874684755913e-07,
"logits/chosen": -3.4948666095733643,
"logits/rejected": -3.6496846675872803,
"logps/chosen": -4.0476837158203125,
"logps/rejected": -115.38198852539062,
"loss": 0.1469,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.433524489402771,
"rewards/margins": 1.433524489402771,
"rewards/rejected": 0.0,
"step": 1407
},
{
"epoch": 7.865921787709497,
"grad_norm": 2.789367083541046,
"learning_rate": 6.013906632766946e-07,
"logits/chosen": -3.7730135917663574,
"logits/rejected": -3.7376444339752197,
"logps/chosen": -4.7420172691345215,
"logps/rejected": -45.96259307861328,
"loss": 0.0994,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.087348461151123,
"rewards/margins": 3.087348461151123,
"rewards/rejected": 0.0,
"step": 1408
},
{
"epoch": 7.871508379888268,
"grad_norm": 1.6642517172457674,
"learning_rate": 6.007937073932156e-07,
"logits/chosen": -3.703899621963501,
"logits/rejected": -3.7811319828033447,
"logps/chosen": -2.2926485538482666,
"logps/rejected": -42.525848388671875,
"loss": 0.1109,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5533089637756348,
"rewards/margins": 2.5533089637756348,
"rewards/rejected": 0.0,
"step": 1409
},
{
"epoch": 7.877094972067039,
"grad_norm": 2.5132314691328,
"learning_rate": 6.001966017123371e-07,
"logits/chosen": -3.6104767322540283,
"logits/rejected": -3.619384765625,
"logps/chosen": -1.8226158618927002,
"logps/rejected": -48.86735534667969,
"loss": 0.1258,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5055662393569946,
"rewards/margins": 1.5055662393569946,
"rewards/rejected": 0.0,
"step": 1410
},
{
"epoch": 7.88268156424581,
"grad_norm": 3.5448643787501206,
"learning_rate": 5.995993471214644e-07,
"logits/chosen": -3.394293785095215,
"logits/rejected": -3.4889090061187744,
"logps/chosen": -21.46180534362793,
"logps/rejected": -35.470645904541016,
"loss": 0.1531,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3129191398620605,
"rewards/margins": 3.3129191398620605,
"rewards/rejected": 0.0,
"step": 1411
},
{
"epoch": 7.888268156424581,
"grad_norm": 2.71436756552808,
"learning_rate": 5.990019445082241e-07,
"logits/chosen": -3.2157909870147705,
"logits/rejected": -3.5748908519744873,
"logps/chosen": -8.530254364013672,
"logps/rejected": -38.43312072753906,
"loss": 0.1243,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5410425662994385,
"rewards/margins": 2.5410425662994385,
"rewards/rejected": 0.0,
"step": 1412
},
{
"epoch": 7.893854748603352,
"grad_norm": 3.4204042283735383,
"learning_rate": 5.98404394760463e-07,
"logits/chosen": -3.390915870666504,
"logits/rejected": -3.157789945602417,
"logps/chosen": -10.765122413635254,
"logps/rejected": -43.94075393676758,
"loss": 0.1314,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5835838317871094,
"rewards/margins": 2.5835838317871094,
"rewards/rejected": 0.0,
"step": 1413
},
{
"epoch": 7.899441340782123,
"grad_norm": 2.5722225871384077,
"learning_rate": 5.978066987662464e-07,
"logits/chosen": -3.525294542312622,
"logits/rejected": -3.432190418243408,
"logps/chosen": -8.895794868469238,
"logps/rejected": -32.79235076904297,
"loss": 0.118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.473673105239868,
"rewards/margins": 2.473673105239868,
"rewards/rejected": 0.0,
"step": 1414
},
{
"epoch": 7.905027932960894,
"grad_norm": 2.356667195650094,
"learning_rate": 5.97208857413857e-07,
"logits/chosen": -3.508132219314575,
"logits/rejected": -3.1443114280700684,
"logps/chosen": -3.209148645401001,
"logps/rejected": -38.9202995300293,
"loss": 0.103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4470789432525635,
"rewards/margins": 2.4470789432525635,
"rewards/rejected": 0.0,
"step": 1415
},
{
"epoch": 7.910614525139665,
"grad_norm": 2.4473535198646243,
"learning_rate": 5.966108715917937e-07,
"logits/chosen": -3.510397434234619,
"logits/rejected": -3.4470112323760986,
"logps/chosen": -7.509511947631836,
"logps/rejected": -74.8454360961914,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.598639965057373,
"rewards/margins": 2.598639965057373,
"rewards/rejected": 0.0,
"step": 1416
},
{
"epoch": 7.916201117318436,
"grad_norm": 2.4295257304259916,
"learning_rate": 5.960127421887696e-07,
"logits/chosen": -3.386652946472168,
"logits/rejected": -3.4341964721679688,
"logps/chosen": -11.838761329650879,
"logps/rejected": -29.63746452331543,
"loss": 0.13,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.595767021179199,
"rewards/margins": 2.595767021179199,
"rewards/rejected": 0.0,
"step": 1417
},
{
"epoch": 7.921787709497207,
"grad_norm": 2.2345989719667037,
"learning_rate": 5.954144700937118e-07,
"logits/chosen": -3.4850287437438965,
"logits/rejected": -3.368241548538208,
"logps/chosen": -14.28183364868164,
"logps/rejected": -23.823863983154297,
"loss": 0.1171,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.896121025085449,
"rewards/margins": 2.896121025085449,
"rewards/rejected": 0.0,
"step": 1418
},
{
"epoch": 7.927374301675978,
"grad_norm": 3.0241236298958203,
"learning_rate": 5.948160561957591e-07,
"logits/chosen": -3.1309311389923096,
"logits/rejected": -3.131141185760498,
"logps/chosen": -3.4051127433776855,
"logps/rejected": -28.47091293334961,
"loss": 0.1267,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3636093139648438,
"rewards/margins": 2.3636093139648438,
"rewards/rejected": 0.0,
"step": 1419
},
{
"epoch": 7.932960893854749,
"grad_norm": 2.158676825979555,
"learning_rate": 5.942175013842609e-07,
"logits/chosen": -3.1837263107299805,
"logits/rejected": -3.4182281494140625,
"logps/chosen": -20.090307235717773,
"logps/rejected": -59.078121185302734,
"loss": 0.1169,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.475255012512207,
"rewards/margins": 2.475255012512207,
"rewards/rejected": 0.0,
"step": 1420
},
{
"epoch": 7.932960893854749,
"eval_logits/chosen": -3.361480236053467,
"eval_logits/rejected": -3.488327741622925,
"eval_logps/chosen": -25.533275604248047,
"eval_logps/rejected": -45.56114196777344,
"eval_loss": 0.6961826086044312,
"eval_rewards/accuracies": 0.7749999761581421,
"eval_rewards/chosen": 0.4536762833595276,
"eval_rewards/margins": 0.4536762833595276,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7276,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 1420
},
{
"epoch": 7.93854748603352,
"grad_norm": 3.6104957537270526,
"learning_rate": 5.936188065487764e-07,
"logits/chosen": -3.586928129196167,
"logits/rejected": -3.7177696228027344,
"logps/chosen": -1.4119430780410767,
"logps/rejected": -24.606647491455078,
"loss": 0.1305,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.733350396156311,
"rewards/margins": 1.733350396156311,
"rewards/rejected": 0.0,
"step": 1421
},
{
"epoch": 7.94413407821229,
"grad_norm": 1.5794556899989902,
"learning_rate": 5.930199725790729e-07,
"logits/chosen": -3.2176473140716553,
"logits/rejected": -3.1848368644714355,
"logps/chosen": -21.64169692993164,
"logps/rejected": -41.134521484375,
"loss": 0.0846,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4605870246887207,
"rewards/margins": 3.4605870246887207,
"rewards/rejected": 0.0,
"step": 1422
},
{
"epoch": 7.949720670391061,
"grad_norm": 1.8753331860776365,
"learning_rate": 5.924210003651241e-07,
"logits/chosen": -3.482959032058716,
"logits/rejected": -3.593095541000366,
"logps/chosen": -4.3200578689575195,
"logps/rejected": -32.17010498046875,
"loss": 0.096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.347722053527832,
"rewards/margins": 2.347722053527832,
"rewards/rejected": 0.0,
"step": 1423
},
{
"epoch": 7.955307262569832,
"grad_norm": 2.255555162649948,
"learning_rate": 5.918218907971094e-07,
"logits/chosen": -3.6305172443389893,
"logits/rejected": -3.607184886932373,
"logps/chosen": -1.2554359436035156,
"logps/rejected": -29.569927215576172,
"loss": 0.1317,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.031468152999878,
"rewards/margins": 2.031468152999878,
"rewards/rejected": 0.0,
"step": 1424
},
{
"epoch": 7.960893854748603,
"grad_norm": 1.597704858025818,
"learning_rate": 5.912226447654126e-07,
"logits/chosen": -3.569333553314209,
"logits/rejected": -3.6216049194335938,
"logps/chosen": -7.23366641998291,
"logps/rejected": -30.640758514404297,
"loss": 0.1068,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3793201446533203,
"rewards/margins": 3.3793201446533203,
"rewards/rejected": 0.0,
"step": 1425
},
{
"epoch": 7.966480446927374,
"grad_norm": 1.8122627061811327,
"learning_rate": 5.906232631606197e-07,
"logits/chosen": -3.5693445205688477,
"logits/rejected": -3.5513341426849365,
"logps/chosen": -10.951475143432617,
"logps/rejected": -31.499202728271484,
"loss": 0.0955,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9858345985412598,
"rewards/margins": 2.9858345985412598,
"rewards/rejected": 0.0,
"step": 1426
},
{
"epoch": 7.972067039106145,
"grad_norm": 2.6827762380234916,
"learning_rate": 5.900237468735187e-07,
"logits/chosen": -3.6298258304595947,
"logits/rejected": -3.8624138832092285,
"logps/chosen": -27.023897171020508,
"logps/rejected": -43.194000244140625,
"loss": 0.1022,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.627375841140747,
"rewards/margins": 2.627375841140747,
"rewards/rejected": 0.0,
"step": 1427
},
{
"epoch": 7.977653631284916,
"grad_norm": 5.03487213920907,
"learning_rate": 5.894240967950975e-07,
"logits/chosen": -3.6355860233306885,
"logits/rejected": -3.4657349586486816,
"logps/chosen": -4.8664655685424805,
"logps/rejected": -72.21208190917969,
"loss": 0.1381,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4866435527801514,
"rewards/margins": 2.4866435527801514,
"rewards/rejected": 0.0,
"step": 1428
},
{
"epoch": 7.983240223463687,
"grad_norm": 6.345048662850746,
"learning_rate": 5.888243138165428e-07,
"logits/chosen": -3.5810298919677734,
"logits/rejected": -3.5718092918395996,
"logps/chosen": -8.52374267578125,
"logps/rejected": -21.86713409423828,
"loss": 0.1264,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.503370761871338,
"rewards/margins": 2.503370761871338,
"rewards/rejected": 0.0,
"step": 1429
},
{
"epoch": 7.988826815642458,
"grad_norm": 2.591814369716358,
"learning_rate": 5.882243988292392e-07,
"logits/chosen": -3.4187800884246826,
"logits/rejected": -3.482560396194458,
"logps/chosen": -2.0377821922302246,
"logps/rejected": -29.07729148864746,
"loss": 0.0949,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1618528366088867,
"rewards/margins": 2.1618528366088867,
"rewards/rejected": 0.0,
"step": 1430
},
{
"epoch": 7.994413407821229,
"grad_norm": 1.2980300131545568,
"learning_rate": 5.876243527247668e-07,
"logits/chosen": -3.6745760440826416,
"logits/rejected": -3.680694341659546,
"logps/chosen": -10.520817756652832,
"logps/rejected": -22.926095962524414,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.663787603378296,
"rewards/margins": 2.663787603378296,
"rewards/rejected": 0.0,
"step": 1431
},
{
"epoch": 8.0,
"grad_norm": 1.6699100499013064,
"learning_rate": 5.870241763949014e-07,
"logits/chosen": -3.602869987487793,
"logits/rejected": -3.5776760578155518,
"logps/chosen": -2.982872486114502,
"logps/rejected": -26.168415069580078,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.163339614868164,
"rewards/margins": 2.163339614868164,
"rewards/rejected": 0.0,
"step": 1432
},
{
"epoch": 8.005586592178771,
"grad_norm": 0.9050705109730053,
"learning_rate": 5.864238707316116e-07,
"logits/chosen": -3.2913479804992676,
"logits/rejected": -3.4438178539276123,
"logps/chosen": -26.037870407104492,
"logps/rejected": -39.11576461791992,
"loss": 0.0951,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1983556747436523,
"rewards/margins": 3.1983556747436523,
"rewards/rejected": 0.0,
"step": 1433
},
{
"epoch": 8.011173184357542,
"grad_norm": 1.127936173615925,
"learning_rate": 5.858234366270586e-07,
"logits/chosen": -3.483159065246582,
"logits/rejected": -3.3761470317840576,
"logps/chosen": -1.068917989730835,
"logps/rejected": -38.09598922729492,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7994022369384766,
"rewards/margins": 1.7994022369384766,
"rewards/rejected": 0.0,
"step": 1434
},
{
"epoch": 8.016759776536313,
"grad_norm": 1.8724140846559836,
"learning_rate": 5.852228749735946e-07,
"logits/chosen": -3.496731758117676,
"logits/rejected": -3.6163177490234375,
"logps/chosen": -1.5850613117218018,
"logps/rejected": -40.112396240234375,
"loss": 0.1264,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8730049133300781,
"rewards/margins": 1.8730049133300781,
"rewards/rejected": 0.0,
"step": 1435
},
{
"epoch": 8.022346368715084,
"grad_norm": 1.5935093998237166,
"learning_rate": 5.84622186663761e-07,
"logits/chosen": -3.453112840652466,
"logits/rejected": -3.3465051651000977,
"logps/chosen": -2.498274803161621,
"logps/rejected": -21.617000579833984,
"loss": 0.0952,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6029903888702393,
"rewards/margins": 2.6029903888702393,
"rewards/rejected": 0.0,
"step": 1436
},
{
"epoch": 8.027932960893855,
"grad_norm": 2.15951044683322,
"learning_rate": 5.840213725902876e-07,
"logits/chosen": -3.4027249813079834,
"logits/rejected": -3.431945562362671,
"logps/chosen": -5.285755157470703,
"logps/rejected": -34.72210693359375,
"loss": 0.1311,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.708683490753174,
"rewards/margins": 2.708683490753174,
"rewards/rejected": 0.0,
"step": 1437
},
{
"epoch": 8.033519553072626,
"grad_norm": 0.8011175434960901,
"learning_rate": 5.834204336460911e-07,
"logits/chosen": -3.5468287467956543,
"logits/rejected": -3.7178070545196533,
"logps/chosen": -5.940998077392578,
"logps/rejected": -49.089820861816406,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1336450576782227,
"rewards/margins": 3.1336450576782227,
"rewards/rejected": 0.0,
"step": 1438
},
{
"epoch": 8.039106145251397,
"grad_norm": 1.3280364220804615,
"learning_rate": 5.828193707242739e-07,
"logits/chosen": -3.666799306869507,
"logits/rejected": -3.7120518684387207,
"logps/chosen": -21.294273376464844,
"logps/rejected": -39.77464294433594,
"loss": 0.0972,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.398981094360352,
"rewards/margins": 4.398981094360352,
"rewards/rejected": 0.0,
"step": 1439
},
{
"epoch": 8.044692737430168,
"grad_norm": 3.402853983729217,
"learning_rate": 5.822181847181225e-07,
"logits/chosen": -3.434978485107422,
"logits/rejected": -3.446974515914917,
"logps/chosen": -0.9313244819641113,
"logps/rejected": -58.56283950805664,
"loss": 0.1522,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3463492393493652,
"rewards/margins": 1.3463492393493652,
"rewards/rejected": 0.0,
"step": 1440
},
{
"epoch": 8.044692737430168,
"eval_logits/chosen": -3.360203504562378,
"eval_logits/rejected": -3.488086700439453,
"eval_logps/chosen": -25.463388442993164,
"eval_logps/rejected": -45.98029327392578,
"eval_loss": 0.7024779915809631,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.46066561341285706,
"eval_rewards/margins": 0.46066561341285706,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7122,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 1440
},
{
"epoch": 8.050279329608939,
"grad_norm": 1.0956868530919193,
"learning_rate": 5.816168765211062e-07,
"logits/chosen": -3.486517906188965,
"logits/rejected": -3.427119255065918,
"logps/chosen": -9.0908203125,
"logps/rejected": -30.563962936401367,
"loss": 0.1106,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1291401386260986,
"rewards/margins": 3.1291401386260986,
"rewards/rejected": 0.0,
"step": 1441
},
{
"epoch": 8.05586592178771,
"grad_norm": 1.4963688156216446,
"learning_rate": 5.810154470268763e-07,
"logits/chosen": -3.4776787757873535,
"logits/rejected": -3.3631644248962402,
"logps/chosen": -6.872175693511963,
"logps/rejected": -41.6898193359375,
"loss": 0.1404,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4865894317626953,
"rewards/margins": 3.4865894317626953,
"rewards/rejected": 0.0,
"step": 1442
},
{
"epoch": 8.061452513966481,
"grad_norm": 2.350593152421529,
"learning_rate": 5.804138971292641e-07,
"logits/chosen": -3.5967485904693604,
"logits/rejected": -3.61181902885437,
"logps/chosen": -8.008771896362305,
"logps/rejected": -44.63813781738281,
"loss": 0.1545,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2435052394866943,
"rewards/margins": 3.2435052394866943,
"rewards/rejected": 0.0,
"step": 1443
},
{
"epoch": 8.067039106145252,
"grad_norm": 0.8201971929711597,
"learning_rate": 5.798122277222797e-07,
"logits/chosen": -3.495875597000122,
"logits/rejected": -3.578230142593384,
"logps/chosen": -4.396419525146484,
"logps/rejected": -61.46167755126953,
"loss": 0.078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4080634117126465,
"rewards/margins": 2.4080634117126465,
"rewards/rejected": 0.0,
"step": 1444
},
{
"epoch": 8.072625698324023,
"grad_norm": 2.6149905749508098,
"learning_rate": 5.792104397001113e-07,
"logits/chosen": -3.2644505500793457,
"logits/rejected": -3.475562334060669,
"logps/chosen": -1.256878137588501,
"logps/rejected": -65.22468566894531,
"loss": 0.125,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.986826777458191,
"rewards/margins": 1.986826777458191,
"rewards/rejected": 0.0,
"step": 1445
},
{
"epoch": 8.078212290502794,
"grad_norm": 1.7701652989609762,
"learning_rate": 5.786085339571229e-07,
"logits/chosen": -3.4310989379882812,
"logits/rejected": -3.526287794113159,
"logps/chosen": -1.9093594551086426,
"logps/rejected": -46.967960357666016,
"loss": 0.1098,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1475911140441895,
"rewards/margins": 2.1475911140441895,
"rewards/rejected": 0.0,
"step": 1446
},
{
"epoch": 8.083798882681565,
"grad_norm": 2.0776576727262817,
"learning_rate": 5.780065113878537e-07,
"logits/chosen": -3.533348321914673,
"logits/rejected": -3.5505824089050293,
"logps/chosen": -3.810591220855713,
"logps/rejected": -56.04825973510742,
"loss": 0.1019,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.585858106613159,
"rewards/margins": 2.585858106613159,
"rewards/rejected": 0.0,
"step": 1447
},
{
"epoch": 8.089385474860336,
"grad_norm": 0.9398781507882047,
"learning_rate": 5.774043728870161e-07,
"logits/chosen": -3.622270345687866,
"logits/rejected": -3.5502419471740723,
"logps/chosen": -8.514244079589844,
"logps/rejected": -24.470233917236328,
"loss": 0.0982,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.955456256866455,
"rewards/margins": 2.955456256866455,
"rewards/rejected": 0.0,
"step": 1448
},
{
"epoch": 8.094972067039107,
"grad_norm": 1.0188300956345018,
"learning_rate": 5.768021193494956e-07,
"logits/chosen": -3.660067081451416,
"logits/rejected": -3.6409153938293457,
"logps/chosen": -0.8675261735916138,
"logps/rejected": -57.86646270751953,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0977039337158203,
"rewards/margins": 2.0977039337158203,
"rewards/rejected": 0.0,
"step": 1449
},
{
"epoch": 8.100558659217878,
"grad_norm": 3.1775383354857127,
"learning_rate": 5.761997516703478e-07,
"logits/chosen": -3.269120931625366,
"logits/rejected": -3.2979540824890137,
"logps/chosen": -1.818122386932373,
"logps/rejected": -73.30386352539062,
"loss": 0.1367,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.148003339767456,
"rewards/margins": 2.148003339767456,
"rewards/rejected": 0.0,
"step": 1450
},
{
"epoch": 8.106145251396647,
"grad_norm": 1.1074461025220506,
"learning_rate": 5.755972707447989e-07,
"logits/chosen": -3.720050096511841,
"logits/rejected": -3.878394365310669,
"logps/chosen": -1.9917762279510498,
"logps/rejected": -46.12985610961914,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4722256660461426,
"rewards/margins": 2.4722256660461426,
"rewards/rejected": 0.0,
"step": 1451
},
{
"epoch": 8.111731843575418,
"grad_norm": 0.8836695493814023,
"learning_rate": 5.749946774682425e-07,
"logits/chosen": -3.4956867694854736,
"logits/rejected": -3.3236804008483887,
"logps/chosen": -18.465164184570312,
"logps/rejected": -59.880489349365234,
"loss": 0.1071,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.942150115966797,
"rewards/margins": 3.942150115966797,
"rewards/rejected": 0.0,
"step": 1452
},
{
"epoch": 8.11731843575419,
"grad_norm": 0.9583283573562732,
"learning_rate": 5.743919727362394e-07,
"logits/chosen": -3.217168092727661,
"logits/rejected": -3.538447856903076,
"logps/chosen": -1.6311407089233398,
"logps/rejected": -62.83324432373047,
"loss": 0.118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1093578338623047,
"rewards/margins": 2.1093578338623047,
"rewards/rejected": 0.0,
"step": 1453
},
{
"epoch": 8.12290502793296,
"grad_norm": 0.8329803477244008,
"learning_rate": 5.737891574445164e-07,
"logits/chosen": -3.36934232711792,
"logits/rejected": -3.589510679244995,
"logps/chosen": -11.139848709106445,
"logps/rejected": -36.51717758178711,
"loss": 0.0731,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.744891405105591,
"rewards/margins": 2.744891405105591,
"rewards/rejected": 0.0,
"step": 1454
},
{
"epoch": 8.128491620111731,
"grad_norm": 5.529502196973929,
"learning_rate": 5.731862324889644e-07,
"logits/chosen": -3.3760480880737305,
"logits/rejected": -3.513336420059204,
"logps/chosen": -2.8150508403778076,
"logps/rejected": -29.952720642089844,
"loss": 0.1324,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7353410720825195,
"rewards/margins": 2.7353410720825195,
"rewards/rejected": 0.0,
"step": 1455
},
{
"epoch": 8.134078212290502,
"grad_norm": 2.6192638848792242,
"learning_rate": 5.72583198765637e-07,
"logits/chosen": -3.425025463104248,
"logits/rejected": -3.4201722145080566,
"logps/chosen": -0.9650635123252869,
"logps/rejected": -38.63450241088867,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8116285800933838,
"rewards/margins": 1.8116285800933838,
"rewards/rejected": 0.0,
"step": 1456
},
{
"epoch": 8.139664804469273,
"grad_norm": 1.0286521930047465,
"learning_rate": 5.7198005717075e-07,
"logits/chosen": -3.4640731811523438,
"logits/rejected": -3.6357429027557373,
"logps/chosen": -1.1994825601577759,
"logps/rejected": -65.90774536132812,
"loss": 0.1141,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2714264392852783,
"rewards/margins": 2.2714264392852783,
"rewards/rejected": 0.0,
"step": 1457
},
{
"epoch": 8.145251396648044,
"grad_norm": 0.9117180594633306,
"learning_rate": 5.71376808600679e-07,
"logits/chosen": -3.1062095165252686,
"logits/rejected": -3.5535805225372314,
"logps/chosen": -4.527098178863525,
"logps/rejected": -61.18075942993164,
"loss": 0.1038,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.646533727645874,
"rewards/margins": 2.646533727645874,
"rewards/rejected": 0.0,
"step": 1458
},
{
"epoch": 8.150837988826815,
"grad_norm": 1.3223691265385114,
"learning_rate": 5.707734539519591e-07,
"logits/chosen": -3.413194417953491,
"logits/rejected": -3.099423885345459,
"logps/chosen": -13.974654197692871,
"logps/rejected": -33.39994812011719,
"loss": 0.095,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.1584930419921875,
"rewards/margins": 4.1584930419921875,
"rewards/rejected": 0.0,
"step": 1459
},
{
"epoch": 8.156424581005586,
"grad_norm": 2.409433570849459,
"learning_rate": 5.701699941212823e-07,
"logits/chosen": -3.4567711353302,
"logits/rejected": -3.5697836875915527,
"logps/chosen": -1.32706880569458,
"logps/rejected": -30.027732849121094,
"loss": 0.1275,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.199556827545166,
"rewards/margins": 2.199556827545166,
"rewards/rejected": 0.0,
"step": 1460
},
{
"epoch": 8.156424581005586,
"eval_logits/chosen": -3.3588969707489014,
"eval_logits/rejected": -3.4880967140197754,
"eval_logps/chosen": -26.15518569946289,
"eval_logps/rejected": -48.0563850402832,
"eval_loss": 0.7525140047073364,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.3914857506752014,
"eval_rewards/margins": 0.3914857506752014,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.703,
"eval_samples_per_second": 9.479,
"eval_steps_per_second": 0.306,
"step": 1460
},
{
"epoch": 8.162011173184357,
"grad_norm": 0.8386509025147895,
"learning_rate": 5.695664300054978e-07,
"logits/chosen": -3.4141733646392822,
"logits/rejected": -3.585333824157715,
"logps/chosen": -12.4519681930542,
"logps/rejected": -68.42787170410156,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.225080966949463,
"rewards/margins": 3.225080966949463,
"rewards/rejected": 0.0,
"step": 1461
},
{
"epoch": 8.167597765363128,
"grad_norm": 3.1654528953590955,
"learning_rate": 5.689627625016091e-07,
"logits/chosen": -3.5402932167053223,
"logits/rejected": -3.541752338409424,
"logps/chosen": -1.338120937347412,
"logps/rejected": -75.27066040039062,
"loss": 0.1238,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8199812173843384,
"rewards/margins": 1.8199812173843384,
"rewards/rejected": 0.0,
"step": 1462
},
{
"epoch": 8.1731843575419,
"grad_norm": 0.920781738470232,
"learning_rate": 5.683589925067738e-07,
"logits/chosen": -3.5387191772460938,
"logits/rejected": -3.5149905681610107,
"logps/chosen": -4.906529903411865,
"logps/rejected": -55.59393310546875,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0590710639953613,
"rewards/margins": 3.0590710639953613,
"rewards/rejected": 0.0,
"step": 1463
},
{
"epoch": 8.17877094972067,
"grad_norm": 1.2810884456437077,
"learning_rate": 5.677551209183016e-07,
"logits/chosen": -3.6051743030548096,
"logits/rejected": -3.4584357738494873,
"logps/chosen": -3.035280227661133,
"logps/rejected": -31.97887420654297,
"loss": 0.1005,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.869049549102783,
"rewards/margins": 2.869049549102783,
"rewards/rejected": 0.0,
"step": 1464
},
{
"epoch": 8.184357541899441,
"grad_norm": 0.8879067099029657,
"learning_rate": 5.671511486336531e-07,
"logits/chosen": -3.5902838706970215,
"logits/rejected": -3.6346681118011475,
"logps/chosen": -0.8220909833908081,
"logps/rejected": -31.008304595947266,
"loss": 0.1005,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3344693183898926,
"rewards/margins": 2.3344693183898926,
"rewards/rejected": 0.0,
"step": 1465
},
{
"epoch": 8.189944134078212,
"grad_norm": 1.8151028436159289,
"learning_rate": 5.665470765504388e-07,
"logits/chosen": -3.507279872894287,
"logits/rejected": -3.320805311203003,
"logps/chosen": -12.214398384094238,
"logps/rejected": -28.21538543701172,
"loss": 0.1373,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6122233867645264,
"rewards/margins": 2.6122233867645264,
"rewards/rejected": 0.0,
"step": 1466
},
{
"epoch": 8.195530726256983,
"grad_norm": 3.9022381059470432,
"learning_rate": 5.659429055664175e-07,
"logits/chosen": -3.4230411052703857,
"logits/rejected": -3.436345338821411,
"logps/chosen": -9.788567543029785,
"logps/rejected": -41.93543243408203,
"loss": 0.123,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5234522819519043,
"rewards/margins": 3.5234522819519043,
"rewards/rejected": 0.0,
"step": 1467
},
{
"epoch": 8.201117318435754,
"grad_norm": 2.587203604927202,
"learning_rate": 5.653386365794946e-07,
"logits/chosen": -3.4539031982421875,
"logits/rejected": -3.4906692504882812,
"logps/chosen": -4.39523458480835,
"logps/rejected": -39.78955078125,
"loss": 0.0902,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.061246871948242,
"rewards/margins": 3.061246871948242,
"rewards/rejected": 0.0,
"step": 1468
},
{
"epoch": 8.206703910614525,
"grad_norm": 2.4062376878197487,
"learning_rate": 5.647342704877216e-07,
"logits/chosen": -3.698641538619995,
"logits/rejected": -3.6472220420837402,
"logps/chosen": -3.5880136489868164,
"logps/rejected": -31.792198181152344,
"loss": 0.1164,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.213914155960083,
"rewards/margins": 2.213914155960083,
"rewards/rejected": 0.0,
"step": 1469
},
{
"epoch": 8.212290502793296,
"grad_norm": 1.4506575969019242,
"learning_rate": 5.64129808189294e-07,
"logits/chosen": -3.744248151779175,
"logits/rejected": -3.6063907146453857,
"logps/chosen": -6.610927581787109,
"logps/rejected": -19.785215377807617,
"loss": 0.0858,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8722078800201416,
"rewards/margins": 2.8722078800201416,
"rewards/rejected": 0.0,
"step": 1470
},
{
"epoch": 8.217877094972067,
"grad_norm": 0.8290463451813451,
"learning_rate": 5.635252505825507e-07,
"logits/chosen": -3.140641927719116,
"logits/rejected": -3.2642526626586914,
"logps/chosen": -1.733540654182434,
"logps/rejected": -94.34074401855469,
"loss": 0.1291,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3417115211486816,
"rewards/margins": 2.3417115211486816,
"rewards/rejected": 0.0,
"step": 1471
},
{
"epoch": 8.223463687150838,
"grad_norm": 1.2338899754429675,
"learning_rate": 5.629205985659718e-07,
"logits/chosen": -3.5912399291992188,
"logits/rejected": -3.4617409706115723,
"logps/chosen": -16.728078842163086,
"logps/rejected": -44.76894760131836,
"loss": 0.0951,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.295330286026001,
"rewards/margins": 3.295330286026001,
"rewards/rejected": 0.0,
"step": 1472
},
{
"epoch": 8.22905027932961,
"grad_norm": 3.4023832547711845,
"learning_rate": 5.623158530381777e-07,
"logits/chosen": -3.5220608711242676,
"logits/rejected": -3.5308523178100586,
"logps/chosen": -2.3623828887939453,
"logps/rejected": -31.084972381591797,
"loss": 0.1355,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6224255561828613,
"rewards/margins": 2.6224255561828613,
"rewards/rejected": 0.0,
"step": 1473
},
{
"epoch": 8.23463687150838,
"grad_norm": 4.063475271499999,
"learning_rate": 5.617110148979283e-07,
"logits/chosen": -3.2087531089782715,
"logits/rejected": -3.1776955127716064,
"logps/chosen": -0.906381368637085,
"logps/rejected": -37.97704315185547,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9931944608688354,
"rewards/margins": 1.9931944608688354,
"rewards/rejected": 0.0,
"step": 1474
},
{
"epoch": 8.240223463687151,
"grad_norm": 2.315438183563028,
"learning_rate": 5.611060850441206e-07,
"logits/chosen": -3.4199037551879883,
"logits/rejected": -3.3875744342803955,
"logps/chosen": -7.336108207702637,
"logps/rejected": -31.82926368713379,
"loss": 0.1327,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.218153715133667,
"rewards/margins": 3.218153715133667,
"rewards/rejected": 0.0,
"step": 1475
},
{
"epoch": 8.245810055865922,
"grad_norm": 2.101473370594678,
"learning_rate": 5.60501064375788e-07,
"logits/chosen": -3.6835854053497314,
"logits/rejected": -3.642129898071289,
"logps/chosen": -6.935665607452393,
"logps/rejected": -57.88630294799805,
"loss": 0.1028,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5883429050445557,
"rewards/margins": 2.5883429050445557,
"rewards/rejected": 0.0,
"step": 1476
},
{
"epoch": 8.251396648044693,
"grad_norm": 8.024140845797877,
"learning_rate": 5.598959537920994e-07,
"logits/chosen": -3.7375729084014893,
"logits/rejected": -3.604775905609131,
"logps/chosen": -9.299234390258789,
"logps/rejected": -40.53462219238281,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.622415542602539,
"rewards/margins": 2.622415542602539,
"rewards/rejected": 0.0,
"step": 1477
},
{
"epoch": 8.256983240223464,
"grad_norm": 1.1968602033493951,
"learning_rate": 5.592907541923564e-07,
"logits/chosen": -3.542449951171875,
"logits/rejected": -3.4163997173309326,
"logps/chosen": -0.6911807656288147,
"logps/rejected": -55.949310302734375,
"loss": 0.1026,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0382418632507324,
"rewards/margins": 2.0382418632507324,
"rewards/rejected": 0.0,
"step": 1478
},
{
"epoch": 8.262569832402235,
"grad_norm": 1.7473793253098002,
"learning_rate": 5.586854664759935e-07,
"logits/chosen": -3.5408735275268555,
"logits/rejected": -3.641183376312256,
"logps/chosen": -10.276369094848633,
"logps/rejected": -46.6556510925293,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.896777629852295,
"rewards/margins": 3.896777629852295,
"rewards/rejected": 0.0,
"step": 1479
},
{
"epoch": 8.268156424581006,
"grad_norm": 0.8557682911360565,
"learning_rate": 5.580800915425762e-07,
"logits/chosen": -3.7345876693725586,
"logits/rejected": -3.5522172451019287,
"logps/chosen": -11.918123245239258,
"logps/rejected": -31.203411102294922,
"loss": 0.0848,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.502218723297119,
"rewards/margins": 3.502218723297119,
"rewards/rejected": 0.0,
"step": 1480
},
{
"epoch": 8.268156424581006,
"eval_logits/chosen": -3.3670451641082764,
"eval_logits/rejected": -3.493870258331299,
"eval_logps/chosen": -25.99761390686035,
"eval_logps/rejected": -47.492557525634766,
"eval_loss": 0.7414113283157349,
"eval_rewards/accuracies": 0.699999988079071,
"eval_rewards/chosen": 0.4072425961494446,
"eval_rewards/margins": 0.4072425961494446,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7446,
"eval_samples_per_second": 9.467,
"eval_steps_per_second": 0.305,
"step": 1480
},
{
"epoch": 8.273743016759777,
"grad_norm": 1.5173921999512312,
"learning_rate": 5.574746302917993e-07,
"logits/chosen": -3.706956386566162,
"logits/rejected": -3.369967460632324,
"logps/chosen": -1.1656317710876465,
"logps/rejected": -44.58979797363281,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2945690155029297,
"rewards/margins": 2.2945690155029297,
"rewards/rejected": 0.0,
"step": 1481
},
{
"epoch": 8.279329608938548,
"grad_norm": 1.7484709661490887,
"learning_rate": 5.56869083623486e-07,
"logits/chosen": -3.258481025695801,
"logits/rejected": -3.519138813018799,
"logps/chosen": -1.1044570207595825,
"logps/rejected": -66.29925537109375,
"loss": 0.1047,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.114016056060791,
"rewards/margins": 2.114016056060791,
"rewards/rejected": 0.0,
"step": 1482
},
{
"epoch": 8.28491620111732,
"grad_norm": 1.1513836623784581,
"learning_rate": 5.562634524375869e-07,
"logits/chosen": -3.3325798511505127,
"logits/rejected": -3.3959743976593018,
"logps/chosen": -3.7329177856445312,
"logps/rejected": -28.587566375732422,
"loss": 0.0785,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9908037185668945,
"rewards/margins": 2.9908037185668945,
"rewards/rejected": 0.0,
"step": 1483
},
{
"epoch": 8.29050279329609,
"grad_norm": 0.6982183167014285,
"learning_rate": 5.556577376341773e-07,
"logits/chosen": -3.776313304901123,
"logits/rejected": -3.878551721572876,
"logps/chosen": -5.873186111450195,
"logps/rejected": -36.530662536621094,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.72420597076416,
"rewards/margins": 2.72420597076416,
"rewards/rejected": 0.0,
"step": 1484
},
{
"epoch": 8.296089385474861,
"grad_norm": 1.493891468823741,
"learning_rate": 5.550519401134576e-07,
"logits/chosen": -3.0950396060943604,
"logits/rejected": -3.0590198040008545,
"logps/chosen": -18.803293228149414,
"logps/rejected": -29.01365852355957,
"loss": 0.1231,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.442892551422119,
"rewards/margins": 2.442892551422119,
"rewards/rejected": 0.0,
"step": 1485
},
{
"epoch": 8.30167597765363,
"grad_norm": 0.8501588378546953,
"learning_rate": 5.544460607757505e-07,
"logits/chosen": -3.559340238571167,
"logits/rejected": -3.6847593784332275,
"logps/chosen": -24.36299705505371,
"logps/rejected": -51.866939544677734,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3191587924957275,
"rewards/margins": 3.3191587924957275,
"rewards/rejected": 0.0,
"step": 1486
},
{
"epoch": 8.307262569832401,
"grad_norm": 1.912154215172588,
"learning_rate": 5.53840100521501e-07,
"logits/chosen": -3.541095018386841,
"logits/rejected": -3.482886552810669,
"logps/chosen": -1.7762627601623535,
"logps/rejected": -65.16224670410156,
"loss": 0.1138,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2329888343811035,
"rewards/margins": 2.2329888343811035,
"rewards/rejected": 0.0,
"step": 1487
},
{
"epoch": 8.312849162011172,
"grad_norm": 0.8990791179958932,
"learning_rate": 5.532340602512736e-07,
"logits/chosen": -3.5542407035827637,
"logits/rejected": -3.6396937370300293,
"logps/chosen": -3.978266716003418,
"logps/rejected": -63.02491760253906,
"loss": 0.1247,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5638134479522705,
"rewards/margins": 2.5638134479522705,
"rewards/rejected": 0.0,
"step": 1488
},
{
"epoch": 8.318435754189943,
"grad_norm": 2.317155548097579,
"learning_rate": 5.526279408657522e-07,
"logits/chosen": -3.3917346000671387,
"logits/rejected": -3.515409231185913,
"logps/chosen": -11.306192398071289,
"logps/rejected": -67.20194244384766,
"loss": 0.0961,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.041069507598877,
"rewards/margins": 3.041069507598877,
"rewards/rejected": 0.0,
"step": 1489
},
{
"epoch": 8.324022346368714,
"grad_norm": 1.0533836877127634,
"learning_rate": 5.520217432657381e-07,
"logits/chosen": -3.6714675426483154,
"logits/rejected": -3.4912898540496826,
"logps/chosen": -6.029515266418457,
"logps/rejected": -71.71287536621094,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8722400665283203,
"rewards/margins": 2.8722400665283203,
"rewards/rejected": 0.0,
"step": 1490
},
{
"epoch": 8.329608938547485,
"grad_norm": 0.9798544392097811,
"learning_rate": 5.51415468352149e-07,
"logits/chosen": -3.5855624675750732,
"logits/rejected": -3.446227788925171,
"logps/chosen": -4.108002185821533,
"logps/rejected": -44.90166091918945,
"loss": 0.1151,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6292355060577393,
"rewards/margins": 2.6292355060577393,
"rewards/rejected": 0.0,
"step": 1491
},
{
"epoch": 8.335195530726256,
"grad_norm": 2.018163556164869,
"learning_rate": 5.508091170260173e-07,
"logits/chosen": -3.568167209625244,
"logits/rejected": -3.5685997009277344,
"logps/chosen": -0.9147201776504517,
"logps/rejected": -71.11444091796875,
"loss": 0.093,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.080580234527588,
"rewards/margins": 2.080580234527588,
"rewards/rejected": 0.0,
"step": 1492
},
{
"epoch": 8.340782122905027,
"grad_norm": 1.0815963292564568,
"learning_rate": 5.502026901884892e-07,
"logits/chosen": -3.4329142570495605,
"logits/rejected": -3.4202609062194824,
"logps/chosen": -2.4329872131347656,
"logps/rejected": -53.15180206298828,
"loss": 0.0895,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.744094133377075,
"rewards/margins": 2.744094133377075,
"rewards/rejected": 0.0,
"step": 1493
},
{
"epoch": 8.346368715083798,
"grad_norm": 0.9285472556353264,
"learning_rate": 5.495961887408228e-07,
"logits/chosen": -3.2632803916931152,
"logits/rejected": -3.1433935165405273,
"logps/chosen": -5.980401515960693,
"logps/rejected": -30.302764892578125,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8719334602355957,
"rewards/margins": 2.8719334602355957,
"rewards/rejected": 0.0,
"step": 1494
},
{
"epoch": 8.35195530726257,
"grad_norm": 0.9039569381629475,
"learning_rate": 5.489896135843873e-07,
"logits/chosen": -3.6132705211639404,
"logits/rejected": -3.616997480392456,
"logps/chosen": -1.6922589540481567,
"logps/rejected": -57.058143615722656,
"loss": 0.1323,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.035038948059082,
"rewards/margins": 2.035038948059082,
"rewards/rejected": 0.0,
"step": 1495
},
{
"epoch": 8.35754189944134,
"grad_norm": 1.8677324812628862,
"learning_rate": 5.483829656206614e-07,
"logits/chosen": -3.5454249382019043,
"logits/rejected": -3.6025378704071045,
"logps/chosen": -26.181739807128906,
"logps/rejected": -51.99799346923828,
"loss": 0.1369,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.705493211746216,
"rewards/margins": 3.705493211746216,
"rewards/rejected": 0.0,
"step": 1496
},
{
"epoch": 8.363128491620111,
"grad_norm": 0.8850994610735713,
"learning_rate": 5.477762457512322e-07,
"logits/chosen": -3.386732816696167,
"logits/rejected": -3.421370506286621,
"logps/chosen": -15.596943855285645,
"logps/rejected": -49.773773193359375,
"loss": 0.0746,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1644387245178223,
"rewards/margins": 3.1644387245178223,
"rewards/rejected": 0.0,
"step": 1497
},
{
"epoch": 8.368715083798882,
"grad_norm": 1.8447022932238195,
"learning_rate": 5.471694548777933e-07,
"logits/chosen": -3.4856345653533936,
"logits/rejected": -3.5015311241149902,
"logps/chosen": -11.751615524291992,
"logps/rejected": -30.04815101623535,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.336698293685913,
"rewards/margins": 3.336698293685913,
"rewards/rejected": 0.0,
"step": 1498
},
{
"epoch": 8.374301675977653,
"grad_norm": 1.3753032193008334,
"learning_rate": 5.46562593902144e-07,
"logits/chosen": -3.624182939529419,
"logits/rejected": -3.7773618698120117,
"logps/chosen": -7.332164287567139,
"logps/rejected": -43.15809631347656,
"loss": 0.081,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1540799140930176,
"rewards/margins": 3.1540799140930176,
"rewards/rejected": 0.0,
"step": 1499
},
{
"epoch": 8.379888268156424,
"grad_norm": 2.678452490793768,
"learning_rate": 5.459556637261878e-07,
"logits/chosen": -3.5382821559906006,
"logits/rejected": -3.62601375579834,
"logps/chosen": -0.8343517780303955,
"logps/rejected": -106.4999771118164,
"loss": 0.1039,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4513757228851318,
"rewards/margins": 1.4513757228851318,
"rewards/rejected": 0.0,
"step": 1500
},
{
"epoch": 8.379888268156424,
"eval_logits/chosen": -3.3645050525665283,
"eval_logits/rejected": -3.4921786785125732,
"eval_logps/chosen": -26.272689819335938,
"eval_logps/rejected": -48.18663024902344,
"eval_loss": 0.7770476341247559,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.37973541021347046,
"eval_rewards/margins": 0.37973541021347046,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7228,
"eval_samples_per_second": 9.474,
"eval_steps_per_second": 0.306,
"step": 1500
},
{
"epoch": 8.385474860335195,
"grad_norm": 2.290756959944643,
"learning_rate": 5.45348665251931e-07,
"logits/chosen": -3.5833957195281982,
"logits/rejected": -3.587348699569702,
"logps/chosen": -18.493114471435547,
"logps/rejected": -57.7999267578125,
"loss": 0.0911,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.490065574645996,
"rewards/margins": 3.490065574645996,
"rewards/rejected": 0.0,
"step": 1501
},
{
"epoch": 8.391061452513966,
"grad_norm": 1.3910650255217563,
"learning_rate": 5.447415993814815e-07,
"logits/chosen": -3.4692883491516113,
"logits/rejected": -3.611524820327759,
"logps/chosen": -9.678472518920898,
"logps/rejected": -27.444807052612305,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2595767974853516,
"rewards/margins": 3.2595767974853516,
"rewards/rejected": 0.0,
"step": 1502
},
{
"epoch": 8.396648044692737,
"grad_norm": 1.6357281876397822,
"learning_rate": 5.441344670170474e-07,
"logits/chosen": -3.481549024581909,
"logits/rejected": -3.4805691242218018,
"logps/chosen": -14.577237129211426,
"logps/rejected": -57.74549865722656,
"loss": 0.1266,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.264359474182129,
"rewards/margins": 3.264359474182129,
"rewards/rejected": 0.0,
"step": 1503
},
{
"epoch": 8.402234636871508,
"grad_norm": 2.0084390245276076,
"learning_rate": 5.435272690609352e-07,
"logits/chosen": -2.948986768722534,
"logits/rejected": -2.76161789894104,
"logps/chosen": -15.853133201599121,
"logps/rejected": -30.536693572998047,
"loss": 0.0982,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1021533012390137,
"rewards/margins": 3.1021533012390137,
"rewards/rejected": 0.0,
"step": 1504
},
{
"epoch": 8.40782122905028,
"grad_norm": 2.0121402411498,
"learning_rate": 5.429200064155495e-07,
"logits/chosen": -2.9251699447631836,
"logits/rejected": -2.935551643371582,
"logps/chosen": -6.076290130615234,
"logps/rejected": -37.09535217285156,
"loss": 0.1138,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.805009603500366,
"rewards/margins": 2.805009603500366,
"rewards/rejected": 0.0,
"step": 1505
},
{
"epoch": 8.41340782122905,
"grad_norm": 3.389742466031773,
"learning_rate": 5.423126799833906e-07,
"logits/chosen": -3.3590941429138184,
"logits/rejected": -3.596940040588379,
"logps/chosen": -2.6399240493774414,
"logps/rejected": -69.94694519042969,
"loss": 0.1042,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5809128284454346,
"rewards/margins": 2.5809128284454346,
"rewards/rejected": 0.0,
"step": 1506
},
{
"epoch": 8.418994413407821,
"grad_norm": 1.358979664747821,
"learning_rate": 5.41705290667054e-07,
"logits/chosen": -3.5871829986572266,
"logits/rejected": -3.5777151584625244,
"logps/chosen": -12.903154373168945,
"logps/rejected": -46.774505615234375,
"loss": 0.1079,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6193244457244873,
"rewards/margins": 3.6193244457244873,
"rewards/rejected": 0.0,
"step": 1507
},
{
"epoch": 8.424581005586592,
"grad_norm": 5.539810397595974,
"learning_rate": 5.410978393692278e-07,
"logits/chosen": -3.370941400527954,
"logits/rejected": -3.4300525188446045,
"logps/chosen": -5.34040641784668,
"logps/rejected": -72.93252563476562,
"loss": 0.1034,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.609839916229248,
"rewards/margins": 2.609839916229248,
"rewards/rejected": 0.0,
"step": 1508
},
{
"epoch": 8.430167597765363,
"grad_norm": 1.8268379553037781,
"learning_rate": 5.404903269926933e-07,
"logits/chosen": -3.537126064300537,
"logits/rejected": -3.7760980129241943,
"logps/chosen": -8.276466369628906,
"logps/rejected": -29.803316116333008,
"loss": 0.1081,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.295217514038086,
"rewards/margins": 3.295217514038086,
"rewards/rejected": 0.0,
"step": 1509
},
{
"epoch": 8.435754189944134,
"grad_norm": 1.6483309748748936,
"learning_rate": 5.398827544403221e-07,
"logits/chosen": -3.3367722034454346,
"logits/rejected": -3.454108476638794,
"logps/chosen": -2.6563539505004883,
"logps/rejected": -61.56664276123047,
"loss": 0.1038,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2975759506225586,
"rewards/margins": 2.2975759506225586,
"rewards/rejected": 0.0,
"step": 1510
},
{
"epoch": 8.441340782122905,
"grad_norm": 1.2754593474707578,
"learning_rate": 5.392751226150748e-07,
"logits/chosen": -3.487358570098877,
"logits/rejected": -3.4356279373168945,
"logps/chosen": -1.3859034776687622,
"logps/rejected": -25.730405807495117,
"loss": 0.1079,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.415024995803833,
"rewards/margins": 1.415024995803833,
"rewards/rejected": 0.0,
"step": 1511
},
{
"epoch": 8.446927374301676,
"grad_norm": 1.9779814605988295,
"learning_rate": 5.386674324200009e-07,
"logits/chosen": -3.5959315299987793,
"logits/rejected": -3.629880428314209,
"logps/chosen": -1.9470138549804688,
"logps/rejected": -47.28021240234375,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8410720825195312,
"rewards/margins": 2.8410720825195312,
"rewards/rejected": 0.0,
"step": 1512
},
{
"epoch": 8.452513966480447,
"grad_norm": 1.0735096680239324,
"learning_rate": 5.380596847582362e-07,
"logits/chosen": -3.4882895946502686,
"logits/rejected": -3.619725227355957,
"logps/chosen": -2.756035089492798,
"logps/rejected": -37.28264236450195,
"loss": 0.0848,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.608781337738037,
"rewards/margins": 2.608781337738037,
"rewards/rejected": 0.0,
"step": 1513
},
{
"epoch": 8.458100558659218,
"grad_norm": 3.0054957128503537,
"learning_rate": 5.374518805330019e-07,
"logits/chosen": -3.7351033687591553,
"logits/rejected": -3.6582224369049072,
"logps/chosen": -1.2357591390609741,
"logps/rejected": -46.32286071777344,
"loss": 0.0819,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.195836067199707,
"rewards/margins": 2.195836067199707,
"rewards/rejected": 0.0,
"step": 1514
},
{
"epoch": 8.46368715083799,
"grad_norm": 1.702956122371466,
"learning_rate": 5.368440206476032e-07,
"logits/chosen": -3.5749435424804688,
"logits/rejected": -3.3852646350860596,
"logps/chosen": -2.3170430660247803,
"logps/rejected": -30.341320037841797,
"loss": 0.1247,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3695919513702393,
"rewards/margins": 2.3695919513702393,
"rewards/rejected": 0.0,
"step": 1515
},
{
"epoch": 8.46927374301676,
"grad_norm": 3.202145899584805,
"learning_rate": 5.362361060054284e-07,
"logits/chosen": -3.424440622329712,
"logits/rejected": -3.652538299560547,
"logps/chosen": -1.2449885606765747,
"logps/rejected": -80.08869934082031,
"loss": 0.1181,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3632490634918213,
"rewards/margins": 2.3632490634918213,
"rewards/rejected": 0.0,
"step": 1516
},
{
"epoch": 8.474860335195531,
"grad_norm": 1.1199726304529054,
"learning_rate": 5.356281375099467e-07,
"logits/chosen": -3.2655513286590576,
"logits/rejected": -3.421264171600342,
"logps/chosen": -6.837052345275879,
"logps/rejected": -21.229103088378906,
"loss": 0.1155,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.299281120300293,
"rewards/margins": 2.299281120300293,
"rewards/rejected": 0.0,
"step": 1517
},
{
"epoch": 8.480446927374302,
"grad_norm": 2.959212075028282,
"learning_rate": 5.350201160647077e-07,
"logits/chosen": -3.680194854736328,
"logits/rejected": -3.5862388610839844,
"logps/chosen": -6.810611724853516,
"logps/rejected": -26.74586296081543,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6470818519592285,
"rewards/margins": 2.6470818519592285,
"rewards/rejected": 0.0,
"step": 1518
},
{
"epoch": 8.486033519553073,
"grad_norm": 3.517136317026258,
"learning_rate": 5.344120425733394e-07,
"logits/chosen": -3.3773891925811768,
"logits/rejected": -3.3142926692962646,
"logps/chosen": -1.5903003215789795,
"logps/rejected": -86.06102752685547,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4226956367492676,
"rewards/margins": 2.4226956367492676,
"rewards/rejected": 0.0,
"step": 1519
},
{
"epoch": 8.491620111731844,
"grad_norm": 1.0342319624873688,
"learning_rate": 5.338039179395474e-07,
"logits/chosen": -3.496845245361328,
"logits/rejected": -3.676945447921753,
"logps/chosen": -0.4623601734638214,
"logps/rejected": -49.321067810058594,
"loss": 0.1203,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5963795185089111,
"rewards/margins": 1.5963795185089111,
"rewards/rejected": 0.0,
"step": 1520
},
{
"epoch": 8.491620111731844,
"eval_logits/chosen": -3.3499298095703125,
"eval_logits/rejected": -3.480783462524414,
"eval_logps/chosen": -26.07053565979004,
"eval_logps/rejected": -47.708984375,
"eval_loss": 0.7492290139198303,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.39995068311691284,
"eval_rewards/margins": 0.39995068311691284,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7066,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 1520
},
{
"epoch": 8.497206703910614,
"grad_norm": 1.1080581426802347,
"learning_rate": 5.331957430671131e-07,
"logits/chosen": -3.567438840866089,
"logits/rejected": -3.6613609790802,
"logps/chosen": -5.783607482910156,
"logps/rejected": -44.21714782714844,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9833672046661377,
"rewards/margins": 2.9833672046661377,
"rewards/rejected": 0.0,
"step": 1521
},
{
"epoch": 8.502793296089386,
"grad_norm": 0.7701425543102947,
"learning_rate": 5.325875188598928e-07,
"logits/chosen": -3.6467268466949463,
"logits/rejected": -3.6299521923065186,
"logps/chosen": -2.856154441833496,
"logps/rejected": -51.156734466552734,
"loss": 0.1139,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.678068161010742,
"rewards/margins": 2.678068161010742,
"rewards/rejected": 0.0,
"step": 1522
},
{
"epoch": 8.508379888268156,
"grad_norm": 0.8195655538118419,
"learning_rate": 5.319792462218162e-07,
"logits/chosen": -3.4476943016052246,
"logits/rejected": -3.5049538612365723,
"logps/chosen": -7.8609185218811035,
"logps/rejected": -33.09964370727539,
"loss": 0.0884,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5700316429138184,
"rewards/margins": 2.5700316429138184,
"rewards/rejected": 0.0,
"step": 1523
},
{
"epoch": 8.513966480446927,
"grad_norm": 3.8949954334163324,
"learning_rate": 5.313709260568842e-07,
"logits/chosen": -3.5582966804504395,
"logits/rejected": -3.485438346862793,
"logps/chosen": -10.91259479522705,
"logps/rejected": -30.21276092529297,
"loss": 0.0936,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.235612392425537,
"rewards/margins": 3.235612392425537,
"rewards/rejected": 0.0,
"step": 1524
},
{
"epoch": 8.519553072625698,
"grad_norm": 1.1650742674810737,
"learning_rate": 5.307625592691694e-07,
"logits/chosen": -3.4802331924438477,
"logits/rejected": -3.549264669418335,
"logps/chosen": -4.471827507019043,
"logps/rejected": -28.547901153564453,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.785182237625122,
"rewards/margins": 2.785182237625122,
"rewards/rejected": 0.0,
"step": 1525
},
{
"epoch": 8.525139664804469,
"grad_norm": 2.9184506790702223,
"learning_rate": 5.30154146762813e-07,
"logits/chosen": -3.5332322120666504,
"logits/rejected": -3.2711181640625,
"logps/chosen": -6.034396648406982,
"logps/rejected": -29.913333892822266,
"loss": 0.0907,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1569344997406006,
"rewards/margins": 3.1569344997406006,
"rewards/rejected": 0.0,
"step": 1526
},
{
"epoch": 8.53072625698324,
"grad_norm": 2.0859174781379504,
"learning_rate": 5.295456894420243e-07,
"logits/chosen": -3.5591535568237305,
"logits/rejected": -3.5449156761169434,
"logps/chosen": -4.0158162117004395,
"logps/rejected": -25.790225982666016,
"loss": 0.115,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8487677574157715,
"rewards/margins": 2.8487677574157715,
"rewards/rejected": 0.0,
"step": 1527
},
{
"epoch": 8.53631284916201,
"grad_norm": 2.575040652943191,
"learning_rate": 5.289371882110794e-07,
"logits/chosen": -3.381300687789917,
"logits/rejected": -3.3595049381256104,
"logps/chosen": -0.8191465139389038,
"logps/rejected": -40.855125427246094,
"loss": 0.129,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7980883121490479,
"rewards/margins": 1.7980883121490479,
"rewards/rejected": 0.0,
"step": 1528
},
{
"epoch": 8.541899441340782,
"grad_norm": 2.0603094432918603,
"learning_rate": 5.283286439743194e-07,
"logits/chosen": -3.4769701957702637,
"logits/rejected": -3.696000337600708,
"logps/chosen": -2.3960256576538086,
"logps/rejected": -79.72267150878906,
"loss": 0.1039,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.579245090484619,
"rewards/margins": 2.579245090484619,
"rewards/rejected": 0.0,
"step": 1529
},
{
"epoch": 8.547486033519553,
"grad_norm": 3.296843118035138,
"learning_rate": 5.277200576361492e-07,
"logits/chosen": -3.4191207885742188,
"logits/rejected": -3.3528926372528076,
"logps/chosen": -3.467456579208374,
"logps/rejected": -77.3072280883789,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1921169757843018,
"rewards/margins": 2.1921169757843018,
"rewards/rejected": 0.0,
"step": 1530
},
{
"epoch": 8.553072625698324,
"grad_norm": 4.3778375320671366,
"learning_rate": 5.271114301010368e-07,
"logits/chosen": -3.7481722831726074,
"logits/rejected": -3.469683885574341,
"logps/chosen": -4.541937828063965,
"logps/rejected": -50.02873992919922,
"loss": 0.0998,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.622313976287842,
"rewards/margins": 2.622313976287842,
"rewards/rejected": 0.0,
"step": 1531
},
{
"epoch": 8.558659217877095,
"grad_norm": 1.442013610427837,
"learning_rate": 5.265027622735109e-07,
"logits/chosen": -3.201296806335449,
"logits/rejected": -3.4194226264953613,
"logps/chosen": -0.5189584493637085,
"logps/rejected": -107.17063903808594,
"loss": 0.1156,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6658756732940674,
"rewards/margins": 1.6658756732940674,
"rewards/rejected": 0.0,
"step": 1532
},
{
"epoch": 8.564245810055866,
"grad_norm": 8.037115998414105,
"learning_rate": 5.258940550581605e-07,
"logits/chosen": -3.4911601543426514,
"logits/rejected": -3.5738885402679443,
"logps/chosen": -5.360485076904297,
"logps/rejected": -48.8051643371582,
"loss": 0.1379,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3148505687713623,
"rewards/margins": 2.3148505687713623,
"rewards/rejected": 0.0,
"step": 1533
},
{
"epoch": 8.569832402234637,
"grad_norm": 0.8811270446557269,
"learning_rate": 5.252853093596327e-07,
"logits/chosen": -3.525219678878784,
"logits/rejected": -3.527994155883789,
"logps/chosen": -29.17947006225586,
"logps/rejected": -39.194740295410156,
"loss": 0.0784,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.913750648498535,
"rewards/margins": 3.913750648498535,
"rewards/rejected": 0.0,
"step": 1534
},
{
"epoch": 8.575418994413408,
"grad_norm": 1.4367890402400878,
"learning_rate": 5.246765260826321e-07,
"logits/chosen": -3.659507989883423,
"logits/rejected": -3.5021913051605225,
"logps/chosen": -14.039175987243652,
"logps/rejected": -39.691810607910156,
"loss": 0.1033,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6174304485321045,
"rewards/margins": 3.6174304485321045,
"rewards/rejected": 0.0,
"step": 1535
},
{
"epoch": 8.581005586592179,
"grad_norm": 1.6099255248292248,
"learning_rate": 5.240677061319193e-07,
"logits/chosen": -3.3976521492004395,
"logits/rejected": -3.2862162590026855,
"logps/chosen": -14.649343490600586,
"logps/rejected": -43.704246520996094,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6249542236328125,
"rewards/margins": 2.6249542236328125,
"rewards/rejected": 0.0,
"step": 1536
},
{
"epoch": 8.58659217877095,
"grad_norm": 1.438840290175727,
"learning_rate": 5.234588504123088e-07,
"logits/chosen": -3.433427095413208,
"logits/rejected": -3.3993639945983887,
"logps/chosen": -1.6107079982757568,
"logps/rejected": -46.80220031738281,
"loss": 0.1226,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3125686645507812,
"rewards/margins": 2.3125686645507812,
"rewards/rejected": 0.0,
"step": 1537
},
{
"epoch": 8.59217877094972,
"grad_norm": 2.1215231779282946,
"learning_rate": 5.228499598286689e-07,
"logits/chosen": -3.644007921218872,
"logits/rejected": -3.645432472229004,
"logps/chosen": -3.7853474617004395,
"logps/rejected": -36.43778610229492,
"loss": 0.1082,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.251805305480957,
"rewards/margins": 3.251805305480957,
"rewards/rejected": 0.0,
"step": 1538
},
{
"epoch": 8.597765363128492,
"grad_norm": 3.0671268932467886,
"learning_rate": 5.222410352859192e-07,
"logits/chosen": -3.644315719604492,
"logits/rejected": -3.7836644649505615,
"logps/chosen": -5.186097145080566,
"logps/rejected": -33.71754455566406,
"loss": 0.113,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1996922492980957,
"rewards/margins": 3.1996922492980957,
"rewards/rejected": 0.0,
"step": 1539
},
{
"epoch": 8.603351955307263,
"grad_norm": 0.8552394511306705,
"learning_rate": 5.216320776890303e-07,
"logits/chosen": -3.527188539505005,
"logits/rejected": -3.445709705352783,
"logps/chosen": -14.652050018310547,
"logps/rejected": -43.6121826171875,
"loss": 0.0618,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5236968994140625,
"rewards/margins": 3.5236968994140625,
"rewards/rejected": 0.0,
"step": 1540
},
{
"epoch": 8.603351955307263,
"eval_logits/chosen": -3.355664014816284,
"eval_logits/rejected": -3.4829509258270264,
"eval_logps/chosen": -26.36115074157715,
"eval_logps/rejected": -48.14537811279297,
"eval_loss": 0.7530530095100403,
"eval_rewards/accuracies": 0.75,
"eval_rewards/chosen": 0.3708893656730652,
"eval_rewards/margins": 0.3708893656730652,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6917,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 1540
},
{
"epoch": 8.608938547486034,
"grad_norm": 2.0342816055264823,
"learning_rate": 5.210230879430213e-07,
"logits/chosen": -3.7195377349853516,
"logits/rejected": -3.627046585083008,
"logps/chosen": -6.564638614654541,
"logps/rejected": -45.38517761230469,
"loss": 0.0918,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.422337770462036,
"rewards/margins": 3.422337770462036,
"rewards/rejected": 0.0,
"step": 1541
},
{
"epoch": 8.614525139664805,
"grad_norm": 2.3634223161676147,
"learning_rate": 5.204140669529597e-07,
"logits/chosen": -3.469510316848755,
"logits/rejected": -3.533547878265381,
"logps/chosen": -1.5332813262939453,
"logps/rejected": -40.70933151245117,
"loss": 0.1201,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.298218011856079,
"rewards/margins": 2.298218011856079,
"rewards/rejected": 0.0,
"step": 1542
},
{
"epoch": 8.620111731843576,
"grad_norm": 4.001016468988307,
"learning_rate": 5.198050156239592e-07,
"logits/chosen": -3.583134651184082,
"logits/rejected": -3.668936014175415,
"logps/chosen": -2.190661668777466,
"logps/rejected": -43.959571838378906,
"loss": 0.1196,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7127172946929932,
"rewards/margins": 1.7127172946929932,
"rewards/rejected": 0.0,
"step": 1543
},
{
"epoch": 8.625698324022347,
"grad_norm": 3.2715619364999884,
"learning_rate": 5.191959348611782e-07,
"logits/chosen": -3.462294816970825,
"logits/rejected": -3.3381593227386475,
"logps/chosen": -7.955991268157959,
"logps/rejected": -46.20689392089844,
"loss": 0.1126,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.412771701812744,
"rewards/margins": 3.412771701812744,
"rewards/rejected": 0.0,
"step": 1544
},
{
"epoch": 8.631284916201118,
"grad_norm": 1.2037941282863056,
"learning_rate": 5.185868255698195e-07,
"logits/chosen": -3.3791279792785645,
"logits/rejected": -3.088306188583374,
"logps/chosen": -9.527669906616211,
"logps/rejected": -71.26609802246094,
"loss": 0.0806,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5406413078308105,
"rewards/margins": 2.5406413078308105,
"rewards/rejected": 0.0,
"step": 1545
},
{
"epoch": 8.636871508379889,
"grad_norm": 1.313322479436088,
"learning_rate": 5.179776886551279e-07,
"logits/chosen": -3.0730252265930176,
"logits/rejected": -3.005383253097534,
"logps/chosen": -6.606561660766602,
"logps/rejected": -43.8348388671875,
"loss": 0.0817,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1686437129974365,
"rewards/margins": 3.1686437129974365,
"rewards/rejected": 0.0,
"step": 1546
},
{
"epoch": 8.64245810055866,
"grad_norm": 2.0170485666166607,
"learning_rate": 5.173685250223893e-07,
"logits/chosen": -3.499685764312744,
"logits/rejected": -3.5609824657440186,
"logps/chosen": -8.598729133605957,
"logps/rejected": -36.26488494873047,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3983078002929688,
"rewards/margins": 3.3983078002929688,
"rewards/rejected": 0.0,
"step": 1547
},
{
"epoch": 8.64804469273743,
"grad_norm": 0.9900053587305689,
"learning_rate": 5.167593355769293e-07,
"logits/chosen": -3.0416650772094727,
"logits/rejected": -2.926699638366699,
"logps/chosen": -21.06253433227539,
"logps/rejected": -38.30773162841797,
"loss": 0.1033,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.208059787750244,
"rewards/margins": 3.208059787750244,
"rewards/rejected": 0.0,
"step": 1548
},
{
"epoch": 8.653631284916202,
"grad_norm": 1.560846482349804,
"learning_rate": 5.161501212241119e-07,
"logits/chosen": -3.668612241744995,
"logits/rejected": -3.6418755054473877,
"logps/chosen": -0.6443605422973633,
"logps/rejected": -52.83890151977539,
"loss": 0.1239,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.734978437423706,
"rewards/margins": 1.734978437423706,
"rewards/rejected": 0.0,
"step": 1549
},
{
"epoch": 8.659217877094973,
"grad_norm": 2.687339764265376,
"learning_rate": 5.155408828693382e-07,
"logits/chosen": -3.4790260791778564,
"logits/rejected": -3.5892174243927,
"logps/chosen": -6.555553913116455,
"logps/rejected": -25.03845977783203,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.771357536315918,
"rewards/margins": 2.771357536315918,
"rewards/rejected": 0.0,
"step": 1550
},
{
"epoch": 8.664804469273744,
"grad_norm": 2.065569564093539,
"learning_rate": 5.14931621418045e-07,
"logits/chosen": -3.499326229095459,
"logits/rejected": -3.357387065887451,
"logps/chosen": -5.375399589538574,
"logps/rejected": -26.244762420654297,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.560816764831543,
"rewards/margins": 2.560816764831543,
"rewards/rejected": 0.0,
"step": 1551
},
{
"epoch": 8.670391061452515,
"grad_norm": 1.4612654511791863,
"learning_rate": 5.143223377757031e-07,
"logits/chosen": -3.5222458839416504,
"logits/rejected": -3.454744338989258,
"logps/chosen": -5.21843147277832,
"logps/rejected": -53.71820068359375,
"loss": 0.093,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6841795444488525,
"rewards/margins": 2.6841795444488525,
"rewards/rejected": 0.0,
"step": 1552
},
{
"epoch": 8.675977653631286,
"grad_norm": 2.888368778684556,
"learning_rate": 5.137130328478166e-07,
"logits/chosen": -3.3584721088409424,
"logits/rejected": -3.3054399490356445,
"logps/chosen": -16.143930435180664,
"logps/rejected": -52.69926452636719,
"loss": 0.1346,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1023361682891846,
"rewards/margins": 3.1023361682891846,
"rewards/rejected": 0.0,
"step": 1553
},
{
"epoch": 8.681564245810057,
"grad_norm": 3.815984947336513,
"learning_rate": 5.131037075399212e-07,
"logits/chosen": -3.327307939529419,
"logits/rejected": -3.222965717315674,
"logps/chosen": -24.722620010375977,
"logps/rejected": -67.58235168457031,
"loss": 0.1097,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7870500087738037,
"rewards/margins": 2.7870500087738037,
"rewards/rejected": 0.0,
"step": 1554
},
{
"epoch": 8.687150837988828,
"grad_norm": 1.8643236715290348,
"learning_rate": 5.124943627575826e-07,
"logits/chosen": -3.320669412612915,
"logits/rejected": -3.542205810546875,
"logps/chosen": -14.131210327148438,
"logps/rejected": -45.428321838378906,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.646136999130249,
"rewards/margins": 3.646136999130249,
"rewards/rejected": 0.0,
"step": 1555
},
{
"epoch": 8.692737430167599,
"grad_norm": 1.9161077564537476,
"learning_rate": 5.11884999406396e-07,
"logits/chosen": -3.2250707149505615,
"logits/rejected": -3.2313990592956543,
"logps/chosen": -23.096805572509766,
"logps/rejected": -60.486968994140625,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9371800422668457,
"rewards/margins": 3.9371800422668457,
"rewards/rejected": 0.0,
"step": 1556
},
{
"epoch": 8.69832402234637,
"grad_norm": 3.096848599623186,
"learning_rate": 5.112756183919836e-07,
"logits/chosen": -3.2024033069610596,
"logits/rejected": -3.2210841178894043,
"logps/chosen": -13.349126815795898,
"logps/rejected": -54.73284912109375,
"loss": 0.1103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.460592746734619,
"rewards/margins": 2.460592746734619,
"rewards/rejected": 0.0,
"step": 1557
},
{
"epoch": 8.703910614525139,
"grad_norm": 1.3174950052810543,
"learning_rate": 5.106662206199942e-07,
"logits/chosen": -3.3897392749786377,
"logits/rejected": -3.579496145248413,
"logps/chosen": -3.686384677886963,
"logps/rejected": -26.034996032714844,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.913018226623535,
"rewards/margins": 2.913018226623535,
"rewards/rejected": 0.0,
"step": 1558
},
{
"epoch": 8.70949720670391,
"grad_norm": 1.1337723992522022,
"learning_rate": 5.100568069961014e-07,
"logits/chosen": -3.577808141708374,
"logits/rejected": -3.5318191051483154,
"logps/chosen": -2.9288718700408936,
"logps/rejected": -63.374305725097656,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5752320289611816,
"rewards/margins": 2.5752320289611816,
"rewards/rejected": 0.0,
"step": 1559
},
{
"epoch": 8.71508379888268,
"grad_norm": 3.3617005770425044,
"learning_rate": 5.094473784260023e-07,
"logits/chosen": -3.412342071533203,
"logits/rejected": -3.5374176502227783,
"logps/chosen": -6.275794506072998,
"logps/rejected": -37.99629211425781,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3646249771118164,
"rewards/margins": 2.3646249771118164,
"rewards/rejected": 0.0,
"step": 1560
},
{
"epoch": 8.71508379888268,
"eval_logits/chosen": -3.343669891357422,
"eval_logits/rejected": -3.472111463546753,
"eval_logps/chosen": -26.061229705810547,
"eval_logps/rejected": -47.825870513916016,
"eval_loss": 0.7493130564689636,
"eval_rewards/accuracies": 0.7250000238418579,
"eval_rewards/chosen": 0.4008810520172119,
"eval_rewards/margins": 0.4008810520172119,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7379,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 1560
},
{
"epoch": 8.720670391061452,
"grad_norm": 1.1258929545927845,
"learning_rate": 5.088379358154165e-07,
"logits/chosen": -3.7648444175720215,
"logits/rejected": -3.540130376815796,
"logps/chosen": -3.2312402725219727,
"logps/rejected": -41.85802459716797,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6411352157592773,
"rewards/margins": 2.6411352157592773,
"rewards/rejected": 0.0,
"step": 1561
},
{
"epoch": 8.726256983240223,
"grad_norm": 1.2156419190124832,
"learning_rate": 5.082284800700841e-07,
"logits/chosen": -3.464627981185913,
"logits/rejected": -3.509732961654663,
"logps/chosen": -3.6714587211608887,
"logps/rejected": -36.64527130126953,
"loss": 0.1028,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.553143262863159,
"rewards/margins": 2.553143262863159,
"rewards/rejected": 0.0,
"step": 1562
},
{
"epoch": 8.731843575418994,
"grad_norm": 1.3044713481705497,
"learning_rate": 5.076190120957649e-07,
"logits/chosen": -3.347734212875366,
"logits/rejected": -3.411790370941162,
"logps/chosen": -3.0454139709472656,
"logps/rejected": -41.397701263427734,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.37506103515625,
"rewards/margins": 2.37506103515625,
"rewards/rejected": 0.0,
"step": 1563
},
{
"epoch": 8.737430167597765,
"grad_norm": 1.3422598830715409,
"learning_rate": 5.070095327982368e-07,
"logits/chosen": -3.61736798286438,
"logits/rejected": -3.5823631286621094,
"logps/chosen": -15.124504089355469,
"logps/rejected": -33.697410583496094,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5412120819091797,
"rewards/margins": 3.5412120819091797,
"rewards/rejected": 0.0,
"step": 1564
},
{
"epoch": 8.743016759776536,
"grad_norm": 0.9727540910407199,
"learning_rate": 5.064000430832947e-07,
"logits/chosen": -3.644531011581421,
"logits/rejected": -3.639143943786621,
"logps/chosen": -8.086284637451172,
"logps/rejected": -24.990882873535156,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3636388778686523,
"rewards/margins": 3.3636388778686523,
"rewards/rejected": 0.0,
"step": 1565
},
{
"epoch": 8.748603351955307,
"grad_norm": 1.7613299882922084,
"learning_rate": 5.057905438567488e-07,
"logits/chosen": -3.5720908641815186,
"logits/rejected": -3.421658515930176,
"logps/chosen": -4.2872819900512695,
"logps/rejected": -46.355926513671875,
"loss": 0.1236,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.071406364440918,
"rewards/margins": 2.071406364440918,
"rewards/rejected": 0.0,
"step": 1566
},
{
"epoch": 8.754189944134078,
"grad_norm": 2.2865459606433034,
"learning_rate": 5.051810360244234e-07,
"logits/chosen": -3.486600875854492,
"logits/rejected": -3.6979551315307617,
"logps/chosen": -19.66204071044922,
"logps/rejected": -56.792850494384766,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.337367534637451,
"rewards/margins": 3.337367534637451,
"rewards/rejected": 0.0,
"step": 1567
},
{
"epoch": 8.759776536312849,
"grad_norm": 7.629526182105958,
"learning_rate": 5.045715204921559e-07,
"logits/chosen": -3.6171090602874756,
"logits/rejected": -3.5304319858551025,
"logps/chosen": -16.706317901611328,
"logps/rejected": -51.3306999206543,
"loss": 0.1213,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9988741874694824,
"rewards/margins": 2.9988741874694824,
"rewards/rejected": 0.0,
"step": 1568
},
{
"epoch": 8.76536312849162,
"grad_norm": 0.8949060173920261,
"learning_rate": 5.039619981657948e-07,
"logits/chosen": -3.300743818283081,
"logits/rejected": -3.1763076782226562,
"logps/chosen": -18.87421989440918,
"logps/rejected": -66.6238021850586,
"loss": 0.0815,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.896085023880005,
"rewards/margins": 3.896085023880005,
"rewards/rejected": 0.0,
"step": 1569
},
{
"epoch": 8.77094972067039,
"grad_norm": 1.015738087328273,
"learning_rate": 5.033524699511986e-07,
"logits/chosen": -3.715214729309082,
"logits/rejected": -3.7003979682922363,
"logps/chosen": -5.552937030792236,
"logps/rejected": -28.38661766052246,
"loss": 0.0878,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.587810754776001,
"rewards/margins": 2.587810754776001,
"rewards/rejected": 0.0,
"step": 1570
},
{
"epoch": 8.776536312849162,
"grad_norm": 3.058420619499173,
"learning_rate": 5.027429367542352e-07,
"logits/chosen": -3.286238193511963,
"logits/rejected": -3.4611129760742188,
"logps/chosen": -6.222301483154297,
"logps/rejected": -40.57895278930664,
"loss": 0.1318,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6173553466796875,
"rewards/margins": 2.6173553466796875,
"rewards/rejected": 0.0,
"step": 1571
},
{
"epoch": 8.782122905027933,
"grad_norm": 2.266191007271309,
"learning_rate": 5.021333994807793e-07,
"logits/chosen": -3.624561309814453,
"logits/rejected": -3.5173733234405518,
"logps/chosen": -2.490200996398926,
"logps/rejected": -42.061866760253906,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5689961910247803,
"rewards/margins": 2.5689961910247803,
"rewards/rejected": 0.0,
"step": 1572
},
{
"epoch": 8.787709497206704,
"grad_norm": 1.1402949409937073,
"learning_rate": 5.015238590367117e-07,
"logits/chosen": -3.3423678874969482,
"logits/rejected": -3.575061798095703,
"logps/chosen": -3.269780158996582,
"logps/rejected": -32.531211853027344,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.163800001144409,
"rewards/margins": 3.163800001144409,
"rewards/rejected": 0.0,
"step": 1573
},
{
"epoch": 8.793296089385475,
"grad_norm": 3.5322342832499434,
"learning_rate": 5.009143163279183e-07,
"logits/chosen": -3.370183229446411,
"logits/rejected": -3.5936026573181152,
"logps/chosen": -0.7177491784095764,
"logps/rejected": -57.73527526855469,
"loss": 0.1397,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7981486320495605,
"rewards/margins": 1.7981486320495605,
"rewards/rejected": 0.0,
"step": 1574
},
{
"epoch": 8.798882681564246,
"grad_norm": 3.2384878827922097,
"learning_rate": 5.003047722602881e-07,
"logits/chosen": -3.5188417434692383,
"logits/rejected": -3.4905717372894287,
"logps/chosen": -7.418839454650879,
"logps/rejected": -43.177005767822266,
"loss": 0.1091,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.034696102142334,
"rewards/margins": 3.034696102142334,
"rewards/rejected": 0.0,
"step": 1575
},
{
"epoch": 8.804469273743017,
"grad_norm": 1.3958213066627738,
"learning_rate": 4.99695227739712e-07,
"logits/chosen": -3.5734057426452637,
"logits/rejected": -3.4808127880096436,
"logps/chosen": -4.616374969482422,
"logps/rejected": -71.2945556640625,
"loss": 0.103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.652176856994629,
"rewards/margins": 2.652176856994629,
"rewards/rejected": 0.0,
"step": 1576
},
{
"epoch": 8.810055865921788,
"grad_norm": 1.1838207134161463,
"learning_rate": 4.990856836720816e-07,
"logits/chosen": -3.5023655891418457,
"logits/rejected": -3.396958112716675,
"logps/chosen": -3.69647216796875,
"logps/rejected": -28.05923080444336,
"loss": 0.106,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5006635189056396,
"rewards/margins": 2.5006635189056396,
"rewards/rejected": 0.0,
"step": 1577
},
{
"epoch": 8.815642458100559,
"grad_norm": 1.417720591473832,
"learning_rate": 4.984761409632881e-07,
"logits/chosen": -3.582775592803955,
"logits/rejected": -3.546322822570801,
"logps/chosen": -5.0747480392456055,
"logps/rejected": -47.944156646728516,
"loss": 0.1128,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0520803928375244,
"rewards/margins": 3.0520803928375244,
"rewards/rejected": 0.0,
"step": 1578
},
{
"epoch": 8.82122905027933,
"grad_norm": 2.2313394840329717,
"learning_rate": 4.978666005192208e-07,
"logits/chosen": -3.461977958679199,
"logits/rejected": -3.450249671936035,
"logps/chosen": -30.081253051757812,
"logps/rejected": -53.075565338134766,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.263123035430908,
"rewards/margins": 3.263123035430908,
"rewards/rejected": 0.0,
"step": 1579
},
{
"epoch": 8.8268156424581,
"grad_norm": 4.088177642631041,
"learning_rate": 4.972570632457648e-07,
"logits/chosen": -3.4786570072174072,
"logits/rejected": -3.452631711959839,
"logps/chosen": -15.857526779174805,
"logps/rejected": -98.74070739746094,
"loss": 0.1591,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5767078399658203,
"rewards/margins": 2.5767078399658203,
"rewards/rejected": 0.0,
"step": 1580
},
{
"epoch": 8.8268156424581,
"eval_logits/chosen": -3.3498008251190186,
"eval_logits/rejected": -3.4777941703796387,
"eval_logps/chosen": -26.655345916748047,
"eval_logps/rejected": -48.50056838989258,
"eval_loss": 0.7682304978370667,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.3414694666862488,
"eval_rewards/margins": 0.3414694666862488,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7177,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 1580
},
{
"epoch": 8.832402234636872,
"grad_norm": 4.485409113305424,
"learning_rate": 4.966475300488013e-07,
"logits/chosen": -3.4706907272338867,
"logits/rejected": -3.287583112716675,
"logps/chosen": -5.174438953399658,
"logps/rejected": -85.02562713623047,
"loss": 0.1581,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.533134937286377,
"rewards/margins": 2.533134937286377,
"rewards/rejected": 0.0,
"step": 1581
},
{
"epoch": 8.837988826815643,
"grad_norm": 2.2256516999035076,
"learning_rate": 4.960380018342054e-07,
"logits/chosen": -3.248904228210449,
"logits/rejected": -3.4592931270599365,
"logps/chosen": -4.772720813751221,
"logps/rejected": -38.422508239746094,
"loss": 0.1207,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3661322593688965,
"rewards/margins": 2.3661322593688965,
"rewards/rejected": 0.0,
"step": 1582
},
{
"epoch": 8.843575418994414,
"grad_norm": 1.5716208939373597,
"learning_rate": 4.954284795078441e-07,
"logits/chosen": -3.2725048065185547,
"logits/rejected": -3.2471745014190674,
"logps/chosen": -15.56840705871582,
"logps/rejected": -39.41374969482422,
"loss": 0.1194,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6685779094696045,
"rewards/margins": 3.6685779094696045,
"rewards/rejected": 0.0,
"step": 1583
},
{
"epoch": 8.849162011173185,
"grad_norm": 3.1934728620877713,
"learning_rate": 4.948189639755766e-07,
"logits/chosen": -3.3488597869873047,
"logits/rejected": -3.3317267894744873,
"logps/chosen": -37.42424011230469,
"logps/rejected": -30.104129791259766,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6193454265594482,
"rewards/margins": 3.6193454265594482,
"rewards/rejected": 0.0,
"step": 1584
},
{
"epoch": 8.854748603351956,
"grad_norm": 1.8094745652787352,
"learning_rate": 4.942094561432512e-07,
"logits/chosen": -3.3481955528259277,
"logits/rejected": -3.2755868434906006,
"logps/chosen": -13.137596130371094,
"logps/rejected": -29.392597198486328,
"loss": 0.0965,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5587501525878906,
"rewards/margins": 3.5587501525878906,
"rewards/rejected": 0.0,
"step": 1585
},
{
"epoch": 8.860335195530727,
"grad_norm": 1.6502555052491021,
"learning_rate": 4.935999569167054e-07,
"logits/chosen": -3.6667110919952393,
"logits/rejected": -3.4561636447906494,
"logps/chosen": -1.662327766418457,
"logps/rejected": -52.39588165283203,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.203054189682007,
"rewards/margins": 2.203054189682007,
"rewards/rejected": 0.0,
"step": 1586
},
{
"epoch": 8.865921787709498,
"grad_norm": 1.357996872364283,
"learning_rate": 4.929904672017632e-07,
"logits/chosen": -3.220386028289795,
"logits/rejected": -3.1334450244903564,
"logps/chosen": -7.67182731628418,
"logps/rejected": -44.112003326416016,
"loss": 0.1114,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.803406000137329,
"rewards/margins": 3.803406000137329,
"rewards/rejected": 0.0,
"step": 1587
},
{
"epoch": 8.871508379888269,
"grad_norm": 2.3309656231730322,
"learning_rate": 4.923809879042352e-07,
"logits/chosen": -3.4371016025543213,
"logits/rejected": -3.160398483276367,
"logps/chosen": -6.324312210083008,
"logps/rejected": -31.49691390991211,
"loss": 0.0875,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0658135414123535,
"rewards/margins": 3.0658135414123535,
"rewards/rejected": 0.0,
"step": 1588
},
{
"epoch": 8.87709497206704,
"grad_norm": 1.2529366060778842,
"learning_rate": 4.917715199299159e-07,
"logits/chosen": -3.572672128677368,
"logits/rejected": -3.508350372314453,
"logps/chosen": -1.9186047315597534,
"logps/rejected": -27.849653244018555,
"loss": 0.1235,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.938552975654602,
"rewards/margins": 1.938552975654602,
"rewards/rejected": 0.0,
"step": 1589
},
{
"epoch": 8.88268156424581,
"grad_norm": 1.2374703915768543,
"learning_rate": 4.911620641845836e-07,
"logits/chosen": -3.663745403289795,
"logits/rejected": -3.6889708042144775,
"logps/chosen": -0.6582353115081787,
"logps/rejected": -80.0638427734375,
"loss": 0.1495,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2671637535095215,
"rewards/margins": 1.2671637535095215,
"rewards/rejected": 0.0,
"step": 1590
},
{
"epoch": 8.888268156424582,
"grad_norm": 2.035017885923048,
"learning_rate": 4.905526215739977e-07,
"logits/chosen": -3.5884945392608643,
"logits/rejected": -3.4904232025146484,
"logps/chosen": -18.197912216186523,
"logps/rejected": -37.696136474609375,
"loss": 0.1212,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4336118698120117,
"rewards/margins": 3.4336118698120117,
"rewards/rejected": 0.0,
"step": 1591
},
{
"epoch": 8.893854748603353,
"grad_norm": 4.704868496625444,
"learning_rate": 4.899431930038986e-07,
"logits/chosen": -3.2843191623687744,
"logits/rejected": -3.3270316123962402,
"logps/chosen": -17.2952823638916,
"logps/rejected": -68.72885131835938,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5416369438171387,
"rewards/margins": 2.5416369438171387,
"rewards/rejected": 0.0,
"step": 1592
},
{
"epoch": 8.899441340782122,
"grad_norm": 3.9141381699334965,
"learning_rate": 4.893337793800059e-07,
"logits/chosen": -3.5973403453826904,
"logits/rejected": -3.699655532836914,
"logps/chosen": -7.495997428894043,
"logps/rejected": -61.90679168701172,
"loss": 0.1119,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1217434406280518,
"rewards/margins": 3.1217434406280518,
"rewards/rejected": 0.0,
"step": 1593
},
{
"epoch": 8.905027932960895,
"grad_norm": 4.424650691551774,
"learning_rate": 4.887243816080165e-07,
"logits/chosen": -3.300236225128174,
"logits/rejected": -3.3679394721984863,
"logps/chosen": -8.020062446594238,
"logps/rejected": -63.595943450927734,
"loss": 0.1349,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.423440456390381,
"rewards/margins": 2.423440456390381,
"rewards/rejected": 0.0,
"step": 1594
},
{
"epoch": 8.910614525139664,
"grad_norm": 1.0632118047821677,
"learning_rate": 4.881150005936041e-07,
"logits/chosen": -2.886989116668701,
"logits/rejected": -2.9253146648406982,
"logps/chosen": -8.046568870544434,
"logps/rejected": -35.37300109863281,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.350053310394287,
"rewards/margins": 2.350053310394287,
"rewards/rejected": 0.0,
"step": 1595
},
{
"epoch": 8.916201117318435,
"grad_norm": 3.7207138444524617,
"learning_rate": 4.875056372424173e-07,
"logits/chosen": -3.399993896484375,
"logits/rejected": -3.3559515476226807,
"logps/chosen": -1.6196277141571045,
"logps/rejected": -125.1887435913086,
"loss": 0.0939,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.29793643951416,
"rewards/margins": 2.29793643951416,
"rewards/rejected": 0.0,
"step": 1596
},
{
"epoch": 8.921787709497206,
"grad_norm": 2.668075368170375,
"learning_rate": 4.868962924600789e-07,
"logits/chosen": -3.4909353256225586,
"logits/rejected": -3.4668478965759277,
"logps/chosen": -9.527432441711426,
"logps/rejected": -28.397212982177734,
"loss": 0.095,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1990573406219482,
"rewards/margins": 3.1990573406219482,
"rewards/rejected": 0.0,
"step": 1597
},
{
"epoch": 8.927374301675977,
"grad_norm": 2.0253597516958934,
"learning_rate": 4.862869671521834e-07,
"logits/chosen": -3.2897651195526123,
"logits/rejected": -3.4349746704101562,
"logps/chosen": -3.4705586433410645,
"logps/rejected": -33.25648498535156,
"loss": 0.1157,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.474574089050293,
"rewards/margins": 2.474574089050293,
"rewards/rejected": 0.0,
"step": 1598
},
{
"epoch": 8.932960893854748,
"grad_norm": 2.038151410844929,
"learning_rate": 4.85677662224297e-07,
"logits/chosen": -3.5211784839630127,
"logits/rejected": -3.6392548084259033,
"logps/chosen": -5.188231468200684,
"logps/rejected": -89.66943359375,
"loss": 0.1228,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9291810989379883,
"rewards/margins": 2.9291810989379883,
"rewards/rejected": 0.0,
"step": 1599
},
{
"epoch": 8.938547486033519,
"grad_norm": 0.9683098958025135,
"learning_rate": 4.85068378581955e-07,
"logits/chosen": -3.423570394515991,
"logits/rejected": -3.555612564086914,
"logps/chosen": -15.116783142089844,
"logps/rejected": -31.30979347229004,
"loss": 0.0961,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.50461745262146,
"rewards/margins": 3.50461745262146,
"rewards/rejected": 0.0,
"step": 1600
},
{
"epoch": 8.938547486033519,
"eval_logits/chosen": -3.344449996948242,
"eval_logits/rejected": -3.470318555831909,
"eval_logps/chosen": -27.05573081970215,
"eval_logps/rejected": -48.810630798339844,
"eval_loss": 0.7781283259391785,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.30143120884895325,
"eval_rewards/margins": 0.30143120884895325,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7151,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 1600
},
{
"epoch": 8.94413407821229,
"grad_norm": 2.923952822851652,
"learning_rate": 4.844591171306617e-07,
"logits/chosen": -3.3482167720794678,
"logits/rejected": -3.3730177879333496,
"logps/chosen": -12.375945091247559,
"logps/rejected": -41.884944915771484,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0999507904052734,
"rewards/margins": 3.0999507904052734,
"rewards/rejected": 0.0,
"step": 1601
},
{
"epoch": 8.949720670391061,
"grad_norm": 1.99215640247595,
"learning_rate": 4.838498787758881e-07,
"logits/chosen": -3.27443265914917,
"logits/rejected": -3.4482321739196777,
"logps/chosen": -1.7216635942459106,
"logps/rejected": -42.25780487060547,
"loss": 0.0974,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9816725254058838,
"rewards/margins": 1.9816725254058838,
"rewards/rejected": 0.0,
"step": 1602
},
{
"epoch": 8.955307262569832,
"grad_norm": 2.984793762440282,
"learning_rate": 4.832406644230707e-07,
"logits/chosen": -3.319697141647339,
"logits/rejected": -3.5623459815979004,
"logps/chosen": -4.52077579498291,
"logps/rejected": -47.441158294677734,
"loss": 0.1156,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.188465118408203,
"rewards/margins": 2.188465118408203,
"rewards/rejected": 0.0,
"step": 1603
},
{
"epoch": 8.960893854748603,
"grad_norm": 1.7438154477280894,
"learning_rate": 4.826314749776108e-07,
"logits/chosen": -3.4082627296447754,
"logits/rejected": -3.320340156555176,
"logps/chosen": -13.580062866210938,
"logps/rejected": -42.55960464477539,
"loss": 0.0786,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.284619092941284,
"rewards/margins": 3.284619092941284,
"rewards/rejected": 0.0,
"step": 1604
},
{
"epoch": 8.966480446927374,
"grad_norm": 1.8153576446453652,
"learning_rate": 4.820223113448722e-07,
"logits/chosen": -3.4017300605773926,
"logits/rejected": -3.4077565670013428,
"logps/chosen": -0.5121971368789673,
"logps/rejected": -56.97151184082031,
"loss": 0.1552,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4818192720413208,
"rewards/margins": 1.4818192720413208,
"rewards/rejected": 0.0,
"step": 1605
},
{
"epoch": 8.972067039106145,
"grad_norm": 3.761434969368255,
"learning_rate": 4.814131744301805e-07,
"logits/chosen": -3.4453651905059814,
"logits/rejected": -3.4377174377441406,
"logps/chosen": -3.7002198696136475,
"logps/rejected": -69.67536926269531,
"loss": 0.1248,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8244028091430664,
"rewards/margins": 2.8244028091430664,
"rewards/rejected": 0.0,
"step": 1606
},
{
"epoch": 8.977653631284916,
"grad_norm": 2.1432809831255684,
"learning_rate": 4.808040651388218e-07,
"logits/chosen": -3.3408074378967285,
"logits/rejected": -3.4254205226898193,
"logps/chosen": -1.5616624355316162,
"logps/rejected": -40.005062103271484,
"loss": 0.1328,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.43265438079834,
"rewards/margins": 2.43265438079834,
"rewards/rejected": 0.0,
"step": 1607
},
{
"epoch": 8.983240223463687,
"grad_norm": 2.3523641894817895,
"learning_rate": 4.801949843760409e-07,
"logits/chosen": -3.645590305328369,
"logits/rejected": -3.5588388442993164,
"logps/chosen": -6.469679355621338,
"logps/rejected": -34.304443359375,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7767696380615234,
"rewards/margins": 2.7767696380615234,
"rewards/rejected": 0.0,
"step": 1608
},
{
"epoch": 8.988826815642458,
"grad_norm": 1.2175327141580934,
"learning_rate": 4.795859330470402e-07,
"logits/chosen": -3.355254888534546,
"logits/rejected": -3.4555022716522217,
"logps/chosen": -29.797555923461914,
"logps/rejected": -46.1976318359375,
"loss": 0.0897,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2099294662475586,
"rewards/margins": 3.2099294662475586,
"rewards/rejected": 0.0,
"step": 1609
},
{
"epoch": 8.994413407821229,
"grad_norm": 1.8427440846233556,
"learning_rate": 4.789769120569785e-07,
"logits/chosen": -3.239454746246338,
"logits/rejected": -3.0870308876037598,
"logps/chosen": -5.3824357986450195,
"logps/rejected": -60.95530319213867,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3480262756347656,
"rewards/margins": 2.3480262756347656,
"rewards/rejected": 0.0,
"step": 1610
},
{
"epoch": 9.0,
"grad_norm": 3.696445974241549,
"learning_rate": 4.783679223109699e-07,
"logits/chosen": -3.4663498401641846,
"logits/rejected": -3.6648917198181152,
"logps/chosen": -2.7009425163269043,
"logps/rejected": -34.224090576171875,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5710911750793457,
"rewards/margins": 1.5710911750793457,
"rewards/rejected": 0.0,
"step": 1611
},
{
"epoch": 9.005586592178771,
"grad_norm": 1.7836807376205073,
"learning_rate": 4.777589647140808e-07,
"logits/chosen": -3.4373486042022705,
"logits/rejected": -3.411127805709839,
"logps/chosen": -6.093512535095215,
"logps/rejected": -47.09807586669922,
"loss": 0.108,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.146267890930176,
"rewards/margins": 3.146267890930176,
"rewards/rejected": 0.0,
"step": 1612
},
{
"epoch": 9.011173184357542,
"grad_norm": 1.7596652181504093,
"learning_rate": 4.77150040171331e-07,
"logits/chosen": -3.6059412956237793,
"logits/rejected": -3.623866319656372,
"logps/chosen": -2.8918423652648926,
"logps/rejected": -74.71697998046875,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2367377281188965,
"rewards/margins": 2.2367377281188965,
"rewards/rejected": 0.0,
"step": 1613
},
{
"epoch": 9.016759776536313,
"grad_norm": 3.097676263988404,
"learning_rate": 4.765411495876912e-07,
"logits/chosen": -3.3083105087280273,
"logits/rejected": -3.602318048477173,
"logps/chosen": -0.7240970134735107,
"logps/rejected": -40.536834716796875,
"loss": 0.1244,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0945539474487305,
"rewards/margins": 2.0945539474487305,
"rewards/rejected": 0.0,
"step": 1614
},
{
"epoch": 9.022346368715084,
"grad_norm": 1.4345374265099398,
"learning_rate": 4.7593229386808077e-07,
"logits/chosen": -3.74735426902771,
"logits/rejected": -3.740783929824829,
"logps/chosen": -2.399965763092041,
"logps/rejected": -33.345176696777344,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3207335472106934,
"rewards/margins": 2.3207335472106934,
"rewards/rejected": 0.0,
"step": 1615
},
{
"epoch": 9.027932960893855,
"grad_norm": 2.0122321395956786,
"learning_rate": 4.7532347391736775e-07,
"logits/chosen": -3.1974687576293945,
"logits/rejected": -3.291553020477295,
"logps/chosen": -3.7264633178710938,
"logps/rejected": -57.64251708984375,
"loss": 0.1114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3138155937194824,
"rewards/margins": 2.3138155937194824,
"rewards/rejected": 0.0,
"step": 1616
},
{
"epoch": 9.033519553072626,
"grad_norm": 1.5822959394116287,
"learning_rate": 4.7471469064036735e-07,
"logits/chosen": -3.2524325847625732,
"logits/rejected": -3.3430263996124268,
"logps/chosen": -0.9469399452209473,
"logps/rejected": -35.770469665527344,
"loss": 0.1363,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1638243198394775,
"rewards/margins": 2.1638243198394775,
"rewards/rejected": 0.0,
"step": 1617
},
{
"epoch": 9.039106145251397,
"grad_norm": 3.2426545116767067,
"learning_rate": 4.7410594494183955e-07,
"logits/chosen": -3.6374306678771973,
"logits/rejected": -3.6596550941467285,
"logps/chosen": -4.790374279022217,
"logps/rejected": -36.828216552734375,
"loss": 0.0927,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5422682762145996,
"rewards/margins": 2.5422682762145996,
"rewards/rejected": 0.0,
"step": 1618
},
{
"epoch": 9.044692737430168,
"grad_norm": 0.6697225002492078,
"learning_rate": 4.734972377264892e-07,
"logits/chosen": -3.4723641872406006,
"logits/rejected": -3.276477336883545,
"logps/chosen": -2.1712141036987305,
"logps/rejected": -37.70220947265625,
"loss": 0.1064,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.700287342071533,
"rewards/margins": 2.700287342071533,
"rewards/rejected": 0.0,
"step": 1619
},
{
"epoch": 9.050279329608939,
"grad_norm": 0.6852575228381395,
"learning_rate": 4.728885698989633e-07,
"logits/chosen": -3.3522133827209473,
"logits/rejected": -3.481334924697876,
"logps/chosen": -1.090745449066162,
"logps/rejected": -39.78439712524414,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2486796379089355,
"rewards/margins": 2.2486796379089355,
"rewards/rejected": 0.0,
"step": 1620
},
{
"epoch": 9.050279329608939,
"eval_logits/chosen": -3.3324215412139893,
"eval_logits/rejected": -3.4577934741973877,
"eval_logps/chosen": -26.813121795654297,
"eval_logps/rejected": -48.46649932861328,
"eval_loss": 0.7766733765602112,
"eval_rewards/accuracies": 0.6499999761581421,
"eval_rewards/chosen": 0.325691819190979,
"eval_rewards/margins": 0.325691819190979,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7163,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 1620
},
{
"epoch": 9.05586592178771,
"grad_norm": 1.5025125740946996,
"learning_rate": 4.7227994236385083e-07,
"logits/chosen": -3.576542854309082,
"logits/rejected": -3.614333391189575,
"logps/chosen": -3.1437058448791504,
"logps/rejected": -38.23810577392578,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.709089994430542,
"rewards/margins": 2.709089994430542,
"rewards/rejected": 0.0,
"step": 1621
},
{
"epoch": 9.061452513966481,
"grad_norm": 0.8305849529527441,
"learning_rate": 4.7167135602568086e-07,
"logits/chosen": -3.600114583969116,
"logits/rejected": -3.417306900024414,
"logps/chosen": -0.8178118467330933,
"logps/rejected": -37.192344665527344,
"loss": 0.0803,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.218064546585083,
"rewards/margins": 2.218064546585083,
"rewards/rejected": 0.0,
"step": 1622
},
{
"epoch": 9.067039106145252,
"grad_norm": 1.104244303328039,
"learning_rate": 4.710628117889207e-07,
"logits/chosen": -3.3144118785858154,
"logits/rejected": -3.446066379547119,
"logps/chosen": -7.979313850402832,
"logps/rejected": -66.49369812011719,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.54119610786438,
"rewards/margins": 3.54119610786438,
"rewards/rejected": 0.0,
"step": 1623
},
{
"epoch": 9.072625698324023,
"grad_norm": 0.6169748980165974,
"learning_rate": 4.704543105579757e-07,
"logits/chosen": -3.5144424438476562,
"logits/rejected": -3.677825927734375,
"logps/chosen": -6.750329971313477,
"logps/rejected": -47.67042922973633,
"loss": 0.0918,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5624661445617676,
"rewards/margins": 3.5624661445617676,
"rewards/rejected": 0.0,
"step": 1624
},
{
"epoch": 9.078212290502794,
"grad_norm": 1.6088784590881093,
"learning_rate": 4.6984585323718705e-07,
"logits/chosen": -3.200291633605957,
"logits/rejected": -3.248264789581299,
"logps/chosen": -1.1295030117034912,
"logps/rejected": -40.96461486816406,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1124207973480225,
"rewards/margins": 2.1124207973480225,
"rewards/rejected": 0.0,
"step": 1625
},
{
"epoch": 9.083798882681565,
"grad_norm": 1.4466208468231971,
"learning_rate": 4.692374407308307e-07,
"logits/chosen": -3.436687707901001,
"logits/rejected": -3.4873499870300293,
"logps/chosen": -6.9484543800354,
"logps/rejected": -45.24601364135742,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7573416233062744,
"rewards/margins": 2.7573416233062744,
"rewards/rejected": 0.0,
"step": 1626
},
{
"epoch": 9.089385474860336,
"grad_norm": 1.031143712645248,
"learning_rate": 4.686290739431157e-07,
"logits/chosen": -3.438779354095459,
"logits/rejected": -3.37857985496521,
"logps/chosen": -11.256845474243164,
"logps/rejected": -40.809532165527344,
"loss": 0.113,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0104825496673584,
"rewards/margins": 3.0104825496673584,
"rewards/rejected": 0.0,
"step": 1627
},
{
"epoch": 9.094972067039107,
"grad_norm": 1.5021763413826712,
"learning_rate": 4.68020753778184e-07,
"logits/chosen": -3.7217040061950684,
"logits/rejected": -3.7122786045074463,
"logps/chosen": -2.4026668071746826,
"logps/rejected": -53.36454772949219,
"loss": 0.0957,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.717082977294922,
"rewards/margins": 2.717082977294922,
"rewards/rejected": 0.0,
"step": 1628
},
{
"epoch": 9.100558659217878,
"grad_norm": 0.6685998811669954,
"learning_rate": 4.674124811401071e-07,
"logits/chosen": -3.4131436347961426,
"logits/rejected": -3.5219039916992188,
"logps/chosen": -3.5351905822753906,
"logps/rejected": -88.4783706665039,
"loss": 0.1036,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.042217493057251,
"rewards/margins": 3.042217493057251,
"rewards/rejected": 0.0,
"step": 1629
},
{
"epoch": 9.106145251396647,
"grad_norm": 1.0102987999404707,
"learning_rate": 4.6680425693288683e-07,
"logits/chosen": -3.4946987628936768,
"logits/rejected": -3.326328992843628,
"logps/chosen": -3.51169753074646,
"logps/rejected": -70.41123962402344,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.480116844177246,
"rewards/margins": 2.480116844177246,
"rewards/rejected": 0.0,
"step": 1630
},
{
"epoch": 9.111731843575418,
"grad_norm": 0.8886492547646674,
"learning_rate": 4.6619608206045275e-07,
"logits/chosen": -3.7399768829345703,
"logits/rejected": -3.654049873352051,
"logps/chosen": -10.962431907653809,
"logps/rejected": -51.43917465209961,
"loss": 0.0729,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.465006113052368,
"rewards/margins": 3.465006113052368,
"rewards/rejected": 0.0,
"step": 1631
},
{
"epoch": 9.11731843575419,
"grad_norm": 1.4639677826571171,
"learning_rate": 4.655879574266607e-07,
"logits/chosen": -3.641827344894409,
"logits/rejected": -3.5533337593078613,
"logps/chosen": -17.22907257080078,
"logps/rejected": -40.89700698852539,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.298551321029663,
"rewards/margins": 3.298551321029663,
"rewards/rejected": 0.0,
"step": 1632
},
{
"epoch": 9.12290502793296,
"grad_norm": 0.8847487547987789,
"learning_rate": 4.649798839352923e-07,
"logits/chosen": -3.5922327041625977,
"logits/rejected": -3.620138645172119,
"logps/chosen": -3.5823731422424316,
"logps/rejected": -13.908525466918945,
"loss": 0.094,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8184237480163574,
"rewards/margins": 2.8184237480163574,
"rewards/rejected": 0.0,
"step": 1633
},
{
"epoch": 9.128491620111731,
"grad_norm": 0.983025461046181,
"learning_rate": 4.643718624900534e-07,
"logits/chosen": -3.3321423530578613,
"logits/rejected": -3.5353293418884277,
"logps/chosen": -5.316637992858887,
"logps/rejected": -54.03192138671875,
"loss": 0.1063,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1418213844299316,
"rewards/margins": 3.1418213844299316,
"rewards/rejected": 0.0,
"step": 1634
},
{
"epoch": 9.134078212290502,
"grad_norm": 1.0651140276522517,
"learning_rate": 4.6376389399457175e-07,
"logits/chosen": -3.2532856464385986,
"logits/rejected": -3.3750452995300293,
"logps/chosen": -27.882795333862305,
"logps/rejected": -73.39027404785156,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.414515495300293,
"rewards/margins": 2.414515495300293,
"rewards/rejected": 0.0,
"step": 1635
},
{
"epoch": 9.139664804469273,
"grad_norm": 1.0352375591863479,
"learning_rate": 4.6315597935239673e-07,
"logits/chosen": -3.720578670501709,
"logits/rejected": -3.4379751682281494,
"logps/chosen": -23.957807540893555,
"logps/rejected": -45.56193923950195,
"loss": 0.1411,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.235013961791992,
"rewards/margins": 4.235013961791992,
"rewards/rejected": 0.0,
"step": 1636
},
{
"epoch": 9.145251396648044,
"grad_norm": 0.8748613766094245,
"learning_rate": 4.6254811946699825e-07,
"logits/chosen": -3.546139717102051,
"logits/rejected": -3.5422675609588623,
"logps/chosen": -4.041231632232666,
"logps/rejected": -56.439640045166016,
"loss": 0.0857,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1619534492492676,
"rewards/margins": 3.1619534492492676,
"rewards/rejected": 0.0,
"step": 1637
},
{
"epoch": 9.150837988826815,
"grad_norm": 0.9517846054201481,
"learning_rate": 4.619403152417638e-07,
"logits/chosen": -3.819573163986206,
"logits/rejected": -3.7253293991088867,
"logps/chosen": -1.1498639583587646,
"logps/rejected": -24.196958541870117,
"loss": 0.1014,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2579398155212402,
"rewards/margins": 2.2579398155212402,
"rewards/rejected": 0.0,
"step": 1638
},
{
"epoch": 9.156424581005586,
"grad_norm": 1.0271047737056191,
"learning_rate": 4.6133256757999894e-07,
"logits/chosen": -3.3910739421844482,
"logits/rejected": -3.5881876945495605,
"logps/chosen": -4.272151470184326,
"logps/rejected": -36.314308166503906,
"loss": 0.0789,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5699243545532227,
"rewards/margins": 2.5699243545532227,
"rewards/rejected": 0.0,
"step": 1639
},
{
"epoch": 9.162011173184357,
"grad_norm": 0.7567277389197201,
"learning_rate": 4.607248773849252e-07,
"logits/chosen": -3.5559473037719727,
"logits/rejected": -3.5395753383636475,
"logps/chosen": -2.048529624938965,
"logps/rejected": -62.623008728027344,
"loss": 0.1226,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2905678749084473,
"rewards/margins": 2.2905678749084473,
"rewards/rejected": 0.0,
"step": 1640
},
{
"epoch": 9.162011173184357,
"eval_logits/chosen": -3.3215668201446533,
"eval_logits/rejected": -3.450575590133667,
"eval_logps/chosen": -27.634836196899414,
"eval_logps/rejected": -50.139163970947266,
"eval_loss": 0.8401439189910889,
"eval_rewards/accuracies": 0.6499999761581421,
"eval_rewards/chosen": 0.2435208559036255,
"eval_rewards/margins": 0.2435208559036255,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6858,
"eval_samples_per_second": 9.484,
"eval_steps_per_second": 0.306,
"step": 1640
},
{
"epoch": 9.167597765363128,
"grad_norm": 0.6652376165818646,
"learning_rate": 4.60117245559678e-07,
"logits/chosen": -3.4068853855133057,
"logits/rejected": -3.4603068828582764,
"logps/chosen": -3.901829957962036,
"logps/rejected": -61.692291259765625,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3450779914855957,
"rewards/margins": 2.3450779914855957,
"rewards/rejected": 0.0,
"step": 1641
},
{
"epoch": 9.1731843575419,
"grad_norm": 2.0880314630346977,
"learning_rate": 4.595096730073066e-07,
"logits/chosen": -3.1529829502105713,
"logits/rejected": -3.1645779609680176,
"logps/chosen": -23.6221981048584,
"logps/rejected": -31.06586265563965,
"loss": 0.1288,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6671934127807617,
"rewards/margins": 2.6671934127807617,
"rewards/rejected": 0.0,
"step": 1642
},
{
"epoch": 9.17877094972067,
"grad_norm": 2.6801130903246286,
"learning_rate": 4.5890216063077227e-07,
"logits/chosen": -3.447528123855591,
"logits/rejected": -3.3506224155426025,
"logps/chosen": -1.7273507118225098,
"logps/rejected": -56.97646713256836,
"loss": 0.1228,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.118861198425293,
"rewards/margins": 2.118861198425293,
"rewards/rejected": 0.0,
"step": 1643
},
{
"epoch": 9.184357541899441,
"grad_norm": 0.6879957323627479,
"learning_rate": 4.582947093329461e-07,
"logits/chosen": -3.5328001976013184,
"logits/rejected": -3.387763738632202,
"logps/chosen": -7.785747528076172,
"logps/rejected": -25.19640350341797,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3997373580932617,
"rewards/margins": 3.3997373580932617,
"rewards/rejected": 0.0,
"step": 1644
},
{
"epoch": 9.189944134078212,
"grad_norm": 0.6458114477243525,
"learning_rate": 4.576873200166093e-07,
"logits/chosen": -3.577589750289917,
"logits/rejected": -3.549164295196533,
"logps/chosen": -15.643972396850586,
"logps/rejected": -36.84091567993164,
"loss": 0.1004,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.667699098587036,
"rewards/margins": 3.667699098587036,
"rewards/rejected": 0.0,
"step": 1645
},
{
"epoch": 9.195530726256983,
"grad_norm": 0.653742575908524,
"learning_rate": 4.570799935844506e-07,
"logits/chosen": -3.426997184753418,
"logits/rejected": -3.263404369354248,
"logps/chosen": -19.646692276000977,
"logps/rejected": -34.58079147338867,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6539430618286133,
"rewards/margins": 3.6539430618286133,
"rewards/rejected": 0.0,
"step": 1646
},
{
"epoch": 9.201117318435754,
"grad_norm": 0.6338413469102935,
"learning_rate": 4.564727309390648e-07,
"logits/chosen": -3.516860246658325,
"logits/rejected": -3.3906006813049316,
"logps/chosen": -23.571407318115234,
"logps/rejected": -71.54456329345703,
"loss": 0.0761,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.067446708679199,
"rewards/margins": 4.067446708679199,
"rewards/rejected": 0.0,
"step": 1647
},
{
"epoch": 9.206703910614525,
"grad_norm": 1.3750926237769714,
"learning_rate": 4.5586553298295284e-07,
"logits/chosen": -3.391157388687134,
"logits/rejected": -3.4704341888427734,
"logps/chosen": -0.6388237476348877,
"logps/rejected": -45.884033203125,
"loss": 0.1029,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.884291410446167,
"rewards/margins": 1.884291410446167,
"rewards/rejected": 0.0,
"step": 1648
},
{
"epoch": 9.212290502793296,
"grad_norm": 2.2531424336147983,
"learning_rate": 4.552584006185186e-07,
"logits/chosen": -3.6195030212402344,
"logits/rejected": -3.72220778465271,
"logps/chosen": -1.3236113786697388,
"logps/rejected": -73.94574737548828,
"loss": 0.0874,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1073408126831055,
"rewards/margins": 2.1073408126831055,
"rewards/rejected": 0.0,
"step": 1649
},
{
"epoch": 9.217877094972067,
"grad_norm": 1.483944660137312,
"learning_rate": 4.54651334748069e-07,
"logits/chosen": -3.4388511180877686,
"logits/rejected": -3.500291109085083,
"logps/chosen": -1.9012879133224487,
"logps/rejected": -39.856170654296875,
"loss": 0.0968,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5029850006103516,
"rewards/margins": 2.5029850006103516,
"rewards/rejected": 0.0,
"step": 1650
},
{
"epoch": 9.223463687150838,
"grad_norm": 1.4906533168379197,
"learning_rate": 4.540443362738123e-07,
"logits/chosen": -3.5285568237304688,
"logits/rejected": -3.536672353744507,
"logps/chosen": -7.189120769500732,
"logps/rejected": -36.30305480957031,
"loss": 0.0956,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1984524726867676,
"rewards/margins": 3.1984524726867676,
"rewards/rejected": 0.0,
"step": 1651
},
{
"epoch": 9.22905027932961,
"grad_norm": 1.009822880276252,
"learning_rate": 4.534374060978561e-07,
"logits/chosen": -3.6347761154174805,
"logits/rejected": -3.5859129428863525,
"logps/chosen": -3.876652240753174,
"logps/rejected": -57.586734771728516,
"loss": 0.0873,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.904188632965088,
"rewards/margins": 2.904188632965088,
"rewards/rejected": 0.0,
"step": 1652
},
{
"epoch": 9.23463687150838,
"grad_norm": 1.0512488552773978,
"learning_rate": 4.5283054512220666e-07,
"logits/chosen": -3.4411423206329346,
"logits/rejected": -3.469578981399536,
"logps/chosen": -0.7947466373443604,
"logps/rejected": -97.12767028808594,
"loss": 0.1436,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.72902250289917,
"rewards/margins": 1.72902250289917,
"rewards/rejected": 0.0,
"step": 1653
},
{
"epoch": 9.240223463687151,
"grad_norm": 1.046013283113762,
"learning_rate": 4.5222375424876783e-07,
"logits/chosen": -3.492800712585449,
"logits/rejected": -3.6472015380859375,
"logps/chosen": -13.19853687286377,
"logps/rejected": -42.79853820800781,
"loss": 0.0968,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3905725479125977,
"rewards/margins": 3.3905725479125977,
"rewards/rejected": 0.0,
"step": 1654
},
{
"epoch": 9.245810055865922,
"grad_norm": 0.6472985246797178,
"learning_rate": 4.516170343793386e-07,
"logits/chosen": -3.4796149730682373,
"logits/rejected": -3.7505271434783936,
"logps/chosen": -3.8719398975372314,
"logps/rejected": -54.36867141723633,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4371376037597656,
"rewards/margins": 3.4371376037597656,
"rewards/rejected": 0.0,
"step": 1655
},
{
"epoch": 9.251396648044693,
"grad_norm": 1.1471450171632256,
"learning_rate": 4.510103864156127e-07,
"logits/chosen": -3.6489460468292236,
"logits/rejected": -3.6754863262176514,
"logps/chosen": -4.7925190925598145,
"logps/rejected": -50.89821243286133,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.939988851547241,
"rewards/margins": 2.939988851547241,
"rewards/rejected": 0.0,
"step": 1656
},
{
"epoch": 9.256983240223464,
"grad_norm": 1.5275573364432178,
"learning_rate": 4.5040381125917733e-07,
"logits/chosen": -3.6592698097229004,
"logits/rejected": -3.6023573875427246,
"logps/chosen": -6.153960227966309,
"logps/rejected": -22.277875900268555,
"loss": 0.0762,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3746848106384277,
"rewards/margins": 3.3746848106384277,
"rewards/rejected": 0.0,
"step": 1657
},
{
"epoch": 9.262569832402235,
"grad_norm": 3.659832937503148,
"learning_rate": 4.497973098115108e-07,
"logits/chosen": -3.221414804458618,
"logits/rejected": -3.2071645259857178,
"logps/chosen": -1.912853479385376,
"logps/rejected": -43.9027099609375,
"loss": 0.1187,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.023972511291504,
"rewards/margins": 2.023972511291504,
"rewards/rejected": 0.0,
"step": 1658
},
{
"epoch": 9.268156424581006,
"grad_norm": 0.9419207384234801,
"learning_rate": 4.491908829739826e-07,
"logits/chosen": -3.550814628601074,
"logits/rejected": -3.6960623264312744,
"logps/chosen": -1.5808019638061523,
"logps/rejected": -105.10787963867188,
"loss": 0.126,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.259671449661255,
"rewards/margins": 2.259671449661255,
"rewards/rejected": 0.0,
"step": 1659
},
{
"epoch": 9.273743016759777,
"grad_norm": 0.9918028558746579,
"learning_rate": 4.48584531647851e-07,
"logits/chosen": -3.4608771800994873,
"logits/rejected": -3.5884146690368652,
"logps/chosen": -2.104559898376465,
"logps/rejected": -50.12146759033203,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.90057373046875,
"rewards/margins": 2.90057373046875,
"rewards/rejected": 0.0,
"step": 1660
},
{
"epoch": 9.273743016759777,
"eval_logits/chosen": -3.320514678955078,
"eval_logits/rejected": -3.4494194984436035,
"eval_logps/chosen": -27.093725204467773,
"eval_logps/rejected": -50.0571403503418,
"eval_loss": 0.822568416595459,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.2976315915584564,
"eval_rewards/margins": 0.2976315915584564,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.728,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 1660
},
{
"epoch": 9.279329608938548,
"grad_norm": 0.7978783107009588,
"learning_rate": 4.479782567342619e-07,
"logits/chosen": -3.4343059062957764,
"logits/rejected": -3.4542601108551025,
"logps/chosen": -2.98053240776062,
"logps/rejected": -68.02059936523438,
"loss": 0.0815,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.34826397895813,
"rewards/margins": 2.34826397895813,
"rewards/rejected": 0.0,
"step": 1661
},
{
"epoch": 9.28491620111732,
"grad_norm": 1.215827653811134,
"learning_rate": 4.4737205913424775e-07,
"logits/chosen": -3.477102041244507,
"logits/rejected": -3.492522954940796,
"logps/chosen": -1.10588538646698,
"logps/rejected": -28.390304565429688,
"loss": 0.1224,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6332898139953613,
"rewards/margins": 2.6332898139953613,
"rewards/rejected": 0.0,
"step": 1662
},
{
"epoch": 9.29050279329609,
"grad_norm": 4.523770466035961,
"learning_rate": 4.4676593974872647e-07,
"logits/chosen": -3.2639989852905273,
"logits/rejected": -3.5039234161376953,
"logps/chosen": -1.6709858179092407,
"logps/rejected": -32.406158447265625,
"loss": 0.1265,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.325526237487793,
"rewards/margins": 2.325526237487793,
"rewards/rejected": 0.0,
"step": 1663
},
{
"epoch": 9.296089385474861,
"grad_norm": 1.1902724242122853,
"learning_rate": 4.46159899478499e-07,
"logits/chosen": -3.3042995929718018,
"logits/rejected": -3.3759191036224365,
"logps/chosen": -1.20881187915802,
"logps/rejected": -48.647247314453125,
"loss": 0.1238,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.325157403945923,
"rewards/margins": 2.325157403945923,
"rewards/rejected": 0.0,
"step": 1664
},
{
"epoch": 9.30167597765363,
"grad_norm": 3.9169471879173385,
"learning_rate": 4.455539392242494e-07,
"logits/chosen": -3.4400534629821777,
"logits/rejected": -3.3671622276306152,
"logps/chosen": -4.663588047027588,
"logps/rejected": -29.19081687927246,
"loss": 0.0897,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0808463096618652,
"rewards/margins": 3.0808463096618652,
"rewards/rejected": 0.0,
"step": 1665
},
{
"epoch": 9.307262569832401,
"grad_norm": 1.4183777468015464,
"learning_rate": 4.4494805988654256e-07,
"logits/chosen": -3.6820247173309326,
"logits/rejected": -3.5284478664398193,
"logps/chosen": -0.8771708011627197,
"logps/rejected": -41.76000213623047,
"loss": 0.1395,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9157788753509521,
"rewards/margins": 1.9157788753509521,
"rewards/rejected": 0.0,
"step": 1666
},
{
"epoch": 9.312849162011172,
"grad_norm": 0.6651252353283887,
"learning_rate": 4.443422623658227e-07,
"logits/chosen": -3.6375091075897217,
"logits/rejected": -3.6346709728240967,
"logps/chosen": -7.7690935134887695,
"logps/rejected": -25.248586654663086,
"loss": 0.1312,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6729512214660645,
"rewards/margins": 2.6729512214660645,
"rewards/rejected": 0.0,
"step": 1667
},
{
"epoch": 9.318435754189943,
"grad_norm": 0.9617451051965464,
"learning_rate": 4.437365475624131e-07,
"logits/chosen": -3.404390335083008,
"logits/rejected": -3.368523597717285,
"logps/chosen": -5.0989508628845215,
"logps/rejected": -44.76288604736328,
"loss": 0.1224,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.140861988067627,
"rewards/margins": 3.140861988067627,
"rewards/rejected": 0.0,
"step": 1668
},
{
"epoch": 9.324022346368714,
"grad_norm": 1.001034236506721,
"learning_rate": 4.4313091637651394e-07,
"logits/chosen": -3.5085628032684326,
"logits/rejected": -3.540684938430786,
"logps/chosen": -5.532584190368652,
"logps/rejected": -43.663787841796875,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.869607925415039,
"rewards/margins": 2.869607925415039,
"rewards/rejected": 0.0,
"step": 1669
},
{
"epoch": 9.329608938547485,
"grad_norm": 0.705774360061405,
"learning_rate": 4.425253697082007e-07,
"logits/chosen": -3.2584965229034424,
"logits/rejected": -3.2584667205810547,
"logps/chosen": -12.138728141784668,
"logps/rejected": -36.21665954589844,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.29658579826355,
"rewards/margins": 3.29658579826355,
"rewards/rejected": 0.0,
"step": 1670
},
{
"epoch": 9.335195530726256,
"grad_norm": 0.8504428691008185,
"learning_rate": 4.419199084574237e-07,
"logits/chosen": -3.736549139022827,
"logits/rejected": -3.687662124633789,
"logps/chosen": -4.930518627166748,
"logps/rejected": -64.18195343017578,
"loss": 0.117,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6274194717407227,
"rewards/margins": 2.6274194717407227,
"rewards/rejected": 0.0,
"step": 1671
},
{
"epoch": 9.340782122905027,
"grad_norm": 0.9255245693297752,
"learning_rate": 4.4131453352400656e-07,
"logits/chosen": -3.119297981262207,
"logits/rejected": -3.45367431640625,
"logps/chosen": -1.8677343130111694,
"logps/rejected": -39.497676849365234,
"loss": 0.0993,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3898773193359375,
"rewards/margins": 2.3898773193359375,
"rewards/rejected": 0.0,
"step": 1672
},
{
"epoch": 9.346368715083798,
"grad_norm": 1.0796381261006927,
"learning_rate": 4.407092458076436e-07,
"logits/chosen": -3.3355917930603027,
"logits/rejected": -3.4962973594665527,
"logps/chosen": -9.620408058166504,
"logps/rejected": -50.528167724609375,
"loss": 0.1316,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2583227157592773,
"rewards/margins": 3.2583227157592773,
"rewards/rejected": 0.0,
"step": 1673
},
{
"epoch": 9.35195530726257,
"grad_norm": 0.979557178370967,
"learning_rate": 4.401040462079006e-07,
"logits/chosen": -3.594233512878418,
"logits/rejected": -3.315873384475708,
"logps/chosen": -1.0419033765792847,
"logps/rejected": -37.96501922607422,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4149818420410156,
"rewards/margins": 2.4149818420410156,
"rewards/rejected": 0.0,
"step": 1674
},
{
"epoch": 9.35754189944134,
"grad_norm": 2.3850754165564654,
"learning_rate": 4.394989356242119e-07,
"logits/chosen": -3.544698476791382,
"logits/rejected": -3.6464734077453613,
"logps/chosen": -1.730089545249939,
"logps/rejected": -61.054046630859375,
"loss": 0.1262,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.486802577972412,
"rewards/margins": 2.486802577972412,
"rewards/rejected": 0.0,
"step": 1675
},
{
"epoch": 9.363128491620111,
"grad_norm": 0.7804655522429714,
"learning_rate": 4.388939149558794e-07,
"logits/chosen": -3.38547945022583,
"logits/rejected": -3.43357253074646,
"logps/chosen": -0.42281386256217957,
"logps/rejected": -62.40565872192383,
"loss": 0.1378,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6719470024108887,
"rewards/margins": 1.6719470024108887,
"rewards/rejected": 0.0,
"step": 1676
},
{
"epoch": 9.368715083798882,
"grad_norm": 0.6369642249548819,
"learning_rate": 4.3828898510207177e-07,
"logits/chosen": -3.666167974472046,
"logits/rejected": -3.686500310897827,
"logps/chosen": -9.158576011657715,
"logps/rejected": -47.06145477294922,
"loss": 0.0987,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7776994705200195,
"rewards/margins": 2.7776994705200195,
"rewards/rejected": 0.0,
"step": 1677
},
{
"epoch": 9.374301675977653,
"grad_norm": 0.7228971001669937,
"learning_rate": 4.376841469618223e-07,
"logits/chosen": -3.5081560611724854,
"logits/rejected": -3.5118141174316406,
"logps/chosen": -2.3876757621765137,
"logps/rejected": -44.77592849731445,
"loss": 0.0959,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8737363815307617,
"rewards/margins": 2.8737363815307617,
"rewards/rejected": 0.0,
"step": 1678
},
{
"epoch": 9.379888268156424,
"grad_norm": 2.1531995223331353,
"learning_rate": 4.370794014340282e-07,
"logits/chosen": -3.298464298248291,
"logits/rejected": -3.3193044662475586,
"logps/chosen": -19.938934326171875,
"logps/rejected": -40.75102996826172,
"loss": 0.1425,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9173717498779297,
"rewards/margins": 2.9173717498779297,
"rewards/rejected": 0.0,
"step": 1679
},
{
"epoch": 9.385474860335195,
"grad_norm": 3.2467322257414124,
"learning_rate": 4.3647474941744945e-07,
"logits/chosen": -3.4913649559020996,
"logits/rejected": -3.5214719772338867,
"logps/chosen": -0.6102027297019958,
"logps/rejected": -52.77445602416992,
"loss": 0.0825,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9872210025787354,
"rewards/margins": 1.9872210025787354,
"rewards/rejected": 0.0,
"step": 1680
},
{
"epoch": 9.385474860335195,
"eval_logits/chosen": -3.311461925506592,
"eval_logits/rejected": -3.4402499198913574,
"eval_logps/chosen": -27.66586685180664,
"eval_logps/rejected": -50.32343292236328,
"eval_loss": 0.8500019907951355,
"eval_rewards/accuracies": 0.6499999761581421,
"eval_rewards/chosen": 0.24041762948036194,
"eval_rewards/margins": 0.24041762948036194,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6971,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 1680
},
{
"epoch": 9.391061452513966,
"grad_norm": 1.137546018759431,
"learning_rate": 4.3587019181070593e-07,
"logits/chosen": -3.4446256160736084,
"logits/rejected": -3.4689605236053467,
"logps/chosen": -0.7464298605918884,
"logps/rejected": -40.93949890136719,
"loss": 0.133,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8554067611694336,
"rewards/margins": 1.8554067611694336,
"rewards/rejected": 0.0,
"step": 1681
},
{
"epoch": 9.396648044692737,
"grad_norm": 2.124718311740144,
"learning_rate": 4.3526572951227835e-07,
"logits/chosen": -3.2802059650421143,
"logits/rejected": -3.5219709873199463,
"logps/chosen": -2.9198992252349854,
"logps/rejected": -52.50535202026367,
"loss": 0.0816,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9849014282226562,
"rewards/margins": 2.9849014282226562,
"rewards/rejected": 0.0,
"step": 1682
},
{
"epoch": 9.402234636871508,
"grad_norm": 1.1790212523455037,
"learning_rate": 4.3466136342050545e-07,
"logits/chosen": -3.5371663570404053,
"logits/rejected": -3.54402494430542,
"logps/chosen": -0.9409477710723877,
"logps/rejected": -40.8814697265625,
"loss": 0.1204,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1401760578155518,
"rewards/margins": 2.1401760578155518,
"rewards/rejected": 0.0,
"step": 1683
},
{
"epoch": 9.40782122905028,
"grad_norm": 0.9343191536862998,
"learning_rate": 4.340570944335825e-07,
"logits/chosen": -3.3438329696655273,
"logits/rejected": -3.513079881668091,
"logps/chosen": -2.165937900543213,
"logps/rejected": -43.00746154785156,
"loss": 0.1067,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3748960494995117,
"rewards/margins": 2.3748960494995117,
"rewards/rejected": 0.0,
"step": 1684
},
{
"epoch": 9.41340782122905,
"grad_norm": 1.889334067259178,
"learning_rate": 4.334529234495611e-07,
"logits/chosen": -3.2260501384735107,
"logits/rejected": -3.549238443374634,
"logps/chosen": -4.397539138793945,
"logps/rejected": -42.89970397949219,
"loss": 0.1042,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7786693572998047,
"rewards/margins": 2.7786693572998047,
"rewards/rejected": 0.0,
"step": 1685
},
{
"epoch": 9.418994413407821,
"grad_norm": 1.9003856935501429,
"learning_rate": 4.3284885136634697e-07,
"logits/chosen": -3.3984479904174805,
"logits/rejected": -3.4613521099090576,
"logps/chosen": -11.049154281616211,
"logps/rejected": -48.33641052246094,
"loss": 0.1566,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3414509296417236,
"rewards/margins": 2.3414509296417236,
"rewards/rejected": 0.0,
"step": 1686
},
{
"epoch": 9.424581005586592,
"grad_norm": 0.6644543604790718,
"learning_rate": 4.322448790816984e-07,
"logits/chosen": -3.161001682281494,
"logits/rejected": -3.209975242614746,
"logps/chosen": -24.673967361450195,
"logps/rejected": -42.61569595336914,
"loss": 0.1202,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0251245498657227,
"rewards/margins": 3.0251245498657227,
"rewards/rejected": 0.0,
"step": 1687
},
{
"epoch": 9.430167597765363,
"grad_norm": 0.8299100994106664,
"learning_rate": 4.3164100749322614e-07,
"logits/chosen": -3.7852985858917236,
"logits/rejected": -3.730530023574829,
"logps/chosen": -7.574895858764648,
"logps/rejected": -30.36754035949707,
"loss": 0.0703,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8281028270721436,
"rewards/margins": 3.8281028270721436,
"rewards/rejected": 0.0,
"step": 1688
},
{
"epoch": 9.435754189944134,
"grad_norm": 2.6184409693735207,
"learning_rate": 4.310372374983909e-07,
"logits/chosen": -3.561497926712036,
"logits/rejected": -3.6270360946655273,
"logps/chosen": -1.438293218612671,
"logps/rejected": -54.53354263305664,
"loss": 0.1259,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8125982284545898,
"rewards/margins": 1.8125982284545898,
"rewards/rejected": 0.0,
"step": 1689
},
{
"epoch": 9.441340782122905,
"grad_norm": 0.6631884411592908,
"learning_rate": 4.3043356999450223e-07,
"logits/chosen": -3.448148727416992,
"logits/rejected": -3.5284643173217773,
"logps/chosen": -2.6424715518951416,
"logps/rejected": -100.07594299316406,
"loss": 0.0941,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3525705337524414,
"rewards/margins": 2.3525705337524414,
"rewards/rejected": 0.0,
"step": 1690
},
{
"epoch": 9.446927374301676,
"grad_norm": 1.1318458693737592,
"learning_rate": 4.298300058787176e-07,
"logits/chosen": -3.44157075881958,
"logits/rejected": -3.594092607498169,
"logps/chosen": -2.2108848094940186,
"logps/rejected": -79.8294677734375,
"loss": 0.103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4404940605163574,
"rewards/margins": 2.4404940605163574,
"rewards/rejected": 0.0,
"step": 1691
},
{
"epoch": 9.452513966480447,
"grad_norm": 1.4703698634310705,
"learning_rate": 4.292265460480411e-07,
"logits/chosen": -3.5306217670440674,
"logits/rejected": -3.624938726425171,
"logps/chosen": -0.5446736812591553,
"logps/rejected": -133.71543884277344,
"loss": 0.0967,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7631876468658447,
"rewards/margins": 1.7631876468658447,
"rewards/rejected": 0.0,
"step": 1692
},
{
"epoch": 9.458100558659218,
"grad_norm": 3.305936388380418,
"learning_rate": 4.2862319139932094e-07,
"logits/chosen": -3.560810089111328,
"logits/rejected": -3.534581184387207,
"logps/chosen": -4.433304309844971,
"logps/rejected": -50.900962829589844,
"loss": 0.0845,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.496093273162842,
"rewards/margins": 3.496093273162842,
"rewards/rejected": 0.0,
"step": 1693
},
{
"epoch": 9.46368715083799,
"grad_norm": 0.7327068207557769,
"learning_rate": 4.2801994282925e-07,
"logits/chosen": -3.4338550567626953,
"logits/rejected": -3.4327404499053955,
"logps/chosen": -7.381734848022461,
"logps/rejected": -40.26165771484375,
"loss": 0.1295,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.829134941101074,
"rewards/margins": 2.829134941101074,
"rewards/rejected": 0.0,
"step": 1694
},
{
"epoch": 9.46927374301676,
"grad_norm": 4.163846340676333,
"learning_rate": 4.27416801234363e-07,
"logits/chosen": -3.3018383979797363,
"logits/rejected": -3.4342386722564697,
"logps/chosen": -2.5537142753601074,
"logps/rejected": -42.40884017944336,
"loss": 0.1354,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7506661415100098,
"rewards/margins": 2.7506661415100098,
"rewards/rejected": 0.0,
"step": 1695
},
{
"epoch": 9.474860335195531,
"grad_norm": 0.6920824321518683,
"learning_rate": 4.2681376751103573e-07,
"logits/chosen": -3.2477638721466064,
"logits/rejected": -3.4153714179992676,
"logps/chosen": -1.4177420139312744,
"logps/rejected": -82.40536499023438,
"loss": 0.1037,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5145020484924316,
"rewards/margins": 2.5145020484924316,
"rewards/rejected": 0.0,
"step": 1696
},
{
"epoch": 9.480446927374302,
"grad_norm": 0.6608240332889246,
"learning_rate": 4.2621084255548356e-07,
"logits/chosen": -3.209749698638916,
"logits/rejected": -3.36749005317688,
"logps/chosen": -4.849580764770508,
"logps/rejected": -69.09649658203125,
"loss": 0.0756,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.505600929260254,
"rewards/margins": 2.505600929260254,
"rewards/rejected": 0.0,
"step": 1697
},
{
"epoch": 9.486033519553073,
"grad_norm": 0.8748741223269052,
"learning_rate": 4.2560802726376066e-07,
"logits/chosen": -3.658421754837036,
"logits/rejected": -3.5597047805786133,
"logps/chosen": -7.290219306945801,
"logps/rejected": -51.058555603027344,
"loss": 0.0835,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.701432466506958,
"rewards/margins": 3.701432466506958,
"rewards/rejected": 0.0,
"step": 1698
},
{
"epoch": 9.491620111731844,
"grad_norm": 0.7820758644196301,
"learning_rate": 4.250053225317576e-07,
"logits/chosen": -3.325294017791748,
"logits/rejected": -3.598661184310913,
"logps/chosen": -4.855356693267822,
"logps/rejected": -27.665653228759766,
"loss": 0.0898,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.507937431335449,
"rewards/margins": 2.507937431335449,
"rewards/rejected": 0.0,
"step": 1699
},
{
"epoch": 9.497206703910614,
"grad_norm": 4.139095095118339,
"learning_rate": 4.24402729255201e-07,
"logits/chosen": -3.417232036590576,
"logits/rejected": -3.365792751312256,
"logps/chosen": -13.969809532165527,
"logps/rejected": -66.77743530273438,
"loss": 0.1345,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.26210355758667,
"rewards/margins": 2.26210355758667,
"rewards/rejected": 0.0,
"step": 1700
},
{
"epoch": 9.497206703910614,
"eval_logits/chosen": -3.3096587657928467,
"eval_logits/rejected": -3.436217784881592,
"eval_logps/chosen": -27.44838523864746,
"eval_logps/rejected": -50.46533966064453,
"eval_loss": 0.8455008268356323,
"eval_rewards/accuracies": 0.699999988079071,
"eval_rewards/chosen": 0.2621658444404602,
"eval_rewards/margins": 0.2621658444404602,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7289,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 1700
},
{
"epoch": 9.502793296089386,
"grad_norm": 1.412572136058337,
"learning_rate": 4.238002483296521e-07,
"logits/chosen": -3.4035112857818604,
"logits/rejected": -3.3024539947509766,
"logps/chosen": -3.373434066772461,
"logps/rejected": -29.926616668701172,
"loss": 0.1308,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4663188457489014,
"rewards/margins": 2.4663188457489014,
"rewards/rejected": 0.0,
"step": 1701
},
{
"epoch": 9.508379888268156,
"grad_norm": 0.6523008592710757,
"learning_rate": 4.2319788065050443e-07,
"logits/chosen": -3.4485371112823486,
"logits/rejected": -3.6596977710723877,
"logps/chosen": -15.020283699035645,
"logps/rejected": -46.4030876159668,
"loss": 0.1032,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8614203929901123,
"rewards/margins": 2.8614203929901123,
"rewards/rejected": 0.0,
"step": 1702
},
{
"epoch": 9.513966480446927,
"grad_norm": 0.9840730198114828,
"learning_rate": 4.2259562711298393e-07,
"logits/chosen": -3.4605963230133057,
"logits/rejected": -3.5639140605926514,
"logps/chosen": -1.4424254894256592,
"logps/rejected": -37.53175354003906,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5046846866607666,
"rewards/margins": 2.5046846866607666,
"rewards/rejected": 0.0,
"step": 1703
},
{
"epoch": 9.519553072625698,
"grad_norm": 3.3210097153062015,
"learning_rate": 4.2199348861214646e-07,
"logits/chosen": -3.572042226791382,
"logits/rejected": -3.51839017868042,
"logps/chosen": -2.0930020809173584,
"logps/rejected": -56.61003875732422,
"loss": 0.1057,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.452495574951172,
"rewards/margins": 2.452495574951172,
"rewards/rejected": 0.0,
"step": 1704
},
{
"epoch": 9.525139664804469,
"grad_norm": 2.8093863564024937,
"learning_rate": 4.2139146604287714e-07,
"logits/chosen": -3.474696159362793,
"logits/rejected": -3.5628952980041504,
"logps/chosen": -3.7944483757019043,
"logps/rejected": -56.560157775878906,
"loss": 0.1393,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9909389019012451,
"rewards/margins": 1.9909389019012451,
"rewards/rejected": 0.0,
"step": 1705
},
{
"epoch": 9.53072625698324,
"grad_norm": 1.2376177190857576,
"learning_rate": 4.2078956029988883e-07,
"logits/chosen": -3.311756134033203,
"logits/rejected": -3.4236950874328613,
"logps/chosen": -5.494135856628418,
"logps/rejected": -78.06456756591797,
"loss": 0.129,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.181332588195801,
"rewards/margins": 2.181332588195801,
"rewards/rejected": 0.0,
"step": 1706
},
{
"epoch": 9.53631284916201,
"grad_norm": 0.8740634913580586,
"learning_rate": 4.2018777227772026e-07,
"logits/chosen": -3.3100528717041016,
"logits/rejected": -3.2620058059692383,
"logps/chosen": -15.985023498535156,
"logps/rejected": -26.272205352783203,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.03938102722168,
"rewards/margins": 4.03938102722168,
"rewards/rejected": 0.0,
"step": 1707
},
{
"epoch": 9.541899441340782,
"grad_norm": 0.7045059680653859,
"learning_rate": 4.1958610287073585e-07,
"logits/chosen": -3.4263386726379395,
"logits/rejected": -3.48311185836792,
"logps/chosen": -2.8320908546447754,
"logps/rejected": -47.739253997802734,
"loss": 0.0883,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8688912391662598,
"rewards/margins": 2.8688912391662598,
"rewards/rejected": 0.0,
"step": 1708
},
{
"epoch": 9.547486033519553,
"grad_norm": 2.939591053489811,
"learning_rate": 4.189845529731238e-07,
"logits/chosen": -3.4358630180358887,
"logits/rejected": -3.4611194133758545,
"logps/chosen": -7.619137763977051,
"logps/rejected": -36.52315902709961,
"loss": 0.0997,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6181259155273438,
"rewards/margins": 2.6181259155273438,
"rewards/rejected": 0.0,
"step": 1709
},
{
"epoch": 9.553072625698324,
"grad_norm": 1.7312824374240363,
"learning_rate": 4.183831234788938e-07,
"logits/chosen": -3.3011412620544434,
"logits/rejected": -3.387990951538086,
"logps/chosen": -1.489870309829712,
"logps/rejected": -75.22355651855469,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.148293972015381,
"rewards/margins": 2.148293972015381,
"rewards/rejected": 0.0,
"step": 1710
},
{
"epoch": 9.558659217877095,
"grad_norm": 6.792215059319358,
"learning_rate": 4.177818152818775e-07,
"logits/chosen": -3.49019193649292,
"logits/rejected": -3.5170786380767822,
"logps/chosen": -16.729494094848633,
"logps/rejected": -36.510337829589844,
"loss": 0.1207,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3461015224456787,
"rewards/margins": 3.3461015224456787,
"rewards/rejected": 0.0,
"step": 1711
},
{
"epoch": 9.564245810055866,
"grad_norm": 2.597248724827953,
"learning_rate": 4.171806292757262e-07,
"logits/chosen": -3.3735554218292236,
"logits/rejected": -3.332463264465332,
"logps/chosen": -1.008131742477417,
"logps/rejected": -58.622379302978516,
"loss": 0.0825,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7787175178527832,
"rewards/margins": 1.7787175178527832,
"rewards/rejected": 0.0,
"step": 1712
},
{
"epoch": 9.569832402234637,
"grad_norm": 0.6999497925872811,
"learning_rate": 4.165795663539089e-07,
"logits/chosen": -3.59106707572937,
"logits/rejected": -3.7462661266326904,
"logps/chosen": -2.865769386291504,
"logps/rejected": -43.900978088378906,
"loss": 0.0867,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.748168468475342,
"rewards/margins": 2.748168468475342,
"rewards/rejected": 0.0,
"step": 1713
},
{
"epoch": 9.575418994413408,
"grad_norm": 1.299737306466016,
"learning_rate": 4.159786274097125e-07,
"logits/chosen": -3.423650026321411,
"logits/rejected": -3.5724918842315674,
"logps/chosen": -2.597463369369507,
"logps/rejected": -55.465576171875,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.100991725921631,
"rewards/margins": 2.100991725921631,
"rewards/rejected": 0.0,
"step": 1714
},
{
"epoch": 9.581005586592179,
"grad_norm": 1.0454743027026128,
"learning_rate": 4.1537781333623906e-07,
"logits/chosen": -3.250215530395508,
"logits/rejected": -3.497128486633301,
"logps/chosen": -0.8876857757568359,
"logps/rejected": -72.73077392578125,
"loss": 0.1251,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6695135831832886,
"rewards/margins": 1.6695135831832886,
"rewards/rejected": 0.0,
"step": 1715
},
{
"epoch": 9.58659217877095,
"grad_norm": 1.0617451284723087,
"learning_rate": 4.147771250264054e-07,
"logits/chosen": -3.3230700492858887,
"logits/rejected": -3.5200843811035156,
"logps/chosen": -1.789273977279663,
"logps/rejected": -30.924949645996094,
"loss": 0.0834,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6298274993896484,
"rewards/margins": 2.6298274993896484,
"rewards/rejected": 0.0,
"step": 1716
},
{
"epoch": 9.59217877094972,
"grad_norm": 2.0966938880006265,
"learning_rate": 4.1417656337294124e-07,
"logits/chosen": -3.3895654678344727,
"logits/rejected": -3.434917449951172,
"logps/chosen": -5.88671350479126,
"logps/rejected": -39.347511291503906,
"loss": 0.0794,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.851647138595581,
"rewards/margins": 3.851647138595581,
"rewards/rejected": 0.0,
"step": 1717
},
{
"epoch": 9.597765363128492,
"grad_norm": 1.0592453797314934,
"learning_rate": 4.1357612926838855e-07,
"logits/chosen": -3.12894344329834,
"logits/rejected": -3.056067943572998,
"logps/chosen": -0.6702836155891418,
"logps/rejected": -49.90453338623047,
"loss": 0.1124,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6518197059631348,
"rewards/margins": 1.6518197059631348,
"rewards/rejected": 0.0,
"step": 1718
},
{
"epoch": 9.603351955307263,
"grad_norm": 0.9466857657811402,
"learning_rate": 4.1297582360509866e-07,
"logits/chosen": -3.5946905612945557,
"logits/rejected": -3.457310199737549,
"logps/chosen": -1.7465816736221313,
"logps/rejected": -28.91625213623047,
"loss": 0.1254,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1947221755981445,
"rewards/margins": 2.1947221755981445,
"rewards/rejected": 0.0,
"step": 1719
},
{
"epoch": 9.608938547486034,
"grad_norm": 3.35693357476571,
"learning_rate": 4.123756472752331e-07,
"logits/chosen": -3.44520902633667,
"logits/rejected": -3.411966323852539,
"logps/chosen": -0.4839590787887573,
"logps/rejected": -38.20967102050781,
"loss": 0.1178,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7154146432876587,
"rewards/margins": 1.7154146432876587,
"rewards/rejected": 0.0,
"step": 1720
},
{
"epoch": 9.608938547486034,
"eval_logits/chosen": -3.305182695388794,
"eval_logits/rejected": -3.4331936836242676,
"eval_logps/chosen": -27.447872161865234,
"eval_logps/rejected": -50.44144821166992,
"eval_loss": 0.8560366630554199,
"eval_rewards/accuracies": 0.699999988079071,
"eval_rewards/chosen": 0.26221686601638794,
"eval_rewards/margins": 0.26221686601638794,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6959,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 1720
},
{
"epoch": 9.614525139664805,
"grad_norm": 0.728579747223021,
"learning_rate": 4.117756011707609e-07,
"logits/chosen": -3.18867826461792,
"logits/rejected": -3.124232292175293,
"logps/chosen": -1.6616671085357666,
"logps/rejected": -35.0905647277832,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0302085876464844,
"rewards/margins": 2.0302085876464844,
"rewards/rejected": 0.0,
"step": 1721
},
{
"epoch": 9.620111731843576,
"grad_norm": 1.4853680298014906,
"learning_rate": 4.111756861834571e-07,
"logits/chosen": -3.761784553527832,
"logits/rejected": -3.7591989040374756,
"logps/chosen": -2.763706684112549,
"logps/rejected": -29.031394958496094,
"loss": 0.1343,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1953248977661133,
"rewards/margins": 2.1953248977661133,
"rewards/rejected": 0.0,
"step": 1722
},
{
"epoch": 9.625698324022347,
"grad_norm": 2.1376789566184704,
"learning_rate": 4.105759032049025e-07,
"logits/chosen": -3.2697160243988037,
"logits/rejected": -3.1297335624694824,
"logps/chosen": -6.834595680236816,
"logps/rejected": -24.282554626464844,
"loss": 0.1242,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6610021591186523,
"rewards/margins": 2.6610021591186523,
"rewards/rejected": 0.0,
"step": 1723
},
{
"epoch": 9.631284916201118,
"grad_norm": 1.8294215741490434,
"learning_rate": 4.0997625312648127e-07,
"logits/chosen": -3.337998867034912,
"logits/rejected": -3.355957269668579,
"logps/chosen": -2.4976565837860107,
"logps/rejected": -99.48643493652344,
"loss": 0.1084,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2600810527801514,
"rewards/margins": 2.2600810527801514,
"rewards/rejected": 0.0,
"step": 1724
},
{
"epoch": 9.636871508379889,
"grad_norm": 0.8759311338583263,
"learning_rate": 4.093767368393802e-07,
"logits/chosen": -2.968350887298584,
"logits/rejected": -2.944314479827881,
"logps/chosen": -9.29737377166748,
"logps/rejected": -84.48825073242188,
"loss": 0.0756,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7920689582824707,
"rewards/margins": 2.7920689582824707,
"rewards/rejected": 0.0,
"step": 1725
},
{
"epoch": 9.64245810055866,
"grad_norm": 1.4459508312142024,
"learning_rate": 4.087773552345873e-07,
"logits/chosen": -3.326263666152954,
"logits/rejected": -3.622400999069214,
"logps/chosen": -13.418744087219238,
"logps/rejected": -74.73787689208984,
"loss": 0.0911,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2520549297332764,
"rewards/margins": 3.2520549297332764,
"rewards/rejected": 0.0,
"step": 1726
},
{
"epoch": 9.64804469273743,
"grad_norm": 1.0096702618259616,
"learning_rate": 4.081781092028905e-07,
"logits/chosen": -3.570830821990967,
"logits/rejected": -3.3866488933563232,
"logps/chosen": -8.488333702087402,
"logps/rejected": -47.86735534667969,
"loss": 0.1,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7169766426086426,
"rewards/margins": 2.7169766426086426,
"rewards/rejected": 0.0,
"step": 1727
},
{
"epoch": 9.653631284916202,
"grad_norm": 0.7058010028158032,
"learning_rate": 4.075789996348759e-07,
"logits/chosen": -3.570552349090576,
"logits/rejected": -3.7115981578826904,
"logps/chosen": -6.230530738830566,
"logps/rejected": -28.94961929321289,
"loss": 0.0842,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.987460136413574,
"rewards/margins": 2.987460136413574,
"rewards/rejected": 0.0,
"step": 1728
},
{
"epoch": 9.659217877094973,
"grad_norm": 1.3403671628087432,
"learning_rate": 4.069800274209271e-07,
"logits/chosen": -3.5410234928131104,
"logits/rejected": -3.5444400310516357,
"logps/chosen": -36.93230438232422,
"logps/rejected": -43.82135772705078,
"loss": 0.0638,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.602138042449951,
"rewards/margins": 3.602138042449951,
"rewards/rejected": 0.0,
"step": 1729
},
{
"epoch": 9.664804469273744,
"grad_norm": 1.3274871794888756,
"learning_rate": 4.063811934512236e-07,
"logits/chosen": -3.160902261734009,
"logits/rejected": -3.1793432235717773,
"logps/chosen": -7.068883419036865,
"logps/rejected": -61.582122802734375,
"loss": 0.0925,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.330369472503662,
"rewards/margins": 3.330369472503662,
"rewards/rejected": 0.0,
"step": 1730
},
{
"epoch": 9.670391061452515,
"grad_norm": 0.7071744118530354,
"learning_rate": 4.0578249861573915e-07,
"logits/chosen": -3.534485340118408,
"logits/rejected": -3.7163448333740234,
"logps/chosen": -2.4362969398498535,
"logps/rejected": -60.17002868652344,
"loss": 0.0799,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.759557008743286,
"rewards/margins": 2.759557008743286,
"rewards/rejected": 0.0,
"step": 1731
},
{
"epoch": 9.675977653631286,
"grad_norm": 1.3149072273229911,
"learning_rate": 4.0518394380424104e-07,
"logits/chosen": -3.4997012615203857,
"logits/rejected": -3.3523571491241455,
"logps/chosen": -4.800792694091797,
"logps/rejected": -92.15414428710938,
"loss": 0.1212,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.196422815322876,
"rewards/margins": 2.196422815322876,
"rewards/rejected": 0.0,
"step": 1732
},
{
"epoch": 9.681564245810057,
"grad_norm": 1.1150452822720032,
"learning_rate": 4.0458552990628823e-07,
"logits/chosen": -3.3276965618133545,
"logits/rejected": -3.60642671585083,
"logps/chosen": -3.349475860595703,
"logps/rejected": -36.168418884277344,
"loss": 0.077,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.071549654006958,
"rewards/margins": 3.071549654006958,
"rewards/rejected": 0.0,
"step": 1733
},
{
"epoch": 9.687150837988828,
"grad_norm": 1.4258397485068366,
"learning_rate": 4.039872578112304e-07,
"logits/chosen": -3.601954460144043,
"logits/rejected": -3.7227015495300293,
"logps/chosen": -3.7684290409088135,
"logps/rejected": -58.21186828613281,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8843202590942383,
"rewards/margins": 2.8843202590942383,
"rewards/rejected": 0.0,
"step": 1734
},
{
"epoch": 9.692737430167599,
"grad_norm": 3.422911984874883,
"learning_rate": 4.033891284082064e-07,
"logits/chosen": -3.7799041271209717,
"logits/rejected": -3.6273603439331055,
"logps/chosen": -2.0787601470947266,
"logps/rejected": -100.98727416992188,
"loss": 0.1114,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9293601512908936,
"rewards/margins": 1.9293601512908936,
"rewards/rejected": 0.0,
"step": 1735
},
{
"epoch": 9.69832402234637,
"grad_norm": 0.7344493698509696,
"learning_rate": 4.0279114258614297e-07,
"logits/chosen": -3.370419502258301,
"logits/rejected": -3.5623855590820312,
"logps/chosen": -21.76702880859375,
"logps/rejected": -47.909759521484375,
"loss": 0.0888,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.385744094848633,
"rewards/margins": 3.385744094848633,
"rewards/rejected": 0.0,
"step": 1736
},
{
"epoch": 9.703910614525139,
"grad_norm": 3.7414667340922865,
"learning_rate": 4.0219330123375355e-07,
"logits/chosen": -3.397857904434204,
"logits/rejected": -3.313786268234253,
"logps/chosen": -0.6954784393310547,
"logps/rejected": -41.59431838989258,
"loss": 0.1144,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0533552169799805,
"rewards/margins": 2.0533552169799805,
"rewards/rejected": 0.0,
"step": 1737
},
{
"epoch": 9.70949720670391,
"grad_norm": 1.2846233750707605,
"learning_rate": 4.015956052395371e-07,
"logits/chosen": -3.5319018363952637,
"logits/rejected": -3.5256760120391846,
"logps/chosen": -8.303479194641113,
"logps/rejected": -34.281288146972656,
"loss": 0.0993,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6803646087646484,
"rewards/margins": 3.6803646087646484,
"rewards/rejected": 0.0,
"step": 1738
},
{
"epoch": 9.71508379888268,
"grad_norm": 3.287051408358936,
"learning_rate": 4.0099805549177594e-07,
"logits/chosen": -3.7380051612854004,
"logits/rejected": -3.736550807952881,
"logps/chosen": -12.186216354370117,
"logps/rejected": -41.61170196533203,
"loss": 0.1099,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.883512258529663,
"rewards/margins": 3.883512258529663,
"rewards/rejected": 0.0,
"step": 1739
},
{
"epoch": 9.720670391061452,
"grad_norm": 2.496423080365672,
"learning_rate": 4.0040065287853563e-07,
"logits/chosen": -3.4628448486328125,
"logits/rejected": -3.5115394592285156,
"logps/chosen": -0.5052372217178345,
"logps/rejected": -94.57827758789062,
"loss": 0.138,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9333962202072144,
"rewards/margins": 1.9333962202072144,
"rewards/rejected": 0.0,
"step": 1740
},
{
"epoch": 9.720670391061452,
"eval_logits/chosen": -3.311786651611328,
"eval_logits/rejected": -3.438049793243408,
"eval_logps/chosen": -27.7125301361084,
"eval_logps/rejected": -50.265037536621094,
"eval_loss": 0.8467325568199158,
"eval_rewards/accuracies": 0.625,
"eval_rewards/chosen": 0.23575131595134735,
"eval_rewards/margins": 0.23575131595134735,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7079,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 1740
},
{
"epoch": 9.726256983240223,
"grad_norm": 0.7484806715848626,
"learning_rate": 3.9980339828766304e-07,
"logits/chosen": -3.3347198963165283,
"logits/rejected": -3.5363245010375977,
"logps/chosen": -0.5856008529663086,
"logps/rejected": -56.053314208984375,
"loss": 0.0893,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6698448657989502,
"rewards/margins": 1.6698448657989502,
"rewards/rejected": 0.0,
"step": 1741
},
{
"epoch": 9.731843575418994,
"grad_norm": 0.6762585368094688,
"learning_rate": 3.9920629260678437e-07,
"logits/chosen": -3.442697525024414,
"logits/rejected": -3.5688514709472656,
"logps/chosen": -4.242497444152832,
"logps/rejected": -44.710716247558594,
"loss": 0.0903,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1763532161712646,
"rewards/margins": 3.1763532161712646,
"rewards/rejected": 0.0,
"step": 1742
},
{
"epoch": 9.737430167597765,
"grad_norm": 0.7985715083369194,
"learning_rate": 3.986093367233054e-07,
"logits/chosen": -3.028599977493286,
"logits/rejected": -3.074604034423828,
"logps/chosen": -9.273598670959473,
"logps/rejected": -37.633121490478516,
"loss": 0.0905,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.960770845413208,
"rewards/margins": 2.960770845413208,
"rewards/rejected": 0.0,
"step": 1743
},
{
"epoch": 9.743016759776536,
"grad_norm": 2.082457001958603,
"learning_rate": 3.980125315244087e-07,
"logits/chosen": -3.359621047973633,
"logits/rejected": -3.2890217304229736,
"logps/chosen": -9.379022598266602,
"logps/rejected": -40.036903381347656,
"loss": 0.1105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7783303260803223,
"rewards/margins": 2.7783303260803223,
"rewards/rejected": 0.0,
"step": 1744
},
{
"epoch": 9.748603351955307,
"grad_norm": 0.866636462205181,
"learning_rate": 3.9741587789705315e-07,
"logits/chosen": -3.580343723297119,
"logits/rejected": -3.5654988288879395,
"logps/chosen": -18.33171272277832,
"logps/rejected": -43.90819549560547,
"loss": 0.1017,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1261768341064453,
"rewards/margins": 3.1261768341064453,
"rewards/rejected": 0.0,
"step": 1745
},
{
"epoch": 9.754189944134078,
"grad_norm": 1.7091966122334377,
"learning_rate": 3.968193767279722e-07,
"logits/chosen": -3.5168099403381348,
"logits/rejected": -3.483940601348877,
"logps/chosen": -0.9022572040557861,
"logps/rejected": -52.251983642578125,
"loss": 0.095,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.030081272125244,
"rewards/margins": 2.030081272125244,
"rewards/rejected": 0.0,
"step": 1746
},
{
"epoch": 9.759776536312849,
"grad_norm": 1.1050512299361617,
"learning_rate": 3.962230289036731e-07,
"logits/chosen": -3.5157039165496826,
"logits/rejected": -3.4680283069610596,
"logps/chosen": -1.9303457736968994,
"logps/rejected": -41.15681457519531,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6821866035461426,
"rewards/margins": 2.6821866035461426,
"rewards/rejected": 0.0,
"step": 1747
},
{
"epoch": 9.76536312849162,
"grad_norm": 1.1568480157771526,
"learning_rate": 3.956268353104345e-07,
"logits/chosen": -3.50109601020813,
"logits/rejected": -3.5288844108581543,
"logps/chosen": -7.669013977050781,
"logps/rejected": -69.0399398803711,
"loss": 0.1037,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.297898292541504,
"rewards/margins": 2.297898292541504,
"rewards/rejected": 0.0,
"step": 1748
},
{
"epoch": 9.77094972067039,
"grad_norm": 0.8890581971010301,
"learning_rate": 3.950307968343065e-07,
"logits/chosen": -3.638458251953125,
"logits/rejected": -3.654754161834717,
"logps/chosen": -6.63097620010376,
"logps/rejected": -56.3765869140625,
"loss": 0.1013,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.622446060180664,
"rewards/margins": 2.622446060180664,
"rewards/rejected": 0.0,
"step": 1749
},
{
"epoch": 9.776536312849162,
"grad_norm": 3.358159262414772,
"learning_rate": 3.944349143611085e-07,
"logits/chosen": -3.5754904747009277,
"logits/rejected": -3.7089226245880127,
"logps/chosen": -5.279230117797852,
"logps/rejected": -46.64799880981445,
"loss": 0.1168,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.894991874694824,
"rewards/margins": 2.894991874694824,
"rewards/rejected": 0.0,
"step": 1750
},
{
"epoch": 9.782122905027933,
"grad_norm": 1.3118787614294551,
"learning_rate": 3.938391887764277e-07,
"logits/chosen": -3.4574172496795654,
"logits/rejected": -3.321378469467163,
"logps/chosen": -16.185279846191406,
"logps/rejected": -48.61097717285156,
"loss": 0.0913,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.327827453613281,
"rewards/margins": 4.327827453613281,
"rewards/rejected": 0.0,
"step": 1751
},
{
"epoch": 9.787709497206704,
"grad_norm": 1.2003247191612298,
"learning_rate": 3.932436209656188e-07,
"logits/chosen": -3.395296096801758,
"logits/rejected": -3.4303250312805176,
"logps/chosen": -20.829021453857422,
"logps/rejected": -52.34408187866211,
"loss": 0.0862,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.022113800048828,
"rewards/margins": 3.022113800048828,
"rewards/rejected": 0.0,
"step": 1752
},
{
"epoch": 9.793296089385475,
"grad_norm": 1.243628704239324,
"learning_rate": 3.926482118138013e-07,
"logits/chosen": -3.3079471588134766,
"logits/rejected": -3.23549222946167,
"logps/chosen": -0.5884277820587158,
"logps/rejected": -66.92959594726562,
"loss": 0.1268,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4585797786712646,
"rewards/margins": 1.4585797786712646,
"rewards/rejected": 0.0,
"step": 1753
},
{
"epoch": 9.798882681564246,
"grad_norm": 2.271405711594028,
"learning_rate": 3.920529622058594e-07,
"logits/chosen": -3.2228918075561523,
"logits/rejected": -3.3771908283233643,
"logps/chosen": -31.38953971862793,
"logps/rejected": -34.92992401123047,
"loss": 0.1034,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8909780979156494,
"rewards/margins": 3.8909780979156494,
"rewards/rejected": 0.0,
"step": 1754
},
{
"epoch": 9.804469273743017,
"grad_norm": 0.7039622569173002,
"learning_rate": 3.914578730264399e-07,
"logits/chosen": -3.1822798252105713,
"logits/rejected": -3.574373722076416,
"logps/chosen": -1.4600975513458252,
"logps/rejected": -57.503257751464844,
"loss": 0.0749,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.233499765396118,
"rewards/margins": 2.233499765396118,
"rewards/rejected": 0.0,
"step": 1755
},
{
"epoch": 9.810055865921788,
"grad_norm": 5.120097178309147,
"learning_rate": 3.9086294515995157e-07,
"logits/chosen": -3.4335861206054688,
"logits/rejected": -3.2751214504241943,
"logps/chosen": -18.258726119995117,
"logps/rejected": -35.95403289794922,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.150092601776123,
"rewards/margins": 3.150092601776123,
"rewards/rejected": 0.0,
"step": 1756
},
{
"epoch": 9.815642458100559,
"grad_norm": 0.9911180404327901,
"learning_rate": 3.9026817949056276e-07,
"logits/chosen": -3.4206182956695557,
"logits/rejected": -3.3530704975128174,
"logps/chosen": -3.826144218444824,
"logps/rejected": -99.31694030761719,
"loss": 0.1315,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.372318744659424,
"rewards/margins": 2.372318744659424,
"rewards/rejected": 0.0,
"step": 1757
},
{
"epoch": 9.82122905027933,
"grad_norm": 0.9858380401630963,
"learning_rate": 3.896735769022014e-07,
"logits/chosen": -3.5911967754364014,
"logits/rejected": -3.4111180305480957,
"logps/chosen": -6.001222610473633,
"logps/rejected": -39.18571090698242,
"loss": 0.1164,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1782052516937256,
"rewards/margins": 3.1782052516937256,
"rewards/rejected": 0.0,
"step": 1758
},
{
"epoch": 9.8268156424581,
"grad_norm": 1.20868578492993,
"learning_rate": 3.8907913827855296e-07,
"logits/chosen": -3.6774656772613525,
"logits/rejected": -3.4747474193573,
"logps/chosen": -1.6732978820800781,
"logps/rejected": -38.50730514526367,
"loss": 0.1052,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.597053050994873,
"rewards/margins": 2.597053050994873,
"rewards/rejected": 0.0,
"step": 1759
},
{
"epoch": 9.832402234636872,
"grad_norm": 2.5127368290338428,
"learning_rate": 3.8848486450305886e-07,
"logits/chosen": -3.605839490890503,
"logits/rejected": -3.558403968811035,
"logps/chosen": -5.57737922668457,
"logps/rejected": -33.1185188293457,
"loss": 0.1465,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.048252582550049,
"rewards/margins": 3.048252582550049,
"rewards/rejected": 0.0,
"step": 1760
},
{
"epoch": 9.832402234636872,
"eval_logits/chosen": -3.3077073097229004,
"eval_logits/rejected": -3.4372220039367676,
"eval_logps/chosen": -27.46579360961914,
"eval_logps/rejected": -49.97900390625,
"eval_loss": 0.8316324353218079,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.26042503118515015,
"eval_rewards/margins": 0.26042503118515015,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6853,
"eval_samples_per_second": 9.484,
"eval_steps_per_second": 0.306,
"step": 1760
},
{
"epoch": 9.837988826815643,
"grad_norm": 0.8045617035115878,
"learning_rate": 3.878907564589158e-07,
"logits/chosen": -3.4086174964904785,
"logits/rejected": -3.4591829776763916,
"logps/chosen": -0.9813145399093628,
"logps/rejected": -36.769046783447266,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1813831329345703,
"rewards/margins": 2.1813831329345703,
"rewards/rejected": 0.0,
"step": 1761
},
{
"epoch": 9.843575418994414,
"grad_norm": 1.7989369383734568,
"learning_rate": 3.872968150290745e-07,
"logits/chosen": -3.2892017364501953,
"logits/rejected": -3.3393900394439697,
"logps/chosen": -3.108140230178833,
"logps/rejected": -91.55709838867188,
"loss": 0.0809,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5772852897644043,
"rewards/margins": 2.5772852897644043,
"rewards/rejected": 0.0,
"step": 1762
},
{
"epoch": 9.849162011173185,
"grad_norm": 1.9401687622075436,
"learning_rate": 3.867030410962372e-07,
"logits/chosen": -3.1430904865264893,
"logits/rejected": -2.9523463249206543,
"logps/chosen": -27.47216796875,
"logps/rejected": -23.676525115966797,
"loss": 0.0956,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5253515243530273,
"rewards/margins": 3.5253515243530273,
"rewards/rejected": 0.0,
"step": 1763
},
{
"epoch": 9.854748603351956,
"grad_norm": 0.8372219159775339,
"learning_rate": 3.86109435542858e-07,
"logits/chosen": -3.4679312705993652,
"logits/rejected": -3.3298377990722656,
"logps/chosen": -8.737722396850586,
"logps/rejected": -47.09326934814453,
"loss": 0.0957,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.532306671142578,
"rewards/margins": 3.532306671142578,
"rewards/rejected": 0.0,
"step": 1764
},
{
"epoch": 9.860335195530727,
"grad_norm": 0.8523774604103546,
"learning_rate": 3.855159992511403e-07,
"logits/chosen": -3.185344696044922,
"logits/rejected": -3.3517708778381348,
"logps/chosen": -5.29677152633667,
"logps/rejected": -92.0924072265625,
"loss": 0.1312,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8936386108398438,
"rewards/margins": 1.8936386108398438,
"rewards/rejected": 0.0,
"step": 1765
},
{
"epoch": 9.865921787709498,
"grad_norm": 1.680458582515685,
"learning_rate": 3.8492273310303615e-07,
"logits/chosen": -3.3776400089263916,
"logits/rejected": -3.503953218460083,
"logps/chosen": -0.49245208501815796,
"logps/rejected": -85.63204956054688,
"loss": 0.1052,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7712440490722656,
"rewards/margins": 1.7712440490722656,
"rewards/rejected": 0.0,
"step": 1766
},
{
"epoch": 9.871508379888269,
"grad_norm": 2.7525121422880634,
"learning_rate": 3.8432963798024493e-07,
"logits/chosen": -3.604893445968628,
"logits/rejected": -3.671189308166504,
"logps/chosen": -9.61042594909668,
"logps/rejected": -50.47281265258789,
"loss": 0.1172,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.054548978805542,
"rewards/margins": 3.054548978805542,
"rewards/rejected": 0.0,
"step": 1767
},
{
"epoch": 9.87709497206704,
"grad_norm": 1.263736681740419,
"learning_rate": 3.837367147642112e-07,
"logits/chosen": -3.3494186401367188,
"logits/rejected": -3.480822801589966,
"logps/chosen": -1.6545333862304688,
"logps/rejected": -57.14366149902344,
"loss": 0.1269,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.903753399848938,
"rewards/margins": 1.903753399848938,
"rewards/rejected": 0.0,
"step": 1768
},
{
"epoch": 9.88268156424581,
"grad_norm": 1.9278952942148668,
"learning_rate": 3.8314396433612476e-07,
"logits/chosen": -3.361748456954956,
"logits/rejected": -3.4422905445098877,
"logps/chosen": -11.146872520446777,
"logps/rejected": -44.38884353637695,
"loss": 0.1103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3656280040740967,
"rewards/margins": 2.3656280040740967,
"rewards/rejected": 0.0,
"step": 1769
},
{
"epoch": 9.888268156424582,
"grad_norm": 0.9324815106028379,
"learning_rate": 3.825513875769184e-07,
"logits/chosen": -3.4899258613586426,
"logits/rejected": -3.5031919479370117,
"logps/chosen": -3.4244465827941895,
"logps/rejected": -52.209190368652344,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.807156801223755,
"rewards/margins": 2.807156801223755,
"rewards/rejected": 0.0,
"step": 1770
},
{
"epoch": 9.893854748603353,
"grad_norm": 0.7663472924188264,
"learning_rate": 3.819589853672664e-07,
"logits/chosen": -3.441516160964966,
"logits/rejected": -3.580808162689209,
"logps/chosen": -2.0890583992004395,
"logps/rejected": -41.098140716552734,
"loss": 0.0653,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.743102550506592,
"rewards/margins": 2.743102550506592,
"rewards/rejected": 0.0,
"step": 1771
},
{
"epoch": 9.899441340782122,
"grad_norm": 2.1268683325530326,
"learning_rate": 3.8136675858758413e-07,
"logits/chosen": -3.400416851043701,
"logits/rejected": -3.420621156692505,
"logps/chosen": -7.7488837242126465,
"logps/rejected": -56.89590072631836,
"loss": 0.1307,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7369866371154785,
"rewards/margins": 2.7369866371154785,
"rewards/rejected": 0.0,
"step": 1772
},
{
"epoch": 9.905027932960895,
"grad_norm": 0.6675627638984232,
"learning_rate": 3.8077470811802607e-07,
"logits/chosen": -3.4226248264312744,
"logits/rejected": -3.3723502159118652,
"logps/chosen": -6.033211708068848,
"logps/rejected": -28.20562744140625,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.184857130050659,
"rewards/margins": 3.184857130050659,
"rewards/rejected": 0.0,
"step": 1773
},
{
"epoch": 9.910614525139664,
"grad_norm": 1.530320240689921,
"learning_rate": 3.801828348384846e-07,
"logits/chosen": -3.1324992179870605,
"logits/rejected": -3.481651782989502,
"logps/chosen": -1.784105896949768,
"logps/rejected": -68.61576843261719,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.685253143310547,
"rewards/margins": 2.685253143310547,
"rewards/rejected": 0.0,
"step": 1774
},
{
"epoch": 9.916201117318435,
"grad_norm": 1.722119950356398,
"learning_rate": 3.7959113962858874e-07,
"logits/chosen": -3.3707611560821533,
"logits/rejected": -3.356269359588623,
"logps/chosen": -1.1857571601867676,
"logps/rejected": -35.922237396240234,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.640213966369629,
"rewards/margins": 2.640213966369629,
"rewards/rejected": 0.0,
"step": 1775
},
{
"epoch": 9.921787709497206,
"grad_norm": 0.8925646178900609,
"learning_rate": 3.789996233677032e-07,
"logits/chosen": -3.47371506690979,
"logits/rejected": -3.368332862854004,
"logps/chosen": -6.826603412628174,
"logps/rejected": -26.780426025390625,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.414897918701172,
"rewards/margins": 2.414897918701172,
"rewards/rejected": 0.0,
"step": 1776
},
{
"epoch": 9.927374301675977,
"grad_norm": 0.6691219573900442,
"learning_rate": 3.7840828693492616e-07,
"logits/chosen": -3.4892804622650146,
"logits/rejected": -3.559783458709717,
"logps/chosen": -12.715020179748535,
"logps/rejected": -58.34233093261719,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7501840591430664,
"rewards/margins": 3.7501840591430664,
"rewards/rejected": 0.0,
"step": 1777
},
{
"epoch": 9.932960893854748,
"grad_norm": 2.2085090168636925,
"learning_rate": 3.77817131209089e-07,
"logits/chosen": -3.3483548164367676,
"logits/rejected": -3.464453935623169,
"logps/chosen": -3.2515854835510254,
"logps/rejected": -60.47914123535156,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.664233684539795,
"rewards/margins": 2.664233684539795,
"rewards/rejected": 0.0,
"step": 1778
},
{
"epoch": 9.938547486033519,
"grad_norm": 0.9067412657774024,
"learning_rate": 3.772261570687547e-07,
"logits/chosen": -3.5294055938720703,
"logits/rejected": -3.6384940147399902,
"logps/chosen": -0.8120299577713013,
"logps/rejected": -51.77969741821289,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.22739315032959,
"rewards/margins": 2.22739315032959,
"rewards/rejected": 0.0,
"step": 1779
},
{
"epoch": 9.94413407821229,
"grad_norm": 0.792412217652063,
"learning_rate": 3.7663536539221557e-07,
"logits/chosen": -2.7787868976593018,
"logits/rejected": -2.9306488037109375,
"logps/chosen": -43.92116165161133,
"logps/rejected": -47.89521026611328,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.404733657836914,
"rewards/margins": 3.404733657836914,
"rewards/rejected": 0.0,
"step": 1780
},
{
"epoch": 9.94413407821229,
"eval_logits/chosen": -3.3067257404327393,
"eval_logits/rejected": -3.434464693069458,
"eval_logps/chosen": -27.879383087158203,
"eval_logps/rejected": -50.4825439453125,
"eval_loss": 0.8560438752174377,
"eval_rewards/accuracies": 0.625,
"eval_rewards/chosen": 0.2190660983324051,
"eval_rewards/margins": 0.2190660983324051,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7005,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 1780
},
{
"epoch": 9.949720670391061,
"grad_norm": 4.858421110396251,
"learning_rate": 3.760447570574935e-07,
"logits/chosen": -3.4732677936553955,
"logits/rejected": -3.3451426029205322,
"logps/chosen": -3.260953664779663,
"logps/rejected": -59.285213470458984,
"loss": 0.0985,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7885258197784424,
"rewards/margins": 1.7885258197784424,
"rewards/rejected": 0.0,
"step": 1781
},
{
"epoch": 9.955307262569832,
"grad_norm": 1.2430823107046305,
"learning_rate": 3.754543329423379e-07,
"logits/chosen": -3.2819435596466064,
"logits/rejected": -3.44335675239563,
"logps/chosen": -2.0860066413879395,
"logps/rejected": -70.48025512695312,
"loss": 0.1349,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7055273056030273,
"rewards/margins": 2.7055273056030273,
"rewards/rejected": 0.0,
"step": 1782
},
{
"epoch": 9.960893854748603,
"grad_norm": 1.5458676899178516,
"learning_rate": 3.748640939242238e-07,
"logits/chosen": -3.4624369144439697,
"logits/rejected": -3.2279250621795654,
"logps/chosen": -1.4880242347717285,
"logps/rejected": -44.38206481933594,
"loss": 0.0902,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.151301145553589,
"rewards/margins": 2.151301145553589,
"rewards/rejected": 0.0,
"step": 1783
},
{
"epoch": 9.966480446927374,
"grad_norm": 2.8207054308446797,
"learning_rate": 3.7427404088035154e-07,
"logits/chosen": -3.529907703399658,
"logits/rejected": -3.399103879928589,
"logps/chosen": -7.378881454467773,
"logps/rejected": -29.37787437438965,
"loss": 0.1076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.654884099960327,
"rewards/margins": 3.654884099960327,
"rewards/rejected": 0.0,
"step": 1784
},
{
"epoch": 9.972067039106145,
"grad_norm": 2.365062685396757,
"learning_rate": 3.736841746876452e-07,
"logits/chosen": -3.386728286743164,
"logits/rejected": -3.290909767150879,
"logps/chosen": -1.703054428100586,
"logps/rejected": -28.475000381469727,
"loss": 0.1203,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1432785987854004,
"rewards/margins": 2.1432785987854004,
"rewards/rejected": 0.0,
"step": 1785
},
{
"epoch": 9.977653631284916,
"grad_norm": 1.6920144311296252,
"learning_rate": 3.730944962227508e-07,
"logits/chosen": -3.4041218757629395,
"logits/rejected": -3.3558971881866455,
"logps/chosen": -0.975854754447937,
"logps/rejected": -37.28769302368164,
"loss": 0.107,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.864650011062622,
"rewards/margins": 1.864650011062622,
"rewards/rejected": 0.0,
"step": 1786
},
{
"epoch": 9.983240223463687,
"grad_norm": 0.7704995271772535,
"learning_rate": 3.7250500636203563e-07,
"logits/chosen": -3.4895503520965576,
"logits/rejected": -3.2797303199768066,
"logps/chosen": -13.055721282958984,
"logps/rejected": -37.99407196044922,
"loss": 0.0876,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.027369499206543,
"rewards/margins": 4.027369499206543,
"rewards/rejected": 0.0,
"step": 1787
},
{
"epoch": 9.988826815642458,
"grad_norm": 1.0171152776714127,
"learning_rate": 3.719157059815868e-07,
"logits/chosen": -3.2032382488250732,
"logits/rejected": -3.1955862045288086,
"logps/chosen": -3.1329994201660156,
"logps/rejected": -41.41448211669922,
"loss": 0.1139,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.500286817550659,
"rewards/margins": 2.500286817550659,
"rewards/rejected": 0.0,
"step": 1788
},
{
"epoch": 9.994413407821229,
"grad_norm": 1.1703063868926848,
"learning_rate": 3.713265959572093e-07,
"logits/chosen": -3.1190359592437744,
"logits/rejected": -3.178412437438965,
"logps/chosen": -3.4634060859680176,
"logps/rejected": -37.48868179321289,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.124701499938965,
"rewards/margins": 2.124701499938965,
"rewards/rejected": 0.0,
"step": 1789
},
{
"epoch": 10.0,
"grad_norm": 2.2025068365730167,
"learning_rate": 3.707376771644256e-07,
"logits/chosen": -3.533526659011841,
"logits/rejected": -3.4222147464752197,
"logps/chosen": -2.5332818031311035,
"logps/rejected": -44.49677276611328,
"loss": 0.1077,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1045467853546143,
"rewards/margins": 3.1045467853546143,
"rewards/rejected": 0.0,
"step": 1790
},
{
"epoch": 10.005586592178771,
"grad_norm": 0.6971282634946807,
"learning_rate": 3.7014895047847427e-07,
"logits/chosen": -3.6349709033966064,
"logits/rejected": -3.6591176986694336,
"logps/chosen": -20.891502380371094,
"logps/rejected": -64.14535522460938,
"loss": 0.1042,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.370220184326172,
"rewards/margins": 3.370220184326172,
"rewards/rejected": 0.0,
"step": 1791
},
{
"epoch": 10.011173184357542,
"grad_norm": 2.7256861857969237,
"learning_rate": 3.695604167743076e-07,
"logits/chosen": -3.316932439804077,
"logits/rejected": -3.2103734016418457,
"logps/chosen": -5.798719882965088,
"logps/rejected": -46.49449157714844,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.738903522491455,
"rewards/margins": 2.738903522491455,
"rewards/rejected": 0.0,
"step": 1792
},
{
"epoch": 10.016759776536313,
"grad_norm": 0.4488725480341928,
"learning_rate": 3.689720769265916e-07,
"logits/chosen": -3.404007911682129,
"logits/rejected": -3.4754297733306885,
"logps/chosen": -3.416945695877075,
"logps/rejected": -38.77027893066406,
"loss": 0.0761,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.912581443786621,
"rewards/margins": 2.912581443786621,
"rewards/rejected": 0.0,
"step": 1793
},
{
"epoch": 10.022346368715084,
"grad_norm": 1.348873498036711,
"learning_rate": 3.6838393180970414e-07,
"logits/chosen": -3.72420072555542,
"logits/rejected": -3.533327341079712,
"logps/chosen": -2.403259754180908,
"logps/rejected": -54.006866455078125,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9669742584228516,
"rewards/margins": 2.9669742584228516,
"rewards/rejected": 0.0,
"step": 1794
},
{
"epoch": 10.027932960893855,
"grad_norm": 0.6901470402549703,
"learning_rate": 3.677959822977334e-07,
"logits/chosen": -3.4903969764709473,
"logits/rejected": -3.695781707763672,
"logps/chosen": -16.70089340209961,
"logps/rejected": -43.73395538330078,
"loss": 0.0944,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4970295429229736,
"rewards/margins": 3.4970295429229736,
"rewards/rejected": 0.0,
"step": 1795
},
{
"epoch": 10.033519553072626,
"grad_norm": 0.874014065062377,
"learning_rate": 3.6720822926447746e-07,
"logits/chosen": -3.149998664855957,
"logits/rejected": -3.092202663421631,
"logps/chosen": -5.896096229553223,
"logps/rejected": -41.47645568847656,
"loss": 0.1202,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8974430561065674,
"rewards/margins": 2.8974430561065674,
"rewards/rejected": 0.0,
"step": 1796
},
{
"epoch": 10.039106145251397,
"grad_norm": 0.6459482810611393,
"learning_rate": 3.666206735834415e-07,
"logits/chosen": -3.2483434677124023,
"logits/rejected": -3.4119620323181152,
"logps/chosen": -0.2977844476699829,
"logps/rejected": -72.59209442138672,
"loss": 0.0998,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.855495810508728,
"rewards/margins": 1.855495810508728,
"rewards/rejected": 0.0,
"step": 1797
},
{
"epoch": 10.044692737430168,
"grad_norm": 4.657927756822812,
"learning_rate": 3.6603331612783814e-07,
"logits/chosen": -3.57955002784729,
"logits/rejected": -3.7270655632019043,
"logps/chosen": -5.491798400878906,
"logps/rejected": -32.26445770263672,
"loss": 0.1423,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0879874229431152,
"rewards/margins": 3.0879874229431152,
"rewards/rejected": 0.0,
"step": 1798
},
{
"epoch": 10.050279329608939,
"grad_norm": 0.5170493262407111,
"learning_rate": 3.654461577705855e-07,
"logits/chosen": -3.653297185897827,
"logits/rejected": -3.5519185066223145,
"logps/chosen": -0.7011286616325378,
"logps/rejected": -48.50175857543945,
"loss": 0.0914,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2897849082946777,
"rewards/margins": 2.2897849082946777,
"rewards/rejected": 0.0,
"step": 1799
},
{
"epoch": 10.05586592178771,
"grad_norm": 0.5477330112233832,
"learning_rate": 3.6485919938430484e-07,
"logits/chosen": -3.3516428470611572,
"logits/rejected": -3.263317823410034,
"logps/chosen": -0.7090689539909363,
"logps/rejected": -63.48958206176758,
"loss": 0.1189,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2917845249176025,
"rewards/margins": 2.2917845249176025,
"rewards/rejected": 0.0,
"step": 1800
},
{
"epoch": 10.05586592178771,
"eval_logits/chosen": -3.306220293045044,
"eval_logits/rejected": -3.4329609870910645,
"eval_logps/chosen": -27.864831924438477,
"eval_logps/rejected": -50.33716583251953,
"eval_loss": 0.8530665040016174,
"eval_rewards/accuracies": 0.625,
"eval_rewards/chosen": 0.22052082419395447,
"eval_rewards/margins": 0.22052082419395447,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7075,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 1800
},
{
"epoch": 10.061452513966481,
"grad_norm": 0.597354188490836,
"learning_rate": 3.642724418413211e-07,
"logits/chosen": -3.14695405960083,
"logits/rejected": -3.36464786529541,
"logps/chosen": -7.296316623687744,
"logps/rejected": -83.26950073242188,
"loss": 0.0847,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0922064781188965,
"rewards/margins": 3.0922064781188965,
"rewards/rejected": 0.0,
"step": 1801
},
{
"epoch": 10.067039106145252,
"grad_norm": 3.7260784031422487,
"learning_rate": 3.6368588601366084e-07,
"logits/chosen": -3.3331053256988525,
"logits/rejected": -3.6264700889587402,
"logps/chosen": -0.9284124970436096,
"logps/rejected": -79.09478759765625,
"loss": 0.139,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1486902236938477,
"rewards/margins": 2.1486902236938477,
"rewards/rejected": 0.0,
"step": 1802
},
{
"epoch": 10.072625698324023,
"grad_norm": 0.5138674110258655,
"learning_rate": 3.6309953277305003e-07,
"logits/chosen": -3.390658140182495,
"logits/rejected": -3.3350045680999756,
"logps/chosen": -0.8583160638809204,
"logps/rejected": -34.64715576171875,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.389380931854248,
"rewards/margins": 2.389380931854248,
"rewards/rejected": 0.0,
"step": 1803
},
{
"epoch": 10.078212290502794,
"grad_norm": 1.5476097008815286,
"learning_rate": 3.625133829909141e-07,
"logits/chosen": -3.593475580215454,
"logits/rejected": -3.447777271270752,
"logps/chosen": -3.3375725746154785,
"logps/rejected": -33.17520523071289,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4503748416900635,
"rewards/margins": 2.4503748416900635,
"rewards/rejected": 0.0,
"step": 1804
},
{
"epoch": 10.083798882681565,
"grad_norm": 0.5830468945337419,
"learning_rate": 3.619274375383763e-07,
"logits/chosen": -3.523662567138672,
"logits/rejected": -3.5923807621002197,
"logps/chosen": -19.54915428161621,
"logps/rejected": -27.53266143798828,
"loss": 0.0825,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7670845985412598,
"rewards/margins": 3.7670845985412598,
"rewards/rejected": 0.0,
"step": 1805
},
{
"epoch": 10.089385474860336,
"grad_norm": 0.9087389475341742,
"learning_rate": 3.613416972862554e-07,
"logits/chosen": -3.2287206649780273,
"logits/rejected": -3.3611888885498047,
"logps/chosen": -3.0734684467315674,
"logps/rejected": -77.2863998413086,
"loss": 0.0955,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.500432014465332,
"rewards/margins": 2.500432014465332,
"rewards/rejected": 0.0,
"step": 1806
},
{
"epoch": 10.094972067039107,
"grad_norm": 0.584162716982449,
"learning_rate": 3.607561631050661e-07,
"logits/chosen": -3.4512503147125244,
"logits/rejected": -3.354774236679077,
"logps/chosen": -6.594293594360352,
"logps/rejected": -22.98871612548828,
"loss": 0.0968,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.304326057434082,
"rewards/margins": 3.304326057434082,
"rewards/rejected": 0.0,
"step": 1807
},
{
"epoch": 10.100558659217878,
"grad_norm": 1.9744675269338456,
"learning_rate": 3.601708358650165e-07,
"logits/chosen": -3.525938034057617,
"logits/rejected": -3.686091423034668,
"logps/chosen": -11.050585746765137,
"logps/rejected": -44.24974822998047,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7290430068969727,
"rewards/margins": 2.7290430068969727,
"rewards/rejected": 0.0,
"step": 1808
},
{
"epoch": 10.106145251396647,
"grad_norm": 1.1717321376977452,
"learning_rate": 3.5958571643600696e-07,
"logits/chosen": -3.4633126258850098,
"logits/rejected": -3.702258586883545,
"logps/chosen": -2.6135754585266113,
"logps/rejected": -45.58203887939453,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6394338607788086,
"rewards/margins": 2.6394338607788086,
"rewards/rejected": 0.0,
"step": 1809
},
{
"epoch": 10.111731843575418,
"grad_norm": 0.8484919348861357,
"learning_rate": 3.5900080568762916e-07,
"logits/chosen": -3.5047709941864014,
"logits/rejected": -3.480854034423828,
"logps/chosen": -47.91346740722656,
"logps/rejected": -30.867521286010742,
"loss": 0.0729,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.134696006774902,
"rewards/margins": 4.134696006774902,
"rewards/rejected": 0.0,
"step": 1810
},
{
"epoch": 10.11731843575419,
"grad_norm": 3.902059987595593,
"learning_rate": 3.584161044891648e-07,
"logits/chosen": -3.7783257961273193,
"logits/rejected": -3.6784465312957764,
"logps/chosen": -3.2678468227386475,
"logps/rejected": -38.271728515625,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1960103511810303,
"rewards/margins": 3.1960103511810303,
"rewards/rejected": 0.0,
"step": 1811
},
{
"epoch": 10.12290502793296,
"grad_norm": 3.712507208825318,
"learning_rate": 3.5783161370958374e-07,
"logits/chosen": -3.2993576526641846,
"logits/rejected": -3.4634902477264404,
"logps/chosen": -0.25519460439682007,
"logps/rejected": -98.33456420898438,
"loss": 0.1408,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7136197090148926,
"rewards/margins": 1.7136197090148926,
"rewards/rejected": 0.0,
"step": 1812
},
{
"epoch": 10.128491620111731,
"grad_norm": 1.321000307062693,
"learning_rate": 3.572473342175436e-07,
"logits/chosen": -3.3377838134765625,
"logits/rejected": -3.445711374282837,
"logps/chosen": -2.820636749267578,
"logps/rejected": -41.340660095214844,
"loss": 0.0904,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.604174852371216,
"rewards/margins": 2.604174852371216,
"rewards/rejected": 0.0,
"step": 1813
},
{
"epoch": 10.134078212290502,
"grad_norm": 1.5393829717655072,
"learning_rate": 3.566632668813878e-07,
"logits/chosen": -3.440643787384033,
"logits/rejected": -3.596426010131836,
"logps/chosen": -18.2464542388916,
"logps/rejected": -34.368446350097656,
"loss": 0.0963,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2566850185394287,
"rewards/margins": 3.2566850185394287,
"rewards/rejected": 0.0,
"step": 1814
},
{
"epoch": 10.139664804469273,
"grad_norm": 1.9601218109100802,
"learning_rate": 3.560794125691441e-07,
"logits/chosen": -3.508427858352661,
"logits/rejected": -3.4402294158935547,
"logps/chosen": -1.8796428442001343,
"logps/rejected": -39.425445556640625,
"loss": 0.0841,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0107765197753906,
"rewards/margins": 3.0107765197753906,
"rewards/rejected": 0.0,
"step": 1815
},
{
"epoch": 10.145251396648044,
"grad_norm": 1.26301681295152,
"learning_rate": 3.5549577214852415e-07,
"logits/chosen": -3.3128695487976074,
"logits/rejected": -3.274829864501953,
"logps/chosen": -18.262205123901367,
"logps/rejected": -34.38519287109375,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6635007858276367,
"rewards/margins": 3.6635007858276367,
"rewards/rejected": 0.0,
"step": 1816
},
{
"epoch": 10.150837988826815,
"grad_norm": 0.5983902495600948,
"learning_rate": 3.549123464869218e-07,
"logits/chosen": -3.3344552516937256,
"logits/rejected": -3.501915693283081,
"logps/chosen": -1.0359522104263306,
"logps/rejected": -86.96023559570312,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0813112258911133,
"rewards/margins": 2.0813112258911133,
"rewards/rejected": 0.0,
"step": 1817
},
{
"epoch": 10.156424581005586,
"grad_norm": 0.5444704846871248,
"learning_rate": 3.5432913645141103e-07,
"logits/chosen": -3.424379825592041,
"logits/rejected": -3.5716700553894043,
"logps/chosen": -1.2069406509399414,
"logps/rejected": -49.48115158081055,
"loss": 0.0938,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.398582696914673,
"rewards/margins": 2.398582696914673,
"rewards/rejected": 0.0,
"step": 1818
},
{
"epoch": 10.162011173184357,
"grad_norm": 0.7436498642751891,
"learning_rate": 3.537461429087464e-07,
"logits/chosen": -3.156156539916992,
"logits/rejected": -2.994645833969116,
"logps/chosen": -3.9752585887908936,
"logps/rejected": -39.014522552490234,
"loss": 0.0765,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7616236209869385,
"rewards/margins": 2.7616236209869385,
"rewards/rejected": 0.0,
"step": 1819
},
{
"epoch": 10.167597765363128,
"grad_norm": 0.6418105324988261,
"learning_rate": 3.531633667253594e-07,
"logits/chosen": -3.4729127883911133,
"logits/rejected": -3.5277273654937744,
"logps/chosen": -4.411098003387451,
"logps/rejected": -68.09147644042969,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.95358943939209,
"rewards/margins": 2.95358943939209,
"rewards/rejected": 0.0,
"step": 1820
},
{
"epoch": 10.167597765363128,
"eval_logits/chosen": -3.306750535964966,
"eval_logits/rejected": -3.43415904045105,
"eval_logps/chosen": -28.0472354888916,
"eval_logps/rejected": -50.623146057128906,
"eval_loss": 0.8597971200942993,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.20228078961372375,
"eval_rewards/margins": 0.20228078961372375,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6905,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 1820
},
{
"epoch": 10.1731843575419,
"grad_norm": 0.5378953337576965,
"learning_rate": 3.525808087673597e-07,
"logits/chosen": -3.7045185565948486,
"logits/rejected": -3.641814947128296,
"logps/chosen": -1.8553470373153687,
"logps/rejected": -30.911592483520508,
"loss": 0.0801,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6366233825683594,
"rewards/margins": 2.6366233825683594,
"rewards/rejected": 0.0,
"step": 1821
},
{
"epoch": 10.17877094972067,
"grad_norm": 1.4765511353150538,
"learning_rate": 3.5199846990053227e-07,
"logits/chosen": -3.639594078063965,
"logits/rejected": -3.5415031909942627,
"logps/chosen": -15.877860069274902,
"logps/rejected": -27.99404525756836,
"loss": 0.115,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.320511817932129,
"rewards/margins": 3.320511817932129,
"rewards/rejected": 0.0,
"step": 1822
},
{
"epoch": 10.184357541899441,
"grad_norm": 0.5777010850299931,
"learning_rate": 3.5141635099033615e-07,
"logits/chosen": -3.521170139312744,
"logits/rejected": -3.6726415157318115,
"logps/chosen": -1.5075368881225586,
"logps/rejected": -33.77130889892578,
"loss": 0.1013,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0790767669677734,
"rewards/margins": 2.0790767669677734,
"rewards/rejected": 0.0,
"step": 1823
},
{
"epoch": 10.189944134078212,
"grad_norm": 1.0164424489320223,
"learning_rate": 3.508344529019036e-07,
"logits/chosen": -3.419110059738159,
"logits/rejected": -3.4553027153015137,
"logps/chosen": -3.7882261276245117,
"logps/rejected": -39.779170989990234,
"loss": 0.0785,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.990870714187622,
"rewards/margins": 2.990870714187622,
"rewards/rejected": 0.0,
"step": 1824
},
{
"epoch": 10.195530726256983,
"grad_norm": 0.4876300906458077,
"learning_rate": 3.5025277650003924e-07,
"logits/chosen": -3.7110509872436523,
"logits/rejected": -3.6583542823791504,
"logps/chosen": -0.771490216255188,
"logps/rejected": -33.36054229736328,
"loss": 0.1132,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.275239944458008,
"rewards/margins": 2.275239944458008,
"rewards/rejected": 0.0,
"step": 1825
},
{
"epoch": 10.201117318435754,
"grad_norm": 0.4735647527961175,
"learning_rate": 3.496713226492173e-07,
"logits/chosen": -3.451616048812866,
"logits/rejected": -3.6345643997192383,
"logps/chosen": -4.551021099090576,
"logps/rejected": -41.84909439086914,
"loss": 0.0826,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.69582462310791,
"rewards/margins": 2.69582462310791,
"rewards/rejected": 0.0,
"step": 1826
},
{
"epoch": 10.206703910614525,
"grad_norm": 0.5634175579177542,
"learning_rate": 3.49090092213582e-07,
"logits/chosen": -3.3760182857513428,
"logits/rejected": -3.4502804279327393,
"logps/chosen": -3.664790630340576,
"logps/rejected": -55.21084976196289,
"loss": 0.0889,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8722262382507324,
"rewards/margins": 2.8722262382507324,
"rewards/rejected": 0.0,
"step": 1827
},
{
"epoch": 10.212290502793296,
"grad_norm": 0.5762721791745985,
"learning_rate": 3.485090860569454e-07,
"logits/chosen": -3.5474774837493896,
"logits/rejected": -3.703139543533325,
"logps/chosen": -8.969099044799805,
"logps/rejected": -69.3248291015625,
"loss": 0.0969,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.520266532897949,
"rewards/margins": 3.520266532897949,
"rewards/rejected": 0.0,
"step": 1828
},
{
"epoch": 10.217877094972067,
"grad_norm": 0.5241805173650772,
"learning_rate": 3.4792830504278576e-07,
"logits/chosen": -3.5396921634674072,
"logits/rejected": -3.5115535259246826,
"logps/chosen": -1.3261332511901855,
"logps/rejected": -61.18022918701172,
"loss": 0.1027,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5625224113464355,
"rewards/margins": 2.5625224113464355,
"rewards/rejected": 0.0,
"step": 1829
},
{
"epoch": 10.223463687150838,
"grad_norm": 1.2202257954137754,
"learning_rate": 3.4734775003424733e-07,
"logits/chosen": -3.3850276470184326,
"logits/rejected": -3.4800312519073486,
"logps/chosen": -1.661970615386963,
"logps/rejected": -66.8587646484375,
"loss": 0.103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.148455858230591,
"rewards/margins": 2.148455858230591,
"rewards/rejected": 0.0,
"step": 1830
},
{
"epoch": 10.22905027932961,
"grad_norm": 0.9781709482587243,
"learning_rate": 3.4676742189413845e-07,
"logits/chosen": -3.4181766510009766,
"logits/rejected": -3.443974256515503,
"logps/chosen": -0.8105766773223877,
"logps/rejected": -30.953338623046875,
"loss": 0.0763,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.427340507507324,
"rewards/margins": 2.427340507507324,
"rewards/rejected": 0.0,
"step": 1831
},
{
"epoch": 10.23463687150838,
"grad_norm": 1.0589211914881005,
"learning_rate": 3.4618732148492967e-07,
"logits/chosen": -3.196134328842163,
"logits/rejected": -3.3517751693725586,
"logps/chosen": -7.323477268218994,
"logps/rejected": -57.46253967285156,
"loss": 0.1267,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3929636478424072,
"rewards/margins": 3.3929636478424072,
"rewards/rejected": 0.0,
"step": 1832
},
{
"epoch": 10.240223463687151,
"grad_norm": 0.5288517081599233,
"learning_rate": 3.456074496687538e-07,
"logits/chosen": -3.1686713695526123,
"logits/rejected": -3.2956857681274414,
"logps/chosen": -2.033550262451172,
"logps/rejected": -42.6346435546875,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9380688667297363,
"rewards/margins": 2.9380688667297363,
"rewards/rejected": 0.0,
"step": 1833
},
{
"epoch": 10.245810055865922,
"grad_norm": 1.1668173711901333,
"learning_rate": 3.4502780730740356e-07,
"logits/chosen": -3.2127201557159424,
"logits/rejected": -3.16595721244812,
"logps/chosen": -6.029697895050049,
"logps/rejected": -33.397308349609375,
"loss": 0.1084,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8888542652130127,
"rewards/margins": 2.8888542652130127,
"rewards/rejected": 0.0,
"step": 1834
},
{
"epoch": 10.251396648044693,
"grad_norm": 1.2717630202495658,
"learning_rate": 3.444483952623306e-07,
"logits/chosen": -3.254587173461914,
"logits/rejected": -3.267111301422119,
"logps/chosen": -12.352456092834473,
"logps/rejected": -44.57154846191406,
"loss": 0.0763,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.622567892074585,
"rewards/margins": 3.622567892074585,
"rewards/rejected": 0.0,
"step": 1835
},
{
"epoch": 10.256983240223464,
"grad_norm": 3.219122668949424,
"learning_rate": 3.4386921439464443e-07,
"logits/chosen": -3.407097578048706,
"logits/rejected": -3.4369802474975586,
"logps/chosen": -16.4822998046875,
"logps/rejected": -44.601016998291016,
"loss": 0.1251,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7790446281433105,
"rewards/margins": 3.7790446281433105,
"rewards/rejected": 0.0,
"step": 1836
},
{
"epoch": 10.262569832402235,
"grad_norm": 1.3450125333718765,
"learning_rate": 3.432902655651113e-07,
"logits/chosen": -3.3426268100738525,
"logits/rejected": -3.3855786323547363,
"logps/chosen": -5.525186538696289,
"logps/rejected": -82.24298095703125,
"loss": 0.1268,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.879131317138672,
"rewards/margins": 2.879131317138672,
"rewards/rejected": 0.0,
"step": 1837
},
{
"epoch": 10.268156424581006,
"grad_norm": 0.6178454890474379,
"learning_rate": 3.427115496341518e-07,
"logits/chosen": -3.4179000854492188,
"logits/rejected": -3.521860361099243,
"logps/chosen": -8.484330177307129,
"logps/rejected": -32.62311553955078,
"loss": 0.0881,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.538153648376465,
"rewards/margins": 3.538153648376465,
"rewards/rejected": 0.0,
"step": 1838
},
{
"epoch": 10.273743016759777,
"grad_norm": 1.7629890611681795,
"learning_rate": 3.421330674618411e-07,
"logits/chosen": -3.66129994392395,
"logits/rejected": -3.7412681579589844,
"logps/chosen": -0.4885304570198059,
"logps/rejected": -49.49581527709961,
"loss": 0.1142,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7407662868499756,
"rewards/margins": 1.7407662868499756,
"rewards/rejected": 0.0,
"step": 1839
},
{
"epoch": 10.279329608938548,
"grad_norm": 0.8929129133825531,
"learning_rate": 3.4155481990790687e-07,
"logits/chosen": -3.2866475582122803,
"logits/rejected": -3.336662769317627,
"logps/chosen": -1.3250962495803833,
"logps/rejected": -55.27764892578125,
"loss": 0.0897,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3882293701171875,
"rewards/margins": 2.3882293701171875,
"rewards/rejected": 0.0,
"step": 1840
},
{
"epoch": 10.279329608938548,
"eval_logits/chosen": -3.2945797443389893,
"eval_logits/rejected": -3.4231491088867188,
"eval_logps/chosen": -28.080852508544922,
"eval_logps/rejected": -51.409461975097656,
"eval_loss": 0.8763619661331177,
"eval_rewards/accuracies": 0.625,
"eval_rewards/chosen": 0.19891873002052307,
"eval_rewards/margins": 0.19891873002052307,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6942,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 1840
},
{
"epoch": 10.28491620111732,
"grad_norm": 0.708406499896884,
"learning_rate": 3.4097680783172763e-07,
"logits/chosen": -3.390831470489502,
"logits/rejected": -3.603947877883911,
"logps/chosen": -0.34076374769210815,
"logps/rejected": -57.07746505737305,
"loss": 0.1065,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.540673851966858,
"rewards/margins": 1.540673851966858,
"rewards/rejected": 0.0,
"step": 1841
},
{
"epoch": 10.29050279329609,
"grad_norm": 1.2933475544436752,
"learning_rate": 3.403990320923325e-07,
"logits/chosen": -3.458280086517334,
"logits/rejected": -3.5281569957733154,
"logps/chosen": -1.0999212265014648,
"logps/rejected": -42.88365936279297,
"loss": 0.0902,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.469132423400879,
"rewards/margins": 2.469132423400879,
"rewards/rejected": 0.0,
"step": 1842
},
{
"epoch": 10.296089385474861,
"grad_norm": 0.5619472867290808,
"learning_rate": 3.3982149354839904e-07,
"logits/chosen": -3.5612244606018066,
"logits/rejected": -3.3477532863616943,
"logps/chosen": -3.444309711456299,
"logps/rejected": -82.17162322998047,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4129276275634766,
"rewards/margins": 3.4129276275634766,
"rewards/rejected": 0.0,
"step": 1843
},
{
"epoch": 10.30167597765363,
"grad_norm": 0.50006164653966,
"learning_rate": 3.392441930582525e-07,
"logits/chosen": -3.3518800735473633,
"logits/rejected": -3.2428812980651855,
"logps/chosen": -2.3187599182128906,
"logps/rejected": -53.334068298339844,
"loss": 0.1265,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5861599445343018,
"rewards/margins": 2.5861599445343018,
"rewards/rejected": 0.0,
"step": 1844
},
{
"epoch": 10.307262569832401,
"grad_norm": 1.164540651723829,
"learning_rate": 3.3866713147986393e-07,
"logits/chosen": -3.6116364002227783,
"logits/rejected": -3.65683650970459,
"logps/chosen": -2.0827600955963135,
"logps/rejected": -35.13486862182617,
"loss": 0.1174,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.676473379135132,
"rewards/margins": 2.676473379135132,
"rewards/rejected": 0.0,
"step": 1845
},
{
"epoch": 10.312849162011172,
"grad_norm": 0.47667747136271305,
"learning_rate": 3.380903096708501e-07,
"logits/chosen": -3.578843116760254,
"logits/rejected": -3.4206881523132324,
"logps/chosen": -3.8796091079711914,
"logps/rejected": -28.51175308227539,
"loss": 0.0552,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4883346557617188,
"rewards/margins": 3.4883346557617188,
"rewards/rejected": 0.0,
"step": 1846
},
{
"epoch": 10.318435754189943,
"grad_norm": 0.6376174597724137,
"learning_rate": 3.375137284884703e-07,
"logits/chosen": -3.4132537841796875,
"logits/rejected": -3.46468186378479,
"logps/chosen": -4.338701248168945,
"logps/rejected": -51.54572677612305,
"loss": 0.1098,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1232264041900635,
"rewards/margins": 2.1232264041900635,
"rewards/rejected": 0.0,
"step": 1847
},
{
"epoch": 10.324022346368714,
"grad_norm": 0.8348576151942313,
"learning_rate": 3.369373887896274e-07,
"logits/chosen": -3.403993606567383,
"logits/rejected": -3.4272985458374023,
"logps/chosen": -1.2190645933151245,
"logps/rejected": -32.052894592285156,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.384263277053833,
"rewards/margins": 2.384263277053833,
"rewards/rejected": 0.0,
"step": 1848
},
{
"epoch": 10.329608938547485,
"grad_norm": 1.064786522288118,
"learning_rate": 3.363612914308645e-07,
"logits/chosen": -3.5142853260040283,
"logits/rejected": -3.445352554321289,
"logps/chosen": -0.7066118121147156,
"logps/rejected": -51.051021575927734,
"loss": 0.1144,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0623998641967773,
"rewards/margins": 2.0623998641967773,
"rewards/rejected": 0.0,
"step": 1849
},
{
"epoch": 10.335195530726256,
"grad_norm": 0.5238202917225792,
"learning_rate": 3.3578543726836495e-07,
"logits/chosen": -2.6879019737243652,
"logits/rejected": -2.939012050628662,
"logps/chosen": -4.14566707611084,
"logps/rejected": -97.30796813964844,
"loss": 0.1184,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2669615745544434,
"rewards/margins": 2.2669615745544434,
"rewards/rejected": 0.0,
"step": 1850
},
{
"epoch": 10.340782122905027,
"grad_norm": 0.683918591125821,
"learning_rate": 3.352098271579509e-07,
"logits/chosen": -3.4271247386932373,
"logits/rejected": -3.4677963256835938,
"logps/chosen": -1.8000907897949219,
"logps/rejected": -54.379390716552734,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.590089797973633,
"rewards/margins": 2.590089797973633,
"rewards/rejected": 0.0,
"step": 1851
},
{
"epoch": 10.346368715083798,
"grad_norm": 1.2349938001511842,
"learning_rate": 3.3463446195508096e-07,
"logits/chosen": -3.3436098098754883,
"logits/rejected": -3.552900552749634,
"logps/chosen": -2.0247182846069336,
"logps/rejected": -39.91437911987305,
"loss": 0.1227,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3063344955444336,
"rewards/margins": 2.3063344955444336,
"rewards/rejected": 0.0,
"step": 1852
},
{
"epoch": 10.35195530726257,
"grad_norm": 0.9950521200296845,
"learning_rate": 3.340593425148506e-07,
"logits/chosen": -3.435642719268799,
"logits/rejected": -3.6248693466186523,
"logps/chosen": -3.458634376525879,
"logps/rejected": -62.87099838256836,
"loss": 0.1006,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.806347370147705,
"rewards/margins": 2.806347370147705,
"rewards/rejected": 0.0,
"step": 1853
},
{
"epoch": 10.35754189944134,
"grad_norm": 0.47763450859602763,
"learning_rate": 3.3348446969198987e-07,
"logits/chosen": -3.394721746444702,
"logits/rejected": -3.4811933040618896,
"logps/chosen": -3.7385308742523193,
"logps/rejected": -42.472137451171875,
"loss": 0.088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.706210136413574,
"rewards/margins": 2.706210136413574,
"rewards/rejected": 0.0,
"step": 1854
},
{
"epoch": 10.363128491620111,
"grad_norm": 1.3451192899406743,
"learning_rate": 3.3290984434086167e-07,
"logits/chosen": -2.9988791942596436,
"logits/rejected": -3.119492530822754,
"logps/chosen": -9.244361877441406,
"logps/rejected": -43.439910888671875,
"loss": 0.0946,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.273994207382202,
"rewards/margins": 3.273994207382202,
"rewards/rejected": 0.0,
"step": 1855
},
{
"epoch": 10.368715083798882,
"grad_norm": 2.619526409237513,
"learning_rate": 3.32335467315462e-07,
"logits/chosen": -3.699422836303711,
"logits/rejected": -3.593337059020996,
"logps/chosen": -7.9409332275390625,
"logps/rejected": -41.3218994140625,
"loss": 0.0959,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.015827178955078,
"rewards/margins": 4.015827178955078,
"rewards/rejected": 0.0,
"step": 1856
},
{
"epoch": 10.374301675977653,
"grad_norm": 0.5557399530140986,
"learning_rate": 3.317613394694175e-07,
"logits/chosen": -3.137514114379883,
"logits/rejected": -3.0139694213867188,
"logps/chosen": -0.49404364824295044,
"logps/rejected": -95.09521484375,
"loss": 0.1072,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7863245010375977,
"rewards/margins": 1.7863245010375977,
"rewards/rejected": 0.0,
"step": 1857
},
{
"epoch": 10.379888268156424,
"grad_norm": 0.6493478016639063,
"learning_rate": 3.3118746165598397e-07,
"logits/chosen": -3.3223178386688232,
"logits/rejected": -3.1887118816375732,
"logps/chosen": -2.961448907852173,
"logps/rejected": -81.54209899902344,
"loss": 0.1037,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4504408836364746,
"rewards/margins": 2.4504408836364746,
"rewards/rejected": 0.0,
"step": 1858
},
{
"epoch": 10.385474860335195,
"grad_norm": 1.9270303079384823,
"learning_rate": 3.306138347280463e-07,
"logits/chosen": -3.484477996826172,
"logits/rejected": -3.3443188667297363,
"logps/chosen": -2.8306005001068115,
"logps/rejected": -41.490234375,
"loss": 0.1235,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.867560386657715,
"rewards/margins": 2.867560386657715,
"rewards/rejected": 0.0,
"step": 1859
},
{
"epoch": 10.391061452513966,
"grad_norm": 0.8294739883535532,
"learning_rate": 3.3004045953811646e-07,
"logits/chosen": -3.254361629486084,
"logits/rejected": -3.349991798400879,
"logps/chosen": -1.9156274795532227,
"logps/rejected": -110.4948501586914,
"loss": 0.0831,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4569637775421143,
"rewards/margins": 2.4569637775421143,
"rewards/rejected": 0.0,
"step": 1860
},
{
"epoch": 10.391061452513966,
"eval_logits/chosen": -3.29571533203125,
"eval_logits/rejected": -3.4226455688476562,
"eval_logps/chosen": -28.695659637451172,
"eval_logps/rejected": -52.63666534423828,
"eval_loss": 0.9198136925697327,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.13743813335895538,
"eval_rewards/margins": 0.13743813335895538,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6799,
"eval_samples_per_second": 9.486,
"eval_steps_per_second": 0.306,
"step": 1860
},
{
"epoch": 10.396648044692737,
"grad_norm": 0.6064433511872755,
"learning_rate": 3.2946733693833174e-07,
"logits/chosen": -3.3379640579223633,
"logits/rejected": -3.3924098014831543,
"logps/chosen": -2.839841365814209,
"logps/rejected": -57.46247100830078,
"loss": 0.1402,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3916800022125244,
"rewards/margins": 2.3916800022125244,
"rewards/rejected": 0.0,
"step": 1861
},
{
"epoch": 10.402234636871508,
"grad_norm": 2.1087870060840728,
"learning_rate": 3.288944677804546e-07,
"logits/chosen": -3.341815233230591,
"logits/rejected": -3.523813009262085,
"logps/chosen": -5.720947265625,
"logps/rejected": -101.80484008789062,
"loss": 0.1137,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.270772933959961,
"rewards/margins": 3.270772933959961,
"rewards/rejected": 0.0,
"step": 1862
},
{
"epoch": 10.40782122905028,
"grad_norm": 0.4689909392959233,
"learning_rate": 3.2832185291587053e-07,
"logits/chosen": -3.3669724464416504,
"logits/rejected": -3.6839537620544434,
"logps/chosen": -0.6228213310241699,
"logps/rejected": -40.415008544921875,
"loss": 0.1286,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0144503116607666,
"rewards/margins": 2.0144503116607666,
"rewards/rejected": 0.0,
"step": 1863
},
{
"epoch": 10.41340782122905,
"grad_norm": 4.798617740793981,
"learning_rate": 3.2774949319558723e-07,
"logits/chosen": -3.4009153842926025,
"logits/rejected": -3.3676819801330566,
"logps/chosen": -6.605515956878662,
"logps/rejected": -32.565277099609375,
"loss": 0.1368,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.265873432159424,
"rewards/margins": 3.265873432159424,
"rewards/rejected": 0.0,
"step": 1864
},
{
"epoch": 10.418994413407821,
"grad_norm": 0.6096916993770324,
"learning_rate": 3.271773894702331e-07,
"logits/chosen": -3.2294697761535645,
"logits/rejected": -3.4293994903564453,
"logps/chosen": -0.6215426921844482,
"logps/rejected": -75.20904541015625,
"loss": 0.1021,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0924994945526123,
"rewards/margins": 2.0924994945526123,
"rewards/rejected": 0.0,
"step": 1865
},
{
"epoch": 10.424581005586592,
"grad_norm": 0.6886767599347046,
"learning_rate": 3.2660554259005645e-07,
"logits/chosen": -3.4089062213897705,
"logits/rejected": -3.3943350315093994,
"logps/chosen": -6.764495372772217,
"logps/rejected": -39.120704650878906,
"loss": 0.1189,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6976699829101562,
"rewards/margins": 3.6976699829101562,
"rewards/rejected": 0.0,
"step": 1866
},
{
"epoch": 10.430167597765363,
"grad_norm": 0.5453366272332151,
"learning_rate": 3.260339534049232e-07,
"logits/chosen": -3.6877708435058594,
"logits/rejected": -3.6817972660064697,
"logps/chosen": -13.224329948425293,
"logps/rejected": -42.58197021484375,
"loss": 0.1195,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.314650535583496,
"rewards/margins": 3.314650535583496,
"rewards/rejected": 0.0,
"step": 1867
},
{
"epoch": 10.435754189944134,
"grad_norm": 0.7275898467664447,
"learning_rate": 3.254626227643168e-07,
"logits/chosen": -3.495012044906616,
"logits/rejected": -3.6272099018096924,
"logps/chosen": -8.475492477416992,
"logps/rejected": -24.994121551513672,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5595078468322754,
"rewards/margins": 3.5595078468322754,
"rewards/rejected": 0.0,
"step": 1868
},
{
"epoch": 10.441340782122905,
"grad_norm": 2.484363144221312,
"learning_rate": 3.248915515173366e-07,
"logits/chosen": -3.5623674392700195,
"logits/rejected": -3.687246799468994,
"logps/chosen": -7.592851638793945,
"logps/rejected": -52.94874954223633,
"loss": 0.1213,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.052880764007568,
"rewards/margins": 4.052880764007568,
"rewards/rejected": 0.0,
"step": 1869
},
{
"epoch": 10.446927374301676,
"grad_norm": 0.603735940648404,
"learning_rate": 3.243207405126958e-07,
"logits/chosen": -3.6547141075134277,
"logits/rejected": -3.6865251064300537,
"logps/chosen": -6.05224609375,
"logps/rejected": -21.353927612304688,
"loss": 0.1025,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1725809574127197,
"rewards/margins": 2.1725809574127197,
"rewards/rejected": 0.0,
"step": 1870
},
{
"epoch": 10.452513966480447,
"grad_norm": 1.1344001752076176,
"learning_rate": 3.237501905987214e-07,
"logits/chosen": -2.693904161453247,
"logits/rejected": -2.6437413692474365,
"logps/chosen": -24.93901824951172,
"logps/rejected": -94.80763244628906,
"loss": 0.0998,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0128211975097656,
"rewards/margins": 3.0128211975097656,
"rewards/rejected": 0.0,
"step": 1871
},
{
"epoch": 10.458100558659218,
"grad_norm": 0.7219791491200507,
"learning_rate": 3.2317990262335217e-07,
"logits/chosen": -3.526754856109619,
"logits/rejected": -3.445040464401245,
"logps/chosen": -8.690518379211426,
"logps/rejected": -58.01318359375,
"loss": 0.0873,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.334439754486084,
"rewards/margins": 3.334439754486084,
"rewards/rejected": 0.0,
"step": 1872
},
{
"epoch": 10.46368715083799,
"grad_norm": 0.6052807808897211,
"learning_rate": 3.2260987743413756e-07,
"logits/chosen": -3.3755505084991455,
"logits/rejected": -3.410623073577881,
"logps/chosen": -1.857917308807373,
"logps/rejected": -54.81926345825195,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1309776306152344,
"rewards/margins": 2.1309776306152344,
"rewards/rejected": 0.0,
"step": 1873
},
{
"epoch": 10.46927374301676,
"grad_norm": 0.597599134085182,
"learning_rate": 3.220401158782364e-07,
"logits/chosen": -3.3665552139282227,
"logits/rejected": -3.3034322261810303,
"logps/chosen": -8.424214363098145,
"logps/rejected": -34.91681671142578,
"loss": 0.0913,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7273669242858887,
"rewards/margins": 2.7273669242858887,
"rewards/rejected": 0.0,
"step": 1874
},
{
"epoch": 10.474860335195531,
"grad_norm": 0.5489151263112975,
"learning_rate": 3.214706188024162e-07,
"logits/chosen": -3.3839004039764404,
"logits/rejected": -3.606938123703003,
"logps/chosen": -1.805442214012146,
"logps/rejected": -63.92960739135742,
"loss": 0.1057,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.525209903717041,
"rewards/margins": 2.525209903717041,
"rewards/rejected": 0.0,
"step": 1875
},
{
"epoch": 10.480446927374302,
"grad_norm": 0.703639741134572,
"learning_rate": 3.2090138705305065e-07,
"logits/chosen": -3.439547300338745,
"logits/rejected": -3.591928005218506,
"logps/chosen": -2.223017454147339,
"logps/rejected": -49.858394622802734,
"loss": 0.1127,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5340662002563477,
"rewards/margins": 2.5340662002563477,
"rewards/rejected": 0.0,
"step": 1876
},
{
"epoch": 10.486033519553073,
"grad_norm": 1.854366482925165,
"learning_rate": 3.203324214761198e-07,
"logits/chosen": -3.5880889892578125,
"logits/rejected": -3.425997257232666,
"logps/chosen": -3.504964590072632,
"logps/rejected": -44.9100341796875,
"loss": 0.1284,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7333126068115234,
"rewards/margins": 2.7333126068115234,
"rewards/rejected": 0.0,
"step": 1877
},
{
"epoch": 10.491620111731844,
"grad_norm": 0.6785987682925823,
"learning_rate": 3.1976372291720756e-07,
"logits/chosen": -3.468360662460327,
"logits/rejected": -3.4381210803985596,
"logps/chosen": -3.877368211746216,
"logps/rejected": -41.90599060058594,
"loss": 0.09,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.057978630065918,
"rewards/margins": 3.057978630065918,
"rewards/rejected": 0.0,
"step": 1878
},
{
"epoch": 10.497206703910614,
"grad_norm": 2.0267640363629873,
"learning_rate": 3.191952922215013e-07,
"logits/chosen": -3.4153389930725098,
"logits/rejected": -3.2908735275268555,
"logps/chosen": -17.036602020263672,
"logps/rejected": -59.45793151855469,
"loss": 0.082,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0718488693237305,
"rewards/margins": 3.0718488693237305,
"rewards/rejected": 0.0,
"step": 1879
},
{
"epoch": 10.502793296089386,
"grad_norm": 1.2286848174016436,
"learning_rate": 3.186271302337906e-07,
"logits/chosen": -3.3930225372314453,
"logits/rejected": -3.4970273971557617,
"logps/chosen": -6.700510501861572,
"logps/rejected": -53.76878356933594,
"loss": 0.0949,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4969258308410645,
"rewards/margins": 2.4969258308410645,
"rewards/rejected": 0.0,
"step": 1880
},
{
"epoch": 10.502793296089386,
"eval_logits/chosen": -3.3038909435272217,
"eval_logits/rejected": -3.43035626411438,
"eval_logps/chosen": -28.814382553100586,
"eval_logps/rejected": -53.03582000732422,
"eval_loss": 0.9273196458816528,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.12556585669517517,
"eval_rewards/margins": 0.12556585669517517,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.722,
"eval_samples_per_second": 9.474,
"eval_steps_per_second": 0.306,
"step": 1880
},
{
"epoch": 10.508379888268156,
"grad_norm": 1.3767565364552767,
"learning_rate": 3.180592377984649e-07,
"logits/chosen": -3.4134857654571533,
"logits/rejected": -3.629716396331787,
"logps/chosen": -3.9373466968536377,
"logps/rejected": -50.44145202636719,
"loss": 0.098,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.643033742904663,
"rewards/margins": 2.643033742904663,
"rewards/rejected": 0.0,
"step": 1881
},
{
"epoch": 10.513966480446927,
"grad_norm": 0.5177749659849222,
"learning_rate": 3.174916157595138e-07,
"logits/chosen": -3.090735912322998,
"logits/rejected": -3.250361919403076,
"logps/chosen": -12.252424240112305,
"logps/rejected": -33.36803436279297,
"loss": 0.1161,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7724671363830566,
"rewards/margins": 3.7724671363830566,
"rewards/rejected": 0.0,
"step": 1882
},
{
"epoch": 10.519553072625698,
"grad_norm": 0.8496090813239161,
"learning_rate": 3.169242649605246e-07,
"logits/chosen": -3.366504669189453,
"logits/rejected": -3.426129102706909,
"logps/chosen": -1.0884718894958496,
"logps/rejected": -38.06122589111328,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3657970428466797,
"rewards/margins": 2.3657970428466797,
"rewards/rejected": 0.0,
"step": 1883
},
{
"epoch": 10.525139664804469,
"grad_norm": 0.8796284574737169,
"learning_rate": 3.163571862446818e-07,
"logits/chosen": -3.7220442295074463,
"logits/rejected": -3.6978707313537598,
"logps/chosen": -7.9090800285339355,
"logps/rejected": -24.3577938079834,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8309731483459473,
"rewards/margins": 2.8309731483459473,
"rewards/rejected": 0.0,
"step": 1884
},
{
"epoch": 10.53072625698324,
"grad_norm": 0.5312805536348916,
"learning_rate": 3.15790380454765e-07,
"logits/chosen": -3.2869033813476562,
"logits/rejected": -3.2248241901397705,
"logps/chosen": -2.2369425296783447,
"logps/rejected": -33.74311828613281,
"loss": 0.098,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.060045003890991,
"rewards/margins": 3.060045003890991,
"rewards/rejected": 0.0,
"step": 1885
},
{
"epoch": 10.53631284916201,
"grad_norm": 0.7302794320445865,
"learning_rate": 3.152238484331491e-07,
"logits/chosen": -3.4044575691223145,
"logits/rejected": -3.6623547077178955,
"logps/chosen": -6.0326924324035645,
"logps/rejected": -89.37249755859375,
"loss": 0.0956,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8519020080566406,
"rewards/margins": 2.8519020080566406,
"rewards/rejected": 0.0,
"step": 1886
},
{
"epoch": 10.541899441340782,
"grad_norm": 0.5663235521646847,
"learning_rate": 3.14657591021801e-07,
"logits/chosen": -3.168351173400879,
"logits/rejected": -3.2040045261383057,
"logps/chosen": -2.657283067703247,
"logps/rejected": -60.280731201171875,
"loss": 0.1132,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.359914779663086,
"rewards/margins": 2.359914779663086,
"rewards/rejected": 0.0,
"step": 1887
},
{
"epoch": 10.547486033519553,
"grad_norm": 0.9185305140268429,
"learning_rate": 3.140916090622803e-07,
"logits/chosen": -3.514221668243408,
"logits/rejected": -3.578711748123169,
"logps/chosen": -0.9986670017242432,
"logps/rejected": -43.1521110534668,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.387401580810547,
"rewards/margins": 2.387401580810547,
"rewards/rejected": 0.0,
"step": 1888
},
{
"epoch": 10.553072625698324,
"grad_norm": 1.4701073629648287,
"learning_rate": 3.135259033957373e-07,
"logits/chosen": -3.2775659561157227,
"logits/rejected": -3.515453338623047,
"logps/chosen": -8.058919906616211,
"logps/rejected": -49.19865417480469,
"loss": 0.1299,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6242642402648926,
"rewards/margins": 3.6242642402648926,
"rewards/rejected": 0.0,
"step": 1889
},
{
"epoch": 10.558659217877095,
"grad_norm": 2.8172893416366644,
"learning_rate": 3.129604748629108e-07,
"logits/chosen": -3.2192795276641846,
"logits/rejected": -3.2048914432525635,
"logps/chosen": -0.8653272390365601,
"logps/rejected": -58.21757507324219,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.102515697479248,
"rewards/margins": 2.102515697479248,
"rewards/rejected": 0.0,
"step": 1890
},
{
"epoch": 10.564245810055866,
"grad_norm": 0.5448801466788024,
"learning_rate": 3.123953243041286e-07,
"logits/chosen": -3.1389858722686768,
"logits/rejected": -3.3476922512054443,
"logps/chosen": -1.8807554244995117,
"logps/rejected": -46.98809814453125,
"loss": 0.1075,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8406925201416016,
"rewards/margins": 2.8406925201416016,
"rewards/rejected": 0.0,
"step": 1891
},
{
"epoch": 10.569832402234637,
"grad_norm": 0.6888829582532957,
"learning_rate": 3.118304525593052e-07,
"logits/chosen": -3.44079327583313,
"logits/rejected": -3.4656615257263184,
"logps/chosen": -0.6652265191078186,
"logps/rejected": -87.39227294921875,
"loss": 0.0885,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.701551079750061,
"rewards/margins": 1.701551079750061,
"rewards/rejected": 0.0,
"step": 1892
},
{
"epoch": 10.575418994413408,
"grad_norm": 0.8904888977785465,
"learning_rate": 3.1126586046794036e-07,
"logits/chosen": -2.9938697814941406,
"logits/rejected": -3.435041666030884,
"logps/chosen": -0.6238469481468201,
"logps/rejected": -48.46592330932617,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8748118877410889,
"rewards/margins": 1.8748118877410889,
"rewards/rejected": 0.0,
"step": 1893
},
{
"epoch": 10.581005586592179,
"grad_norm": 1.4878413639749024,
"learning_rate": 3.107015488691185e-07,
"logits/chosen": -2.7491397857666016,
"logits/rejected": -2.68963623046875,
"logps/chosen": -19.811344146728516,
"logps/rejected": -62.037689208984375,
"loss": 0.1163,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.192837953567505,
"rewards/margins": 2.192837953567505,
"rewards/rejected": 0.0,
"step": 1894
},
{
"epoch": 10.58659217877095,
"grad_norm": 0.6106390540040627,
"learning_rate": 3.101375186015074e-07,
"logits/chosen": -3.508349657058716,
"logits/rejected": -3.42073655128479,
"logps/chosen": -10.890157699584961,
"logps/rejected": -28.693105697631836,
"loss": 0.0979,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7642245292663574,
"rewards/margins": 3.7642245292663574,
"rewards/rejected": 0.0,
"step": 1895
},
{
"epoch": 10.59217877094972,
"grad_norm": 0.8537535394462678,
"learning_rate": 3.095737705033562e-07,
"logits/chosen": -3.3676419258117676,
"logits/rejected": -3.559569835662842,
"logps/chosen": -1.8120158910751343,
"logps/rejected": -109.5850830078125,
"loss": 0.0747,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8345835208892822,
"rewards/margins": 2.8345835208892822,
"rewards/rejected": 0.0,
"step": 1896
},
{
"epoch": 10.597765363128492,
"grad_norm": 0.555774535175715,
"learning_rate": 3.0901030541249507e-07,
"logits/chosen": -3.679433822631836,
"logits/rejected": -3.6491522789001465,
"logps/chosen": -5.315507411956787,
"logps/rejected": -32.05292510986328,
"loss": 0.0868,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4725875854492188,
"rewards/margins": 3.4725875854492188,
"rewards/rejected": 0.0,
"step": 1897
},
{
"epoch": 10.603351955307263,
"grad_norm": 1.2410684993215815,
"learning_rate": 3.084471241663337e-07,
"logits/chosen": -3.521740198135376,
"logits/rejected": -3.4961280822753906,
"logps/chosen": -6.980363845825195,
"logps/rejected": -42.24245071411133,
"loss": 0.085,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.250321388244629,
"rewards/margins": 3.250321388244629,
"rewards/rejected": 0.0,
"step": 1898
},
{
"epoch": 10.608938547486034,
"grad_norm": 0.5777248211280508,
"learning_rate": 3.078842276018592e-07,
"logits/chosen": -3.5176894664764404,
"logits/rejected": -3.681659698486328,
"logps/chosen": -1.005691409111023,
"logps/rejected": -30.57900619506836,
"loss": 0.1313,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.057600975036621,
"rewards/margins": 2.057600975036621,
"rewards/rejected": 0.0,
"step": 1899
},
{
"epoch": 10.614525139664805,
"grad_norm": 0.6841981164333766,
"learning_rate": 3.0732161655563663e-07,
"logits/chosen": -3.368708848953247,
"logits/rejected": -3.388643503189087,
"logps/chosen": -1.4272022247314453,
"logps/rejected": -39.79710388183594,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8871188163757324,
"rewards/margins": 2.8871188163757324,
"rewards/rejected": 0.0,
"step": 1900
},
{
"epoch": 10.614525139664805,
"eval_logits/chosen": -3.3017399311065674,
"eval_logits/rejected": -3.4273781776428223,
"eval_logps/chosen": -28.869882583618164,
"eval_logps/rejected": -52.3021240234375,
"eval_loss": 0.930737316608429,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.12001585960388184,
"eval_rewards/margins": 0.12001585960388184,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7227,
"eval_samples_per_second": 9.474,
"eval_steps_per_second": 0.306,
"step": 1900
},
{
"epoch": 10.620111731843576,
"grad_norm": 0.873393200558506,
"learning_rate": 3.06759291863806e-07,
"logits/chosen": -3.3183014392852783,
"logits/rejected": -3.3055436611175537,
"logps/chosen": -1.8771289587020874,
"logps/rejected": -41.003761291503906,
"loss": 0.0651,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.37975811958313,
"rewards/margins": 2.37975811958313,
"rewards/rejected": 0.0,
"step": 1901
},
{
"epoch": 10.625698324022347,
"grad_norm": 1.0898740440146328,
"learning_rate": 3.06197254362082e-07,
"logits/chosen": -3.221456289291382,
"logits/rejected": -3.405565023422241,
"logps/chosen": -6.760497093200684,
"logps/rejected": -75.37287902832031,
"loss": 0.1035,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8092257976531982,
"rewards/margins": 2.8092257976531982,
"rewards/rejected": 0.0,
"step": 1902
},
{
"epoch": 10.631284916201118,
"grad_norm": 0.7698520182240444,
"learning_rate": 3.0563550488575217e-07,
"logits/chosen": -3.7989161014556885,
"logits/rejected": -3.5736124515533447,
"logps/chosen": -0.87165367603302,
"logps/rejected": -71.2402114868164,
"loss": 0.0835,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5508410930633545,
"rewards/margins": 2.5508410930633545,
"rewards/rejected": 0.0,
"step": 1903
},
{
"epoch": 10.636871508379889,
"grad_norm": 1.2176089987866257,
"learning_rate": 3.0507404426967684e-07,
"logits/chosen": -3.381575107574463,
"logits/rejected": -3.1284801959991455,
"logps/chosen": -0.5544835329055786,
"logps/rejected": -59.278507232666016,
"loss": 0.124,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.175469398498535,
"rewards/margins": 2.175469398498535,
"rewards/rejected": 0.0,
"step": 1904
},
{
"epoch": 10.64245810055866,
"grad_norm": 0.7624922702151887,
"learning_rate": 3.045128733482858e-07,
"logits/chosen": -3.4666285514831543,
"logits/rejected": -3.5378479957580566,
"logps/chosen": -0.3660850524902344,
"logps/rejected": -49.37598419189453,
"loss": 0.0987,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7370697259902954,
"rewards/margins": 1.7370697259902954,
"rewards/rejected": 0.0,
"step": 1905
},
{
"epoch": 10.64804469273743,
"grad_norm": 1.1756598573532124,
"learning_rate": 3.0395199295557963e-07,
"logits/chosen": -3.420257806777954,
"logits/rejected": -3.4432029724121094,
"logps/chosen": -29.889463424682617,
"logps/rejected": -50.085670471191406,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.109011650085449,
"rewards/margins": 3.109011650085449,
"rewards/rejected": 0.0,
"step": 1906
},
{
"epoch": 10.653631284916202,
"grad_norm": 1.3235234717404352,
"learning_rate": 3.0339140392512597e-07,
"logits/chosen": -3.469914674758911,
"logits/rejected": -3.234611988067627,
"logps/chosen": -0.6308854818344116,
"logps/rejected": -52.49540710449219,
"loss": 0.0976,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2448477745056152,
"rewards/margins": 2.2448477745056152,
"rewards/rejected": 0.0,
"step": 1907
},
{
"epoch": 10.659217877094973,
"grad_norm": 1.232860094957608,
"learning_rate": 3.028311070900601e-07,
"logits/chosen": -3.393388509750366,
"logits/rejected": -3.316394567489624,
"logps/chosen": -0.7641682624816895,
"logps/rejected": -44.09846115112305,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9598150253295898,
"rewards/margins": 1.9598150253295898,
"rewards/rejected": 0.0,
"step": 1908
},
{
"epoch": 10.664804469273744,
"grad_norm": 0.7945585085668834,
"learning_rate": 3.022711032830831e-07,
"logits/chosen": -3.463810443878174,
"logits/rejected": -3.5105905532836914,
"logps/chosen": -13.247488975524902,
"logps/rejected": -35.6832275390625,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5260543823242188,
"rewards/margins": 2.5260543823242188,
"rewards/rejected": 0.0,
"step": 1909
},
{
"epoch": 10.670391061452515,
"grad_norm": 1.8275428433326804,
"learning_rate": 3.017113933364601e-07,
"logits/chosen": -3.427177667617798,
"logits/rejected": -3.3653926849365234,
"logps/chosen": -1.6435902118682861,
"logps/rejected": -40.15003204345703,
"loss": 0.079,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7192788124084473,
"rewards/margins": 2.7192788124084473,
"rewards/rejected": 0.0,
"step": 1910
},
{
"epoch": 10.675977653631286,
"grad_norm": 1.301829605088348,
"learning_rate": 3.0115197808202e-07,
"logits/chosen": -3.458570957183838,
"logits/rejected": -3.323073387145996,
"logps/chosen": -3.0687127113342285,
"logps/rejected": -42.68876266479492,
"loss": 0.1218,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2481980323791504,
"rewards/margins": 3.2481980323791504,
"rewards/rejected": 0.0,
"step": 1911
},
{
"epoch": 10.681564245810057,
"grad_norm": 1.1111341206972045,
"learning_rate": 3.005928583511533e-07,
"logits/chosen": -3.2323250770568848,
"logits/rejected": -3.360978126525879,
"logps/chosen": -24.130712509155273,
"logps/rejected": -93.03076934814453,
"loss": 0.1063,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4027507305145264,
"rewards/margins": 3.4027507305145264,
"rewards/rejected": 0.0,
"step": 1912
},
{
"epoch": 10.687150837988828,
"grad_norm": 0.597553110429445,
"learning_rate": 3.000340349748115e-07,
"logits/chosen": -3.2457923889160156,
"logits/rejected": -3.084268808364868,
"logps/chosen": -4.707247734069824,
"logps/rejected": -32.77264404296875,
"loss": 0.1144,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.896733283996582,
"rewards/margins": 2.896733283996582,
"rewards/rejected": 0.0,
"step": 1913
},
{
"epoch": 10.692737430167599,
"grad_norm": 1.8222346411955612,
"learning_rate": 2.994755087835057e-07,
"logits/chosen": -3.381150722503662,
"logits/rejected": -3.5762670040130615,
"logps/chosen": -1.53989839553833,
"logps/rejected": -61.47698974609375,
"loss": 0.1051,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.131669044494629,
"rewards/margins": 2.131669044494629,
"rewards/rejected": 0.0,
"step": 1914
},
{
"epoch": 10.69832402234637,
"grad_norm": 0.606089882489306,
"learning_rate": 2.9891728060730554e-07,
"logits/chosen": -3.354189157485962,
"logits/rejected": -3.413299560546875,
"logps/chosen": -4.776266574859619,
"logps/rejected": -37.680641174316406,
"loss": 0.1025,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6719112396240234,
"rewards/margins": 3.6719112396240234,
"rewards/rejected": 0.0,
"step": 1915
},
{
"epoch": 10.703910614525139,
"grad_norm": 0.6255216250633863,
"learning_rate": 2.98359351275837e-07,
"logits/chosen": -3.3042688369750977,
"logits/rejected": -3.2384936809539795,
"logps/chosen": -21.000625610351562,
"logps/rejected": -35.76881790161133,
"loss": 0.0752,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.684452533721924,
"rewards/margins": 3.684452533721924,
"rewards/rejected": 0.0,
"step": 1916
},
{
"epoch": 10.70949720670391,
"grad_norm": 0.5183001117662647,
"learning_rate": 2.978017216182828e-07,
"logits/chosen": -3.7192203998565674,
"logits/rejected": -3.6571333408355713,
"logps/chosen": -0.5691882371902466,
"logps/rejected": -41.70458984375,
"loss": 0.0879,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7690117359161377,
"rewards/margins": 1.7690117359161377,
"rewards/rejected": 0.0,
"step": 1917
},
{
"epoch": 10.71508379888268,
"grad_norm": 0.5499181775208685,
"learning_rate": 2.9724439246337983e-07,
"logits/chosen": -3.563899040222168,
"logits/rejected": -3.4503607749938965,
"logps/chosen": -31.96733283996582,
"logps/rejected": -61.7823600769043,
"loss": 0.1077,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.206428050994873,
"rewards/margins": 4.206428050994873,
"rewards/rejected": 0.0,
"step": 1918
},
{
"epoch": 10.720670391061452,
"grad_norm": 1.5115312931237417,
"learning_rate": 2.966873646394184e-07,
"logits/chosen": -3.1484599113464355,
"logits/rejected": -3.1858108043670654,
"logps/chosen": -3.586225748062134,
"logps/rejected": -45.302642822265625,
"loss": 0.0997,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2733733654022217,
"rewards/margins": 2.2733733654022217,
"rewards/rejected": 0.0,
"step": 1919
},
{
"epoch": 10.726256983240223,
"grad_norm": 0.62356296755553,
"learning_rate": 2.961306389742412e-07,
"logits/chosen": -3.3710219860076904,
"logits/rejected": -3.560894250869751,
"logps/chosen": -2.2886223793029785,
"logps/rejected": -25.385658264160156,
"loss": 0.0955,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3927793502807617,
"rewards/margins": 2.3927793502807617,
"rewards/rejected": 0.0,
"step": 1920
},
{
"epoch": 10.726256983240223,
"eval_logits/chosen": -3.2988133430480957,
"eval_logits/rejected": -3.424295425415039,
"eval_logps/chosen": -28.57001304626465,
"eval_logps/rejected": -52.16538619995117,
"eval_loss": 0.9159795045852661,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.15000297129154205,
"eval_rewards/margins": 0.15000297129154205,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7401,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 1920
},
{
"epoch": 10.731843575418994,
"grad_norm": 1.0503261425352703,
"learning_rate": 2.955742162952416e-07,
"logits/chosen": -3.2798068523406982,
"logits/rejected": -3.387129783630371,
"logps/chosen": -0.26641911268234253,
"logps/rejected": -74.62503814697266,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4750702381134033,
"rewards/margins": 1.4750702381134033,
"rewards/rejected": 0.0,
"step": 1921
},
{
"epoch": 10.737430167597765,
"grad_norm": 2.100259964729435,
"learning_rate": 2.950180974293629e-07,
"logits/chosen": -3.501065969467163,
"logits/rejected": -3.4448275566101074,
"logps/chosen": -1.0389482975006104,
"logps/rejected": -31.143098831176758,
"loss": 0.1366,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0664217472076416,
"rewards/margins": 2.0664217472076416,
"rewards/rejected": 0.0,
"step": 1922
},
{
"epoch": 10.743016759776536,
"grad_norm": 2.1238573896044284,
"learning_rate": 2.944622832030965e-07,
"logits/chosen": -3.509662628173828,
"logits/rejected": -3.5671849250793457,
"logps/chosen": -1.2824070453643799,
"logps/rejected": -116.43682098388672,
"loss": 0.1177,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1577115058898926,
"rewards/margins": 2.1577115058898926,
"rewards/rejected": 0.0,
"step": 1923
},
{
"epoch": 10.748603351955307,
"grad_norm": 0.839912529658205,
"learning_rate": 2.939067744424818e-07,
"logits/chosen": -3.3863625526428223,
"logits/rejected": -3.307372570037842,
"logps/chosen": -4.680670261383057,
"logps/rejected": -69.56145477294922,
"loss": 0.1003,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3522374629974365,
"rewards/margins": 2.3522374629974365,
"rewards/rejected": 0.0,
"step": 1924
},
{
"epoch": 10.754189944134078,
"grad_norm": 0.5323115968588062,
"learning_rate": 2.933515719731032e-07,
"logits/chosen": -3.5454235076904297,
"logits/rejected": -3.3904130458831787,
"logps/chosen": -1.270723819732666,
"logps/rejected": -50.822425842285156,
"loss": 0.0943,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1869544982910156,
"rewards/margins": 2.1869544982910156,
"rewards/rejected": 0.0,
"step": 1925
},
{
"epoch": 10.759776536312849,
"grad_norm": 0.5828229340637391,
"learning_rate": 2.927966766200908e-07,
"logits/chosen": -3.409114122390747,
"logits/rejected": -3.580754280090332,
"logps/chosen": -1.3685938119888306,
"logps/rejected": -67.45565032958984,
"loss": 0.1067,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.452620267868042,
"rewards/margins": 2.452620267868042,
"rewards/rejected": 0.0,
"step": 1926
},
{
"epoch": 10.76536312849162,
"grad_norm": 0.5205560299987109,
"learning_rate": 2.92242089208118e-07,
"logits/chosen": -3.6759276390075684,
"logits/rejected": -3.5998117923736572,
"logps/chosen": -19.0526123046875,
"logps/rejected": -69.45529174804688,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8056230545043945,
"rewards/margins": 3.8056230545043945,
"rewards/rejected": 0.0,
"step": 1927
},
{
"epoch": 10.77094972067039,
"grad_norm": 1.6213822192455811,
"learning_rate": 2.916878105614002e-07,
"logits/chosen": -3.3806276321411133,
"logits/rejected": -3.4644851684570312,
"logps/chosen": -0.5472477674484253,
"logps/rejected": -49.34136962890625,
"loss": 0.1178,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5164899826049805,
"rewards/margins": 1.5164899826049805,
"rewards/rejected": 0.0,
"step": 1928
},
{
"epoch": 10.776536312849162,
"grad_norm": 0.5263298691262972,
"learning_rate": 2.91133841503694e-07,
"logits/chosen": -3.7453887462615967,
"logits/rejected": -3.672898292541504,
"logps/chosen": -1.9932701587677002,
"logps/rejected": -46.162261962890625,
"loss": 0.1077,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4026293754577637,
"rewards/margins": 2.4026293754577637,
"rewards/rejected": 0.0,
"step": 1929
},
{
"epoch": 10.782122905027933,
"grad_norm": 1.0631947297476212,
"learning_rate": 2.9058018285829687e-07,
"logits/chosen": -3.469196081161499,
"logits/rejected": -3.448237895965576,
"logps/chosen": -2.9169487953186035,
"logps/rejected": -35.945030212402344,
"loss": 0.0893,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2502312660217285,
"rewards/margins": 2.2502312660217285,
"rewards/rejected": 0.0,
"step": 1930
},
{
"epoch": 10.787709497206704,
"grad_norm": 1.058273935507051,
"learning_rate": 2.900268354480432e-07,
"logits/chosen": -3.6098685264587402,
"logits/rejected": -3.6558334827423096,
"logps/chosen": -2.7207980155944824,
"logps/rejected": -37.187965393066406,
"loss": 0.1115,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.113663673400879,
"rewards/margins": 3.113663673400879,
"rewards/rejected": 0.0,
"step": 1931
},
{
"epoch": 10.793296089385475,
"grad_norm": 1.5996199889833045,
"learning_rate": 2.8947380009530643e-07,
"logits/chosen": -3.5827581882476807,
"logits/rejected": -3.387035846710205,
"logps/chosen": -3.351130485534668,
"logps/rejected": -29.75267791748047,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2918968200683594,
"rewards/margins": 2.2918968200683594,
"rewards/rejected": 0.0,
"step": 1932
},
{
"epoch": 10.798882681564246,
"grad_norm": 0.9484081597408495,
"learning_rate": 2.8892107762199536e-07,
"logits/chosen": -3.4980547428131104,
"logits/rejected": -3.5675454139709473,
"logps/chosen": -2.8483567237854004,
"logps/rejected": -50.05510711669922,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7646398544311523,
"rewards/margins": 2.7646398544311523,
"rewards/rejected": 0.0,
"step": 1933
},
{
"epoch": 10.804469273743017,
"grad_norm": 1.644980661624941,
"learning_rate": 2.8836866884955405e-07,
"logits/chosen": -3.267019271850586,
"logits/rejected": -3.480748414993286,
"logps/chosen": -6.18753719329834,
"logps/rejected": -50.49284362792969,
"loss": 0.1081,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6509337425231934,
"rewards/margins": 2.6509337425231934,
"rewards/rejected": 0.0,
"step": 1934
},
{
"epoch": 10.810055865921788,
"grad_norm": 0.5187585647974681,
"learning_rate": 2.8781657459896035e-07,
"logits/chosen": -3.3421170711517334,
"logits/rejected": -3.4056236743927,
"logps/chosen": -9.775256156921387,
"logps/rejected": -48.3869514465332,
"loss": 0.1076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.34415864944458,
"rewards/margins": 3.34415864944458,
"rewards/rejected": 0.0,
"step": 1935
},
{
"epoch": 10.815642458100559,
"grad_norm": 0.5438588348285353,
"learning_rate": 2.872647956907246e-07,
"logits/chosen": -3.78772234916687,
"logits/rejected": -3.7548537254333496,
"logps/chosen": -0.6591119170188904,
"logps/rejected": -86.69601440429688,
"loss": 0.0881,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.208374500274658,
"rewards/margins": 2.208374500274658,
"rewards/rejected": 0.0,
"step": 1936
},
{
"epoch": 10.82122905027933,
"grad_norm": 1.9386996272615775,
"learning_rate": 2.8671333294488846e-07,
"logits/chosen": -3.363569974899292,
"logits/rejected": -3.1417958736419678,
"logps/chosen": -1.0323398113250732,
"logps/rejected": -63.85051345825195,
"loss": 0.1271,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.382911443710327,
"rewards/margins": 2.382911443710327,
"rewards/rejected": 0.0,
"step": 1937
},
{
"epoch": 10.8268156424581,
"grad_norm": 0.9552522085373845,
"learning_rate": 2.8616218718102387e-07,
"logits/chosen": -3.3511106967926025,
"logits/rejected": -3.5561373233795166,
"logps/chosen": -17.523204803466797,
"logps/rejected": -42.81861114501953,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9075281620025635,
"rewards/margins": 2.9075281620025635,
"rewards/rejected": 0.0,
"step": 1938
},
{
"epoch": 10.832402234636872,
"grad_norm": 2.3717920822897516,
"learning_rate": 2.8561135921823156e-07,
"logits/chosen": -3.3507566452026367,
"logits/rejected": -3.2384164333343506,
"logps/chosen": -1.3525452613830566,
"logps/rejected": -77.60067749023438,
"loss": 0.1043,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6592729091644287,
"rewards/margins": 1.6592729091644287,
"rewards/rejected": 0.0,
"step": 1939
},
{
"epoch": 10.837988826815643,
"grad_norm": 1.4011181177985683,
"learning_rate": 2.8506084987513977e-07,
"logits/chosen": -3.2683188915252686,
"logits/rejected": -3.394500732421875,
"logps/chosen": -1.9470837116241455,
"logps/rejected": -30.30990982055664,
"loss": 0.1188,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.501068592071533,
"rewards/margins": 2.501068592071533,
"rewards/rejected": 0.0,
"step": 1940
},
{
"epoch": 10.837988826815643,
"eval_logits/chosen": -3.294837236404419,
"eval_logits/rejected": -3.4212710857391357,
"eval_logps/chosen": -28.783550262451172,
"eval_logps/rejected": -52.66058349609375,
"eval_loss": 0.921066164970398,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.12864917516708374,
"eval_rewards/margins": 0.12864917516708374,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6845,
"eval_samples_per_second": 9.485,
"eval_steps_per_second": 0.306,
"step": 1940
},
{
"epoch": 10.843575418994414,
"grad_norm": 1.4302198050688797,
"learning_rate": 2.845106599699041e-07,
"logits/chosen": -3.2402455806732178,
"logits/rejected": -3.3550610542297363,
"logps/chosen": -3.4866740703582764,
"logps/rejected": -29.693632125854492,
"loss": 0.1133,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1640701293945312,
"rewards/margins": 2.1640701293945312,
"rewards/rejected": 0.0,
"step": 1941
},
{
"epoch": 10.849162011173185,
"grad_norm": 0.7268407995244136,
"learning_rate": 2.8396079032020383e-07,
"logits/chosen": -3.356489419937134,
"logits/rejected": -3.349102020263672,
"logps/chosen": -2.3652195930480957,
"logps/rejected": -64.1489486694336,
"loss": 0.0812,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.05629301071167,
"rewards/margins": 3.05629301071167,
"rewards/rejected": 0.0,
"step": 1942
},
{
"epoch": 10.854748603351956,
"grad_norm": 0.9204822561237457,
"learning_rate": 2.834112417432437e-07,
"logits/chosen": -2.7726292610168457,
"logits/rejected": -2.7120933532714844,
"logps/chosen": -6.252779006958008,
"logps/rejected": -46.49617004394531,
"loss": 0.0784,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0688788890838623,
"rewards/margins": 3.0688788890838623,
"rewards/rejected": 0.0,
"step": 1943
},
{
"epoch": 10.860335195530727,
"grad_norm": 2.1601964192839476,
"learning_rate": 2.828620150557508e-07,
"logits/chosen": -3.430494785308838,
"logits/rejected": -3.464099407196045,
"logps/chosen": -1.9049228429794312,
"logps/rejected": -42.39170455932617,
"loss": 0.0863,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5023398399353027,
"rewards/margins": 2.5023398399353027,
"rewards/rejected": 0.0,
"step": 1944
},
{
"epoch": 10.865921787709498,
"grad_norm": 2.032897164494274,
"learning_rate": 2.823131110739737e-07,
"logits/chosen": -3.3368167877197266,
"logits/rejected": -3.2068445682525635,
"logps/chosen": -0.6238577365875244,
"logps/rejected": -100.99341583251953,
"loss": 0.12,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0053141117095947,
"rewards/margins": 2.0053141117095947,
"rewards/rejected": 0.0,
"step": 1945
},
{
"epoch": 10.871508379888269,
"grad_norm": 0.535290823782499,
"learning_rate": 2.8176453061368143e-07,
"logits/chosen": -3.3537046909332275,
"logits/rejected": -3.333434581756592,
"logps/chosen": -6.67498779296875,
"logps/rejected": -36.00105285644531,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6048285961151123,
"rewards/margins": 2.6048285961151123,
"rewards/rejected": 0.0,
"step": 1946
},
{
"epoch": 10.87709497206704,
"grad_norm": 0.568975661018824,
"learning_rate": 2.8121627449016225e-07,
"logits/chosen": -3.5854060649871826,
"logits/rejected": -3.5121772289276123,
"logps/chosen": -9.655231475830078,
"logps/rejected": -59.148563385009766,
"loss": 0.1263,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.508772850036621,
"rewards/margins": 3.508772850036621,
"rewards/rejected": 0.0,
"step": 1947
},
{
"epoch": 10.88268156424581,
"grad_norm": 0.9634577235067475,
"learning_rate": 2.8066834351822255e-07,
"logits/chosen": -3.4804861545562744,
"logits/rejected": -3.5888350009918213,
"logps/chosen": -2.7011873722076416,
"logps/rejected": -75.55279541015625,
"loss": 0.1107,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.418036937713623,
"rewards/margins": 2.418036937713623,
"rewards/rejected": 0.0,
"step": 1948
},
{
"epoch": 10.888268156424582,
"grad_norm": 0.7512882568369851,
"learning_rate": 2.8012073851218486e-07,
"logits/chosen": -3.5435726642608643,
"logits/rejected": -3.4737308025360107,
"logps/chosen": -2.4472789764404297,
"logps/rejected": -44.22047805786133,
"loss": 0.0943,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8572998046875,
"rewards/margins": 2.8572998046875,
"rewards/rejected": 0.0,
"step": 1949
},
{
"epoch": 10.893854748603353,
"grad_norm": 0.5463667182865234,
"learning_rate": 2.7957346028588856e-07,
"logits/chosen": -3.351771831512451,
"logits/rejected": -3.396040916442871,
"logps/chosen": -24.655719757080078,
"logps/rejected": -61.6947021484375,
"loss": 0.0799,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0992283821105957,
"rewards/margins": 3.0992283821105957,
"rewards/rejected": 0.0,
"step": 1950
},
{
"epoch": 10.899441340782122,
"grad_norm": 0.9779824265490464,
"learning_rate": 2.790265096526857e-07,
"logits/chosen": -3.2073376178741455,
"logits/rejected": -3.3103933334350586,
"logps/chosen": -0.8609297871589661,
"logps/rejected": -63.67443084716797,
"loss": 0.1221,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9942567348480225,
"rewards/margins": 1.9942567348480225,
"rewards/rejected": 0.0,
"step": 1951
},
{
"epoch": 10.905027932960895,
"grad_norm": 3.878224074611816,
"learning_rate": 2.7847988742544295e-07,
"logits/chosen": -3.391176462173462,
"logits/rejected": -3.37831711769104,
"logps/chosen": -1.989864706993103,
"logps/rejected": -58.87008285522461,
"loss": 0.1191,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.503532886505127,
"rewards/margins": 2.503532886505127,
"rewards/rejected": 0.0,
"step": 1952
},
{
"epoch": 10.910614525139664,
"grad_norm": 1.061200239757773,
"learning_rate": 2.7793359441653805e-07,
"logits/chosen": -3.3318586349487305,
"logits/rejected": -3.3976187705993652,
"logps/chosen": -1.9514379501342773,
"logps/rejected": -41.914207458496094,
"loss": 0.1216,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.640911102294922,
"rewards/margins": 2.640911102294922,
"rewards/rejected": 0.0,
"step": 1953
},
{
"epoch": 10.916201117318435,
"grad_norm": 0.6202107189716216,
"learning_rate": 2.773876314378597e-07,
"logits/chosen": -3.4760282039642334,
"logits/rejected": -3.516247510910034,
"logps/chosen": -19.76568031311035,
"logps/rejected": -42.70884704589844,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.23059606552124,
"rewards/margins": 4.23059606552124,
"rewards/rejected": 0.0,
"step": 1954
},
{
"epoch": 10.921787709497206,
"grad_norm": 0.6151641360085743,
"learning_rate": 2.768419993008061e-07,
"logits/chosen": -3.4956982135772705,
"logits/rejected": -3.478393793106079,
"logps/chosen": -0.6943643689155579,
"logps/rejected": -74.71638488769531,
"loss": 0.124,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9081413745880127,
"rewards/margins": 1.9081413745880127,
"rewards/rejected": 0.0,
"step": 1955
},
{
"epoch": 10.927374301675977,
"grad_norm": 0.6040467280421871,
"learning_rate": 2.762966988162838e-07,
"logits/chosen": -3.4565107822418213,
"logits/rejected": -3.611046314239502,
"logps/chosen": -1.6663256883621216,
"logps/rejected": -27.423904418945312,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8244056701660156,
"rewards/margins": 2.8244056701660156,
"rewards/rejected": 0.0,
"step": 1956
},
{
"epoch": 10.932960893854748,
"grad_norm": 4.431823116814799,
"learning_rate": 2.757517307947065e-07,
"logits/chosen": -3.713186264038086,
"logits/rejected": -3.288609266281128,
"logps/chosen": -2.0880815982818604,
"logps/rejected": -53.6541633605957,
"loss": 0.103,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3980512619018555,
"rewards/margins": 2.3980512619018555,
"rewards/rejected": 0.0,
"step": 1957
},
{
"epoch": 10.938547486033519,
"grad_norm": 1.4618321912785606,
"learning_rate": 2.752070960459934e-07,
"logits/chosen": -3.5102028846740723,
"logits/rejected": -3.429625988006592,
"logps/chosen": -1.5148488283157349,
"logps/rejected": -29.034435272216797,
"loss": 0.0979,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9058281183242798,
"rewards/margins": 1.9058281183242798,
"rewards/rejected": 0.0,
"step": 1958
},
{
"epoch": 10.94413407821229,
"grad_norm": 1.0733951619882203,
"learning_rate": 2.746627953795694e-07,
"logits/chosen": -3.506091356277466,
"logits/rejected": -3.4129672050476074,
"logps/chosen": -3.1617989540100098,
"logps/rejected": -41.68645095825195,
"loss": 0.0754,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.317117214202881,
"rewards/margins": 3.317117214202881,
"rewards/rejected": 0.0,
"step": 1959
},
{
"epoch": 10.949720670391061,
"grad_norm": 0.8766754576708049,
"learning_rate": 2.741188296043615e-07,
"logits/chosen": -3.3641016483306885,
"logits/rejected": -3.35844087600708,
"logps/chosen": -3.9195914268493652,
"logps/rejected": -46.63465881347656,
"loss": 0.1093,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.267271041870117,
"rewards/margins": 2.267271041870117,
"rewards/rejected": 0.0,
"step": 1960
},
{
"epoch": 10.949720670391061,
"eval_logits/chosen": -3.294306516647339,
"eval_logits/rejected": -3.421966075897217,
"eval_logps/chosen": -28.862768173217773,
"eval_logps/rejected": -52.48956298828125,
"eval_loss": 0.9208307266235352,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.12072701752185822,
"eval_rewards/margins": 0.12072701752185822,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6643,
"eval_samples_per_second": 9.49,
"eval_steps_per_second": 0.306,
"step": 1960
},
{
"epoch": 10.955307262569832,
"grad_norm": 1.0814898778876383,
"learning_rate": 2.7357519952880023e-07,
"logits/chosen": -3.5372488498687744,
"logits/rejected": -3.697329044342041,
"logps/chosen": -12.813082695007324,
"logps/rejected": -74.08599853515625,
"loss": 0.0946,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.092062473297119,
"rewards/margins": 3.092062473297119,
"rewards/rejected": 0.0,
"step": 1961
},
{
"epoch": 10.960893854748603,
"grad_norm": 0.9277978725629857,
"learning_rate": 2.7303190596081645e-07,
"logits/chosen": -3.6069722175598145,
"logits/rejected": -3.2105302810668945,
"logps/chosen": -0.25405922532081604,
"logps/rejected": -116.97374725341797,
"loss": 0.1159,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.2921063899993896,
"rewards/margins": 1.2921063899993896,
"rewards/rejected": 0.0,
"step": 1962
},
{
"epoch": 10.966480446927374,
"grad_norm": 5.685547353250434,
"learning_rate": 2.7248894970784127e-07,
"logits/chosen": -3.3292641639709473,
"logits/rejected": -3.362541913986206,
"logps/chosen": -0.9024280309677124,
"logps/rejected": -110.84246826171875,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2061171531677246,
"rewards/margins": 2.2061171531677246,
"rewards/rejected": 0.0,
"step": 1963
},
{
"epoch": 10.972067039106145,
"grad_norm": 1.3361292153172168,
"learning_rate": 2.719463315768043e-07,
"logits/chosen": -3.4417128562927246,
"logits/rejected": -3.208587408065796,
"logps/chosen": -1.2682263851165771,
"logps/rejected": -44.37663269042969,
"loss": 0.0849,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.668027639389038,
"rewards/margins": 2.668027639389038,
"rewards/rejected": 0.0,
"step": 1964
},
{
"epoch": 10.977653631284916,
"grad_norm": 1.0865640081898422,
"learning_rate": 2.7140405237413275e-07,
"logits/chosen": -3.6909279823303223,
"logits/rejected": -3.6696698665618896,
"logps/chosen": -1.5324313640594482,
"logps/rejected": -29.734695434570312,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.615619659423828,
"rewards/margins": 2.615619659423828,
"rewards/rejected": 0.0,
"step": 1965
},
{
"epoch": 10.983240223463687,
"grad_norm": 0.5593576686892301,
"learning_rate": 2.708621129057501e-07,
"logits/chosen": -3.3819236755371094,
"logits/rejected": -3.462059497833252,
"logps/chosen": -14.919900894165039,
"logps/rejected": -67.44153594970703,
"loss": 0.0965,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.56858491897583,
"rewards/margins": 3.56858491897583,
"rewards/rejected": 0.0,
"step": 1966
},
{
"epoch": 10.988826815642458,
"grad_norm": 0.6010040579794459,
"learning_rate": 2.703205139770749e-07,
"logits/chosen": -3.402266263961792,
"logits/rejected": -3.3524739742279053,
"logps/chosen": -7.726075649261475,
"logps/rejected": -73.15519714355469,
"loss": 0.0894,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.823573589324951,
"rewards/margins": 2.823573589324951,
"rewards/rejected": 0.0,
"step": 1967
},
{
"epoch": 10.994413407821229,
"grad_norm": 0.5636415023647893,
"learning_rate": 2.6977925639301955e-07,
"logits/chosen": -3.5721802711486816,
"logits/rejected": -3.4628896713256836,
"logps/chosen": -1.6515507698059082,
"logps/rejected": -62.12843322753906,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9263627529144287,
"rewards/margins": 2.9263627529144287,
"rewards/rejected": 0.0,
"step": 1968
},
{
"epoch": 11.0,
"grad_norm": 0.6608148885043948,
"learning_rate": 2.6923834095798914e-07,
"logits/chosen": -3.3973584175109863,
"logits/rejected": -3.4416000843048096,
"logps/chosen": -3.7127230167388916,
"logps/rejected": -43.00478744506836,
"loss": 0.1061,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.04290771484375,
"rewards/margins": 2.04290771484375,
"rewards/rejected": 0.0,
"step": 1969
},
{
"epoch": 11.005586592178771,
"grad_norm": 0.7570754454846276,
"learning_rate": 2.6869776847588076e-07,
"logits/chosen": -3.223130464553833,
"logits/rejected": -3.216498613357544,
"logps/chosen": -4.856444835662842,
"logps/rejected": -37.17080307006836,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8240160942077637,
"rewards/margins": 2.8240160942077637,
"rewards/rejected": 0.0,
"step": 1970
},
{
"epoch": 11.011173184357542,
"grad_norm": 0.7587609785066302,
"learning_rate": 2.681575397500808e-07,
"logits/chosen": -3.4802913665771484,
"logits/rejected": -3.5363049507141113,
"logps/chosen": -12.489826202392578,
"logps/rejected": -59.16836166381836,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.173032760620117,
"rewards/margins": 3.173032760620117,
"rewards/rejected": 0.0,
"step": 1971
},
{
"epoch": 11.016759776536313,
"grad_norm": 0.5487584039154622,
"learning_rate": 2.676176555834658e-07,
"logits/chosen": -3.7237424850463867,
"logits/rejected": -3.539865493774414,
"logps/chosen": -7.172542572021484,
"logps/rejected": -32.795982360839844,
"loss": 0.0883,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.917008876800537,
"rewards/margins": 3.917008876800537,
"rewards/rejected": 0.0,
"step": 1972
},
{
"epoch": 11.022346368715084,
"grad_norm": 1.3067089064331106,
"learning_rate": 2.6707811677839975e-07,
"logits/chosen": -3.0273399353027344,
"logits/rejected": -3.053793430328369,
"logps/chosen": -1.631102204322815,
"logps/rejected": -64.28227996826172,
"loss": 0.1006,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1646173000335693,
"rewards/margins": 2.1646173000335693,
"rewards/rejected": 0.0,
"step": 1973
},
{
"epoch": 11.027932960893855,
"grad_norm": 0.45809697672224176,
"learning_rate": 2.6653892413673316e-07,
"logits/chosen": -3.6722640991210938,
"logits/rejected": -3.7272047996520996,
"logps/chosen": -4.221530914306641,
"logps/rejected": -37.29861068725586,
"loss": 0.0878,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3270933628082275,
"rewards/margins": 3.3270933628082275,
"rewards/rejected": 0.0,
"step": 1974
},
{
"epoch": 11.033519553072626,
"grad_norm": 0.4210681518111144,
"learning_rate": 2.6600007845980256e-07,
"logits/chosen": -3.356821298599243,
"logits/rejected": -3.310434341430664,
"logps/chosen": -25.088289260864258,
"logps/rejected": -42.456703186035156,
"loss": 0.0824,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.16225528717041,
"rewards/margins": 3.16225528717041,
"rewards/rejected": 0.0,
"step": 1975
},
{
"epoch": 11.039106145251397,
"grad_norm": 1.9968140360414173,
"learning_rate": 2.654615805484284e-07,
"logits/chosen": -3.5969486236572266,
"logits/rejected": -3.580840587615967,
"logps/chosen": -12.102442741394043,
"logps/rejected": -32.60572814941406,
"loss": 0.0859,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4988389015197754,
"rewards/margins": 3.4988389015197754,
"rewards/rejected": 0.0,
"step": 1976
},
{
"epoch": 11.044692737430168,
"grad_norm": 0.4183318629887757,
"learning_rate": 2.649234312029145e-07,
"logits/chosen": -3.3813321590423584,
"logits/rejected": -3.486619234085083,
"logps/chosen": -3.1055541038513184,
"logps/rejected": -53.136512756347656,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6955175399780273,
"rewards/margins": 2.6955175399780273,
"rewards/rejected": 0.0,
"step": 1977
},
{
"epoch": 11.050279329608939,
"grad_norm": 0.4650920392732018,
"learning_rate": 2.6438563122304657e-07,
"logits/chosen": -3.3683040142059326,
"logits/rejected": -3.384139060974121,
"logps/chosen": -9.655780792236328,
"logps/rejected": -37.85082244873047,
"loss": 0.0789,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1659979820251465,
"rewards/margins": 3.1659979820251465,
"rewards/rejected": 0.0,
"step": 1978
},
{
"epoch": 11.05586592178771,
"grad_norm": 0.6060572375804102,
"learning_rate": 2.638481814080916e-07,
"logits/chosen": -3.5283408164978027,
"logits/rejected": -3.427849292755127,
"logps/chosen": -1.1676838397979736,
"logps/rejected": -48.28765106201172,
"loss": 0.1021,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4302749633789062,
"rewards/margins": 2.4302749633789062,
"rewards/rejected": 0.0,
"step": 1979
},
{
"epoch": 11.061452513966481,
"grad_norm": 0.4322984282245594,
"learning_rate": 2.633110825567951e-07,
"logits/chosen": -3.6758594512939453,
"logits/rejected": -3.6464524269104004,
"logps/chosen": -7.674620151519775,
"logps/rejected": -60.0107421875,
"loss": 0.133,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.139512300491333,
"rewards/margins": 3.139512300491333,
"rewards/rejected": 0.0,
"step": 1980
},
{
"epoch": 11.061452513966481,
"eval_logits/chosen": -3.2864978313446045,
"eval_logits/rejected": -3.4144539833068848,
"eval_logps/chosen": -29.15273666381836,
"eval_logps/rejected": -53.541419982910156,
"eval_loss": 0.9443025588989258,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.09173060208559036,
"eval_rewards/margins": 0.09173060208559036,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6928,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 1980
},
{
"epoch": 11.067039106145252,
"grad_norm": 0.4754917243396108,
"learning_rate": 2.6277433546738205e-07,
"logits/chosen": -3.230081558227539,
"logits/rejected": -3.004481554031372,
"logps/chosen": -0.15751391649246216,
"logps/rejected": -70.2298583984375,
"loss": 0.1023,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.1462054252624512,
"rewards/margins": 1.1462054252624512,
"rewards/rejected": 0.0,
"step": 1981
},
{
"epoch": 11.072625698324023,
"grad_norm": 0.48976582866360574,
"learning_rate": 2.62237940937554e-07,
"logits/chosen": -3.5290935039520264,
"logits/rejected": -3.5418808460235596,
"logps/chosen": -4.300559043884277,
"logps/rejected": -35.070491790771484,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1748969554901123,
"rewards/margins": 3.1748969554901123,
"rewards/rejected": 0.0,
"step": 1982
},
{
"epoch": 11.078212290502794,
"grad_norm": 0.4792929877417928,
"learning_rate": 2.617018997644889e-07,
"logits/chosen": -3.114351511001587,
"logits/rejected": -3.112081527709961,
"logps/chosen": -10.51168155670166,
"logps/rejected": -52.60894775390625,
"loss": 0.1098,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.372950792312622,
"rewards/margins": 3.372950792312622,
"rewards/rejected": 0.0,
"step": 1983
},
{
"epoch": 11.083798882681565,
"grad_norm": 0.6160879245652107,
"learning_rate": 2.611662127448393e-07,
"logits/chosen": -3.4955079555511475,
"logits/rejected": -3.5562283992767334,
"logps/chosen": -15.152576446533203,
"logps/rejected": -68.93836975097656,
"loss": 0.1133,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.657057523727417,
"rewards/margins": 2.657057523727417,
"rewards/rejected": 0.0,
"step": 1984
},
{
"epoch": 11.089385474860336,
"grad_norm": 0.41305624555883874,
"learning_rate": 2.6063088067473156e-07,
"logits/chosen": -3.2018439769744873,
"logits/rejected": -3.2682015895843506,
"logps/chosen": -8.8161039352417,
"logps/rejected": -29.48605728149414,
"loss": 0.1102,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0993926525115967,
"rewards/margins": 3.0993926525115967,
"rewards/rejected": 0.0,
"step": 1985
},
{
"epoch": 11.094972067039107,
"grad_norm": 0.5248289945571151,
"learning_rate": 2.6009590434976444e-07,
"logits/chosen": -3.2796809673309326,
"logits/rejected": -3.429037094116211,
"logps/chosen": -4.598927974700928,
"logps/rejected": -62.207794189453125,
"loss": 0.088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3135581016540527,
"rewards/margins": 2.3135581016540527,
"rewards/rejected": 0.0,
"step": 1986
},
{
"epoch": 11.100558659217878,
"grad_norm": 0.46065714416705367,
"learning_rate": 2.595612845650078e-07,
"logits/chosen": -3.3993310928344727,
"logits/rejected": -3.231445074081421,
"logps/chosen": -7.13292121887207,
"logps/rejected": -73.23664855957031,
"loss": 0.0851,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9397664070129395,
"rewards/margins": 2.9397664070129395,
"rewards/rejected": 0.0,
"step": 1987
},
{
"epoch": 11.106145251396647,
"grad_norm": 0.4565078942258675,
"learning_rate": 2.5902702211500246e-07,
"logits/chosen": -3.4503095149993896,
"logits/rejected": -3.3747165203094482,
"logps/chosen": -3.5082180500030518,
"logps/rejected": -67.02780151367188,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.288248062133789,
"rewards/margins": 2.288248062133789,
"rewards/rejected": 0.0,
"step": 1988
},
{
"epoch": 11.111731843575418,
"grad_norm": 0.4719089458995008,
"learning_rate": 2.584931177937569e-07,
"logits/chosen": -3.2662644386291504,
"logits/rejected": -3.3980376720428467,
"logps/chosen": -1.2309666872024536,
"logps/rejected": -61.435935974121094,
"loss": 0.0827,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.695246934890747,
"rewards/margins": 2.695246934890747,
"rewards/rejected": 0.0,
"step": 1989
},
{
"epoch": 11.11731843575419,
"grad_norm": 0.6205757416787661,
"learning_rate": 2.5795957239474853e-07,
"logits/chosen": -3.261859893798828,
"logits/rejected": -3.39766526222229,
"logps/chosen": -0.8327498435974121,
"logps/rejected": -61.17940139770508,
"loss": 0.1069,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.063709020614624,
"rewards/margins": 2.063709020614624,
"rewards/rejected": 0.0,
"step": 1990
},
{
"epoch": 11.12290502793296,
"grad_norm": 1.8285638134381723,
"learning_rate": 2.5742638671092054e-07,
"logits/chosen": -3.409914016723633,
"logits/rejected": -3.5615062713623047,
"logps/chosen": -4.146988868713379,
"logps/rejected": -44.768638610839844,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5233983993530273,
"rewards/margins": 3.5233983993530273,
"rewards/rejected": 0.0,
"step": 1991
},
{
"epoch": 11.128491620111731,
"grad_norm": 0.5714640768278682,
"learning_rate": 2.568935615346819e-07,
"logits/chosen": -3.1631546020507812,
"logits/rejected": -3.1505377292633057,
"logps/chosen": -4.494215488433838,
"logps/rejected": -57.05901336669922,
"loss": 0.0826,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5487918853759766,
"rewards/margins": 3.5487918853759766,
"rewards/rejected": 0.0,
"step": 1992
},
{
"epoch": 11.134078212290502,
"grad_norm": 2.102452150592736,
"learning_rate": 2.563610976579057e-07,
"logits/chosen": -3.577125072479248,
"logits/rejected": -3.68088698387146,
"logps/chosen": -0.7124360799789429,
"logps/rejected": -57.468414306640625,
"loss": 0.1291,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1272451877593994,
"rewards/margins": 2.1272451877593994,
"rewards/rejected": 0.0,
"step": 1993
},
{
"epoch": 11.139664804469273,
"grad_norm": 1.9838899032802388,
"learning_rate": 2.558289958719282e-07,
"logits/chosen": -3.1666100025177,
"logits/rejected": -3.02303409576416,
"logps/chosen": -2.4945926666259766,
"logps/rejected": -40.52445983886719,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.230691909790039,
"rewards/margins": 3.230691909790039,
"rewards/rejected": 0.0,
"step": 1994
},
{
"epoch": 11.145251396648044,
"grad_norm": 0.4207349330247922,
"learning_rate": 2.552972569675472e-07,
"logits/chosen": -3.5099363327026367,
"logits/rejected": -3.418095350265503,
"logps/chosen": -17.87454605102539,
"logps/rejected": -29.387659072875977,
"loss": 0.0829,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9897732734680176,
"rewards/margins": 3.9897732734680176,
"rewards/rejected": 0.0,
"step": 1995
},
{
"epoch": 11.150837988826815,
"grad_norm": 1.3450358018556177,
"learning_rate": 2.5476588173502156e-07,
"logits/chosen": -3.435154438018799,
"logits/rejected": -3.3458993434906006,
"logps/chosen": -9.573295593261719,
"logps/rejected": -66.47622680664062,
"loss": 0.0837,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4141845703125,
"rewards/margins": 3.4141845703125,
"rewards/rejected": 0.0,
"step": 1996
},
{
"epoch": 11.156424581005586,
"grad_norm": 2.1526471206314595,
"learning_rate": 2.542348709640695e-07,
"logits/chosen": -3.580726385116577,
"logits/rejected": -3.455291748046875,
"logps/chosen": -1.1961976289749146,
"logps/rejected": -44.997676849365234,
"loss": 0.0969,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.362941026687622,
"rewards/margins": 2.362941026687622,
"rewards/rejected": 0.0,
"step": 1997
},
{
"epoch": 11.162011173184357,
"grad_norm": 0.5094837035971311,
"learning_rate": 2.5370422544386726e-07,
"logits/chosen": -3.2765495777130127,
"logits/rejected": -3.47189998626709,
"logps/chosen": -0.45980337262153625,
"logps/rejected": -67.9575424194336,
"loss": 0.0942,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.857908010482788,
"rewards/margins": 1.857908010482788,
"rewards/rejected": 0.0,
"step": 1998
},
{
"epoch": 11.167597765363128,
"grad_norm": 0.6650143731143028,
"learning_rate": 2.531739459630494e-07,
"logits/chosen": -3.3117434978485107,
"logits/rejected": -3.202538013458252,
"logps/chosen": -5.910820484161377,
"logps/rejected": -44.21772003173828,
"loss": 0.077,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5671982765197754,
"rewards/margins": 3.5671982765197754,
"rewards/rejected": 0.0,
"step": 1999
},
{
"epoch": 11.1731843575419,
"grad_norm": 1.0015887536491135,
"learning_rate": 2.5264403330970466e-07,
"logits/chosen": -3.384261131286621,
"logits/rejected": -3.4819624423980713,
"logps/chosen": -1.515886902809143,
"logps/rejected": -64.92242431640625,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.557046890258789,
"rewards/margins": 2.557046890258789,
"rewards/rejected": 0.0,
"step": 2000
},
{
"epoch": 11.1731843575419,
"eval_logits/chosen": -3.290687084197998,
"eval_logits/rejected": -3.415971279144287,
"eval_logps/chosen": -29.300617218017578,
"eval_logps/rejected": -53.619422912597656,
"eval_loss": 0.9535685777664185,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.07694248855113983,
"eval_rewards/margins": 0.07694248855113983,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6671,
"eval_samples_per_second": 9.49,
"eval_steps_per_second": 0.306,
"step": 2000
},
{
"epoch": 11.17877094972067,
"grad_norm": 3.398848744419848,
"learning_rate": 2.5211448827137837e-07,
"logits/chosen": -3.575794219970703,
"logits/rejected": -3.443030834197998,
"logps/chosen": -1.1924731731414795,
"logps/rejected": -48.99359130859375,
"loss": 0.1191,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.109663486480713,
"rewards/margins": 2.109663486480713,
"rewards/rejected": 0.0,
"step": 2001
},
{
"epoch": 11.184357541899441,
"grad_norm": 0.4464897914448476,
"learning_rate": 2.5158531163506854e-07,
"logits/chosen": -3.6060190200805664,
"logits/rejected": -3.4800190925598145,
"logps/chosen": -0.4656723141670227,
"logps/rejected": -80.21681213378906,
"loss": 0.1018,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1573047637939453,
"rewards/margins": 2.1573047637939453,
"rewards/rejected": 0.0,
"step": 2002
},
{
"epoch": 11.189944134078212,
"grad_norm": 0.5631812622021007,
"learning_rate": 2.510565041872258e-07,
"logits/chosen": -3.494626760482788,
"logits/rejected": -3.3245034217834473,
"logps/chosen": -7.909552574157715,
"logps/rejected": -34.34803009033203,
"loss": 0.1069,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4123058319091797,
"rewards/margins": 3.4123058319091797,
"rewards/rejected": 0.0,
"step": 2003
},
{
"epoch": 11.195530726256983,
"grad_norm": 1.1083021166285107,
"learning_rate": 2.5052806671375224e-07,
"logits/chosen": -3.471977949142456,
"logits/rejected": -3.5947768688201904,
"logps/chosen": -0.12925735116004944,
"logps/rejected": -33.51356506347656,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4285383224487305,
"rewards/margins": 1.4285383224487305,
"rewards/rejected": 0.0,
"step": 2004
},
{
"epoch": 11.201117318435754,
"grad_norm": 0.5050854730725913,
"learning_rate": 2.500000000000001e-07,
"logits/chosen": -3.2237627506256104,
"logits/rejected": -3.2795255184173584,
"logps/chosen": -1.5750045776367188,
"logps/rejected": -49.049198150634766,
"loss": 0.0981,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1445679664611816,
"rewards/margins": 2.1445679664611816,
"rewards/rejected": 0.0,
"step": 2005
},
{
"epoch": 11.206703910614525,
"grad_norm": 0.7780193463055313,
"learning_rate": 2.4947230483077046e-07,
"logits/chosen": -3.7185072898864746,
"logits/rejected": -3.555849552154541,
"logps/chosen": -2.027606725692749,
"logps/rejected": -31.118919372558594,
"loss": 0.106,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.578911304473877,
"rewards/margins": 2.578911304473877,
"rewards/rejected": 0.0,
"step": 2006
},
{
"epoch": 11.212290502793296,
"grad_norm": 0.7515147029504262,
"learning_rate": 2.489449819903123e-07,
"logits/chosen": -3.58176326751709,
"logits/rejected": -3.5337986946105957,
"logps/chosen": -6.2703022956848145,
"logps/rejected": -36.084991455078125,
"loss": 0.0806,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4396705627441406,
"rewards/margins": 3.4396705627441406,
"rewards/rejected": 0.0,
"step": 2007
},
{
"epoch": 11.217877094972067,
"grad_norm": 0.6890743153357715,
"learning_rate": 2.4841803226232166e-07,
"logits/chosen": -3.3658447265625,
"logits/rejected": -3.2889177799224854,
"logps/chosen": -1.3058695793151855,
"logps/rejected": -74.79275512695312,
"loss": 0.1306,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2486634254455566,
"rewards/margins": 2.2486634254455566,
"rewards/rejected": 0.0,
"step": 2008
},
{
"epoch": 11.223463687150838,
"grad_norm": 0.43004423077332365,
"learning_rate": 2.47891456429939e-07,
"logits/chosen": -3.491614818572998,
"logits/rejected": -3.4320404529571533,
"logps/chosen": -0.6702783107757568,
"logps/rejected": -48.801658630371094,
"loss": 0.0873,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3724660873413086,
"rewards/margins": 2.3724660873413086,
"rewards/rejected": 0.0,
"step": 2009
},
{
"epoch": 11.22905027932961,
"grad_norm": 0.5405604310322775,
"learning_rate": 2.473652552757504e-07,
"logits/chosen": -3.5797393321990967,
"logits/rejected": -3.596543550491333,
"logps/chosen": -1.1575409173965454,
"logps/rejected": -67.98393249511719,
"loss": 0.1069,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9111378192901611,
"rewards/margins": 1.9111378192901611,
"rewards/rejected": 0.0,
"step": 2010
},
{
"epoch": 11.23463687150838,
"grad_norm": 1.6557292561068724,
"learning_rate": 2.4683942958178416e-07,
"logits/chosen": -3.2692625522613525,
"logits/rejected": -3.3795197010040283,
"logps/chosen": -2.376598358154297,
"logps/rejected": -32.81221008300781,
"loss": 0.0886,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5531296730041504,
"rewards/margins": 2.5531296730041504,
"rewards/rejected": 0.0,
"step": 2011
},
{
"epoch": 11.240223463687151,
"grad_norm": 0.45609308312774133,
"learning_rate": 2.4631398012951104e-07,
"logits/chosen": -3.3126449584960938,
"logits/rejected": -3.3992831707000732,
"logps/chosen": -14.70008659362793,
"logps/rejected": -40.075138092041016,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3907418251037598,
"rewards/margins": 3.3907418251037598,
"rewards/rejected": 0.0,
"step": 2012
},
{
"epoch": 11.245810055865922,
"grad_norm": 0.5591609311118848,
"learning_rate": 2.4578890769984227e-07,
"logits/chosen": -3.0882582664489746,
"logits/rejected": -3.1441426277160645,
"logps/chosen": -3.592134475708008,
"logps/rejected": -38.634803771972656,
"loss": 0.0891,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2177412509918213,
"rewards/margins": 3.2177412509918213,
"rewards/rejected": 0.0,
"step": 2013
},
{
"epoch": 11.251396648044693,
"grad_norm": 0.44512968524400404,
"learning_rate": 2.4526421307312954e-07,
"logits/chosen": -3.2984673976898193,
"logits/rejected": -3.168893337249756,
"logps/chosen": -8.026363372802734,
"logps/rejected": -43.317626953125,
"loss": 0.1167,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.16433048248291,
"rewards/margins": 3.16433048248291,
"rewards/rejected": 0.0,
"step": 2014
},
{
"epoch": 11.256983240223464,
"grad_norm": 0.5264800157859246,
"learning_rate": 2.4473989702916197e-07,
"logits/chosen": -3.5242347717285156,
"logits/rejected": -3.5590171813964844,
"logps/chosen": -1.7915736436843872,
"logps/rejected": -26.157859802246094,
"loss": 0.117,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.661710739135742,
"rewards/margins": 2.661710739135742,
"rewards/rejected": 0.0,
"step": 2015
},
{
"epoch": 11.262569832402235,
"grad_norm": 0.48931397758179795,
"learning_rate": 2.4421596034716657e-07,
"logits/chosen": -3.1743080615997314,
"logits/rejected": -3.1876299381256104,
"logps/chosen": -7.518608093261719,
"logps/rejected": -57.348182678222656,
"loss": 0.0849,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2221872806549072,
"rewards/margins": 3.2221872806549072,
"rewards/rejected": 0.0,
"step": 2016
},
{
"epoch": 11.268156424581006,
"grad_norm": 0.47034890465607826,
"learning_rate": 2.4369240380580713e-07,
"logits/chosen": -3.6134889125823975,
"logits/rejected": -3.5242602825164795,
"logps/chosen": -1.8971567153930664,
"logps/rejected": -62.129615783691406,
"loss": 0.0792,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5112192630767822,
"rewards/margins": 2.5112192630767822,
"rewards/rejected": 0.0,
"step": 2017
},
{
"epoch": 11.273743016759777,
"grad_norm": 2.982781794529716,
"learning_rate": 2.431692281831811e-07,
"logits/chosen": -3.328136682510376,
"logits/rejected": -3.329343795776367,
"logps/chosen": -3.3141212463378906,
"logps/rejected": -64.44425201416016,
"loss": 0.1221,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9037461280822754,
"rewards/margins": 2.9037461280822754,
"rewards/rejected": 0.0,
"step": 2018
},
{
"epoch": 11.279329608938548,
"grad_norm": 0.880899449492303,
"learning_rate": 2.426464342568213e-07,
"logits/chosen": -3.4399499893188477,
"logits/rejected": -3.456308603286743,
"logps/chosen": -0.7736425995826721,
"logps/rejected": -87.82368469238281,
"loss": 0.1031,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.341750383377075,
"rewards/margins": 2.341750383377075,
"rewards/rejected": 0.0,
"step": 2019
},
{
"epoch": 11.28491620111732,
"grad_norm": 0.4050577518047395,
"learning_rate": 2.421240228036923e-07,
"logits/chosen": -3.4010915756225586,
"logits/rejected": -3.3362855911254883,
"logps/chosen": -7.492812633514404,
"logps/rejected": -59.437522888183594,
"loss": 0.0795,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1338376998901367,
"rewards/margins": 3.1338376998901367,
"rewards/rejected": 0.0,
"step": 2020
},
{
"epoch": 11.28491620111732,
"eval_logits/chosen": -3.2706871032714844,
"eval_logits/rejected": -3.3993258476257324,
"eval_logps/chosen": -29.3426456451416,
"eval_logps/rejected": -53.73309326171875,
"eval_loss": 0.9600690007209778,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": 0.07273955643177032,
"eval_rewards/margins": 0.07273955643177032,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6658,
"eval_samples_per_second": 9.49,
"eval_steps_per_second": 0.306,
"step": 2020
},
{
"epoch": 11.29050279329609,
"grad_norm": 0.4855923805945946,
"learning_rate": 2.4160199460019063e-07,
"logits/chosen": -3.0380606651306152,
"logits/rejected": -3.221691131591797,
"logps/chosen": -0.1503109633922577,
"logps/rejected": -49.60858917236328,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4069033861160278,
"rewards/margins": 1.4069033861160278,
"rewards/rejected": 0.0,
"step": 2021
},
{
"epoch": 11.296089385474861,
"grad_norm": 0.4525412670100124,
"learning_rate": 2.4108035042214315e-07,
"logits/chosen": -3.411052703857422,
"logits/rejected": -3.3881309032440186,
"logps/chosen": -7.243567943572998,
"logps/rejected": -83.95309448242188,
"loss": 0.1271,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.962278127670288,
"rewards/margins": 2.962278127670288,
"rewards/rejected": 0.0,
"step": 2022
},
{
"epoch": 11.30167597765363,
"grad_norm": 0.4495443200088319,
"learning_rate": 2.40559091044806e-07,
"logits/chosen": -3.3428659439086914,
"logits/rejected": -3.438525915145874,
"logps/chosen": -2.012706995010376,
"logps/rejected": -36.666404724121094,
"loss": 0.1004,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.595395565032959,
"rewards/margins": 2.595395565032959,
"rewards/rejected": 0.0,
"step": 2023
},
{
"epoch": 11.307262569832401,
"grad_norm": 0.4579302566909401,
"learning_rate": 2.4003821724286347e-07,
"logits/chosen": -3.5048301219940186,
"logits/rejected": -3.567451000213623,
"logps/chosen": -4.573941230773926,
"logps/rejected": -49.41082000732422,
"loss": 0.0851,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8217461109161377,
"rewards/margins": 2.8217461109161377,
"rewards/rejected": 0.0,
"step": 2024
},
{
"epoch": 11.312849162011172,
"grad_norm": 0.9202875260967602,
"learning_rate": 2.395177297904268e-07,
"logits/chosen": -3.6099133491516113,
"logits/rejected": -3.168043851852417,
"logps/chosen": -0.5238317251205444,
"logps/rejected": -100.91718292236328,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2232165336608887,
"rewards/margins": 2.2232165336608887,
"rewards/rejected": 0.0,
"step": 2025
},
{
"epoch": 11.318435754189943,
"grad_norm": 0.4451221198149984,
"learning_rate": 2.38997629461033e-07,
"logits/chosen": -3.102287769317627,
"logits/rejected": -3.227853775024414,
"logps/chosen": -20.150047302246094,
"logps/rejected": -40.576385498046875,
"loss": 0.07,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.901444911956787,
"rewards/margins": 3.901444911956787,
"rewards/rejected": 0.0,
"step": 2026
},
{
"epoch": 11.324022346368714,
"grad_norm": 0.4696703530585902,
"learning_rate": 2.3847791702764357e-07,
"logits/chosen": -3.080986261367798,
"logits/rejected": -3.215866804122925,
"logps/chosen": -0.3314264714717865,
"logps/rejected": -102.53214263916016,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6432881355285645,
"rewards/margins": 1.6432881355285645,
"rewards/rejected": 0.0,
"step": 2027
},
{
"epoch": 11.329608938547485,
"grad_norm": 0.5094241705056239,
"learning_rate": 2.379585932626444e-07,
"logits/chosen": -3.2015206813812256,
"logits/rejected": -3.4322457313537598,
"logps/chosen": -25.20183753967285,
"logps/rejected": -53.60784912109375,
"loss": 0.0783,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.472428321838379,
"rewards/margins": 3.472428321838379,
"rewards/rejected": 0.0,
"step": 2028
},
{
"epoch": 11.335195530726256,
"grad_norm": 0.4306664131712824,
"learning_rate": 2.3743965893784223e-07,
"logits/chosen": -3.070831775665283,
"logits/rejected": -3.1770286560058594,
"logps/chosen": -29.29344940185547,
"logps/rejected": -38.18990707397461,
"loss": 0.076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9771416187286377,
"rewards/margins": 3.9771416187286377,
"rewards/rejected": 0.0,
"step": 2029
},
{
"epoch": 11.340782122905027,
"grad_norm": 0.714306424060186,
"learning_rate": 2.3692111482446658e-07,
"logits/chosen": -3.3322372436523438,
"logits/rejected": -3.544323444366455,
"logps/chosen": -1.9236233234405518,
"logps/rejected": -33.58332061767578,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7325119972229004,
"rewards/margins": 2.7325119972229004,
"rewards/rejected": 0.0,
"step": 2030
},
{
"epoch": 11.346368715083798,
"grad_norm": 0.773819892480267,
"learning_rate": 2.3640296169316604e-07,
"logits/chosen": -3.1517975330352783,
"logits/rejected": -3.1840906143188477,
"logps/chosen": -4.88399076461792,
"logps/rejected": -32.34602355957031,
"loss": 0.1014,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.597146511077881,
"rewards/margins": 2.597146511077881,
"rewards/rejected": 0.0,
"step": 2031
},
{
"epoch": 11.35195530726257,
"grad_norm": 1.0386674687752542,
"learning_rate": 2.358852003140085e-07,
"logits/chosen": -3.435654401779175,
"logits/rejected": -3.552208185195923,
"logps/chosen": -35.12843322753906,
"logps/rejected": -41.45343017578125,
"loss": 0.0595,
"rewards/accuracies": 1.0,
"rewards/chosen": 5.088462829589844,
"rewards/margins": 5.088462829589844,
"rewards/rejected": 0.0,
"step": 2032
},
{
"epoch": 11.35754189944134,
"grad_norm": 0.6542811410942738,
"learning_rate": 2.3536783145647936e-07,
"logits/chosen": -3.254070520401001,
"logits/rejected": -3.3398876190185547,
"logps/chosen": -1.2409307956695557,
"logps/rejected": -28.919286727905273,
"loss": 0.1184,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3911044597625732,
"rewards/margins": 2.3911044597625732,
"rewards/rejected": 0.0,
"step": 2033
},
{
"epoch": 11.363128491620111,
"grad_norm": 0.730378674888996,
"learning_rate": 2.348508558894815e-07,
"logits/chosen": -3.287074089050293,
"logits/rejected": -3.207710027694702,
"logps/chosen": -1.7696616649627686,
"logps/rejected": -36.400367736816406,
"loss": 0.0801,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2627506256103516,
"rewards/margins": 2.2627506256103516,
"rewards/rejected": 0.0,
"step": 2034
},
{
"epoch": 11.368715083798882,
"grad_norm": 0.42622717376859515,
"learning_rate": 2.3433427438133206e-07,
"logits/chosen": -3.5788822174072266,
"logits/rejected": -3.491795778274536,
"logps/chosen": -1.0302319526672363,
"logps/rejected": -40.77783203125,
"loss": 0.093,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.491347312927246,
"rewards/margins": 2.491347312927246,
"rewards/rejected": 0.0,
"step": 2035
},
{
"epoch": 11.374301675977653,
"grad_norm": 0.49680662202885134,
"learning_rate": 2.338180876997632e-07,
"logits/chosen": -3.430384874343872,
"logits/rejected": -3.5522563457489014,
"logps/chosen": -2.5603208541870117,
"logps/rejected": -56.43990707397461,
"loss": 0.0887,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1166903972625732,
"rewards/margins": 3.1166903972625732,
"rewards/rejected": 0.0,
"step": 2036
},
{
"epoch": 11.379888268156424,
"grad_norm": 0.4629377401366435,
"learning_rate": 2.3330229661192074e-07,
"logits/chosen": -3.4896841049194336,
"logits/rejected": -3.30959153175354,
"logps/chosen": -1.999104619026184,
"logps/rejected": -45.836875915527344,
"loss": 0.0841,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.052429676055908,
"rewards/margins": 3.052429676055908,
"rewards/rejected": 0.0,
"step": 2037
},
{
"epoch": 11.385474860335195,
"grad_norm": 0.4416742911024336,
"learning_rate": 2.3278690188436144e-07,
"logits/chosen": -3.3385913372039795,
"logits/rejected": -3.5497634410858154,
"logps/chosen": -0.18844816088676453,
"logps/rejected": -62.22562789916992,
"loss": 0.1158,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5747473239898682,
"rewards/margins": 1.5747473239898682,
"rewards/rejected": 0.0,
"step": 2038
},
{
"epoch": 11.391061452513966,
"grad_norm": 0.6254421054874332,
"learning_rate": 2.3227190428305422e-07,
"logits/chosen": -3.536216974258423,
"logits/rejected": -3.433133363723755,
"logps/chosen": -7.839025974273682,
"logps/rejected": -46.35868453979492,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7869088649749756,
"rewards/margins": 3.7869088649749756,
"rewards/rejected": 0.0,
"step": 2039
},
{
"epoch": 11.396648044692737,
"grad_norm": 1.4167930649729465,
"learning_rate": 2.3175730457337695e-07,
"logits/chosen": -3.319622278213501,
"logits/rejected": -2.999826669692993,
"logps/chosen": -15.87757682800293,
"logps/rejected": -42.64402770996094,
"loss": 0.1097,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7572197914123535,
"rewards/margins": 2.7572197914123535,
"rewards/rejected": 0.0,
"step": 2040
},
{
"epoch": 11.396648044692737,
"eval_logits/chosen": -3.2879624366760254,
"eval_logits/rejected": -3.4116337299346924,
"eval_logps/chosen": -29.63619613647461,
"eval_logps/rejected": -53.59400177001953,
"eval_loss": 0.9723474383354187,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.043384671211242676,
"eval_rewards/margins": 0.043384671211242676,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6912,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 2040
},
{
"epoch": 11.402234636871508,
"grad_norm": 0.4308853248830393,
"learning_rate": 2.312431035201165e-07,
"logits/chosen": -3.4698472023010254,
"logits/rejected": -3.575444221496582,
"logps/chosen": -23.711624145507812,
"logps/rejected": -43.78266143798828,
"loss": 0.0953,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.2616167068481445,
"rewards/margins": 4.2616167068481445,
"rewards/rejected": 0.0,
"step": 2041
},
{
"epoch": 11.40782122905028,
"grad_norm": 1.3220967883114643,
"learning_rate": 2.3072930188746697e-07,
"logits/chosen": -3.6334421634674072,
"logits/rejected": -3.668956995010376,
"logps/chosen": -16.505786895751953,
"logps/rejected": -42.4336051940918,
"loss": 0.1357,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0035178661346436,
"rewards/margins": 3.0035178661346436,
"rewards/rejected": 0.0,
"step": 2042
},
{
"epoch": 11.41340782122905,
"grad_norm": 1.6326144170751449,
"learning_rate": 2.3021590043902979e-07,
"logits/chosen": -3.575676918029785,
"logits/rejected": -3.6062214374542236,
"logps/chosen": -2.799795150756836,
"logps/rejected": -36.262168884277344,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2209248542785645,
"rewards/margins": 3.2209248542785645,
"rewards/rejected": 0.0,
"step": 2043
},
{
"epoch": 11.418994413407821,
"grad_norm": 0.8682698856020982,
"learning_rate": 2.2970289993781022e-07,
"logits/chosen": -3.4891762733459473,
"logits/rejected": -3.435276746749878,
"logps/chosen": -27.726642608642578,
"logps/rejected": -50.93583679199219,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8662686347961426,
"rewards/margins": 2.8662686347961426,
"rewards/rejected": 0.0,
"step": 2044
},
{
"epoch": 11.424581005586592,
"grad_norm": 0.3732082545979339,
"learning_rate": 2.2919030114621846e-07,
"logits/chosen": -3.3657753467559814,
"logits/rejected": -3.402918577194214,
"logps/chosen": -6.243122577667236,
"logps/rejected": -55.91102600097656,
"loss": 0.0984,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.098987340927124,
"rewards/margins": 3.098987340927124,
"rewards/rejected": 0.0,
"step": 2045
},
{
"epoch": 11.430167597765363,
"grad_norm": 0.7583505967837766,
"learning_rate": 2.2867810482606815e-07,
"logits/chosen": -3.4816181659698486,
"logits/rejected": -3.5671916007995605,
"logps/chosen": -7.070084571838379,
"logps/rejected": -71.01090240478516,
"loss": 0.1139,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6898624897003174,
"rewards/margins": 3.6898624897003174,
"rewards/rejected": 0.0,
"step": 2046
},
{
"epoch": 11.435754189944134,
"grad_norm": 0.45539848253727405,
"learning_rate": 2.2816631173857347e-07,
"logits/chosen": -3.114863634109497,
"logits/rejected": -3.171290636062622,
"logps/chosen": -1.5149728059768677,
"logps/rejected": -43.40138244628906,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.839877128601074,
"rewards/margins": 2.839877128601074,
"rewards/rejected": 0.0,
"step": 2047
},
{
"epoch": 11.441340782122905,
"grad_norm": 0.44648626009766645,
"learning_rate": 2.2765492264435087e-07,
"logits/chosen": -3.459627151489258,
"logits/rejected": -3.62988543510437,
"logps/chosen": -0.8486126065254211,
"logps/rejected": -46.58003616333008,
"loss": 0.1024,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.526358127593994,
"rewards/margins": 2.526358127593994,
"rewards/rejected": 0.0,
"step": 2048
},
{
"epoch": 11.446927374301676,
"grad_norm": 0.8420872506929729,
"learning_rate": 2.2714393830341482e-07,
"logits/chosen": -3.106123924255371,
"logits/rejected": -3.256624698638916,
"logps/chosen": -2.9391133785247803,
"logps/rejected": -64.49932098388672,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3291449546813965,
"rewards/margins": 2.3291449546813965,
"rewards/rejected": 0.0,
"step": 2049
},
{
"epoch": 11.452513966480447,
"grad_norm": 0.4427576769856584,
"learning_rate": 2.266333594751797e-07,
"logits/chosen": -3.3989005088806152,
"logits/rejected": -2.927534341812134,
"logps/chosen": -0.3201085925102234,
"logps/rejected": -91.62421417236328,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.822981357574463,
"rewards/margins": 1.822981357574463,
"rewards/rejected": 0.0,
"step": 2050
},
{
"epoch": 11.458100558659218,
"grad_norm": 1.9602573980316333,
"learning_rate": 2.2612318691845628e-07,
"logits/chosen": -3.3661935329437256,
"logits/rejected": -3.5726842880249023,
"logps/chosen": -0.6012658476829529,
"logps/rejected": -58.52061462402344,
"loss": 0.1406,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9937279224395752,
"rewards/margins": 1.9937279224395752,
"rewards/rejected": 0.0,
"step": 2051
},
{
"epoch": 11.46368715083799,
"grad_norm": 0.515007798693676,
"learning_rate": 2.256134213914519e-07,
"logits/chosen": -3.470977783203125,
"logits/rejected": -3.5958571434020996,
"logps/chosen": -6.127221584320068,
"logps/rejected": -31.5274715423584,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0713388919830322,
"rewards/margins": 3.0713388919830322,
"rewards/rejected": 0.0,
"step": 2052
},
{
"epoch": 11.46927374301676,
"grad_norm": 0.4701438832254253,
"learning_rate": 2.2510406365176876e-07,
"logits/chosen": -3.633430004119873,
"logits/rejected": -3.618941068649292,
"logps/chosen": -9.452795028686523,
"logps/rejected": -26.49870491027832,
"loss": 0.0943,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.020148515701294,
"rewards/margins": 3.020148515701294,
"rewards/rejected": 0.0,
"step": 2053
},
{
"epoch": 11.474860335195531,
"grad_norm": 0.48878517668358795,
"learning_rate": 2.245951144564036e-07,
"logits/chosen": -3.1848111152648926,
"logits/rejected": -3.3400304317474365,
"logps/chosen": -1.1764954328536987,
"logps/rejected": -49.090797424316406,
"loss": 0.0908,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.439659833908081,
"rewards/margins": 2.439659833908081,
"rewards/rejected": 0.0,
"step": 2054
},
{
"epoch": 11.480446927374302,
"grad_norm": 0.5292091561947605,
"learning_rate": 2.2408657456174519e-07,
"logits/chosen": -3.138007640838623,
"logits/rejected": -3.077355146408081,
"logps/chosen": -19.65635108947754,
"logps/rejected": -77.37872314453125,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.171947479248047,
"rewards/margins": 3.171947479248047,
"rewards/rejected": 0.0,
"step": 2055
},
{
"epoch": 11.486033519553073,
"grad_norm": 0.7808373197064731,
"learning_rate": 2.2357844472357428e-07,
"logits/chosen": -3.347581148147583,
"logits/rejected": -3.231194019317627,
"logps/chosen": -0.8362330198287964,
"logps/rejected": -72.31649780273438,
"loss": 0.1392,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3305838108062744,
"rewards/margins": 2.3305838108062744,
"rewards/rejected": 0.0,
"step": 2056
},
{
"epoch": 11.491620111731844,
"grad_norm": 0.5000031262306723,
"learning_rate": 2.2307072569706286e-07,
"logits/chosen": -3.4568393230438232,
"logits/rejected": -3.389314889907837,
"logps/chosen": -1.3895785808563232,
"logps/rejected": -46.76854705810547,
"loss": 0.1029,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2795677185058594,
"rewards/margins": 2.2795677185058594,
"rewards/rejected": 0.0,
"step": 2057
},
{
"epoch": 11.497206703910614,
"grad_norm": 0.6276310130319155,
"learning_rate": 2.2256341823677105e-07,
"logits/chosen": -3.4142253398895264,
"logits/rejected": -3.395514965057373,
"logps/chosen": -0.7626144289970398,
"logps/rejected": -52.26962661743164,
"loss": 0.0947,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.44633150100708,
"rewards/margins": 2.44633150100708,
"rewards/rejected": 0.0,
"step": 2058
},
{
"epoch": 11.502793296089386,
"grad_norm": 0.7585110112800841,
"learning_rate": 2.2205652309664874e-07,
"logits/chosen": -3.391388177871704,
"logits/rejected": -3.6604607105255127,
"logps/chosen": -3.459272623062134,
"logps/rejected": -98.18557739257812,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.000126838684082,
"rewards/margins": 2.000126838684082,
"rewards/rejected": 0.0,
"step": 2059
},
{
"epoch": 11.508379888268156,
"grad_norm": 0.5827440897026235,
"learning_rate": 2.2155004103003204e-07,
"logits/chosen": -3.507307529449463,
"logits/rejected": -3.482379913330078,
"logps/chosen": -1.2515177726745605,
"logps/rejected": -47.592872619628906,
"loss": 0.1379,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8853281736373901,
"rewards/margins": 1.8853281736373901,
"rewards/rejected": 0.0,
"step": 2060
},
{
"epoch": 11.508379888268156,
"eval_logits/chosen": -3.2820420265197754,
"eval_logits/rejected": -3.407228469848633,
"eval_logps/chosen": -29.63604164123535,
"eval_logps/rejected": -53.693016052246094,
"eval_loss": 0.9693609476089478,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.04340028017759323,
"eval_rewards/margins": 0.04340028017759323,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7388,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 2060
},
{
"epoch": 11.513966480446927,
"grad_norm": 0.7229903736753329,
"learning_rate": 2.210439727896437e-07,
"logits/chosen": -3.416884183883667,
"logits/rejected": -3.483350992202759,
"logps/chosen": -0.4949205815792084,
"logps/rejected": -26.10192108154297,
"loss": 0.0943,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.139338970184326,
"rewards/margins": 2.139338970184326,
"rewards/rejected": 0.0,
"step": 2061
},
{
"epoch": 11.519553072625698,
"grad_norm": 2.0827388719160456,
"learning_rate": 2.205383191275909e-07,
"logits/chosen": -3.3750455379486084,
"logits/rejected": -3.5473861694335938,
"logps/chosen": -9.99547290802002,
"logps/rejected": -52.88782501220703,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.794489622116089,
"rewards/margins": 2.794489622116089,
"rewards/rejected": 0.0,
"step": 2062
},
{
"epoch": 11.525139664804469,
"grad_norm": 0.4764702527149911,
"learning_rate": 2.200330807953657e-07,
"logits/chosen": -3.4228296279907227,
"logits/rejected": -3.34614634513855,
"logps/chosen": -1.506871223449707,
"logps/rejected": -73.5494384765625,
"loss": 0.0976,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.776228427886963,
"rewards/margins": 1.776228427886963,
"rewards/rejected": 0.0,
"step": 2063
},
{
"epoch": 11.53072625698324,
"grad_norm": 0.6727045742513333,
"learning_rate": 2.1952825854384138e-07,
"logits/chosen": -3.231311082839966,
"logits/rejected": -3.223719358444214,
"logps/chosen": -4.259906768798828,
"logps/rejected": -47.73058319091797,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.09010648727417,
"rewards/margins": 3.09010648727417,
"rewards/rejected": 0.0,
"step": 2064
},
{
"epoch": 11.53631284916201,
"grad_norm": 0.5606730494853186,
"learning_rate": 2.1902385312327438e-07,
"logits/chosen": -3.5459465980529785,
"logits/rejected": -3.4588568210601807,
"logps/chosen": -4.553582668304443,
"logps/rejected": -40.20221710205078,
"loss": 0.0726,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.609736442565918,
"rewards/margins": 3.609736442565918,
"rewards/rejected": 0.0,
"step": 2065
},
{
"epoch": 11.541899441340782,
"grad_norm": 0.42592803024726883,
"learning_rate": 2.185198652833009e-07,
"logits/chosen": -3.2113282680511475,
"logits/rejected": -3.3077077865600586,
"logps/chosen": -3.825059175491333,
"logps/rejected": -82.86166381835938,
"loss": 0.0893,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2419495582580566,
"rewards/margins": 2.2419495582580566,
"rewards/rejected": 0.0,
"step": 2066
},
{
"epoch": 11.547486033519553,
"grad_norm": 0.4329193833006763,
"learning_rate": 2.1801629577293617e-07,
"logits/chosen": -3.5197689533233643,
"logits/rejected": -3.620059013366699,
"logps/chosen": -1.2800936698913574,
"logps/rejected": -47.56212615966797,
"loss": 0.0902,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.994467258453369,
"rewards/margins": 2.994467258453369,
"rewards/rejected": 0.0,
"step": 2067
},
{
"epoch": 11.553072625698324,
"grad_norm": 0.6671559109647464,
"learning_rate": 2.1751314534057458e-07,
"logits/chosen": -3.589369773864746,
"logits/rejected": -3.6188716888427734,
"logps/chosen": -1.5386724472045898,
"logps/rejected": -43.215354919433594,
"loss": 0.0966,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6964192390441895,
"rewards/margins": 2.6964192390441895,
"rewards/rejected": 0.0,
"step": 2068
},
{
"epoch": 11.558659217877095,
"grad_norm": 1.173315727758856,
"learning_rate": 2.1701041473398724e-07,
"logits/chosen": -3.56231427192688,
"logits/rejected": -3.6502504348754883,
"logps/chosen": -2.079115629196167,
"logps/rejected": -54.19228744506836,
"loss": 0.086,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3071088790893555,
"rewards/margins": 2.3071088790893555,
"rewards/rejected": 0.0,
"step": 2069
},
{
"epoch": 11.564245810055866,
"grad_norm": 0.40321424019889146,
"learning_rate": 2.1650810470032126e-07,
"logits/chosen": -3.4973809719085693,
"logits/rejected": -3.4846181869506836,
"logps/chosen": -0.876288652420044,
"logps/rejected": -48.02561569213867,
"loss": 0.0777,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3294496536254883,
"rewards/margins": 2.3294496536254883,
"rewards/rejected": 0.0,
"step": 2070
},
{
"epoch": 11.569832402234637,
"grad_norm": 0.7211952196032335,
"learning_rate": 2.1600621598609863e-07,
"logits/chosen": -3.3570892810821533,
"logits/rejected": -3.3892922401428223,
"logps/chosen": -0.2817106544971466,
"logps/rejected": -75.76512908935547,
"loss": 0.1143,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5755364894866943,
"rewards/margins": 1.5755364894866943,
"rewards/rejected": 0.0,
"step": 2071
},
{
"epoch": 11.575418994413408,
"grad_norm": 0.5148112052635592,
"learning_rate": 2.1550474933721603e-07,
"logits/chosen": -3.1936018466949463,
"logits/rejected": -3.2042415142059326,
"logps/chosen": -2.460733652114868,
"logps/rejected": -67.72477722167969,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5301566123962402,
"rewards/margins": 2.5301566123962402,
"rewards/rejected": 0.0,
"step": 2072
},
{
"epoch": 11.581005586592179,
"grad_norm": 0.6167113232137355,
"learning_rate": 2.1500370549894132e-07,
"logits/chosen": -3.265223503112793,
"logits/rejected": -3.202566146850586,
"logps/chosen": -1.2127795219421387,
"logps/rejected": -49.323612213134766,
"loss": 0.0969,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4704246520996094,
"rewards/margins": 2.4704246520996094,
"rewards/rejected": 0.0,
"step": 2073
},
{
"epoch": 11.58659217877095,
"grad_norm": 0.6130793910733988,
"learning_rate": 2.145030852159155e-07,
"logits/chosen": -3.730278730392456,
"logits/rejected": -3.687791109085083,
"logps/chosen": -0.1538744866847992,
"logps/rejected": -42.28436279296875,
"loss": 0.0868,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4262964725494385,
"rewards/margins": 1.4262964725494385,
"rewards/rejected": 0.0,
"step": 2074
},
{
"epoch": 11.59217877094972,
"grad_norm": 1.1180693351795155,
"learning_rate": 2.1400288923214937e-07,
"logits/chosen": -3.408236026763916,
"logits/rejected": -3.3714475631713867,
"logps/chosen": -1.658205509185791,
"logps/rejected": -54.93598175048828,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5506534576416016,
"rewards/margins": 2.5506534576416016,
"rewards/rejected": 0.0,
"step": 2075
},
{
"epoch": 11.597765363128492,
"grad_norm": 0.5064447962549264,
"learning_rate": 2.1350311829102285e-07,
"logits/chosen": -3.3614866733551025,
"logits/rejected": -3.36944580078125,
"logps/chosen": -0.326558381319046,
"logps/rejected": -53.19791030883789,
"loss": 0.0934,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6810168027877808,
"rewards/margins": 1.6810168027877808,
"rewards/rejected": 0.0,
"step": 2076
},
{
"epoch": 11.603351955307263,
"grad_norm": 0.5463711792928004,
"learning_rate": 2.1300377313528523e-07,
"logits/chosen": -3.37495493888855,
"logits/rejected": -3.3020448684692383,
"logps/chosen": -7.588983535766602,
"logps/rejected": -59.491539001464844,
"loss": 0.0947,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7611913681030273,
"rewards/margins": 2.7611913681030273,
"rewards/rejected": 0.0,
"step": 2077
},
{
"epoch": 11.608938547486034,
"grad_norm": 0.8164411536318138,
"learning_rate": 2.1250485450705159e-07,
"logits/chosen": -3.532632350921631,
"logits/rejected": -3.720940589904785,
"logps/chosen": -5.543283939361572,
"logps/rejected": -85.24209594726562,
"loss": 0.0807,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0614562034606934,
"rewards/margins": 3.0614562034606934,
"rewards/rejected": 0.0,
"step": 2078
},
{
"epoch": 11.614525139664805,
"grad_norm": 0.927976280513748,
"learning_rate": 2.120063631478044e-07,
"logits/chosen": -3.5740299224853516,
"logits/rejected": -3.7173476219177246,
"logps/chosen": -10.757414817810059,
"logps/rejected": -35.99129867553711,
"loss": 0.1068,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0895583629608154,
"rewards/margins": 3.0895583629608154,
"rewards/rejected": 0.0,
"step": 2079
},
{
"epoch": 11.620111731843576,
"grad_norm": 0.5243803748945338,
"learning_rate": 2.115082997983904e-07,
"logits/chosen": -3.521507740020752,
"logits/rejected": -3.1975574493408203,
"logps/chosen": -3.892530679702759,
"logps/rejected": -26.3160400390625,
"loss": 0.0688,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4351625442504883,
"rewards/margins": 3.4351625442504883,
"rewards/rejected": 0.0,
"step": 2080
},
{
"epoch": 11.620111731843576,
"eval_logits/chosen": -3.2764930725097656,
"eval_logits/rejected": -3.4013266563415527,
"eval_logps/chosen": -29.51177978515625,
"eval_logps/rejected": -54.050270080566406,
"eval_loss": 0.9726657271385193,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.05582625791430473,
"eval_rewards/margins": 0.05582625791430473,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7021,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 2080
},
{
"epoch": 11.625698324022347,
"grad_norm": 0.5444743947234505,
"learning_rate": 2.110106651990205e-07,
"logits/chosen": -3.456770658493042,
"logits/rejected": -3.2560853958129883,
"logps/chosen": -0.5817017555236816,
"logps/rejected": -118.65254211425781,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1886162757873535,
"rewards/margins": 2.1886162757873535,
"rewards/rejected": 0.0,
"step": 2081
},
{
"epoch": 11.631284916201118,
"grad_norm": 1.1128860956078888,
"learning_rate": 2.1051346008926812e-07,
"logits/chosen": -3.5402708053588867,
"logits/rejected": -3.5771753787994385,
"logps/chosen": -7.171589374542236,
"logps/rejected": -23.094192504882812,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2549638748168945,
"rewards/margins": 3.2549638748168945,
"rewards/rejected": 0.0,
"step": 2082
},
{
"epoch": 11.636871508379889,
"grad_norm": 0.5792323796491307,
"learning_rate": 2.1001668520806913e-07,
"logits/chosen": -3.3181416988372803,
"logits/rejected": -3.6179141998291016,
"logps/chosen": -10.795852661132812,
"logps/rejected": -59.83184051513672,
"loss": 0.0908,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2504091262817383,
"rewards/margins": 3.2504091262817383,
"rewards/rejected": 0.0,
"step": 2083
},
{
"epoch": 11.64245810055866,
"grad_norm": 0.4761403785744252,
"learning_rate": 2.0952034129371882e-07,
"logits/chosen": -3.581909418106079,
"logits/rejected": -3.1815035343170166,
"logps/chosen": -11.74966049194336,
"logps/rejected": -73.74712371826172,
"loss": 0.0758,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.506298542022705,
"rewards/margins": 3.506298542022705,
"rewards/rejected": 0.0,
"step": 2084
},
{
"epoch": 11.64804469273743,
"grad_norm": 0.5056695575110497,
"learning_rate": 2.0902442908387307e-07,
"logits/chosen": -3.187776803970337,
"logits/rejected": -3.6853528022766113,
"logps/chosen": -1.2484012842178345,
"logps/rejected": -62.97319793701172,
"loss": 0.0901,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.032311201095581,
"rewards/margins": 2.032311201095581,
"rewards/rejected": 0.0,
"step": 2085
},
{
"epoch": 11.653631284916202,
"grad_norm": 1.9086239252426467,
"learning_rate": 2.0852894931554582e-07,
"logits/chosen": -3.5712294578552246,
"logits/rejected": -3.558100700378418,
"logps/chosen": -7.533488750457764,
"logps/rejected": -67.64964294433594,
"loss": 0.1437,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.256208896636963,
"rewards/margins": 3.256208896636963,
"rewards/rejected": 0.0,
"step": 2086
},
{
"epoch": 11.659217877094973,
"grad_norm": 0.5605342456016813,
"learning_rate": 2.0803390272510762e-07,
"logits/chosen": -3.5309107303619385,
"logits/rejected": -3.030372381210327,
"logps/chosen": -10.332486152648926,
"logps/rejected": -64.68547821044922,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2190003395080566,
"rewards/margins": 3.2190003395080566,
"rewards/rejected": 0.0,
"step": 2087
},
{
"epoch": 11.664804469273744,
"grad_norm": 0.530977087774688,
"learning_rate": 2.0753929004828653e-07,
"logits/chosen": -3.2939679622650146,
"logits/rejected": -3.307115077972412,
"logps/chosen": -1.5258204936981201,
"logps/rejected": -64.69021606445312,
"loss": 0.0944,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4115960597991943,
"rewards/margins": 2.4115960597991943,
"rewards/rejected": 0.0,
"step": 2088
},
{
"epoch": 11.670391061452515,
"grad_norm": 0.4456606009275189,
"learning_rate": 2.0704511202016483e-07,
"logits/chosen": -3.461855173110962,
"logits/rejected": -3.3375463485717773,
"logps/chosen": -0.309223473072052,
"logps/rejected": -48.380401611328125,
"loss": 0.115,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6402618885040283,
"rewards/margins": 1.6402618885040283,
"rewards/rejected": 0.0,
"step": 2089
},
{
"epoch": 11.675977653631286,
"grad_norm": 0.41630393104425495,
"learning_rate": 2.0655136937517903e-07,
"logits/chosen": -3.4845030307769775,
"logits/rejected": -3.4499826431274414,
"logps/chosen": -0.4920295774936676,
"logps/rejected": -35.75161361694336,
"loss": 0.0941,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.52449893951416,
"rewards/margins": 2.52449893951416,
"rewards/rejected": 0.0,
"step": 2090
},
{
"epoch": 11.681564245810057,
"grad_norm": 0.5409384675075213,
"learning_rate": 2.0605806284711857e-07,
"logits/chosen": -3.332338333129883,
"logits/rejected": -3.3106095790863037,
"logps/chosen": -1.0673799514770508,
"logps/rejected": -61.278194427490234,
"loss": 0.1149,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2988715171813965,
"rewards/margins": 2.2988715171813965,
"rewards/rejected": 0.0,
"step": 2091
},
{
"epoch": 11.687150837988828,
"grad_norm": 2.4057489678597936,
"learning_rate": 2.0556519316912523e-07,
"logits/chosen": -3.5256285667419434,
"logits/rejected": -3.313889265060425,
"logps/chosen": -1.9876480102539062,
"logps/rejected": -53.89966583251953,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1563377380371094,
"rewards/margins": 3.1563377380371094,
"rewards/rejected": 0.0,
"step": 2092
},
{
"epoch": 11.692737430167599,
"grad_norm": 0.5666254690362961,
"learning_rate": 2.0507276107369047e-07,
"logits/chosen": -3.126408338546753,
"logits/rejected": -3.331571102142334,
"logps/chosen": -0.6741228103637695,
"logps/rejected": -52.774986267089844,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9165393114089966,
"rewards/margins": 1.9165393114089966,
"rewards/rejected": 0.0,
"step": 2093
},
{
"epoch": 11.69832402234637,
"grad_norm": 1.180640827580221,
"learning_rate": 2.045807672926566e-07,
"logits/chosen": -3.6363162994384766,
"logits/rejected": -3.535933494567871,
"logps/chosen": -0.7559529542922974,
"logps/rejected": -32.57082748413086,
"loss": 0.1104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1516265869140625,
"rewards/margins": 2.1516265869140625,
"rewards/rejected": 0.0,
"step": 2094
},
{
"epoch": 11.703910614525139,
"grad_norm": 1.0739088949985398,
"learning_rate": 2.040892125572138e-07,
"logits/chosen": -3.442941904067993,
"logits/rejected": -3.2288362979888916,
"logps/chosen": -0.630698561668396,
"logps/rejected": -53.339263916015625,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0331945419311523,
"rewards/margins": 2.0331945419311523,
"rewards/rejected": 0.0,
"step": 2095
},
{
"epoch": 11.70949720670391,
"grad_norm": 1.968379364228944,
"learning_rate": 2.0359809759789997e-07,
"logits/chosen": -3.225104331970215,
"logits/rejected": -3.3189284801483154,
"logps/chosen": -1.058898687362671,
"logps/rejected": -53.428955078125,
"loss": 0.1065,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.358876943588257,
"rewards/margins": 2.358876943588257,
"rewards/rejected": 0.0,
"step": 2096
},
{
"epoch": 11.71508379888268,
"grad_norm": 0.8065051146922257,
"learning_rate": 2.0310742314459945e-07,
"logits/chosen": -3.1914432048797607,
"logits/rejected": -3.2663049697875977,
"logps/chosen": -9.212248802185059,
"logps/rejected": -52.66050338745117,
"loss": 0.0733,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.001647472381592,
"rewards/margins": 3.001647472381592,
"rewards/rejected": 0.0,
"step": 2097
},
{
"epoch": 11.720670391061452,
"grad_norm": 2.1276834324619203,
"learning_rate": 2.0261718992654175e-07,
"logits/chosen": -3.418238639831543,
"logits/rejected": -3.5759994983673096,
"logps/chosen": -1.724967360496521,
"logps/rejected": -67.7520751953125,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7611351013183594,
"rewards/margins": 2.7611351013183594,
"rewards/rejected": 0.0,
"step": 2098
},
{
"epoch": 11.726256983240223,
"grad_norm": 1.9786927674124697,
"learning_rate": 2.0212739867230083e-07,
"logits/chosen": -3.475306510925293,
"logits/rejected": -3.4787042140960693,
"logps/chosen": -3.1913681030273438,
"logps/rejected": -29.67127227783203,
"loss": 0.1106,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3035426139831543,
"rewards/margins": 3.3035426139831543,
"rewards/rejected": 0.0,
"step": 2099
},
{
"epoch": 11.731843575418994,
"grad_norm": 1.1462877697832752,
"learning_rate": 2.0163805010979351e-07,
"logits/chosen": -3.322868824005127,
"logits/rejected": -3.2770586013793945,
"logps/chosen": -7.151961803436279,
"logps/rejected": -94.27562713623047,
"loss": 0.1238,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.769965171813965,
"rewards/margins": 2.769965171813965,
"rewards/rejected": 0.0,
"step": 2100
},
{
"epoch": 11.731843575418994,
"eval_logits/chosen": -3.273817777633667,
"eval_logits/rejected": -3.3991897106170654,
"eval_logps/chosen": -29.351842880249023,
"eval_logps/rejected": -53.964881896972656,
"eval_loss": 0.9642612338066101,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.07181987911462784,
"eval_rewards/margins": 0.07181987911462784,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7177,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 2100
},
{
"epoch": 11.737430167597765,
"grad_norm": 0.4820005838869611,
"learning_rate": 2.0114914496627943e-07,
"logits/chosen": -3.5352282524108887,
"logits/rejected": -3.244537830352783,
"logps/chosen": -0.8986507058143616,
"logps/rejected": -57.75574493408203,
"loss": 0.1192,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.129504919052124,
"rewards/margins": 2.129504919052124,
"rewards/rejected": 0.0,
"step": 2101
},
{
"epoch": 11.743016759776536,
"grad_norm": 0.5069767628399827,
"learning_rate": 2.00660683968358e-07,
"logits/chosen": -3.225170612335205,
"logits/rejected": -3.566986322402954,
"logps/chosen": -0.30834734439849854,
"logps/rejected": -101.76130676269531,
"loss": 0.1223,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7267638444900513,
"rewards/margins": 1.7267638444900513,
"rewards/rejected": 0.0,
"step": 2102
},
{
"epoch": 11.748603351955307,
"grad_norm": 0.5131499011156985,
"learning_rate": 2.001726678419698e-07,
"logits/chosen": -3.2986557483673096,
"logits/rejected": -3.2515666484832764,
"logps/chosen": -0.8364378213882446,
"logps/rejected": -45.09752655029297,
"loss": 0.0952,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3112003803253174,
"rewards/margins": 2.3112003803253174,
"rewards/rejected": 0.0,
"step": 2103
},
{
"epoch": 11.754189944134078,
"grad_norm": 1.3827191962840166,
"learning_rate": 1.9968509731239352e-07,
"logits/chosen": -3.241009473800659,
"logits/rejected": -3.2275688648223877,
"logps/chosen": -0.8142117261886597,
"logps/rejected": -60.17033767700195,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4198660850524902,
"rewards/margins": 2.4198660850524902,
"rewards/rejected": 0.0,
"step": 2104
},
{
"epoch": 11.759776536312849,
"grad_norm": 4.5464145286381195,
"learning_rate": 1.991979731042458e-07,
"logits/chosen": -3.138615369796753,
"logits/rejected": -3.4954192638397217,
"logps/chosen": -4.946340560913086,
"logps/rejected": -149.63125610351562,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6405844688415527,
"rewards/margins": 2.6405844688415527,
"rewards/rejected": 0.0,
"step": 2105
},
{
"epoch": 11.76536312849162,
"grad_norm": 1.816761052938629,
"learning_rate": 1.987112959414801e-07,
"logits/chosen": -3.4835660457611084,
"logits/rejected": -3.4168593883514404,
"logps/chosen": -2.545900821685791,
"logps/rejected": -33.79948425292969,
"loss": 0.1533,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.324270009994507,
"rewards/margins": 2.324270009994507,
"rewards/rejected": 0.0,
"step": 2106
},
{
"epoch": 11.77094972067039,
"grad_norm": 1.021166904232079,
"learning_rate": 1.9822506654738487e-07,
"logits/chosen": -3.2605655193328857,
"logits/rejected": -3.407984733581543,
"logps/chosen": -1.2512481212615967,
"logps/rejected": -46.00942611694336,
"loss": 0.1047,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4974803924560547,
"rewards/margins": 2.4974803924560547,
"rewards/rejected": 0.0,
"step": 2107
},
{
"epoch": 11.776536312849162,
"grad_norm": 2.827081417890233,
"learning_rate": 1.9773928564458408e-07,
"logits/chosen": -3.1404619216918945,
"logits/rejected": -3.169956684112549,
"logps/chosen": -9.449261665344238,
"logps/rejected": -33.91954803466797,
"loss": 0.0939,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.326777458190918,
"rewards/margins": 3.326777458190918,
"rewards/rejected": 0.0,
"step": 2108
},
{
"epoch": 11.782122905027933,
"grad_norm": 0.9347712030633651,
"learning_rate": 1.972539539550346e-07,
"logits/chosen": -3.3544299602508545,
"logits/rejected": -3.3321897983551025,
"logps/chosen": -4.494087219238281,
"logps/rejected": -33.97837829589844,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.867629051208496,
"rewards/margins": 2.867629051208496,
"rewards/rejected": 0.0,
"step": 2109
},
{
"epoch": 11.787709497206704,
"grad_norm": 0.44190054236605475,
"learning_rate": 1.967690722000256e-07,
"logits/chosen": -3.0685782432556152,
"logits/rejected": -3.000790596008301,
"logps/chosen": -3.3948845863342285,
"logps/rejected": -47.722129821777344,
"loss": 0.1099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6497321128845215,
"rewards/margins": 2.6497321128845215,
"rewards/rejected": 0.0,
"step": 2110
},
{
"epoch": 11.793296089385475,
"grad_norm": 0.40784498624049237,
"learning_rate": 1.9628464110017756e-07,
"logits/chosen": -3.157773017883301,
"logits/rejected": -3.2839903831481934,
"logps/chosen": -6.024378776550293,
"logps/rejected": -52.742496490478516,
"loss": 0.1117,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5025463104248047,
"rewards/margins": 2.5025463104248047,
"rewards/rejected": 0.0,
"step": 2111
},
{
"epoch": 11.798882681564246,
"grad_norm": 0.9165257709534127,
"learning_rate": 1.9580066137544204e-07,
"logits/chosen": -3.252223014831543,
"logits/rejected": -3.277956008911133,
"logps/chosen": -0.549738883972168,
"logps/rejected": -31.06273078918457,
"loss": 0.1134,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8361353874206543,
"rewards/margins": 1.8361353874206543,
"rewards/rejected": 0.0,
"step": 2112
},
{
"epoch": 11.804469273743017,
"grad_norm": 0.6663716498748804,
"learning_rate": 1.9531713374509823e-07,
"logits/chosen": -3.544379472732544,
"logits/rejected": -3.51424241065979,
"logps/chosen": -1.7592977285385132,
"logps/rejected": -36.52344512939453,
"loss": 0.0691,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8453869819641113,
"rewards/margins": 2.8453869819641113,
"rewards/rejected": 0.0,
"step": 2113
},
{
"epoch": 11.810055865921788,
"grad_norm": 0.5070481300564695,
"learning_rate": 1.9483405892775496e-07,
"logits/chosen": -3.4501495361328125,
"logits/rejected": -3.3233320713043213,
"logps/chosen": -2.323268413543701,
"logps/rejected": -49.021034240722656,
"loss": 0.0926,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.024831533432007,
"rewards/margins": 3.024831533432007,
"rewards/rejected": 0.0,
"step": 2114
},
{
"epoch": 11.815642458100559,
"grad_norm": 0.5281047641423829,
"learning_rate": 1.9435143764134716e-07,
"logits/chosen": -3.6315793991088867,
"logits/rejected": -3.5739078521728516,
"logps/chosen": -1.9289957284927368,
"logps/rejected": -43.96451950073242,
"loss": 0.1205,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6348986625671387,
"rewards/margins": 2.6348986625671387,
"rewards/rejected": 0.0,
"step": 2115
},
{
"epoch": 11.82122905027933,
"grad_norm": 1.869397832797094,
"learning_rate": 1.9386927060313606e-07,
"logits/chosen": -3.134777307510376,
"logits/rejected": -3.1398351192474365,
"logps/chosen": -8.755775451660156,
"logps/rejected": -29.74664878845215,
"loss": 0.1416,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.388301372528076,
"rewards/margins": 3.388301372528076,
"rewards/rejected": 0.0,
"step": 2116
},
{
"epoch": 11.8268156424581,
"grad_norm": 0.43745260018046894,
"learning_rate": 1.9338755852970777e-07,
"logits/chosen": -3.228424549102783,
"logits/rejected": -3.3705661296844482,
"logps/chosen": -1.1017125844955444,
"logps/rejected": -50.154632568359375,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.395521640777588,
"rewards/margins": 2.395521640777588,
"rewards/rejected": 0.0,
"step": 2117
},
{
"epoch": 11.832402234636872,
"grad_norm": 0.465204292742605,
"learning_rate": 1.9290630213697217e-07,
"logits/chosen": -3.456580638885498,
"logits/rejected": -3.3084499835968018,
"logps/chosen": -4.237008094787598,
"logps/rejected": -41.89889907836914,
"loss": 0.1071,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5508902072906494,
"rewards/margins": 3.5508902072906494,
"rewards/rejected": 0.0,
"step": 2118
},
{
"epoch": 11.837988826815643,
"grad_norm": 1.9742680386399745,
"learning_rate": 1.92425502140162e-07,
"logits/chosen": -3.3315179347991943,
"logits/rejected": -3.209519147872925,
"logps/chosen": -4.020634651184082,
"logps/rejected": -30.618637084960938,
"loss": 0.1219,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.715378999710083,
"rewards/margins": 2.715378999710083,
"rewards/rejected": 0.0,
"step": 2119
},
{
"epoch": 11.843575418994414,
"grad_norm": 0.9255626394952459,
"learning_rate": 1.9194515925383149e-07,
"logits/chosen": -3.1456193923950195,
"logits/rejected": -3.241938829421997,
"logps/chosen": -10.063833236694336,
"logps/rejected": -51.778404235839844,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.76076340675354,
"rewards/margins": 2.76076340675354,
"rewards/rejected": 0.0,
"step": 2120
},
{
"epoch": 11.843575418994414,
"eval_logits/chosen": -3.272770643234253,
"eval_logits/rejected": -3.3985888957977295,
"eval_logps/chosen": -29.64059066772461,
"eval_logps/rejected": -54.02800750732422,
"eval_loss": 0.9608362913131714,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.042945243418216705,
"eval_rewards/margins": 0.042945243418216705,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7156,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 2120
},
{
"epoch": 11.849162011173185,
"grad_norm": 1.3394621049516837,
"learning_rate": 1.9146527419185626e-07,
"logits/chosen": -3.4177379608154297,
"logits/rejected": -3.447481393814087,
"logps/chosen": -1.6461918354034424,
"logps/rejected": -57.266231536865234,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7926669120788574,
"rewards/margins": 2.7926669120788574,
"rewards/rejected": 0.0,
"step": 2121
},
{
"epoch": 11.854748603351956,
"grad_norm": 0.4624204756202754,
"learning_rate": 1.9098584766743027e-07,
"logits/chosen": -3.424178123474121,
"logits/rejected": -3.4454879760742188,
"logps/chosen": -1.4741344451904297,
"logps/rejected": -52.27875518798828,
"loss": 0.1387,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6950955390930176,
"rewards/margins": 2.6950955390930176,
"rewards/rejected": 0.0,
"step": 2122
},
{
"epoch": 11.860335195530727,
"grad_norm": 0.7365250401734182,
"learning_rate": 1.9050688039306716e-07,
"logits/chosen": -3.210947275161743,
"logits/rejected": -3.3241193294525146,
"logps/chosen": -2.2139992713928223,
"logps/rejected": -64.00433349609375,
"loss": 0.0718,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1638004779815674,
"rewards/margins": 3.1638004779815674,
"rewards/rejected": 0.0,
"step": 2123
},
{
"epoch": 11.865921787709498,
"grad_norm": 0.7778971142705587,
"learning_rate": 1.9002837308059756e-07,
"logits/chosen": -2.925323009490967,
"logits/rejected": -2.8885862827301025,
"logps/chosen": -8.409188270568848,
"logps/rejected": -55.37174606323242,
"loss": 0.1143,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9058103561401367,
"rewards/margins": 2.9058103561401367,
"rewards/rejected": 0.0,
"step": 2124
},
{
"epoch": 11.871508379888269,
"grad_norm": 0.8504688428084934,
"learning_rate": 1.8955032644116847e-07,
"logits/chosen": -3.2839128971099854,
"logits/rejected": -3.3986287117004395,
"logps/chosen": -3.0384836196899414,
"logps/rejected": -57.346824645996094,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.340529441833496,
"rewards/margins": 2.340529441833496,
"rewards/rejected": 0.0,
"step": 2125
},
{
"epoch": 11.87709497206704,
"grad_norm": 4.1360967384429115,
"learning_rate": 1.8907274118524235e-07,
"logits/chosen": -3.41344952583313,
"logits/rejected": -3.3344621658325195,
"logps/chosen": -5.416000843048096,
"logps/rejected": -33.39531707763672,
"loss": 0.1251,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1773829460144043,
"rewards/margins": 3.1773829460144043,
"rewards/rejected": 0.0,
"step": 2126
},
{
"epoch": 11.88268156424581,
"grad_norm": 0.49328609463654216,
"learning_rate": 1.8859561802259595e-07,
"logits/chosen": -3.2809367179870605,
"logits/rejected": -3.4360008239746094,
"logps/chosen": -0.2677420973777771,
"logps/rejected": -45.579261779785156,
"loss": 0.1304,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.699220895767212,
"rewards/margins": 1.699220895767212,
"rewards/rejected": 0.0,
"step": 2127
},
{
"epoch": 11.888268156424582,
"grad_norm": 0.9312043475282246,
"learning_rate": 1.8811895766231928e-07,
"logits/chosen": -3.3856492042541504,
"logits/rejected": -3.301811456680298,
"logps/chosen": -0.7853468656539917,
"logps/rejected": -97.83235168457031,
"loss": 0.0943,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.086432456970215,
"rewards/margins": 2.086432456970215,
"rewards/rejected": 0.0,
"step": 2128
},
{
"epoch": 11.893854748603353,
"grad_norm": 0.6598547775012298,
"learning_rate": 1.8764276081281428e-07,
"logits/chosen": -3.3419158458709717,
"logits/rejected": -3.520700693130493,
"logps/chosen": -3.2810051441192627,
"logps/rejected": -53.71620559692383,
"loss": 0.1139,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.639993667602539,
"rewards/margins": 3.639993667602539,
"rewards/rejected": 0.0,
"step": 2129
},
{
"epoch": 11.899441340782122,
"grad_norm": 0.4169267337223477,
"learning_rate": 1.8716702818179486e-07,
"logits/chosen": -3.5588531494140625,
"logits/rejected": -3.3309450149536133,
"logps/chosen": -1.766896367073059,
"logps/rejected": -52.8946533203125,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.736774444580078,
"rewards/margins": 2.736774444580078,
"rewards/rejected": 0.0,
"step": 2130
},
{
"epoch": 11.905027932960895,
"grad_norm": 0.863608474306918,
"learning_rate": 1.866917604762837e-07,
"logits/chosen": -3.491826057434082,
"logits/rejected": -3.6879568099975586,
"logps/chosen": -26.05544662475586,
"logps/rejected": -52.53486633300781,
"loss": 0.0987,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1738553047180176,
"rewards/margins": 3.1738553047180176,
"rewards/rejected": 0.0,
"step": 2131
},
{
"epoch": 11.910614525139664,
"grad_norm": 0.45433539472533707,
"learning_rate": 1.8621695840261386e-07,
"logits/chosen": -3.1819138526916504,
"logits/rejected": -3.31738018989563,
"logps/chosen": -0.6118383407592773,
"logps/rejected": -62.76938247680664,
"loss": 0.1046,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1953125,
"rewards/margins": 2.1953125,
"rewards/rejected": 0.0,
"step": 2132
},
{
"epoch": 11.916201117318435,
"grad_norm": 0.46966855925399337,
"learning_rate": 1.857426226664255e-07,
"logits/chosen": -3.4747657775878906,
"logits/rejected": -3.4594054222106934,
"logps/chosen": -3.7471964359283447,
"logps/rejected": -34.69524002075195,
"loss": 0.1022,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.984175205230713,
"rewards/margins": 2.984175205230713,
"rewards/rejected": 0.0,
"step": 2133
},
{
"epoch": 11.921787709497206,
"grad_norm": 1.0275136441730288,
"learning_rate": 1.85268753972666e-07,
"logits/chosen": -3.4865078926086426,
"logits/rejected": -3.601943016052246,
"logps/chosen": -9.224930763244629,
"logps/rejected": -97.36491394042969,
"loss": 0.1125,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7399888038635254,
"rewards/margins": 2.7399888038635254,
"rewards/rejected": 0.0,
"step": 2134
},
{
"epoch": 11.927374301675977,
"grad_norm": 0.5682073761034893,
"learning_rate": 1.847953530255887e-07,
"logits/chosen": -3.10056734085083,
"logits/rejected": -3.3187789916992188,
"logps/chosen": -28.744565963745117,
"logps/rejected": -72.4722900390625,
"loss": 0.1077,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4625368118286133,
"rewards/margins": 3.4625368118286133,
"rewards/rejected": 0.0,
"step": 2135
},
{
"epoch": 11.932960893854748,
"grad_norm": 1.6231014668350887,
"learning_rate": 1.8432242052875162e-07,
"logits/chosen": -3.2790393829345703,
"logits/rejected": -3.3346593379974365,
"logps/chosen": -4.857345104217529,
"logps/rejected": -88.18710327148438,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2367210388183594,
"rewards/margins": 3.2367210388183594,
"rewards/rejected": 0.0,
"step": 2136
},
{
"epoch": 11.938547486033519,
"grad_norm": 2.255345980568635,
"learning_rate": 1.8384995718501682e-07,
"logits/chosen": -3.2585246562957764,
"logits/rejected": -3.4492573738098145,
"logps/chosen": -0.42797067761421204,
"logps/rejected": -95.33000183105469,
"loss": 0.1403,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.56181800365448,
"rewards/margins": 1.56181800365448,
"rewards/rejected": 0.0,
"step": 2137
},
{
"epoch": 11.94413407821229,
"grad_norm": 1.023505762488989,
"learning_rate": 1.8337796369654885e-07,
"logits/chosen": -3.2255258560180664,
"logits/rejected": -3.2198212146759033,
"logps/chosen": -1.1346759796142578,
"logps/rejected": -33.88743591308594,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6318230628967285,
"rewards/margins": 2.6318230628967285,
"rewards/rejected": 0.0,
"step": 2138
},
{
"epoch": 11.949720670391061,
"grad_norm": 1.167077931668197,
"learning_rate": 1.829064407648141e-07,
"logits/chosen": -3.289717197418213,
"logits/rejected": -3.258373737335205,
"logps/chosen": -0.7531739473342896,
"logps/rejected": -72.3611068725586,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3357958793640137,
"rewards/margins": 2.3357958793640137,
"rewards/rejected": 0.0,
"step": 2139
},
{
"epoch": 11.955307262569832,
"grad_norm": 1.1151009280089255,
"learning_rate": 1.8243538909057955e-07,
"logits/chosen": -3.5916812419891357,
"logits/rejected": -3.5592756271362305,
"logps/chosen": -1.4563828706741333,
"logps/rejected": -48.035823822021484,
"loss": 0.0781,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.417919158935547,
"rewards/margins": 2.417919158935547,
"rewards/rejected": 0.0,
"step": 2140
},
{
"epoch": 11.955307262569832,
"eval_logits/chosen": -3.276703357696533,
"eval_logits/rejected": -3.4007294178009033,
"eval_logps/chosen": -29.58773422241211,
"eval_logps/rejected": -54.25297164916992,
"eval_loss": 0.9734715819358826,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.04823073744773865,
"eval_rewards/margins": 0.04823073744773865,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7264,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 2140
},
{
"epoch": 11.960893854748603,
"grad_norm": 1.9002789397551343,
"learning_rate": 1.8196480937391224e-07,
"logits/chosen": -3.371736526489258,
"logits/rejected": -3.214059829711914,
"logps/chosen": -10.563570022583008,
"logps/rejected": -41.80162048339844,
"loss": 0.0872,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8670542240142822,
"rewards/margins": 3.8670542240142822,
"rewards/rejected": 0.0,
"step": 2141
},
{
"epoch": 11.966480446927374,
"grad_norm": 0.5989836192601836,
"learning_rate": 1.8149470231417686e-07,
"logits/chosen": -3.370497703552246,
"logits/rejected": -3.3959856033325195,
"logps/chosen": -0.544033408164978,
"logps/rejected": -60.811676025390625,
"loss": 0.108,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1266367435455322,
"rewards/margins": 2.1266367435455322,
"rewards/rejected": 0.0,
"step": 2142
},
{
"epoch": 11.972067039106145,
"grad_norm": 1.4740528255270782,
"learning_rate": 1.8102506861003668e-07,
"logits/chosen": -3.4351704120635986,
"logits/rejected": -3.348499059677124,
"logps/chosen": -3.8457369804382324,
"logps/rejected": -48.02412414550781,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6251397132873535,
"rewards/margins": 2.6251397132873535,
"rewards/rejected": 0.0,
"step": 2143
},
{
"epoch": 11.977653631284916,
"grad_norm": 0.584902785218383,
"learning_rate": 1.805559089594509e-07,
"logits/chosen": -3.5643978118896484,
"logits/rejected": -3.5030927658081055,
"logps/chosen": -0.784415602684021,
"logps/rejected": -40.671119689941406,
"loss": 0.0994,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2984249591827393,
"rewards/margins": 2.2984249591827393,
"rewards/rejected": 0.0,
"step": 2144
},
{
"epoch": 11.983240223463687,
"grad_norm": 0.878489920214445,
"learning_rate": 1.800872240596743e-07,
"logits/chosen": -3.441546678543091,
"logits/rejected": -3.4520084857940674,
"logps/chosen": -3.2121951580047607,
"logps/rejected": -45.37566375732422,
"loss": 0.1155,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6495347023010254,
"rewards/margins": 2.6495347023010254,
"rewards/rejected": 0.0,
"step": 2145
},
{
"epoch": 11.988826815642458,
"grad_norm": 0.7719241937540747,
"learning_rate": 1.7961901460725604e-07,
"logits/chosen": -3.2772867679595947,
"logits/rejected": -3.240518569946289,
"logps/chosen": -16.807783126831055,
"logps/rejected": -37.623146057128906,
"loss": 0.0992,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.625868320465088,
"rewards/margins": 2.625868320465088,
"rewards/rejected": 0.0,
"step": 2146
},
{
"epoch": 11.994413407821229,
"grad_norm": 0.7699004627039944,
"learning_rate": 1.7915128129803887e-07,
"logits/chosen": -3.148439645767212,
"logits/rejected": -3.2155673503875732,
"logps/chosen": -17.334806442260742,
"logps/rejected": -39.368770599365234,
"loss": 0.1111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.160996437072754,
"rewards/margins": 2.160996437072754,
"rewards/rejected": 0.0,
"step": 2147
},
{
"epoch": 12.0,
"grad_norm": 0.37342330187378703,
"learning_rate": 1.7868402482715766e-07,
"logits/chosen": -3.322864294052124,
"logits/rejected": -3.254760503768921,
"logps/chosen": -0.28817659616470337,
"logps/rejected": -69.99283599853516,
"loss": 0.1228,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7711372375488281,
"rewards/margins": 1.7711372375488281,
"rewards/rejected": 0.0,
"step": 2148
},
{
"epoch": 12.005586592178771,
"grad_norm": 0.40378364708525616,
"learning_rate": 1.782172458890387e-07,
"logits/chosen": -3.5557851791381836,
"logits/rejected": -3.6075425148010254,
"logps/chosen": -1.883641242980957,
"logps/rejected": -52.34925079345703,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.914915084838867,
"rewards/margins": 2.914915084838867,
"rewards/rejected": 0.0,
"step": 2149
},
{
"epoch": 12.011173184357542,
"grad_norm": 0.5213657222741441,
"learning_rate": 1.7775094517739903e-07,
"logits/chosen": -3.403524160385132,
"logits/rejected": -3.5024876594543457,
"logps/chosen": -2.2176942825317383,
"logps/rejected": -42.349082946777344,
"loss": 0.0872,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.812084674835205,
"rewards/margins": 2.812084674835205,
"rewards/rejected": 0.0,
"step": 2150
},
{
"epoch": 12.016759776536313,
"grad_norm": 0.6444221789051341,
"learning_rate": 1.7728512338524388e-07,
"logits/chosen": -3.3550872802734375,
"logits/rejected": -3.214698076248169,
"logps/chosen": -5.084292411804199,
"logps/rejected": -62.02729797363281,
"loss": 0.0877,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.57625150680542,
"rewards/margins": 3.57625150680542,
"rewards/rejected": 0.0,
"step": 2151
},
{
"epoch": 12.022346368715084,
"grad_norm": 0.670653075788408,
"learning_rate": 1.7681978120486796e-07,
"logits/chosen": -3.240642547607422,
"logits/rejected": -3.228584051132202,
"logps/chosen": -12.051186561584473,
"logps/rejected": -27.182872772216797,
"loss": 0.1127,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1799678802490234,
"rewards/margins": 3.1799678802490234,
"rewards/rejected": 0.0,
"step": 2152
},
{
"epoch": 12.027932960893855,
"grad_norm": 0.4131077591617559,
"learning_rate": 1.763549193278524e-07,
"logits/chosen": -3.4236247539520264,
"logits/rejected": -3.319427490234375,
"logps/chosen": -0.9676990509033203,
"logps/rejected": -47.9969596862793,
"loss": 0.085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5235748291015625,
"rewards/margins": 2.5235748291015625,
"rewards/rejected": 0.0,
"step": 2153
},
{
"epoch": 12.033519553072626,
"grad_norm": 0.48318206462326607,
"learning_rate": 1.7589053844506474e-07,
"logits/chosen": -3.0199475288391113,
"logits/rejected": -3.0718603134155273,
"logps/chosen": -16.391725540161133,
"logps/rejected": -41.51329803466797,
"loss": 0.1158,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.944697380065918,
"rewards/margins": 3.944697380065918,
"rewards/rejected": 0.0,
"step": 2154
},
{
"epoch": 12.039106145251397,
"grad_norm": 0.4376015475095301,
"learning_rate": 1.7542663924665773e-07,
"logits/chosen": -3.278794765472412,
"logits/rejected": -3.369462490081787,
"logps/chosen": -29.321495056152344,
"logps/rejected": -74.3553695678711,
"loss": 0.0815,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.3270392417907715,
"rewards/margins": 4.3270392417907715,
"rewards/rejected": 0.0,
"step": 2155
},
{
"epoch": 12.044692737430168,
"grad_norm": 0.5227145883328304,
"learning_rate": 1.7496322242206811e-07,
"logits/chosen": -3.485487222671509,
"logits/rejected": -3.5119569301605225,
"logps/chosen": -0.7452935576438904,
"logps/rejected": -54.82063293457031,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.328976631164551,
"rewards/margins": 2.328976631164551,
"rewards/rejected": 0.0,
"step": 2156
},
{
"epoch": 12.050279329608939,
"grad_norm": 0.3774624657314431,
"learning_rate": 1.7450028866001587e-07,
"logits/chosen": -3.4001994132995605,
"logits/rejected": -3.350005865097046,
"logps/chosen": -9.71610164642334,
"logps/rejected": -54.90071105957031,
"loss": 0.1009,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7138869762420654,
"rewards/margins": 2.7138869762420654,
"rewards/rejected": 0.0,
"step": 2157
},
{
"epoch": 12.05586592178771,
"grad_norm": 0.36388899604965774,
"learning_rate": 1.7403783864850286e-07,
"logits/chosen": -3.3635263442993164,
"logits/rejected": -3.4219889640808105,
"logps/chosen": -1.1147568225860596,
"logps/rejected": -96.10930633544922,
"loss": 0.1021,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4907445907592773,
"rewards/margins": 2.4907445907592773,
"rewards/rejected": 0.0,
"step": 2158
},
{
"epoch": 12.061452513966481,
"grad_norm": 0.4618707459634858,
"learning_rate": 1.7357587307481263e-07,
"logits/chosen": -3.41139554977417,
"logits/rejected": -3.4588818550109863,
"logps/chosen": -0.8434318900108337,
"logps/rejected": -45.489280700683594,
"loss": 0.0951,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.515263795852661,
"rewards/margins": 2.515263795852661,
"rewards/rejected": 0.0,
"step": 2159
},
{
"epoch": 12.067039106145252,
"grad_norm": 0.38213439243139324,
"learning_rate": 1.7311439262550764e-07,
"logits/chosen": -3.2825281620025635,
"logits/rejected": -3.3209400177001953,
"logps/chosen": -1.6238597631454468,
"logps/rejected": -44.853328704833984,
"loss": 0.0932,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.048689842224121,
"rewards/margins": 3.048689842224121,
"rewards/rejected": 0.0,
"step": 2160
},
{
"epoch": 12.067039106145252,
"eval_logits/chosen": -3.26629638671875,
"eval_logits/rejected": -3.391310453414917,
"eval_logps/chosen": -29.958614349365234,
"eval_logps/rejected": -54.96844482421875,
"eval_loss": 0.9988755583763123,
"eval_rewards/accuracies": 0.6000000238418579,
"eval_rewards/chosen": 0.011142569594085217,
"eval_rewards/margins": 0.011142569594085217,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6987,
"eval_samples_per_second": 9.481,
"eval_steps_per_second": 0.306,
"step": 2160
},
{
"epoch": 12.072625698324023,
"grad_norm": 0.3799615625511484,
"learning_rate": 1.7265339798643047e-07,
"logits/chosen": -3.2318854331970215,
"logits/rejected": -3.3783950805664062,
"logps/chosen": -4.772999286651611,
"logps/rejected": -53.933372497558594,
"loss": 0.0986,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.409533739089966,
"rewards/margins": 2.409533739089966,
"rewards/rejected": 0.0,
"step": 2161
},
{
"epoch": 12.078212290502794,
"grad_norm": 0.4618446017978479,
"learning_rate": 1.721928898427012e-07,
"logits/chosen": -3.3426854610443115,
"logits/rejected": -3.2753188610076904,
"logps/chosen": -9.115558624267578,
"logps/rejected": -36.56756591796875,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.01203989982605,
"rewards/margins": 3.01203989982605,
"rewards/rejected": 0.0,
"step": 2162
},
{
"epoch": 12.083798882681565,
"grad_norm": 0.5644777010435362,
"learning_rate": 1.7173286887871684e-07,
"logits/chosen": -3.2985708713531494,
"logits/rejected": -3.3940656185150146,
"logps/chosen": -0.3955370783805847,
"logps/rejected": -80.76856231689453,
"loss": 0.1251,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7360330820083618,
"rewards/margins": 1.7360330820083618,
"rewards/rejected": 0.0,
"step": 2163
},
{
"epoch": 12.089385474860336,
"grad_norm": 0.9941797723077999,
"learning_rate": 1.712733357781505e-07,
"logits/chosen": -2.990569829940796,
"logits/rejected": -2.997157573699951,
"logps/chosen": -12.115921020507812,
"logps/rejected": -45.2780876159668,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1058850288391113,
"rewards/margins": 3.1058850288391113,
"rewards/rejected": 0.0,
"step": 2164
},
{
"epoch": 12.094972067039107,
"grad_norm": 0.37519598750511796,
"learning_rate": 1.708142912239502e-07,
"logits/chosen": -3.445521593093872,
"logits/rejected": -3.2191734313964844,
"logps/chosen": -7.310822486877441,
"logps/rejected": -33.084388732910156,
"loss": 0.0828,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.684077501296997,
"rewards/margins": 3.684077501296997,
"rewards/rejected": 0.0,
"step": 2165
},
{
"epoch": 12.100558659217878,
"grad_norm": 0.428464822203489,
"learning_rate": 1.7035573589833796e-07,
"logits/chosen": -3.2800114154815674,
"logits/rejected": -3.3751020431518555,
"logps/chosen": -1.4960510730743408,
"logps/rejected": -66.55987548828125,
"loss": 0.0788,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6144371032714844,
"rewards/margins": 2.6144371032714844,
"rewards/rejected": 0.0,
"step": 2166
},
{
"epoch": 12.106145251396647,
"grad_norm": 0.4087638051616013,
"learning_rate": 1.6989767048280857e-07,
"logits/chosen": -3.595515727996826,
"logits/rejected": -3.7190680503845215,
"logps/chosen": -5.6592116355896,
"logps/rejected": -55.2605094909668,
"loss": 0.0707,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3244736194610596,
"rewards/margins": 3.3244736194610596,
"rewards/rejected": 0.0,
"step": 2167
},
{
"epoch": 12.111731843575418,
"grad_norm": 0.35076977748399546,
"learning_rate": 1.6944009565812884e-07,
"logits/chosen": -3.5053322315216064,
"logits/rejected": -3.5027060508728027,
"logps/chosen": -10.765336036682129,
"logps/rejected": -28.595090866088867,
"loss": 0.0925,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.610229730606079,
"rewards/margins": 3.610229730606079,
"rewards/rejected": 0.0,
"step": 2168
},
{
"epoch": 12.11731843575419,
"grad_norm": 0.36300012325593645,
"learning_rate": 1.689830121043363e-07,
"logits/chosen": -3.317830801010132,
"logits/rejected": -3.3645992279052734,
"logps/chosen": -4.212851524353027,
"logps/rejected": -26.253808975219727,
"loss": 0.1199,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.080103874206543,
"rewards/margins": 3.080103874206543,
"rewards/rejected": 0.0,
"step": 2169
},
{
"epoch": 12.12290502793296,
"grad_norm": 0.8000529602967763,
"learning_rate": 1.68526420500739e-07,
"logits/chosen": -3.5707921981811523,
"logits/rejected": -3.5279901027679443,
"logps/chosen": -5.280650615692139,
"logps/rejected": -57.73506164550781,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6453027725219727,
"rewards/margins": 2.6453027725219727,
"rewards/rejected": 0.0,
"step": 2170
},
{
"epoch": 12.128491620111731,
"grad_norm": 0.41276710037417147,
"learning_rate": 1.6807032152591277e-07,
"logits/chosen": -3.3153977394104004,
"logits/rejected": -3.4051661491394043,
"logps/chosen": -2.830223560333252,
"logps/rejected": -44.051605224609375,
"loss": 0.0722,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.325134038925171,
"rewards/margins": 3.325134038925171,
"rewards/rejected": 0.0,
"step": 2171
},
{
"epoch": 12.134078212290502,
"grad_norm": 0.6782927881683671,
"learning_rate": 1.6761471585770231e-07,
"logits/chosen": -3.4325907230377197,
"logits/rejected": -3.4208483695983887,
"logps/chosen": -7.119795799255371,
"logps/rejected": -50.88927459716797,
"loss": 0.0809,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3752222061157227,
"rewards/margins": 3.3752222061157227,
"rewards/rejected": 0.0,
"step": 2172
},
{
"epoch": 12.139664804469273,
"grad_norm": 0.440721933774724,
"learning_rate": 1.6715960417321868e-07,
"logits/chosen": -3.5034115314483643,
"logits/rejected": -3.3925890922546387,
"logps/chosen": -1.6475284099578857,
"logps/rejected": -30.561010360717773,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1842782497406006,
"rewards/margins": 3.1842782497406006,
"rewards/rejected": 0.0,
"step": 2173
},
{
"epoch": 12.145251396648044,
"grad_norm": 0.4353003495233225,
"learning_rate": 1.667049871488389e-07,
"logits/chosen": -3.606475830078125,
"logits/rejected": -3.7171545028686523,
"logps/chosen": -1.5183839797973633,
"logps/rejected": -64.20674133300781,
"loss": 0.1239,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.561896562576294,
"rewards/margins": 2.561896562576294,
"rewards/rejected": 0.0,
"step": 2174
},
{
"epoch": 12.150837988826815,
"grad_norm": 0.5052694588263025,
"learning_rate": 1.662508654602048e-07,
"logits/chosen": -3.3743844032287598,
"logits/rejected": -3.4152326583862305,
"logps/chosen": -10.79573917388916,
"logps/rejected": -37.08635330200195,
"loss": 0.0945,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2049059867858887,
"rewards/margins": 3.2049059867858887,
"rewards/rejected": 0.0,
"step": 2175
},
{
"epoch": 12.156424581005586,
"grad_norm": 0.3844838734093418,
"learning_rate": 1.657972397822221e-07,
"logits/chosen": -3.402432680130005,
"logits/rejected": -3.5067174434661865,
"logps/chosen": -0.4482359290122986,
"logps/rejected": -54.469364166259766,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9763888120651245,
"rewards/margins": 1.9763888120651245,
"rewards/rejected": 0.0,
"step": 2176
},
{
"epoch": 12.162011173184357,
"grad_norm": 0.42614555087996847,
"learning_rate": 1.653441107890593e-07,
"logits/chosen": -3.187455654144287,
"logits/rejected": -3.3040640354156494,
"logps/chosen": -43.201087951660156,
"logps/rejected": -39.88694763183594,
"loss": 0.0965,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3057801723480225,
"rewards/margins": 3.3057801723480225,
"rewards/rejected": 0.0,
"step": 2177
},
{
"epoch": 12.167597765363128,
"grad_norm": 0.4477183244334658,
"learning_rate": 1.648914791541467e-07,
"logits/chosen": -3.285576581954956,
"logits/rejected": -3.524308919906616,
"logps/chosen": -1.0072684288024902,
"logps/rejected": -71.30606079101562,
"loss": 0.1106,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2392828464508057,
"rewards/margins": 2.2392828464508057,
"rewards/rejected": 0.0,
"step": 2178
},
{
"epoch": 12.1731843575419,
"grad_norm": 0.3276722931826065,
"learning_rate": 1.6443934555017592e-07,
"logits/chosen": -3.368210792541504,
"logits/rejected": -3.342644214630127,
"logps/chosen": -0.6389519572257996,
"logps/rejected": -31.776371002197266,
"loss": 0.1173,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2634239196777344,
"rewards/margins": 2.2634239196777344,
"rewards/rejected": 0.0,
"step": 2179
},
{
"epoch": 12.17877094972067,
"grad_norm": 1.4139689292348572,
"learning_rate": 1.6398771064909745e-07,
"logits/chosen": -3.434065103530884,
"logits/rejected": -3.4983153343200684,
"logps/chosen": -6.767345428466797,
"logps/rejected": -36.096797943115234,
"loss": 0.0887,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.419342041015625,
"rewards/margins": 3.419342041015625,
"rewards/rejected": 0.0,
"step": 2180
},
{
"epoch": 12.17877094972067,
"eval_logits/chosen": -3.2602896690368652,
"eval_logits/rejected": -3.3862922191619873,
"eval_logps/chosen": -30.069255828857422,
"eval_logps/rejected": -55.100860595703125,
"eval_loss": 1.0059175491333008,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": 7.850527617847547e-05,
"eval_rewards/margins": 7.850527617847547e-05,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 2180
},
{
"epoch": 12.184357541899441,
"grad_norm": 1.333253235266498,
"learning_rate": 1.6353657512212165e-07,
"logits/chosen": -3.5264642238616943,
"logits/rejected": -3.4460244178771973,
"logps/chosen": -4.27070426940918,
"logps/rejected": -65.77252960205078,
"loss": 0.0974,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6007394790649414,
"rewards/margins": 2.6007394790649414,
"rewards/rejected": 0.0,
"step": 2181
},
{
"epoch": 12.189944134078212,
"grad_norm": 0.45016741182450515,
"learning_rate": 1.6308593963971602e-07,
"logits/chosen": -3.445392370223999,
"logits/rejected": -3.356531858444214,
"logps/chosen": -10.348440170288086,
"logps/rejected": -49.000614166259766,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6431450843811035,
"rewards/margins": 3.6431450843811035,
"rewards/rejected": 0.0,
"step": 2182
},
{
"epoch": 12.195530726256983,
"grad_norm": 0.3812873777314003,
"learning_rate": 1.6263580487160522e-07,
"logits/chosen": -3.4518959522247314,
"logits/rejected": -3.5236244201660156,
"logps/chosen": -2.4868862628936768,
"logps/rejected": -35.648406982421875,
"loss": 0.0817,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.286860942840576,
"rewards/margins": 3.286860942840576,
"rewards/rejected": 0.0,
"step": 2183
},
{
"epoch": 12.201117318435754,
"grad_norm": 0.3654832108414279,
"learning_rate": 1.6218617148676967e-07,
"logits/chosen": -3.2200927734375,
"logits/rejected": -3.3731446266174316,
"logps/chosen": -1.4337997436523438,
"logps/rejected": -48.387725830078125,
"loss": 0.1456,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6443228721618652,
"rewards/margins": 2.6443228721618652,
"rewards/rejected": 0.0,
"step": 2184
},
{
"epoch": 12.206703910614525,
"grad_norm": 0.3652935435315875,
"learning_rate": 1.6173704015344462e-07,
"logits/chosen": -3.314303398132324,
"logits/rejected": -3.428551435470581,
"logps/chosen": -0.7392306327819824,
"logps/rejected": -63.2274055480957,
"loss": 0.1074,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.33054780960083,
"rewards/margins": 2.33054780960083,
"rewards/rejected": 0.0,
"step": 2185
},
{
"epoch": 12.212290502793296,
"grad_norm": 0.43048003457963097,
"learning_rate": 1.612884115391193e-07,
"logits/chosen": -3.207775354385376,
"logits/rejected": -3.3207106590270996,
"logps/chosen": -0.2809603214263916,
"logps/rejected": -66.76077270507812,
"loss": 0.0909,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8991446495056152,
"rewards/margins": 1.8991446495056152,
"rewards/rejected": 0.0,
"step": 2186
},
{
"epoch": 12.217877094972067,
"grad_norm": 0.406424038962057,
"learning_rate": 1.6084028631053547e-07,
"logits/chosen": -3.457875967025757,
"logits/rejected": -3.4442527294158936,
"logps/chosen": -14.073575019836426,
"logps/rejected": -75.53553771972656,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7035882472991943,
"rewards/margins": 3.7035882472991943,
"rewards/rejected": 0.0,
"step": 2187
},
{
"epoch": 12.223463687150838,
"grad_norm": 0.3896109584985441,
"learning_rate": 1.6039266513368756e-07,
"logits/chosen": -3.3706226348876953,
"logits/rejected": -3.5377938747406006,
"logps/chosen": -0.7984238862991333,
"logps/rejected": -86.73091125488281,
"loss": 0.0614,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.480565071105957,
"rewards/margins": 2.480565071105957,
"rewards/rejected": 0.0,
"step": 2188
},
{
"epoch": 12.22905027932961,
"grad_norm": 0.591910687950212,
"learning_rate": 1.5994554867381966e-07,
"logits/chosen": -3.386492967605591,
"logits/rejected": -3.1366889476776123,
"logps/chosen": -4.666633129119873,
"logps/rejected": -84.95759582519531,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.382068157196045,
"rewards/margins": 2.382068157196045,
"rewards/rejected": 0.0,
"step": 2189
},
{
"epoch": 12.23463687150838,
"grad_norm": 0.3807122108778428,
"learning_rate": 1.5949893759542697e-07,
"logits/chosen": -3.2206106185913086,
"logits/rejected": -3.2755048274993896,
"logps/chosen": -0.6934922933578491,
"logps/rejected": -24.947383880615234,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2132568359375,
"rewards/margins": 2.2132568359375,
"rewards/rejected": 0.0,
"step": 2190
},
{
"epoch": 12.240223463687151,
"grad_norm": 0.4109311447771417,
"learning_rate": 1.59052832562253e-07,
"logits/chosen": -3.467851400375366,
"logits/rejected": -3.601161003112793,
"logps/chosen": -1.1642745733261108,
"logps/rejected": -31.09759521484375,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.292839527130127,
"rewards/margins": 2.292839527130127,
"rewards/rejected": 0.0,
"step": 2191
},
{
"epoch": 12.245810055865922,
"grad_norm": 0.6455000209211309,
"learning_rate": 1.5860723423728912e-07,
"logits/chosen": -3.2219746112823486,
"logits/rejected": -3.1025068759918213,
"logps/chosen": -20.236587524414062,
"logps/rejected": -45.65630340576172,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6771585941314697,
"rewards/margins": 2.6771585941314697,
"rewards/rejected": 0.0,
"step": 2192
},
{
"epoch": 12.251396648044693,
"grad_norm": 0.4645737219040393,
"learning_rate": 1.5816214328277393e-07,
"logits/chosen": -3.4909138679504395,
"logits/rejected": -3.520278215408325,
"logps/chosen": -0.46695539355278015,
"logps/rejected": -47.601051330566406,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9735751152038574,
"rewards/margins": 1.9735751152038574,
"rewards/rejected": 0.0,
"step": 2193
},
{
"epoch": 12.256983240223464,
"grad_norm": 0.4397582110207123,
"learning_rate": 1.5771756036019184e-07,
"logits/chosen": -3.3240466117858887,
"logits/rejected": -3.5705275535583496,
"logps/chosen": -3.9083380699157715,
"logps/rejected": -42.57920455932617,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4163131713867188,
"rewards/margins": 3.4163131713867188,
"rewards/rejected": 0.0,
"step": 2194
},
{
"epoch": 12.262569832402235,
"grad_norm": 0.5262597377670216,
"learning_rate": 1.5727348613027218e-07,
"logits/chosen": -3.5049076080322266,
"logits/rejected": -3.6485280990600586,
"logps/chosen": -5.9262590408325195,
"logps/rejected": -37.675071716308594,
"loss": 0.1141,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1949758529663086,
"rewards/margins": 3.1949758529663086,
"rewards/rejected": 0.0,
"step": 2195
},
{
"epoch": 12.268156424581006,
"grad_norm": 0.416532771090321,
"learning_rate": 1.5682992125298845e-07,
"logits/chosen": -3.2953948974609375,
"logits/rejected": -3.368100881576538,
"logps/chosen": -0.5940881371498108,
"logps/rejected": -68.54150390625,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.233419418334961,
"rewards/margins": 2.233419418334961,
"rewards/rejected": 0.0,
"step": 2196
},
{
"epoch": 12.273743016759777,
"grad_norm": 0.5294551958350034,
"learning_rate": 1.5638686638755693e-07,
"logits/chosen": -3.361480474472046,
"logits/rejected": -3.4927079677581787,
"logps/chosen": -0.32955583930015564,
"logps/rejected": -58.109832763671875,
"loss": 0.1237,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5935286283493042,
"rewards/margins": 1.5935286283493042,
"rewards/rejected": 0.0,
"step": 2197
},
{
"epoch": 12.279329608938548,
"grad_norm": 0.44180384245356047,
"learning_rate": 1.5594432219243597e-07,
"logits/chosen": -3.352954149246216,
"logits/rejected": -3.4404146671295166,
"logps/chosen": -5.233445167541504,
"logps/rejected": -59.0228157043457,
"loss": 0.083,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1550750732421875,
"rewards/margins": 3.1550750732421875,
"rewards/rejected": 0.0,
"step": 2198
},
{
"epoch": 12.28491620111732,
"grad_norm": 0.4837843686530828,
"learning_rate": 1.555022893253255e-07,
"logits/chosen": -3.3813109397888184,
"logits/rejected": -3.3913769721984863,
"logps/chosen": -2.2488725185394287,
"logps/rejected": -39.50756072998047,
"loss": 0.0837,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2228925228118896,
"rewards/margins": 3.2228925228118896,
"rewards/rejected": 0.0,
"step": 2199
},
{
"epoch": 12.29050279329609,
"grad_norm": 0.38008790278698096,
"learning_rate": 1.5506076844316446e-07,
"logits/chosen": -3.2536447048187256,
"logits/rejected": -3.1121184825897217,
"logps/chosen": -3.3936192989349365,
"logps/rejected": -37.43659973144531,
"loss": 0.1067,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1023712158203125,
"rewards/margins": 3.1023712158203125,
"rewards/rejected": 0.0,
"step": 2200
},
{
"epoch": 12.29050279329609,
"eval_logits/chosen": -3.273606777191162,
"eval_logits/rejected": -3.3964200019836426,
"eval_logps/chosen": -30.072830200195312,
"eval_logps/rejected": -54.90795135498047,
"eval_loss": 1.008197546005249,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.000278729188721627,
"eval_rewards/margins": -0.000278729188721627,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7007,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 2200
},
{
"epoch": 12.296089385474861,
"grad_norm": 0.39689128481062536,
"learning_rate": 1.5461976020213202e-07,
"logits/chosen": -3.4687182903289795,
"logits/rejected": -3.4634604454040527,
"logps/chosen": -25.262069702148438,
"logps/rejected": -52.00890350341797,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5829715728759766,
"rewards/margins": 3.5829715728759766,
"rewards/rejected": 0.0,
"step": 2201
},
{
"epoch": 12.30167597765363,
"grad_norm": 0.553139249662218,
"learning_rate": 1.5417926525764474e-07,
"logits/chosen": -3.5827085971832275,
"logits/rejected": -3.570051670074463,
"logps/chosen": -1.1240618228912354,
"logps/rejected": -77.54736328125,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.891718864440918,
"rewards/margins": 2.891718864440918,
"rewards/rejected": 0.0,
"step": 2202
},
{
"epoch": 12.307262569832401,
"grad_norm": 0.43779318577112997,
"learning_rate": 1.5373928426435662e-07,
"logits/chosen": -3.643719434738159,
"logits/rejected": -3.5031206607818604,
"logps/chosen": -4.085589408874512,
"logps/rejected": -45.71474838256836,
"loss": 0.1176,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.81341552734375,
"rewards/margins": 2.81341552734375,
"rewards/rejected": 0.0,
"step": 2203
},
{
"epoch": 12.312849162011172,
"grad_norm": 1.1185142752538035,
"learning_rate": 1.532998178761577e-07,
"logits/chosen": -3.420722007751465,
"logits/rejected": -3.2349698543548584,
"logps/chosen": -2.943021774291992,
"logps/rejected": -63.329444885253906,
"loss": 0.1255,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.979187488555908,
"rewards/margins": 2.979187488555908,
"rewards/rejected": 0.0,
"step": 2204
},
{
"epoch": 12.318435754189943,
"grad_norm": 0.6442421529523719,
"learning_rate": 1.5286086674617337e-07,
"logits/chosen": -3.457326650619507,
"logits/rejected": -3.456826686859131,
"logps/chosen": -14.299467086791992,
"logps/rejected": -37.61988830566406,
"loss": 0.0908,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.071661472320557,
"rewards/margins": 4.071661472320557,
"rewards/rejected": 0.0,
"step": 2205
},
{
"epoch": 12.324022346368714,
"grad_norm": 1.8035445348879455,
"learning_rate": 1.5242243152676316e-07,
"logits/chosen": -3.164618492126465,
"logits/rejected": -3.2809653282165527,
"logps/chosen": -0.4628209173679352,
"logps/rejected": -85.3101577758789,
"loss": 0.1203,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0777363777160645,
"rewards/margins": 2.0777363777160645,
"rewards/rejected": 0.0,
"step": 2206
},
{
"epoch": 12.329608938547485,
"grad_norm": 1.2268782360442905,
"learning_rate": 1.5198451286951974e-07,
"logits/chosen": -3.4171974658966064,
"logits/rejected": -3.4579007625579834,
"logps/chosen": -2.159998893737793,
"logps/rejected": -32.2164306640625,
"loss": 0.1027,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5198140144348145,
"rewards/margins": 2.5198140144348145,
"rewards/rejected": 0.0,
"step": 2207
},
{
"epoch": 12.335195530726256,
"grad_norm": 0.3357318551819423,
"learning_rate": 1.5154711142526865e-07,
"logits/chosen": -3.39219331741333,
"logits/rejected": -3.5031139850616455,
"logps/chosen": -0.35548001527786255,
"logps/rejected": -82.95361328125,
"loss": 0.1168,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0930593013763428,
"rewards/margins": 2.0930593013763428,
"rewards/rejected": 0.0,
"step": 2208
},
{
"epoch": 12.340782122905027,
"grad_norm": 0.4103618008361082,
"learning_rate": 1.5111022784406575e-07,
"logits/chosen": -3.3791322708129883,
"logits/rejected": -3.258723735809326,
"logps/chosen": -0.6859988570213318,
"logps/rejected": -33.16255187988281,
"loss": 0.0831,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4664194583892822,
"rewards/margins": 2.4664194583892822,
"rewards/rejected": 0.0,
"step": 2209
},
{
"epoch": 12.346368715083798,
"grad_norm": 0.49985270363340084,
"learning_rate": 1.5067386277519827e-07,
"logits/chosen": -3.5391788482666016,
"logits/rejected": -3.704902410507202,
"logps/chosen": -9.990312576293945,
"logps/rejected": -43.68678283691406,
"loss": 0.0733,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8827545642852783,
"rewards/margins": 3.8827545642852783,
"rewards/rejected": 0.0,
"step": 2210
},
{
"epoch": 12.35195530726257,
"grad_norm": 0.368229749968429,
"learning_rate": 1.502380168671823e-07,
"logits/chosen": -3.7137563228607178,
"logits/rejected": -3.5952017307281494,
"logps/chosen": -3.101325750350952,
"logps/rejected": -38.811553955078125,
"loss": 0.0957,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0027360916137695,
"rewards/margins": 3.0027360916137695,
"rewards/rejected": 0.0,
"step": 2211
},
{
"epoch": 12.35754189944134,
"grad_norm": 1.620329892973557,
"learning_rate": 1.4980269076776246e-07,
"logits/chosen": -3.2845687866210938,
"logits/rejected": -3.166053295135498,
"logps/chosen": -0.5835750102996826,
"logps/rejected": -54.8746452331543,
"loss": 0.1073,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3022303581237793,
"rewards/margins": 2.3022303581237793,
"rewards/rejected": 0.0,
"step": 2212
},
{
"epoch": 12.363128491620111,
"grad_norm": 1.4818341578006364,
"learning_rate": 1.4936788512391086e-07,
"logits/chosen": -3.530430316925049,
"logits/rejected": -3.3934195041656494,
"logps/chosen": -0.3126392960548401,
"logps/rejected": -39.160057067871094,
"loss": 0.1285,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7503440380096436,
"rewards/margins": 1.7503440380096436,
"rewards/rejected": 0.0,
"step": 2213
},
{
"epoch": 12.368715083798882,
"grad_norm": 0.4162038412431759,
"learning_rate": 1.489336005818262e-07,
"logits/chosen": -3.309678316116333,
"logits/rejected": -3.298839569091797,
"logps/chosen": -3.787191867828369,
"logps/rejected": -68.62223815917969,
"loss": 0.1097,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5112075805664062,
"rewards/margins": 2.5112075805664062,
"rewards/rejected": 0.0,
"step": 2214
},
{
"epoch": 12.374301675977653,
"grad_norm": 0.3863444820312,
"learning_rate": 1.4849983778693247e-07,
"logits/chosen": -3.56412410736084,
"logits/rejected": -3.5686590671539307,
"logps/chosen": -1.782627820968628,
"logps/rejected": -30.9729061126709,
"loss": 0.0904,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.731541633605957,
"rewards/margins": 2.731541633605957,
"rewards/rejected": 0.0,
"step": 2215
},
{
"epoch": 12.379888268156424,
"grad_norm": 0.4670909926411719,
"learning_rate": 1.4806659738387845e-07,
"logits/chosen": -3.4526968002319336,
"logits/rejected": -3.2827701568603516,
"logps/chosen": -8.086905479431152,
"logps/rejected": -61.43003463745117,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2623023986816406,
"rewards/margins": 3.2623023986816406,
"rewards/rejected": 0.0,
"step": 2216
},
{
"epoch": 12.385474860335195,
"grad_norm": 0.5216245990843851,
"learning_rate": 1.47633880016537e-07,
"logits/chosen": -3.4484896659851074,
"logits/rejected": -3.606757640838623,
"logps/chosen": -1.9328851699829102,
"logps/rejected": -61.44506072998047,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7314529418945312,
"rewards/margins": 2.7314529418945312,
"rewards/rejected": 0.0,
"step": 2217
},
{
"epoch": 12.391061452513966,
"grad_norm": 0.39406110944684036,
"learning_rate": 1.472016863280024e-07,
"logits/chosen": -3.373232841491699,
"logits/rejected": -3.3617072105407715,
"logps/chosen": -1.397723913192749,
"logps/rejected": -33.85374069213867,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.688080310821533,
"rewards/margins": 2.688080310821533,
"rewards/rejected": 0.0,
"step": 2218
},
{
"epoch": 12.396648044692737,
"grad_norm": 0.5872700290830913,
"learning_rate": 1.4677001696059205e-07,
"logits/chosen": -3.268261671066284,
"logits/rejected": -3.110149621963501,
"logps/chosen": -49.851505279541016,
"logps/rejected": -38.20856475830078,
"loss": 0.0935,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.620477676391602,
"rewards/margins": 4.620477676391602,
"rewards/rejected": 0.0,
"step": 2219
},
{
"epoch": 12.402234636871508,
"grad_norm": 0.7493306505323228,
"learning_rate": 1.463388725558433e-07,
"logits/chosen": -3.0824813842773438,
"logits/rejected": -3.2189035415649414,
"logps/chosen": -1.364945650100708,
"logps/rejected": -90.79068756103516,
"loss": 0.0828,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3378281593322754,
"rewards/margins": 2.3378281593322754,
"rewards/rejected": 0.0,
"step": 2220
},
{
"epoch": 12.402234636871508,
"eval_logits/chosen": -3.263772964477539,
"eval_logits/rejected": -3.389078140258789,
"eval_logps/chosen": -29.9830379486084,
"eval_logps/rejected": -54.61565399169922,
"eval_loss": 1.0082952976226807,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": 0.00870051421225071,
"eval_rewards/margins": 0.00870051421225071,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7282,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 2220
},
{
"epoch": 12.40782122905028,
"grad_norm": 0.7994012378277839,
"learning_rate": 1.459082537545136e-07,
"logits/chosen": -3.4357802867889404,
"logits/rejected": -3.5319669246673584,
"logps/chosen": -10.154078483581543,
"logps/rejected": -25.528575897216797,
"loss": 0.1334,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.206284523010254,
"rewards/margins": 3.206284523010254,
"rewards/rejected": 0.0,
"step": 2221
},
{
"epoch": 12.41340782122905,
"grad_norm": 0.4414507972582413,
"learning_rate": 1.4547816119657906e-07,
"logits/chosen": -3.3708724975585938,
"logits/rejected": -3.5275392532348633,
"logps/chosen": -2.411421298980713,
"logps/rejected": -50.24624252319336,
"loss": 0.0913,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.822587013244629,
"rewards/margins": 2.822587013244629,
"rewards/rejected": 0.0,
"step": 2222
},
{
"epoch": 12.418994413407821,
"grad_norm": 2.28493377914312,
"learning_rate": 1.4504859552123388e-07,
"logits/chosen": -3.313854217529297,
"logits/rejected": -3.4889936447143555,
"logps/chosen": -3.167032480239868,
"logps/rejected": -65.09809112548828,
"loss": 0.1161,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.334043502807617,
"rewards/margins": 2.334043502807617,
"rewards/rejected": 0.0,
"step": 2223
},
{
"epoch": 12.424581005586592,
"grad_norm": 0.4608324185014521,
"learning_rate": 1.4461955736688913e-07,
"logits/chosen": -3.000213146209717,
"logits/rejected": -2.950758218765259,
"logps/chosen": -9.52685832977295,
"logps/rejected": -50.034454345703125,
"loss": 0.0919,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.961146354675293,
"rewards/margins": 2.961146354675293,
"rewards/rejected": 0.0,
"step": 2224
},
{
"epoch": 12.430167597765363,
"grad_norm": 0.7029217378704538,
"learning_rate": 1.4419104737117194e-07,
"logits/chosen": -3.2592275142669678,
"logits/rejected": -3.2010481357574463,
"logps/chosen": -8.226953506469727,
"logps/rejected": -46.574825286865234,
"loss": 0.0949,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6288509368896484,
"rewards/margins": 3.6288509368896484,
"rewards/rejected": 0.0,
"step": 2225
},
{
"epoch": 12.435754189944134,
"grad_norm": 1.8061372757179883,
"learning_rate": 1.4376306617092444e-07,
"logits/chosen": -3.177401542663574,
"logits/rejected": -3.282233238220215,
"logps/chosen": -1.1220273971557617,
"logps/rejected": -81.22883605957031,
"loss": 0.1089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7618532180786133,
"rewards/margins": 2.7618532180786133,
"rewards/rejected": 0.0,
"step": 2226
},
{
"epoch": 12.441340782122905,
"grad_norm": 0.4328558614661112,
"learning_rate": 1.4333561440220283e-07,
"logits/chosen": -3.417804002761841,
"logits/rejected": -3.6785597801208496,
"logps/chosen": -3.133924961090088,
"logps/rejected": -59.23695755004883,
"loss": 0.0753,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.837836742401123,
"rewards/margins": 2.837836742401123,
"rewards/rejected": 0.0,
"step": 2227
},
{
"epoch": 12.446927374301676,
"grad_norm": 0.3953785524468112,
"learning_rate": 1.4290869270027694e-07,
"logits/chosen": -2.9185352325439453,
"logits/rejected": -2.9892590045928955,
"logps/chosen": -30.449909210205078,
"logps/rejected": -82.8421630859375,
"loss": 0.0983,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1973934173583984,
"rewards/margins": 3.1973934173583984,
"rewards/rejected": 0.0,
"step": 2228
},
{
"epoch": 12.452513966480447,
"grad_norm": 0.7888034554133668,
"learning_rate": 1.4248230169962788e-07,
"logits/chosen": -3.5354950428009033,
"logits/rejected": -3.5087695121765137,
"logps/chosen": -1.1169776916503906,
"logps/rejected": -40.777870178222656,
"loss": 0.1596,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6863861083984375,
"rewards/margins": 2.6863861083984375,
"rewards/rejected": 0.0,
"step": 2229
},
{
"epoch": 12.458100558659218,
"grad_norm": 0.5604318581308836,
"learning_rate": 1.4205644203394905e-07,
"logits/chosen": -3.388920783996582,
"logits/rejected": -3.4258475303649902,
"logps/chosen": -1.7158160209655762,
"logps/rejected": -24.538524627685547,
"loss": 0.0931,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9679946899414062,
"rewards/margins": 2.9679946899414062,
"rewards/rejected": 0.0,
"step": 2230
},
{
"epoch": 12.46368715083799,
"grad_norm": 0.45429736244359015,
"learning_rate": 1.4163111433614367e-07,
"logits/chosen": -3.5283403396606445,
"logits/rejected": -3.4119808673858643,
"logps/chosen": -6.153657913208008,
"logps/rejected": -61.084014892578125,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.651522159576416,
"rewards/margins": 2.651522159576416,
"rewards/rejected": 0.0,
"step": 2231
},
{
"epoch": 12.46927374301676,
"grad_norm": 0.7081597777192936,
"learning_rate": 1.4120631923832433e-07,
"logits/chosen": -3.5893843173980713,
"logits/rejected": -3.5564889907836914,
"logps/chosen": -1.6719268560409546,
"logps/rejected": -39.64484405517578,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.543139696121216,
"rewards/margins": 2.543139696121216,
"rewards/rejected": 0.0,
"step": 2232
},
{
"epoch": 12.474860335195531,
"grad_norm": 1.7343545818297634,
"learning_rate": 1.407820573718123e-07,
"logits/chosen": -3.1008386611938477,
"logits/rejected": -3.3952138423919678,
"logps/chosen": -0.5091478824615479,
"logps/rejected": -76.26481628417969,
"loss": 0.1341,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9277145862579346,
"rewards/margins": 1.9277145862579346,
"rewards/rejected": 0.0,
"step": 2233
},
{
"epoch": 12.480446927374302,
"grad_norm": 0.9048386788322109,
"learning_rate": 1.403583293671362e-07,
"logits/chosen": -3.2435061931610107,
"logits/rejected": -3.1556649208068848,
"logps/chosen": -4.584765434265137,
"logps/rejected": -57.08665466308594,
"loss": 0.1173,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6094436645507812,
"rewards/margins": 2.6094436645507812,
"rewards/rejected": 0.0,
"step": 2234
},
{
"epoch": 12.486033519553073,
"grad_norm": 0.6167273944700344,
"learning_rate": 1.3993513585403137e-07,
"logits/chosen": -3.394244432449341,
"logits/rejected": -3.3959386348724365,
"logps/chosen": -0.9997613430023193,
"logps/rejected": -29.299489974975586,
"loss": 0.0986,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.547755241394043,
"rewards/margins": 2.547755241394043,
"rewards/rejected": 0.0,
"step": 2235
},
{
"epoch": 12.491620111731844,
"grad_norm": 1.2274380369073263,
"learning_rate": 1.3951247746143858e-07,
"logits/chosen": -3.270373821258545,
"logits/rejected": -3.2850375175476074,
"logps/chosen": -4.181968688964844,
"logps/rejected": -44.3603515625,
"loss": 0.0791,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7681868076324463,
"rewards/margins": 2.7681868076324463,
"rewards/rejected": 0.0,
"step": 2236
},
{
"epoch": 12.497206703910614,
"grad_norm": 0.4390033977859069,
"learning_rate": 1.3909035481750387e-07,
"logits/chosen": -3.350935459136963,
"logits/rejected": -3.4299426078796387,
"logps/chosen": -0.43021515011787415,
"logps/rejected": -37.922393798828125,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0837368965148926,
"rewards/margins": 2.0837368965148926,
"rewards/rejected": 0.0,
"step": 2237
},
{
"epoch": 12.502793296089386,
"grad_norm": 0.6648856335827535,
"learning_rate": 1.386687685495761e-07,
"logits/chosen": -3.3990509510040283,
"logits/rejected": -3.4585108757019043,
"logps/chosen": -0.5690751671791077,
"logps/rejected": -63.70109558105469,
"loss": 0.0868,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9457981586456299,
"rewards/margins": 1.9457981586456299,
"rewards/rejected": 0.0,
"step": 2238
},
{
"epoch": 12.508379888268156,
"grad_norm": 0.4289420248302391,
"learning_rate": 1.3824771928420798e-07,
"logits/chosen": -3.1850552558898926,
"logits/rejected": -3.1720054149627686,
"logps/chosen": -1.643048644065857,
"logps/rejected": -43.59968185424805,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7690720558166504,
"rewards/margins": 2.7690720558166504,
"rewards/rejected": 0.0,
"step": 2239
},
{
"epoch": 12.513966480446927,
"grad_norm": 0.3732674150498357,
"learning_rate": 1.3782720764715354e-07,
"logits/chosen": -3.4731249809265137,
"logits/rejected": -3.4296131134033203,
"logps/chosen": -0.17604614794254303,
"logps/rejected": -56.595123291015625,
"loss": 0.1253,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5817192792892456,
"rewards/margins": 1.5817192792892456,
"rewards/rejected": 0.0,
"step": 2240
},
{
"epoch": 12.513966480446927,
"eval_logits/chosen": -3.2657618522644043,
"eval_logits/rejected": -3.390824556350708,
"eval_logps/chosen": -30.071054458618164,
"eval_logps/rejected": -55.22324752807617,
"eval_loss": 1.0132746696472168,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.00010088682029163465,
"eval_rewards/margins": -0.00010088682029163465,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6892,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 2240
},
{
"epoch": 12.519553072625698,
"grad_norm": 0.7790209618270445,
"learning_rate": 1.37407234263368e-07,
"logits/chosen": -3.5272529125213623,
"logits/rejected": -3.371617555618286,
"logps/chosen": -2.3697447776794434,
"logps/rejected": -40.91530227661133,
"loss": 0.1124,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.849588394165039,
"rewards/margins": 2.849588394165039,
"rewards/rejected": 0.0,
"step": 2241
},
{
"epoch": 12.525139664804469,
"grad_norm": 1.0634639607925207,
"learning_rate": 1.3698779975700657e-07,
"logits/chosen": -3.3085103034973145,
"logits/rejected": -3.4670214653015137,
"logps/chosen": -2.182471513748169,
"logps/rejected": -73.75332641601562,
"loss": 0.0701,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5379199981689453,
"rewards/margins": 2.5379199981689453,
"rewards/rejected": 0.0,
"step": 2242
},
{
"epoch": 12.53072625698324,
"grad_norm": 2.0743889979879833,
"learning_rate": 1.3656890475142363e-07,
"logits/chosen": -3.423936367034912,
"logits/rejected": -3.3345353603363037,
"logps/chosen": -1.023058295249939,
"logps/rejected": -61.91513442993164,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6286544799804688,
"rewards/margins": 2.6286544799804688,
"rewards/rejected": 0.0,
"step": 2243
},
{
"epoch": 12.53631284916201,
"grad_norm": 0.39536921993716084,
"learning_rate": 1.3615054986917179e-07,
"logits/chosen": -3.265951156616211,
"logits/rejected": -3.256979465484619,
"logps/chosen": -11.967455863952637,
"logps/rejected": -78.19622802734375,
"loss": 0.1265,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3327698707580566,
"rewards/margins": 3.3327698707580566,
"rewards/rejected": 0.0,
"step": 2244
},
{
"epoch": 12.541899441340782,
"grad_norm": 1.8837471000302437,
"learning_rate": 1.3573273573200072e-07,
"logits/chosen": -3.3350565433502197,
"logits/rejected": -3.3303418159484863,
"logps/chosen": -1.2605342864990234,
"logps/rejected": -43.27762222290039,
"loss": 0.1406,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.628943681716919,
"rewards/margins": 2.628943681716919,
"rewards/rejected": 0.0,
"step": 2245
},
{
"epoch": 12.547486033519553,
"grad_norm": 0.4651543847476161,
"learning_rate": 1.3531546296085715e-07,
"logits/chosen": -3.2130534648895264,
"logits/rejected": -3.4951980113983154,
"logps/chosen": -1.6723864078521729,
"logps/rejected": -40.059288024902344,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.673156261444092,
"rewards/margins": 2.673156261444092,
"rewards/rejected": 0.0,
"step": 2246
},
{
"epoch": 12.553072625698324,
"grad_norm": 0.6138720038777592,
"learning_rate": 1.3489873217588205e-07,
"logits/chosen": -3.507383108139038,
"logits/rejected": -3.515362024307251,
"logps/chosen": -1.6765773296356201,
"logps/rejected": -67.03004455566406,
"loss": 0.1105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8080921173095703,
"rewards/margins": 2.8080921173095703,
"rewards/rejected": 0.0,
"step": 2247
},
{
"epoch": 12.558659217877095,
"grad_norm": 0.7722542782114533,
"learning_rate": 1.3448254399641206e-07,
"logits/chosen": -3.5299947261810303,
"logits/rejected": -3.495619058609009,
"logps/chosen": -14.777970314025879,
"logps/rejected": -62.427162170410156,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.873805522918701,
"rewards/margins": 3.873805522918701,
"rewards/rejected": 0.0,
"step": 2248
},
{
"epoch": 12.564245810055866,
"grad_norm": 0.41163074403727734,
"learning_rate": 1.340668990409768e-07,
"logits/chosen": -3.5440049171447754,
"logits/rejected": -3.554119825363159,
"logps/chosen": -3.7114877700805664,
"logps/rejected": -36.919189453125,
"loss": 0.0868,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0860977172851562,
"rewards/margins": 3.0860977172851562,
"rewards/rejected": 0.0,
"step": 2249
},
{
"epoch": 12.569832402234637,
"grad_norm": 0.4857325688319073,
"learning_rate": 1.336517979272987e-07,
"logits/chosen": -3.6649792194366455,
"logits/rejected": -3.497124433517456,
"logps/chosen": -1.410256266593933,
"logps/rejected": -38.30725860595703,
"loss": 0.0928,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8645949363708496,
"rewards/margins": 2.8645949363708496,
"rewards/rejected": 0.0,
"step": 2250
},
{
"epoch": 12.575418994413408,
"grad_norm": 0.42074558799151407,
"learning_rate": 1.3323724127229191e-07,
"logits/chosen": -3.2385005950927734,
"logits/rejected": -3.319828987121582,
"logps/chosen": -0.31787440180778503,
"logps/rejected": -84.67262268066406,
"loss": 0.0919,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.686431884765625,
"rewards/margins": 1.686431884765625,
"rewards/rejected": 0.0,
"step": 2251
},
{
"epoch": 12.581005586592179,
"grad_norm": 0.518755336213781,
"learning_rate": 1.3282322969206145e-07,
"logits/chosen": -2.9969799518585205,
"logits/rejected": -2.7330167293548584,
"logps/chosen": -18.212379455566406,
"logps/rejected": -60.43743133544922,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.314410448074341,
"rewards/margins": 3.314410448074341,
"rewards/rejected": 0.0,
"step": 2252
},
{
"epoch": 12.58659217877095,
"grad_norm": 0.38469720487246467,
"learning_rate": 1.3240976380190228e-07,
"logits/chosen": -3.724048614501953,
"logits/rejected": -3.806608200073242,
"logps/chosen": -2.855919361114502,
"logps/rejected": -27.25371551513672,
"loss": 0.096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6647424697875977,
"rewards/margins": 2.6647424697875977,
"rewards/rejected": 0.0,
"step": 2253
},
{
"epoch": 12.59217877094972,
"grad_norm": 0.9301604504936327,
"learning_rate": 1.319968442162983e-07,
"logits/chosen": -3.6221389770507812,
"logits/rejected": -3.578037977218628,
"logps/chosen": -7.850653171539307,
"logps/rejected": -50.1861572265625,
"loss": 0.1345,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3544840812683105,
"rewards/margins": 2.3544840812683105,
"rewards/rejected": 0.0,
"step": 2254
},
{
"epoch": 12.597765363128492,
"grad_norm": 0.41255745904527696,
"learning_rate": 1.3158447154892167e-07,
"logits/chosen": -3.176616907119751,
"logits/rejected": -3.4189412593841553,
"logps/chosen": -6.886829853057861,
"logps/rejected": -48.617454528808594,
"loss": 0.0909,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4059970378875732,
"rewards/margins": 3.4059970378875732,
"rewards/rejected": 0.0,
"step": 2255
},
{
"epoch": 12.603351955307263,
"grad_norm": 0.8510964543267436,
"learning_rate": 1.3117264641263138e-07,
"logits/chosen": -3.5320487022399902,
"logits/rejected": -3.3638176918029785,
"logps/chosen": -15.711894035339355,
"logps/rejected": -36.03840255737305,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.414116382598877,
"rewards/margins": 3.414116382598877,
"rewards/rejected": 0.0,
"step": 2256
},
{
"epoch": 12.608938547486034,
"grad_norm": 0.4011281051392134,
"learning_rate": 1.307613694194733e-07,
"logits/chosen": -3.2588086128234863,
"logits/rejected": -3.465930461883545,
"logps/chosen": -2.835179328918457,
"logps/rejected": -98.97663879394531,
"loss": 0.1149,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0574088096618652,
"rewards/margins": 3.0574088096618652,
"rewards/rejected": 0.0,
"step": 2257
},
{
"epoch": 12.614525139664805,
"grad_norm": 0.380910060419441,
"learning_rate": 1.3035064118067784e-07,
"logits/chosen": -3.342238426208496,
"logits/rejected": -3.1885581016540527,
"logps/chosen": -11.553668975830078,
"logps/rejected": -27.367694854736328,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.3055620193481445,
"rewards/margins": 4.3055620193481445,
"rewards/rejected": 0.0,
"step": 2258
},
{
"epoch": 12.620111731843576,
"grad_norm": 2.3723394233967396,
"learning_rate": 1.2994046230666067e-07,
"logits/chosen": -3.3266143798828125,
"logits/rejected": -3.239595651626587,
"logps/chosen": -1.1125938892364502,
"logps/rejected": -56.21533966064453,
"loss": 0.0855,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.46859073638916,
"rewards/margins": 2.46859073638916,
"rewards/rejected": 0.0,
"step": 2259
},
{
"epoch": 12.625698324022347,
"grad_norm": 0.40838752606489764,
"learning_rate": 1.2953083340702048e-07,
"logits/chosen": -3.198375940322876,
"logits/rejected": -3.3978817462921143,
"logps/chosen": -0.4842168688774109,
"logps/rejected": -103.18638610839844,
"loss": 0.106,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0734763145446777,
"rewards/margins": 2.0734763145446777,
"rewards/rejected": 0.0,
"step": 2260
},
{
"epoch": 12.625698324022347,
"eval_logits/chosen": -3.2565982341766357,
"eval_logits/rejected": -3.3827126026153564,
"eval_logps/chosen": -30.2506103515625,
"eval_logps/rejected": -55.432945251464844,
"eval_loss": 1.0206115245819092,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.01805674470961094,
"eval_rewards/margins": -0.01805674470961094,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7009,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 2260
},
{
"epoch": 12.631284916201118,
"grad_norm": 1.0435656728977338,
"learning_rate": 1.291217550905388e-07,
"logits/chosen": -3.0976061820983887,
"logits/rejected": -3.229267120361328,
"logps/chosen": -17.147357940673828,
"logps/rejected": -43.47943878173828,
"loss": 0.0882,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.126455307006836,
"rewards/margins": 3.126455307006836,
"rewards/rejected": 0.0,
"step": 2261
},
{
"epoch": 12.636871508379889,
"grad_norm": 0.6080571860345122,
"learning_rate": 1.2871322796517885e-07,
"logits/chosen": -3.450392007827759,
"logits/rejected": -3.6465983390808105,
"logps/chosen": -0.8923044204711914,
"logps/rejected": -60.1802864074707,
"loss": 0.0885,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3599853515625,
"rewards/margins": 2.3599853515625,
"rewards/rejected": 0.0,
"step": 2262
},
{
"epoch": 12.64245810055866,
"grad_norm": 0.9806283048446751,
"learning_rate": 1.283052526380848e-07,
"logits/chosen": -3.5287232398986816,
"logits/rejected": -3.418076992034912,
"logps/chosen": -0.7432454824447632,
"logps/rejected": -89.81851196289062,
"loss": 0.1434,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3667078018188477,
"rewards/margins": 2.3667078018188477,
"rewards/rejected": 0.0,
"step": 2263
},
{
"epoch": 12.64804469273743,
"grad_norm": 0.4467889703973007,
"learning_rate": 1.2789782971558044e-07,
"logits/chosen": -3.540520191192627,
"logits/rejected": -3.454953908920288,
"logps/chosen": -5.883553981781006,
"logps/rejected": -87.08979797363281,
"loss": 0.1113,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.080805778503418,
"rewards/margins": 3.080805778503418,
"rewards/rejected": 0.0,
"step": 2264
},
{
"epoch": 12.653631284916202,
"grad_norm": 0.4474252456646257,
"learning_rate": 1.2749095980316876e-07,
"logits/chosen": -3.170991897583008,
"logits/rejected": -3.045565605163574,
"logps/chosen": -4.559657573699951,
"logps/rejected": -65.60655212402344,
"loss": 0.1128,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.680539608001709,
"rewards/margins": 2.680539608001709,
"rewards/rejected": 0.0,
"step": 2265
},
{
"epoch": 12.659217877094973,
"grad_norm": 0.42478662256223054,
"learning_rate": 1.2708464350553134e-07,
"logits/chosen": -3.2953999042510986,
"logits/rejected": -3.3270702362060547,
"logps/chosen": -6.541364669799805,
"logps/rejected": -37.65180206298828,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0329442024230957,
"rewards/margins": 3.0329442024230957,
"rewards/rejected": 0.0,
"step": 2266
},
{
"epoch": 12.664804469273744,
"grad_norm": 0.4749508217388263,
"learning_rate": 1.26678881426526e-07,
"logits/chosen": -3.16400146484375,
"logits/rejected": -3.2337470054626465,
"logps/chosen": -0.4564622640609741,
"logps/rejected": -67.92149353027344,
"loss": 0.0947,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0054678916931152,
"rewards/margins": 2.0054678916931152,
"rewards/rejected": 0.0,
"step": 2267
},
{
"epoch": 12.670391061452515,
"grad_norm": 0.7671061924001767,
"learning_rate": 1.2627367416918783e-07,
"logits/chosen": -3.6778130531311035,
"logits/rejected": -3.4792375564575195,
"logps/chosen": -1.3930641412734985,
"logps/rejected": -71.66496276855469,
"loss": 0.0969,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.701521873474121,
"rewards/margins": 2.701521873474121,
"rewards/rejected": 0.0,
"step": 2268
},
{
"epoch": 12.675977653631286,
"grad_norm": 0.40730423026594653,
"learning_rate": 1.2586902233572688e-07,
"logits/chosen": -3.3053927421569824,
"logits/rejected": -3.326603412628174,
"logps/chosen": -2.162668228149414,
"logps/rejected": -41.80281448364258,
"loss": 0.0794,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7441184520721436,
"rewards/margins": 2.7441184520721436,
"rewards/rejected": 0.0,
"step": 2269
},
{
"epoch": 12.681564245810057,
"grad_norm": 0.9292182777541721,
"learning_rate": 1.2546492652752782e-07,
"logits/chosen": -3.3834147453308105,
"logits/rejected": -3.273434638977051,
"logps/chosen": -0.6231309771537781,
"logps/rejected": -61.2021484375,
"loss": 0.1277,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9415831565856934,
"rewards/margins": 1.9415831565856934,
"rewards/rejected": 0.0,
"step": 2270
},
{
"epoch": 12.687150837988828,
"grad_norm": 1.0463400068040039,
"learning_rate": 1.2506138734514897e-07,
"logits/chosen": -3.2636947631835938,
"logits/rejected": -3.337916135787964,
"logps/chosen": -8.795644760131836,
"logps/rejected": -53.609832763671875,
"loss": 0.094,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.550365447998047,
"rewards/margins": 3.550365447998047,
"rewards/rejected": 0.0,
"step": 2271
},
{
"epoch": 12.692737430167599,
"grad_norm": 0.42881096784347483,
"learning_rate": 1.2465840538832145e-07,
"logits/chosen": -3.402754306793213,
"logits/rejected": -3.577317476272583,
"logps/chosen": -8.689858436584473,
"logps/rejected": -49.614295959472656,
"loss": 0.1294,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9228854179382324,
"rewards/margins": 2.9228854179382324,
"rewards/rejected": 0.0,
"step": 2272
},
{
"epoch": 12.69832402234637,
"grad_norm": 1.6028214497713202,
"learning_rate": 1.2425598125594822e-07,
"logits/chosen": -3.5034029483795166,
"logits/rejected": -3.358663320541382,
"logps/chosen": -1.7330137491226196,
"logps/rejected": -42.20705795288086,
"loss": 0.1177,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.675685405731201,
"rewards/margins": 2.675685405731201,
"rewards/rejected": 0.0,
"step": 2273
},
{
"epoch": 12.703910614525139,
"grad_norm": 0.5623449768736747,
"learning_rate": 1.2385411554610304e-07,
"logits/chosen": -3.3504233360290527,
"logits/rejected": -3.390580177307129,
"logps/chosen": -5.595243453979492,
"logps/rejected": -42.18009948730469,
"loss": 0.0749,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6517512798309326,
"rewards/margins": 2.6517512798309326,
"rewards/rejected": 0.0,
"step": 2274
},
{
"epoch": 12.70949720670391,
"grad_norm": 0.686978584004586,
"learning_rate": 1.2345280885603037e-07,
"logits/chosen": -3.412993907928467,
"logits/rejected": -3.3707876205444336,
"logps/chosen": -1.2643492221832275,
"logps/rejected": -33.132591247558594,
"loss": 0.1081,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.125216484069824,
"rewards/margins": 2.125216484069824,
"rewards/rejected": 0.0,
"step": 2275
},
{
"epoch": 12.71508379888268,
"grad_norm": 0.34437154460398195,
"learning_rate": 1.2305206178214278e-07,
"logits/chosen": -3.456272602081299,
"logits/rejected": -3.514094114303589,
"logps/chosen": -0.8294367790222168,
"logps/rejected": -36.98713684082031,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3648805618286133,
"rewards/margins": 2.3648805618286133,
"rewards/rejected": 0.0,
"step": 2276
},
{
"epoch": 12.720670391061452,
"grad_norm": 0.4517748131352269,
"learning_rate": 1.2265187492002238e-07,
"logits/chosen": -3.48850154876709,
"logits/rejected": -3.203737497329712,
"logps/chosen": -0.5373401045799255,
"logps/rejected": -53.24457550048828,
"loss": 0.1338,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1982016563415527,
"rewards/margins": 2.1982016563415527,
"rewards/rejected": 0.0,
"step": 2277
},
{
"epoch": 12.726256983240223,
"grad_norm": 0.9630359520256496,
"learning_rate": 1.2225224886441782e-07,
"logits/chosen": -3.467395544052124,
"logits/rejected": -3.5238499641418457,
"logps/chosen": -6.839809894561768,
"logps/rejected": -32.9104118347168,
"loss": 0.0913,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6489503383636475,
"rewards/margins": 3.6489503383636475,
"rewards/rejected": 0.0,
"step": 2278
},
{
"epoch": 12.731843575418994,
"grad_norm": 0.5032997328960539,
"learning_rate": 1.2185318420924462e-07,
"logits/chosen": -3.3734967708587646,
"logits/rejected": -3.332409381866455,
"logps/chosen": -0.4818335771560669,
"logps/rejected": -82.39198303222656,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0698351860046387,
"rewards/margins": 2.0698351860046387,
"rewards/rejected": 0.0,
"step": 2279
},
{
"epoch": 12.737430167597765,
"grad_norm": 0.40200559581104356,
"learning_rate": 1.21454681547584e-07,
"logits/chosen": -3.384718656539917,
"logits/rejected": -3.402092695236206,
"logps/chosen": -4.479341506958008,
"logps/rejected": -38.881500244140625,
"loss": 0.0846,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.584264039993286,
"rewards/margins": 2.584264039993286,
"rewards/rejected": 0.0,
"step": 2280
},
{
"epoch": 12.737430167597765,
"eval_logits/chosen": -3.2632031440734863,
"eval_logits/rejected": -3.387631893157959,
"eval_logps/chosen": -30.60529136657715,
"eval_logps/rejected": -55.6917610168457,
"eval_loss": 1.0288151502609253,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.053524959832429886,
"eval_rewards/margins": -0.053524959832429886,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7164,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 2280
},
{
"epoch": 12.743016759776536,
"grad_norm": 0.46464037528578844,
"learning_rate": 1.2105674147168177e-07,
"logits/chosen": -3.386658191680908,
"logits/rejected": -3.4391565322875977,
"logps/chosen": -3.7839372158050537,
"logps/rejected": -32.402061462402344,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8743462562561035,
"rewards/margins": 2.8743462562561035,
"rewards/rejected": 0.0,
"step": 2281
},
{
"epoch": 12.748603351955307,
"grad_norm": 0.5061322845280998,
"learning_rate": 1.2065936457294785e-07,
"logits/chosen": -3.6489040851593018,
"logits/rejected": -3.607834815979004,
"logps/chosen": -5.303872585296631,
"logps/rejected": -50.91162872314453,
"loss": 0.0827,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.063188076019287,
"rewards/margins": 3.063188076019287,
"rewards/rejected": 0.0,
"step": 2282
},
{
"epoch": 12.754189944134078,
"grad_norm": 0.4144482019368093,
"learning_rate": 1.2026255144195507e-07,
"logits/chosen": -3.567878246307373,
"logits/rejected": -3.5910308361053467,
"logps/chosen": -2.981426239013672,
"logps/rejected": -48.233543395996094,
"loss": 0.09,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2606635093688965,
"rewards/margins": 3.2606635093688965,
"rewards/rejected": 0.0,
"step": 2283
},
{
"epoch": 12.759776536312849,
"grad_norm": 1.138898640724903,
"learning_rate": 1.1986630266843833e-07,
"logits/chosen": -3.4588491916656494,
"logits/rejected": -3.533299684524536,
"logps/chosen": -2.726400375366211,
"logps/rejected": -59.0339469909668,
"loss": 0.091,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9490859508514404,
"rewards/margins": 2.9490859508514404,
"rewards/rejected": 0.0,
"step": 2284
},
{
"epoch": 12.76536312849162,
"grad_norm": 0.5321998402593737,
"learning_rate": 1.1947061884129384e-07,
"logits/chosen": -3.5031237602233887,
"logits/rejected": -3.5955991744995117,
"logps/chosen": -0.3842340111732483,
"logps/rejected": -78.43391418457031,
"loss": 0.099,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9946914911270142,
"rewards/margins": 1.9946914911270142,
"rewards/rejected": 0.0,
"step": 2285
},
{
"epoch": 12.77094972067039,
"grad_norm": 0.5736551056197052,
"learning_rate": 1.1907550054857862e-07,
"logits/chosen": -3.4828784465789795,
"logits/rejected": -3.4474778175354004,
"logps/chosen": -12.686964988708496,
"logps/rejected": -37.39101028442383,
"loss": 0.1182,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.75471830368042,
"rewards/margins": 3.75471830368042,
"rewards/rejected": 0.0,
"step": 2286
},
{
"epoch": 12.776536312849162,
"grad_norm": 0.41298213307299225,
"learning_rate": 1.186809483775082e-07,
"logits/chosen": -3.2222204208374023,
"logits/rejected": -3.294398784637451,
"logps/chosen": -5.704745292663574,
"logps/rejected": -47.13985061645508,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5716466903686523,
"rewards/margins": 3.5716466903686523,
"rewards/rejected": 0.0,
"step": 2287
},
{
"epoch": 12.782122905027933,
"grad_norm": 0.4950033879661093,
"learning_rate": 1.182869629144579e-07,
"logits/chosen": -3.2026522159576416,
"logits/rejected": -3.1474239826202393,
"logps/chosen": -12.411076545715332,
"logps/rejected": -49.24952697753906,
"loss": 0.0871,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.957362174987793,
"rewards/margins": 2.957362174987793,
"rewards/rejected": 0.0,
"step": 2288
},
{
"epoch": 12.787709497206704,
"grad_norm": 0.3400573035366642,
"learning_rate": 1.1789354474496e-07,
"logits/chosen": -3.648020029067993,
"logits/rejected": -3.613192319869995,
"logps/chosen": -1.627375841140747,
"logps/rejected": -44.16062927246094,
"loss": 0.121,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3433022499084473,
"rewards/margins": 2.3433022499084473,
"rewards/rejected": 0.0,
"step": 2289
},
{
"epoch": 12.793296089385475,
"grad_norm": 0.454376022046106,
"learning_rate": 1.1750069445370408e-07,
"logits/chosen": -3.5968801975250244,
"logits/rejected": -3.6151320934295654,
"logps/chosen": -7.9256272315979,
"logps/rejected": -36.523895263671875,
"loss": 0.1153,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9409613609313965,
"rewards/margins": 3.9409613609313965,
"rewards/rejected": 0.0,
"step": 2290
},
{
"epoch": 12.798882681564246,
"grad_norm": 1.0140073801191927,
"learning_rate": 1.1710841262453563e-07,
"logits/chosen": -3.7124454975128174,
"logits/rejected": -3.6189231872558594,
"logps/chosen": -1.2007167339324951,
"logps/rejected": -38.000518798828125,
"loss": 0.0988,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7239294052124023,
"rewards/margins": 2.7239294052124023,
"rewards/rejected": 0.0,
"step": 2291
},
{
"epoch": 12.804469273743017,
"grad_norm": 0.3992200071958301,
"learning_rate": 1.1671669984045529e-07,
"logits/chosen": -3.361949920654297,
"logits/rejected": -3.4557816982269287,
"logps/chosen": -0.41469088196754456,
"logps/rejected": -96.77440643310547,
"loss": 0.1227,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8790662288665771,
"rewards/margins": 1.8790662288665771,
"rewards/rejected": 0.0,
"step": 2292
},
{
"epoch": 12.810055865921788,
"grad_norm": 0.45543717625752306,
"learning_rate": 1.1632555668361799e-07,
"logits/chosen": -3.2611303329467773,
"logits/rejected": -3.474437952041626,
"logps/chosen": -0.6212449073791504,
"logps/rejected": -107.97794342041016,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.037877321243286,
"rewards/margins": 2.037877321243286,
"rewards/rejected": 0.0,
"step": 2293
},
{
"epoch": 12.815642458100559,
"grad_norm": 0.39206772516047084,
"learning_rate": 1.1593498373533195e-07,
"logits/chosen": -3.4295456409454346,
"logits/rejected": -3.521411895751953,
"logps/chosen": -4.333605766296387,
"logps/rejected": -51.17797088623047,
"loss": 0.0834,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9523544311523438,
"rewards/margins": 3.9523544311523438,
"rewards/rejected": 0.0,
"step": 2294
},
{
"epoch": 12.82122905027933,
"grad_norm": 0.37064044990678563,
"learning_rate": 1.1554498157605863e-07,
"logits/chosen": -3.3964834213256836,
"logits/rejected": -3.3317043781280518,
"logps/chosen": -1.1701773405075073,
"logps/rejected": -41.469482421875,
"loss": 0.1057,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.688302516937256,
"rewards/margins": 2.688302516937256,
"rewards/rejected": 0.0,
"step": 2295
},
{
"epoch": 12.8268156424581,
"grad_norm": 0.5249714895001094,
"learning_rate": 1.1515555078541012e-07,
"logits/chosen": -3.524517774581909,
"logits/rejected": -3.5933830738067627,
"logps/chosen": -6.073652267456055,
"logps/rejected": -40.80901336669922,
"loss": 0.0896,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.266629695892334,
"rewards/margins": 2.266629695892334,
"rewards/rejected": 0.0,
"step": 2296
},
{
"epoch": 12.832402234636872,
"grad_norm": 0.4645679581551184,
"learning_rate": 1.1476669194215033e-07,
"logits/chosen": -3.3193717002868652,
"logits/rejected": -3.1490135192871094,
"logps/chosen": -1.5837781429290771,
"logps/rejected": -48.53924560546875,
"loss": 0.082,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7303173542022705,
"rewards/margins": 2.7303173542022705,
"rewards/rejected": 0.0,
"step": 2297
},
{
"epoch": 12.837988826815643,
"grad_norm": 1.3326266721744495,
"learning_rate": 1.143784056241926e-07,
"logits/chosen": -3.339778184890747,
"logits/rejected": -3.4562277793884277,
"logps/chosen": -0.9490244388580322,
"logps/rejected": -43.138038635253906,
"loss": 0.0887,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.599933385848999,
"rewards/margins": 2.599933385848999,
"rewards/rejected": 0.0,
"step": 2298
},
{
"epoch": 12.843575418994414,
"grad_norm": 0.41659324142136256,
"learning_rate": 1.1399069240859955e-07,
"logits/chosen": -3.285973310470581,
"logits/rejected": -3.210517406463623,
"logps/chosen": -1.9052813053131104,
"logps/rejected": -40.415252685546875,
"loss": 0.0824,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0174355506896973,
"rewards/margins": 3.0174355506896973,
"rewards/rejected": 0.0,
"step": 2299
},
{
"epoch": 12.849162011173185,
"grad_norm": 0.7115066262589483,
"learning_rate": 1.1360355287158212e-07,
"logits/chosen": -3.2919743061065674,
"logits/rejected": -3.3648951053619385,
"logps/chosen": -4.958246231079102,
"logps/rejected": -43.85498809814453,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.034874439239502,
"rewards/margins": 3.034874439239502,
"rewards/rejected": 0.0,
"step": 2300
},
{
"epoch": 12.849162011173185,
"eval_logits/chosen": -3.2657992839813232,
"eval_logits/rejected": -3.389568328857422,
"eval_logps/chosen": -30.225671768188477,
"eval_logps/rejected": -55.22808837890625,
"eval_loss": 1.0184718370437622,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.01556316576898098,
"eval_rewards/margins": -0.01556316576898098,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7301,
"eval_samples_per_second": 9.471,
"eval_steps_per_second": 0.306,
"step": 2300
},
{
"epoch": 12.854748603351956,
"grad_norm": 0.44602212592943713,
"learning_rate": 1.1321698758849851e-07,
"logits/chosen": -3.4965732097625732,
"logits/rejected": -3.648874282836914,
"logps/chosen": -1.9803823232650757,
"logps/rejected": -81.80268859863281,
"loss": 0.1273,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.669301986694336,
"rewards/margins": 2.669301986694336,
"rewards/rejected": 0.0,
"step": 2301
},
{
"epoch": 12.860335195530727,
"grad_norm": 0.460457654972344,
"learning_rate": 1.128309971338537e-07,
"logits/chosen": -3.4491028785705566,
"logits/rejected": -3.6618459224700928,
"logps/chosen": -0.38219979405403137,
"logps/rejected": -105.92637634277344,
"loss": 0.1088,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5562540292739868,
"rewards/margins": 1.5562540292739868,
"rewards/rejected": 0.0,
"step": 2302
},
{
"epoch": 12.865921787709498,
"grad_norm": 0.5539161184618399,
"learning_rate": 1.1244558208129795e-07,
"logits/chosen": -3.259394407272339,
"logits/rejected": -3.2880139350891113,
"logps/chosen": -0.3285684287548065,
"logps/rejected": -31.84539222717285,
"loss": 0.1353,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7426209449768066,
"rewards/margins": 1.7426209449768066,
"rewards/rejected": 0.0,
"step": 2303
},
{
"epoch": 12.871508379888269,
"grad_norm": 0.44330712618631835,
"learning_rate": 1.1206074300362723e-07,
"logits/chosen": -3.51483416557312,
"logits/rejected": -3.612828254699707,
"logps/chosen": -0.490126371383667,
"logps/rejected": -43.235599517822266,
"loss": 0.0939,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2221014499664307,
"rewards/margins": 2.2221014499664307,
"rewards/rejected": 0.0,
"step": 2304
},
{
"epoch": 12.87709497206704,
"grad_norm": 0.4274635041965022,
"learning_rate": 1.1167648047278033e-07,
"logits/chosen": -3.399958372116089,
"logits/rejected": -3.4519026279449463,
"logps/chosen": -0.6736325025558472,
"logps/rejected": -34.915618896484375,
"loss": 0.0786,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.456824541091919,
"rewards/margins": 2.456824541091919,
"rewards/rejected": 0.0,
"step": 2305
},
{
"epoch": 12.88268156424581,
"grad_norm": 1.1598479550425806,
"learning_rate": 1.1129279505984035e-07,
"logits/chosen": -3.4720335006713867,
"logits/rejected": -3.435961961746216,
"logps/chosen": -2.019890308380127,
"logps/rejected": -50.04304504394531,
"loss": 0.0854,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0951738357543945,
"rewards/margins": 2.0951738357543945,
"rewards/rejected": 0.0,
"step": 2306
},
{
"epoch": 12.888268156424582,
"grad_norm": 0.4425863194920016,
"learning_rate": 1.1090968733503158e-07,
"logits/chosen": -3.3264081478118896,
"logits/rejected": -3.538029193878174,
"logps/chosen": -0.9513140916824341,
"logps/rejected": -86.88038635253906,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8230986595153809,
"rewards/margins": 1.8230986595153809,
"rewards/rejected": 0.0,
"step": 2307
},
{
"epoch": 12.893854748603353,
"grad_norm": 0.6762900506062927,
"learning_rate": 1.1052715786772077e-07,
"logits/chosen": -3.3086395263671875,
"logits/rejected": -3.4097607135772705,
"logps/chosen": -1.1275460720062256,
"logps/rejected": -51.35859680175781,
"loss": 0.1287,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0423707962036133,
"rewards/margins": 2.0423707962036133,
"rewards/rejected": 0.0,
"step": 2308
},
{
"epoch": 12.899441340782122,
"grad_norm": 1.1668855063735013,
"learning_rate": 1.1014520722641473e-07,
"logits/chosen": -3.527906894683838,
"logits/rejected": -3.312272548675537,
"logps/chosen": -15.089126586914062,
"logps/rejected": -22.657346725463867,
"loss": 0.0978,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.603621244430542,
"rewards/margins": 2.603621244430542,
"rewards/rejected": 0.0,
"step": 2309
},
{
"epoch": 12.905027932960895,
"grad_norm": 0.4243446357729765,
"learning_rate": 1.0976383597876004e-07,
"logits/chosen": -3.4055118560791016,
"logits/rejected": -3.5252139568328857,
"logps/chosen": -3.22821044921875,
"logps/rejected": -42.4783935546875,
"loss": 0.1049,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2797393798828125,
"rewards/margins": 3.2797393798828125,
"rewards/rejected": 0.0,
"step": 2310
},
{
"epoch": 12.910614525139664,
"grad_norm": 0.5526143207246139,
"learning_rate": 1.0938304469154236e-07,
"logits/chosen": -3.4045262336730957,
"logits/rejected": -3.383348226547241,
"logps/chosen": -1.4761393070220947,
"logps/rejected": -56.56805419921875,
"loss": 0.1181,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.531733512878418,
"rewards/margins": 2.531733512878418,
"rewards/rejected": 0.0,
"step": 2311
},
{
"epoch": 12.916201117318435,
"grad_norm": 0.5165698942231859,
"learning_rate": 1.0900283393068538e-07,
"logits/chosen": -3.5687220096588135,
"logits/rejected": -3.64994740486145,
"logps/chosen": -0.9706995487213135,
"logps/rejected": -50.04160690307617,
"loss": 0.0942,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.31058406829834,
"rewards/margins": 2.31058406829834,
"rewards/rejected": 0.0,
"step": 2312
},
{
"epoch": 12.921787709497206,
"grad_norm": 1.3137916392199143,
"learning_rate": 1.0862320426124987e-07,
"logits/chosen": -3.4807398319244385,
"logits/rejected": -3.5688562393188477,
"logps/chosen": -3.280137538909912,
"logps/rejected": -51.393150329589844,
"loss": 0.1115,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0259995460510254,
"rewards/margins": 3.0259995460510254,
"rewards/rejected": 0.0,
"step": 2313
},
{
"epoch": 12.927374301675977,
"grad_norm": 0.5793005922518515,
"learning_rate": 1.0824415624743316e-07,
"logits/chosen": -3.4329700469970703,
"logits/rejected": -3.5415170192718506,
"logps/chosen": -5.662870407104492,
"logps/rejected": -44.23683547973633,
"loss": 0.0636,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.847524642944336,
"rewards/margins": 3.847524642944336,
"rewards/rejected": 0.0,
"step": 2314
},
{
"epoch": 12.932960893854748,
"grad_norm": 0.688741724035675,
"learning_rate": 1.0786569045256844e-07,
"logits/chosen": -3.2243945598602295,
"logits/rejected": -3.1893362998962402,
"logps/chosen": -36.53736877441406,
"logps/rejected": -35.05921173095703,
"loss": 0.0955,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.05674934387207,
"rewards/margins": 4.05674934387207,
"rewards/rejected": 0.0,
"step": 2315
},
{
"epoch": 12.938547486033519,
"grad_norm": 0.7024087167910086,
"learning_rate": 1.0748780743912273e-07,
"logits/chosen": -3.17535662651062,
"logits/rejected": -3.2973382472991943,
"logps/chosen": -0.6445251703262329,
"logps/rejected": -37.588218688964844,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3380391597747803,
"rewards/margins": 2.3380391597747803,
"rewards/rejected": 0.0,
"step": 2316
},
{
"epoch": 12.94413407821229,
"grad_norm": 0.43512439717326357,
"learning_rate": 1.071105077686979e-07,
"logits/chosen": -3.64111328125,
"logits/rejected": -3.4275224208831787,
"logps/chosen": -0.5026963949203491,
"logps/rejected": -82.1552505493164,
"loss": 0.0932,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1187191009521484,
"rewards/margins": 2.1187191009521484,
"rewards/rejected": 0.0,
"step": 2317
},
{
"epoch": 12.949720670391061,
"grad_norm": 0.6007210329834282,
"learning_rate": 1.0673379200202831e-07,
"logits/chosen": -3.512960195541382,
"logits/rejected": -3.60532283782959,
"logps/chosen": -0.7710492610931396,
"logps/rejected": -56.94520568847656,
"loss": 0.0759,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1803629398345947,
"rewards/margins": 2.1803629398345947,
"rewards/rejected": 0.0,
"step": 2318
},
{
"epoch": 12.955307262569832,
"grad_norm": 0.7332608521876424,
"learning_rate": 1.0635766069898067e-07,
"logits/chosen": -3.2716264724731445,
"logits/rejected": -3.2488861083984375,
"logps/chosen": -1.935168981552124,
"logps/rejected": -39.56070327758789,
"loss": 0.0909,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5785341262817383,
"rewards/margins": 2.5785341262817383,
"rewards/rejected": 0.0,
"step": 2319
},
{
"epoch": 12.960893854748603,
"grad_norm": 0.8285272950785425,
"learning_rate": 1.0598211441855309e-07,
"logits/chosen": -3.2604289054870605,
"logits/rejected": -3.347188949584961,
"logps/chosen": -0.8810999989509583,
"logps/rejected": -73.14479064941406,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.319912910461426,
"rewards/margins": 2.319912910461426,
"rewards/rejected": 0.0,
"step": 2320
},
{
"epoch": 12.960893854748603,
"eval_logits/chosen": -3.2581944465637207,
"eval_logits/rejected": -3.3839733600616455,
"eval_logps/chosen": -30.50138282775879,
"eval_logps/rejected": -55.073570251464844,
"eval_loss": 1.0289722681045532,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.04313389211893082,
"eval_rewards/margins": -0.04313389211893082,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6713,
"eval_samples_per_second": 9.488,
"eval_steps_per_second": 0.306,
"step": 2320
},
{
"epoch": 12.966480446927374,
"grad_norm": 0.5036770849735132,
"learning_rate": 1.0560715371887424e-07,
"logits/chosen": -3.299856185913086,
"logits/rejected": -3.561100482940674,
"logps/chosen": -0.48619604110717773,
"logps/rejected": -47.678043365478516,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1039106845855713,
"rewards/margins": 2.1039106845855713,
"rewards/rejected": 0.0,
"step": 2321
},
{
"epoch": 12.972067039106145,
"grad_norm": 0.4041259081197625,
"learning_rate": 1.0523277915720252e-07,
"logits/chosen": -3.390162229537964,
"logits/rejected": -3.370681047439575,
"logps/chosen": -0.8836253881454468,
"logps/rejected": -79.54744720458984,
"loss": 0.1099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.232394218444824,
"rewards/margins": 2.232394218444824,
"rewards/rejected": 0.0,
"step": 2322
},
{
"epoch": 12.977653631284916,
"grad_norm": 0.4098458996762994,
"learning_rate": 1.0485899128992498e-07,
"logits/chosen": -3.3942453861236572,
"logits/rejected": -3.4706342220306396,
"logps/chosen": -21.32162857055664,
"logps/rejected": -35.426292419433594,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.139532089233398,
"rewards/margins": 4.139532089233398,
"rewards/rejected": 0.0,
"step": 2323
},
{
"epoch": 12.983240223463687,
"grad_norm": 0.5340033200407157,
"learning_rate": 1.0448579067255747e-07,
"logits/chosen": -3.272456407546997,
"logits/rejected": -3.3082878589630127,
"logps/chosen": -6.20727014541626,
"logps/rejected": -62.77374267578125,
"loss": 0.0906,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2633891105651855,
"rewards/margins": 3.2633891105651855,
"rewards/rejected": 0.0,
"step": 2324
},
{
"epoch": 12.988826815642458,
"grad_norm": 0.45993193104965063,
"learning_rate": 1.0411317785974194e-07,
"logits/chosen": -3.577439785003662,
"logits/rejected": -3.5963082313537598,
"logps/chosen": -2.768301486968994,
"logps/rejected": -51.79939270019531,
"loss": 0.0899,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9437942504882812,
"rewards/margins": 2.9437942504882812,
"rewards/rejected": 0.0,
"step": 2325
},
{
"epoch": 12.994413407821229,
"grad_norm": 0.44211725643679217,
"learning_rate": 1.0374115340524787e-07,
"logits/chosen": -3.2956740856170654,
"logits/rejected": -3.453584671020508,
"logps/chosen": -2.895425796508789,
"logps/rejected": -37.691078186035156,
"loss": 0.0739,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1172409057617188,
"rewards/margins": 3.1172409057617188,
"rewards/rejected": 0.0,
"step": 2326
},
{
"epoch": 13.0,
"grad_norm": 0.3660531302685357,
"learning_rate": 1.0336971786196963e-07,
"logits/chosen": -3.4428699016571045,
"logits/rejected": -3.473644733428955,
"logps/chosen": -0.23429515957832336,
"logps/rejected": -100.61065673828125,
"loss": 0.1103,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6497079133987427,
"rewards/margins": 1.6497079133987427,
"rewards/rejected": 0.0,
"step": 2327
},
{
"epoch": 13.005586592178771,
"grad_norm": 0.6408412166809896,
"learning_rate": 1.0299887178192668e-07,
"logits/chosen": -3.5169355869293213,
"logits/rejected": -3.649749755859375,
"logps/chosen": -0.6415660977363586,
"logps/rejected": -124.60030364990234,
"loss": 0.1416,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.137908458709717,
"rewards/margins": 2.137908458709717,
"rewards/rejected": 0.0,
"step": 2328
},
{
"epoch": 13.011173184357542,
"grad_norm": 0.4911671107424767,
"learning_rate": 1.026286157162623e-07,
"logits/chosen": -3.168640375137329,
"logits/rejected": -3.291517496109009,
"logps/chosen": -1.6391764879226685,
"logps/rejected": -44.04465866088867,
"loss": 0.0824,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.718446969985962,
"rewards/margins": 2.718446969985962,
"rewards/rejected": 0.0,
"step": 2329
},
{
"epoch": 13.016759776536313,
"grad_norm": 0.36539337437173375,
"learning_rate": 1.0225895021524288e-07,
"logits/chosen": -3.586638927459717,
"logits/rejected": -3.3814616203308105,
"logps/chosen": -1.656554102897644,
"logps/rejected": -53.50571060180664,
"loss": 0.0965,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1035714149475098,
"rewards/margins": 3.1035714149475098,
"rewards/rejected": 0.0,
"step": 2330
},
{
"epoch": 13.022346368715084,
"grad_norm": 0.41900180730220143,
"learning_rate": 1.0188987582825731e-07,
"logits/chosen": -3.539759635925293,
"logits/rejected": -3.4876558780670166,
"logps/chosen": -0.5386089086532593,
"logps/rejected": -53.84013366699219,
"loss": 0.1126,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0928072929382324,
"rewards/margins": 2.0928072929382324,
"rewards/rejected": 0.0,
"step": 2331
},
{
"epoch": 13.027932960893855,
"grad_norm": 0.41260798214807526,
"learning_rate": 1.0152139310381564e-07,
"logits/chosen": -3.0246479511260986,
"logits/rejected": -3.1097795963287354,
"logps/chosen": -16.28483772277832,
"logps/rejected": -50.55168914794922,
"loss": 0.0947,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.687743663787842,
"rewards/margins": 3.687743663787842,
"rewards/rejected": 0.0,
"step": 2332
},
{
"epoch": 13.033519553072626,
"grad_norm": 0.40657110392164086,
"learning_rate": 1.0115350258954924e-07,
"logits/chosen": -3.4643912315368652,
"logits/rejected": -3.541138172149658,
"logps/chosen": -6.377556800842285,
"logps/rejected": -50.0582389831543,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3987014293670654,
"rewards/margins": 3.3987014293670654,
"rewards/rejected": 0.0,
"step": 2333
},
{
"epoch": 13.039106145251397,
"grad_norm": 0.5654692336015206,
"learning_rate": 1.007862048322084e-07,
"logits/chosen": -3.3608760833740234,
"logits/rejected": -3.413088321685791,
"logps/chosen": -0.3426097333431244,
"logps/rejected": -43.96967315673828,
"loss": 0.1014,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8785548210144043,
"rewards/margins": 1.8785548210144043,
"rewards/rejected": 0.0,
"step": 2334
},
{
"epoch": 13.044692737430168,
"grad_norm": 1.1834090094153376,
"learning_rate": 1.004195003776635e-07,
"logits/chosen": -3.2762203216552734,
"logits/rejected": -3.2310073375701904,
"logps/chosen": -21.713489532470703,
"logps/rejected": -64.16056823730469,
"loss": 0.099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.722824811935425,
"rewards/margins": 2.722824811935425,
"rewards/rejected": 0.0,
"step": 2335
},
{
"epoch": 13.050279329608939,
"grad_norm": 0.37757238915543967,
"learning_rate": 1.0005338977090222e-07,
"logits/chosen": -3.4060440063476562,
"logits/rejected": -3.3707094192504883,
"logps/chosen": -0.9788174629211426,
"logps/rejected": -70.06330108642578,
"loss": 0.114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3599703311920166,
"rewards/margins": 2.3599703311920166,
"rewards/rejected": 0.0,
"step": 2336
},
{
"epoch": 13.05586592178771,
"grad_norm": 0.8538934050291332,
"learning_rate": 9.968787355603043e-08,
"logits/chosen": -3.287316083908081,
"logits/rejected": -3.288396120071411,
"logps/chosen": -2.2251811027526855,
"logps/rejected": -25.12554168701172,
"loss": 0.1251,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1232540607452393,
"rewards/margins": 3.1232540607452393,
"rewards/rejected": 0.0,
"step": 2337
},
{
"epoch": 13.061452513966481,
"grad_norm": 0.41511079809678714,
"learning_rate": 9.932295227627024e-08,
"logits/chosen": -3.3394670486450195,
"logits/rejected": -3.3566854000091553,
"logps/chosen": -6.414815902709961,
"logps/rejected": -53.56317901611328,
"loss": 0.1009,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6157889366149902,
"rewards/margins": 3.6157889366149902,
"rewards/rejected": 0.0,
"step": 2338
},
{
"epoch": 13.067039106145252,
"grad_norm": 0.34623811350305467,
"learning_rate": 9.895862647395964e-08,
"logits/chosen": -3.355541467666626,
"logits/rejected": -3.5320823192596436,
"logps/chosen": -8.17490005493164,
"logps/rejected": -104.54173278808594,
"loss": 0.1158,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5288174152374268,
"rewards/margins": 3.5288174152374268,
"rewards/rejected": 0.0,
"step": 2339
},
{
"epoch": 13.072625698324023,
"grad_norm": 0.4178735197864293,
"learning_rate": 9.859489669055165e-08,
"logits/chosen": -2.971041440963745,
"logits/rejected": -3.082794666290283,
"logps/chosen": -3.396512269973755,
"logps/rejected": -40.988922119140625,
"loss": 0.109,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.658754587173462,
"rewards/margins": 2.658754587173462,
"rewards/rejected": 0.0,
"step": 2340
},
{
"epoch": 13.072625698324023,
"eval_logits/chosen": -3.2651684284210205,
"eval_logits/rejected": -3.3902339935302734,
"eval_logps/chosen": -30.2309513092041,
"eval_logps/rejected": -54.971168518066406,
"eval_loss": 1.0176070928573608,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.01609114371240139,
"eval_rewards/margins": -0.01609114371240139,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6447,
"eval_samples_per_second": 9.496,
"eval_steps_per_second": 0.306,
"step": 2340
},
{
"epoch": 13.078212290502794,
"grad_norm": 0.40969094850763416,
"learning_rate": 9.823176346661355e-08,
"logits/chosen": -3.2507095336914062,
"logits/rejected": -3.461381673812866,
"logps/chosen": -5.09861421585083,
"logps/rejected": -50.253456115722656,
"loss": 0.0888,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.959385395050049,
"rewards/margins": 2.959385395050049,
"rewards/rejected": 0.0,
"step": 2341
},
{
"epoch": 13.083798882681565,
"grad_norm": 0.39790536006023935,
"learning_rate": 9.786922734182596e-08,
"logits/chosen": -3.341721534729004,
"logits/rejected": -3.2856099605560303,
"logps/chosen": -3.377207040786743,
"logps/rejected": -36.50450897216797,
"loss": 0.0969,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.119828224182129,
"rewards/margins": 3.119828224182129,
"rewards/rejected": 0.0,
"step": 2342
},
{
"epoch": 13.089385474860336,
"grad_norm": 0.44848991439820296,
"learning_rate": 9.750728885498194e-08,
"logits/chosen": -3.4456944465637207,
"logits/rejected": -3.5190508365631104,
"logps/chosen": -20.230289459228516,
"logps/rejected": -60.31085968017578,
"loss": 0.0777,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.820690393447876,
"rewards/margins": 3.820690393447876,
"rewards/rejected": 0.0,
"step": 2343
},
{
"epoch": 13.094972067039107,
"grad_norm": 0.4169931954396705,
"learning_rate": 9.714594854398695e-08,
"logits/chosen": -3.7361228466033936,
"logits/rejected": -3.552189588546753,
"logps/chosen": -0.8857628107070923,
"logps/rejected": -43.35779571533203,
"loss": 0.1337,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4181251525878906,
"rewards/margins": 2.4181251525878906,
"rewards/rejected": 0.0,
"step": 2344
},
{
"epoch": 13.100558659217878,
"grad_norm": 0.3927370456785596,
"learning_rate": 9.678520694585662e-08,
"logits/chosen": -3.4572887420654297,
"logits/rejected": -3.352020502090454,
"logps/chosen": -17.534107208251953,
"logps/rejected": -52.65359115600586,
"loss": 0.0942,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.997272491455078,
"rewards/margins": 3.997272491455078,
"rewards/rejected": 0.0,
"step": 2345
},
{
"epoch": 13.106145251396647,
"grad_norm": 0.46886156405210677,
"learning_rate": 9.642506459671745e-08,
"logits/chosen": -3.3222928047180176,
"logits/rejected": -3.3929309844970703,
"logps/chosen": -3.5490429401397705,
"logps/rejected": -98.44245147705078,
"loss": 0.1173,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.640000581741333,
"rewards/margins": 2.640000581741333,
"rewards/rejected": 0.0,
"step": 2346
},
{
"epoch": 13.111731843575418,
"grad_norm": 0.4245962211325615,
"learning_rate": 9.606552203180513e-08,
"logits/chosen": -3.337649345397949,
"logits/rejected": -3.1500914096832275,
"logps/chosen": -14.195104598999023,
"logps/rejected": -40.21344757080078,
"loss": 0.1036,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5140798091888428,
"rewards/margins": 3.5140798091888428,
"rewards/rejected": 0.0,
"step": 2347
},
{
"epoch": 13.11731843575419,
"grad_norm": 0.5806114984389156,
"learning_rate": 9.570657978546381e-08,
"logits/chosen": -3.372487783432007,
"logits/rejected": -3.514693021774292,
"logps/chosen": -0.7192268371582031,
"logps/rejected": -105.02080535888672,
"loss": 0.1272,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.049988269805908,
"rewards/margins": 2.049988269805908,
"rewards/rejected": 0.0,
"step": 2348
},
{
"epoch": 13.12290502793296,
"grad_norm": 0.39375042646087816,
"learning_rate": 9.534823839114553e-08,
"logits/chosen": -3.4985358715057373,
"logits/rejected": -3.4309816360473633,
"logps/chosen": -19.601560592651367,
"logps/rejected": -43.45613098144531,
"loss": 0.0898,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.677501678466797,
"rewards/margins": 3.677501678466797,
"rewards/rejected": 0.0,
"step": 2349
},
{
"epoch": 13.128491620111731,
"grad_norm": 0.4178634388752394,
"learning_rate": 9.499049838140982e-08,
"logits/chosen": -3.300751209259033,
"logits/rejected": -3.357069969177246,
"logps/chosen": -3.3639509677886963,
"logps/rejected": -34.51044464111328,
"loss": 0.0848,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.083294153213501,
"rewards/margins": 3.083294153213501,
"rewards/rejected": 0.0,
"step": 2350
},
{
"epoch": 13.134078212290502,
"grad_norm": 0.33561040614202914,
"learning_rate": 9.463336028792157e-08,
"logits/chosen": -3.408461570739746,
"logits/rejected": -3.563394069671631,
"logps/chosen": -2.475998878479004,
"logps/rejected": -43.663578033447266,
"loss": 0.1234,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.060432195663452,
"rewards/margins": 3.060432195663452,
"rewards/rejected": 0.0,
"step": 2351
},
{
"epoch": 13.139664804469273,
"grad_norm": 0.3351565583774232,
"learning_rate": 9.42768246414517e-08,
"logits/chosen": -3.5358974933624268,
"logits/rejected": -3.603031635284424,
"logps/chosen": -1.1546649932861328,
"logps/rejected": -43.19342803955078,
"loss": 0.097,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2943787574768066,
"rewards/margins": 2.2943787574768066,
"rewards/rejected": 0.0,
"step": 2352
},
{
"epoch": 13.145251396648044,
"grad_norm": 0.4063516832940742,
"learning_rate": 9.392089197187602e-08,
"logits/chosen": -3.252317428588867,
"logits/rejected": -3.304638624191284,
"logps/chosen": -0.6251400709152222,
"logps/rejected": -60.52002716064453,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2664456367492676,
"rewards/margins": 2.2664456367492676,
"rewards/rejected": 0.0,
"step": 2353
},
{
"epoch": 13.150837988826815,
"grad_norm": 0.7339427043590783,
"learning_rate": 9.356556280817334e-08,
"logits/chosen": -3.2447080612182617,
"logits/rejected": -3.2067127227783203,
"logps/chosen": -0.7351337671279907,
"logps/rejected": -80.97441864013672,
"loss": 0.147,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.164402723312378,
"rewards/margins": 2.164402723312378,
"rewards/rejected": 0.0,
"step": 2354
},
{
"epoch": 13.156424581005586,
"grad_norm": 0.3783231588225719,
"learning_rate": 9.321083767842646e-08,
"logits/chosen": -3.2772772312164307,
"logits/rejected": -3.340531587600708,
"logps/chosen": -4.022555351257324,
"logps/rejected": -37.15563201904297,
"loss": 0.085,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.373973846435547,
"rewards/margins": 3.373973846435547,
"rewards/rejected": 0.0,
"step": 2355
},
{
"epoch": 13.162011173184357,
"grad_norm": 0.5567838268088865,
"learning_rate": 9.285671710981995e-08,
"logits/chosen": -3.241022825241089,
"logits/rejected": -3.380427598953247,
"logps/chosen": -1.5501210689544678,
"logps/rejected": -42.20450210571289,
"loss": 0.121,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.869105339050293,
"rewards/margins": 2.869105339050293,
"rewards/rejected": 0.0,
"step": 2356
},
{
"epoch": 13.167597765363128,
"grad_norm": 0.41097528962627683,
"learning_rate": 9.250320162864012e-08,
"logits/chosen": -3.313119888305664,
"logits/rejected": -3.4442129135131836,
"logps/chosen": -5.751883029937744,
"logps/rejected": -36.29100799560547,
"loss": 0.1063,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2788279056549072,
"rewards/margins": 2.2788279056549072,
"rewards/rejected": 0.0,
"step": 2357
},
{
"epoch": 13.1731843575419,
"grad_norm": 0.38552319656934547,
"learning_rate": 9.215029176027373e-08,
"logits/chosen": -3.304882049560547,
"logits/rejected": -3.3625175952911377,
"logps/chosen": -1.073160171508789,
"logps/rejected": -75.98919677734375,
"loss": 0.075,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6493020057678223,
"rewards/margins": 2.6493020057678223,
"rewards/rejected": 0.0,
"step": 2358
},
{
"epoch": 13.17877094972067,
"grad_norm": 0.4618925236399504,
"learning_rate": 9.179798802920813e-08,
"logits/chosen": -3.4035518169403076,
"logits/rejected": -3.3589015007019043,
"logps/chosen": -6.312320709228516,
"logps/rejected": -42.05010223388672,
"loss": 0.0911,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.631145715713501,
"rewards/margins": 3.631145715713501,
"rewards/rejected": 0.0,
"step": 2359
},
{
"epoch": 13.184357541899441,
"grad_norm": 0.3400826453101628,
"learning_rate": 9.144629095902894e-08,
"logits/chosen": -3.422144889831543,
"logits/rejected": -3.3887534141540527,
"logps/chosen": -18.418508529663086,
"logps/rejected": -41.16917037963867,
"loss": 0.0913,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.146446466445923,
"rewards/margins": 3.146446466445923,
"rewards/rejected": 0.0,
"step": 2360
},
{
"epoch": 13.184357541899441,
"eval_logits/chosen": -3.256208896636963,
"eval_logits/rejected": -3.3817107677459717,
"eval_logps/chosen": -30.51275634765625,
"eval_logps/rejected": -55.494361877441406,
"eval_loss": 1.0322561264038086,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.04427126422524452,
"eval_rewards/margins": -0.04427126422524452,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6781,
"eval_samples_per_second": 9.486,
"eval_steps_per_second": 0.306,
"step": 2360
},
{
"epoch": 13.189944134078212,
"grad_norm": 1.002263887913487,
"learning_rate": 9.109520107242069e-08,
"logits/chosen": -3.4199793338775635,
"logits/rejected": -3.5535426139831543,
"logps/chosen": -3.609666347503662,
"logps/rejected": -48.57051086425781,
"loss": 0.0818,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0819921493530273,
"rewards/margins": 3.0819921493530273,
"rewards/rejected": 0.0,
"step": 2361
},
{
"epoch": 13.195530726256983,
"grad_norm": 0.340160470860633,
"learning_rate": 9.074471889116591e-08,
"logits/chosen": -3.2756481170654297,
"logits/rejected": -3.430361270904541,
"logps/chosen": -0.9844825267791748,
"logps/rejected": -47.28947830200195,
"loss": 0.1052,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6441168785095215,
"rewards/margins": 2.6441168785095215,
"rewards/rejected": 0.0,
"step": 2362
},
{
"epoch": 13.201117318435754,
"grad_norm": 0.4376289377535241,
"learning_rate": 9.039484493614297e-08,
"logits/chosen": -3.513704299926758,
"logits/rejected": -3.4845776557922363,
"logps/chosen": -15.548395156860352,
"logps/rejected": -56.104347229003906,
"loss": 0.0836,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.270206928253174,
"rewards/margins": 3.270206928253174,
"rewards/rejected": 0.0,
"step": 2363
},
{
"epoch": 13.206703910614525,
"grad_norm": 0.4568918698457708,
"learning_rate": 9.00455797273274e-08,
"logits/chosen": -3.138018846511841,
"logits/rejected": -3.2717790603637695,
"logps/chosen": -1.6022212505340576,
"logps/rejected": -51.163604736328125,
"loss": 0.1047,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.738399028778076,
"rewards/margins": 2.738399028778076,
"rewards/rejected": 0.0,
"step": 2364
},
{
"epoch": 13.212290502793296,
"grad_norm": 0.40291158651426134,
"learning_rate": 8.969692378378895e-08,
"logits/chosen": -3.2863690853118896,
"logits/rejected": -3.287386655807495,
"logps/chosen": -25.854572296142578,
"logps/rejected": -42.86607360839844,
"loss": 0.0766,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.345571994781494,
"rewards/margins": 4.345571994781494,
"rewards/rejected": 0.0,
"step": 2365
},
{
"epoch": 13.217877094972067,
"grad_norm": 0.42250807313026095,
"learning_rate": 8.934887762369292e-08,
"logits/chosen": -3.4446041584014893,
"logits/rejected": -3.593846321105957,
"logps/chosen": -0.8210120797157288,
"logps/rejected": -45.549713134765625,
"loss": 0.1078,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.494363307952881,
"rewards/margins": 2.494363307952881,
"rewards/rejected": 0.0,
"step": 2366
},
{
"epoch": 13.223463687150838,
"grad_norm": 0.4410715974925357,
"learning_rate": 8.900144176429763e-08,
"logits/chosen": -3.172802448272705,
"logits/rejected": -3.222987174987793,
"logps/chosen": -6.237480640411377,
"logps/rejected": -56.25542449951172,
"loss": 0.111,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.55985689163208,
"rewards/margins": 2.55985689163208,
"rewards/rejected": 0.0,
"step": 2367
},
{
"epoch": 13.22905027932961,
"grad_norm": 1.9814781233024836,
"learning_rate": 8.865461672195479e-08,
"logits/chosen": -3.2542126178741455,
"logits/rejected": -3.2465641498565674,
"logps/chosen": -8.509102821350098,
"logps/rejected": -52.00053787231445,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2179455757141113,
"rewards/margins": 3.2179455757141113,
"rewards/rejected": 0.0,
"step": 2368
},
{
"epoch": 13.23463687150838,
"grad_norm": 0.5108016115370785,
"learning_rate": 8.830840301210796e-08,
"logits/chosen": -3.230532169342041,
"logits/rejected": -3.542701244354248,
"logps/chosen": -1.4257872104644775,
"logps/rejected": -74.46788024902344,
"loss": 0.1489,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5936241149902344,
"rewards/margins": 2.5936241149902344,
"rewards/rejected": 0.0,
"step": 2369
},
{
"epoch": 13.240223463687151,
"grad_norm": 0.43404165553903007,
"learning_rate": 8.796280114929272e-08,
"logits/chosen": -3.2910354137420654,
"logits/rejected": -3.26667857170105,
"logps/chosen": -5.609430313110352,
"logps/rejected": -37.08899688720703,
"loss": 0.1039,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2219905853271484,
"rewards/margins": 3.2219905853271484,
"rewards/rejected": 0.0,
"step": 2370
},
{
"epoch": 13.245810055865922,
"grad_norm": 0.4531421096419252,
"learning_rate": 8.761781164713472e-08,
"logits/chosen": -3.318911075592041,
"logits/rejected": -3.3356807231903076,
"logps/chosen": -7.822868824005127,
"logps/rejected": -53.15188980102539,
"loss": 0.0794,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9234256744384766,
"rewards/margins": 2.9234256744384766,
"rewards/rejected": 0.0,
"step": 2371
},
{
"epoch": 13.251396648044693,
"grad_norm": 0.4393544810482372,
"learning_rate": 8.727343501834972e-08,
"logits/chosen": -3.2168662548065186,
"logits/rejected": -3.3560662269592285,
"logps/chosen": -25.274131774902344,
"logps/rejected": -55.020294189453125,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.256657123565674,
"rewards/margins": 3.256657123565674,
"rewards/rejected": 0.0,
"step": 2372
},
{
"epoch": 13.256983240223464,
"grad_norm": 0.3822046822850603,
"learning_rate": 8.692967177474309e-08,
"logits/chosen": -3.4313976764678955,
"logits/rejected": -3.4952926635742188,
"logps/chosen": -6.944964408874512,
"logps/rejected": -48.05303192138672,
"loss": 0.0862,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.522512912750244,
"rewards/margins": 4.522512912750244,
"rewards/rejected": 0.0,
"step": 2373
},
{
"epoch": 13.262569832402235,
"grad_norm": 1.780335975045036,
"learning_rate": 8.658652242720777e-08,
"logits/chosen": -2.8541641235351562,
"logits/rejected": -3.011066198348999,
"logps/chosen": -3.7367441654205322,
"logps/rejected": -40.481956481933594,
"loss": 0.1302,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9329137802124023,
"rewards/margins": 2.9329137802124023,
"rewards/rejected": 0.0,
"step": 2374
},
{
"epoch": 13.268156424581006,
"grad_norm": 0.615246519240483,
"learning_rate": 8.62439874857252e-08,
"logits/chosen": -3.517827033996582,
"logits/rejected": -3.4203219413757324,
"logps/chosen": -4.895214080810547,
"logps/rejected": -49.74782180786133,
"loss": 0.0919,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.790727138519287,
"rewards/margins": 2.790727138519287,
"rewards/rejected": 0.0,
"step": 2375
},
{
"epoch": 13.273743016759777,
"grad_norm": 0.3921191278380291,
"learning_rate": 8.590206745936307e-08,
"logits/chosen": -3.570204734802246,
"logits/rejected": -3.51688551902771,
"logps/chosen": -5.914793491363525,
"logps/rejected": -104.97129821777344,
"loss": 0.0927,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8447258472442627,
"rewards/margins": 3.8447258472442627,
"rewards/rejected": 0.0,
"step": 2376
},
{
"epoch": 13.279329608938548,
"grad_norm": 0.8084549264345066,
"learning_rate": 8.556076285627556e-08,
"logits/chosen": -2.666715145111084,
"logits/rejected": -2.7342615127563477,
"logps/chosen": -8.738677978515625,
"logps/rejected": -100.0520248413086,
"loss": 0.1035,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3125147819519043,
"rewards/margins": 2.3125147819519043,
"rewards/rejected": 0.0,
"step": 2377
},
{
"epoch": 13.28491620111732,
"grad_norm": 0.3602284916437586,
"learning_rate": 8.522007418370187e-08,
"logits/chosen": -3.507559061050415,
"logits/rejected": -3.5460026264190674,
"logps/chosen": -0.5510622262954712,
"logps/rejected": -39.30126190185547,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.086501359939575,
"rewards/margins": 2.086501359939575,
"rewards/rejected": 0.0,
"step": 2378
},
{
"epoch": 13.29050279329609,
"grad_norm": 0.449789327678422,
"learning_rate": 8.488000194796646e-08,
"logits/chosen": -3.4018757343292236,
"logits/rejected": -3.5999956130981445,
"logps/chosen": -2.413813352584839,
"logps/rejected": -36.636959075927734,
"loss": 0.0592,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.452755928039551,
"rewards/margins": 3.452755928039551,
"rewards/rejected": 0.0,
"step": 2379
},
{
"epoch": 13.296089385474861,
"grad_norm": 1.131102020935119,
"learning_rate": 8.45405466544768e-08,
"logits/chosen": -3.4479286670684814,
"logits/rejected": -3.5594661235809326,
"logps/chosen": -2.1537208557128906,
"logps/rejected": -59.7096061706543,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1846818923950195,
"rewards/margins": 2.1846818923950195,
"rewards/rejected": 0.0,
"step": 2380
},
{
"epoch": 13.296089385474861,
"eval_logits/chosen": -3.2597923278808594,
"eval_logits/rejected": -3.3839545249938965,
"eval_logps/chosen": -30.492467880249023,
"eval_logps/rejected": -55.89141082763672,
"eval_loss": 1.0290724039077759,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.04224271699786186,
"eval_rewards/margins": -0.04224271699786186,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7095,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 2380
},
{
"epoch": 13.30167597765363,
"grad_norm": 0.4456367654266066,
"learning_rate": 8.420170880772415e-08,
"logits/chosen": -3.3668711185455322,
"logits/rejected": -3.6564130783081055,
"logps/chosen": -0.4650506377220154,
"logps/rejected": -63.997406005859375,
"loss": 0.1067,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1184096336364746,
"rewards/margins": 2.1184096336364746,
"rewards/rejected": 0.0,
"step": 2381
},
{
"epoch": 13.307262569832401,
"grad_norm": 0.36569527767192056,
"learning_rate": 8.386348891128198e-08,
"logits/chosen": -3.3553197383880615,
"logits/rejected": -3.3262808322906494,
"logps/chosen": -0.3164428770542145,
"logps/rejected": -34.770713806152344,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7639267444610596,
"rewards/margins": 1.7639267444610596,
"rewards/rejected": 0.0,
"step": 2382
},
{
"epoch": 13.312849162011172,
"grad_norm": 0.40184913178279863,
"learning_rate": 8.352588746780487e-08,
"logits/chosen": -3.364434003829956,
"logits/rejected": -3.323281764984131,
"logps/chosen": -5.599338531494141,
"logps/rejected": -35.86624526977539,
"loss": 0.0869,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.009035587310791,
"rewards/margins": 3.009035587310791,
"rewards/rejected": 0.0,
"step": 2383
},
{
"epoch": 13.318435754189943,
"grad_norm": 0.3866840029712693,
"learning_rate": 8.318890497902914e-08,
"logits/chosen": -3.225160837173462,
"logits/rejected": -3.2643654346466064,
"logps/chosen": -6.571699142456055,
"logps/rejected": -31.512510299682617,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1004815101623535,
"rewards/margins": 3.1004815101623535,
"rewards/rejected": 0.0,
"step": 2384
},
{
"epoch": 13.324022346368714,
"grad_norm": 0.41761581252563856,
"learning_rate": 8.285254194577057e-08,
"logits/chosen": -3.3451530933380127,
"logits/rejected": -3.414379358291626,
"logps/chosen": -0.5896964073181152,
"logps/rejected": -47.427093505859375,
"loss": 0.1089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.404115676879883,
"rewards/margins": 2.404115676879883,
"rewards/rejected": 0.0,
"step": 2385
},
{
"epoch": 13.329608938547485,
"grad_norm": 0.45632078737894827,
"learning_rate": 8.251679886792457e-08,
"logits/chosen": -3.568117380142212,
"logits/rejected": -3.6124401092529297,
"logps/chosen": -7.401118278503418,
"logps/rejected": -71.91776275634766,
"loss": 0.0893,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2571563720703125,
"rewards/margins": 3.2571563720703125,
"rewards/rejected": 0.0,
"step": 2386
},
{
"epoch": 13.335195530726256,
"grad_norm": 0.5625236710883265,
"learning_rate": 8.218167624446504e-08,
"logits/chosen": -3.2756669521331787,
"logits/rejected": -3.131768226623535,
"logps/chosen": -0.7099722623825073,
"logps/rejected": -32.888465881347656,
"loss": 0.1087,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3063101768493652,
"rewards/margins": 2.3063101768493652,
"rewards/rejected": 0.0,
"step": 2387
},
{
"epoch": 13.340782122905027,
"grad_norm": 0.7671117523165139,
"learning_rate": 8.18471745734442e-08,
"logits/chosen": -3.6030690670013428,
"logits/rejected": -3.4550702571868896,
"logps/chosen": -0.3922824263572693,
"logps/rejected": -50.66614532470703,
"loss": 0.1121,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2038583755493164,
"rewards/margins": 2.2038583755493164,
"rewards/rejected": 0.0,
"step": 2388
},
{
"epoch": 13.346368715083798,
"grad_norm": 0.4001003233855212,
"learning_rate": 8.151329435199067e-08,
"logits/chosen": -3.3189687728881836,
"logits/rejected": -3.348407745361328,
"logps/chosen": -0.6709566116333008,
"logps/rejected": -64.72965240478516,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.204169273376465,
"rewards/margins": 2.204169273376465,
"rewards/rejected": 0.0,
"step": 2389
},
{
"epoch": 13.35195530726257,
"grad_norm": 0.3865552716765946,
"learning_rate": 8.118003607631024e-08,
"logits/chosen": -3.5786283016204834,
"logits/rejected": -3.579033374786377,
"logps/chosen": -2.694607734680176,
"logps/rejected": -81.97444152832031,
"loss": 0.1167,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.57879900932312,
"rewards/margins": 2.57879900932312,
"rewards/rejected": 0.0,
"step": 2390
},
{
"epoch": 13.35754189944134,
"grad_norm": 0.48414906708529765,
"learning_rate": 8.084740024168407e-08,
"logits/chosen": -3.3128931522369385,
"logits/rejected": -3.3622987270355225,
"logps/chosen": -0.8101568222045898,
"logps/rejected": -55.46433639526367,
"loss": 0.0783,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5311551094055176,
"rewards/margins": 2.5311551094055176,
"rewards/rejected": 0.0,
"step": 2391
},
{
"epoch": 13.363128491620111,
"grad_norm": 0.4407183387585617,
"learning_rate": 8.051538734246793e-08,
"logits/chosen": -3.258774995803833,
"logits/rejected": -3.069319486618042,
"logps/chosen": -2.305711030960083,
"logps/rejected": -43.63422393798828,
"loss": 0.0942,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1751363277435303,
"rewards/margins": 3.1751363277435303,
"rewards/rejected": 0.0,
"step": 2392
},
{
"epoch": 13.368715083798882,
"grad_norm": 0.40758371727240833,
"learning_rate": 8.018399787209256e-08,
"logits/chosen": -3.353487491607666,
"logits/rejected": -3.504014492034912,
"logps/chosen": -0.7513201236724854,
"logps/rejected": -46.85102081298828,
"loss": 0.0854,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2565627098083496,
"rewards/margins": 2.2565627098083496,
"rewards/rejected": 0.0,
"step": 2393
},
{
"epoch": 13.374301675977653,
"grad_norm": 1.5687597986504542,
"learning_rate": 7.985323232306124e-08,
"logits/chosen": -3.660555124282837,
"logits/rejected": -3.657735586166382,
"logps/chosen": -5.776373386383057,
"logps/rejected": -43.98847961425781,
"loss": 0.1194,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4119997024536133,
"rewards/margins": 3.4119997024536133,
"rewards/rejected": 0.0,
"step": 2394
},
{
"epoch": 13.379888268156424,
"grad_norm": 0.3220517556472039,
"learning_rate": 7.952309118695083e-08,
"logits/chosen": -3.4327592849731445,
"logits/rejected": -3.3201828002929688,
"logps/chosen": -4.240723133087158,
"logps/rejected": -66.45988464355469,
"loss": 0.136,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1899328231811523,
"rewards/margins": 3.1899328231811523,
"rewards/rejected": 0.0,
"step": 2395
},
{
"epoch": 13.385474860335195,
"grad_norm": 0.35886696892881076,
"learning_rate": 7.919357495440976e-08,
"logits/chosen": -3.4310574531555176,
"logits/rejected": -3.372286558151245,
"logps/chosen": -5.797863483428955,
"logps/rejected": -46.23900604248047,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9535272121429443,
"rewards/margins": 3.9535272121429443,
"rewards/rejected": 0.0,
"step": 2396
},
{
"epoch": 13.391061452513966,
"grad_norm": 0.396275412424554,
"learning_rate": 7.886468411515784e-08,
"logits/chosen": -3.400118350982666,
"logits/rejected": -3.1456902027130127,
"logps/chosen": -8.911643028259277,
"logps/rejected": -44.610939025878906,
"loss": 0.1109,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2362236976623535,
"rewards/margins": 3.2362236976623535,
"rewards/rejected": 0.0,
"step": 2397
},
{
"epoch": 13.396648044692737,
"grad_norm": 0.4708518096849073,
"learning_rate": 7.853641915798531e-08,
"logits/chosen": -2.7484145164489746,
"logits/rejected": -2.6871142387390137,
"logps/chosen": -4.712937831878662,
"logps/rejected": -31.69449806213379,
"loss": 0.0988,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9272632598876953,
"rewards/margins": 2.9272632598876953,
"rewards/rejected": 0.0,
"step": 2398
},
{
"epoch": 13.402234636871508,
"grad_norm": 0.38999874305736726,
"learning_rate": 7.820878057075275e-08,
"logits/chosen": -3.495323419570923,
"logits/rejected": -3.4236326217651367,
"logps/chosen": -6.782710552215576,
"logps/rejected": -52.033782958984375,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9194343090057373,
"rewards/margins": 2.9194343090057373,
"rewards/rejected": 0.0,
"step": 2399
},
{
"epoch": 13.40782122905028,
"grad_norm": 0.41039560712904977,
"learning_rate": 7.788176884038889e-08,
"logits/chosen": -3.5283381938934326,
"logits/rejected": -3.547929048538208,
"logps/chosen": -0.29387354850769043,
"logps/rejected": -49.664527893066406,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6123600006103516,
"rewards/margins": 1.6123600006103516,
"rewards/rejected": 0.0,
"step": 2400
},
{
"epoch": 13.40782122905028,
"eval_logits/chosen": -3.258481502532959,
"eval_logits/rejected": -3.381990909576416,
"eval_logps/chosen": -30.63241195678711,
"eval_logps/rejected": -56.020240783691406,
"eval_loss": 1.0348480939865112,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.056236814707517624,
"eval_rewards/margins": -0.056236814707517624,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7249,
"eval_samples_per_second": 9.473,
"eval_steps_per_second": 0.306,
"step": 2400
},
{
"epoch": 13.41340782122905,
"grad_norm": 1.2377268202018599,
"learning_rate": 7.755538445289179e-08,
"logits/chosen": -3.432081699371338,
"logits/rejected": -3.4291303157806396,
"logps/chosen": -5.873993396759033,
"logps/rejected": -46.123741149902344,
"loss": 0.0891,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1080503463745117,
"rewards/margins": 3.1080503463745117,
"rewards/rejected": 0.0,
"step": 2401
},
{
"epoch": 13.418994413407821,
"grad_norm": 0.40422115430184585,
"learning_rate": 7.722962789332676e-08,
"logits/chosen": -3.4764888286590576,
"logits/rejected": -3.468198776245117,
"logps/chosen": -5.431130886077881,
"logps/rejected": -34.02408218383789,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5071778297424316,
"rewards/margins": 3.5071778297424316,
"rewards/rejected": 0.0,
"step": 2402
},
{
"epoch": 13.424581005586592,
"grad_norm": 0.41757046049489394,
"learning_rate": 7.69044996458258e-08,
"logits/chosen": -3.4845077991485596,
"logits/rejected": -3.4692227840423584,
"logps/chosen": -0.43687373399734497,
"logps/rejected": -37.74633026123047,
"loss": 0.0858,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.040178060531616,
"rewards/margins": 2.040178060531616,
"rewards/rejected": 0.0,
"step": 2403
},
{
"epoch": 13.430167597765363,
"grad_norm": 1.1194557115758397,
"learning_rate": 7.658000019358762e-08,
"logits/chosen": -3.3184359073638916,
"logits/rejected": -3.2005765438079834,
"logps/chosen": -0.553019642829895,
"logps/rejected": -111.14012145996094,
"loss": 0.1347,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7884210348129272,
"rewards/margins": 1.7884210348129272,
"rewards/rejected": 0.0,
"step": 2404
},
{
"epoch": 13.435754189944134,
"grad_norm": 0.4788821439759304,
"learning_rate": 7.62561300188761e-08,
"logits/chosen": -3.4542956352233887,
"logits/rejected": -3.4098803997039795,
"logps/chosen": -4.152714729309082,
"logps/rejected": -39.71040344238281,
"loss": 0.1296,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5099453926086426,
"rewards/margins": 3.5099453926086426,
"rewards/rejected": 0.0,
"step": 2405
},
{
"epoch": 13.441340782122905,
"grad_norm": 0.49335687149670227,
"learning_rate": 7.593288960302008e-08,
"logits/chosen": -3.4500067234039307,
"logits/rejected": -3.3621268272399902,
"logps/chosen": -0.4197743237018585,
"logps/rejected": -51.166534423828125,
"loss": 0.0916,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.170515775680542,
"rewards/margins": 2.170515775680542,
"rewards/rejected": 0.0,
"step": 2406
},
{
"epoch": 13.446927374301676,
"grad_norm": 0.41892482487072363,
"learning_rate": 7.561027942641218e-08,
"logits/chosen": -3.198925495147705,
"logits/rejected": -3.25957989692688,
"logps/chosen": -8.140546798706055,
"logps/rejected": -102.78670501708984,
"loss": 0.0732,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6075353622436523,
"rewards/margins": 3.6075353622436523,
"rewards/rejected": 0.0,
"step": 2407
},
{
"epoch": 13.452513966480447,
"grad_norm": 0.8317060791614108,
"learning_rate": 7.528829996850899e-08,
"logits/chosen": -3.583693504333496,
"logits/rejected": -3.5818145275115967,
"logps/chosen": -14.598260879516602,
"logps/rejected": -41.034305572509766,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.366270542144775,
"rewards/margins": 4.366270542144775,
"rewards/rejected": 0.0,
"step": 2408
},
{
"epoch": 13.458100558659218,
"grad_norm": 0.4295777228419252,
"learning_rate": 7.496695170782879e-08,
"logits/chosen": -3.62851619720459,
"logits/rejected": -3.5785186290740967,
"logps/chosen": -2.085975408554077,
"logps/rejected": -66.775634765625,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.925323963165283,
"rewards/margins": 2.925323963165283,
"rewards/rejected": 0.0,
"step": 2409
},
{
"epoch": 13.46368715083799,
"grad_norm": 0.37220209217410677,
"learning_rate": 7.464623512195278e-08,
"logits/chosen": -3.257615327835083,
"logits/rejected": -3.296339988708496,
"logps/chosen": -0.5343749523162842,
"logps/rejected": -43.92362976074219,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4658780097961426,
"rewards/margins": 2.4658780097961426,
"rewards/rejected": 0.0,
"step": 2410
},
{
"epoch": 13.46927374301676,
"grad_norm": 0.3584433295653612,
"learning_rate": 7.432615068752279e-08,
"logits/chosen": -3.3008205890655518,
"logits/rejected": -3.3128490447998047,
"logps/chosen": -8.472904205322266,
"logps/rejected": -41.18153381347656,
"loss": 0.0959,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.400026321411133,
"rewards/margins": 3.400026321411133,
"rewards/rejected": 0.0,
"step": 2411
},
{
"epoch": 13.474860335195531,
"grad_norm": 0.39790237405570594,
"learning_rate": 7.400669888024108e-08,
"logits/chosen": -3.236492395401001,
"logits/rejected": -3.2329535484313965,
"logps/chosen": -7.113299369812012,
"logps/rejected": -41.26221466064453,
"loss": 0.082,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.763033866882324,
"rewards/margins": 2.763033866882324,
"rewards/rejected": 0.0,
"step": 2412
},
{
"epoch": 13.480446927374302,
"grad_norm": 0.4062685015336655,
"learning_rate": 7.368788017487016e-08,
"logits/chosen": -3.2081613540649414,
"logits/rejected": -3.2045440673828125,
"logps/chosen": -1.781062364578247,
"logps/rejected": -66.75691223144531,
"loss": 0.0927,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8163845539093018,
"rewards/margins": 2.8163845539093018,
"rewards/rejected": 0.0,
"step": 2413
},
{
"epoch": 13.486033519553073,
"grad_norm": 0.3760925307992988,
"learning_rate": 7.336969504523133e-08,
"logits/chosen": -3.3029186725616455,
"logits/rejected": -3.1940293312072754,
"logps/chosen": -1.1946035623550415,
"logps/rejected": -51.66459655761719,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4233834743499756,
"rewards/margins": 2.4233834743499756,
"rewards/rejected": 0.0,
"step": 2414
},
{
"epoch": 13.491620111731844,
"grad_norm": 0.38560132937368724,
"learning_rate": 7.305214396420439e-08,
"logits/chosen": -3.598036527633667,
"logits/rejected": -3.5693676471710205,
"logps/chosen": -33.32712936401367,
"logps/rejected": -30.71826934814453,
"loss": 0.0922,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.256027698516846,
"rewards/margins": 4.256027698516846,
"rewards/rejected": 0.0,
"step": 2415
},
{
"epoch": 13.497206703910614,
"grad_norm": 0.4257576326590357,
"learning_rate": 7.273522740372662e-08,
"logits/chosen": -3.414893388748169,
"logits/rejected": -3.609196186065674,
"logps/chosen": -0.678642749786377,
"logps/rejected": -100.84706115722656,
"loss": 0.139,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.088613510131836,
"rewards/margins": 2.088613510131836,
"rewards/rejected": 0.0,
"step": 2416
},
{
"epoch": 13.502793296089386,
"grad_norm": 0.352367799804756,
"learning_rate": 7.241894583479285e-08,
"logits/chosen": -3.357116937637329,
"logits/rejected": -3.5004019737243652,
"logps/chosen": -1.0748229026794434,
"logps/rejected": -55.708335876464844,
"loss": 0.1256,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.753563404083252,
"rewards/margins": 2.753563404083252,
"rewards/rejected": 0.0,
"step": 2417
},
{
"epoch": 13.508379888268156,
"grad_norm": 0.46647670703766125,
"learning_rate": 7.210329972745343e-08,
"logits/chosen": -3.197568893432617,
"logits/rejected": -3.4688611030578613,
"logps/chosen": -1.6286230087280273,
"logps/rejected": -61.687870025634766,
"loss": 0.1062,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.151333808898926,
"rewards/margins": 2.151333808898926,
"rewards/rejected": 0.0,
"step": 2418
},
{
"epoch": 13.513966480446927,
"grad_norm": 0.37209910965383863,
"learning_rate": 7.17882895508149e-08,
"logits/chosen": -3.263106346130371,
"logits/rejected": -3.3420896530151367,
"logps/chosen": -0.9882630109786987,
"logps/rejected": -48.91582489013672,
"loss": 0.1219,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.689342498779297,
"rewards/margins": 2.689342498779297,
"rewards/rejected": 0.0,
"step": 2419
},
{
"epoch": 13.519553072625698,
"grad_norm": 0.5386244488849635,
"learning_rate": 7.147391577303846e-08,
"logits/chosen": -3.374979257583618,
"logits/rejected": -3.3277366161346436,
"logps/chosen": -0.9086631536483765,
"logps/rejected": -60.96214294433594,
"loss": 0.0812,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7497658729553223,
"rewards/margins": 2.7497658729553223,
"rewards/rejected": 0.0,
"step": 2420
},
{
"epoch": 13.519553072625698,
"eval_logits/chosen": -3.2651188373565674,
"eval_logits/rejected": -3.3867595195770264,
"eval_logps/chosen": -30.666950225830078,
"eval_logps/rejected": -55.848052978515625,
"eval_loss": 1.037139654159546,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.059690773487091064,
"eval_rewards/margins": -0.059690773487091064,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6905,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 2420
},
{
"epoch": 13.525139664804469,
"grad_norm": 0.3462712885292635,
"learning_rate": 7.116017886133946e-08,
"logits/chosen": -3.36868953704834,
"logits/rejected": -3.415334463119507,
"logps/chosen": -1.1047523021697998,
"logps/rejected": -50.67715835571289,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1574814319610596,
"rewards/margins": 2.1574814319610596,
"rewards/rejected": 0.0,
"step": 2421
},
{
"epoch": 13.53072625698324,
"grad_norm": 0.8207276516407748,
"learning_rate": 7.084707928198702e-08,
"logits/chosen": -3.461683511734009,
"logits/rejected": -3.4847850799560547,
"logps/chosen": -0.8327699899673462,
"logps/rejected": -43.814453125,
"loss": 0.101,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.402855396270752,
"rewards/margins": 2.402855396270752,
"rewards/rejected": 0.0,
"step": 2422
},
{
"epoch": 13.53631284916201,
"grad_norm": 1.3697531927469424,
"learning_rate": 7.053461750030249e-08,
"logits/chosen": -3.266954183578491,
"logits/rejected": -3.3406190872192383,
"logps/chosen": -0.3930329382419586,
"logps/rejected": -58.03802490234375,
"loss": 0.1051,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7540619373321533,
"rewards/margins": 1.7540619373321533,
"rewards/rejected": 0.0,
"step": 2423
},
{
"epoch": 13.541899441340782,
"grad_norm": 0.3806872794265044,
"learning_rate": 7.022279398066005e-08,
"logits/chosen": -3.372802734375,
"logits/rejected": -3.563987970352173,
"logps/chosen": -3.2917611598968506,
"logps/rejected": -73.0836181640625,
"loss": 0.0725,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.173635959625244,
"rewards/margins": 3.173635959625244,
"rewards/rejected": 0.0,
"step": 2424
},
{
"epoch": 13.547486033519553,
"grad_norm": 0.8882674169538836,
"learning_rate": 6.991160918648493e-08,
"logits/chosen": -3.353987455368042,
"logits/rejected": -3.4794535636901855,
"logps/chosen": -1.3371047973632812,
"logps/rejected": -49.239593505859375,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.450688362121582,
"rewards/margins": 2.450688362121582,
"rewards/rejected": 0.0,
"step": 2425
},
{
"epoch": 13.553072625698324,
"grad_norm": 0.3142022219595674,
"learning_rate": 6.960106358025308e-08,
"logits/chosen": -3.6025381088256836,
"logits/rejected": -3.450307846069336,
"logps/chosen": -3.1776933670043945,
"logps/rejected": -82.77340698242188,
"loss": 0.1186,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1456172466278076,
"rewards/margins": 3.1456172466278076,
"rewards/rejected": 0.0,
"step": 2426
},
{
"epoch": 13.558659217877095,
"grad_norm": 0.3932636939664874,
"learning_rate": 6.929115762349047e-08,
"logits/chosen": -3.5441834926605225,
"logits/rejected": -3.4236481189727783,
"logps/chosen": -0.7303559184074402,
"logps/rejected": -36.51518630981445,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1503782272338867,
"rewards/margins": 2.1503782272338867,
"rewards/rejected": 0.0,
"step": 2427
},
{
"epoch": 13.564245810055866,
"grad_norm": 0.40096723201484263,
"learning_rate": 6.898189177677294e-08,
"logits/chosen": -3.488233804702759,
"logits/rejected": -3.387334108352661,
"logps/chosen": -4.267492771148682,
"logps/rejected": -40.590179443359375,
"loss": 0.0789,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.034738540649414,
"rewards/margins": 3.034738540649414,
"rewards/rejected": 0.0,
"step": 2428
},
{
"epoch": 13.569832402234637,
"grad_norm": 0.4350445433608281,
"learning_rate": 6.867326649972415e-08,
"logits/chosen": -3.425281286239624,
"logits/rejected": -3.5342533588409424,
"logps/chosen": -0.9734011888504028,
"logps/rejected": -107.08503723144531,
"loss": 0.1467,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7057478427886963,
"rewards/margins": 2.7057478427886963,
"rewards/rejected": 0.0,
"step": 2429
},
{
"epoch": 13.575418994413408,
"grad_norm": 0.8901314257917506,
"learning_rate": 6.836528225101651e-08,
"logits/chosen": -3.3475356101989746,
"logits/rejected": -3.3032736778259277,
"logps/chosen": -1.2079683542251587,
"logps/rejected": -50.296966552734375,
"loss": 0.0984,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3507351875305176,
"rewards/margins": 2.3507351875305176,
"rewards/rejected": 0.0,
"step": 2430
},
{
"epoch": 13.581005586592179,
"grad_norm": 0.42334171374625834,
"learning_rate": 6.805793948836941e-08,
"logits/chosen": -3.5956475734710693,
"logits/rejected": -3.687025547027588,
"logps/chosen": -4.879007339477539,
"logps/rejected": -56.3758430480957,
"loss": 0.0739,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.328376054763794,
"rewards/margins": 3.328376054763794,
"rewards/rejected": 0.0,
"step": 2431
},
{
"epoch": 13.58659217877095,
"grad_norm": 0.47845070758021246,
"learning_rate": 6.775123866854888e-08,
"logits/chosen": -3.5741028785705566,
"logits/rejected": -3.652073383331299,
"logps/chosen": -2.6571004390716553,
"logps/rejected": -65.98702239990234,
"loss": 0.0884,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9428372383117676,
"rewards/margins": 2.9428372383117676,
"rewards/rejected": 0.0,
"step": 2432
},
{
"epoch": 13.59217877094972,
"grad_norm": 0.901852773461313,
"learning_rate": 6.744518024736696e-08,
"logits/chosen": -3.3788886070251465,
"logits/rejected": -3.3145554065704346,
"logps/chosen": -0.9111521244049072,
"logps/rejected": -66.0184097290039,
"loss": 0.0901,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.013181686401367,
"rewards/margins": 2.013181686401367,
"rewards/rejected": 0.0,
"step": 2433
},
{
"epoch": 13.597765363128492,
"grad_norm": 0.43246246536836536,
"learning_rate": 6.713976467968102e-08,
"logits/chosen": -3.3495986461639404,
"logits/rejected": -3.4264187812805176,
"logps/chosen": -0.4869656562805176,
"logps/rejected": -58.441123962402344,
"loss": 0.0954,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1961636543273926,
"rewards/margins": 2.1961636543273926,
"rewards/rejected": 0.0,
"step": 2434
},
{
"epoch": 13.603351955307263,
"grad_norm": 0.5290181129690829,
"learning_rate": 6.683499241939295e-08,
"logits/chosen": -3.6454989910125732,
"logits/rejected": -3.300635814666748,
"logps/chosen": -1.7670848369598389,
"logps/rejected": -61.70159912109375,
"loss": 0.108,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9578745365142822,
"rewards/margins": 2.9578745365142822,
"rewards/rejected": 0.0,
"step": 2435
},
{
"epoch": 13.608938547486034,
"grad_norm": 0.6807000900549025,
"learning_rate": 6.653086391944851e-08,
"logits/chosen": -2.8142776489257812,
"logits/rejected": -2.8938748836517334,
"logps/chosen": -41.72840118408203,
"logps/rejected": -94.65515899658203,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8236005306243896,
"rewards/margins": 2.8236005306243896,
"rewards/rejected": 0.0,
"step": 2436
},
{
"epoch": 13.614525139664805,
"grad_norm": 0.42846793246064874,
"learning_rate": 6.62273796318371e-08,
"logits/chosen": -3.383944034576416,
"logits/rejected": -3.426647424697876,
"logps/chosen": -0.9192503094673157,
"logps/rejected": -69.60661315917969,
"loss": 0.0979,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3855140209198,
"rewards/margins": 2.3855140209198,
"rewards/rejected": 0.0,
"step": 2437
},
{
"epoch": 13.620111731843576,
"grad_norm": 0.3407561917723307,
"learning_rate": 6.592454000759007e-08,
"logits/chosen": -3.527812957763672,
"logits/rejected": -3.5327913761138916,
"logps/chosen": -3.513021469116211,
"logps/rejected": -22.4188175201416,
"loss": 0.1275,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.804161548614502,
"rewards/margins": 2.804161548614502,
"rewards/rejected": 0.0,
"step": 2438
},
{
"epoch": 13.625698324022347,
"grad_norm": 1.4595379049514006,
"learning_rate": 6.56223454967813e-08,
"logits/chosen": -3.3430094718933105,
"logits/rejected": -3.432039737701416,
"logps/chosen": -0.7493870258331299,
"logps/rejected": -61.24337387084961,
"loss": 0.1196,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4020402431488037,
"rewards/margins": 2.4020402431488037,
"rewards/rejected": 0.0,
"step": 2439
},
{
"epoch": 13.631284916201118,
"grad_norm": 0.4298410453976905,
"learning_rate": 6.532079654852551e-08,
"logits/chosen": -3.794787883758545,
"logits/rejected": -3.6870479583740234,
"logps/chosen": -1.8893942832946777,
"logps/rejected": -34.15863037109375,
"loss": 0.1006,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4425315856933594,
"rewards/margins": 3.4425315856933594,
"rewards/rejected": 0.0,
"step": 2440
},
{
"epoch": 13.631284916201118,
"eval_logits/chosen": -3.2630057334899902,
"eval_logits/rejected": -3.386068820953369,
"eval_logps/chosen": -30.32887840270996,
"eval_logps/rejected": -55.73801803588867,
"eval_loss": 1.0398709774017334,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.025883492082357407,
"eval_rewards/margins": -0.025883492082357407,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6927,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 2440
},
{
"epoch": 13.636871508379889,
"grad_norm": 0.48349027312865905,
"learning_rate": 6.501989361097821e-08,
"logits/chosen": -3.492025375366211,
"logits/rejected": -3.4451024532318115,
"logps/chosen": -1.0900187492370605,
"logps/rejected": -40.548744201660156,
"loss": 0.094,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4581050872802734,
"rewards/margins": 2.4581050872802734,
"rewards/rejected": 0.0,
"step": 2441
},
{
"epoch": 13.64245810055866,
"grad_norm": 0.31571172431084366,
"learning_rate": 6.471963713133471e-08,
"logits/chosen": -3.242710828781128,
"logits/rejected": -3.1574764251708984,
"logps/chosen": -1.191739797592163,
"logps/rejected": -74.82826232910156,
"loss": 0.1194,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.147900342941284,
"rewards/margins": 2.147900342941284,
"rewards/rejected": 0.0,
"step": 2442
},
{
"epoch": 13.64804469273743,
"grad_norm": 0.3764505868960777,
"learning_rate": 6.442002755582965e-08,
"logits/chosen": -3.401458978652954,
"logits/rejected": -3.4201838970184326,
"logps/chosen": -1.2702481746673584,
"logps/rejected": -29.992412567138672,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3849074840545654,
"rewards/margins": 2.3849074840545654,
"rewards/rejected": 0.0,
"step": 2443
},
{
"epoch": 13.653631284916202,
"grad_norm": 0.3700985671014757,
"learning_rate": 6.412106532973616e-08,
"logits/chosen": -3.3580126762390137,
"logits/rejected": -3.4226503372192383,
"logps/chosen": -3.0248465538024902,
"logps/rejected": -48.47951126098633,
"loss": 0.0812,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.143270969390869,
"rewards/margins": 3.143270969390869,
"rewards/rejected": 0.0,
"step": 2444
},
{
"epoch": 13.659217877094973,
"grad_norm": 0.3835411769332013,
"learning_rate": 6.382275089736533e-08,
"logits/chosen": -3.449427604675293,
"logits/rejected": -3.3362691402435303,
"logps/chosen": -0.9317790269851685,
"logps/rejected": -48.91193389892578,
"loss": 0.0948,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.655461311340332,
"rewards/margins": 2.655461311340332,
"rewards/rejected": 0.0,
"step": 2445
},
{
"epoch": 13.664804469273744,
"grad_norm": 0.4041573936421392,
"learning_rate": 6.352508470206574e-08,
"logits/chosen": -3.378014087677002,
"logits/rejected": -3.3828506469726562,
"logps/chosen": -0.7811074256896973,
"logps/rejected": -68.95384979248047,
"loss": 0.1156,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2991526126861572,
"rewards/margins": 2.2991526126861572,
"rewards/rejected": 0.0,
"step": 2446
},
{
"epoch": 13.670391061452515,
"grad_norm": 0.381333318267043,
"learning_rate": 6.322806718622204e-08,
"logits/chosen": -3.376889944076538,
"logits/rejected": -3.3511784076690674,
"logps/chosen": -9.594871520996094,
"logps/rejected": -79.86514282226562,
"loss": 0.0936,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.468109607696533,
"rewards/margins": 3.468109607696533,
"rewards/rejected": 0.0,
"step": 2447
},
{
"epoch": 13.675977653631286,
"grad_norm": 0.43726183992411216,
"learning_rate": 6.293169879125538e-08,
"logits/chosen": -3.6175291538238525,
"logits/rejected": -3.491042375564575,
"logps/chosen": -0.26917219161987305,
"logps/rejected": -42.95506286621094,
"loss": 0.0773,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0813467502593994,
"rewards/margins": 2.0813467502593994,
"rewards/rejected": 0.0,
"step": 2448
},
{
"epoch": 13.681564245810057,
"grad_norm": 0.3479759291901259,
"learning_rate": 6.263597995762199e-08,
"logits/chosen": -3.379270553588867,
"logits/rejected": -3.3774287700653076,
"logps/chosen": -1.5853116512298584,
"logps/rejected": -22.441299438476562,
"loss": 0.0664,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8696866035461426,
"rewards/margins": 2.8696866035461426,
"rewards/rejected": 0.0,
"step": 2449
},
{
"epoch": 13.687150837988828,
"grad_norm": 0.8532118856100707,
"learning_rate": 6.23409111248126e-08,
"logits/chosen": -3.2234463691711426,
"logits/rejected": -3.2407162189483643,
"logps/chosen": -0.7098406553268433,
"logps/rejected": -96.90626525878906,
"loss": 0.1352,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1984286308288574,
"rewards/margins": 2.1984286308288574,
"rewards/rejected": 0.0,
"step": 2450
},
{
"epoch": 13.692737430167599,
"grad_norm": 0.3862349518604081,
"learning_rate": 6.204649273135209e-08,
"logits/chosen": -3.3121843338012695,
"logits/rejected": -3.284202814102173,
"logps/chosen": -6.544798851013184,
"logps/rejected": -29.416772842407227,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.399103879928589,
"rewards/margins": 3.399103879928589,
"rewards/rejected": 0.0,
"step": 2451
},
{
"epoch": 13.69832402234637,
"grad_norm": 0.4318334352488587,
"learning_rate": 6.175272521479868e-08,
"logits/chosen": -3.348818302154541,
"logits/rejected": -3.2802324295043945,
"logps/chosen": -0.7584010362625122,
"logps/rejected": -87.66751098632812,
"loss": 0.0838,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5336742401123047,
"rewards/margins": 2.5336742401123047,
"rewards/rejected": 0.0,
"step": 2452
},
{
"epoch": 13.703910614525139,
"grad_norm": 0.3912107462193857,
"learning_rate": 6.145960901174313e-08,
"logits/chosen": -3.1701173782348633,
"logits/rejected": -3.342404365539551,
"logps/chosen": -1.729007363319397,
"logps/rejected": -80.1790771484375,
"loss": 0.0918,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.575544834136963,
"rewards/margins": 2.575544834136963,
"rewards/rejected": 0.0,
"step": 2453
},
{
"epoch": 13.70949720670391,
"grad_norm": 0.33826820534948204,
"learning_rate": 6.116714455780841e-08,
"logits/chosen": -3.6022965908050537,
"logits/rejected": -3.5730085372924805,
"logps/chosen": -0.2734066843986511,
"logps/rejected": -37.372413635253906,
"loss": 0.0972,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6353272199630737,
"rewards/margins": 1.6353272199630737,
"rewards/rejected": 0.0,
"step": 2454
},
{
"epoch": 13.71508379888268,
"grad_norm": 0.4282417536358434,
"learning_rate": 6.087533228764869e-08,
"logits/chosen": -3.5499305725097656,
"logits/rejected": -3.4979584217071533,
"logps/chosen": -3.1904568672180176,
"logps/rejected": -50.225433349609375,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.881998062133789,
"rewards/margins": 2.881998062133789,
"rewards/rejected": 0.0,
"step": 2455
},
{
"epoch": 13.720670391061452,
"grad_norm": 0.4670839027146506,
"learning_rate": 6.058417263494892e-08,
"logits/chosen": -3.258791446685791,
"logits/rejected": -3.4401707649230957,
"logps/chosen": -1.52991783618927,
"logps/rejected": -37.89375305175781,
"loss": 0.0864,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6789238452911377,
"rewards/margins": 2.6789238452911377,
"rewards/rejected": 0.0,
"step": 2456
},
{
"epoch": 13.726256983240223,
"grad_norm": 0.4269025969426856,
"learning_rate": 6.029366603242453e-08,
"logits/chosen": -3.4440407752990723,
"logits/rejected": -3.2248666286468506,
"logps/chosen": -6.943201065063477,
"logps/rejected": -71.95067596435547,
"loss": 0.095,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6368632316589355,
"rewards/margins": 2.6368632316589355,
"rewards/rejected": 0.0,
"step": 2457
},
{
"epoch": 13.731843575418994,
"grad_norm": 0.47097584071658793,
"learning_rate": 6.00038129118196e-08,
"logits/chosen": -3.3258748054504395,
"logits/rejected": -3.1533684730529785,
"logps/chosen": -0.4051774740219116,
"logps/rejected": -113.83723449707031,
"loss": 0.1042,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7320462465286255,
"rewards/margins": 1.7320462465286255,
"rewards/rejected": 0.0,
"step": 2458
},
{
"epoch": 13.737430167597765,
"grad_norm": 0.3694995698827105,
"learning_rate": 5.971461370390779e-08,
"logits/chosen": -3.4326939582824707,
"logits/rejected": -3.613823652267456,
"logps/chosen": -4.124021053314209,
"logps/rejected": -35.067256927490234,
"loss": 0.0764,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1807050704956055,
"rewards/margins": 3.1807050704956055,
"rewards/rejected": 0.0,
"step": 2459
},
{
"epoch": 13.743016759776536,
"grad_norm": 0.3412927447785637,
"learning_rate": 5.942606883849061e-08,
"logits/chosen": -3.3147404193878174,
"logits/rejected": -3.4530279636383057,
"logps/chosen": -10.147854804992676,
"logps/rejected": -50.02666091918945,
"loss": 0.076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6943697929382324,
"rewards/margins": 3.6943697929382324,
"rewards/rejected": 0.0,
"step": 2460
},
{
"epoch": 13.743016759776536,
"eval_logits/chosen": -3.2568747997283936,
"eval_logits/rejected": -3.381835460662842,
"eval_logps/chosen": -30.75702476501465,
"eval_logps/rejected": -55.996238708496094,
"eval_loss": 1.0451282262802124,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.06869839131832123,
"eval_rewards/margins": -0.06869839131832123,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7157,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 2460
},
{
"epoch": 13.748603351955307,
"grad_norm": 0.5056231660460868,
"learning_rate": 5.9138178744397026e-08,
"logits/chosen": -3.45645809173584,
"logits/rejected": -3.566572666168213,
"logps/chosen": -0.503404438495636,
"logps/rejected": -55.78275680541992,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.135225772857666,
"rewards/margins": 2.135225772857666,
"rewards/rejected": 0.0,
"step": 2461
},
{
"epoch": 13.754189944134078,
"grad_norm": 0.38893605521270735,
"learning_rate": 5.8850943849483024e-08,
"logits/chosen": -3.5577340126037598,
"logits/rejected": -3.289719820022583,
"logps/chosen": -1.7229567766189575,
"logps/rejected": -35.91162109375,
"loss": 0.0972,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.82952880859375,
"rewards/margins": 2.82952880859375,
"rewards/rejected": 0.0,
"step": 2462
},
{
"epoch": 13.759776536312849,
"grad_norm": 1.391989449429673,
"learning_rate": 5.856436458063085e-08,
"logits/chosen": -2.970513343811035,
"logits/rejected": -3.20465350151062,
"logps/chosen": -31.5120792388916,
"logps/rejected": -67.71055603027344,
"loss": 0.1104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.864408254623413,
"rewards/margins": 2.864408254623413,
"rewards/rejected": 0.0,
"step": 2463
},
{
"epoch": 13.76536312849162,
"grad_norm": 0.3956122948278953,
"learning_rate": 5.8278441363748266e-08,
"logits/chosen": -3.3388075828552246,
"logits/rejected": -3.2895302772521973,
"logps/chosen": -1.098218560218811,
"logps/rejected": -84.32713317871094,
"loss": 0.0985,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.219403028488159,
"rewards/margins": 2.219403028488159,
"rewards/rejected": 0.0,
"step": 2464
},
{
"epoch": 13.77094972067039,
"grad_norm": 0.5040526789435122,
"learning_rate": 5.799317462376802e-08,
"logits/chosen": -3.611959457397461,
"logits/rejected": -3.395556688308716,
"logps/chosen": -9.191370964050293,
"logps/rejected": -23.500995635986328,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.60721755027771,
"rewards/margins": 3.60721755027771,
"rewards/rejected": 0.0,
"step": 2465
},
{
"epoch": 13.776536312849162,
"grad_norm": 0.3623352437962981,
"learning_rate": 5.7708564784647585e-08,
"logits/chosen": -3.712226629257202,
"logits/rejected": -3.5574452877044678,
"logps/chosen": -2.2654924392700195,
"logps/rejected": -69.63509368896484,
"loss": 0.0815,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.731916904449463,
"rewards/margins": 2.731916904449463,
"rewards/rejected": 0.0,
"step": 2466
},
{
"epoch": 13.782122905027933,
"grad_norm": 0.4276455071005255,
"learning_rate": 5.742461226936746e-08,
"logits/chosen": -3.530557632446289,
"logits/rejected": -3.3399665355682373,
"logps/chosen": -15.764871597290039,
"logps/rejected": -51.784812927246094,
"loss": 0.0619,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7098512649536133,
"rewards/margins": 3.7098512649536133,
"rewards/rejected": 0.0,
"step": 2467
},
{
"epoch": 13.787709497206704,
"grad_norm": 1.0922418529887215,
"learning_rate": 5.714131749993201e-08,
"logits/chosen": -3.275423049926758,
"logits/rejected": -3.168363094329834,
"logps/chosen": -6.508008003234863,
"logps/rejected": -30.521949768066406,
"loss": 0.0678,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.0412983894348145,
"rewards/margins": 4.0412983894348145,
"rewards/rejected": 0.0,
"step": 2468
},
{
"epoch": 13.793296089385475,
"grad_norm": 0.3808300020693675,
"learning_rate": 5.685868089736756e-08,
"logits/chosen": -3.273233652114868,
"logits/rejected": -3.2076663970947266,
"logps/chosen": -2.4366536140441895,
"logps/rejected": -45.89551544189453,
"loss": 0.0669,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.258554458618164,
"rewards/margins": 3.258554458618164,
"rewards/rejected": 0.0,
"step": 2469
},
{
"epoch": 13.798882681564246,
"grad_norm": 0.3625829305182228,
"learning_rate": 5.657670288172245e-08,
"logits/chosen": -3.233433246612549,
"logits/rejected": -3.353445053100586,
"logps/chosen": -0.8617339730262756,
"logps/rejected": -46.014373779296875,
"loss": 0.096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.202939987182617,
"rewards/margins": 2.202939987182617,
"rewards/rejected": 0.0,
"step": 2470
},
{
"epoch": 13.804469273743017,
"grad_norm": 0.7421793665656019,
"learning_rate": 5.6295383872066174e-08,
"logits/chosen": -3.3089280128479004,
"logits/rejected": -3.4478418827056885,
"logps/chosen": -1.426999807357788,
"logps/rejected": -57.93741226196289,
"loss": 0.0941,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1230382919311523,
"rewards/margins": 3.1230382919311523,
"rewards/rejected": 0.0,
"step": 2471
},
{
"epoch": 13.810055865921788,
"grad_norm": 0.5560215504161522,
"learning_rate": 5.601472428648901e-08,
"logits/chosen": -3.2834036350250244,
"logits/rejected": -3.293836832046509,
"logps/chosen": -5.384027004241943,
"logps/rejected": -43.36772918701172,
"loss": 0.0979,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.096794605255127,
"rewards/margins": 3.096794605255127,
"rewards/rejected": 0.0,
"step": 2472
},
{
"epoch": 13.815642458100559,
"grad_norm": 0.48068691739954705,
"learning_rate": 5.573472454210099e-08,
"logits/chosen": -3.3021554946899414,
"logits/rejected": -3.2614760398864746,
"logps/chosen": -0.7331230640411377,
"logps/rejected": -48.53559875488281,
"loss": 0.1139,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.484909772872925,
"rewards/margins": 2.484909772872925,
"rewards/rejected": 0.0,
"step": 2473
},
{
"epoch": 13.82122905027933,
"grad_norm": 0.4228967725332,
"learning_rate": 5.54553850550315e-08,
"logits/chosen": -3.1667332649230957,
"logits/rejected": -3.569082260131836,
"logps/chosen": -4.470986366271973,
"logps/rejected": -49.33929443359375,
"loss": 0.1316,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8256728649139404,
"rewards/margins": 2.8256728649139404,
"rewards/rejected": 0.0,
"step": 2474
},
{
"epoch": 13.8268156424581,
"grad_norm": 0.5758302695747335,
"learning_rate": 5.517670624042908e-08,
"logits/chosen": -3.196789503097534,
"logits/rejected": -3.2065672874450684,
"logps/chosen": -2.022822856903076,
"logps/rejected": -60.15814208984375,
"loss": 0.1015,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.174039363861084,
"rewards/margins": 2.174039363861084,
"rewards/rejected": 0.0,
"step": 2475
},
{
"epoch": 13.832402234636872,
"grad_norm": 0.3537566198334273,
"learning_rate": 5.489868851245966e-08,
"logits/chosen": -3.4229576587677,
"logits/rejected": -3.4131813049316406,
"logps/chosen": -0.5902296304702759,
"logps/rejected": -44.242584228515625,
"loss": 0.1254,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.101193428039551,
"rewards/margins": 2.101193428039551,
"rewards/rejected": 0.0,
"step": 2476
},
{
"epoch": 13.837988826815643,
"grad_norm": 0.5164478856248721,
"learning_rate": 5.4621332284307485e-08,
"logits/chosen": -3.5133912563323975,
"logits/rejected": -3.3796379566192627,
"logps/chosen": -0.5123422145843506,
"logps/rejected": -72.57955932617188,
"loss": 0.1038,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9728453159332275,
"rewards/margins": 1.9728453159332275,
"rewards/rejected": 0.0,
"step": 2477
},
{
"epoch": 13.843575418994414,
"grad_norm": 0.34060610282066395,
"learning_rate": 5.4344637968173066e-08,
"logits/chosen": -3.4484715461730957,
"logits/rejected": -3.355842113494873,
"logps/chosen": -4.761620044708252,
"logps/rejected": -63.75336456298828,
"loss": 0.1036,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.49881649017334,
"rewards/margins": 3.49881649017334,
"rewards/rejected": 0.0,
"step": 2478
},
{
"epoch": 13.849162011173185,
"grad_norm": 0.36252659299686363,
"learning_rate": 5.406860597527352e-08,
"logits/chosen": -3.4152116775512695,
"logits/rejected": -3.427280902862549,
"logps/chosen": -1.1045894622802734,
"logps/rejected": -71.25437927246094,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2757999897003174,
"rewards/margins": 2.2757999897003174,
"rewards/rejected": 0.0,
"step": 2479
},
{
"epoch": 13.854748603351956,
"grad_norm": 0.4692979948850874,
"learning_rate": 5.379323671584157e-08,
"logits/chosen": -3.529876470565796,
"logits/rejected": -3.5923233032226562,
"logps/chosen": -28.87140655517578,
"logps/rejected": -52.27913284301758,
"loss": 0.0781,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.453380584716797,
"rewards/margins": 4.453380584716797,
"rewards/rejected": 0.0,
"step": 2480
},
{
"epoch": 13.854748603351956,
"eval_logits/chosen": -3.259183168411255,
"eval_logits/rejected": -3.3825583457946777,
"eval_logps/chosen": -30.66366195678711,
"eval_logps/rejected": -55.99504852294922,
"eval_loss": 1.0457611083984375,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.059362124651670456,
"eval_rewards/margins": -0.059362124651670456,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7272,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 2480
},
{
"epoch": 13.860335195530727,
"grad_norm": 0.3591490248366364,
"learning_rate": 5.351853059912492e-08,
"logits/chosen": -3.2581000328063965,
"logits/rejected": -3.3914709091186523,
"logps/chosen": -3.7509045600891113,
"logps/rejected": -85.30378723144531,
"loss": 0.1322,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.855607509613037,
"rewards/margins": 2.855607509613037,
"rewards/rejected": 0.0,
"step": 2481
},
{
"epoch": 13.865921787709498,
"grad_norm": 0.42055219665901744,
"learning_rate": 5.324448803338577e-08,
"logits/chosen": -3.6724579334259033,
"logits/rejected": -3.55611515045166,
"logps/chosen": -1.2097103595733643,
"logps/rejected": -57.648094177246094,
"loss": 0.1112,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.73215389251709,
"rewards/margins": 2.73215389251709,
"rewards/rejected": 0.0,
"step": 2482
},
{
"epoch": 13.871508379888269,
"grad_norm": 0.3887641847552534,
"learning_rate": 5.297110942590027e-08,
"logits/chosen": -3.3861851692199707,
"logits/rejected": -3.496201276779175,
"logps/chosen": -4.638120651245117,
"logps/rejected": -32.964237213134766,
"loss": 0.1047,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4208145141601562,
"rewards/margins": 3.4208145141601562,
"rewards/rejected": 0.0,
"step": 2483
},
{
"epoch": 13.87709497206704,
"grad_norm": 0.5033224240333831,
"learning_rate": 5.26983951829576e-08,
"logits/chosen": -3.4654667377471924,
"logits/rejected": -3.3526651859283447,
"logps/chosen": -19.81593132019043,
"logps/rejected": -29.83510971069336,
"loss": 0.0788,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.035046577453613,
"rewards/margins": 4.035046577453613,
"rewards/rejected": 0.0,
"step": 2484
},
{
"epoch": 13.88268156424581,
"grad_norm": 0.5211346855195558,
"learning_rate": 5.242634570985966e-08,
"logits/chosen": -3.5142388343811035,
"logits/rejected": -3.537492275238037,
"logps/chosen": -9.576385498046875,
"logps/rejected": -43.98325729370117,
"loss": 0.1114,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6981780529022217,
"rewards/margins": 2.6981780529022217,
"rewards/rejected": 0.0,
"step": 2485
},
{
"epoch": 13.888268156424582,
"grad_norm": 0.394665246695912,
"learning_rate": 5.215496141092074e-08,
"logits/chosen": -3.624476671218872,
"logits/rejected": -3.6229772567749023,
"logps/chosen": -1.8939690589904785,
"logps/rejected": -43.50250244140625,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.396291494369507,
"rewards/margins": 3.396291494369507,
"rewards/rejected": 0.0,
"step": 2486
},
{
"epoch": 13.893854748603353,
"grad_norm": 0.4739673899151669,
"learning_rate": 5.188424268946573e-08,
"logits/chosen": -3.3838634490966797,
"logits/rejected": -3.3739559650421143,
"logps/chosen": -7.8452277183532715,
"logps/rejected": -49.918739318847656,
"loss": 0.1023,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.364142417907715,
"rewards/margins": 3.364142417907715,
"rewards/rejected": 0.0,
"step": 2487
},
{
"epoch": 13.899441340782122,
"grad_norm": 0.4319820305631335,
"learning_rate": 5.161418994783129e-08,
"logits/chosen": -3.3190858364105225,
"logits/rejected": -3.4019381999969482,
"logps/chosen": -5.985474586486816,
"logps/rejected": -72.33639526367188,
"loss": 0.0974,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8086514472961426,
"rewards/margins": 2.8086514472961426,
"rewards/rejected": 0.0,
"step": 2488
},
{
"epoch": 13.905027932960895,
"grad_norm": 0.5339954859468932,
"learning_rate": 5.13448035873637e-08,
"logits/chosen": -3.6297895908355713,
"logits/rejected": -3.5210063457489014,
"logps/chosen": -17.91924285888672,
"logps/rejected": -60.66377639770508,
"loss": 0.0889,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.580571174621582,
"rewards/margins": 3.580571174621582,
"rewards/rejected": 0.0,
"step": 2489
},
{
"epoch": 13.910614525139664,
"grad_norm": 0.4504867638513347,
"learning_rate": 5.107608400841912e-08,
"logits/chosen": -3.335150718688965,
"logits/rejected": -3.196349859237671,
"logps/chosen": -5.946123123168945,
"logps/rejected": -45.173431396484375,
"loss": 0.1225,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9648706912994385,
"rewards/margins": 3.9648706912994385,
"rewards/rejected": 0.0,
"step": 2490
},
{
"epoch": 13.916201117318435,
"grad_norm": 0.34157498460295826,
"learning_rate": 5.080803161036268e-08,
"logits/chosen": -3.4631834030151367,
"logits/rejected": -3.5812928676605225,
"logps/chosen": -0.24410250782966614,
"logps/rejected": -34.043128967285156,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.688918113708496,
"rewards/margins": 1.688918113708496,
"rewards/rejected": 0.0,
"step": 2491
},
{
"epoch": 13.921787709497206,
"grad_norm": 0.4207208697244406,
"learning_rate": 5.054064679156794e-08,
"logits/chosen": -3.413464069366455,
"logits/rejected": -3.4899497032165527,
"logps/chosen": -0.3154837191104889,
"logps/rejected": -39.27338409423828,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7795913219451904,
"rewards/margins": 1.7795913219451904,
"rewards/rejected": 0.0,
"step": 2492
},
{
"epoch": 13.927374301675977,
"grad_norm": 0.36446031525425,
"learning_rate": 5.0273929949416425e-08,
"logits/chosen": -3.2451226711273193,
"logits/rejected": -3.5481081008911133,
"logps/chosen": -0.687081515789032,
"logps/rejected": -55.72673416137695,
"loss": 0.1032,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5647451877593994,
"rewards/margins": 2.5647451877593994,
"rewards/rejected": 0.0,
"step": 2493
},
{
"epoch": 13.932960893854748,
"grad_norm": 1.5084722771222736,
"learning_rate": 5.0007881480296777e-08,
"logits/chosen": -3.366252899169922,
"logits/rejected": -3.440880537033081,
"logps/chosen": -0.37535303831100464,
"logps/rejected": -87.42872619628906,
"loss": 0.1351,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7993223667144775,
"rewards/margins": 1.7993223667144775,
"rewards/rejected": 0.0,
"step": 2494
},
{
"epoch": 13.938547486033519,
"grad_norm": 0.5251464349693189,
"learning_rate": 4.974250177960459e-08,
"logits/chosen": -3.3979170322418213,
"logits/rejected": -3.4452855587005615,
"logps/chosen": -9.844120025634766,
"logps/rejected": -55.45526123046875,
"loss": 0.0968,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4103057384490967,
"rewards/margins": 3.4103057384490967,
"rewards/rejected": 0.0,
"step": 2495
},
{
"epoch": 13.94413407821229,
"grad_norm": 0.4172781926412672,
"learning_rate": 4.9477791241741116e-08,
"logits/chosen": -3.177208185195923,
"logits/rejected": -3.424086809158325,
"logps/chosen": -30.380510330200195,
"logps/rejected": -44.14557647705078,
"loss": 0.1112,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7149908542633057,
"rewards/margins": 2.7149908542633057,
"rewards/rejected": 0.0,
"step": 2496
},
{
"epoch": 13.949720670391061,
"grad_norm": 0.48841241166711075,
"learning_rate": 4.921375026011349e-08,
"logits/chosen": -3.4830703735351562,
"logits/rejected": -3.333214521408081,
"logps/chosen": -1.4240703582763672,
"logps/rejected": -56.699066162109375,
"loss": 0.0829,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1017422676086426,
"rewards/margins": 2.1017422676086426,
"rewards/rejected": 0.0,
"step": 2497
},
{
"epoch": 13.955307262569832,
"grad_norm": 0.6062750944641184,
"learning_rate": 4.895037922713369e-08,
"logits/chosen": -3.237408399581909,
"logits/rejected": -3.168778419494629,
"logps/chosen": -0.853169858455658,
"logps/rejected": -77.13175964355469,
"loss": 0.127,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.408933162689209,
"rewards/margins": 2.408933162689209,
"rewards/rejected": 0.0,
"step": 2498
},
{
"epoch": 13.960893854748603,
"grad_norm": 0.5009395657043406,
"learning_rate": 4.8687678534217845e-08,
"logits/chosen": -3.6134917736053467,
"logits/rejected": -3.5716700553894043,
"logps/chosen": -0.602191150188446,
"logps/rejected": -55.38518524169922,
"loss": 0.0896,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.096320629119873,
"rewards/margins": 2.096320629119873,
"rewards/rejected": 0.0,
"step": 2499
},
{
"epoch": 13.966480446927374,
"grad_norm": 0.5193637019708659,
"learning_rate": 4.842564857178605e-08,
"logits/chosen": -3.6649270057678223,
"logits/rejected": -3.5219991207122803,
"logps/chosen": -6.3779802322387695,
"logps/rejected": -44.12037658691406,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.800743341445923,
"rewards/margins": 3.800743341445923,
"rewards/rejected": 0.0,
"step": 2500
},
{
"epoch": 13.966480446927374,
"eval_logits/chosen": -3.26655912399292,
"eval_logits/rejected": -3.389557361602783,
"eval_logps/chosen": -30.509235382080078,
"eval_logps/rejected": -55.891700744628906,
"eval_loss": 1.043331265449524,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.043919093906879425,
"eval_rewards/margins": -0.043919093906879425,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.705,
"eval_samples_per_second": 9.479,
"eval_steps_per_second": 0.306,
"step": 2500
},
{
"epoch": 13.972067039106145,
"grad_norm": 1.3732558235325156,
"learning_rate": 4.8164289729261456e-08,
"logits/chosen": -3.3140311241149902,
"logits/rejected": -3.2370216846466064,
"logps/chosen": -9.072028160095215,
"logps/rejected": -29.328149795532227,
"loss": 0.1126,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3890528678894043,
"rewards/margins": 3.3890528678894043,
"rewards/rejected": 0.0,
"step": 2501
},
{
"epoch": 13.977653631284916,
"grad_norm": 0.4150946977059628,
"learning_rate": 4.7903602395069854e-08,
"logits/chosen": -3.29394268989563,
"logits/rejected": -3.1019275188446045,
"logps/chosen": -0.9298399090766907,
"logps/rejected": -29.049501419067383,
"loss": 0.1016,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.187602996826172,
"rewards/margins": 2.187602996826172,
"rewards/rejected": 0.0,
"step": 2502
},
{
"epoch": 13.983240223463687,
"grad_norm": 2.046627753222545,
"learning_rate": 4.7643586956639004e-08,
"logits/chosen": -3.40679669380188,
"logits/rejected": -3.2866289615631104,
"logps/chosen": -1.019418716430664,
"logps/rejected": -25.22226333618164,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2473247051239014,
"rewards/margins": 2.2473247051239014,
"rewards/rejected": 0.0,
"step": 2503
},
{
"epoch": 13.988826815642458,
"grad_norm": 0.4593675037651337,
"learning_rate": 4.738424380039835e-08,
"logits/chosen": -3.554079055786133,
"logits/rejected": -3.388239860534668,
"logps/chosen": -3.9275014400482178,
"logps/rejected": -40.449485778808594,
"loss": 0.1013,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1833853721618652,
"rewards/margins": 3.1833853721618652,
"rewards/rejected": 0.0,
"step": 2504
},
{
"epoch": 13.994413407821229,
"grad_norm": 0.380511098651879,
"learning_rate": 4.712557331177775e-08,
"logits/chosen": -3.6009199619293213,
"logits/rejected": -3.547635078430176,
"logps/chosen": -0.5908699035644531,
"logps/rejected": -67.65447235107422,
"loss": 0.0844,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.278252601623535,
"rewards/margins": 2.278252601623535,
"rewards/rejected": 0.0,
"step": 2505
},
{
"epoch": 14.0,
"grad_norm": 0.6200767041535287,
"learning_rate": 4.686757587520773e-08,
"logits/chosen": -3.1858692169189453,
"logits/rejected": -3.4097728729248047,
"logps/chosen": -0.5349952578544617,
"logps/rejected": -65.05491638183594,
"loss": 0.149,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9272640943527222,
"rewards/margins": 1.9272640943527222,
"rewards/rejected": 0.0,
"step": 2506
},
{
"epoch": 14.005586592178771,
"grad_norm": 0.39174736869501914,
"learning_rate": 4.661025187411849e-08,
"logits/chosen": -2.706138849258423,
"logits/rejected": -2.865696430206299,
"logps/chosen": -25.698196411132812,
"logps/rejected": -51.548553466796875,
"loss": 0.1079,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.400902271270752,
"rewards/margins": 3.400902271270752,
"rewards/rejected": 0.0,
"step": 2507
},
{
"epoch": 14.011173184357542,
"grad_norm": 0.4515259932811653,
"learning_rate": 4.635360169093927e-08,
"logits/chosen": -3.237030506134033,
"logits/rejected": -3.3163609504699707,
"logps/chosen": -0.3199694752693176,
"logps/rejected": -51.186988830566406,
"loss": 0.0837,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9741865396499634,
"rewards/margins": 1.9741865396499634,
"rewards/rejected": 0.0,
"step": 2508
},
{
"epoch": 14.016759776536313,
"grad_norm": 0.34533738462280605,
"learning_rate": 4.6097625707097906e-08,
"logits/chosen": -3.1123125553131104,
"logits/rejected": -3.3256561756134033,
"logps/chosen": -10.399734497070312,
"logps/rejected": -43.868534088134766,
"loss": 0.1201,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3708457946777344,
"rewards/margins": 3.3708457946777344,
"rewards/rejected": 0.0,
"step": 2509
},
{
"epoch": 14.022346368715084,
"grad_norm": 0.5019482966361488,
"learning_rate": 4.584232430302032e-08,
"logits/chosen": -3.5791730880737305,
"logits/rejected": -3.4997928142547607,
"logps/chosen": -8.169157028198242,
"logps/rejected": -37.524234771728516,
"loss": 0.0758,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1666946411132812,
"rewards/margins": 3.1666946411132812,
"rewards/rejected": 0.0,
"step": 2510
},
{
"epoch": 14.027932960893855,
"grad_norm": 0.36052654630885134,
"learning_rate": 4.5587697858129936e-08,
"logits/chosen": -3.343183994293213,
"logits/rejected": -3.2706618309020996,
"logps/chosen": -5.515871047973633,
"logps/rejected": -42.919532775878906,
"loss": 0.1026,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.694736957550049,
"rewards/margins": 3.694736957550049,
"rewards/rejected": 0.0,
"step": 2511
},
{
"epoch": 14.033519553072626,
"grad_norm": 0.4548371099956356,
"learning_rate": 4.533374675084689e-08,
"logits/chosen": -3.6490325927734375,
"logits/rejected": -3.58361554145813,
"logps/chosen": -3.832155466079712,
"logps/rejected": -53.35199737548828,
"loss": 0.0833,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9735910892486572,
"rewards/margins": 2.9735910892486572,
"rewards/rejected": 0.0,
"step": 2512
},
{
"epoch": 14.039106145251397,
"grad_norm": 0.31969567415625777,
"learning_rate": 4.508047135858778e-08,
"logits/chosen": -2.882420063018799,
"logits/rejected": -2.9554548263549805,
"logps/chosen": -2.9381048679351807,
"logps/rejected": -37.72036361694336,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9198451042175293,
"rewards/margins": 2.9198451042175293,
"rewards/rejected": 0.0,
"step": 2513
},
{
"epoch": 14.044692737430168,
"grad_norm": 0.4494046124884269,
"learning_rate": 4.482787205776495e-08,
"logits/chosen": -3.3964457511901855,
"logits/rejected": -3.201207160949707,
"logps/chosen": -0.42687922716140747,
"logps/rejected": -38.09913635253906,
"loss": 0.1204,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7753281593322754,
"rewards/margins": 1.7753281593322754,
"rewards/rejected": 0.0,
"step": 2514
},
{
"epoch": 14.050279329608939,
"grad_norm": 0.39334101637067403,
"learning_rate": 4.45759492237861e-08,
"logits/chosen": -3.2897698879241943,
"logits/rejected": -3.1025137901306152,
"logps/chosen": -3.072049617767334,
"logps/rejected": -51.14334487915039,
"loss": 0.0997,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.66015887260437,
"rewards/margins": 2.66015887260437,
"rewards/rejected": 0.0,
"step": 2515
},
{
"epoch": 14.05586592178771,
"grad_norm": 0.44879904847492463,
"learning_rate": 4.432470323105309e-08,
"logits/chosen": -3.555020332336426,
"logits/rejected": -3.409587860107422,
"logps/chosen": -2.543583393096924,
"logps/rejected": -87.54447174072266,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9039902687072754,
"rewards/margins": 2.9039902687072754,
"rewards/rejected": 0.0,
"step": 2516
},
{
"epoch": 14.061452513966481,
"grad_norm": 0.3836887953363112,
"learning_rate": 4.407413445296254e-08,
"logits/chosen": -3.2767131328582764,
"logits/rejected": -3.217470169067383,
"logps/chosen": -1.1344311237335205,
"logps/rejected": -57.810550689697266,
"loss": 0.1074,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.732820510864258,
"rewards/margins": 2.732820510864258,
"rewards/rejected": 0.0,
"step": 2517
},
{
"epoch": 14.067039106145252,
"grad_norm": 0.45793046586598585,
"learning_rate": 4.3824243261904146e-08,
"logits/chosen": -3.396780490875244,
"logits/rejected": -3.420109510421753,
"logps/chosen": -0.9700617790222168,
"logps/rejected": -62.61637496948242,
"loss": 0.126,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1917288303375244,
"rewards/margins": 2.1917288303375244,
"rewards/rejected": 0.0,
"step": 2518
},
{
"epoch": 14.072625698324023,
"grad_norm": 0.40737209774679406,
"learning_rate": 4.3575030029260715e-08,
"logits/chosen": -3.4730730056762695,
"logits/rejected": -3.267848491668701,
"logps/chosen": -13.06028938293457,
"logps/rejected": -40.35478973388672,
"loss": 0.0829,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1192760467529297,
"rewards/margins": 3.1192760467529297,
"rewards/rejected": 0.0,
"step": 2519
},
{
"epoch": 14.078212290502794,
"grad_norm": 0.3516096457488634,
"learning_rate": 4.332649512540748e-08,
"logits/chosen": -3.439720392227173,
"logits/rejected": -3.3801205158233643,
"logps/chosen": -1.9102089405059814,
"logps/rejected": -59.78977966308594,
"loss": 0.1012,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7264504432678223,
"rewards/margins": 2.7264504432678223,
"rewards/rejected": 0.0,
"step": 2520
},
{
"epoch": 14.078212290502794,
"eval_logits/chosen": -3.254176616668701,
"eval_logits/rejected": -3.378744602203369,
"eval_logps/chosen": -30.575481414794922,
"eval_logps/rejected": -56.211090087890625,
"eval_loss": 1.0447423458099365,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.05054422467947006,
"eval_rewards/margins": -0.05054422467947006,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7152,
"eval_samples_per_second": 9.476,
"eval_steps_per_second": 0.306,
"step": 2520
},
{
"epoch": 14.083798882681565,
"grad_norm": 0.47111034592068946,
"learning_rate": 4.307863891971164e-08,
"logits/chosen": -3.168804407119751,
"logits/rejected": -3.2795393466949463,
"logps/chosen": -11.020282745361328,
"logps/rejected": -35.376094818115234,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.0943098068237305,
"rewards/margins": 4.0943098068237305,
"rewards/rejected": 0.0,
"step": 2521
},
{
"epoch": 14.089385474860336,
"grad_norm": 0.40081552329055525,
"learning_rate": 4.283146178053154e-08,
"logits/chosen": -3.5196638107299805,
"logits/rejected": -3.485856771469116,
"logps/chosen": -0.3709907829761505,
"logps/rejected": -83.14396667480469,
"loss": 0.0767,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1558685302734375,
"rewards/margins": 2.1558685302734375,
"rewards/rejected": 0.0,
"step": 2522
},
{
"epoch": 14.094972067039107,
"grad_norm": 0.34594186164033264,
"learning_rate": 4.258496407521644e-08,
"logits/chosen": -3.193652391433716,
"logits/rejected": -3.2667601108551025,
"logps/chosen": -27.787944793701172,
"logps/rejected": -39.35036087036133,
"loss": 0.0808,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.521094799041748,
"rewards/margins": 3.521094799041748,
"rewards/rejected": 0.0,
"step": 2523
},
{
"epoch": 14.100558659217878,
"grad_norm": 0.3721317257821369,
"learning_rate": 4.2339146170106086e-08,
"logits/chosen": -3.4976680278778076,
"logits/rejected": -3.675879716873169,
"logps/chosen": -0.825690746307373,
"logps/rejected": -30.738506317138672,
"loss": 0.0961,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6704134941101074,
"rewards/margins": 2.6704134941101074,
"rewards/rejected": 0.0,
"step": 2524
},
{
"epoch": 14.106145251396647,
"grad_norm": 0.38951100725223997,
"learning_rate": 4.209400843052924e-08,
"logits/chosen": -3.378603458404541,
"logits/rejected": -3.3314805030822754,
"logps/chosen": -1.0683079957962036,
"logps/rejected": -33.909095764160156,
"loss": 0.0852,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.552832841873169,
"rewards/margins": 2.552832841873169,
"rewards/rejected": 0.0,
"step": 2525
},
{
"epoch": 14.111731843575418,
"grad_norm": 0.43354575033828296,
"learning_rate": 4.184955122080458e-08,
"logits/chosen": -3.3340320587158203,
"logits/rejected": -3.439610004425049,
"logps/chosen": -0.18201176822185516,
"logps/rejected": -59.50272750854492,
"loss": 0.1094,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.106680154800415,
"rewards/margins": 1.106680154800415,
"rewards/rejected": 0.0,
"step": 2526
},
{
"epoch": 14.11731843575419,
"grad_norm": 0.663533112419587,
"learning_rate": 4.160577490423894e-08,
"logits/chosen": -3.37919282913208,
"logits/rejected": -3.4174959659576416,
"logps/chosen": -4.811269760131836,
"logps/rejected": -30.667572021484375,
"loss": 0.1119,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9435300827026367,
"rewards/margins": 2.9435300827026367,
"rewards/rejected": 0.0,
"step": 2527
},
{
"epoch": 14.12290502793296,
"grad_norm": 0.3576851415028276,
"learning_rate": 4.1362679843127356e-08,
"logits/chosen": -3.211427688598633,
"logits/rejected": -3.2749810218811035,
"logps/chosen": -28.984590530395508,
"logps/rejected": -51.678653717041016,
"loss": 0.1046,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3913588523864746,
"rewards/margins": 3.3913588523864746,
"rewards/rejected": 0.0,
"step": 2528
},
{
"epoch": 14.128491620111731,
"grad_norm": 0.4741016444318694,
"learning_rate": 4.1120266398752326e-08,
"logits/chosen": -3.5302581787109375,
"logits/rejected": -3.514535427093506,
"logps/chosen": -8.391327857971191,
"logps/rejected": -37.496742248535156,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.247954368591309,
"rewards/margins": 4.247954368591309,
"rewards/rejected": 0.0,
"step": 2529
},
{
"epoch": 14.134078212290502,
"grad_norm": 0.4395296616487134,
"learning_rate": 4.0878534931383435e-08,
"logits/chosen": -3.499288558959961,
"logits/rejected": -3.4895715713500977,
"logps/chosen": -1.2045276165008545,
"logps/rejected": -32.941505432128906,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3039512634277344,
"rewards/margins": 2.3039512634277344,
"rewards/rejected": 0.0,
"step": 2530
},
{
"epoch": 14.139664804469273,
"grad_norm": 0.35539218186692995,
"learning_rate": 4.063748580027676e-08,
"logits/chosen": -3.5145745277404785,
"logits/rejected": -3.4668242931365967,
"logps/chosen": -0.7444800138473511,
"logps/rejected": -56.768516540527344,
"loss": 0.0792,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1890501976013184,
"rewards/margins": 2.1890501976013184,
"rewards/rejected": 0.0,
"step": 2531
},
{
"epoch": 14.145251396648044,
"grad_norm": 0.5206543439377311,
"learning_rate": 4.0397119363674015e-08,
"logits/chosen": -3.6280972957611084,
"logits/rejected": -3.4742085933685303,
"logps/chosen": -8.813718795776367,
"logps/rejected": -34.00957489013672,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.191366195678711,
"rewards/margins": 4.191366195678711,
"rewards/rejected": 0.0,
"step": 2532
},
{
"epoch": 14.150837988826815,
"grad_norm": 0.43278429496198656,
"learning_rate": 4.015743597880289e-08,
"logits/chosen": -3.387723207473755,
"logits/rejected": -3.6379129886627197,
"logps/chosen": -0.9776702523231506,
"logps/rejected": -97.9456787109375,
"loss": 0.1267,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0290231704711914,
"rewards/margins": 2.0290231704711914,
"rewards/rejected": 0.0,
"step": 2533
},
{
"epoch": 14.156424581005586,
"grad_norm": 0.5068155687807977,
"learning_rate": 3.991843600187522e-08,
"logits/chosen": -3.269091844558716,
"logits/rejected": -3.4184460639953613,
"logps/chosen": -1.414226770401001,
"logps/rejected": -63.77281188964844,
"loss": 0.1273,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5663299560546875,
"rewards/margins": 2.5663299560546875,
"rewards/rejected": 0.0,
"step": 2534
},
{
"epoch": 14.162011173184357,
"grad_norm": 0.40955927938213654,
"learning_rate": 3.968011978808783e-08,
"logits/chosen": -3.5155231952667236,
"logits/rejected": -3.675851345062256,
"logps/chosen": -0.7108304500579834,
"logps/rejected": -74.70698547363281,
"loss": 0.1118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.098842144012451,
"rewards/margins": 2.098842144012451,
"rewards/rejected": 0.0,
"step": 2535
},
{
"epoch": 14.167597765363128,
"grad_norm": 0.40162096835373096,
"learning_rate": 3.944248769162079e-08,
"logits/chosen": -3.493814706802368,
"logits/rejected": -3.4598453044891357,
"logps/chosen": -21.722007751464844,
"logps/rejected": -58.04270553588867,
"loss": 0.0687,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4508676528930664,
"rewards/margins": 3.4508676528930664,
"rewards/rejected": 0.0,
"step": 2536
},
{
"epoch": 14.1731843575419,
"grad_norm": 0.43248293866437015,
"learning_rate": 3.920554006563792e-08,
"logits/chosen": -3.0406687259674072,
"logits/rejected": -2.966635227203369,
"logps/chosen": -1.9222922325134277,
"logps/rejected": -38.26771545410156,
"loss": 0.105,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8196706771850586,
"rewards/margins": 2.8196706771850586,
"rewards/rejected": 0.0,
"step": 2537
},
{
"epoch": 14.17877094972067,
"grad_norm": 0.40370307551024015,
"learning_rate": 3.8969277262285495e-08,
"logits/chosen": -3.1226117610931396,
"logits/rejected": -3.253596782684326,
"logps/chosen": -7.3527936935424805,
"logps/rejected": -64.65121459960938,
"loss": 0.0959,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6517767906188965,
"rewards/margins": 2.6517767906188965,
"rewards/rejected": 0.0,
"step": 2538
},
{
"epoch": 14.184357541899441,
"grad_norm": 0.42280144704831885,
"learning_rate": 3.873369963269219e-08,
"logits/chosen": -3.481581687927246,
"logits/rejected": -3.59002947807312,
"logps/chosen": -0.7101815938949585,
"logps/rejected": -84.5839614868164,
"loss": 0.0849,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.281764030456543,
"rewards/margins": 2.281764030456543,
"rewards/rejected": 0.0,
"step": 2539
},
{
"epoch": 14.189944134078212,
"grad_norm": 0.3405423202530319,
"learning_rate": 3.8498807526968215e-08,
"logits/chosen": -3.2610485553741455,
"logits/rejected": -3.4200282096862793,
"logps/chosen": -0.4112340807914734,
"logps/rejected": -55.193363189697266,
"loss": 0.1257,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.174483299255371,
"rewards/margins": 2.174483299255371,
"rewards/rejected": 0.0,
"step": 2540
},
{
"epoch": 14.189944134078212,
"eval_logits/chosen": -3.257338285446167,
"eval_logits/rejected": -3.381732940673828,
"eval_logps/chosen": -30.871402740478516,
"eval_logps/rejected": -56.050811767578125,
"eval_loss": 1.0482580661773682,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.08013595640659332,
"eval_rewards/margins": -0.08013595640659332,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6937,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 2540
},
{
"epoch": 14.195530726256983,
"grad_norm": 0.38538809453258466,
"learning_rate": 3.826460129420511e-08,
"logits/chosen": -3.4766619205474854,
"logits/rejected": -3.399369478225708,
"logps/chosen": -22.428661346435547,
"logps/rejected": -32.86236572265625,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9801149368286133,
"rewards/margins": 3.9801149368286133,
"rewards/rejected": 0.0,
"step": 2541
},
{
"epoch": 14.201117318435754,
"grad_norm": 0.35870652344020576,
"learning_rate": 3.803108128247512e-08,
"logits/chosen": -3.4228286743164062,
"logits/rejected": -3.4164814949035645,
"logps/chosen": -1.1249449253082275,
"logps/rejected": -62.07542419433594,
"loss": 0.1017,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.915952682495117,
"rewards/margins": 2.915952682495117,
"rewards/rejected": 0.0,
"step": 2542
},
{
"epoch": 14.206703910614525,
"grad_norm": 0.38860705277975816,
"learning_rate": 3.77982478388304e-08,
"logits/chosen": -3.1100687980651855,
"logits/rejected": -3.129997491836548,
"logps/chosen": -8.455366134643555,
"logps/rejected": -24.278011322021484,
"loss": 0.1109,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.617151975631714,
"rewards/margins": 2.617151975631714,
"rewards/rejected": 0.0,
"step": 2543
},
{
"epoch": 14.212290502793296,
"grad_norm": 0.4470156301698857,
"learning_rate": 3.7566101309303156e-08,
"logits/chosen": -3.403496265411377,
"logits/rejected": -3.533383369445801,
"logps/chosen": -12.462495803833008,
"logps/rejected": -74.77552795410156,
"loss": 0.1232,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8615341186523438,
"rewards/margins": 2.8615341186523438,
"rewards/rejected": 0.0,
"step": 2544
},
{
"epoch": 14.217877094972067,
"grad_norm": 0.42620057037924514,
"learning_rate": 3.7334642038904185e-08,
"logits/chosen": -3.377558469772339,
"logits/rejected": -3.22792649269104,
"logps/chosen": -0.30472633242607117,
"logps/rejected": -60.303619384765625,
"loss": 0.0904,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7432624101638794,
"rewards/margins": 1.7432624101638794,
"rewards/rejected": 0.0,
"step": 2545
},
{
"epoch": 14.223463687150838,
"grad_norm": 0.35496394065499254,
"learning_rate": 3.7103870371623434e-08,
"logits/chosen": -3.5415914058685303,
"logits/rejected": -3.546121597290039,
"logps/chosen": -0.2317655086517334,
"logps/rejected": -44.095947265625,
"loss": 0.1163,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.514262080192566,
"rewards/margins": 1.514262080192566,
"rewards/rejected": 0.0,
"step": 2546
},
{
"epoch": 14.22905027932961,
"grad_norm": 0.5830025937584982,
"learning_rate": 3.687378665042862e-08,
"logits/chosen": -3.3116681575775146,
"logits/rejected": -3.3289129734039307,
"logps/chosen": -0.45762497186660767,
"logps/rejected": -55.56715393066406,
"loss": 0.0994,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9170093536376953,
"rewards/margins": 1.9170093536376953,
"rewards/rejected": 0.0,
"step": 2547
},
{
"epoch": 14.23463687150838,
"grad_norm": 0.36813760464536227,
"learning_rate": 3.664439121726509e-08,
"logits/chosen": -3.4270031452178955,
"logits/rejected": -3.4405107498168945,
"logps/chosen": -3.3159165382385254,
"logps/rejected": -52.413536071777344,
"loss": 0.0867,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.170335292816162,
"rewards/margins": 3.170335292816162,
"rewards/rejected": 0.0,
"step": 2548
},
{
"epoch": 14.240223463687151,
"grad_norm": 0.40096396614975366,
"learning_rate": 3.641568441305532e-08,
"logits/chosen": -3.6501471996307373,
"logits/rejected": -3.588867425918579,
"logps/chosen": -31.158584594726562,
"logps/rejected": -20.77127456665039,
"loss": 0.0702,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6337647438049316,
"rewards/margins": 3.6337647438049316,
"rewards/rejected": 0.0,
"step": 2549
},
{
"epoch": 14.245810055865922,
"grad_norm": 0.3740658947458769,
"learning_rate": 3.618766657769834e-08,
"logits/chosen": -3.2856688499450684,
"logits/rejected": -3.244035005569458,
"logps/chosen": -1.9044477939605713,
"logps/rejected": -32.862998962402344,
"loss": 0.1112,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0152015686035156,
"rewards/margins": 3.0152015686035156,
"rewards/rejected": 0.0,
"step": 2550
},
{
"epoch": 14.251396648044693,
"grad_norm": 0.4314091033780081,
"learning_rate": 3.5960338050069316e-08,
"logits/chosen": -3.368588447570801,
"logits/rejected": -3.219712257385254,
"logps/chosen": -3.6364493370056152,
"logps/rejected": -63.97681427001953,
"loss": 0.1024,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1884875297546387,
"rewards/margins": 3.1884875297546387,
"rewards/rejected": 0.0,
"step": 2551
},
{
"epoch": 14.256983240223464,
"grad_norm": 0.37774614647382987,
"learning_rate": 3.5733699168018826e-08,
"logits/chosen": -3.426694393157959,
"logits/rejected": -3.3392457962036133,
"logps/chosen": -4.514124393463135,
"logps/rejected": -33.23273849487305,
"loss": 0.076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8817319869995117,
"rewards/margins": 3.8817319869995117,
"rewards/rejected": 0.0,
"step": 2552
},
{
"epoch": 14.262569832402235,
"grad_norm": 0.4382203653417537,
"learning_rate": 3.550775026837283e-08,
"logits/chosen": -3.526038408279419,
"logits/rejected": -3.4986283779144287,
"logps/chosen": -3.162095546722412,
"logps/rejected": -37.188873291015625,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.527355670928955,
"rewards/margins": 2.527355670928955,
"rewards/rejected": 0.0,
"step": 2553
},
{
"epoch": 14.268156424581006,
"grad_norm": 0.3481932633927292,
"learning_rate": 3.528249168693132e-08,
"logits/chosen": -3.265181064605713,
"logits/rejected": -3.219682455062866,
"logps/chosen": -0.8243135809898376,
"logps/rejected": -44.84501266479492,
"loss": 0.1148,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9604628086090088,
"rewards/margins": 1.9604628086090088,
"rewards/rejected": 0.0,
"step": 2554
},
{
"epoch": 14.273743016759777,
"grad_norm": 0.35303876429225944,
"learning_rate": 3.505792375846889e-08,
"logits/chosen": -3.349062204360962,
"logits/rejected": -3.506606101989746,
"logps/chosen": -4.68916130065918,
"logps/rejected": -51.4373893737793,
"loss": 0.1179,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8279829025268555,
"rewards/margins": 2.8279829025268555,
"rewards/rejected": 0.0,
"step": 2555
},
{
"epoch": 14.279329608938548,
"grad_norm": 0.40274354528791934,
"learning_rate": 3.483404681673341e-08,
"logits/chosen": -3.368523597717285,
"logits/rejected": -3.2170631885528564,
"logps/chosen": -5.536831378936768,
"logps/rejected": -51.38372802734375,
"loss": 0.0755,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8051626682281494,
"rewards/margins": 3.8051626682281494,
"rewards/rejected": 0.0,
"step": 2556
},
{
"epoch": 14.28491620111732,
"grad_norm": 0.38131181541146014,
"learning_rate": 3.461086119444589e-08,
"logits/chosen": -3.321594715118408,
"logits/rejected": -3.440338134765625,
"logps/chosen": -12.703758239746094,
"logps/rejected": -84.10555267333984,
"loss": 0.0924,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1931815147399902,
"rewards/margins": 3.1931815147399902,
"rewards/rejected": 0.0,
"step": 2557
},
{
"epoch": 14.29050279329609,
"grad_norm": 0.34319082466630746,
"learning_rate": 3.4388367223299906e-08,
"logits/chosen": -3.3441390991210938,
"logits/rejected": -3.4788408279418945,
"logps/chosen": -2.0679593086242676,
"logps/rejected": -112.08697509765625,
"loss": 0.0886,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.617946147918701,
"rewards/margins": 2.617946147918701,
"rewards/rejected": 0.0,
"step": 2558
},
{
"epoch": 14.296089385474861,
"grad_norm": 0.5708486696705353,
"learning_rate": 3.416656523396111e-08,
"logits/chosen": -3.172494649887085,
"logits/rejected": -3.3157424926757812,
"logps/chosen": -0.9829146862030029,
"logps/rejected": -70.50743103027344,
"loss": 0.1169,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.420619249343872,
"rewards/margins": 2.420619249343872,
"rewards/rejected": 0.0,
"step": 2559
},
{
"epoch": 14.30167597765363,
"grad_norm": 0.3337373841054444,
"learning_rate": 3.394545555606681e-08,
"logits/chosen": -3.390951633453369,
"logits/rejected": -3.4657745361328125,
"logps/chosen": -1.568161129951477,
"logps/rejected": -43.94268035888672,
"loss": 0.1197,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1630754470825195,
"rewards/margins": 3.1630754470825195,
"rewards/rejected": 0.0,
"step": 2560
},
{
"epoch": 14.30167597765363,
"eval_logits/chosen": -3.255465030670166,
"eval_logits/rejected": -3.378861665725708,
"eval_logps/chosen": -30.90744400024414,
"eval_logps/rejected": -56.270057678222656,
"eval_loss": 1.056686520576477,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.08374010026454926,
"eval_rewards/margins": -0.08374010026454926,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7032,
"eval_samples_per_second": 9.479,
"eval_steps_per_second": 0.306,
"step": 2560
},
{
"epoch": 14.307262569832401,
"grad_norm": 0.4062184874937163,
"learning_rate": 3.3725038518225224e-08,
"logits/chosen": -2.9874773025512695,
"logits/rejected": -3.006955623626709,
"logps/chosen": -14.823481559753418,
"logps/rejected": -57.64149856567383,
"loss": 0.1158,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2055070400238037,
"rewards/margins": 3.2055070400238037,
"rewards/rejected": 0.0,
"step": 2561
},
{
"epoch": 14.312849162011172,
"grad_norm": 0.35126790298425564,
"learning_rate": 3.3505314448015574e-08,
"logits/chosen": -3.2318477630615234,
"logits/rejected": -3.3444159030914307,
"logps/chosen": -0.5783787965774536,
"logps/rejected": -36.68168640136719,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0725185871124268,
"rewards/margins": 2.0725185871124268,
"rewards/rejected": 0.0,
"step": 2562
},
{
"epoch": 14.318435754189943,
"grad_norm": 0.5177101421785699,
"learning_rate": 3.328628367198666e-08,
"logits/chosen": -3.232680559158325,
"logits/rejected": -3.465559244155884,
"logps/chosen": -0.8780964612960815,
"logps/rejected": -35.866153717041016,
"loss": 0.1048,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.195763111114502,
"rewards/margins": 2.195763111114502,
"rewards/rejected": 0.0,
"step": 2563
},
{
"epoch": 14.324022346368714,
"grad_norm": 0.4864073422162365,
"learning_rate": 3.306794651565753e-08,
"logits/chosen": -3.5679423809051514,
"logits/rejected": -3.545628786087036,
"logps/chosen": -3.0100579261779785,
"logps/rejected": -58.98371124267578,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.029555559158325,
"rewards/margins": 2.029555559158325,
"rewards/rejected": 0.0,
"step": 2564
},
{
"epoch": 14.329608938547485,
"grad_norm": 0.35815007533728227,
"learning_rate": 3.2850303303515794e-08,
"logits/chosen": -3.2007193565368652,
"logits/rejected": -3.4571046829223633,
"logps/chosen": -2.7985048294067383,
"logps/rejected": -53.030723571777344,
"loss": 0.0917,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0638461112976074,
"rewards/margins": 3.0638461112976074,
"rewards/rejected": 0.0,
"step": 2565
},
{
"epoch": 14.335195530726256,
"grad_norm": 0.4100680157995391,
"learning_rate": 3.263335435901821e-08,
"logits/chosen": -3.204197645187378,
"logits/rejected": -3.0503411293029785,
"logps/chosen": -1.3740668296813965,
"logps/rejected": -47.070106506347656,
"loss": 0.0854,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.610954999923706,
"rewards/margins": 2.610954999923706,
"rewards/rejected": 0.0,
"step": 2566
},
{
"epoch": 14.340782122905027,
"grad_norm": 1.2958371240100712,
"learning_rate": 3.2417100004589526e-08,
"logits/chosen": -3.090535879135132,
"logits/rejected": -3.2727270126342773,
"logps/chosen": -2.032452344894409,
"logps/rejected": -106.64765167236328,
"loss": 0.1032,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0451555252075195,
"rewards/margins": 2.0451555252075195,
"rewards/rejected": 0.0,
"step": 2567
},
{
"epoch": 14.346368715083798,
"grad_norm": 0.5461939516497486,
"learning_rate": 3.220154056162216e-08,
"logits/chosen": -3.3030285835266113,
"logits/rejected": -3.4315459728240967,
"logps/chosen": -3.357105255126953,
"logps/rejected": -42.3886604309082,
"loss": 0.1331,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.027128219604492,
"rewards/margins": 3.027128219604492,
"rewards/rejected": 0.0,
"step": 2568
},
{
"epoch": 14.35195530726257,
"grad_norm": 0.3496657848254997,
"learning_rate": 3.198667635047592e-08,
"logits/chosen": -3.306248664855957,
"logits/rejected": -3.3260040283203125,
"logps/chosen": -4.4225053787231445,
"logps/rejected": -50.34362030029297,
"loss": 0.0817,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.587620973587036,
"rewards/margins": 3.587620973587036,
"rewards/rejected": 0.0,
"step": 2569
},
{
"epoch": 14.35754189944134,
"grad_norm": 0.3618740932211241,
"learning_rate": 3.177250769047718e-08,
"logits/chosen": -3.3666021823883057,
"logits/rejected": -3.229046106338501,
"logps/chosen": -8.65818977355957,
"logps/rejected": -52.667110443115234,
"loss": 0.0673,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7629356384277344,
"rewards/margins": 3.7629356384277344,
"rewards/rejected": 0.0,
"step": 2570
},
{
"epoch": 14.363128491620111,
"grad_norm": 0.39086312368819265,
"learning_rate": 3.155903489991879e-08,
"logits/chosen": -3.4258573055267334,
"logits/rejected": -3.39311146736145,
"logps/chosen": -3.4805006980895996,
"logps/rejected": -29.188730239868164,
"loss": 0.0955,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.843630313873291,
"rewards/margins": 2.843630313873291,
"rewards/rejected": 0.0,
"step": 2571
},
{
"epoch": 14.368715083798882,
"grad_norm": 0.430391431592419,
"learning_rate": 3.134625829605925e-08,
"logits/chosen": -3.3774266242980957,
"logits/rejected": -3.485069513320923,
"logps/chosen": -1.3629817962646484,
"logps/rejected": -63.495811462402344,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.650252103805542,
"rewards/margins": 2.650252103805542,
"rewards/rejected": 0.0,
"step": 2572
},
{
"epoch": 14.374301675977653,
"grad_norm": 0.4148997561832557,
"learning_rate": 3.113417819512265e-08,
"logits/chosen": -3.3535616397857666,
"logits/rejected": -3.6837968826293945,
"logps/chosen": -2.745608329772949,
"logps/rejected": -50.464202880859375,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9353556632995605,
"rewards/margins": 2.9353556632995605,
"rewards/rejected": 0.0,
"step": 2573
},
{
"epoch": 14.379888268156424,
"grad_norm": 0.40928154109029374,
"learning_rate": 3.09227949122976e-08,
"logits/chosen": -3.064005136489868,
"logits/rejected": -3.35892915725708,
"logps/chosen": -0.336436927318573,
"logps/rejected": -98.37713623046875,
"loss": 0.1195,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0511691570281982,
"rewards/margins": 2.0511691570281982,
"rewards/rejected": 0.0,
"step": 2574
},
{
"epoch": 14.385474860335195,
"grad_norm": 0.3885864644716112,
"learning_rate": 3.071210876173741e-08,
"logits/chosen": -3.5664894580841064,
"logits/rejected": -3.6670656204223633,
"logps/chosen": -1.5626771450042725,
"logps/rejected": -50.25803756713867,
"loss": 0.11,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.824280261993408,
"rewards/margins": 2.824280261993408,
"rewards/rejected": 0.0,
"step": 2575
},
{
"epoch": 14.391061452513966,
"grad_norm": 0.39290590716934026,
"learning_rate": 3.0502120056559225e-08,
"logits/chosen": -3.382290840148926,
"logits/rejected": -3.4783260822296143,
"logps/chosen": -3.0286645889282227,
"logps/rejected": -51.1189079284668,
"loss": 0.1018,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.124781847000122,
"rewards/margins": 3.124781847000122,
"rewards/rejected": 0.0,
"step": 2576
},
{
"epoch": 14.396648044692737,
"grad_norm": 0.8561349864475638,
"learning_rate": 3.029282910884362e-08,
"logits/chosen": -3.3393774032592773,
"logits/rejected": -3.3798530101776123,
"logps/chosen": -0.8968712091445923,
"logps/rejected": -67.7798080444336,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1824569702148438,
"rewards/margins": 2.1824569702148438,
"rewards/rejected": 0.0,
"step": 2577
},
{
"epoch": 14.402234636871508,
"grad_norm": 0.38494307401336597,
"learning_rate": 3.008423622963424e-08,
"logits/chosen": -3.4091882705688477,
"logits/rejected": -3.287452459335327,
"logps/chosen": -0.607304036617279,
"logps/rejected": -57.99492645263672,
"loss": 0.0917,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7982110977172852,
"rewards/margins": 1.7982110977172852,
"rewards/rejected": 0.0,
"step": 2578
},
{
"epoch": 14.40782122905028,
"grad_norm": 0.39067061708928996,
"learning_rate": 2.987634172893727e-08,
"logits/chosen": -3.2327842712402344,
"logits/rejected": -3.473045587539673,
"logps/chosen": -1.3123928308486938,
"logps/rejected": -74.77027893066406,
"loss": 0.0793,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2215685844421387,
"rewards/margins": 2.2215685844421387,
"rewards/rejected": 0.0,
"step": 2579
},
{
"epoch": 14.41340782122905,
"grad_norm": 0.35512891358892545,
"learning_rate": 2.9669145915720896e-08,
"logits/chosen": -3.2294089794158936,
"logits/rejected": -3.275222063064575,
"logps/chosen": -0.3199467062950134,
"logps/rejected": -84.38740539550781,
"loss": 0.1024,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7708466053009033,
"rewards/margins": 1.7708466053009033,
"rewards/rejected": 0.0,
"step": 2580
},
{
"epoch": 14.41340782122905,
"eval_logits/chosen": -3.2508492469787598,
"eval_logits/rejected": -3.3756213188171387,
"eval_logps/chosen": -30.67477035522461,
"eval_logps/rejected": -56.07917022705078,
"eval_loss": 1.0568053722381592,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.06047270447015762,
"eval_rewards/margins": -0.06047270447015762,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.693,
"eval_samples_per_second": 9.482,
"eval_steps_per_second": 0.306,
"step": 2580
},
{
"epoch": 14.418994413407821,
"grad_norm": 0.40818006036734,
"learning_rate": 2.9462649097915015e-08,
"logits/chosen": -3.4440417289733887,
"logits/rejected": -3.4083168506622314,
"logps/chosen": -10.291391372680664,
"logps/rejected": -36.64921569824219,
"loss": 0.08,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.854292869567871,
"rewards/margins": 3.854292869567871,
"rewards/rejected": 0.0,
"step": 2581
},
{
"epoch": 14.424581005586592,
"grad_norm": 0.3520335560565263,
"learning_rate": 2.9256851582410756e-08,
"logits/chosen": -3.4099106788635254,
"logits/rejected": -3.474188804626465,
"logps/chosen": -0.756075918674469,
"logps/rejected": -31.355152130126953,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.440232753753662,
"rewards/margins": 2.440232753753662,
"rewards/rejected": 0.0,
"step": 2582
},
{
"epoch": 14.430167597765363,
"grad_norm": 0.4478345951399215,
"learning_rate": 2.9051753675059744e-08,
"logits/chosen": -3.2319183349609375,
"logits/rejected": -3.228318452835083,
"logps/chosen": -0.3813924193382263,
"logps/rejected": -126.91836547851562,
"loss": 0.0816,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2227683067321777,
"rewards/margins": 2.2227683067321777,
"rewards/rejected": 0.0,
"step": 2583
},
{
"epoch": 14.435754189944134,
"grad_norm": 0.4553624792675634,
"learning_rate": 2.884735568067409e-08,
"logits/chosen": -3.5689172744750977,
"logits/rejected": -3.6130666732788086,
"logps/chosen": -1.2657164335250854,
"logps/rejected": -44.229610443115234,
"loss": 0.0886,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.909900188446045,
"rewards/margins": 2.909900188446045,
"rewards/rejected": 0.0,
"step": 2584
},
{
"epoch": 14.441340782122905,
"grad_norm": 0.4124485290524065,
"learning_rate": 2.864365790302564e-08,
"logits/chosen": -3.4473507404327393,
"logits/rejected": -3.515202045440674,
"logps/chosen": -3.42461895942688,
"logps/rejected": -63.259376525878906,
"loss": 0.0743,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9966366291046143,
"rewards/margins": 2.9966366291046143,
"rewards/rejected": 0.0,
"step": 2585
},
{
"epoch": 14.446927374301676,
"grad_norm": 0.38026992578702373,
"learning_rate": 2.844066064484546e-08,
"logits/chosen": -3.2951223850250244,
"logits/rejected": -3.399473190307617,
"logps/chosen": -0.6012113094329834,
"logps/rejected": -53.583499908447266,
"loss": 0.1027,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.072929859161377,
"rewards/margins": 2.072929859161377,
"rewards/rejected": 0.0,
"step": 2586
},
{
"epoch": 14.452513966480447,
"grad_norm": 0.40418546186396853,
"learning_rate": 2.8238364207823717e-08,
"logits/chosen": -3.137958288192749,
"logits/rejected": -3.0798516273498535,
"logps/chosen": -20.386611938476562,
"logps/rejected": -57.91345977783203,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8627772331237793,
"rewards/margins": 2.8627772331237793,
"rewards/rejected": 0.0,
"step": 2587
},
{
"epoch": 14.458100558659218,
"grad_norm": 0.4793948181881813,
"learning_rate": 2.803676889260892e-08,
"logits/chosen": -3.414140462875366,
"logits/rejected": -3.2515978813171387,
"logps/chosen": -5.7660675048828125,
"logps/rejected": -34.76047134399414,
"loss": 0.1344,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0209760665893555,
"rewards/margins": 3.0209760665893555,
"rewards/rejected": 0.0,
"step": 2588
},
{
"epoch": 14.46368715083799,
"grad_norm": 0.36959361743706953,
"learning_rate": 2.7835874998807575e-08,
"logits/chosen": -3.1530601978302,
"logits/rejected": -3.2044665813446045,
"logps/chosen": -6.9558563232421875,
"logps/rejected": -76.71287536621094,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.355774164199829,
"rewards/margins": 3.355774164199829,
"rewards/rejected": 0.0,
"step": 2589
},
{
"epoch": 14.46927374301676,
"grad_norm": 0.414943361556342,
"learning_rate": 2.763568282498374e-08,
"logits/chosen": -2.955747365951538,
"logits/rejected": -3.082019567489624,
"logps/chosen": -11.286369323730469,
"logps/rejected": -44.17339324951172,
"loss": 0.0869,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6280245780944824,
"rewards/margins": 3.6280245780944824,
"rewards/rejected": 0.0,
"step": 2590
},
{
"epoch": 14.474860335195531,
"grad_norm": 0.45495298373564647,
"learning_rate": 2.7436192668658876e-08,
"logits/chosen": -3.0897738933563232,
"logits/rejected": -3.135404109954834,
"logps/chosen": -3.8595781326293945,
"logps/rejected": -71.08548736572266,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9746313095092773,
"rewards/margins": 2.9746313095092773,
"rewards/rejected": 0.0,
"step": 2591
},
{
"epoch": 14.480446927374302,
"grad_norm": 0.38281341793748186,
"learning_rate": 2.7237404826310595e-08,
"logits/chosen": -3.3906188011169434,
"logits/rejected": -3.6005892753601074,
"logps/chosen": -4.082036018371582,
"logps/rejected": -43.455020904541016,
"loss": 0.078,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.56838321685791,
"rewards/margins": 3.56838321685791,
"rewards/rejected": 0.0,
"step": 2592
},
{
"epoch": 14.486033519553073,
"grad_norm": 0.37134089965678546,
"learning_rate": 2.7039319593373245e-08,
"logits/chosen": -3.496018171310425,
"logits/rejected": -3.381351947784424,
"logps/chosen": -2.4031014442443848,
"logps/rejected": -39.70995330810547,
"loss": 0.1094,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.413717746734619,
"rewards/margins": 3.413717746734619,
"rewards/rejected": 0.0,
"step": 2593
},
{
"epoch": 14.491620111731844,
"grad_norm": 0.46900536140110083,
"learning_rate": 2.684193726423656e-08,
"logits/chosen": -3.3678228855133057,
"logits/rejected": -3.2361197471618652,
"logps/chosen": -3.444424867630005,
"logps/rejected": -67.34076690673828,
"loss": 0.0993,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.425002098083496,
"rewards/margins": 2.425002098083496,
"rewards/rejected": 0.0,
"step": 2594
},
{
"epoch": 14.497206703910614,
"grad_norm": 0.36701090376847895,
"learning_rate": 2.6645258132245994e-08,
"logits/chosen": -3.5504260063171387,
"logits/rejected": -3.4358696937561035,
"logps/chosen": -1.103353500366211,
"logps/rejected": -43.75018310546875,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.983163833618164,
"rewards/margins": 2.983163833618164,
"rewards/rejected": 0.0,
"step": 2595
},
{
"epoch": 14.502793296089386,
"grad_norm": 0.39706166504429785,
"learning_rate": 2.644928248970174e-08,
"logits/chosen": -3.490511178970337,
"logits/rejected": -3.5199904441833496,
"logps/chosen": -0.4865378439426422,
"logps/rejected": -50.04668426513672,
"loss": 0.1024,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2721471786499023,
"rewards/margins": 2.2721471786499023,
"rewards/rejected": 0.0,
"step": 2596
},
{
"epoch": 14.508379888268156,
"grad_norm": 0.5558956921166254,
"learning_rate": 2.6254010627858425e-08,
"logits/chosen": -3.393669366836548,
"logits/rejected": -3.3033180236816406,
"logps/chosen": -0.1819501668214798,
"logps/rejected": -62.02227783203125,
"loss": 0.125,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.5980136394500732,
"rewards/margins": 1.5980136394500732,
"rewards/rejected": 0.0,
"step": 2597
},
{
"epoch": 14.513966480446927,
"grad_norm": 0.48786314777861517,
"learning_rate": 2.6059442836924907e-08,
"logits/chosen": -3.3012449741363525,
"logits/rejected": -3.271615505218506,
"logps/chosen": -3.1856186389923096,
"logps/rejected": -50.03583908081055,
"loss": 0.0949,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0265607833862305,
"rewards/margins": 3.0265607833862305,
"rewards/rejected": 0.0,
"step": 2598
},
{
"epoch": 14.519553072625698,
"grad_norm": 0.32651791115638257,
"learning_rate": 2.586557940606349e-08,
"logits/chosen": -3.349930763244629,
"logits/rejected": -3.3481690883636475,
"logps/chosen": -0.39179161190986633,
"logps/rejected": -27.280929565429688,
"loss": 0.094,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.166875123977661,
"rewards/margins": 2.166875123977661,
"rewards/rejected": 0.0,
"step": 2599
},
{
"epoch": 14.525139664804469,
"grad_norm": 0.823962753447027,
"learning_rate": 2.5672420623389867e-08,
"logits/chosen": -3.3070006370544434,
"logits/rejected": -3.481149435043335,
"logps/chosen": -1.2035163640975952,
"logps/rejected": -70.06402587890625,
"loss": 0.0841,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6678123474121094,
"rewards/margins": 2.6678123474121094,
"rewards/rejected": 0.0,
"step": 2600
},
{
"epoch": 14.525139664804469,
"eval_logits/chosen": -3.251675844192505,
"eval_logits/rejected": -3.3768248558044434,
"eval_logps/chosen": -30.821081161499023,
"eval_logps/rejected": -56.12659454345703,
"eval_loss": 1.052950143814087,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.07510419934988022,
"eval_rewards/margins": -0.07510419934988022,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6999,
"eval_samples_per_second": 9.48,
"eval_steps_per_second": 0.306,
"step": 2600
},
{
"epoch": 14.53072625698324,
"grad_norm": 0.35034672718428556,
"learning_rate": 2.5479966775972196e-08,
"logits/chosen": -3.4400594234466553,
"logits/rejected": -3.4662489891052246,
"logps/chosen": -0.2684183418750763,
"logps/rejected": -53.87624740600586,
"loss": 0.1112,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6512014865875244,
"rewards/margins": 1.6512014865875244,
"rewards/rejected": 0.0,
"step": 2601
},
{
"epoch": 14.53631284916201,
"grad_norm": 0.33204673460745693,
"learning_rate": 2.5288218149831398e-08,
"logits/chosen": -2.746803045272827,
"logits/rejected": -2.8619046211242676,
"logps/chosen": -38.81970977783203,
"logps/rejected": -39.522727966308594,
"loss": 0.0621,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.217256546020508,
"rewards/margins": 4.217256546020508,
"rewards/rejected": 0.0,
"step": 2602
},
{
"epoch": 14.541899441340782,
"grad_norm": 0.35216370747195497,
"learning_rate": 2.509717502993991e-08,
"logits/chosen": -3.466219425201416,
"logits/rejected": -3.3800175189971924,
"logps/chosen": -4.04020881652832,
"logps/rejected": -34.599159240722656,
"loss": 0.0841,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.462942600250244,
"rewards/margins": 2.462942600250244,
"rewards/rejected": 0.0,
"step": 2603
},
{
"epoch": 14.547486033519553,
"grad_norm": 0.37850456251304454,
"learning_rate": 2.4906837700221895e-08,
"logits/chosen": -3.2788219451904297,
"logits/rejected": -3.29059100151062,
"logps/chosen": -19.956703186035156,
"logps/rejected": -68.52668762207031,
"loss": 0.11,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.91402530670166,
"rewards/margins": 3.91402530670166,
"rewards/rejected": 0.0,
"step": 2604
},
{
"epoch": 14.553072625698324,
"grad_norm": 0.3620836293269757,
"learning_rate": 2.4717206443552573e-08,
"logits/chosen": -3.2138986587524414,
"logits/rejected": -3.1645960807800293,
"logps/chosen": -7.612255573272705,
"logps/rejected": -41.527992248535156,
"loss": 0.0928,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1727817058563232,
"rewards/margins": 3.1727817058563232,
"rewards/rejected": 0.0,
"step": 2605
},
{
"epoch": 14.558659217877095,
"grad_norm": 0.3769520825069747,
"learning_rate": 2.4528281541757677e-08,
"logits/chosen": -3.6428723335266113,
"logits/rejected": -3.611280679702759,
"logps/chosen": -1.5903761386871338,
"logps/rejected": -39.019962310791016,
"loss": 0.0825,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8951711654663086,
"rewards/margins": 2.8951711654663086,
"rewards/rejected": 0.0,
"step": 2606
},
{
"epoch": 14.564245810055866,
"grad_norm": 0.5338666848084862,
"learning_rate": 2.4340063275613333e-08,
"logits/chosen": -3.184285879135132,
"logits/rejected": -3.4263367652893066,
"logps/chosen": -0.8953671455383301,
"logps/rejected": -40.780067443847656,
"loss": 0.1237,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0662784576416016,
"rewards/margins": 2.0662784576416016,
"rewards/rejected": 0.0,
"step": 2607
},
{
"epoch": 14.569832402234637,
"grad_norm": 0.44377925234671706,
"learning_rate": 2.415255192484533e-08,
"logits/chosen": -3.537266254425049,
"logits/rejected": -3.3981332778930664,
"logps/chosen": -13.35228443145752,
"logps/rejected": -40.480587005615234,
"loss": 0.1022,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.2639312744140625,
"rewards/margins": 4.2639312744140625,
"rewards/rejected": 0.0,
"step": 2608
},
{
"epoch": 14.575418994413408,
"grad_norm": 0.3736243148903746,
"learning_rate": 2.396574776812893e-08,
"logits/chosen": -3.4469752311706543,
"logits/rejected": -3.5112974643707275,
"logps/chosen": -1.0641577243804932,
"logps/rejected": -59.10731887817383,
"loss": 0.0937,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.733170986175537,
"rewards/margins": 2.733170986175537,
"rewards/rejected": 0.0,
"step": 2609
},
{
"epoch": 14.581005586592179,
"grad_norm": 0.36598148041686873,
"learning_rate": 2.377965108308838e-08,
"logits/chosen": -3.415602684020996,
"logits/rejected": -3.525176525115967,
"logps/chosen": -0.1782577633857727,
"logps/rejected": -56.63910675048828,
"loss": 0.122,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8573319911956787,
"rewards/margins": 1.8573319911956787,
"rewards/rejected": 0.0,
"step": 2610
},
{
"epoch": 14.58659217877095,
"grad_norm": 0.37965183369100297,
"learning_rate": 2.3594262146296672e-08,
"logits/chosen": -3.457428216934204,
"logits/rejected": -3.622685670852661,
"logps/chosen": -0.9535687565803528,
"logps/rejected": -60.888526916503906,
"loss": 0.107,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5787744522094727,
"rewards/margins": 2.5787744522094727,
"rewards/rejected": 0.0,
"step": 2611
},
{
"epoch": 14.59217877094972,
"grad_norm": 0.3711794585903184,
"learning_rate": 2.3409581233274466e-08,
"logits/chosen": -3.2968878746032715,
"logits/rejected": -3.3927321434020996,
"logps/chosen": -0.4315730631351471,
"logps/rejected": -76.43840026855469,
"loss": 0.1068,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3661458492279053,
"rewards/margins": 2.3661458492279053,
"rewards/rejected": 0.0,
"step": 2612
},
{
"epoch": 14.597765363128492,
"grad_norm": 0.418805128932327,
"learning_rate": 2.3225608618490766e-08,
"logits/chosen": -3.59462308883667,
"logits/rejected": -3.6426193714141846,
"logps/chosen": -6.757093906402588,
"logps/rejected": -29.405033111572266,
"loss": 0.0849,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.983060598373413,
"rewards/margins": 2.983060598373413,
"rewards/rejected": 0.0,
"step": 2613
},
{
"epoch": 14.603351955307263,
"grad_norm": 0.34918248560991544,
"learning_rate": 2.3042344575361528e-08,
"logits/chosen": -3.340179920196533,
"logits/rejected": -3.330141305923462,
"logps/chosen": -1.3053386211395264,
"logps/rejected": -34.125205993652344,
"loss": 0.092,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5008788108825684,
"rewards/margins": 2.5008788108825684,
"rewards/rejected": 0.0,
"step": 2614
},
{
"epoch": 14.608938547486034,
"grad_norm": 0.4830880413118678,
"learning_rate": 2.2859789376249716e-08,
"logits/chosen": -3.3018860816955566,
"logits/rejected": -3.420578718185425,
"logps/chosen": -1.1772595643997192,
"logps/rejected": -58.97906494140625,
"loss": 0.0822,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2907752990722656,
"rewards/margins": 2.2907752990722656,
"rewards/rejected": 0.0,
"step": 2615
},
{
"epoch": 14.614525139664805,
"grad_norm": 0.5430887252058046,
"learning_rate": 2.2677943292464973e-08,
"logits/chosen": -3.4364094734191895,
"logits/rejected": -3.503571033477783,
"logps/chosen": -0.9159830212593079,
"logps/rejected": -69.68579864501953,
"loss": 0.0908,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6822304725646973,
"rewards/margins": 2.6822304725646973,
"rewards/rejected": 0.0,
"step": 2616
},
{
"epoch": 14.620111731843576,
"grad_norm": 0.48130237763903183,
"learning_rate": 2.249680659426284e-08,
"logits/chosen": -3.2569284439086914,
"logits/rejected": -3.2110025882720947,
"logps/chosen": -3.1655027866363525,
"logps/rejected": -65.03008270263672,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6212363243103027,
"rewards/margins": 2.6212363243103027,
"rewards/rejected": 0.0,
"step": 2617
},
{
"epoch": 14.625698324022347,
"grad_norm": 0.4390675915180137,
"learning_rate": 2.231637955084481e-08,
"logits/chosen": -3.1946511268615723,
"logits/rejected": -3.288429021835327,
"logps/chosen": -18.75676727294922,
"logps/rejected": -54.70408630371094,
"loss": 0.1064,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4994304180145264,
"rewards/margins": 3.4994304180145264,
"rewards/rejected": 0.0,
"step": 2618
},
{
"epoch": 14.631284916201118,
"grad_norm": 1.2330578550313098,
"learning_rate": 2.21366624303575e-08,
"logits/chosen": -3.0974862575531006,
"logits/rejected": -3.4215402603149414,
"logps/chosen": -2.1984615325927734,
"logps/rejected": -55.712890625,
"loss": 0.1385,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8407673835754395,
"rewards/margins": 1.8407673835754395,
"rewards/rejected": 0.0,
"step": 2619
},
{
"epoch": 14.636871508379889,
"grad_norm": 0.4838777218585874,
"learning_rate": 2.195765549989276e-08,
"logits/chosen": -3.335300922393799,
"logits/rejected": -3.259881019592285,
"logps/chosen": -1.1644904613494873,
"logps/rejected": -39.106361389160156,
"loss": 0.1166,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7018516063690186,
"rewards/margins": 2.7018516063690186,
"rewards/rejected": 0.0,
"step": 2620
},
{
"epoch": 14.636871508379889,
"eval_logits/chosen": -3.2521300315856934,
"eval_logits/rejected": -3.377594470977783,
"eval_logps/chosen": -30.75811767578125,
"eval_logps/rejected": -56.20076370239258,
"eval_loss": 1.052037239074707,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.06880731880664825,
"eval_rewards/margins": -0.06880731880664825,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7264,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 2620
},
{
"epoch": 14.64245810055866,
"grad_norm": 0.3687159054789329,
"learning_rate": 2.1779359025486498e-08,
"logits/chosen": -3.552938222885132,
"logits/rejected": -3.731419563293457,
"logps/chosen": -3.871910572052002,
"logps/rejected": -38.07981491088867,
"loss": 0.1068,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2329859733581543,
"rewards/margins": 3.2329859733581543,
"rewards/rejected": 0.0,
"step": 2621
},
{
"epoch": 14.64804469273743,
"grad_norm": 0.3875358266687342,
"learning_rate": 2.1601773272119318e-08,
"logits/chosen": -3.381281614303589,
"logits/rejected": -3.2493088245391846,
"logps/chosen": -0.6715680360794067,
"logps/rejected": -55.6790771484375,
"loss": 0.0999,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4488768577575684,
"rewards/margins": 2.4488768577575684,
"rewards/rejected": 0.0,
"step": 2622
},
{
"epoch": 14.653631284916202,
"grad_norm": 0.4122394220220475,
"learning_rate": 2.1424898503714985e-08,
"logits/chosen": -3.5211215019226074,
"logits/rejected": -3.5921006202697754,
"logps/chosen": -2.2461485862731934,
"logps/rejected": -42.21638488769531,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5515592098236084,
"rewards/margins": 2.5515592098236084,
"rewards/rejected": 0.0,
"step": 2623
},
{
"epoch": 14.659217877094973,
"grad_norm": 0.3859643799676988,
"learning_rate": 2.124873498314106e-08,
"logits/chosen": -3.062135696411133,
"logits/rejected": -3.137162923812866,
"logps/chosen": -3.3561148643493652,
"logps/rejected": -31.430452346801758,
"loss": 0.0919,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6686174869537354,
"rewards/margins": 2.6686174869537354,
"rewards/rejected": 0.0,
"step": 2624
},
{
"epoch": 14.664804469273744,
"grad_norm": 0.4212571230933188,
"learning_rate": 2.1073282972207896e-08,
"logits/chosen": -3.439175605773926,
"logits/rejected": -3.3376693725585938,
"logps/chosen": -5.655553817749023,
"logps/rejected": -68.16014862060547,
"loss": 0.1252,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.194246768951416,
"rewards/margins": 3.194246768951416,
"rewards/rejected": 0.0,
"step": 2625
},
{
"epoch": 14.670391061452515,
"grad_norm": 0.3567387071922027,
"learning_rate": 2.089854273166841e-08,
"logits/chosen": -3.439663887023926,
"logits/rejected": -3.444983959197998,
"logps/chosen": -0.3894228935241699,
"logps/rejected": -29.912153244018555,
"loss": 0.1173,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1726903915405273,
"rewards/margins": 2.1726903915405273,
"rewards/rejected": 0.0,
"step": 2626
},
{
"epoch": 14.675977653631286,
"grad_norm": 0.3520192576168209,
"learning_rate": 2.0724514521217763e-08,
"logits/chosen": -3.247497081756592,
"logits/rejected": -3.4018421173095703,
"logps/chosen": -0.22753706574440002,
"logps/rejected": -39.27146530151367,
"loss": 0.1053,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6547861099243164,
"rewards/margins": 1.6547861099243164,
"rewards/rejected": 0.0,
"step": 2627
},
{
"epoch": 14.681564245810057,
"grad_norm": 0.3970515001799419,
"learning_rate": 2.0551198599492836e-08,
"logits/chosen": -3.331470489501953,
"logits/rejected": -3.4937338829040527,
"logps/chosen": -7.155062675476074,
"logps/rejected": -32.809383392333984,
"loss": 0.0912,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.697918176651001,
"rewards/margins": 3.697918176651001,
"rewards/rejected": 0.0,
"step": 2628
},
{
"epoch": 14.687150837988828,
"grad_norm": 0.45355096438905357,
"learning_rate": 2.0378595224071925e-08,
"logits/chosen": -3.2169480323791504,
"logits/rejected": -3.1807217597961426,
"logps/chosen": -7.081135272979736,
"logps/rejected": -33.9853630065918,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.2905778884887695,
"rewards/margins": 4.2905778884887695,
"rewards/rejected": 0.0,
"step": 2629
},
{
"epoch": 14.692737430167599,
"grad_norm": 0.350916390676908,
"learning_rate": 2.0206704651474336e-08,
"logits/chosen": -3.228888988494873,
"logits/rejected": -3.2587034702301025,
"logps/chosen": -1.4819749593734741,
"logps/rejected": -37.1948127746582,
"loss": 0.0992,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5577054023742676,
"rewards/margins": 2.5577054023742676,
"rewards/rejected": 0.0,
"step": 2630
},
{
"epoch": 14.69832402234637,
"grad_norm": 0.4109007574926235,
"learning_rate": 2.003552713716028e-08,
"logits/chosen": -3.43320631980896,
"logits/rejected": -3.302509307861328,
"logps/chosen": -3.303668975830078,
"logps/rejected": -30.332958221435547,
"loss": 0.0821,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.254763126373291,
"rewards/margins": 3.254763126373291,
"rewards/rejected": 0.0,
"step": 2631
},
{
"epoch": 14.703910614525139,
"grad_norm": 1.0999709243811522,
"learning_rate": 1.9865062935529808e-08,
"logits/chosen": -3.463435173034668,
"logits/rejected": -3.440654754638672,
"logps/chosen": -4.47990608215332,
"logps/rejected": -64.33112335205078,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.208681106567383,
"rewards/margins": 3.208681106567383,
"rewards/rejected": 0.0,
"step": 2632
},
{
"epoch": 14.70949720670391,
"grad_norm": 0.481934449821989,
"learning_rate": 1.9695312299923162e-08,
"logits/chosen": -3.424265146255493,
"logits/rejected": -3.4145216941833496,
"logps/chosen": -0.8707083463668823,
"logps/rejected": -77.55708312988281,
"loss": 0.1271,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0164778232574463,
"rewards/margins": 2.0164778232574463,
"rewards/rejected": 0.0,
"step": 2633
},
{
"epoch": 14.71508379888268,
"grad_norm": 0.3750543951241516,
"learning_rate": 1.952627548262009e-08,
"logits/chosen": -3.446359395980835,
"logits/rejected": -3.3021953105926514,
"logps/chosen": -4.184319972991943,
"logps/rejected": -54.65296936035156,
"loss": 0.0745,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.354356527328491,
"rewards/margins": 3.354356527328491,
"rewards/rejected": 0.0,
"step": 2634
},
{
"epoch": 14.720670391061452,
"grad_norm": 0.44207718376243293,
"learning_rate": 1.9357952734839312e-08,
"logits/chosen": -2.96549654006958,
"logits/rejected": -2.860167980194092,
"logps/chosen": -32.893165588378906,
"logps/rejected": -67.67338562011719,
"loss": 0.0976,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.400967597961426,
"rewards/margins": 4.400967597961426,
"rewards/rejected": 0.0,
"step": 2635
},
{
"epoch": 14.726256983240223,
"grad_norm": 0.38039342394313047,
"learning_rate": 1.919034430673844e-08,
"logits/chosen": -3.536604404449463,
"logits/rejected": -3.4818167686462402,
"logps/chosen": -1.9360244274139404,
"logps/rejected": -59.25251388549805,
"loss": 0.09,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3960742950439453,
"rewards/margins": 2.3960742950439453,
"rewards/rejected": 0.0,
"step": 2636
},
{
"epoch": 14.731843575418994,
"grad_norm": 0.3490895639330724,
"learning_rate": 1.9023450447413436e-08,
"logits/chosen": -3.300827980041504,
"logits/rejected": -3.4647958278656006,
"logps/chosen": -0.49746406078338623,
"logps/rejected": -57.20527648925781,
"loss": 0.0689,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0046162605285645,
"rewards/margins": 2.0046162605285645,
"rewards/rejected": 0.0,
"step": 2637
},
{
"epoch": 14.737430167597765,
"grad_norm": 0.40550564877263295,
"learning_rate": 1.885727140489829e-08,
"logits/chosen": -3.253418445587158,
"logits/rejected": -3.5220563411712646,
"logps/chosen": -0.15144218504428864,
"logps/rejected": -111.05549621582031,
"loss": 0.1321,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.3530927896499634,
"rewards/margins": 1.3530927896499634,
"rewards/rejected": 0.0,
"step": 2638
},
{
"epoch": 14.743016759776536,
"grad_norm": 0.35351196643128785,
"learning_rate": 1.8691807426164607e-08,
"logits/chosen": -3.5048904418945312,
"logits/rejected": -3.4716110229492188,
"logps/chosen": -13.23552417755127,
"logps/rejected": -54.78971862792969,
"loss": 0.0797,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.246061086654663,
"rewards/margins": 3.246061086654663,
"rewards/rejected": 0.0,
"step": 2639
},
{
"epoch": 14.748603351955307,
"grad_norm": 0.4608382579300712,
"learning_rate": 1.852705875712157e-08,
"logits/chosen": -3.4893388748168945,
"logits/rejected": -3.3630337715148926,
"logps/chosen": -4.5363874435424805,
"logps/rejected": -38.974388122558594,
"loss": 0.0786,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.436383008956909,
"rewards/margins": 3.436383008956909,
"rewards/rejected": 0.0,
"step": 2640
},
{
"epoch": 14.748603351955307,
"eval_logits/chosen": -3.2588112354278564,
"eval_logits/rejected": -3.381526470184326,
"eval_logps/chosen": -30.6961727142334,
"eval_logps/rejected": -56.20176315307617,
"eval_loss": 1.0570828914642334,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.06261324137449265,
"eval_rewards/margins": -0.06261324137449265,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7221,
"eval_samples_per_second": 9.474,
"eval_steps_per_second": 0.306,
"step": 2640
},
{
"epoch": 14.754189944134078,
"grad_norm": 0.36795394142857324,
"learning_rate": 1.836302564261466e-08,
"logits/chosen": -3.246004581451416,
"logits/rejected": -3.4199206829071045,
"logps/chosen": -0.334102988243103,
"logps/rejected": -123.59196472167969,
"loss": 0.1066,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6130549907684326,
"rewards/margins": 1.6130549907684326,
"rewards/rejected": 0.0,
"step": 2641
},
{
"epoch": 14.759776536312849,
"grad_norm": 0.405662194125218,
"learning_rate": 1.8199708326426643e-08,
"logits/chosen": -3.481472969055176,
"logits/rejected": -3.556199073791504,
"logps/chosen": -0.28236332535743713,
"logps/rejected": -64.80596923828125,
"loss": 0.1208,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7299556732177734,
"rewards/margins": 1.7299556732177734,
"rewards/rejected": 0.0,
"step": 2642
},
{
"epoch": 14.76536312849162,
"grad_norm": 0.36918646172301467,
"learning_rate": 1.8037107051275923e-08,
"logits/chosen": -3.6038081645965576,
"logits/rejected": -3.556504487991333,
"logps/chosen": -3.957348585128784,
"logps/rejected": -57.72964096069336,
"loss": 0.1031,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.272129774093628,
"rewards/margins": 3.272129774093628,
"rewards/rejected": 0.0,
"step": 2643
},
{
"epoch": 14.77094972067039,
"grad_norm": 0.42410614403411123,
"learning_rate": 1.787522205881703e-08,
"logits/chosen": -3.5083959102630615,
"logits/rejected": -3.3299269676208496,
"logps/chosen": -4.993856430053711,
"logps/rejected": -41.53742980957031,
"loss": 0.1013,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.24393367767334,
"rewards/margins": 3.24393367767334,
"rewards/rejected": 0.0,
"step": 2644
},
{
"epoch": 14.776536312849162,
"grad_norm": 0.43327424424618927,
"learning_rate": 1.77140535896399e-08,
"logits/chosen": -3.5569281578063965,
"logits/rejected": -3.5528361797332764,
"logps/chosen": -3.9627926349639893,
"logps/rejected": -30.01145362854004,
"loss": 0.1239,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1907334327697754,
"rewards/margins": 3.1907334327697754,
"rewards/rejected": 0.0,
"step": 2645
},
{
"epoch": 14.782122905027933,
"grad_norm": 0.3958390691017802,
"learning_rate": 1.7553601883269597e-08,
"logits/chosen": -3.432112455368042,
"logits/rejected": -3.261931896209717,
"logps/chosen": -3.6665518283843994,
"logps/rejected": -92.46833801269531,
"loss": 0.1059,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.111374855041504,
"rewards/margins": 3.111374855041504,
"rewards/rejected": 0.0,
"step": 2646
},
{
"epoch": 14.787709497206704,
"grad_norm": 0.38290720230321823,
"learning_rate": 1.739386717816582e-08,
"logits/chosen": -3.3596928119659424,
"logits/rejected": -3.4711050987243652,
"logps/chosen": -0.6624868512153625,
"logps/rejected": -60.173095703125,
"loss": 0.1174,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2967941761016846,
"rewards/margins": 2.2967941761016846,
"rewards/rejected": 0.0,
"step": 2647
},
{
"epoch": 14.793296089385475,
"grad_norm": 0.37660516483796985,
"learning_rate": 1.7234849711722842e-08,
"logits/chosen": -3.545567512512207,
"logits/rejected": -3.5239815711975098,
"logps/chosen": -2.9479568004608154,
"logps/rejected": -100.97955322265625,
"loss": 0.0839,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.462566375732422,
"rewards/margins": 2.462566375732422,
"rewards/rejected": 0.0,
"step": 2648
},
{
"epoch": 14.798882681564246,
"grad_norm": 0.7143914445117544,
"learning_rate": 1.7076549720269116e-08,
"logits/chosen": -3.246804714202881,
"logits/rejected": -3.3367602825164795,
"logps/chosen": -0.2639075815677643,
"logps/rejected": -54.3329963684082,
"loss": 0.1355,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8515474796295166,
"rewards/margins": 1.8515474796295166,
"rewards/rejected": 0.0,
"step": 2649
},
{
"epoch": 14.804469273743017,
"grad_norm": 0.40277215037576825,
"learning_rate": 1.6918967439066344e-08,
"logits/chosen": -3.342604875564575,
"logits/rejected": -3.419504165649414,
"logps/chosen": -2.6654586791992188,
"logps/rejected": -57.11647033691406,
"loss": 0.0972,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9482383728027344,
"rewards/margins": 2.9482383728027344,
"rewards/rejected": 0.0,
"step": 2650
},
{
"epoch": 14.810055865921788,
"grad_norm": 0.4039025045873746,
"learning_rate": 1.6762103102310077e-08,
"logits/chosen": -3.5208730697631836,
"logits/rejected": -2.899697780609131,
"logps/chosen": -0.5285656452178955,
"logps/rejected": -153.28765869140625,
"loss": 0.106,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.113286256790161,
"rewards/margins": 2.113286256790161,
"rewards/rejected": 0.0,
"step": 2651
},
{
"epoch": 14.815642458100559,
"grad_norm": 0.3418434975746601,
"learning_rate": 1.660595694312844e-08,
"logits/chosen": -3.465338945388794,
"logits/rejected": -3.5613439083099365,
"logps/chosen": -1.449244737625122,
"logps/rejected": -83.42619323730469,
"loss": 0.1193,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.855820655822754,
"rewards/margins": 2.855820655822754,
"rewards/rejected": 0.0,
"step": 2652
},
{
"epoch": 14.82122905027933,
"grad_norm": 0.3613168884981856,
"learning_rate": 1.6450529193582586e-08,
"logits/chosen": -3.3460452556610107,
"logits/rejected": -3.55086350440979,
"logps/chosen": -0.679241418838501,
"logps/rejected": -49.6741943359375,
"loss": 0.1232,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1192545890808105,
"rewards/margins": 2.1192545890808105,
"rewards/rejected": 0.0,
"step": 2653
},
{
"epoch": 14.8268156424581,
"grad_norm": 0.35899306172298456,
"learning_rate": 1.6295820084665845e-08,
"logits/chosen": -3.6450414657592773,
"logits/rejected": -3.6122703552246094,
"logps/chosen": -20.39347267150879,
"logps/rejected": -44.38009262084961,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.465202808380127,
"rewards/margins": 2.465202808380127,
"rewards/rejected": 0.0,
"step": 2654
},
{
"epoch": 14.832402234636872,
"grad_norm": 0.32645729268390417,
"learning_rate": 1.6141829846303468e-08,
"logits/chosen": -3.470165491104126,
"logits/rejected": -3.3938283920288086,
"logps/chosen": -0.5699958205223083,
"logps/rejected": -48.6433219909668,
"loss": 0.1056,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4341936111450195,
"rewards/margins": 2.4341936111450195,
"rewards/rejected": 0.0,
"step": 2655
},
{
"epoch": 14.837988826815643,
"grad_norm": 0.5142779757414533,
"learning_rate": 1.5988558707352327e-08,
"logits/chosen": -3.477855682373047,
"logits/rejected": -3.513197183609009,
"logps/chosen": -48.18681716918945,
"logps/rejected": -48.22716522216797,
"loss": 0.0991,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.520964622497559,
"rewards/margins": 4.520964622497559,
"rewards/rejected": 0.0,
"step": 2656
},
{
"epoch": 14.843575418994414,
"grad_norm": 0.48715230823999733,
"learning_rate": 1.5836006895600717e-08,
"logits/chosen": -3.271955966949463,
"logits/rejected": -3.4795589447021484,
"logps/chosen": -0.3449538052082062,
"logps/rejected": -76.18074035644531,
"loss": 0.1349,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.436866044998169,
"rewards/margins": 1.436866044998169,
"rewards/rejected": 0.0,
"step": 2657
},
{
"epoch": 14.849162011173185,
"grad_norm": 0.3772270483117669,
"learning_rate": 1.5684174637767723e-08,
"logits/chosen": -3.569859266281128,
"logits/rejected": -3.5872387886047363,
"logps/chosen": -40.804996490478516,
"logps/rejected": -63.99356460571289,
"loss": 0.0954,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.621565818786621,
"rewards/margins": 3.621565818786621,
"rewards/rejected": 0.0,
"step": 2658
},
{
"epoch": 14.854748603351956,
"grad_norm": 0.6019501026357243,
"learning_rate": 1.5533062159503064e-08,
"logits/chosen": -3.6645209789276123,
"logits/rejected": -3.5907912254333496,
"logps/chosen": -11.660226821899414,
"logps/rejected": -94.15673065185547,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.636493682861328,
"rewards/margins": 2.636493682861328,
"rewards/rejected": 0.0,
"step": 2659
},
{
"epoch": 14.860335195530727,
"grad_norm": 0.39926621115797817,
"learning_rate": 1.538266968538704e-08,
"logits/chosen": -3.1913115978240967,
"logits/rejected": -3.3153505325317383,
"logps/chosen": -14.5741605758667,
"logps/rejected": -61.53295135498047,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4079737663269043,
"rewards/margins": 3.4079737663269043,
"rewards/rejected": 0.0,
"step": 2660
},
{
"epoch": 14.860335195530727,
"eval_logits/chosen": -3.2547249794006348,
"eval_logits/rejected": -3.3794593811035156,
"eval_logps/chosen": -30.67974281311035,
"eval_logps/rejected": -55.7110481262207,
"eval_loss": 1.0554580688476562,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.060970108956098557,
"eval_rewards/margins": -0.060970108956098557,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7183,
"eval_samples_per_second": 9.475,
"eval_steps_per_second": 0.306,
"step": 2660
},
{
"epoch": 14.865921787709498,
"grad_norm": 0.4576728086417737,
"learning_rate": 1.5232997438929307e-08,
"logits/chosen": -3.0724565982818604,
"logits/rejected": -3.2148144245147705,
"logps/chosen": -2.644747018814087,
"logps/rejected": -51.407508850097656,
"loss": 0.0933,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1311590671539307,
"rewards/margins": 2.1311590671539307,
"rewards/rejected": 0.0,
"step": 2661
},
{
"epoch": 14.871508379888269,
"grad_norm": 0.3438935345685653,
"learning_rate": 1.5084045642569764e-08,
"logits/chosen": -3.4441304206848145,
"logits/rejected": -3.6405422687530518,
"logps/chosen": -32.07526779174805,
"logps/rejected": -113.20217895507812,
"loss": 0.0879,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.755329132080078,
"rewards/margins": 2.755329132080078,
"rewards/rejected": 0.0,
"step": 2662
},
{
"epoch": 14.87709497206704,
"grad_norm": 0.35839125398176075,
"learning_rate": 1.4935814517677225e-08,
"logits/chosen": -3.478083372116089,
"logits/rejected": -3.4834911823272705,
"logps/chosen": -14.865629196166992,
"logps/rejected": -74.30992126464844,
"loss": 0.1284,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5339725017547607,
"rewards/margins": 2.5339725017547607,
"rewards/rejected": 0.0,
"step": 2663
},
{
"epoch": 14.88268156424581,
"grad_norm": 0.417511190026462,
"learning_rate": 1.4788304284549468e-08,
"logits/chosen": -3.2606968879699707,
"logits/rejected": -3.337392568588257,
"logps/chosen": -0.19713765382766724,
"logps/rejected": -74.63046264648438,
"loss": 0.0982,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7466249465942383,
"rewards/margins": 1.7466249465942383,
"rewards/rejected": 0.0,
"step": 2664
},
{
"epoch": 14.888268156424582,
"grad_norm": 0.7624378878488587,
"learning_rate": 1.4641515162413021e-08,
"logits/chosen": -3.3800857067108154,
"logits/rejected": -3.2392077445983887,
"logps/chosen": -5.718407154083252,
"logps/rejected": -102.63508605957031,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.418631076812744,
"rewards/margins": 2.418631076812744,
"rewards/rejected": 0.0,
"step": 2665
},
{
"epoch": 14.893854748603353,
"grad_norm": 0.3408636206728141,
"learning_rate": 1.4495447369422765e-08,
"logits/chosen": -3.474879264831543,
"logits/rejected": -3.3781349658966064,
"logps/chosen": -19.407983779907227,
"logps/rejected": -40.66181564331055,
"loss": 0.0863,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.638751983642578,
"rewards/margins": 3.638751983642578,
"rewards/rejected": 0.0,
"step": 2666
},
{
"epoch": 14.899441340782122,
"grad_norm": 0.3584504928456228,
"learning_rate": 1.4350101122661219e-08,
"logits/chosen": -3.325920820236206,
"logits/rejected": -3.4892213344573975,
"logps/chosen": -0.6450057625770569,
"logps/rejected": -92.8226547241211,
"loss": 0.0924,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.411738634109497,
"rewards/margins": 2.411738634109497,
"rewards/rejected": 0.0,
"step": 2667
},
{
"epoch": 14.905027932960895,
"grad_norm": 0.3456695753014961,
"learning_rate": 1.4205476638138924e-08,
"logits/chosen": -3.2882425785064697,
"logits/rejected": -3.3540515899658203,
"logps/chosen": -0.23389209806919098,
"logps/rejected": -79.29186248779297,
"loss": 0.1086,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7516508102416992,
"rewards/margins": 1.7516508102416992,
"rewards/rejected": 0.0,
"step": 2668
},
{
"epoch": 14.910614525139664,
"grad_norm": 0.36433128295571626,
"learning_rate": 1.406157413079373e-08,
"logits/chosen": -3.3356049060821533,
"logits/rejected": -3.394909620285034,
"logps/chosen": -7.042104721069336,
"logps/rejected": -54.11695098876953,
"loss": 0.0829,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.572568893432617,
"rewards/margins": 3.572568893432617,
"rewards/rejected": 0.0,
"step": 2669
},
{
"epoch": 14.916201117318435,
"grad_norm": 0.3917565352504246,
"learning_rate": 1.391839381449017e-08,
"logits/chosen": -3.6086223125457764,
"logits/rejected": -3.6328821182250977,
"logps/chosen": -17.311904907226562,
"logps/rejected": -64.36416625976562,
"loss": 0.0721,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.573361396789551,
"rewards/margins": 3.573361396789551,
"rewards/rejected": 0.0,
"step": 2670
},
{
"epoch": 14.921787709497206,
"grad_norm": 0.36165601537723724,
"learning_rate": 1.3775935902019808e-08,
"logits/chosen": -3.27400803565979,
"logits/rejected": -3.322840452194214,
"logps/chosen": -0.5292971134185791,
"logps/rejected": -99.6595230102539,
"loss": 0.0904,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.121829032897949,
"rewards/margins": 2.121829032897949,
"rewards/rejected": 0.0,
"step": 2671
},
{
"epoch": 14.927374301675977,
"grad_norm": 0.3660794490358241,
"learning_rate": 1.3634200605100509e-08,
"logits/chosen": -3.366211414337158,
"logits/rejected": -3.605614185333252,
"logps/chosen": -0.6540247201919556,
"logps/rejected": -90.26676940917969,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.179960250854492,
"rewards/margins": 2.179960250854492,
"rewards/rejected": 0.0,
"step": 2672
},
{
"epoch": 14.932960893854748,
"grad_norm": 0.39284028631774637,
"learning_rate": 1.349318813437611e-08,
"logits/chosen": -3.196241617202759,
"logits/rejected": -3.1078386306762695,
"logps/chosen": -5.704566955566406,
"logps/rejected": -43.353031158447266,
"loss": 0.1064,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7156381607055664,
"rewards/margins": 2.7156381607055664,
"rewards/rejected": 0.0,
"step": 2673
},
{
"epoch": 14.938547486033519,
"grad_norm": 1.3109298078540716,
"learning_rate": 1.3352898699416248e-08,
"logits/chosen": -3.215620517730713,
"logits/rejected": -3.4648029804229736,
"logps/chosen": -25.826221466064453,
"logps/rejected": -37.723182678222656,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5699305534362793,
"rewards/margins": 3.5699305534362793,
"rewards/rejected": 0.0,
"step": 2674
},
{
"epoch": 14.94413407821229,
"grad_norm": 0.39928406022927004,
"learning_rate": 1.321333250871609e-08,
"logits/chosen": -3.2574706077575684,
"logits/rejected": -3.259279251098633,
"logps/chosen": -1.2326209545135498,
"logps/rejected": -86.41130828857422,
"loss": 0.089,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9776511192321777,
"rewards/margins": 1.9776511192321777,
"rewards/rejected": 0.0,
"step": 2675
},
{
"epoch": 14.949720670391061,
"grad_norm": 0.4625925721043822,
"learning_rate": 1.3074489769695718e-08,
"logits/chosen": -3.4715349674224854,
"logits/rejected": -3.3700530529022217,
"logps/chosen": -3.6988751888275146,
"logps/rejected": -30.277475357055664,
"loss": 0.1116,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.283562183380127,
"rewards/margins": 3.283562183380127,
"rewards/rejected": 0.0,
"step": 2676
},
{
"epoch": 14.955307262569832,
"grad_norm": 0.35754320889886004,
"learning_rate": 1.2936370688700238e-08,
"logits/chosen": -3.3388655185699463,
"logits/rejected": -3.2582459449768066,
"logps/chosen": -1.3774136304855347,
"logps/rejected": -40.6134033203125,
"loss": 0.0701,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.365615129470825,
"rewards/margins": 3.365615129470825,
"rewards/rejected": 0.0,
"step": 2677
},
{
"epoch": 14.960893854748603,
"grad_norm": 0.40319193458720853,
"learning_rate": 1.2798975470999284e-08,
"logits/chosen": -3.1612911224365234,
"logits/rejected": -3.018653154373169,
"logps/chosen": -0.9814417362213135,
"logps/rejected": -51.494178771972656,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3663196563720703,
"rewards/margins": 2.3663196563720703,
"rewards/rejected": 0.0,
"step": 2678
},
{
"epoch": 14.966480446927374,
"grad_norm": 0.7016258468644363,
"learning_rate": 1.266230432078641e-08,
"logits/chosen": -3.057927131652832,
"logits/rejected": -3.1106045246124268,
"logps/chosen": -1.153242588043213,
"logps/rejected": -37.66864013671875,
"loss": 0.1007,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3504390716552734,
"rewards/margins": 2.3504390716552734,
"rewards/rejected": 0.0,
"step": 2679
},
{
"epoch": 14.972067039106145,
"grad_norm": 0.38532646113777913,
"learning_rate": 1.2526357441179525e-08,
"logits/chosen": -3.4418842792510986,
"logits/rejected": -3.4901740550994873,
"logps/chosen": -3.085214614868164,
"logps/rejected": -59.00658416748047,
"loss": 0.1146,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.921659469604492,
"rewards/margins": 2.921659469604492,
"rewards/rejected": 0.0,
"step": 2680
},
{
"epoch": 14.972067039106145,
"eval_logits/chosen": -3.2502505779266357,
"eval_logits/rejected": -3.375441312789917,
"eval_logps/chosen": -30.633203506469727,
"eval_logps/rejected": -55.960472106933594,
"eval_loss": 1.0556432008743286,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.05631597712635994,
"eval_rewards/margins": -0.05631597712635994,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7228,
"eval_samples_per_second": 9.474,
"eval_steps_per_second": 0.306,
"step": 2680
},
{
"epoch": 14.977653631284916,
"grad_norm": 0.4249891450285254,
"learning_rate": 1.239113503421968e-08,
"logits/chosen": -3.3209052085876465,
"logits/rejected": -3.5237913131713867,
"logps/chosen": -20.31497573852539,
"logps/rejected": -75.81820678710938,
"loss": 0.0874,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.690044403076172,
"rewards/margins": 3.690044403076172,
"rewards/rejected": 0.0,
"step": 2681
},
{
"epoch": 14.983240223463687,
"grad_norm": 0.9186611554219153,
"learning_rate": 1.2256637300871619e-08,
"logits/chosen": -3.2406022548675537,
"logits/rejected": -3.4094669818878174,
"logps/chosen": -0.22653105854988098,
"logps/rejected": -88.873779296875,
"loss": 0.0887,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.779331922531128,
"rewards/margins": 1.779331922531128,
"rewards/rejected": 0.0,
"step": 2682
},
{
"epoch": 14.988826815642458,
"grad_norm": 0.3797020995264553,
"learning_rate": 1.2122864441022896e-08,
"logits/chosen": -3.150249719619751,
"logits/rejected": -3.466625213623047,
"logps/chosen": -2.8326287269592285,
"logps/rejected": -87.0802001953125,
"loss": 0.1132,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5402469635009766,
"rewards/margins": 2.5402469635009766,
"rewards/rejected": 0.0,
"step": 2683
},
{
"epoch": 14.994413407821229,
"grad_norm": 0.5331114849910051,
"learning_rate": 1.1989816653483753e-08,
"logits/chosen": -3.565075159072876,
"logits/rejected": -3.598769426345825,
"logps/chosen": -1.9038105010986328,
"logps/rejected": -47.03626251220703,
"loss": 0.0681,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2623674869537354,
"rewards/margins": 3.2623674869537354,
"rewards/rejected": 0.0,
"step": 2684
},
{
"epoch": 15.0,
"grad_norm": 0.38551657826205105,
"learning_rate": 1.1857494135986857e-08,
"logits/chosen": -3.2997381687164307,
"logits/rejected": -3.363992214202881,
"logps/chosen": -0.6210635900497437,
"logps/rejected": -45.32338333129883,
"loss": 0.1122,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5469043254852295,
"rewards/margins": 2.5469043254852295,
"rewards/rejected": 0.0,
"step": 2685
},
{
"epoch": 15.005586592178771,
"grad_norm": 0.39172324615560666,
"learning_rate": 1.172589708518723e-08,
"logits/chosen": -3.440688133239746,
"logits/rejected": -3.56842041015625,
"logps/chosen": -2.112839698791504,
"logps/rejected": -47.4237174987793,
"loss": 0.0762,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6917471885681152,
"rewards/margins": 2.6917471885681152,
"rewards/rejected": 0.0,
"step": 2686
},
{
"epoch": 15.011173184357542,
"grad_norm": 0.4049448507105249,
"learning_rate": 1.1595025696661375e-08,
"logits/chosen": -3.44490385055542,
"logits/rejected": -3.497697353363037,
"logps/chosen": -4.48445987701416,
"logps/rejected": -65.00106811523438,
"loss": 0.1011,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.183898448944092,
"rewards/margins": 3.183898448944092,
"rewards/rejected": 0.0,
"step": 2687
},
{
"epoch": 15.016759776536313,
"grad_norm": 0.3432792319540246,
"learning_rate": 1.1464880164907376e-08,
"logits/chosen": -3.2594828605651855,
"logits/rejected": -3.5466315746307373,
"logps/chosen": -1.9839377403259277,
"logps/rejected": -102.76213836669922,
"loss": 0.0769,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6706361770629883,
"rewards/margins": 2.6706361770629883,
"rewards/rejected": 0.0,
"step": 2688
},
{
"epoch": 15.022346368715084,
"grad_norm": 0.4694490269016035,
"learning_rate": 1.1335460683344955e-08,
"logits/chosen": -3.1902191638946533,
"logits/rejected": -3.242811679840088,
"logps/chosen": -4.908884048461914,
"logps/rejected": -54.772071838378906,
"loss": 0.0885,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.411595582962036,
"rewards/margins": 3.411595582962036,
"rewards/rejected": 0.0,
"step": 2689
},
{
"epoch": 15.027932960893855,
"grad_norm": 0.4142338154777935,
"learning_rate": 1.1206767444314202e-08,
"logits/chosen": -3.1747968196868896,
"logits/rejected": -3.1738836765289307,
"logps/chosen": -4.6701812744140625,
"logps/rejected": -58.981422424316406,
"loss": 0.1169,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1388354301452637,
"rewards/margins": 2.1388354301452637,
"rewards/rejected": 0.0,
"step": 2690
},
{
"epoch": 15.033519553072626,
"grad_norm": 0.3075111500743282,
"learning_rate": 1.1078800639076458e-08,
"logits/chosen": -3.4383513927459717,
"logits/rejected": -3.5019662380218506,
"logps/chosen": -6.686107158660889,
"logps/rejected": -43.87648010253906,
"loss": 0.0966,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.979459285736084,
"rewards/margins": 2.979459285736084,
"rewards/rejected": 0.0,
"step": 2691
},
{
"epoch": 15.039106145251397,
"grad_norm": 0.4069977774294662,
"learning_rate": 1.0951560457813036e-08,
"logits/chosen": -3.448096990585327,
"logits/rejected": -3.5652003288269043,
"logps/chosen": -4.753368854522705,
"logps/rejected": -47.4362907409668,
"loss": 0.0832,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.358572483062744,
"rewards/margins": 3.358572483062744,
"rewards/rejected": 0.0,
"step": 2692
},
{
"epoch": 15.044692737430168,
"grad_norm": 0.5367508361295136,
"learning_rate": 1.0825047089625505e-08,
"logits/chosen": -3.3518686294555664,
"logits/rejected": -3.2453532218933105,
"logps/chosen": -0.5641560554504395,
"logps/rejected": -35.182167053222656,
"loss": 0.1216,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9657347202301025,
"rewards/margins": 1.9657347202301025,
"rewards/rejected": 0.0,
"step": 2693
},
{
"epoch": 15.050279329608939,
"grad_norm": 0.7600421614889334,
"learning_rate": 1.0699260722535241e-08,
"logits/chosen": -3.469454526901245,
"logits/rejected": -3.1513831615448,
"logps/chosen": -12.951601028442383,
"logps/rejected": -59.583045959472656,
"loss": 0.1334,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.727524757385254,
"rewards/margins": 2.727524757385254,
"rewards/rejected": 0.0,
"step": 2694
},
{
"epoch": 15.05586592178771,
"grad_norm": 0.3705788733523154,
"learning_rate": 1.0574201543483375e-08,
"logits/chosen": -3.3381845951080322,
"logits/rejected": -3.3471267223358154,
"logps/chosen": -2.429077386856079,
"logps/rejected": -49.2909049987793,
"loss": 0.0995,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4597177505493164,
"rewards/margins": 3.4597177505493164,
"rewards/rejected": 0.0,
"step": 2695
},
{
"epoch": 15.061452513966481,
"grad_norm": 0.32981565886700426,
"learning_rate": 1.0449869738329952e-08,
"logits/chosen": -3.3754477500915527,
"logits/rejected": -3.1595022678375244,
"logps/chosen": -19.427053451538086,
"logps/rejected": -65.10870361328125,
"loss": 0.0815,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7762484550476074,
"rewards/margins": 3.7762484550476074,
"rewards/rejected": 0.0,
"step": 2696
},
{
"epoch": 15.067039106145252,
"grad_norm": 0.41375104021716336,
"learning_rate": 1.0326265491854113e-08,
"logits/chosen": -3.4222311973571777,
"logits/rejected": -3.4941632747650146,
"logps/chosen": -11.23513412475586,
"logps/rejected": -70.00746154785156,
"loss": 0.0823,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.793848752975464,
"rewards/margins": 3.793848752975464,
"rewards/rejected": 0.0,
"step": 2697
},
{
"epoch": 15.072625698324023,
"grad_norm": 0.336578667935485,
"learning_rate": 1.0203388987753969e-08,
"logits/chosen": -3.333406925201416,
"logits/rejected": -3.3734652996063232,
"logps/chosen": -1.4314913749694824,
"logps/rejected": -46.64700698852539,
"loss": 0.0924,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6283531188964844,
"rewards/margins": 2.6283531188964844,
"rewards/rejected": 0.0,
"step": 2698
},
{
"epoch": 15.078212290502794,
"grad_norm": 0.5739988082155488,
"learning_rate": 1.008124040864572e-08,
"logits/chosen": -3.5269570350646973,
"logits/rejected": -3.4450862407684326,
"logps/chosen": -3.9541525840759277,
"logps/rejected": -48.523494720458984,
"loss": 0.0776,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.282543659210205,
"rewards/margins": 3.282543659210205,
"rewards/rejected": 0.0,
"step": 2699
},
{
"epoch": 15.083798882681565,
"grad_norm": 0.4819194039282161,
"learning_rate": 9.95981993606404e-09,
"logits/chosen": -3.373324394226074,
"logits/rejected": -3.571223258972168,
"logps/chosen": -0.398854523897171,
"logps/rejected": -76.59867095947266,
"loss": 0.0965,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2527942657470703,
"rewards/margins": 2.2527942657470703,
"rewards/rejected": 0.0,
"step": 2700
},
{
"epoch": 15.083798882681565,
"eval_logits/chosen": -3.2535011768341064,
"eval_logits/rejected": -3.377370834350586,
"eval_logps/chosen": -30.89455223083496,
"eval_logps/rejected": -56.2887077331543,
"eval_loss": 1.0546748638153076,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.08245096355676651,
"eval_rewards/margins": -0.08245096355676651,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6911,
"eval_samples_per_second": 9.483,
"eval_steps_per_second": 0.306,
"step": 2700
},
{
"epoch": 15.089385474860336,
"grad_norm": 0.5165000292690357,
"learning_rate": 9.839127750461307e-09,
"logits/chosen": -3.2869677543640137,
"logits/rejected": -3.3428895473480225,
"logps/chosen": -12.017826080322266,
"logps/rejected": -70.00727844238281,
"loss": 0.1354,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.796191453933716,
"rewards/margins": 2.796191453933716,
"rewards/rejected": 0.0,
"step": 2701
},
{
"epoch": 15.094972067039107,
"grad_norm": 0.9315881207818567,
"learning_rate": 9.719164031207594e-09,
"logits/chosen": -3.3606231212615967,
"logits/rejected": -3.4705748558044434,
"logps/chosen": -4.759427547454834,
"logps/rejected": -79.03294372558594,
"loss": 0.1026,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3535492420196533,
"rewards/margins": 3.3535492420196533,
"rewards/rejected": 0.0,
"step": 2702
},
{
"epoch": 15.100558659217878,
"grad_norm": 0.3500163013017413,
"learning_rate": 9.599928956590341e-09,
"logits/chosen": -3.4071614742279053,
"logits/rejected": -3.410222291946411,
"logps/chosen": -0.421855628490448,
"logps/rejected": -62.42582702636719,
"loss": 0.083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2876079082489014,
"rewards/margins": 2.2876079082489014,
"rewards/rejected": 0.0,
"step": 2703
},
{
"epoch": 15.106145251396647,
"grad_norm": 0.3711339402041741,
"learning_rate": 9.48142270381419e-09,
"logits/chosen": -3.186708688735962,
"logits/rejected": -3.0210330486297607,
"logps/chosen": -0.47165483236312866,
"logps/rejected": -112.34162139892578,
"loss": 0.1204,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1767287254333496,
"rewards/margins": 2.1767287254333496,
"rewards/rejected": 0.0,
"step": 2704
},
{
"epoch": 15.111731843575418,
"grad_norm": 0.3545778721974017,
"learning_rate": 9.363645449000424e-09,
"logits/chosen": -3.1981797218322754,
"logits/rejected": -3.3513071537017822,
"logps/chosen": -2.14025616645813,
"logps/rejected": -47.29647445678711,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.368718385696411,
"rewards/margins": 3.368718385696411,
"rewards/rejected": 0.0,
"step": 2705
},
{
"epoch": 15.11731843575419,
"grad_norm": 0.36042445074666246,
"learning_rate": 9.246597367187137e-09,
"logits/chosen": -3.6300017833709717,
"logits/rejected": -3.638338088989258,
"logps/chosen": -0.6851445436477661,
"logps/rejected": -44.4206657409668,
"loss": 0.1063,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4193100929260254,
"rewards/margins": 2.4193100929260254,
"rewards/rejected": 0.0,
"step": 2706
},
{
"epoch": 15.12290502793296,
"grad_norm": 0.35750649403745627,
"learning_rate": 9.13027863232857e-09,
"logits/chosen": -3.5256896018981934,
"logits/rejected": -3.6527068614959717,
"logps/chosen": -0.9782887697219849,
"logps/rejected": -79.44715118408203,
"loss": 0.0996,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6410274505615234,
"rewards/margins": 2.6410274505615234,
"rewards/rejected": 0.0,
"step": 2707
},
{
"epoch": 15.128491620111731,
"grad_norm": 0.3815683215151135,
"learning_rate": 9.014689417294997e-09,
"logits/chosen": -3.341118335723877,
"logits/rejected": -3.4082393646240234,
"logps/chosen": -4.818462371826172,
"logps/rejected": -28.40052032470703,
"loss": 0.0648,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1663920879364014,
"rewards/margins": 3.1663920879364014,
"rewards/rejected": 0.0,
"step": 2708
},
{
"epoch": 15.134078212290502,
"grad_norm": 0.32520778180684184,
"learning_rate": 8.899829893872723e-09,
"logits/chosen": -3.397822141647339,
"logits/rejected": -3.574009418487549,
"logps/chosen": -0.23265886306762695,
"logps/rejected": -67.71945190429688,
"loss": 0.1075,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8376045227050781,
"rewards/margins": 1.8376045227050781,
"rewards/rejected": 0.0,
"step": 2709
},
{
"epoch": 15.139664804469273,
"grad_norm": 0.4223280329092187,
"learning_rate": 8.785700232763316e-09,
"logits/chosen": -3.484344959259033,
"logits/rejected": -3.507127046585083,
"logps/chosen": -7.165587902069092,
"logps/rejected": -36.34278106689453,
"loss": 0.0869,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9120960235595703,
"rewards/margins": 2.9120960235595703,
"rewards/rejected": 0.0,
"step": 2710
},
{
"epoch": 15.145251396648044,
"grad_norm": 0.35484509768665096,
"learning_rate": 8.672300603583927e-09,
"logits/chosen": -3.5402567386627197,
"logits/rejected": -3.580690860748291,
"logps/chosen": -0.25957396626472473,
"logps/rejected": -25.984642028808594,
"loss": 0.1108,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6842262744903564,
"rewards/margins": 1.6842262744903564,
"rewards/rejected": 0.0,
"step": 2711
},
{
"epoch": 15.150837988826815,
"grad_norm": 0.5961823887947978,
"learning_rate": 8.559631174866466e-09,
"logits/chosen": -3.2367472648620605,
"logits/rejected": -3.441416025161743,
"logps/chosen": -0.6959924697875977,
"logps/rejected": -58.732383728027344,
"loss": 0.1091,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.281392812728882,
"rewards/margins": 2.281392812728882,
"rewards/rejected": 0.0,
"step": 2712
},
{
"epoch": 15.156424581005586,
"grad_norm": 0.3467801645805499,
"learning_rate": 8.447692114057825e-09,
"logits/chosen": -3.3672757148742676,
"logits/rejected": -3.531606912612915,
"logps/chosen": -4.9795637130737305,
"logps/rejected": -62.002288818359375,
"loss": 0.1041,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.696640729904175,
"rewards/margins": 2.696640729904175,
"rewards/rejected": 0.0,
"step": 2713
},
{
"epoch": 15.162011173184357,
"grad_norm": 0.4039314376388327,
"learning_rate": 8.336483587519317e-09,
"logits/chosen": -3.610227346420288,
"logits/rejected": -3.4002838134765625,
"logps/chosen": -0.809124231338501,
"logps/rejected": -57.50293731689453,
"loss": 0.0998,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6413872241973877,
"rewards/margins": 2.6413872241973877,
"rewards/rejected": 0.0,
"step": 2714
},
{
"epoch": 15.167597765363128,
"grad_norm": 0.40278736650436614,
"learning_rate": 8.226005760526788e-09,
"logits/chosen": -3.2914633750915527,
"logits/rejected": -3.4402213096618652,
"logps/chosen": -1.4635589122772217,
"logps/rejected": -75.35786437988281,
"loss": 0.0918,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.584859609603882,
"rewards/margins": 2.584859609603882,
"rewards/rejected": 0.0,
"step": 2715
},
{
"epoch": 15.1731843575419,
"grad_norm": 0.45085494048910235,
"learning_rate": 8.116258797269792e-09,
"logits/chosen": -3.327390432357788,
"logits/rejected": -3.331871271133423,
"logps/chosen": -5.626540184020996,
"logps/rejected": -59.60590362548828,
"loss": 0.1071,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6810355186462402,
"rewards/margins": 2.6810355186462402,
"rewards/rejected": 0.0,
"step": 2716
},
{
"epoch": 15.17877094972067,
"grad_norm": 0.37558104584062896,
"learning_rate": 8.00724286085197e-09,
"logits/chosen": -3.3111743927001953,
"logits/rejected": -3.431515693664551,
"logps/chosen": -0.6694455742835999,
"logps/rejected": -63.245025634765625,
"loss": 0.1032,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.226891279220581,
"rewards/margins": 2.226891279220581,
"rewards/rejected": 0.0,
"step": 2717
},
{
"epoch": 15.184357541899441,
"grad_norm": 0.38529163976970043,
"learning_rate": 7.898958113290389e-09,
"logits/chosen": -3.3776395320892334,
"logits/rejected": -3.3438429832458496,
"logps/chosen": -1.2175061702728271,
"logps/rejected": -47.03485870361328,
"loss": 0.0936,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5619194507598877,
"rewards/margins": 2.5619194507598877,
"rewards/rejected": 0.0,
"step": 2718
},
{
"epoch": 15.189944134078212,
"grad_norm": 0.4038537049443026,
"learning_rate": 7.79140471551537e-09,
"logits/chosen": -3.1527297496795654,
"logits/rejected": -2.9282472133636475,
"logps/chosen": -3.5159504413604736,
"logps/rejected": -117.24052429199219,
"loss": 0.1005,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7162249088287354,
"rewards/margins": 2.7162249088287354,
"rewards/rejected": 0.0,
"step": 2719
},
{
"epoch": 15.195530726256983,
"grad_norm": 0.743766690694123,
"learning_rate": 7.684582827370556e-09,
"logits/chosen": -3.243110179901123,
"logits/rejected": -3.5146403312683105,
"logps/chosen": -6.167523384094238,
"logps/rejected": -46.33700180053711,
"loss": 0.0833,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9911394119262695,
"rewards/margins": 2.9911394119262695,
"rewards/rejected": 0.0,
"step": 2720
},
{
"epoch": 15.195530726256983,
"eval_logits/chosen": -3.2534797191619873,
"eval_logits/rejected": -3.377426862716675,
"eval_logps/chosen": -30.66217613220215,
"eval_logps/rejected": -55.8941764831543,
"eval_loss": 1.0542831420898438,
"eval_rewards/accuracies": 0.574999988079071,
"eval_rewards/chosen": -0.05921328067779541,
"eval_rewards/margins": -0.05921328067779541,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7089,
"eval_samples_per_second": 9.478,
"eval_steps_per_second": 0.306,
"step": 2720
},
{
"epoch": 15.201117318435754,
"grad_norm": 0.3434447764060888,
"learning_rate": 7.57849260761223e-09,
"logits/chosen": -3.464975595474243,
"logits/rejected": -3.4877188205718994,
"logps/chosen": -0.31153929233551025,
"logps/rejected": -44.04620361328125,
"loss": 0.0923,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0053279399871826,
"rewards/margins": 2.0053279399871826,
"rewards/rejected": 0.0,
"step": 2721
},
{
"epoch": 15.206703910614525,
"grad_norm": 0.46505473654441737,
"learning_rate": 7.473134213909326e-09,
"logits/chosen": -3.2479467391967773,
"logits/rejected": -3.1812796592712402,
"logps/chosen": -10.523893356323242,
"logps/rejected": -47.685768127441406,
"loss": 0.1135,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1481878757476807,
"rewards/margins": 3.1481878757476807,
"rewards/rejected": 0.0,
"step": 2722
},
{
"epoch": 15.212290502793296,
"grad_norm": 0.3462974578647901,
"learning_rate": 7.368507802843205e-09,
"logits/chosen": -2.9624855518341064,
"logits/rejected": -3.018097400665283,
"logps/chosen": -8.608769416809082,
"logps/rejected": -70.04875183105469,
"loss": 0.0911,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.850444793701172,
"rewards/margins": 2.850444793701172,
"rewards/rejected": 0.0,
"step": 2723
},
{
"epoch": 15.217877094972067,
"grad_norm": 0.482276154279113,
"learning_rate": 7.264613529907482e-09,
"logits/chosen": -3.4359025955200195,
"logits/rejected": -3.4081170558929443,
"logps/chosen": -6.527714729309082,
"logps/rejected": -41.29267501831055,
"loss": 0.0779,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.407116651535034,
"rewards/margins": 3.407116651535034,
"rewards/rejected": 0.0,
"step": 2724
},
{
"epoch": 15.223463687150838,
"grad_norm": 0.4081627577822981,
"learning_rate": 7.1614515495073134e-09,
"logits/chosen": -3.299471616744995,
"logits/rejected": -3.2404634952545166,
"logps/chosen": -0.8103320598602295,
"logps/rejected": -84.19842529296875,
"loss": 0.0992,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.342984199523926,
"rewards/margins": 2.342984199523926,
"rewards/rejected": 0.0,
"step": 2725
},
{
"epoch": 15.22905027932961,
"grad_norm": 0.5901699901597808,
"learning_rate": 7.059022014959892e-09,
"logits/chosen": -3.3895773887634277,
"logits/rejected": -3.554718494415283,
"logps/chosen": -6.208129405975342,
"logps/rejected": -36.23834228515625,
"loss": 0.0726,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.03847599029541,
"rewards/margins": 4.03847599029541,
"rewards/rejected": 0.0,
"step": 2726
},
{
"epoch": 15.23463687150838,
"grad_norm": 0.3070817502639455,
"learning_rate": 6.957325078493836e-09,
"logits/chosen": -3.4733989238739014,
"logits/rejected": -3.4898452758789062,
"logps/chosen": -4.381089210510254,
"logps/rejected": -35.10344314575195,
"loss": 0.099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7187013626098633,
"rewards/margins": 2.7187013626098633,
"rewards/rejected": 0.0,
"step": 2727
},
{
"epoch": 15.240223463687151,
"grad_norm": 0.32389759578194693,
"learning_rate": 6.856360891248692e-09,
"logits/chosen": -3.2947194576263428,
"logits/rejected": -3.4516637325286865,
"logps/chosen": -4.336068630218506,
"logps/rejected": -58.00398635864258,
"loss": 0.1186,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1628036499023438,
"rewards/margins": 3.1628036499023438,
"rewards/rejected": 0.0,
"step": 2728
},
{
"epoch": 15.245810055865922,
"grad_norm": 0.3574632109553235,
"learning_rate": 6.756129603275373e-09,
"logits/chosen": -3.4045233726501465,
"logits/rejected": -3.68743896484375,
"logps/chosen": -4.604673862457275,
"logps/rejected": -64.30682373046875,
"loss": 0.0872,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2407960891723633,
"rewards/margins": 3.2407960891723633,
"rewards/rejected": 0.0,
"step": 2729
},
{
"epoch": 15.251396648044693,
"grad_norm": 0.3472052756323566,
"learning_rate": 6.656631363535448e-09,
"logits/chosen": -3.3194024562835693,
"logits/rejected": -3.561497211456299,
"logps/chosen": -29.88372039794922,
"logps/rejected": -46.24955368041992,
"loss": 0.0725,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.303365707397461,
"rewards/margins": 3.303365707397461,
"rewards/rejected": 0.0,
"step": 2730
},
{
"epoch": 15.256983240223464,
"grad_norm": 0.39174063896126554,
"learning_rate": 6.557866319900906e-09,
"logits/chosen": -3.5682806968688965,
"logits/rejected": -3.375464916229248,
"logps/chosen": -21.68202018737793,
"logps/rejected": -39.89384460449219,
"loss": 0.1002,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.635549545288086,
"rewards/margins": 3.635549545288086,
"rewards/rejected": 0.0,
"step": 2731
},
{
"epoch": 15.262569832402235,
"grad_norm": 0.35820702823310646,
"learning_rate": 6.459834619154281e-09,
"logits/chosen": -3.4970037937164307,
"logits/rejected": -3.3811521530151367,
"logps/chosen": -6.336273193359375,
"logps/rejected": -83.13807678222656,
"loss": 0.1113,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.971449375152588,
"rewards/margins": 2.971449375152588,
"rewards/rejected": 0.0,
"step": 2732
},
{
"epoch": 15.268156424581006,
"grad_norm": 0.41517351887392095,
"learning_rate": 6.362536406988084e-09,
"logits/chosen": -3.433579683303833,
"logits/rejected": -3.5727155208587646,
"logps/chosen": -6.247230529785156,
"logps/rejected": -26.58795928955078,
"loss": 0.0755,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5267505645751953,
"rewards/margins": 3.5267505645751953,
"rewards/rejected": 0.0,
"step": 2733
},
{
"epoch": 15.273743016759777,
"grad_norm": 0.43341762233507536,
"learning_rate": 6.265971828004812e-09,
"logits/chosen": -3.1191630363464355,
"logits/rejected": -3.0236997604370117,
"logps/chosen": -22.686269760131836,
"logps/rejected": -32.200618743896484,
"loss": 0.0805,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7814812660217285,
"rewards/margins": 3.7814812660217285,
"rewards/rejected": 0.0,
"step": 2734
},
{
"epoch": 15.279329608938548,
"grad_norm": 0.5873944055075848,
"learning_rate": 6.170141025716668e-09,
"logits/chosen": -3.465365409851074,
"logits/rejected": -3.3521463871002197,
"logps/chosen": -0.907586932182312,
"logps/rejected": -32.12206268310547,
"loss": 0.0849,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4782509803771973,
"rewards/margins": 2.4782509803771973,
"rewards/rejected": 0.0,
"step": 2735
},
{
"epoch": 15.28491620111732,
"grad_norm": 0.44320503233625297,
"learning_rate": 6.075044142545172e-09,
"logits/chosen": -3.2940800189971924,
"logits/rejected": -3.42421555519104,
"logps/chosen": -2.1653804779052734,
"logps/rejected": -38.06364822387695,
"loss": 0.0941,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0755441188812256,
"rewards/margins": 3.0755441188812256,
"rewards/rejected": 0.0,
"step": 2736
},
{
"epoch": 15.29050279329609,
"grad_norm": 0.48844686058675907,
"learning_rate": 5.980681319821379e-09,
"logits/chosen": -3.2081522941589355,
"logits/rejected": -3.2111096382141113,
"logps/chosen": -0.3553764820098877,
"logps/rejected": -47.57817077636719,
"loss": 0.1265,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0247111320495605,
"rewards/margins": 2.0247111320495605,
"rewards/rejected": 0.0,
"step": 2737
},
{
"epoch": 15.296089385474861,
"grad_norm": 0.40511396835768404,
"learning_rate": 5.887052697785166e-09,
"logits/chosen": -3.509035110473633,
"logits/rejected": -3.4871180057525635,
"logps/chosen": -1.976684808731079,
"logps/rejected": -31.386011123657227,
"loss": 0.1069,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5718722343444824,
"rewards/margins": 2.5718722343444824,
"rewards/rejected": 0.0,
"step": 2738
},
{
"epoch": 15.30167597765363,
"grad_norm": 0.31613930392954237,
"learning_rate": 5.794158415585282e-09,
"logits/chosen": -3.2557921409606934,
"logits/rejected": -3.3748040199279785,
"logps/chosen": -8.889276504516602,
"logps/rejected": -51.01299285888672,
"loss": 0.1245,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.503488540649414,
"rewards/margins": 3.503488540649414,
"rewards/rejected": 0.0,
"step": 2739
},
{
"epoch": 15.307262569832401,
"grad_norm": 0.35332353763257246,
"learning_rate": 5.701998611279291e-09,
"logits/chosen": -3.317143201828003,
"logits/rejected": -3.417137861251831,
"logps/chosen": -7.209802150726318,
"logps/rejected": -53.06700897216797,
"loss": 0.1128,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.875089645385742,
"rewards/margins": 2.875089645385742,
"rewards/rejected": 0.0,
"step": 2740
},
{
"epoch": 15.307262569832401,
"eval_logits/chosen": -3.248673677444458,
"eval_logits/rejected": -3.3741021156311035,
"eval_logps/chosen": -30.63228988647461,
"eval_logps/rejected": -55.99335479736328,
"eval_loss": 1.0509639978408813,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.05622493475675583,
"eval_rewards/margins": -0.05622493475675583,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7092,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 2740
},
{
"epoch": 15.312849162011172,
"grad_norm": 0.3515655917271136,
"learning_rate": 5.610573421832965e-09,
"logits/chosen": -3.2977492809295654,
"logits/rejected": -3.5348691940307617,
"logps/chosen": -2.017648935317993,
"logps/rejected": -41.454952239990234,
"loss": 0.0871,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.260664939880371,
"rewards/margins": 3.260664939880371,
"rewards/rejected": 0.0,
"step": 2741
},
{
"epoch": 15.318435754189943,
"grad_norm": 0.37468446332625743,
"learning_rate": 5.5198829831205606e-09,
"logits/chosen": -3.413870096206665,
"logits/rejected": -3.137200355529785,
"logps/chosen": -0.5963768362998962,
"logps/rejected": -49.070533752441406,
"loss": 0.0921,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7700161933898926,
"rewards/margins": 2.7700161933898926,
"rewards/rejected": 0.0,
"step": 2742
},
{
"epoch": 15.324022346368714,
"grad_norm": 0.35422052364738515,
"learning_rate": 5.429927429924042e-09,
"logits/chosen": -3.496981143951416,
"logits/rejected": -3.2775912284851074,
"logps/chosen": -6.259299278259277,
"logps/rejected": -80.9996337890625,
"loss": 0.1076,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.123628616333008,
"rewards/margins": 3.123628616333008,
"rewards/rejected": 0.0,
"step": 2743
},
{
"epoch": 15.329608938547485,
"grad_norm": 0.48379818395480784,
"learning_rate": 5.340706895933577e-09,
"logits/chosen": -3.5097808837890625,
"logits/rejected": -3.4879391193389893,
"logps/chosen": -7.59335994720459,
"logps/rejected": -28.682659149169922,
"loss": 0.1044,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4279065132141113,
"rewards/margins": 3.4279065132141113,
"rewards/rejected": 0.0,
"step": 2744
},
{
"epoch": 15.335195530726256,
"grad_norm": 0.4690569174309656,
"learning_rate": 5.2522215137465465e-09,
"logits/chosen": -3.30172061920166,
"logits/rejected": -3.314605712890625,
"logps/chosen": -0.37239521741867065,
"logps/rejected": -46.06013488769531,
"loss": 0.1152,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.7459301948547363,
"rewards/margins": 1.7459301948547363,
"rewards/rejected": 0.0,
"step": 2745
},
{
"epoch": 15.340782122905027,
"grad_norm": 0.36687165605599187,
"learning_rate": 5.164471414868199e-09,
"logits/chosen": -3.320753335952759,
"logits/rejected": -3.253136157989502,
"logps/chosen": -1.0493203401565552,
"logps/rejected": -22.281099319458008,
"loss": 0.0826,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4916725158691406,
"rewards/margins": 2.4916725158691406,
"rewards/rejected": 0.0,
"step": 2746
},
{
"epoch": 15.346368715083798,
"grad_norm": 0.3325703607712872,
"learning_rate": 5.0774567297106606e-09,
"logits/chosen": -3.440880298614502,
"logits/rejected": -3.3833158016204834,
"logps/chosen": -4.225191116333008,
"logps/rejected": -61.686500549316406,
"loss": 0.0958,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.345611095428467,
"rewards/margins": 3.345611095428467,
"rewards/rejected": 0.0,
"step": 2747
},
{
"epoch": 15.35195530726257,
"grad_norm": 0.35702848776093915,
"learning_rate": 4.99117758759332e-09,
"logits/chosen": -3.266876220703125,
"logits/rejected": -3.3380188941955566,
"logps/chosen": -0.42809247970581055,
"logps/rejected": -62.52800750732422,
"loss": 0.1115,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2515180110931396,
"rewards/margins": 2.2515180110931396,
"rewards/rejected": 0.0,
"step": 2748
},
{
"epoch": 15.35754189944134,
"grad_norm": 0.41493908331632684,
"learning_rate": 4.905634116742386e-09,
"logits/chosen": -3.363133430480957,
"logits/rejected": -3.332078695297241,
"logps/chosen": -1.516545295715332,
"logps/rejected": -64.26512145996094,
"loss": 0.1157,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.72746205329895,
"rewards/margins": 2.72746205329895,
"rewards/rejected": 0.0,
"step": 2749
},
{
"epoch": 15.363128491620111,
"grad_norm": 0.3838088181751727,
"learning_rate": 4.820826444290604e-09,
"logits/chosen": -3.287088394165039,
"logits/rejected": -3.380878210067749,
"logps/chosen": -6.814674377441406,
"logps/rejected": -46.192665100097656,
"loss": 0.1093,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.646423816680908,
"rewards/margins": 3.646423816680908,
"rewards/rejected": 0.0,
"step": 2750
},
{
"epoch": 15.368715083798882,
"grad_norm": 0.37050209695463104,
"learning_rate": 4.736754696277268e-09,
"logits/chosen": -3.5849552154541016,
"logits/rejected": -3.4903130531311035,
"logps/chosen": -34.72379684448242,
"logps/rejected": -35.2935791015625,
"loss": 0.1174,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.601708889007568,
"rewards/margins": 4.601708889007568,
"rewards/rejected": 0.0,
"step": 2751
},
{
"epoch": 15.374301675977653,
"grad_norm": 0.3436745519374388,
"learning_rate": 4.65341899764804e-09,
"logits/chosen": -3.405404806137085,
"logits/rejected": -3.4282948970794678,
"logps/chosen": -0.6028783321380615,
"logps/rejected": -52.68398666381836,
"loss": 0.0984,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1569461822509766,
"rewards/margins": 2.1569461822509766,
"rewards/rejected": 0.0,
"step": 2752
},
{
"epoch": 15.379888268156424,
"grad_norm": 0.4644076721999623,
"learning_rate": 4.570819472254628e-09,
"logits/chosen": -3.4488673210144043,
"logits/rejected": -3.379628896713257,
"logps/chosen": -0.7595265507698059,
"logps/rejected": -85.30735778808594,
"loss": 0.1259,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.232293128967285,
"rewards/margins": 2.232293128967285,
"rewards/rejected": 0.0,
"step": 2753
},
{
"epoch": 15.385474860335195,
"grad_norm": 0.34475775741966763,
"learning_rate": 4.488956242854503e-09,
"logits/chosen": -3.2192726135253906,
"logits/rejected": -3.2809946537017822,
"logps/chosen": -1.3715202808380127,
"logps/rejected": -53.71438980102539,
"loss": 0.1068,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.748518943786621,
"rewards/margins": 2.748518943786621,
"rewards/rejected": 0.0,
"step": 2754
},
{
"epoch": 15.391061452513966,
"grad_norm": 0.3098299269784389,
"learning_rate": 4.407829431111176e-09,
"logits/chosen": -3.5013234615325928,
"logits/rejected": -3.4125614166259766,
"logps/chosen": -0.7996332049369812,
"logps/rejected": -140.3853759765625,
"loss": 0.0782,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.476461410522461,
"rewards/margins": 2.476461410522461,
"rewards/rejected": 0.0,
"step": 2755
},
{
"epoch": 15.396648044692737,
"grad_norm": 0.4045607273273071,
"learning_rate": 4.3274391575933714e-09,
"logits/chosen": -3.5030741691589355,
"logits/rejected": -3.4345967769622803,
"logps/chosen": -0.5863586664199829,
"logps/rejected": -41.62118148803711,
"loss": 0.08,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4675443172454834,
"rewards/margins": 2.4675443172454834,
"rewards/rejected": 0.0,
"step": 2756
},
{
"epoch": 15.402234636871508,
"grad_norm": 0.37688208971115295,
"learning_rate": 4.247785541775517e-09,
"logits/chosen": -3.214529275894165,
"logits/rejected": -3.252412796020508,
"logps/chosen": -8.346531867980957,
"logps/rejected": -38.34511184692383,
"loss": 0.1164,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.976717472076416,
"rewards/margins": 3.976717472076416,
"rewards/rejected": 0.0,
"step": 2757
},
{
"epoch": 15.40782122905028,
"grad_norm": 0.3723122284990119,
"learning_rate": 4.168868702036921e-09,
"logits/chosen": -3.5542376041412354,
"logits/rejected": -3.650407314300537,
"logps/chosen": -0.21057796478271484,
"logps/rejected": -90.39620208740234,
"loss": 0.0905,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6911938190460205,
"rewards/margins": 1.6911938190460205,
"rewards/rejected": 0.0,
"step": 2758
},
{
"epoch": 15.41340782122905,
"grad_norm": 0.3624657042072919,
"learning_rate": 4.090688755662153e-09,
"logits/chosen": -3.197291612625122,
"logits/rejected": -2.999779224395752,
"logps/chosen": -0.9325481653213501,
"logps/rejected": -71.47624969482422,
"loss": 0.1334,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.152543544769287,
"rewards/margins": 2.152543544769287,
"rewards/rejected": 0.0,
"step": 2759
},
{
"epoch": 15.418994413407821,
"grad_norm": 0.3472619525538552,
"learning_rate": 4.0132458188405496e-09,
"logits/chosen": -3.5566887855529785,
"logits/rejected": -3.4715640544891357,
"logps/chosen": -8.182580947875977,
"logps/rejected": -33.7274169921875,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.491220712661743,
"rewards/margins": 3.491220712661743,
"rewards/rejected": 0.0,
"step": 2760
},
{
"epoch": 15.418994413407821,
"eval_logits/chosen": -3.256040573120117,
"eval_logits/rejected": -3.3804473876953125,
"eval_logps/chosen": -30.82790756225586,
"eval_logps/rejected": -56.25598907470703,
"eval_loss": 1.0565524101257324,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.0757867619395256,
"eval_rewards/margins": -0.0757867619395256,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6514,
"eval_samples_per_second": 9.494,
"eval_steps_per_second": 0.306,
"step": 2760
},
{
"epoch": 15.424581005586592,
"grad_norm": 0.36330710276209954,
"learning_rate": 3.936540006665989e-09,
"logits/chosen": -3.451162815093994,
"logits/rejected": -3.5979788303375244,
"logps/chosen": -3.8159427642822266,
"logps/rejected": -34.103302001953125,
"loss": 0.0981,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.461691379547119,
"rewards/margins": 3.461691379547119,
"rewards/rejected": 0.0,
"step": 2761
},
{
"epoch": 15.430167597765363,
"grad_norm": 0.3461085529795291,
"learning_rate": 3.860571433137116e-09,
"logits/chosen": -3.5518178939819336,
"logits/rejected": -3.4665303230285645,
"logps/chosen": -2.5919435024261475,
"logps/rejected": -103.93636322021484,
"loss": 0.1023,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.152487277984619,
"rewards/margins": 3.152487277984619,
"rewards/rejected": 0.0,
"step": 2762
},
{
"epoch": 15.435754189944134,
"grad_norm": 0.3751277054303707,
"learning_rate": 3.785340211156618e-09,
"logits/chosen": -3.345266580581665,
"logits/rejected": -3.226581335067749,
"logps/chosen": -14.586767196655273,
"logps/rejected": -28.173521041870117,
"loss": 0.065,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4477198123931885,
"rewards/margins": 3.4477198123931885,
"rewards/rejected": 0.0,
"step": 2763
},
{
"epoch": 15.441340782122905,
"grad_norm": 0.417944946479477,
"learning_rate": 3.7108464525316685e-09,
"logits/chosen": -3.285778045654297,
"logits/rejected": -3.213641405105591,
"logps/chosen": -6.269316673278809,
"logps/rejected": -47.36651611328125,
"loss": 0.0772,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.783946990966797,
"rewards/margins": 3.783946990966797,
"rewards/rejected": 0.0,
"step": 2764
},
{
"epoch": 15.446927374301676,
"grad_norm": 0.37137562846821254,
"learning_rate": 3.637090267973042e-09,
"logits/chosen": -3.311185598373413,
"logits/rejected": -3.404200315475464,
"logps/chosen": -13.746042251586914,
"logps/rejected": -54.992279052734375,
"loss": 0.0921,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.807616710662842,
"rewards/margins": 2.807616710662842,
"rewards/rejected": 0.0,
"step": 2765
},
{
"epoch": 15.452513966480447,
"grad_norm": 0.38958315114479514,
"learning_rate": 3.564071767095722e-09,
"logits/chosen": -3.327028274536133,
"logits/rejected": -3.3071060180664062,
"logps/chosen": -0.4666496515274048,
"logps/rejected": -85.70694732666016,
"loss": 0.097,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.8747763633728027,
"rewards/margins": 1.8747763633728027,
"rewards/rejected": 0.0,
"step": 2766
},
{
"epoch": 15.458100558659218,
"grad_norm": 0.40104745850233675,
"learning_rate": 3.491791058418181e-09,
"logits/chosen": -3.304408073425293,
"logits/rejected": -3.363276720046997,
"logps/chosen": -0.6727848052978516,
"logps/rejected": -38.689842224121094,
"loss": 0.1013,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.456422805786133,
"rewards/margins": 2.456422805786133,
"rewards/rejected": 0.0,
"step": 2767
},
{
"epoch": 15.46368715083799,
"grad_norm": 0.4313312495782813,
"learning_rate": 3.4202482493623785e-09,
"logits/chosen": -3.247502088546753,
"logits/rejected": -3.3012685775756836,
"logps/chosen": -1.0190379619598389,
"logps/rejected": -45.1026611328125,
"loss": 0.0934,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6603927612304688,
"rewards/margins": 2.6603927612304688,
"rewards/rejected": 0.0,
"step": 2768
},
{
"epoch": 15.46927374301676,
"grad_norm": 0.3697268473542088,
"learning_rate": 3.349443446253708e-09,
"logits/chosen": -3.3885860443115234,
"logits/rejected": -3.260544776916504,
"logps/chosen": -5.293089866638184,
"logps/rejected": -54.82860565185547,
"loss": 0.0886,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2731010913848877,
"rewards/margins": 3.2731010913848877,
"rewards/rejected": 0.0,
"step": 2769
},
{
"epoch": 15.474860335195531,
"grad_norm": 0.3749107031852132,
"learning_rate": 3.2793767543207172e-09,
"logits/chosen": -3.2207043170928955,
"logits/rejected": -3.2187039852142334,
"logps/chosen": -13.834918975830078,
"logps/rejected": -55.70347595214844,
"loss": 0.0803,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.689237117767334,
"rewards/margins": 3.689237117767334,
"rewards/rejected": 0.0,
"step": 2770
},
{
"epoch": 15.480446927374302,
"grad_norm": 0.35471904615786953,
"learning_rate": 3.2100482776949432e-09,
"logits/chosen": -3.1515934467315674,
"logits/rejected": -3.304704427719116,
"logps/chosen": -0.6870374083518982,
"logps/rejected": -86.122314453125,
"loss": 0.0911,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0924084186553955,
"rewards/margins": 2.0924084186553955,
"rewards/rejected": 0.0,
"step": 2771
},
{
"epoch": 15.486033519553073,
"grad_norm": 0.44887963252520147,
"learning_rate": 3.141458119410856e-09,
"logits/chosen": -3.3256747722625732,
"logits/rejected": -3.313903331756592,
"logps/chosen": -2.753474473953247,
"logps/rejected": -36.37815475463867,
"loss": 0.0956,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6246652603149414,
"rewards/margins": 2.6246652603149414,
"rewards/rejected": 0.0,
"step": 2772
},
{
"epoch": 15.491620111731844,
"grad_norm": 0.4019342901537875,
"learning_rate": 3.073606381405747e-09,
"logits/chosen": -3.468522071838379,
"logits/rejected": -3.572396993637085,
"logps/chosen": -2.670816421508789,
"logps/rejected": -37.25193405151367,
"loss": 0.0834,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2995452880859375,
"rewards/margins": 3.2995452880859375,
"rewards/rejected": 0.0,
"step": 2773
},
{
"epoch": 15.497206703910614,
"grad_norm": 0.4202327403144555,
"learning_rate": 3.0064931645192306e-09,
"logits/chosen": -3.538756847381592,
"logits/rejected": -3.5643651485443115,
"logps/chosen": -3.070742607116699,
"logps/rejected": -66.23321533203125,
"loss": 0.1099,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.683175563812256,
"rewards/margins": 2.683175563812256,
"rewards/rejected": 0.0,
"step": 2774
},
{
"epoch": 15.502793296089386,
"grad_norm": 0.5020722710923473,
"learning_rate": 2.940118568493577e-09,
"logits/chosen": -3.5868358612060547,
"logits/rejected": -3.4845035076141357,
"logps/chosen": -4.515740394592285,
"logps/rejected": -62.495758056640625,
"loss": 0.1083,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8176839351654053,
"rewards/margins": 2.8176839351654053,
"rewards/rejected": 0.0,
"step": 2775
},
{
"epoch": 15.508379888268156,
"grad_norm": 0.4227194900443271,
"learning_rate": 2.8744826919733233e-09,
"logits/chosen": -3.276412010192871,
"logits/rejected": -3.315068006515503,
"logps/chosen": -2.1350624561309814,
"logps/rejected": -52.504066467285156,
"loss": 0.0988,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8215818405151367,
"rewards/margins": 2.8215818405151367,
"rewards/rejected": 0.0,
"step": 2776
},
{
"epoch": 15.513966480446927,
"grad_norm": 0.41264950439068643,
"learning_rate": 2.8095856325049405e-09,
"logits/chosen": -3.0627875328063965,
"logits/rejected": -3.2191262245178223,
"logps/chosen": -0.2128685712814331,
"logps/rejected": -96.5482177734375,
"loss": 0.1141,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.4802885055541992,
"rewards/margins": 1.4802885055541992,
"rewards/rejected": 0.0,
"step": 2777
},
{
"epoch": 15.519553072625698,
"grad_norm": 0.33991140388893726,
"learning_rate": 2.745427486537111e-09,
"logits/chosen": -3.030118942260742,
"logits/rejected": -3.1139426231384277,
"logps/chosen": -2.5844459533691406,
"logps/rejected": -53.24211120605469,
"loss": 0.0962,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.950360059738159,
"rewards/margins": 2.950360059738159,
"rewards/rejected": 0.0,
"step": 2778
},
{
"epoch": 15.525139664804469,
"grad_norm": 0.33174144887954643,
"learning_rate": 2.6820083494201727e-09,
"logits/chosen": -3.4181268215179443,
"logits/rejected": -3.5480806827545166,
"logps/chosen": -0.7553225755691528,
"logps/rejected": -82.75349426269531,
"loss": 0.1136,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4046425819396973,
"rewards/margins": 2.4046425819396973,
"rewards/rejected": 0.0,
"step": 2779
},
{
"epoch": 15.53072625698324,
"grad_norm": 0.36692690904686837,
"learning_rate": 2.6193283154063996e-09,
"logits/chosen": -3.325366735458374,
"logits/rejected": -3.6006627082824707,
"logps/chosen": -1.3810858726501465,
"logps/rejected": -67.34014892578125,
"loss": 0.1223,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4140939712524414,
"rewards/margins": 2.4140939712524414,
"rewards/rejected": 0.0,
"step": 2780
},
{
"epoch": 15.53072625698324,
"eval_logits/chosen": -3.257312297821045,
"eval_logits/rejected": -3.3804821968078613,
"eval_logps/chosen": -30.872472763061523,
"eval_logps/rejected": -56.256752014160156,
"eval_loss": 1.0601108074188232,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.08024321496486664,
"eval_rewards/margins": -0.08024321496486664,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6862,
"eval_samples_per_second": 9.484,
"eval_steps_per_second": 0.306,
"step": 2780
},
{
"epoch": 15.53631284916201,
"grad_norm": 0.4061069101464907,
"learning_rate": 2.557387477649331e-09,
"logits/chosen": -3.199857234954834,
"logits/rejected": -3.4232935905456543,
"logps/chosen": -4.499742031097412,
"logps/rejected": -53.95488739013672,
"loss": 0.0864,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0268375873565674,
"rewards/margins": 3.0268375873565674,
"rewards/rejected": 0.0,
"step": 2781
},
{
"epoch": 15.541899441340782,
"grad_norm": 0.40446462715684434,
"learning_rate": 2.496185928204164e-09,
"logits/chosen": -3.1261484622955322,
"logits/rejected": -3.2383618354797363,
"logps/chosen": -1.6808207035064697,
"logps/rejected": -50.73039245605469,
"loss": 0.0984,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.562302589416504,
"rewards/margins": 2.562302589416504,
"rewards/rejected": 0.0,
"step": 2782
},
{
"epoch": 15.547486033519553,
"grad_norm": 0.3560462108194252,
"learning_rate": 2.4357237580273084e-09,
"logits/chosen": -3.512803316116333,
"logits/rejected": -3.4058237075805664,
"logps/chosen": -1.1344797611236572,
"logps/rejected": -44.242820739746094,
"loss": 0.073,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1338982582092285,
"rewards/margins": 3.1338982582092285,
"rewards/rejected": 0.0,
"step": 2783
},
{
"epoch": 15.553072625698324,
"grad_norm": 0.363655029715933,
"learning_rate": 2.3760010569762734e-09,
"logits/chosen": -3.3123326301574707,
"logits/rejected": -3.3357810974121094,
"logps/chosen": -2.1729488372802734,
"logps/rejected": -70.17076110839844,
"loss": 0.088,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0184102058410645,
"rewards/margins": 3.0184102058410645,
"rewards/rejected": 0.0,
"step": 2784
},
{
"epoch": 15.558659217877095,
"grad_norm": 0.4839223569530984,
"learning_rate": 2.3170179138097266e-09,
"logits/chosen": -3.253693103790283,
"logits/rejected": -3.337338447570801,
"logps/chosen": -1.2565081119537354,
"logps/rejected": -62.19097900390625,
"loss": 0.1098,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.555415153503418,
"rewards/margins": 2.555415153503418,
"rewards/rejected": 0.0,
"step": 2785
},
{
"epoch": 15.564245810055866,
"grad_norm": 0.4104548003522868,
"learning_rate": 2.2587744161869926e-09,
"logits/chosen": -3.2256131172180176,
"logits/rejected": -3.183342218399048,
"logps/chosen": -2.3459343910217285,
"logps/rejected": -71.43930053710938,
"loss": 0.096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9714996814727783,
"rewards/margins": 2.9714996814727783,
"rewards/rejected": 0.0,
"step": 2786
},
{
"epoch": 15.569832402234637,
"grad_norm": 0.6294585426588813,
"learning_rate": 2.2012706506683297e-09,
"logits/chosen": -3.559893846511841,
"logits/rejected": -3.4145257472991943,
"logps/chosen": -2.7881810665130615,
"logps/rejected": -40.22637176513672,
"loss": 0.1127,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.719789743423462,
"rewards/margins": 2.719789743423462,
"rewards/rejected": 0.0,
"step": 2787
},
{
"epoch": 15.575418994413408,
"grad_norm": 0.33294000108246335,
"learning_rate": 2.144506702714599e-09,
"logits/chosen": -3.468883514404297,
"logits/rejected": -3.5237717628479004,
"logps/chosen": -9.07915210723877,
"logps/rejected": -29.209163665771484,
"loss": 0.0941,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.923649549484253,
"rewards/margins": 3.923649549484253,
"rewards/rejected": 0.0,
"step": 2788
},
{
"epoch": 15.581005586592179,
"grad_norm": 0.3837843837695766,
"learning_rate": 2.0884826566871516e-09,
"logits/chosen": -2.6626386642456055,
"logits/rejected": -2.7222583293914795,
"logps/chosen": -19.42976188659668,
"logps/rejected": -48.52910614013672,
"loss": 0.0972,
"rewards/accuracies": 1.0,
"rewards/chosen": 4.182826042175293,
"rewards/margins": 4.182826042175293,
"rewards/rejected": 0.0,
"step": 2789
},
{
"epoch": 15.58659217877095,
"grad_norm": 0.43376266031743754,
"learning_rate": 2.033198595847607e-09,
"logits/chosen": -3.3460216522216797,
"logits/rejected": -3.280305862426758,
"logps/chosen": -7.48615837097168,
"logps/rejected": -48.15699768066406,
"loss": 0.1229,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.4884071350097656,
"rewards/margins": 3.4884071350097656,
"rewards/rejected": 0.0,
"step": 2790
},
{
"epoch": 15.59217877094972,
"grad_norm": 0.43090268920455477,
"learning_rate": 1.978654602358021e-09,
"logits/chosen": -3.3077661991119385,
"logits/rejected": -3.497183084487915,
"logps/chosen": -1.2378193140029907,
"logps/rejected": -66.26832580566406,
"loss": 0.1393,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.811579704284668,
"rewards/margins": 2.811579704284668,
"rewards/rejected": 0.0,
"step": 2791
},
{
"epoch": 15.597765363128492,
"grad_norm": 0.40436075782981956,
"learning_rate": 1.9248507572804383e-09,
"logits/chosen": -3.3102433681488037,
"logits/rejected": -3.268357038497925,
"logps/chosen": -0.8314454555511475,
"logps/rejected": -39.71807098388672,
"loss": 0.0769,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.867051839828491,
"rewards/margins": 2.867051839828491,
"rewards/rejected": 0.0,
"step": 2792
},
{
"epoch": 15.603351955307263,
"grad_norm": 0.44023941554623247,
"learning_rate": 1.8717871405769526e-09,
"logits/chosen": -3.3736393451690674,
"logits/rejected": -3.412764310836792,
"logps/chosen": -4.920578956604004,
"logps/rejected": -51.50208282470703,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.1983420848846436,
"rewards/margins": 3.1983420848846436,
"rewards/rejected": 0.0,
"step": 2793
},
{
"epoch": 15.608938547486034,
"grad_norm": 0.39669380790497527,
"learning_rate": 1.8194638311095912e-09,
"logits/chosen": -3.205962896347046,
"logits/rejected": -3.237531900405884,
"logps/chosen": -0.654625415802002,
"logps/rejected": -56.472381591796875,
"loss": 0.1173,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1814088821411133,
"rewards/margins": 2.1814088821411133,
"rewards/rejected": 0.0,
"step": 2794
},
{
"epoch": 15.614525139664805,
"grad_norm": 0.6864799818663945,
"learning_rate": 1.7678809066400401e-09,
"logits/chosen": -3.3602356910705566,
"logits/rejected": -3.311319351196289,
"logps/chosen": -1.0779731273651123,
"logps/rejected": -65.07955932617188,
"loss": 0.1042,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3188016414642334,
"rewards/margins": 2.3188016414642334,
"rewards/rejected": 0.0,
"step": 2795
},
{
"epoch": 15.620111731843576,
"grad_norm": 0.36891150579566906,
"learning_rate": 1.7170384438297548e-09,
"logits/chosen": -3.42869234085083,
"logits/rejected": -3.382807731628418,
"logps/chosen": -0.31308987736701965,
"logps/rejected": -74.28213500976562,
"loss": 0.118,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.075309991836548,
"rewards/margins": 2.075309991836548,
"rewards/rejected": 0.0,
"step": 2796
},
{
"epoch": 15.625698324022347,
"grad_norm": 0.45196185506431397,
"learning_rate": 1.6669365182396811e-09,
"logits/chosen": -3.568336248397827,
"logits/rejected": -3.7408010959625244,
"logps/chosen": -0.34792613983154297,
"logps/rejected": -44.23535919189453,
"loss": 0.1178,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.9495230913162231,
"rewards/margins": 1.9495230913162231,
"rewards/rejected": 0.0,
"step": 2797
},
{
"epoch": 15.631284916201118,
"grad_norm": 0.5411387809142836,
"learning_rate": 1.617575204330257e-09,
"logits/chosen": -3.2192094326019287,
"logits/rejected": -3.367995262145996,
"logps/chosen": -0.28346142172813416,
"logps/rejected": -46.0729866027832,
"loss": 0.0939,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0429527759552,
"rewards/margins": 2.0429527759552,
"rewards/rejected": 0.0,
"step": 2798
},
{
"epoch": 15.636871508379889,
"grad_norm": 0.3763849368960105,
"learning_rate": 1.5689545754610788e-09,
"logits/chosen": -3.341198205947876,
"logits/rejected": -3.2854690551757812,
"logps/chosen": -3.698657512664795,
"logps/rejected": -36.41625213623047,
"loss": 0.0843,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.817406177520752,
"rewards/margins": 2.817406177520752,
"rewards/rejected": 0.0,
"step": 2799
},
{
"epoch": 15.64245810055866,
"grad_norm": 0.33605716429498145,
"learning_rate": 1.5210747038911232e-09,
"logits/chosen": -3.370471715927124,
"logits/rejected": -3.373816728591919,
"logps/chosen": -4.742188930511475,
"logps/rejected": -40.91890335083008,
"loss": 0.1231,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.6829161643981934,
"rewards/margins": 2.6829161643981934,
"rewards/rejected": 0.0,
"step": 2800
},
{
"epoch": 15.64245810055866,
"eval_logits/chosen": -3.2499423027038574,
"eval_logits/rejected": -3.375277280807495,
"eval_logps/chosen": -30.777721405029297,
"eval_logps/rejected": -56.15149688720703,
"eval_loss": 1.055828332901001,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.07076773047447205,
"eval_rewards/margins": -0.07076773047447205,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7114,
"eval_samples_per_second": 9.477,
"eval_steps_per_second": 0.306,
"step": 2800
},
{
"epoch": 15.64804469273743,
"grad_norm": 0.4077306927781004,
"learning_rate": 1.4739356607784136e-09,
"logits/chosen": -3.1192891597747803,
"logits/rejected": -3.1471028327941895,
"logps/chosen": -4.933080673217773,
"logps/rejected": -155.74085998535156,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.16139554977417,
"rewards/margins": 3.16139554977417,
"rewards/rejected": 0.0,
"step": 2801
},
{
"epoch": 15.653631284916202,
"grad_norm": 0.4373517979374343,
"learning_rate": 1.427537516179911e-09,
"logits/chosen": -3.622830867767334,
"logits/rejected": -3.3402109146118164,
"logps/chosen": -1.7372000217437744,
"logps/rejected": -47.54938507080078,
"loss": 0.0923,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2426764965057373,
"rewards/margins": 2.2426764965057373,
"rewards/rejected": 0.0,
"step": 2802
},
{
"epoch": 15.659217877094973,
"grad_norm": 0.43996780202222885,
"learning_rate": 1.3818803390515666e-09,
"logits/chosen": -3.3221375942230225,
"logits/rejected": -3.4538419246673584,
"logps/chosen": -0.5822171568870544,
"logps/rejected": -50.92298889160156,
"loss": 0.0956,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.3626534938812256,
"rewards/margins": 2.3626534938812256,
"rewards/rejected": 0.0,
"step": 2803
},
{
"epoch": 15.664804469273744,
"grad_norm": 0.31288169138884964,
"learning_rate": 1.3369641972481027e-09,
"logits/chosen": -3.369594097137451,
"logits/rejected": -3.592721939086914,
"logps/chosen": -5.258505821228027,
"logps/rejected": -54.46895980834961,
"loss": 0.0661,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5000388622283936,
"rewards/margins": 3.5000388622283936,
"rewards/rejected": 0.0,
"step": 2804
},
{
"epoch": 15.670391061452515,
"grad_norm": 0.41871178002196785,
"learning_rate": 1.2927891575228445e-09,
"logits/chosen": -3.2368991374969482,
"logits/rejected": -3.3737740516662598,
"logps/chosen": -1.0429718494415283,
"logps/rejected": -34.85597229003906,
"loss": 0.0957,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5941667556762695,
"rewards/margins": 2.5941667556762695,
"rewards/rejected": 0.0,
"step": 2805
},
{
"epoch": 15.675977653631286,
"grad_norm": 0.9885229141249284,
"learning_rate": 1.249355285527831e-09,
"logits/chosen": -3.5666892528533936,
"logits/rejected": -3.6364030838012695,
"logps/chosen": -3.115077018737793,
"logps/rejected": -45.557373046875,
"loss": 0.1067,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.011629104614258,
"rewards/margins": 3.011629104614258,
"rewards/rejected": 0.0,
"step": 2806
},
{
"epoch": 15.681564245810057,
"grad_norm": 0.43451908623452734,
"learning_rate": 1.2066626458134276e-09,
"logits/chosen": -3.215453863143921,
"logits/rejected": -3.174268960952759,
"logps/chosen": -3.840770721435547,
"logps/rejected": -58.07538986206055,
"loss": 0.1218,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.510533094406128,
"rewards/margins": 3.510533094406128,
"rewards/rejected": 0.0,
"step": 2807
},
{
"epoch": 15.687150837988828,
"grad_norm": 0.34577052547749765,
"learning_rate": 1.1647113018286025e-09,
"logits/chosen": -3.3677327632904053,
"logits/rejected": -3.3671536445617676,
"logps/chosen": -0.43579035997390747,
"logps/rejected": -92.73271179199219,
"loss": 0.0921,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.171261787414551,
"rewards/margins": 2.171261787414551,
"rewards/rejected": 0.0,
"step": 2808
},
{
"epoch": 15.692737430167599,
"grad_norm": 0.401057482924194,
"learning_rate": 1.1235013159205386e-09,
"logits/chosen": -3.073883533477783,
"logits/rejected": -3.161151885986328,
"logps/chosen": -4.56620979309082,
"logps/rejected": -38.5812873840332,
"loss": 0.0865,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3170461654663086,
"rewards/margins": 3.3170461654663086,
"rewards/rejected": 0.0,
"step": 2809
},
{
"epoch": 15.69832402234637,
"grad_norm": 0.4264076689718895,
"learning_rate": 1.083032749334467e-09,
"logits/chosen": -3.21286678314209,
"logits/rejected": -3.2060153484344482,
"logps/chosen": -1.5400182008743286,
"logps/rejected": -42.095340728759766,
"loss": 0.1311,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.428891658782959,
"rewards/margins": 2.428891658782959,
"rewards/rejected": 0.0,
"step": 2810
},
{
"epoch": 15.703910614525139,
"grad_norm": 0.42769534207869314,
"learning_rate": 1.0433056622140557e-09,
"logits/chosen": -3.2073304653167725,
"logits/rejected": -3.4061129093170166,
"logps/chosen": -0.41317540407180786,
"logps/rejected": -56.47168731689453,
"loss": 0.1123,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.034308910369873,
"rewards/margins": 2.034308910369873,
"rewards/rejected": 0.0,
"step": 2811
},
{
"epoch": 15.70949720670391,
"grad_norm": 0.36023229614273433,
"learning_rate": 1.0043201136006874e-09,
"logits/chosen": -3.570122003555298,
"logits/rejected": -3.4414169788360596,
"logps/chosen": -2.0092458724975586,
"logps/rejected": -32.25787353515625,
"loss": 0.107,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2174861431121826,
"rewards/margins": 3.2174861431121826,
"rewards/rejected": 0.0,
"step": 2812
},
{
"epoch": 15.71508379888268,
"grad_norm": 0.403769439570388,
"learning_rate": 9.660761614339597e-10,
"logits/chosen": -3.461387872695923,
"logits/rejected": -3.382807731628418,
"logps/chosen": -4.465088844299316,
"logps/rejected": -70.43394470214844,
"loss": 0.0768,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0921897888183594,
"rewards/margins": 3.0921897888183594,
"rewards/rejected": 0.0,
"step": 2813
},
{
"epoch": 15.720670391061452,
"grad_norm": 0.3666172402209137,
"learning_rate": 9.285738625511297e-10,
"logits/chosen": -3.2861690521240234,
"logits/rejected": -3.5039193630218506,
"logps/chosen": -0.2656005024909973,
"logps/rejected": -48.74907302856445,
"loss": 0.1084,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6739084720611572,
"rewards/margins": 1.6739084720611572,
"rewards/rejected": 0.0,
"step": 2814
},
{
"epoch": 15.726256983240223,
"grad_norm": 0.3917052549432524,
"learning_rate": 8.918132726872807e-10,
"logits/chosen": -3.465508222579956,
"logits/rejected": -3.417743444442749,
"logps/chosen": -17.935710906982422,
"logps/rejected": -80.55555725097656,
"loss": 0.0887,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.849966049194336,
"rewards/margins": 3.849966049194336,
"rewards/rejected": 0.0,
"step": 2815
},
{
"epoch": 15.731843575418994,
"grad_norm": 0.32990131461111255,
"learning_rate": 8.557944464752109e-10,
"logits/chosen": -3.2874207496643066,
"logits/rejected": -3.3129971027374268,
"logps/chosen": -0.7452528476715088,
"logps/rejected": -39.76192092895508,
"loss": 0.1339,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.0606377124786377,
"rewards/margins": 2.0606377124786377,
"rewards/rejected": 0.0,
"step": 2816
},
{
"epoch": 15.737430167597765,
"grad_norm": 0.31626882369629983,
"learning_rate": 8.205174374453227e-10,
"logits/chosen": -3.5613372325897217,
"logits/rejected": -3.5797479152679443,
"logps/chosen": -0.7551592588424683,
"logps/rejected": -57.97498321533203,
"loss": 0.102,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.154473304748535,
"rewards/margins": 2.154473304748535,
"rewards/rejected": 0.0,
"step": 2817
},
{
"epoch": 15.743016759776536,
"grad_norm": 0.3616808663572626,
"learning_rate": 7.859822980255115e-10,
"logits/chosen": -3.7419943809509277,
"logits/rejected": -3.503006935119629,
"logps/chosen": -2.766098976135254,
"logps/rejected": -38.50261688232422,
"loss": 0.1045,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9870448112487793,
"rewards/margins": 2.9870448112487793,
"rewards/rejected": 0.0,
"step": 2818
},
{
"epoch": 15.748603351955307,
"grad_norm": 0.3574739958631293,
"learning_rate": 7.521890795411656e-10,
"logits/chosen": -2.97888445854187,
"logits/rejected": -2.923058271408081,
"logps/chosen": -11.004693984985352,
"logps/rejected": -31.374242782592773,
"loss": 0.0848,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3422088623046875,
"rewards/margins": 3.3422088623046875,
"rewards/rejected": 0.0,
"step": 2819
},
{
"epoch": 15.754189944134078,
"grad_norm": 0.3451826393107498,
"learning_rate": 7.191378322150554e-10,
"logits/chosen": -3.349212884902954,
"logits/rejected": -3.506939172744751,
"logps/chosen": -1.2821694612503052,
"logps/rejected": -51.124549865722656,
"loss": 0.1085,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8572986125946045,
"rewards/margins": 2.8572986125946045,
"rewards/rejected": 0.0,
"step": 2820
},
{
"epoch": 15.754189944134078,
"eval_logits/chosen": -3.257516384124756,
"eval_logits/rejected": -3.381110668182373,
"eval_logps/chosen": -30.6396484375,
"eval_logps/rejected": -56.05371856689453,
"eval_loss": 1.049382209777832,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.056960977613925934,
"eval_rewards/margins": -0.056960977613925934,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7293,
"eval_samples_per_second": 9.472,
"eval_steps_per_second": 0.306,
"step": 2820
},
{
"epoch": 15.759776536312849,
"grad_norm": 0.3546071638052034,
"learning_rate": 6.868286051671668e-10,
"logits/chosen": -3.405954360961914,
"logits/rejected": -3.282923698425293,
"logps/chosen": -0.9710592031478882,
"logps/rejected": -85.11328125,
"loss": 0.074,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.37251615524292,
"rewards/margins": 2.37251615524292,
"rewards/rejected": 0.0,
"step": 2821
},
{
"epoch": 15.76536312849162,
"grad_norm": 0.3724010116990325,
"learning_rate": 6.552614464147566e-10,
"logits/chosen": -3.3313539028167725,
"logits/rejected": -3.620511293411255,
"logps/chosen": -29.941688537597656,
"logps/rejected": -38.74200439453125,
"loss": 0.1327,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.5802860260009766,
"rewards/margins": 3.5802860260009766,
"rewards/rejected": 0.0,
"step": 2822
},
{
"epoch": 15.77094972067039,
"grad_norm": 1.129327964416941,
"learning_rate": 6.244364028722971e-10,
"logits/chosen": -3.35436749458313,
"logits/rejected": -3.415541887283325,
"logps/chosen": -11.93348217010498,
"logps/rejected": -32.43596649169922,
"loss": 0.1275,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7170023918151855,
"rewards/margins": 3.7170023918151855,
"rewards/rejected": 0.0,
"step": 2823
},
{
"epoch": 15.776536312849162,
"grad_norm": 0.3934489794562454,
"learning_rate": 5.94353520351254e-10,
"logits/chosen": -3.326505422592163,
"logits/rejected": -3.2996695041656494,
"logps/chosen": -5.983352184295654,
"logps/rejected": -51.20966339111328,
"loss": 0.1183,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2338926792144775,
"rewards/margins": 3.2338926792144775,
"rewards/rejected": 0.0,
"step": 2824
},
{
"epoch": 15.782122905027933,
"grad_norm": 0.3626987993167899,
"learning_rate": 5.65012843560142e-10,
"logits/chosen": -3.15582537651062,
"logits/rejected": -3.2727463245391846,
"logps/chosen": -1.3835318088531494,
"logps/rejected": -82.92355346679688,
"loss": 0.0825,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4493255615234375,
"rewards/margins": 2.4493255615234375,
"rewards/rejected": 0.0,
"step": 2825
},
{
"epoch": 15.787709497206704,
"grad_norm": 0.3539223343213207,
"learning_rate": 5.364144161044693e-10,
"logits/chosen": -3.2896037101745605,
"logits/rejected": -3.143247127532959,
"logps/chosen": -19.809814453125,
"logps/rejected": -46.5184326171875,
"loss": 0.0789,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.245551586151123,
"rewards/margins": 3.245551586151123,
"rewards/rejected": 0.0,
"step": 2826
},
{
"epoch": 15.793296089385475,
"grad_norm": 0.45218234362373877,
"learning_rate": 5.085582804865706e-10,
"logits/chosen": -3.4615471363067627,
"logits/rejected": -3.479534149169922,
"logps/chosen": -7.718883991241455,
"logps/rejected": -53.56597900390625,
"loss": 0.0755,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8744218349456787,
"rewards/margins": 3.8744218349456787,
"rewards/rejected": 0.0,
"step": 2827
},
{
"epoch": 15.798882681564246,
"grad_norm": 0.47733689901707615,
"learning_rate": 4.814444781056082e-10,
"logits/chosen": -3.46455979347229,
"logits/rejected": -3.5759031772613525,
"logps/chosen": -3.0673773288726807,
"logps/rejected": -76.47246551513672,
"loss": 0.1154,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.164743661880493,
"rewards/margins": 3.164743661880493,
"rewards/rejected": 0.0,
"step": 2828
},
{
"epoch": 15.804469273743017,
"grad_norm": 0.49745046031795465,
"learning_rate": 4.550730492575705e-10,
"logits/chosen": -2.965928316116333,
"logits/rejected": -2.9631450176239014,
"logps/chosen": -10.267690658569336,
"logps/rejected": -35.395652770996094,
"loss": 0.1054,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.2025234699249268,
"rewards/margins": 3.2025234699249268,
"rewards/rejected": 0.0,
"step": 2829
},
{
"epoch": 15.810055865921788,
"grad_norm": 0.37343320783515743,
"learning_rate": 4.294440331350513e-10,
"logits/chosen": -3.3837499618530273,
"logits/rejected": -3.3825016021728516,
"logps/chosen": -5.725942611694336,
"logps/rejected": -31.717445373535156,
"loss": 0.0931,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.850558280944824,
"rewards/margins": 3.850558280944824,
"rewards/rejected": 0.0,
"step": 2830
},
{
"epoch": 15.815642458100559,
"grad_norm": 0.6451118237468618,
"learning_rate": 4.0455746782736e-10,
"logits/chosen": -3.4130003452301025,
"logits/rejected": -3.327850103378296,
"logps/chosen": -0.28376999497413635,
"logps/rejected": -112.74542236328125,
"loss": 0.1483,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.6321723461151123,
"rewards/margins": 1.6321723461151123,
"rewards/rejected": 0.0,
"step": 2831
},
{
"epoch": 15.82122905027933,
"grad_norm": 0.9555876968137641,
"learning_rate": 3.8041339032029993e-10,
"logits/chosen": -3.3221280574798584,
"logits/rejected": -3.451664447784424,
"logps/chosen": -47.78090286254883,
"logps/rejected": -55.13854217529297,
"loss": 0.0975,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.7417855262756348,
"rewards/margins": 3.7417855262756348,
"rewards/rejected": 0.0,
"step": 2832
},
{
"epoch": 15.8268156424581,
"grad_norm": 0.33516527586800354,
"learning_rate": 3.5701183649639035e-10,
"logits/chosen": -3.323343276977539,
"logits/rejected": -3.324136257171631,
"logps/chosen": -1.7849292755126953,
"logps/rejected": -73.18279266357422,
"loss": 0.104,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.9350509643554688,
"rewards/margins": 2.9350509643554688,
"rewards/rejected": 0.0,
"step": 2833
},
{
"epoch": 15.832402234636872,
"grad_norm": 0.34990907835476237,
"learning_rate": 3.343528411344221e-10,
"logits/chosen": -3.246032476425171,
"logits/rejected": -3.4260365962982178,
"logps/chosen": -1.6394751071929932,
"logps/rejected": -78.3773422241211,
"loss": 0.0891,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8834900856018066,
"rewards/margins": 2.8834900856018066,
"rewards/rejected": 0.0,
"step": 2834
},
{
"epoch": 15.837988826815643,
"grad_norm": 0.35826203365915993,
"learning_rate": 3.1243643790968e-10,
"logits/chosen": -3.260681390762329,
"logits/rejected": -3.158780813217163,
"logps/chosen": -0.801013708114624,
"logps/rejected": -84.7547607421875,
"loss": 0.0822,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.4451889991760254,
"rewards/margins": 2.4451889991760254,
"rewards/rejected": 0.0,
"step": 2835
},
{
"epoch": 15.843575418994414,
"grad_norm": 0.37958438137663664,
"learning_rate": 2.912626593938872e-10,
"logits/chosen": -3.321927070617676,
"logits/rejected": -3.245786190032959,
"logps/chosen": -3.858067274093628,
"logps/rejected": -69.91818237304688,
"loss": 0.1202,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.448540210723877,
"rewards/margins": 2.448540210723877,
"rewards/rejected": 0.0,
"step": 2836
},
{
"epoch": 15.849162011173185,
"grad_norm": 0.341561244863116,
"learning_rate": 2.70831537055094e-10,
"logits/chosen": -3.281402349472046,
"logits/rejected": -3.4028379917144775,
"logps/chosen": -1.1686630249023438,
"logps/rejected": -61.36640548706055,
"loss": 0.1279,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.2457661628723145,
"rewards/margins": 2.2457661628723145,
"rewards/rejected": 0.0,
"step": 2837
},
{
"epoch": 15.854748603351956,
"grad_norm": 0.46006948038488005,
"learning_rate": 2.511431012575116e-10,
"logits/chosen": -3.2890331745147705,
"logits/rejected": -3.3994855880737305,
"logps/chosen": -0.8558558225631714,
"logps/rejected": -42.05720138549805,
"loss": 0.1221,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.270813465118408,
"rewards/margins": 2.270813465118408,
"rewards/rejected": 0.0,
"step": 2838
},
{
"epoch": 15.860335195530727,
"grad_norm": 0.425827273136935,
"learning_rate": 2.3219738126162292e-10,
"logits/chosen": -3.5399320125579834,
"logits/rejected": -3.692891836166382,
"logps/chosen": -1.2003713846206665,
"logps/rejected": -37.772796630859375,
"loss": 0.1496,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.314761161804199,
"rewards/margins": 2.314761161804199,
"rewards/rejected": 0.0,
"step": 2839
},
{
"epoch": 15.865921787709498,
"grad_norm": 0.5341951624759204,
"learning_rate": 2.1399440522429367e-10,
"logits/chosen": -3.338050603866577,
"logits/rejected": -3.5315940380096436,
"logps/chosen": -1.6954944133758545,
"logps/rejected": -46.38953399658203,
"loss": 0.0907,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8678371906280518,
"rewards/margins": 2.8678371906280518,
"rewards/rejected": 0.0,
"step": 2840
},
{
"epoch": 15.865921787709498,
"eval_logits/chosen": -3.258183240890503,
"eval_logits/rejected": -3.381404161453247,
"eval_logps/chosen": -30.842792510986328,
"eval_logps/rejected": -56.60395431518555,
"eval_loss": 1.0576844215393066,
"eval_rewards/accuracies": 0.5249999761581421,
"eval_rewards/chosen": -0.0772751122713089,
"eval_rewards/margins": -0.0772751122713089,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.6797,
"eval_samples_per_second": 9.486,
"eval_steps_per_second": 0.306,
"step": 2840
},
{
"epoch": 15.871508379888269,
"grad_norm": 0.3903180939096369,
"learning_rate": 1.965342001982173e-10,
"logits/chosen": -3.5267210006713867,
"logits/rejected": -3.6699228286743164,
"logps/chosen": -1.2052972316741943,
"logps/rejected": -32.841796875,
"loss": 0.1008,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.193192481994629,
"rewards/margins": 2.193192481994629,
"rewards/rejected": 0.0,
"step": 2841
},
{
"epoch": 15.87709497206704,
"grad_norm": 0.3768808213732159,
"learning_rate": 1.7981679213247002e-10,
"logits/chosen": -3.294266700744629,
"logits/rejected": -3.1123218536376953,
"logps/chosen": -0.28513163328170776,
"logps/rejected": -139.90008544921875,
"loss": 0.0926,
"rewards/accuracies": 1.0,
"rewards/chosen": 1.780448079109192,
"rewards/margins": 1.780448079109192,
"rewards/rejected": 0.0,
"step": 2842
},
{
"epoch": 15.88268156424581,
"grad_norm": 0.43974030954277454,
"learning_rate": 1.6384220587212227e-10,
"logits/chosen": -3.217982053756714,
"logits/rejected": -3.167692184448242,
"logps/chosen": -1.0253431797027588,
"logps/rejected": -22.315250396728516,
"loss": 0.1162,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5189590454101562,
"rewards/margins": 2.5189590454101562,
"rewards/rejected": 0.0,
"step": 2843
},
{
"epoch": 15.888268156424582,
"grad_norm": 0.36849146650165276,
"learning_rate": 1.4861046515818322e-10,
"logits/chosen": -3.2889318466186523,
"logits/rejected": -3.4570798873901367,
"logps/chosen": -12.28207015991211,
"logps/rejected": -52.328094482421875,
"loss": 0.0668,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6533098220825195,
"rewards/margins": 3.6533098220825195,
"rewards/rejected": 0.0,
"step": 2844
},
{
"epoch": 15.893854748603353,
"grad_norm": 0.39614195727997487,
"learning_rate": 1.3412159262771173e-10,
"logits/chosen": -3.5022990703582764,
"logits/rejected": -3.5626304149627686,
"logps/chosen": -2.2584948539733887,
"logps/rejected": -39.102142333984375,
"loss": 0.0891,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5635828971862793,
"rewards/margins": 2.5635828971862793,
"rewards/rejected": 0.0,
"step": 2845
},
{
"epoch": 15.899441340782122,
"grad_norm": 0.3592022131781414,
"learning_rate": 1.2037560981381644e-10,
"logits/chosen": -3.4192519187927246,
"logits/rejected": -3.5048091411590576,
"logps/chosen": -4.079451084136963,
"logps/rejected": -79.90035247802734,
"loss": 0.1145,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5213623046875,
"rewards/margins": 2.5213623046875,
"rewards/rejected": 0.0,
"step": 2846
},
{
"epoch": 15.905027932960895,
"grad_norm": 0.3622297963349978,
"learning_rate": 1.0737253714548922e-10,
"logits/chosen": -3.217479944229126,
"logits/rejected": -3.2967512607574463,
"logps/chosen": -2.382681369781494,
"logps/rejected": -67.23530578613281,
"loss": 0.0973,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.307535171508789,
"rewards/margins": 3.307535171508789,
"rewards/rejected": 0.0,
"step": 2847
},
{
"epoch": 15.910614525139664,
"grad_norm": 0.4541421069651796,
"learning_rate": 9.511239394754956e-11,
"logits/chosen": -3.268536329269409,
"logits/rejected": -3.3430495262145996,
"logps/chosen": -15.523447036743164,
"logps/rejected": -48.1860466003418,
"loss": 0.0946,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9595518112182617,
"rewards/margins": 3.9595518112182617,
"rewards/rejected": 0.0,
"step": 2848
},
{
"epoch": 15.916201117318435,
"grad_norm": 0.373783789412415,
"learning_rate": 8.35951984408112e-11,
"logits/chosen": -3.253858804702759,
"logits/rejected": -3.187084674835205,
"logps/chosen": -1.0501309633255005,
"logps/rejected": -67.25447082519531,
"loss": 0.1169,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7255032062530518,
"rewards/margins": 2.7255032062530518,
"rewards/rejected": 0.0,
"step": 2849
},
{
"epoch": 15.921787709497206,
"grad_norm": 0.3555827029802541,
"learning_rate": 7.282096774180457e-11,
"logits/chosen": -3.4511032104492188,
"logits/rejected": -3.380977153778076,
"logps/chosen": -4.205111026763916,
"logps/rejected": -40.449798583984375,
"loss": 0.0988,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.0369317531585693,
"rewards/margins": 3.0369317531585693,
"rewards/rejected": 0.0,
"step": 2850
},
{
"epoch": 15.927374301675977,
"grad_norm": 0.32413343270208034,
"learning_rate": 6.278971786305432e-11,
"logits/chosen": -3.3584396839141846,
"logits/rejected": -3.4900057315826416,
"logps/chosen": -5.657356262207031,
"logps/rejected": -52.383033752441406,
"loss": 0.1151,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.128657341003418,
"rewards/margins": 3.128657341003418,
"rewards/rejected": 0.0,
"step": 2851
},
{
"epoch": 15.932960893854748,
"grad_norm": 0.3237755223361088,
"learning_rate": 5.3501463712690706e-11,
"logits/chosen": -3.318645715713501,
"logits/rejected": -3.4634900093078613,
"logps/chosen": -9.837621688842773,
"logps/rejected": -29.679607391357422,
"loss": 0.0968,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.9674246311187744,
"rewards/margins": 3.9674246311187744,
"rewards/rejected": 0.0,
"step": 2852
},
{
"epoch": 15.938547486033519,
"grad_norm": 0.45349698875495575,
"learning_rate": 4.4956219094782756e-11,
"logits/chosen": -3.209752321243286,
"logits/rejected": -3.299441337585449,
"logps/chosen": -7.105777740478516,
"logps/rejected": -48.06958770751953,
"loss": 0.1182,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.6124625205993652,
"rewards/margins": 3.6124625205993652,
"rewards/rejected": 0.0,
"step": 2853
},
{
"epoch": 15.94413407821229,
"grad_norm": 0.3385902059748917,
"learning_rate": 3.71539967090051e-11,
"logits/chosen": -3.5183067321777344,
"logits/rejected": -3.205066680908203,
"logps/chosen": -0.8646174669265747,
"logps/rejected": -47.16189956665039,
"loss": 0.0923,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.7576231956481934,
"rewards/margins": 2.7576231956481934,
"rewards/rejected": 0.0,
"step": 2854
},
{
"epoch": 15.949720670391061,
"grad_norm": 0.329009735142087,
"learning_rate": 3.009480815091559e-11,
"logits/chosen": -3.21635365486145,
"logits/rejected": -3.232184410095215,
"logps/chosen": -0.7088548541069031,
"logps/rejected": -44.514739990234375,
"loss": 0.0812,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5989913940429688,
"rewards/margins": 2.5989913940429688,
"rewards/rejected": 0.0,
"step": 2855
},
{
"epoch": 15.955307262569832,
"grad_norm": 0.4543984012760132,
"learning_rate": 2.377866391173322e-11,
"logits/chosen": -3.3538575172424316,
"logits/rejected": -3.369420051574707,
"logps/chosen": -0.6714050769805908,
"logps/rejected": -62.79505157470703,
"loss": 0.1132,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.1580557823181152,
"rewards/margins": 2.1580557823181152,
"rewards/rejected": 0.0,
"step": 2856
},
{
"epoch": 15.960893854748603,
"grad_norm": 0.3553760966999862,
"learning_rate": 1.8205573378338167e-11,
"logits/chosen": -3.3686656951904297,
"logits/rejected": -3.355982542037964,
"logps/chosen": -1.023118019104004,
"logps/rejected": -38.32646942138672,
"loss": 0.1175,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.8010873794555664,
"rewards/margins": 2.8010873794555664,
"rewards/rejected": 0.0,
"step": 2857
},
{
"epoch": 15.966480446927374,
"grad_norm": 0.6577935053933072,
"learning_rate": 1.3375544833382769e-11,
"logits/chosen": -3.499521017074585,
"logits/rejected": -3.325239419937134,
"logps/chosen": -2.625056743621826,
"logps/rejected": -35.65730667114258,
"loss": 0.0853,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.995539665222168,
"rewards/margins": 2.995539665222168,
"rewards/rejected": 0.0,
"step": 2858
},
{
"epoch": 15.972067039106145,
"grad_norm": 0.39916437887814754,
"learning_rate": 9.288585455069498e-12,
"logits/chosen": -3.4343879222869873,
"logits/rejected": -3.238798141479492,
"logps/chosen": -1.6715790033340454,
"logps/rejected": -28.844999313354492,
"loss": 0.1096,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.5408287048339844,
"rewards/margins": 2.5408287048339844,
"rewards/rejected": 0.0,
"step": 2859
},
{
"epoch": 15.977653631284916,
"grad_norm": 0.4465943684246888,
"learning_rate": 5.944701317428524e-12,
"logits/chosen": -3.165088415145874,
"logits/rejected": -3.23022723197937,
"logps/chosen": -1.974366545677185,
"logps/rejected": -49.10314178466797,
"loss": 0.1089,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.056807041168213,
"rewards/margins": 2.056807041168213,
"rewards/rejected": 0.0,
"step": 2860
},
{
"epoch": 15.977653631284916,
"eval_logits/chosen": -3.2575747966766357,
"eval_logits/rejected": -3.3810267448425293,
"eval_logps/chosen": -30.767574310302734,
"eval_logps/rejected": -56.184364318847656,
"eval_loss": 1.0558310747146606,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.06975345313549042,
"eval_rewards/margins": -0.06975345313549042,
"eval_rewards/rejected": 0.0,
"eval_runtime": 32.7389,
"eval_samples_per_second": 9.469,
"eval_steps_per_second": 0.305,
"step": 2860
},
{
"epoch": 15.983240223463687,
"grad_norm": 0.67981272571839,
"learning_rate": 3.34389738998464e-12,
"logits/chosen": -3.2139832973480225,
"logits/rejected": -3.546630382537842,
"logps/chosen": -6.066982269287109,
"logps/rejected": -87.87742614746094,
"loss": 0.1009,
"rewards/accuracies": 1.0,
"rewards/chosen": 2.997310161590576,
"rewards/margins": 2.997310161590576,
"rewards/rejected": 0.0,
"step": 2861
},
{
"epoch": 15.988826815642458,
"grad_norm": 0.3724043214672477,
"learning_rate": 1.4861775380903275e-12,
"logits/chosen": -3.2743287086486816,
"logits/rejected": -3.3364179134368896,
"logps/chosen": -11.646625518798828,
"logps/rejected": -46.14485168457031,
"loss": 0.0892,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.116117477416992,
"rewards/margins": 3.116117477416992,
"rewards/rejected": 0.0,
"step": 2862
},
{
"epoch": 15.994413407821229,
"grad_norm": 0.3525394612067773,
"learning_rate": 3.715445225371816e-13,
"logits/chosen": -3.5659289360046387,
"logits/rejected": -3.6955089569091797,
"logps/chosen": -34.07235336303711,
"logps/rejected": -28.369863510131836,
"loss": 0.1055,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.8572776317596436,
"rewards/margins": 3.8572776317596436,
"rewards/rejected": 0.0,
"step": 2863
},
{
"epoch": 16.0,
"grad_norm": 0.41363535239067295,
"learning_rate": 0.0,
"logits/chosen": -3.784046173095703,
"logits/rejected": -3.575810670852661,
"logps/chosen": -5.2440080642700195,
"logps/rejected": -48.66508483886719,
"loss": 0.0748,
"rewards/accuracies": 1.0,
"rewards/chosen": 3.3392105102539062,
"rewards/margins": 3.3392105102539062,
"rewards/rejected": 0.0,
"step": 2864
},
{
"epoch": 16.0,
"step": 2864,
"total_flos": 0.0,
"train_loss": 0.0,
"train_runtime": 0.0303,
"train_samples_per_second": 3013918.631,
"train_steps_per_second": 94647.62
}
],
"logging_steps": 1,
"max_steps": 2864,
"num_input_tokens_seen": 0,
"num_train_epochs": 16,
"save_steps": 200,
"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": 4,
"trial_name": null,
"trial_params": null
}
Reference in New Issue View Git Blame Copy Permalink