taozi555/MN-12B-Mag-Mell-R1-KTO
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity
Files
d167527f563b4aa86375eb2d9fcd6c5dc64b4edf
MN-12B-Mag-Mell-R1-KTO/trainer_state.json

8954 lines
286 KiB
JSON
Raw Normal View History

初始化项目，由ModelHub XC社区提供模型 Model: taozi555/MN-12B-Mag-Mell-R1-KTO Source: Original Platform
2026-07-20 23:46:10 +08:00
{
"best_global_step": null,
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 594,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0016842105263157896,
"grad_norm": 756.2248967148687,
"kl": 0.0,
"learning_rate": 0.0,
"logits/chosen": -1430707968.0,
"logits/rejected": -627221650.2857143,
"logps/chosen": -274.7185974121094,
"logps/rejected": -244.01890345982142,
"loss": 0.5,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.003368421052631579,
"grad_norm": 127.71874425517983,
"kl": 0.0,
"learning_rate": 8.333333333333334e-09,
"logits/chosen": -625243200.0,
"logits/rejected": -765075712.0,
"logps/chosen": -214.222900390625,
"logps/rejected": -240.65220642089844,
"loss": 0.5,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 2
},
{
"epoch": 0.0050526315789473685,
"grad_norm": 181.90697566158104,
"kl": 0.4360809326171875,
"learning_rate": 1.6666666666666667e-08,
"logits/chosen": -1081217792.0,
"logits/rejected": -543856768.0,
"logps/chosen": -203.947265625,
"logps/rejected": -286.96112060546875,
"loss": 0.4983,
"rewards/chosen": -0.00811920315027237,
"rewards/margins": -0.013331988826394081,
"rewards/rejected": 0.005212785676121712,
"step": 3
},
{
"epoch": 0.006736842105263158,
"grad_norm": 157.50669772082293,
"kl": 0.939788818359375,
"learning_rate": 2.5e-08,
"logits/chosen": -665212416.0,
"logits/rejected": -989478912.0,
"logps/chosen": -212.04036458333334,
"logps/rejected": -268.3646484375,
"loss": 0.4997,
"rewards/chosen": -0.022487640380859375,
"rewards/margins": -0.013136444240808487,
"rewards/rejected": -0.009351196140050888,
"step": 4
},
{
"epoch": 0.008421052631578947,
"grad_norm": 323.0915477325882,
"kl": 0.5852317810058594,
"learning_rate": 3.3333333333333334e-08,
"logits/chosen": -1499776170.6666667,
"logits/rejected": -823883468.8,
"logps/chosen": -206.73421223958334,
"logps/rejected": -249.1157470703125,
"loss": 0.4999,
"rewards/chosen": 0.03809356689453125,
"rewards/margins": 0.03527191206812859,
"rewards/rejected": 0.002821654826402664,
"step": 5
},
{
"epoch": 0.010105263157894737,
"grad_norm": 985.1109364386158,
"kl": 1.2429885864257812,
"learning_rate": 4.166666666666666e-08,
"logits/chosen": -830737578.6666666,
"logits/rejected": -1135501516.8,
"logps/chosen": -179.890869140625,
"logps/rejected": -192.8841796875,
"loss": 0.499,
"rewards/chosen": 0.012117513765891394,
"rewards/margins": -0.012128396580616633,
"rewards/rejected": 0.024245910346508026,
"step": 6
},
{
"epoch": 0.011789473684210527,
"grad_norm": 122.52722890356792,
"kl": 2.1596755981445312,
"learning_rate": 5e-08,
"logits/chosen": -917885610.6666666,
"logits/rejected": -1120814080.0,
"logps/chosen": -223.6461181640625,
"logps/rejected": -218.6181396484375,
"loss": 0.5027,
"rewards/chosen": -0.0010091145522892475,
"rewards/margins": -0.017569171357899906,
"rewards/rejected": 0.016560056805610658,
"step": 7
},
{
"epoch": 0.013473684210526317,
"grad_norm": 95.64200187206991,
"kl": 1.8461112976074219,
"learning_rate": 5.833333333333333e-08,
"logits/chosen": -963475648.0,
"logits/rejected": -777407829.3333334,
"logps/chosen": -194.63409423828125,
"logps/rejected": -209.066162109375,
"loss": 0.4996,
"rewards/chosen": 0.036714937537908554,
"rewards/margins": 0.05407130097349484,
"rewards/rejected": -0.017356363435586292,
"step": 8
},
{
"epoch": 0.015157894736842105,
"grad_norm": 230.4658286109559,
"kl": 0.5257720947265625,
"learning_rate": 6.666666666666667e-08,
"logits/chosen": -474723840.0,
"logits/rejected": -627376426.6666666,
"logps/chosen": -223.980908203125,
"logps/rejected": -260.8198649088542,
"loss": 0.4984,
"rewards/chosen": 0.005419158190488815,
"rewards/margins": 0.021833036839962006,
"rewards/rejected": -0.01641387864947319,
"step": 9
},
{
"epoch": 0.016842105263157894,
"grad_norm": 144.31561283203507,
"kl": 2.7238082885742188,
"learning_rate": 7.5e-08,
"logits/chosen": -834623385.6,
"logits/rejected": -1033011200.0,
"logps/chosen": -257.28134765625,
"logps/rejected": -233.40643310546875,
"loss": 0.502,
"rewards/chosen": 0.04759369194507599,
"rewards/margins": 0.024074305097262067,
"rewards/rejected": 0.023519386847813923,
"step": 10
},
{
"epoch": 0.018526315789473686,
"grad_norm": 246.23166775055776,
"kl": 0.989532470703125,
"learning_rate": 8.333333333333333e-08,
"logits/chosen": -933497548.8,
"logits/rejected": -1285298858.6666667,
"logps/chosen": -221.352392578125,
"logps/rejected": -204.99224853515625,
"loss": 0.4965,
"rewards/chosen": 0.0369128406047821,
"rewards/margins": 0.07419260541598002,
"rewards/rejected": -0.037279764811197914,
"step": 11
},
{
"epoch": 0.020210526315789474,
"grad_norm": 173.9408541060697,
"kl": 0.8486251831054688,
"learning_rate": 9.166666666666665e-08,
"logits/chosen": -960234432.0,
"logits/rejected": -981980800.0,
"logps/chosen": -243.4825439453125,
"logps/rejected": -170.26206970214844,
"loss": 0.5027,
"rewards/chosen": -0.004507828503847122,
"rewards/margins": -0.03447256237268448,
"rewards/rejected": 0.029964733868837357,
"step": 12
},
{
"epoch": 0.021894736842105262,
"grad_norm": 175.56996418245757,
"kl": 0.223480224609375,
"learning_rate": 1e-07,
"logits/chosen": -1123834752.0,
"logits/rejected": -974051264.0,
"logps/chosen": -217.30453491210938,
"logps/rejected": -285.2402038574219,
"loss": 0.498,
"rewards/chosen": 0.011188508942723274,
"rewards/margins": 0.04091682657599449,
"rewards/rejected": -0.029728317633271217,
"step": 13
},
{
"epoch": 0.023578947368421053,
"grad_norm": 109.85035410014211,
"kl": 1.2569313049316406,
"learning_rate": 1.0833333333333334e-07,
"logits/chosen": -752794521.6,
"logits/rejected": -981168981.3333334,
"logps/chosen": -259.7815673828125,
"logps/rejected": -145.32327270507812,
"loss": 0.4979,
"rewards/chosen": 0.027963563799858093,
"rewards/margins": 0.05449656397104263,
"rewards/rejected": -0.02653300017118454,
"step": 14
},
{
"epoch": 0.02526315789473684,
"grad_norm": 112.13711806138755,
"kl": 0.5673065185546875,
"learning_rate": 1.1666666666666667e-07,
"logits/chosen": -726509824.0,
"logits/rejected": -840169088.0,
"logps/chosen": -206.02989196777344,
"logps/rejected": -283.2633972167969,
"loss": 0.4991,
"rewards/chosen": 0.005207442678511143,
"rewards/margins": 0.007555770222097635,
"rewards/rejected": -0.0023483275435864925,
"step": 15
},
{
"epoch": 0.026947368421052633,
"grad_norm": 103.03187365170976,
"kl": 0.21651458740234375,
"learning_rate": 1.25e-07,
"logits/chosen": -1134580838.4,
"logits/rejected": -816441002.6666666,
"logps/chosen": -255.2339111328125,
"logps/rejected": -226.8826904296875,
"loss": 0.5015,
"rewards/chosen": -0.0194052129983902,
"rewards/margins": -0.014941000193357468,
"rewards/rejected": -0.00446421280503273,
"step": 16
},
{
"epoch": 0.02863157894736842,
"grad_norm": 775.9066138659358,
"kl": 1.3112754821777344,
"learning_rate": 1.3333333333333334e-07,
"logits/chosen": -1642847744.0,
"logits/rejected": -1082994346.6666667,
"logps/chosen": -223.9878692626953,
"logps/rejected": -277.746337890625,
"loss": 0.4999,
"rewards/chosen": 0.01275329664349556,
"rewards/margins": 0.02697169159849485,
"rewards/rejected": -0.014218394954999289,
"step": 17
},
{
"epoch": 0.03031578947368421,
"grad_norm": 168.9285804872629,
"kl": 0.6275863647460938,
"learning_rate": 1.4166666666666665e-07,
"logits/chosen": -667281792.0,
"logits/rejected": -773589248.0,
"logps/chosen": -258.7420654296875,
"logps/rejected": -229.06549072265625,
"loss": 0.5024,
"rewards/chosen": 8.049095049500465e-05,
"rewards/margins": -0.011579894926398993,
"rewards/rejected": 0.011660385876893997,
"step": 18
},
{
"epoch": 0.032,
"grad_norm": 96.35367098860345,
"kl": 0.8299026489257812,
"learning_rate": 1.5e-07,
"logits/chosen": -1054523520.0,
"logits/rejected": -1119684096.0,
"logps/chosen": -227.5456085205078,
"logps/rejected": -175.5878448486328,
"loss": 0.5028,
"rewards/chosen": -0.006691360846161842,
"rewards/margins": -0.007221984677016735,
"rewards/rejected": 0.0005306238308548927,
"step": 19
},
{
"epoch": 0.03368421052631579,
"grad_norm": 212.06858439091923,
"kl": 0.49672698974609375,
"learning_rate": 1.583333333333333e-07,
"logits/chosen": -1220449484.8,
"logits/rejected": -1335795029.3333333,
"logps/chosen": -192.7220947265625,
"logps/rejected": -213.80257161458334,
"loss": 0.4997,
"rewards/chosen": 0.028179934620857237,
"rewards/margins": 0.10689647098382313,
"rewards/rejected": -0.0787165363629659,
"step": 20
},
{
"epoch": 0.03536842105263158,
"grad_norm": 142.05363612622702,
"kl": 0.5618019104003906,
"learning_rate": 1.6666666666666665e-07,
"logits/chosen": -865751616.0,
"logits/rejected": -749211562.6666666,
"logps/chosen": -274.3310852050781,
"logps/rejected": -161.32567342122397,
"loss": 0.5039,
"rewards/chosen": 0.006252288818359375,
"rewards/margins": -0.04492963353792826,
"rewards/rejected": 0.051181922356287636,
"step": 21
},
{
"epoch": 0.03705263157894737,
"grad_norm": 202.39733674967437,
"kl": 2.1550140380859375,
"learning_rate": 1.75e-07,
"logits/chosen": -852115285.3333334,
"logits/rejected": -1103316787.2,
"logps/chosen": -221.8100382486979,
"logps/rejected": -217.199462890625,
"loss": 0.4971,
"rewards/chosen": 0.03007812798023224,
"rewards/margins": 0.013606876134872437,
"rewards/rejected": 0.016471251845359802,
"step": 22
},
{
"epoch": 0.03873684210526316,
"grad_norm": 205.96536510497384,
"kl": 0.9358940124511719,
"learning_rate": 1.833333333333333e-07,
"logits/chosen": -1277413580.8,
"logits/rejected": -971898112.0,
"logps/chosen": -217.6686767578125,
"logps/rejected": -276.12188720703125,
"loss": 0.5003,
"rewards/chosen": 0.022262269258499147,
"rewards/margins": 0.0022859687606493657,
"rewards/rejected": 0.01997630049784978,
"step": 23
},
{
"epoch": 0.04042105263157895,
"grad_norm": 97.86200798922602,
"kl": 1.6557350158691406,
"learning_rate": 1.9166666666666668e-07,
"logits/chosen": -837048012.8,
"logits/rejected": -1441815552.0,
"logps/chosen": -212.37763671875,
"logps/rejected": -136.08146158854166,
"loss": 0.5045,
"rewards/chosen": -0.0468679815530777,
"rewards/margins": -0.06955506652593613,
"rewards/rejected": 0.02268708497285843,
"step": 24
},
{
"epoch": 0.042105263157894736,
"grad_norm": 146.9171587034632,
"kl": 1.9702682495117188,
"learning_rate": 2e-07,
"logits/chosen": -692624256.0,
"logits/rejected": -764641536.0,
"logps/chosen": -240.31827799479166,
"logps/rejected": -236.6541015625,
"loss": 0.5038,
"rewards/chosen": -0.0033594767252604165,
"rewards/margins": 0.008408406128485998,
"rewards/rejected": -0.011767882853746414,
"step": 25
},
{
"epoch": 0.043789473684210524,
"grad_norm": 372.015222658044,
"kl": 2.9479598999023438,
"learning_rate": 2.0833333333333333e-07,
"logits/chosen": -836751744.0,
"logits/rejected": -1142913280.0,
"logps/chosen": -212.76544189453125,
"logps/rejected": -166.8655548095703,
"loss": 0.4997,
"rewards/chosen": -0.01788177713751793,
"rewards/margins": -0.06701736897230148,
"rewards/rejected": 0.049135591834783554,
"step": 26
},
{
"epoch": 0.04547368421052632,
"grad_norm": 173.67968690260923,
"kl": 0.8008079528808594,
"learning_rate": 2.1666666666666667e-07,
"logits/chosen": -796426240.0,
"logits/rejected": -697571200.0,
"logps/chosen": -200.24410400390624,
"logps/rejected": -241.0162556966146,
"loss": 0.4962,
"rewards/chosen": 0.07569732666015624,
"rewards/margins": 0.07425994885464508,
"rewards/rejected": 0.0014373778055111568,
"step": 27
},
{
"epoch": 0.04715789473684211,
"grad_norm": 179.48473623203154,
"kl": 1.4168357849121094,
"learning_rate": 2.25e-07,
"logits/chosen": -744388736.0,
"logits/rejected": -863063040.0,
"logps/chosen": -290.7626953125,
"logps/rejected": -231.60797119140625,
"loss": 0.504,
"rewards/chosen": -0.019368745386600494,
"rewards/margins": -0.058126453310251236,
"rewards/rejected": 0.03875770792365074,
"step": 28
},
{
"epoch": 0.048842105263157895,
"grad_norm": 130.22057137941954,
"kl": 1.3458938598632812,
"learning_rate": 2.3333333333333333e-07,
"logits/chosen": -691689472.0,
"logits/rejected": -918764885.3333334,
"logps/chosen": -209.3045654296875,
"logps/rejected": -241.4656982421875,
"loss": 0.4978,
"rewards/chosen": 0.07106705009937286,
"rewards/margins": 0.06067454318205515,
"rewards/rejected": 0.010392506917317709,
"step": 29
},
{
"epoch": 0.05052631578947368,
"grad_norm": 130.9807773433376,
"kl": 0.4033660888671875,
"learning_rate": 2.4166666666666665e-07,
"logits/chosen": -920996224.0,
"logits/rejected": -841631424.0,
"logps/chosen": -190.30531311035156,
"logps/rejected": -162.928466796875,
"loss": 0.4982,
"rewards/chosen": 0.035247232764959335,
"rewards/margins": 0.06811809912323952,
"rewards/rejected": -0.03287086635828018,
"step": 30
},
{
"epoch": 0.05221052631578947,
"grad_norm": 76.30157221981797,
"kl": 0.47165679931640625,
"learning_rate": 2.5e-07,
"logits/chosen": -1194017920.0,
"logits/rejected": -877919317.3333334,
"logps/chosen": -306.70257568359375,
"logps/rejected": -252.0715535481771,
"loss": 0.4996,
"rewards/chosen": 0.04151611402630806,
"rewards/margins": 0.03829981635014216,
"rewards/rejected": 0.003216297676165899,
"step": 31
},
{
"epoch": 0.053894736842105266,
"grad_norm": 102.32572244854322,
"kl": 2.0101585388183594,
"learning_rate": 2.5833333333333333e-07,
"logits/chosen": -944446336.0,
"logits/rejected": -1106841088.0,
"logps/chosen": -199.4850311279297,
"logps/rejected": -183.14755249023438,
"loss": 0.5005,
"rewards/chosen": 0.046735379844903946,
"rewards/margins": -0.004496384412050247,
"rewards/rejected": 0.05123176425695419,
"step": 32
},
{
"epoch": 0.055578947368421054,
"grad_norm": 99.04470417971368,
"kl": 2.4130401611328125,
"learning_rate": 2.6666666666666667e-07,
"logits/chosen": -825663078.4,
"logits/rejected": -991676245.3333334,
"logps/chosen": -220.846435546875,
"logps/rejected": -356.6270751953125,
"loss": 0.4947,
"rewards/chosen": 0.052228087186813356,
"rewards/margins": 0.03795552278558413,
"rewards/rejected": 0.014272564401229223,
"step": 33
},
{
"epoch": 0.05726315789473684,
"grad_norm": 246.96656404040365,
"kl": 1.233936071395874,
"learning_rate": 2.75e-07,
"logits/chosen": -1095287296.0,
"logits/rejected": -1060861504.0,
"logps/chosen": -182.49984741210938,
"logps/rejected": -234.69015502929688,
"loss": 0.4936,
"rewards/chosen": 0.06470909714698792,
"rewards/margins": 0.04259873181581497,
"rewards/rejected": 0.022110365331172943,
"step": 34
},
{
"epoch": 0.05894736842105263,
"grad_norm": 152.73517140204396,
"kl": 2.7289276123046875,
"learning_rate": 2.833333333333333e-07,
"logits/chosen": -1097634133.3333333,
"logits/rejected": -1259796582.4,
"logps/chosen": -202.9609375,
"logps/rejected": -189.42052001953124,
"loss": 0.4999,
"rewards/chosen": 0.01765136917432149,
"rewards/margins": -0.034761963287989306,
"rewards/rejected": 0.052413332462310794,
"step": 35
},
{
"epoch": 0.06063157894736842,
"grad_norm": 215.293974300091,
"kl": 2.45306396484375,
"learning_rate": 2.916666666666667e-07,
"logits/chosen": -631062869.3333334,
"logits/rejected": -987482112.0,
"logps/chosen": -213.41678873697916,
"logps/rejected": -221.72783203125,
"loss": 0.5006,
"rewards/chosen": 0.04089965671300888,
"rewards/margins": -0.024423225224018102,
"rewards/rejected": 0.06532288193702698,
"step": 36
},
{
"epoch": 0.06231578947368421,
"grad_norm": 195.90220199948243,
"kl": 2.9997177124023438,
"learning_rate": 3e-07,
"logits/chosen": -848359219.2,
"logits/rejected": -1426793557.3333333,
"logps/chosen": -195.2117431640625,
"logps/rejected": -180.9798787434896,
"loss": 0.4993,
"rewards/chosen": 0.08168579339981079,
"rewards/margins": 0.04379872928063074,
"rewards/rejected": 0.037887064119180046,
"step": 37
},
{
"epoch": 0.064,
"grad_norm": 121.58903894060508,
"kl": 2.0522327423095703,
"learning_rate": 3.0833333333333333e-07,
"logits/chosen": -1282800435.2,
"logits/rejected": -435072682.6666667,
"logps/chosen": -226.767138671875,
"logps/rejected": -249.10555013020834,
"loss": 0.4959,
"rewards/chosen": 0.11044464111328126,
"rewards/margins": 0.0951441449423631,
"rewards/rejected": 0.015300496170918146,
"step": 38
},
{
"epoch": 0.06568421052631579,
"grad_norm": 116.69257759601491,
"kl": 2.5439231395721436,
"learning_rate": 3.166666666666666e-07,
"logits/chosen": -557800192.0,
"logits/rejected": -930213376.0,
"logps/chosen": -211.4619140625,
"logps/rejected": -234.09031677246094,
"loss": 0.4983,
"rewards/chosen": 0.09951868653297424,
"rewards/margins": 0.06241312623023987,
"rewards/rejected": 0.037105560302734375,
"step": 39
},
{
"epoch": 0.06736842105263158,
"grad_norm": 365.02933086523586,
"kl": 2.2659378051757812,
"learning_rate": 3.25e-07,
"logits/chosen": -965182668.8,
"logits/rejected": -642834560.0,
"logps/chosen": -220.280224609375,
"logps/rejected": -222.31856282552084,
"loss": 0.4953,
"rewards/chosen": 0.09896668195724487,
"rewards/margins": 0.024154870708783457,
"rewards/rejected": 0.07481181124846141,
"step": 40
},
{
"epoch": 0.06905263157894737,
"grad_norm": 582.132852175163,
"kl": 2.304011344909668,
"learning_rate": 3.333333333333333e-07,
"logits/chosen": -885514444.8,
"logits/rejected": -748998314.6666666,
"logps/chosen": -227.1810302734375,
"logps/rejected": -196.22355143229166,
"loss": 0.4766,
"rewards/chosen": 0.18464019298553466,
"rewards/margins": 0.5814312537511189,
"rewards/rejected": -0.3967910607655843,
"step": 41
},
{
"epoch": 0.07073684210526315,
"grad_norm": 168.55393612476047,
"kl": 2.7851181030273438,
"learning_rate": 3.4166666666666664e-07,
"logits/chosen": -1012587946.6666666,
"logits/rejected": -900292300.8,
"logps/chosen": -227.2638956705729,
"logps/rejected": -187.71868896484375,
"loss": 0.4928,
"rewards/chosen": 0.1678065061569214,
"rewards/margins": 0.08132548332214355,
"rewards/rejected": 0.08648102283477783,
"step": 42
},
{
"epoch": 0.07242105263157894,
"grad_norm": 236.97744686773765,
"kl": 3.527801513671875,
"learning_rate": 3.5e-07,
"logits/chosen": -970476885.3333334,
"logits/rejected": -398392224.0,
"logps/chosen": -196.3362019856771,
"logps/rejected": -353.429443359375,
"loss": 0.4969,
"rewards/chosen": 0.08341166377067566,
"rewards/margins": 0.04898173362016678,
"rewards/rejected": 0.03442993015050888,
"step": 43
},
{
"epoch": 0.07410526315789474,
"grad_norm": 251.79089726814883,
"kl": 4.6118011474609375,
"learning_rate": 3.583333333333333e-07,
"logits/chosen": -944620885.3333334,
"logits/rejected": -546614579.2,
"logps/chosen": -229.73470052083334,
"logps/rejected": -211.629931640625,
"loss": 0.4902,
"rewards/chosen": 0.1472813884417216,
"rewards/margins": 0.1069905529419581,
"rewards/rejected": 0.04029083549976349,
"step": 44
},
{
"epoch": 0.07578947368421053,
"grad_norm": 115.12623939626701,
"kl": 4.802654266357422,
"learning_rate": 3.666666666666666e-07,
"logits/chosen": -1199768166.4,
"logits/rejected": -1026606506.6666666,
"logps/chosen": -250.584716796875,
"logps/rejected": -256.4547932942708,
"loss": 0.4888,
"rewards/chosen": 0.15660583972930908,
"rewards/margins": 0.056487321853637695,
"rewards/rejected": 0.10011851787567139,
"step": 45
},
{
"epoch": 0.07747368421052632,
"grad_norm": 147.48000207541645,
"kl": 2.1971359252929688,
"learning_rate": 3.75e-07,
"logits/chosen": -779579904.0,
"logits/rejected": -777894058.6666666,
"logps/chosen": -234.33892822265625,
"logps/rejected": -279.74733479817706,
"loss": 0.4941,
"rewards/chosen": 0.2183128446340561,
"rewards/margins": 0.09147873024145761,
"rewards/rejected": 0.12683411439259848,
"step": 46
},
{
"epoch": 0.07915789473684211,
"grad_norm": 172.83190668649902,
"kl": 5.049167633056641,
"learning_rate": 3.8333333333333335e-07,
"logits/chosen": -911320192.0,
"logits/rejected": -1069644800.0,
"logps/chosen": -215.0435791015625,
"logps/rejected": -261.28668212890625,
"loss": 0.4947,
"rewards/chosen": 0.17158202826976776,
"rewards/margins": 0.06581267714500427,
"rewards/rejected": 0.10576935112476349,
"step": 47
},
{
"epoch": 0.0808421052631579,
"grad_norm": 135.23703398521903,
"kl": 3.5211753845214844,
"learning_rate": 3.9166666666666664e-07,
"logits/chosen": -1736729472.0,
"logits/rejected": -939244617.1428572,
"logps/chosen": -226.9305877685547,
"logps/rejected": -232.20169503348214,
"loss": 0.4972,
"rewards/chosen": 0.16886138916015625,
"rewards/margins": 0.05592117990766253,
"rewards/rejected": 0.11294020925249372,
"step": 48
},
{
"epoch": 0.08252631578947368,
"grad_norm": 87.76901964257267,
"kl": 5.832710266113281,
"learning_rate": 4e-07,
"logits/chosen": -794667622.4,
"logits/rejected": -622882474.6666666,
"logps/chosen": -230.16806640625,
"logps/rejected": -224.6505330403646,
"loss": 0.4913,
"rewards/chosen": 0.2949032783508301,
"rewards/margins": 0.20258251825968426,
"rewards/rejected": 0.09232076009114583,
"step": 49
},
{
"epoch": 0.08421052631578947,
"grad_norm": 131.7748262627241,
"kl": 4.04681396484375,
"learning_rate": 4.083333333333333e-07,
"logits/chosen": -1042156714.6666666,
"logits/rejected": -956070912.0,
"logps/chosen": -220.65376790364584,
"logps/rejected": -351.86298828125,
"loss": 0.4921,
"rewards/chosen": 0.1149485210577647,
"rewards/margins": -0.044381918509801235,
"rewards/rejected": 0.15933043956756593,
"step": 50
},
{
"epoch": 0.08589473684210526,
"grad_norm": 93.7193003978435,
"kl": 5.549095153808594,
"learning_rate": 4.1666666666666667e-07,
"logits/chosen": -913337088.0,
"logits/rejected": -1463801514.6666667,
"logps/chosen": -234.42970275878906,
"logps/rejected": -256.32749430338544,
"loss": 0.5037,
"rewards/chosen": 0.12055359780788422,
"rewards/margins": 0.05089595417181651,
"rewards/rejected": 0.0696576436360677,
"step": 51
},
{
"epoch": 0.08757894736842105,
"grad_norm": 124.37458019612187,
"kl": 5.329154968261719,
"learning_rate": 4.2499999999999995e-07,
"logits/chosen": -769787840.0,
"logits/rejected": -873242752.0,
"logps/chosen": -228.21946716308594,
"logps/rejected": -178.02896118164062,
"loss": 0.5018,
"rewards/chosen": 0.1875530183315277,
"rewards/margins": -0.029434025287628174,
"rewards/rejected": 0.21698704361915588,
"step": 52
},
{
"epoch": 0.08926315789473684,
"grad_norm": 151.7193063462552,
"kl": 2.5696582794189453,
"learning_rate": 4.3333333333333335e-07,
"logits/chosen": -966253760.0,
"logits/rejected": -818051840.0,
"logps/chosen": -210.7382354736328,
"logps/rejected": -210.70034790039062,
"loss": 0.4948,
"rewards/chosen": 0.09387054294347763,
"rewards/margins": -0.11254673451185226,
"rewards/rejected": 0.2064172774553299,
"step": 53
},
{
"epoch": 0.09094736842105264,
"grad_norm": 113.57346599809466,
"kl": 2.8759994506835938,
"learning_rate": 4.4166666666666664e-07,
"logits/chosen": -936665600.0,
"logits/rejected": -744050602.6666666,
"logps/chosen": -210.92718505859375,
"logps/rejected": -213.8590087890625,
"loss": 0.4893,
"rewards/chosen": 0.17248383164405823,
"rewards/margins": -0.09370574355125427,
"rewards/rejected": 0.2661895751953125,
"step": 54
},
{
"epoch": 0.09263157894736843,
"grad_norm": 174.61803322959847,
"kl": 4.2318267822265625,
"learning_rate": 4.5e-07,
"logits/chosen": -677702400.0,
"logits/rejected": -692814116.5714285,
"logps/chosen": -218.57711791992188,
"logps/rejected": -213.56180245535714,
"loss": 0.4879,
"rewards/chosen": 0.4497177302837372,
"rewards/margins": 0.41578697732516695,
"rewards/rejected": 0.03393075295857021,
"step": 55
},
{
"epoch": 0.09431578947368421,
"grad_norm": 147.27747224286998,
"kl": 3.260711669921875,
"learning_rate": 4.5833333333333327e-07,
"logits/chosen": -1289244544.0,
"logits/rejected": -916406464.0,
"logps/chosen": -246.1941375732422,
"logps/rejected": -253.98768615722656,
"loss": 0.4921,
"rewards/chosen": 0.16882018744945526,
"rewards/margins": -0.14416618645191193,
"rewards/rejected": 0.3129863739013672,
"step": 56
},
{
"epoch": 0.096,
"grad_norm": 95.08985748052072,
"kl": 1.6493759155273438,
"learning_rate": 4.6666666666666666e-07,
"logits/chosen": -733326400.0,
"logits/rejected": -922427392.0,
"logps/chosen": -178.46246337890625,
"logps/rejected": -258.2501627604167,
"loss": 0.4867,
"rewards/chosen": 0.3112235963344574,
"rewards/margins": 0.17294743657112122,
"rewards/rejected": 0.13827615976333618,
"step": 57
},
{
"epoch": 0.09768421052631579,
"grad_norm": 133.57090255911547,
"kl": 3.168426513671875,
"learning_rate": 4.7499999999999995e-07,
"logits/chosen": -767859712.0,
"logits/rejected": -977725248.0,
"logps/chosen": -228.76634216308594,
"logps/rejected": -284.31427001953125,
"loss": 0.4985,
"rewards/chosen": 0.1971721649169922,
"rewards/margins": 0.03166121244430542,
"rewards/rejected": 0.16551095247268677,
"step": 58
},
{
"epoch": 0.09936842105263158,
"grad_norm": 108.0916864306588,
"kl": 0.6577796936035156,
"learning_rate": 4.833333333333333e-07,
"logits/chosen": -1173459148.8,
"logits/rejected": -655583232.0,
"logps/chosen": -196.4993408203125,
"logps/rejected": -219.53653971354166,
"loss": 0.4757,
"rewards/chosen": 0.23870911598205566,
"rewards/margins": 0.28514466087023416,
"rewards/rejected": -0.046435544888178505,
"step": 59
},
{
"epoch": 0.10105263157894737,
"grad_norm": 87.62449313729059,
"kl": 2.194793701171875,
"learning_rate": 4.916666666666666e-07,
"logits/chosen": -1028928256.0,
"logits/rejected": -540835200.0,
"logps/chosen": -227.88580322265625,
"logps/rejected": -234.1096649169922,
"loss": 0.45,
"rewards/chosen": 0.5302422046661377,
"rewards/margins": 1.1665672659873962,
"rewards/rejected": -0.6363250613212585,
"step": 60
},
{
"epoch": 0.10273684210526315,
"grad_norm": 119.47945666482947,
"kl": 1.5802383422851562,
"learning_rate": 5e-07,
"logits/chosen": -1018257536.0,
"logits/rejected": -1083207040.0,
"logps/chosen": -222.7873992919922,
"logps/rejected": -275.23138427734375,
"loss": 0.5044,
"rewards/chosen": 0.32063254714012146,
"rewards/margins": 0.10322417318820953,
"rewards/rejected": 0.21740837395191193,
"step": 61
},
{
"epoch": 0.10442105263157894,
"grad_norm": 114.74202024192041,
"kl": 1.9775314331054688,
"learning_rate": 4.999956736067562e-07,
"logits/chosen": -720981299.2,
"logits/rejected": -543601152.0,
"logps/chosen": -240.6268798828125,
"logps/rejected": -207.02860514322916,
"loss": 0.4983,
"rewards/chosen": 0.25333800315856936,
"rewards/margins": -0.036342994372049964,
"rewards/rejected": 0.2896809975306193,
"step": 62
},
{
"epoch": 0.10610526315789473,
"grad_norm": 77.65993832754032,
"kl": 0.7235641479492188,
"learning_rate": 4.999826945767664e-07,
"logits/chosen": -692212443.4285715,
"logits/rejected": -1071096576.0,
"logps/chosen": -255.70343889508928,
"logps/rejected": -259.9676208496094,
"loss": 0.5084,
"rewards/chosen": 0.2564025947025844,
"rewards/margins": -0.06277771506990704,
"rewards/rejected": 0.31918030977249146,
"step": 63
},
{
"epoch": 0.10778947368421053,
"grad_norm": 101.9412698769531,
"kl": 2.6254653930664062,
"learning_rate": 4.999610633592496e-07,
"logits/chosen": -1115974656.0,
"logits/rejected": -880023961.6,
"logps/chosen": -227.8680419921875,
"logps/rejected": -210.0842041015625,
"loss": 0.4639,
"rewards/chosen": 0.1862970987955729,
"rewards/margins": 0.7373786608378092,
"rewards/rejected": -0.5510815620422364,
"step": 64
},
{
"epoch": 0.10947368421052632,
"grad_norm": 83.26215203358481,
"kl": 0.0,
"learning_rate": 4.99930780702887e-07,
"logits/chosen": -2037875712.0,
"logits/rejected": -891881472.0,
"logps/chosen": -298.9163818359375,
"logps/rejected": -212.7870076497396,
"loss": 0.5016,
"rewards/chosen": 0.0009033232927322388,
"rewards/margins": -0.28533248603343964,
"rewards/rejected": 0.2862358093261719,
"step": 65
},
{
"epoch": 0.11115789473684211,
"grad_norm": 65.37492119822153,
"kl": 4.6044464111328125,
"learning_rate": 4.998918476557962e-07,
"logits/chosen": -1163713365.3333333,
"logits/rejected": -717799936.0,
"logps/chosen": -291.79477945963544,
"logps/rejected": -137.73323974609374,
"loss": 0.4896,
"rewards/chosen": 0.2572901447614034,
"rewards/margins": 0.1486562927563985,
"rewards/rejected": 0.10863385200500489,
"step": 66
},
{
"epoch": 0.1128421052631579,
"grad_norm": 108.123906294101,
"kl": 0.8921241760253906,
"learning_rate": 4.998442655654946e-07,
"logits/chosen": -1018941312.0,
"logits/rejected": -723264512.0,
"logps/chosen": -227.9298553466797,
"logps/rejected": -196.53343200683594,
"loss": 0.4936,
"rewards/chosen": 0.1972118318080902,
"rewards/margins": -0.028878778219223022,
"rewards/rejected": 0.22609061002731323,
"step": 67
},
{
"epoch": 0.11452631578947368,
"grad_norm": 154.7633558965442,
"kl": 3.6258544921875,
"learning_rate": 4.997880360788526e-07,
"logits/chosen": -1207795302.4,
"logits/rejected": -640148437.3333334,
"logps/chosen": -207.35986328125,
"logps/rejected": -239.59635416666666,
"loss": 0.4908,
"rewards/chosen": 0.24745573997497558,
"rewards/margins": 0.09161568085352578,
"rewards/rejected": 0.1558400591214498,
"step": 68
},
{
"epoch": 0.11621052631578947,
"grad_norm": 189.59978273083877,
"kl": 1.4420166015625,
"learning_rate": 4.997231611420373e-07,
"logits/chosen": -722635929.6,
"logits/rejected": -1086283520.0,
"logps/chosen": -205.415087890625,
"logps/rejected": -223.599853515625,
"loss": 0.4755,
"rewards/chosen": 0.2955571174621582,
"rewards/margins": 0.23686978320280708,
"rewards/rejected": 0.0586873342593511,
"step": 69
},
{
"epoch": 0.11789473684210526,
"grad_norm": 82.88808037398137,
"kl": 0.0,
"learning_rate": 4.996496430004445e-07,
"logits/chosen": -894456704.0,
"logits/rejected": -1008478336.0,
"logps/chosen": -197.12306213378906,
"logps/rejected": -189.72410583496094,
"loss": 0.4798,
"rewards/chosen": 0.5647777318954468,
"rewards/margins": 0.4188384860754013,
"rewards/rejected": 0.14593924582004547,
"step": 70
},
{
"epoch": 0.11957894736842105,
"grad_norm": 116.22452277931659,
"kl": 0.0,
"learning_rate": 4.995674841986217e-07,
"logits/chosen": -1016096426.6666666,
"logits/rejected": -340471296.0,
"logps/chosen": -235.82279459635416,
"logps/rejected": -181.5946807861328,
"loss": 0.4523,
"rewards/chosen": 0.3477465311686198,
"rewards/margins": 0.14004083971182507,
"rewards/rejected": 0.20770569145679474,
"step": 71
},
{
"epoch": 0.12126315789473684,
"grad_norm": 169.23180648930278,
"kl": 2.728069305419922,
"learning_rate": 4.994766875801788e-07,
"logits/chosen": -802654592.0,
"logits/rejected": -754744960.0,
"logps/chosen": -195.60926818847656,
"logps/rejected": -193.22048950195312,
"loss": 0.5093,
"rewards/chosen": 0.20827637612819672,
"rewards/margins": 0.17099057137966156,
"rewards/rejected": 0.037285804748535156,
"step": 72
},
{
"epoch": 0.12294736842105262,
"grad_norm": 97.86580017412098,
"kl": 1.1262283325195312,
"learning_rate": 4.993772562876909e-07,
"logits/chosen": -791456000.0,
"logits/rejected": -912904704.0,
"logps/chosen": -172.02268981933594,
"logps/rejected": -196.89213053385416,
"loss": 0.4554,
"rewards/chosen": 0.7986927032470703,
"rewards/margins": 1.128045916557312,
"rewards/rejected": -0.3293532133102417,
"step": 73
},
{
"epoch": 0.12463157894736843,
"grad_norm": 87.12336042753508,
"kl": 0.322601318359375,
"learning_rate": 4.992691937625891e-07,
"logits/chosen": -579524224.0,
"logits/rejected": -731365120.0,
"logps/chosen": -232.52737426757812,
"logps/rejected": -234.31417846679688,
"loss": 0.4493,
"rewards/chosen": 0.24241486191749573,
"rewards/margins": 1.9988974630832672,
"rewards/rejected": -1.7564826011657715,
"step": 74
},
{
"epoch": 0.12631578947368421,
"grad_norm": 72.27499509744034,
"kl": 0.0,
"learning_rate": 4.991525037450412e-07,
"logits/chosen": -883592192.0,
"logits/rejected": -728281088.0,
"logps/chosen": -225.2747599283854,
"logps/rejected": -199.8900146484375,
"loss": 0.4878,
"rewards/chosen": 0.1509552001953125,
"rewards/margins": 0.05537720024585724,
"rewards/rejected": 0.09557799994945526,
"step": 75
},
{
"epoch": 0.128,
"grad_norm": 107.49035218888741,
"kl": 0.0,
"learning_rate": 4.990271902738222e-07,
"logits/chosen": -625081190.4,
"logits/rejected": -646702250.6666666,
"logps/chosen": -232.6982421875,
"logps/rejected": -326.85638427734375,
"loss": 0.45,
"rewards/chosen": 0.23982207775115966,
"rewards/margins": 0.3932990511258443,
"rewards/rejected": -0.15347697337468466,
"step": 76
},
{
"epoch": 0.1296842105263158,
"grad_norm": 101.17758060728043,
"kl": 0.0,
"learning_rate": 4.988932576861752e-07,
"logits/chosen": -1040184640.0,
"logits/rejected": -690037376.0,
"logps/chosen": -184.14572143554688,
"logps/rejected": -228.58084106445312,
"loss": 0.5148,
"rewards/chosen": 0.24517688155174255,
"rewards/margins": 0.009785279631614685,
"rewards/rejected": 0.23539160192012787,
"step": 77
},
{
"epoch": 0.13136842105263158,
"grad_norm": 84.42483041363506,
"kl": 0.3793144226074219,
"learning_rate": 4.987507106176606e-07,
"logits/chosen": -611603046.4,
"logits/rejected": -684819114.6666666,
"logps/chosen": -241.2910400390625,
"logps/rejected": -251.3832804361979,
"loss": 0.4885,
"rewards/chosen": 0.5874441623687744,
"rewards/margins": -0.2702706495920817,
"rewards/rejected": 0.8577148119608561,
"step": 78
},
{
"epoch": 0.13305263157894737,
"grad_norm": 123.42412793246983,
"kl": 0.0,
"learning_rate": 4.985995540019955e-07,
"logits/chosen": -902578752.0,
"logits/rejected": -1213336917.3333333,
"logps/chosen": -271.75140380859375,
"logps/rejected": -266.7210286458333,
"loss": 0.4488,
"rewards/chosen": 0.22760315239429474,
"rewards/margins": -0.06182275712490082,
"rewards/rejected": 0.28942590951919556,
"step": 79
},
{
"epoch": 0.13473684210526315,
"grad_norm": 70.07470086670996,
"kl": 0.5335845947265625,
"learning_rate": 4.984397930708837e-07,
"logits/chosen": -737288960.0,
"logits/rejected": -944431616.0,
"logps/chosen": -198.37362670898438,
"logps/rejected": -224.55250549316406,
"loss": 0.4547,
"rewards/chosen": 0.5928070545196533,
"rewards/margins": 0.4272579848766327,
"rewards/rejected": 0.16554906964302063,
"step": 80
},
{
"epoch": 0.13642105263157894,
"grad_norm": 66.8382583080533,
"kl": 0.0,
"learning_rate": 4.982714333538342e-07,
"logits/chosen": -604147353.6,
"logits/rejected": -792211114.6666666,
"logps/chosen": -264.7359375,
"logps/rejected": -207.922119140625,
"loss": 0.502,
"rewards/chosen": 0.11271027326583863,
"rewards/margins": -0.04983551700909933,
"rewards/rejected": 0.16254579027493796,
"step": 81
},
{
"epoch": 0.13810526315789473,
"grad_norm": 90.56573316497496,
"kl": 1.1843795776367188,
"learning_rate": 4.980944806779698e-07,
"logits/chosen": -994012032.0,
"logits/rejected": -756074154.6666666,
"logps/chosen": -211.22325134277344,
"logps/rejected": -273.8466389973958,
"loss": 0.503,
"rewards/chosen": 0.21327057480812073,
"rewards/margins": 0.3816507955392202,
"rewards/rejected": -0.16838022073109946,
"step": 82
},
{
"epoch": 0.13978947368421052,
"grad_norm": 95.89588780633147,
"kl": 0.0,
"learning_rate": 4.979089411678251e-07,
"logits/chosen": -1088499200.0,
"logits/rejected": -580556672.0,
"logps/chosen": -201.80905151367188,
"logps/rejected": -149.05142211914062,
"loss": 0.4614,
"rewards/chosen": 0.5321415066719055,
"rewards/margins": 0.3093837797641754,
"rewards/rejected": 0.2227577269077301,
"step": 83
},
{
"epoch": 0.1414736842105263,
"grad_norm": 84.70589504268058,
"kl": 0.0,
"learning_rate": 4.977148212451354e-07,
"logits/chosen": -1006119296.0,
"logits/rejected": -702025600.0,
"logps/chosen": -220.37423706054688,
"logps/rejected": -213.92950439453125,
"loss": 0.4784,
"rewards/chosen": 0.37330934405326843,
"rewards/margins": -0.14478299021720886,
"rewards/rejected": 0.5180923342704773,
"step": 84
},
{
"epoch": 0.1431578947368421,
"grad_norm": 80.24131984783912,
"kl": 0.18048095703125,
"learning_rate": 4.975121276286136e-07,
"logits/chosen": -574469888.0,
"logits/rejected": -1207006720.0,
"logps/chosen": -224.60697428385416,
"logps/rejected": -230.05050659179688,
"loss": 0.4587,
"rewards/chosen": 0.3602706988652547,
"rewards/margins": 0.06148912509282428,
"rewards/rejected": 0.2987815737724304,
"step": 85
},
{
"epoch": 0.14484210526315788,
"grad_norm": 119.4817380814399,
"kl": 0.0,
"learning_rate": 4.97300867333718e-07,
"logits/chosen": -559665344.0,
"logits/rejected": -1335321600.0,
"logps/chosen": -246.3226318359375,
"logps/rejected": -240.66729736328125,
"loss": 0.4545,
"rewards/chosen": 0.27413901686668396,
"rewards/margins": 0.11862906813621521,
"rewards/rejected": 0.15550994873046875,
"step": 86
},
{
"epoch": 0.14652631578947367,
"grad_norm": 76.72262570711749,
"kl": 0.0,
"learning_rate": 4.970810476724096e-07,
"logits/chosen": -794261632.0,
"logits/rejected": -888321194.6666666,
"logps/chosen": -223.9755096435547,
"logps/rejected": -230.53841145833334,
"loss": 0.4915,
"rewards/chosen": 0.6259682178497314,
"rewards/margins": 0.16484379768371582,
"rewards/rejected": 0.4611244201660156,
"step": 87
},
{
"epoch": 0.1482105263157895,
"grad_norm": 103.45011997542589,
"kl": 0.6499404907226562,
"learning_rate": 4.968526762528988e-07,
"logits/chosen": -780058624.0,
"logits/rejected": -916248371.2,
"logps/chosen": -256.363525390625,
"logps/rejected": -207.6240234375,
"loss": 0.4831,
"rewards/chosen": 0.28617451588312787,
"rewards/margins": 0.07391071716944378,
"rewards/rejected": 0.2122637987136841,
"step": 88
},
{
"epoch": 0.14989473684210528,
"grad_norm": 215.23670048541007,
"kl": 0.0,
"learning_rate": 4.96615760979382e-07,
"logits/chosen": -637054976.0,
"logits/rejected": -924139712.0,
"logps/chosen": -238.05352783203125,
"logps/rejected": -168.34219360351562,
"loss": 0.4829,
"rewards/chosen": 0.5608859062194824,
"rewards/margins": -0.13643085956573486,
"rewards/rejected": 0.6973167657852173,
"step": 89
},
{
"epoch": 0.15157894736842106,
"grad_norm": 76.57882355799092,
"kl": 0.0,
"learning_rate": 4.963703100517683e-07,
"logits/chosen": -1086607488.0,
"logits/rejected": -987515264.0,
"logps/chosen": -204.48553466796875,
"logps/rejected": -251.81475830078125,
"loss": 0.4542,
"rewards/chosen": 0.7157703638076782,
"rewards/margins": 1.3066982626914978,
"rewards/rejected": -0.5909278988838196,
"step": 90
},
{
"epoch": 0.15326315789473685,
"grad_norm": 85.3247850434762,
"kl": 0.0,
"learning_rate": 4.961163319653958e-07,
"logits/chosen": -1008856704.0,
"logits/rejected": -618826112.0,
"logps/chosen": -187.7606201171875,
"logps/rejected": -191.36654663085938,
"loss": 0.4556,
"rewards/chosen": 0.2826835811138153,
"rewards/margins": -0.048234134912490845,
"rewards/rejected": 0.33091771602630615,
"step": 91
},
{
"epoch": 0.15494736842105264,
"grad_norm": 172.74293893285517,
"kl": 0.0,
"learning_rate": 4.958538355107369e-07,
"logits/chosen": -771198272.0,
"logits/rejected": -911799936.0,
"logps/chosen": -198.0313262939453,
"logps/rejected": -243.49655151367188,
"loss": 0.4021,
"rewards/chosen": 0.537731945514679,
"rewards/margins": 1.0454849600791931,
"rewards/rejected": -0.5077530145645142,
"step": 92
},
{
"epoch": 0.15663157894736843,
"grad_norm": 115.09597490195979,
"kl": 0.0,
"learning_rate": 4.955828297730948e-07,
"logits/chosen": -578800576.0,
"logits/rejected": -940042752.0,
"logps/chosen": -166.36688232421875,
"logps/rejected": -236.39175415039062,
"loss": 0.4575,
"rewards/chosen": 0.2515438199043274,
"rewards/margins": 0.63800048828125,
"rewards/rejected": -0.3864566683769226,
"step": 93
},
{
"epoch": 0.15831578947368422,
"grad_norm": 189.22015271094992,
"kl": 0.0,
"learning_rate": 4.953033241322886e-07,
"logits/chosen": -971949875.2,
"logits/rejected": -642591701.3333334,
"logps/chosen": -208.137646484375,
"logps/rejected": -267.102294921875,
"loss": 0.4444,
"rewards/chosen": 0.30209686756134035,
"rewards/margins": 0.15472238659858706,
"rewards/rejected": 0.1473744809627533,
"step": 94
},
{
"epoch": 0.16,
"grad_norm": 151.69924056862672,
"kl": 0.8690567016601562,
"learning_rate": 4.950153282623288e-07,
"logits/chosen": -929138029.7142857,
"logits/rejected": -305715200.0,
"logps/chosen": -243.89243861607142,
"logps/rejected": -234.21505737304688,
"loss": 0.4272,
"rewards/chosen": 1.003729956490653,
"rewards/margins": 1.042243140084403,
"rewards/rejected": -0.03851318359375,
"step": 95
},
{
"epoch": 0.1616842105263158,
"grad_norm": 117.75406936780551,
"kl": 0.0,
"learning_rate": 4.947188521310827e-07,
"logits/chosen": -1078521088.0,
"logits/rejected": -754925696.0,
"logps/chosen": -241.37603759765625,
"logps/rejected": -246.62840270996094,
"loss": 0.468,
"rewards/chosen": 0.06325645744800568,
"rewards/margins": 2.5708944350481033,
"rewards/rejected": -2.5076379776000977,
"step": 96
},
{
"epoch": 0.16336842105263158,
"grad_norm": 92.90737005905727,
"kl": 0.0,
"learning_rate": 4.944139059999286e-07,
"logits/chosen": -939986624.0,
"logits/rejected": -757044608.0,
"logps/chosen": -255.14584350585938,
"logps/rejected": -217.84385681152344,
"loss": 0.4852,
"rewards/chosen": 0.14831504225730896,
"rewards/margins": -0.12935563921928406,
"rewards/rejected": 0.277670681476593,
"step": 97
},
{
"epoch": 0.16505263157894737,
"grad_norm": 75.87889683822247,
"kl": 2.0634384155273438,
"learning_rate": 4.941005004234018e-07,
"logits/chosen": -674378444.8,
"logits/rejected": -790328490.6666666,
"logps/chosen": -222.7276123046875,
"logps/rejected": -319.3123372395833,
"loss": 0.4712,
"rewards/chosen": 0.9145062446594239,
"rewards/margins": 1.130675482749939,
"rewards/rejected": -0.21616923809051514,
"step": 98
},
{
"epoch": 0.16673684210526316,
"grad_norm": 83.71279388164065,
"kl": 0.06458663940429688,
"learning_rate": 4.937786462488283e-07,
"logits/chosen": -715301376.0,
"logits/rejected": -695437458.2857143,
"logps/chosen": -187.44473266601562,
"logps/rejected": -221.67951311383928,
"loss": 0.4397,
"rewards/chosen": 0.4498535096645355,
"rewards/margins": 0.21583446860313416,
"rewards/rejected": 0.23401904106140137,
"step": 99
},
{
"epoch": 0.16842105263157894,
"grad_norm": 105.87757666440052,
"kl": 4.265495300292969,
"learning_rate": 4.9344835461595e-07,
"logits/chosen": -985528320.0,
"logits/rejected": -841993301.3333334,
"logps/chosen": -187.4853759765625,
"logps/rejected": -244.225830078125,
"loss": 0.4832,
"rewards/chosen": 0.04033417701721191,
"rewards/margins": -0.09548329114913941,
"rewards/rejected": 0.13581746816635132,
"step": 100
},
{
"epoch": 0.17010526315789473,
"grad_norm": 99.19589292443813,
"kl": 0.0,
"learning_rate": 4.93109636956539e-07,
"logits/chosen": -1094334464.0,
"logits/rejected": -712445226.6666666,
"logps/chosen": -216.4138671875,
"logps/rejected": -227.96439615885416,
"loss": 0.4423,
"rewards/chosen": 0.8524996757507324,
"rewards/margins": 1.196986190478007,
"rewards/rejected": -0.3444865147272746,
"step": 101
},
{
"epoch": 0.17178947368421052,
"grad_norm": 74.15099841568721,
"kl": 2.151416778564453,
"learning_rate": 4.927625049940012e-07,
"logits/chosen": -766361152.0,
"logits/rejected": -333142048.0,
"logps/chosen": -199.77578735351562,
"logps/rejected": -168.93804931640625,
"loss": 0.4554,
"rewards/chosen": 0.7172653079032898,
"rewards/margins": 0.35114437341690063,
"rewards/rejected": 0.36612093448638916,
"step": 102
},
{
"epoch": 0.1734736842105263,
"grad_norm": 107.51116166109057,
"kl": 1.1529388427734375,
"learning_rate": 4.92406970742972e-07,
"logits/chosen": -971440256.0,
"logits/rejected": -828623232.0,
"logps/chosen": -195.1447296142578,
"logps/rejected": -280.0885009765625,
"loss": 0.4425,
"rewards/chosen": 0.5913848876953125,
"rewards/margins": 1.5447425842285156,
"rewards/rejected": -0.9533576965332031,
"step": 103
},
{
"epoch": 0.1751578947368421,
"grad_norm": 74.47709378212676,
"kl": 4.383308410644531,
"learning_rate": 4.920430465088991e-07,
"logits/chosen": -670352640.0,
"logits/rejected": -671316608.0,
"logps/chosen": -255.62147521972656,
"logps/rejected": -200.2668914794922,
"loss": 0.4575,
"rewards/chosen": 0.5768207311630249,
"rewards/margins": 0.4924035966396332,
"rewards/rejected": 0.08441713452339172,
"step": 104
},
{
"epoch": 0.17684210526315788,
"grad_norm": 132.0792640659443,
"kl": 1.2592926025390625,
"learning_rate": 4.916707448876173e-07,
"logits/chosen": -832779328.0,
"logits/rejected": -1296543232.0,
"logps/chosen": -227.14276123046875,
"logps/rejected": -218.0616912841797,
"loss": 0.4254,
"rewards/chosen": 1.1795654296875,
"rewards/margins": 1.3467857092618942,
"rewards/rejected": -0.16722027957439423,
"step": 105
},
{
"epoch": 0.17852631578947367,
"grad_norm": 145.72514277447476,
"kl": 0.0,
"learning_rate": 4.912900787649123e-07,
"logits/chosen": -898515675.4285715,
"logits/rejected": -1168541696.0,
"logps/chosen": -201.853515625,
"logps/rejected": -228.6392822265625,
"loss": 0.4635,
"rewards/chosen": 0.14519500732421875,
"rewards/margins": -0.7220779657363892,
"rewards/rejected": 0.8672729730606079,
"step": 106
},
{
"epoch": 0.18021052631578946,
"grad_norm": 96.96868580065858,
"kl": 0.47565460205078125,
"learning_rate": 4.90901061316075e-07,
"logits/chosen": -979152704.0,
"logits/rejected": -788988416.0,
"logps/chosen": -262.84649658203125,
"logps/rejected": -219.4362589518229,
"loss": 0.4916,
"rewards/chosen": -0.06726683676242828,
"rewards/margins": 1.062198743224144,
"rewards/rejected": -1.1294655799865723,
"step": 107
},
{
"epoch": 0.18189473684210528,
"grad_norm": 131.83686757875427,
"kl": 0.0,
"learning_rate": 4.905037060054449e-07,
"logits/chosen": -1066169344.0,
"logits/rejected": -1416755712.0,
"logps/chosen": -177.3980255126953,
"logps/rejected": -229.30007934570312,
"loss": 0.4426,
"rewards/chosen": 1.1091067790985107,
"rewards/margins": 0.9519591778516769,
"rewards/rejected": 0.1571476012468338,
"step": 108
},
{
"epoch": 0.18357894736842106,
"grad_norm": 127.9527788904727,
"kl": 0.0,
"learning_rate": 4.900980265859448e-07,
"logits/chosen": -559809066.6666666,
"logits/rejected": -611907481.6,
"logps/chosen": -164.3425089518229,
"logps/rejected": -240.176318359375,
"loss": 0.424,
"rewards/chosen": 1.06194003423055,
"rewards/margins": 0.5945489724477131,
"rewards/rejected": 0.4673910617828369,
"step": 109
},
{
"epoch": 0.18526315789473685,
"grad_norm": 84.90732586257637,
"kl": 0.0,
"learning_rate": 4.896840370986042e-07,
"logits/chosen": -976780032.0,
"logits/rejected": -716965632.0,
"logps/chosen": -253.70814514160156,
"logps/rejected": -167.56753540039062,
"loss": 0.4595,
"rewards/chosen": 0.47311899065971375,
"rewards/margins": 0.09406355023384094,
"rewards/rejected": 0.3790554404258728,
"step": 110
},
{
"epoch": 0.18694736842105264,
"grad_norm": 89.53866936328932,
"kl": 0.0,
"learning_rate": 4.892617518720737e-07,
"logits/chosen": -829139285.3333334,
"logits/rejected": -455405536.0,
"logps/chosen": -206.80936686197916,
"logps/rejected": -257.3726806640625,
"loss": 0.4362,
"rewards/chosen": 0.4422444502512614,
"rewards/margins": 2.1721797386805215,
"rewards/rejected": -1.7299352884292603,
"step": 111
},
{
"epoch": 0.18863157894736843,
"grad_norm": 192.37189364365034,
"kl": 0.0,
"learning_rate": 4.888311855221289e-07,
"logits/chosen": -770445312.0,
"logits/rejected": -951467392.0,
"logps/chosen": -199.4364471435547,
"logps/rejected": -215.43002319335938,
"loss": 0.5149,
"rewards/chosen": -1.172553300857544,
"rewards/margins": -1.1502388417720795,
"rewards/rejected": -0.022314459085464478,
"step": 112
},
{
"epoch": 0.19031578947368422,
"grad_norm": 95.28956568265271,
"kl": 0.0,
"learning_rate": 4.883923529511646e-07,
"logits/chosen": -535924416.0,
"logits/rejected": -786336960.0,
"logps/chosen": -247.45355224609375,
"logps/rejected": -203.68768310546875,
"loss": 0.4584,
"rewards/chosen": 0.5863098502159119,
"rewards/margins": -0.6281996369361877,
"rewards/rejected": 1.2145094871520996,
"step": 113
},
{
"epoch": 0.192,
"grad_norm": 141.25750368118935,
"kl": 0.0,
"learning_rate": 4.879452693476789e-07,
"logits/chosen": -1001671372.8,
"logits/rejected": -836295936.0,
"logps/chosen": -201.454052734375,
"logps/rejected": -205.16349283854166,
"loss": 0.4619,
"rewards/chosen": 0.1915292501449585,
"rewards/margins": 1.2447933912277223,
"rewards/rejected": -1.0532641410827637,
"step": 114
},
{
"epoch": 0.1936842105263158,
"grad_norm": 90.01647424789367,
"kl": 1.8182296752929688,
"learning_rate": 4.874899501857477e-07,
"logits/chosen": -642349952.0,
"logits/rejected": -997530176.0,
"logps/chosen": -269.48760986328125,
"logps/rejected": -281.49591064453125,
"loss": 0.4548,
"rewards/chosen": -0.8407798409461975,
"rewards/margins": -0.7594596445560455,
"rewards/rejected": -0.08132019639015198,
"step": 115
},
{
"epoch": 0.19536842105263158,
"grad_norm": 142.4095270366161,
"kl": 0.0,
"learning_rate": 4.87026411224489e-07,
"logits/chosen": -909213798.4,
"logits/rejected": -513591637.3333333,
"logps/chosen": -240.9342041015625,
"logps/rejected": -284.7578531901042,
"loss": 0.405,
"rewards/chosen": 0.716831636428833,
"rewards/margins": 1.0737988074620564,
"rewards/rejected": -0.35696717103322345,
"step": 116
},
{
"epoch": 0.19705263157894737,
"grad_norm": 313.6380304727835,
"kl": 0.0,
"learning_rate": 4.865546685075174e-07,
"logits/chosen": -904379648.0,
"logits/rejected": -820042547.2,
"logps/chosen": -220.26141357421875,
"logps/rejected": -245.7005859375,
"loss": 0.431,
"rewards/chosen": 1.131263256072998,
"rewards/margins": 1.3263026356697083,
"rewards/rejected": -0.1950393795967102,
"step": 117
},
{
"epoch": 0.19873684210526316,
"grad_norm": 108.28624508912246,
"kl": 0.0,
"learning_rate": 4.860747383623889e-07,
"logits/chosen": -723659136.0,
"logits/rejected": -772578752.0,
"logps/chosen": -184.33941650390625,
"logps/rejected": -185.81124877929688,
"loss": 0.3927,
"rewards/chosen": 1.3619561195373535,
"rewards/margins": 1.463188648223877,
"rewards/rejected": -0.10123252868652344,
"step": 118
},
{
"epoch": 0.20042105263157894,
"grad_norm": 314.88581870057607,
"kl": 0.0,
"learning_rate": 4.85586637400036e-07,
"logits/chosen": -458558624.0,
"logits/rejected": -640709376.0,
"logps/chosen": -257.4935302734375,
"logps/rejected": -263.326416015625,
"loss": 0.4136,
"rewards/chosen": 0.4196751117706299,
"rewards/margins": 1.0923622846603394,
"rewards/rejected": -0.6726871728897095,
"step": 119
},
{
"epoch": 0.20210526315789473,
"grad_norm": 138.4365785651591,
"kl": 0.0,
"learning_rate": 4.85090382514192e-07,
"logits/chosen": -712995328.0,
"logits/rejected": -820921941.3333334,
"logps/chosen": -222.21136474609375,
"logps/rejected": -301.34344482421875,
"loss": 0.4317,
"rewards/chosen": 0.45389634370803833,
"rewards/margins": 0.7454813917477925,
"rewards/rejected": -0.2915850480397542,
"step": 120
},
{
"epoch": 0.20378947368421052,
"grad_norm": 92.36506768933428,
"kl": 0.0,
"learning_rate": 4.845859908808073e-07,
"logits/chosen": -695455232.0,
"logits/rejected": -842530406.4,
"logps/chosen": -220.80411783854166,
"logps/rejected": -278.7416015625,
"loss": 0.458,
"rewards/chosen": 0.664484699567159,
"rewards/margins": 1.1589130957921348,
"rewards/rejected": -0.4944283962249756,
"step": 121
},
{
"epoch": 0.2054736842105263,
"grad_norm": 73.63166789116582,
"kl": 0.4639434814453125,
"learning_rate": 4.840734799574546e-07,
"logits/chosen": -1259681877.3333333,
"logits/rejected": -1012162560.0,
"logps/chosen": -189.5477091471354,
"logps/rejected": -160.91326904296875,
"loss": 0.4281,
"rewards/chosen": 0.8242993354797363,
"rewards/margins": 0.5254170298576355,
"rewards/rejected": 0.29888230562210083,
"step": 122
},
{
"epoch": 0.2071578947368421,
"grad_norm": 97.28869763760606,
"kl": 0.318084716796875,
"learning_rate": 4.835528674827239e-07,
"logits/chosen": -732118954.6666666,
"logits/rejected": -770999680.0,
"logps/chosen": -228.71061197916666,
"logps/rejected": -203.45211791992188,
"loss": 0.4352,
"rewards/chosen": 0.7461237907409668,
"rewards/margins": 0.08015847206115723,
"rewards/rejected": 0.6659653186798096,
"step": 123
},
{
"epoch": 0.20884210526315788,
"grad_norm": 103.28487237641944,
"kl": 0.0,
"learning_rate": 4.830241714756098e-07,
"logits/chosen": -580801901.7142857,
"logits/rejected": -687661312.0,
"logps/chosen": -200.93122209821428,
"logps/rejected": -259.4430847167969,
"loss": 0.3573,
"rewards/chosen": 1.0478062629699707,
"rewards/margins": -0.5003474950790405,
"rewards/rejected": 1.5481537580490112,
"step": 124
},
{
"epoch": 0.21052631578947367,
"grad_norm": 80.62146513928802,
"kl": 1.8628311157226562,
"learning_rate": 4.82487410234887e-07,
"logits/chosen": -867455707.4285715,
"logits/rejected": -523073568.0,
"logps/chosen": -173.80622209821428,
"logps/rejected": -181.6681671142578,
"loss": 0.5246,
"rewards/chosen": -0.30137692178998676,
"rewards/margins": -1.384874565260751,
"rewards/rejected": 1.0834976434707642,
"step": 125
},
{
"epoch": 0.21221052631578946,
"grad_norm": 82.55380217967634,
"kl": 2.27947998046875,
"learning_rate": 4.819426023384769e-07,
"logits/chosen": -958819942.4,
"logits/rejected": -718372181.3333334,
"logps/chosen": -212.732763671875,
"logps/rejected": -163.0384724934896,
"loss": 0.5059,
"rewards/chosen": 0.950899314880371,
"rewards/margins": 0.2773471991221109,
"rewards/rejected": 0.6735521157582601,
"step": 126
},
{
"epoch": 0.21389473684210528,
"grad_norm": 134.57523570837543,
"kl": 0.0,
"learning_rate": 4.813897666428053e-07,
"logits/chosen": -582362880.0,
"logits/rejected": -902837043.2,
"logps/chosen": -258.3628743489583,
"logps/rejected": -191.00611572265626,
"loss": 0.4573,
"rewards/chosen": 0.8414382934570312,
"rewards/margins": 0.5884557962417603,
"rewards/rejected": 0.252982497215271,
"step": 127
},
{
"epoch": 0.21557894736842106,
"grad_norm": 77.91215003119481,
"kl": 0.0,
"learning_rate": 4.80828922282149e-07,
"logits/chosen": -715421593.6,
"logits/rejected": -996736597.3333334,
"logps/chosen": -221.0246337890625,
"logps/rejected": -193.3650105794271,
"loss": 0.4897,
"rewards/chosen": 0.5778164863586426,
"rewards/margins": -0.47630953788757324,
"rewards/rejected": 1.0541260242462158,
"step": 128
},
{
"epoch": 0.21726315789473685,
"grad_norm": 84.25313070423411,
"kl": 2.1194534301757812,
"learning_rate": 4.802600886679741e-07,
"logits/chosen": -852968704.0,
"logits/rejected": -692032000.0,
"logps/chosen": -224.84365844726562,
"logps/rejected": -224.15106201171875,
"loss": 0.436,
"rewards/chosen": 0.7637878656387329,
"rewards/margins": 0.6827945858240128,
"rewards/rejected": 0.08099327981472015,
"step": 129
},
{
"epoch": 0.21894736842105264,
"grad_norm": 76.6604526230067,
"kl": 0.0,
"learning_rate": 4.79683285488264e-07,
"logits/chosen": -768312064.0,
"logits/rejected": -553834240.0,
"logps/chosen": -209.3781982421875,
"logps/rejected": -275.8822428385417,
"loss": 0.3727,
"rewards/chosen": 1.0357406616210938,
"rewards/margins": 10.20118408203125,
"rewards/rejected": -9.165443420410156,
"step": 130
},
{
"epoch": 0.22063157894736843,
"grad_norm": 58.96966604279425,
"kl": 0.0,
"learning_rate": 4.790985327068375e-07,
"logits/chosen": -1307087360.0,
"logits/rejected": -688753612.8,
"logps/chosen": -197.68208821614584,
"logps/rejected": -239.55947265625,
"loss": 0.4189,
"rewards/chosen": 1.5207815170288086,
"rewards/margins": 1.0784048080444335,
"rewards/rejected": 0.442376708984375,
"step": 131
},
{
"epoch": 0.22231578947368422,
"grad_norm": 50.487962918238146,
"kl": 0.0,
"learning_rate": 4.785058505626587e-07,
"logits/chosen": -903821824.0,
"logits/rejected": -479294592.0,
"logps/chosen": -188.7058563232422,
"logps/rejected": -270.5711364746094,
"loss": 0.4645,
"rewards/chosen": 0.9261478781700134,
"rewards/margins": 0.3811134696006775,
"rewards/rejected": 0.5450344085693359,
"step": 132
},
{
"epoch": 0.224,
"grad_norm": 73.6874876394,
"kl": 0.84033203125,
"learning_rate": 4.779052595691354e-07,
"logits/chosen": -728201152.0,
"logits/rejected": -1323703552.0,
"logps/chosen": -219.38204956054688,
"logps/rejected": -226.48712158203125,
"loss": 0.4986,
"rewards/chosen": 0.8902133703231812,
"rewards/margins": -0.04759258031845093,
"rewards/rejected": 0.9378059506416321,
"step": 133
},
{
"epoch": 0.2256842105263158,
"grad_norm": 95.59678991510799,
"kl": 0.3926353454589844,
"learning_rate": 4.772967805134105e-07,
"logits/chosen": -1039820992.0,
"logits/rejected": -1321091328.0,
"logps/chosen": -226.0445556640625,
"logps/rejected": -204.591064453125,
"loss": 0.4261,
"rewards/chosen": 1.1810604333877563,
"rewards/margins": 0.4814262390136719,
"rewards/rejected": 0.6996341943740845,
"step": 134
},
{
"epoch": 0.22736842105263158,
"grad_norm": 188.3750547944905,
"kl": 0.0,
"learning_rate": 4.766804344556413e-07,
"logits/chosen": -612920320.0,
"logits/rejected": -678868736.0,
"logps/chosen": -246.20358276367188,
"logps/rejected": -229.70596313476562,
"loss": 0.4773,
"rewards/chosen": 0.6662670373916626,
"rewards/margins": 1.006878674030304,
"rewards/rejected": -0.34061163663864136,
"step": 135
},
{
"epoch": 0.22905263157894737,
"grad_norm": 68.4177176261016,
"kl": 0.0,
"learning_rate": 4.760562427282712e-07,
"logits/chosen": -829823573.3333334,
"logits/rejected": -959791168.0,
"logps/chosen": -276.5669352213542,
"logps/rejected": -458.56427001953125,
"loss": 0.486,
"rewards/chosen": -0.15483548243840536,
"rewards/margins": 0.1716126004854838,
"rewards/rejected": -0.32644808292388916,
"step": 136
},
{
"epoch": 0.23073684210526316,
"grad_norm": 170.91909658805818,
"kl": 0.0,
"learning_rate": 4.754242269352911e-07,
"logits/chosen": -975248237.7142857,
"logits/rejected": -1017260032.0,
"logps/chosen": -223.60107421875,
"logps/rejected": -248.3441162109375,
"loss": 0.3863,
"rewards/chosen": 1.2192728860037667,
"rewards/margins": 1.134688836123262,
"rewards/rejected": 0.08458404988050461,
"step": 137
},
{
"epoch": 0.23242105263157894,
"grad_norm": 49.863175356012206,
"kl": 0.0,
"learning_rate": 4.747844089514919e-07,
"logits/chosen": -1015484825.6,
"logits/rejected": -859580074.6666666,
"logps/chosen": -238.688232421875,
"logps/rejected": -222.31477864583334,
"loss": 0.4322,
"rewards/chosen": 1.2349286079406738,
"rewards/margins": 1.1146054069201152,
"rewards/rejected": 0.12032320102055867,
"step": 138
},
{
"epoch": 0.23410526315789473,
"grad_norm": 253.28180299915874,
"kl": 2.1038894653320312,
"learning_rate": 4.741368109217071e-07,
"logits/chosen": -1175646293.3333333,
"logits/rejected": -688029491.2,
"logps/chosen": -291.4467366536458,
"logps/rejected": -250.909619140625,
"loss": 0.4026,
"rewards/chosen": 0.8213460445404053,
"rewards/margins": 1.0762538671493531,
"rewards/rejected": -0.25490782260894773,
"step": 139
},
{
"epoch": 0.23578947368421052,
"grad_norm": 80.42815443807443,
"kl": 0.0,
"learning_rate": 4.734814552600468e-07,
"logits/chosen": -934606677.3333334,
"logits/rejected": -1007555904.0,
"logps/chosen": -220.41634114583334,
"logps/rejected": -240.6993865966797,
"loss": 0.4317,
"rewards/chosen": 0.8413275877634684,
"rewards/margins": 0.3180929819742838,
"rewards/rejected": 0.5232346057891846,
"step": 140
},
{
"epoch": 0.2374736842105263,
"grad_norm": 466.33070749527764,
"kl": 0.0,
"learning_rate": 4.728183646491214e-07,
"logits/chosen": -753246873.6,
"logits/rejected": -974831786.6666666,
"logps/chosen": -238.600927734375,
"logps/rejected": -243.89959716796875,
"loss": 0.3808,
"rewards/chosen": -0.32029788494110106,
"rewards/margins": -0.08840243816375731,
"rewards/rejected": -0.23189544677734375,
"step": 141
},
{
"epoch": 0.2391578947368421,
"grad_norm": 176.73411471688468,
"kl": 0.0,
"learning_rate": 4.721475620392567e-07,
"logits/chosen": -908911445.3333334,
"logits/rejected": -713944217.6,
"logps/chosen": -227.62554931640625,
"logps/rejected": -191.32474365234376,
"loss": 0.4569,
"rewards/chosen": 1.7199311256408691,
"rewards/margins": 2.082160711288452,
"rewards/rejected": -0.362229585647583,
"step": 142
},
{
"epoch": 0.24084210526315789,
"grad_norm": 566.7557950897859,
"kl": 0.0,
"learning_rate": 4.7146907064769993e-07,
"logits/chosen": -1072672170.6666666,
"logits/rejected": -1189377228.8,
"logps/chosen": -252.31046549479166,
"logps/rejected": -199.03504638671876,
"loss": 0.4128,
"rewards/chosen": 0.5313369830449423,
"rewards/margins": 1.2499615271886189,
"rewards/rejected": -0.7186245441436767,
"step": 143
},
{
"epoch": 0.24252631578947367,
"grad_norm": 4243.997152918305,
"kl": 0.0,
"learning_rate": 4.707829139578155e-07,
"logits/chosen": -873702195.2,
"logits/rejected": -1221907285.3333333,
"logps/chosen": -200.1794921875,
"logps/rejected": -223.1510009765625,
"loss": 0.4001,
"rewards/chosen": 0.7977983474731445,
"rewards/margins": 2.0752524852752687,
"rewards/rejected": -1.277454137802124,
"step": 144
},
{
"epoch": 0.24421052631578946,
"grad_norm": 314.5163932036576,
"kl": 4.798194885253906,
"learning_rate": 4.7008911571827284e-07,
"logits/chosen": -978259712.0,
"logits/rejected": -640961024.0,
"logps/chosen": -256.1488342285156,
"logps/rejected": -242.27796936035156,
"loss": 0.4857,
"rewards/chosen": 0.40538519620895386,
"rewards/margins": 0.4533770680427551,
"rewards/rejected": -0.04799187183380127,
"step": 145
},
{
"epoch": 0.24589473684210525,
"grad_norm": 360.7422844202394,
"kl": 0.0,
"learning_rate": 4.6938769994222406e-07,
"logits/chosen": -825879756.8,
"logits/rejected": -888731989.3333334,
"logps/chosen": -198.52425537109374,
"logps/rejected": -280.38706461588544,
"loss": 0.358,
"rewards/chosen": 0.7168326377868652,
"rewards/margins": 1.5658982594807944,
"rewards/rejected": -0.8490656216939291,
"step": 146
},
{
"epoch": 0.24757894736842107,
"grad_norm": 2156.8379086317163,
"kl": 0.0,
"learning_rate": 4.686786909064729e-07,
"logits/chosen": -410741248.0,
"logits/rejected": -832765440.0,
"logps/chosen": -249.67108154296875,
"logps/rejected": -203.24400329589844,
"loss": 0.421,
"rewards/chosen": -0.19027066230773926,
"rewards/margins": -0.8781961798667908,
"rewards/rejected": 0.6879255175590515,
"step": 147
},
{
"epoch": 0.24926315789473685,
"grad_norm": 286.56368982318264,
"kl": 0.0,
"learning_rate": 4.679621131506346e-07,
"logits/chosen": -1148589421.7142856,
"logits/rejected": -688305152.0,
"logps/chosen": -267.63023158482144,
"logps/rejected": -254.83946228027344,
"loss": 0.4676,
"rewards/chosen": 0.4722416400909424,
"rewards/margins": 0.7534733414649963,
"rewards/rejected": -0.28123170137405396,
"step": 148
},
{
"epoch": 0.25094736842105264,
"grad_norm": 222.97278644560413,
"kl": 0.0,
"learning_rate": 4.6723799147628664e-07,
"logits/chosen": -779184896.0,
"logits/rejected": -826172416.0,
"logps/chosen": -190.68980407714844,
"logps/rejected": -227.09054565429688,
"loss": 0.4862,
"rewards/chosen": 1.420555830001831,
"rewards/margins": 1.2711764872074127,
"rewards/rejected": 0.14937934279441833,
"step": 149
},
{
"epoch": 0.25263157894736843,
"grad_norm": 145.46733652305187,
"kl": 0.0,
"learning_rate": 4.6650635094610966e-07,
"logits/chosen": -831112106.6666666,
"logits/rejected": -727667648.0,
"logps/chosen": -205.18501790364584,
"logps/rejected": -203.62704467773438,
"loss": 0.4654,
"rewards/chosen": 0.4877975384394328,
"rewards/margins": -0.5469383796056111,
"rewards/rejected": 1.034735918045044,
"step": 150
},
{
"epoch": 0.2543157894736842,
"grad_norm": 148.86465225969857,
"kl": 0.0,
"learning_rate": 4.6576721688302103e-07,
"logits/chosen": -616928128.0,
"logits/rejected": -506641365.3333333,
"logps/chosen": -269.32086181640625,
"logps/rejected": -279.0386962890625,
"loss": 0.4548,
"rewards/chosen": -0.4652397334575653,
"rewards/margins": -0.7438713212807972,
"rewards/rejected": 0.278631587823232,
"step": 151
},
{
"epoch": 0.256,
"grad_norm": 80.8642282340819,
"kl": 0.0,
"learning_rate": 4.6502061486929765e-07,
"logits/chosen": -1172402585.6,
"logits/rejected": -611508778.6666666,
"logps/chosen": -199.3252197265625,
"logps/rejected": -200.5145263671875,
"loss": 0.4583,
"rewards/chosen": 0.7260040283203125,
"rewards/margins": 0.8308873613675436,
"rewards/rejected": -0.10488333304723103,
"step": 152
},
{
"epoch": 0.2576842105263158,
"grad_norm": 59.01380093137268,
"kl": 0.0,
"learning_rate": 4.6426657074569076e-07,
"logits/chosen": -679793049.6,
"logits/rejected": -954248618.6666666,
"logps/chosen": -219.611181640625,
"logps/rejected": -154.1990966796875,
"loss": 0.4257,
"rewards/chosen": 0.9013955116271972,
"rewards/margins": 1.4685372829437255,
"rewards/rejected": -0.5671417713165283,
"step": 153
},
{
"epoch": 0.2593684210526316,
"grad_norm": 165.5135211510967,
"kl": 0.0,
"learning_rate": 4.6350511061053155e-07,
"logits/chosen": -996774144.0,
"logits/rejected": -484326496.0,
"logps/chosen": -222.6114501953125,
"logps/rejected": -187.74560546875,
"loss": 0.4011,
"rewards/chosen": 0.8318793773651123,
"rewards/margins": 1.6771708726882935,
"rewards/rejected": -0.8452914953231812,
"step": 154
},
{
"epoch": 0.26105263157894737,
"grad_norm": 62.878815512005815,
"kl": 0.0,
"learning_rate": 4.62736260818828e-07,
"logits/chosen": -1305949610.6666667,
"logits/rejected": -473263104.0,
"logps/chosen": -258.23980712890625,
"logps/rejected": -231.45439453125,
"loss": 0.4867,
"rewards/chosen": -0.5735707680384318,
"rewards/margins": -0.266325561205546,
"rewards/rejected": -0.30724520683288575,
"step": 155
},
{
"epoch": 0.26273684210526316,
"grad_norm": 68.5671614498506,
"kl": 0.0,
"learning_rate": 4.6196004798135236e-07,
"logits/chosen": -818773568.0,
"logits/rejected": -855193408.0,
"logps/chosen": -297.919677734375,
"logps/rejected": -230.89285278320312,
"loss": 0.4601,
"rewards/chosen": -0.12369692325592041,
"rewards/margins": -0.5421707332134247,
"rewards/rejected": 0.4184738099575043,
"step": 156
},
{
"epoch": 0.26442105263157895,
"grad_norm": 49.45697259448593,
"kl": 0.0,
"learning_rate": 4.611764989637205e-07,
"logits/chosen": -671601715.2,
"logits/rejected": -1087509162.6666667,
"logps/chosen": -242.408056640625,
"logps/rejected": -313.653564453125,
"loss": 0.5013,
"rewards/chosen": 1.0554190635681153,
"rewards/margins": 3.382936731974284,
"rewards/rejected": -2.3275176684061685,
"step": 157
},
{
"epoch": 0.26610526315789473,
"grad_norm": 106.47797133432495,
"kl": 0.0,
"learning_rate": 4.603856408854618e-07,
"logits/chosen": -757610240.0,
"logits/rejected": -791967232.0,
"logps/chosen": -249.16958618164062,
"logps/rejected": -192.21998596191406,
"loss": 0.4991,
"rewards/chosen": 1.1682857275009155,
"rewards/margins": 0.8241084814071655,
"rewards/rejected": 0.34417724609375,
"step": 158
},
{
"epoch": 0.2677894736842105,
"grad_norm": 49.18374091004554,
"kl": 0.0,
"learning_rate": 4.5958750111908065e-07,
"logits/chosen": -1117810432.0,
"logits/rejected": -668232345.6,
"logps/chosen": -165.53057861328125,
"logps/rejected": -253.265234375,
"loss": 0.502,
"rewards/chosen": 0.5383493900299072,
"rewards/margins": 0.5533985137939453,
"rewards/rejected": -0.015049123764038086,
"step": 159
},
{
"epoch": 0.2694736842105263,
"grad_norm": 44.5074703062048,
"kl": 0.0,
"learning_rate": 4.5878210728910886e-07,
"logits/chosen": -953598976.0,
"logits/rejected": -1030626918.4,
"logps/chosen": -242.2763875325521,
"logps/rejected": -223.04658203125,
"loss": 0.4662,
"rewards/chosen": 0.856818675994873,
"rewards/margins": 0.2501978874206543,
"rewards/rejected": 0.6066207885742188,
"step": 160
},
{
"epoch": 0.2711578947368421,
"grad_norm": 75.43811234031128,
"kl": 0.0,
"learning_rate": 4.5796948727115e-07,
"logits/chosen": -793493299.2,
"logits/rejected": -714943658.6666666,
"logps/chosen": -278.5044189453125,
"logps/rejected": -209.54473876953125,
"loss": 0.3873,
"rewards/chosen": 0.2761260747909546,
"rewards/margins": -0.06068220933278401,
"rewards/rejected": 0.3368082841237386,
"step": 161
},
{
"epoch": 0.2728421052631579,
"grad_norm": 63.90407966356232,
"kl": 0.0,
"learning_rate": 4.5714966919091414e-07,
"logits/chosen": -482634784.0,
"logits/rejected": -1263959466.6666667,
"logps/chosen": -282.9561767578125,
"logps/rejected": -233.3080037434896,
"loss": 0.398,
"rewards/chosen": 0.31816405057907104,
"rewards/margins": 0.5229944586753845,
"rewards/rejected": -0.20483040809631348,
"step": 162
},
{
"epoch": 0.2745263157894737,
"grad_norm": 121.01434181632379,
"kl": 0.0,
"learning_rate": 4.5632268142324435e-07,
"logits/chosen": -697656746.6666666,
"logits/rejected": -517255552.0,
"logps/chosen": -242.203369140625,
"logps/rejected": -196.41952514648438,
"loss": 0.3492,
"rewards/chosen": 1.0457820097605388,
"rewards/margins": 1.9965357581774394,
"rewards/rejected": -0.9507537484169006,
"step": 163
},
{
"epoch": 0.27621052631578946,
"grad_norm": 59.7684852435887,
"kl": 0.0,
"learning_rate": 4.554885525911351e-07,
"logits/chosen": -932756352.0,
"logits/rejected": -709267626.6666666,
"logps/chosen": -220.3507537841797,
"logps/rejected": -211.04886881510416,
"loss": 0.4335,
"rewards/chosen": 0.9487717151641846,
"rewards/margins": 2.4863777955373125,
"rewards/rejected": -1.5376060803731282,
"step": 164
},
{
"epoch": 0.27789473684210525,
"grad_norm": 84.16432681430886,
"kl": 0.0,
"learning_rate": 4.546473115647409e-07,
"logits/chosen": -594779008.0,
"logits/rejected": -749184102.4,
"logps/chosen": -238.7244873046875,
"logps/rejected": -268.174755859375,
"loss": 0.42,
"rewards/chosen": -1.2170298099517822,
"rewards/margins": 0.033264207839965865,
"rewards/rejected": -1.250294017791748,
"step": 165
},
{
"epoch": 0.27957894736842104,
"grad_norm": 72.08410795046584,
"kl": 0.0,
"learning_rate": 4.53798987460378e-07,
"logits/chosen": -850566144.0,
"logits/rejected": -682344243.2,
"logps/chosen": -284.1813557942708,
"logps/rejected": -257.943896484375,
"loss": 0.3765,
"rewards/chosen": -0.22490237156550089,
"rewards/margins": 2.2622447689374288,
"rewards/rejected": -2.4871471405029295,
"step": 166
},
{
"epoch": 0.2812631578947368,
"grad_norm": 226.29378772227912,
"kl": 0.0,
"learning_rate": 4.529436096395156e-07,
"logits/chosen": -800250816.0,
"logits/rejected": -1130786816.0,
"logps/chosen": -183.79263305664062,
"logps/rejected": -233.85260009765625,
"loss": 0.4077,
"rewards/chosen": 0.2685844302177429,
"rewards/margins": 0.8706353306770325,
"rewards/rejected": -0.6020509004592896,
"step": 167
},
{
"epoch": 0.2829473684210526,
"grad_norm": 69.313319440612,
"kl": 0.0,
"learning_rate": 4.520812077077604e-07,
"logits/chosen": -840975616.0,
"logits/rejected": -762263040.0,
"logps/chosen": -194.9156005859375,
"logps/rejected": -208.27591959635416,
"loss": 0.5229,
"rewards/chosen": -0.07560654282569886,
"rewards/margins": 0.054520648717880246,
"rewards/rejected": -0.1301271915435791,
"step": 168
},
{
"epoch": 0.2846315789473684,
"grad_norm": 93.08308180288073,
"kl": 0.0,
"learning_rate": 4.512118115138314e-07,
"logits/chosen": -760346965.3333334,
"logits/rejected": -865922432.0,
"logps/chosen": -213.150634765625,
"logps/rejected": -177.36080932617188,
"loss": 0.4687,
"rewards/chosen": 0.10082270701726277,
"rewards/margins": 0.7567354639371237,
"rewards/rejected": -0.6559127569198608,
"step": 169
},
{
"epoch": 0.2863157894736842,
"grad_norm": 91.57290262396198,
"kl": 0.0,
"learning_rate": 4.503354511485273e-07,
"logits/chosen": -858243276.8,
"logits/rejected": -1284581632.0,
"logps/chosen": -279.31884765625,
"logps/rejected": -272.38559977213544,
"loss": 0.4222,
"rewards/chosen": -0.1363077163696289,
"rewards/margins": 0.4786560535430908,
"rewards/rejected": -0.6149637699127197,
"step": 170
},
{
"epoch": 0.288,
"grad_norm": 99.04964646632929,
"kl": 0.0,
"learning_rate": 4.4945215694368447e-07,
"logits/chosen": -328858944.0,
"logits/rejected": -713192234.6666666,
"logps/chosen": -172.78839111328125,
"logps/rejected": -212.23014322916666,
"loss": 0.4366,
"rewards/chosen": 0.2652008533477783,
"rewards/margins": 1.0063419342041016,
"rewards/rejected": -0.7411410808563232,
"step": 171
},
{
"epoch": 0.28968421052631577,
"grad_norm": 89.56438737059334,
"kl": 0.0,
"learning_rate": 4.485619594711277e-07,
"logits/chosen": -598744512.0,
"logits/rejected": -598062378.6666666,
"logps/chosen": -262.1839599609375,
"logps/rejected": -212.77608235677084,
"loss": 0.4281,
"rewards/chosen": 0.616127073764801,
"rewards/margins": 0.38657637437184655,
"rewards/rejected": 0.2295506993929545,
"step": 172
},
{
"epoch": 0.29136842105263155,
"grad_norm": 55.09758282232377,
"kl": 0.0,
"learning_rate": 4.4766488954161153e-07,
"logits/chosen": -804758186.6666666,
"logits/rejected": -721973760.0,
"logps/chosen": -261.65061442057294,
"logps/rejected": -265.42978515625,
"loss": 0.4359,
"rewards/chosen": -0.7668237686157227,
"rewards/margins": 0.3059118270874024,
"rewards/rejected": -1.072735595703125,
"step": 173
},
{
"epoch": 0.29305263157894734,
"grad_norm": 69.82107320148485,
"kl": 0.0,
"learning_rate": 4.4676097820375426e-07,
"logits/chosen": -515698304.0,
"logits/rejected": -1235355648.0,
"logps/chosen": -205.48640950520834,
"logps/rejected": -245.5269287109375,
"loss": 0.4254,
"rewards/chosen": 0.9509898026784261,
"rewards/margins": 0.5845218737920126,
"rewards/rejected": 0.36646792888641355,
"step": 174
},
{
"epoch": 0.29473684210526313,
"grad_norm": 78.35333198019593,
"kl": 0.0,
"learning_rate": 4.458502567429631e-07,
"logits/chosen": -498773248.0,
"logits/rejected": -519963648.0,
"logps/chosen": -194.12052408854166,
"logps/rejected": -172.4505615234375,
"loss": 0.4344,
"rewards/chosen": 0.4248570998509725,
"rewards/margins": 1.5527800162633258,
"rewards/rejected": -1.1279229164123534,
"step": 175
},
{
"epoch": 0.296421052631579,
"grad_norm": 169.14882087581336,
"kl": 0.0,
"learning_rate": 4.4493275668035147e-07,
"logits/chosen": -721955242.6666666,
"logits/rejected": -710825932.8,
"logps/chosen": -167.6915079752604,
"logps/rejected": -147.37952880859376,
"loss": 0.4063,
"rewards/chosen": 0.4422839085261027,
"rewards/margins": 2.1021666447321574,
"rewards/rejected": -1.6598827362060546,
"step": 176
},
{
"epoch": 0.29810526315789476,
"grad_norm": 67.3786479497002,
"kl": 0.0,
"learning_rate": 4.440085097716479e-07,
"logits/chosen": -663990016.0,
"logits/rejected": -555922304.0,
"logps/chosen": -259.0157470703125,
"logps/rejected": -192.321533203125,
"loss": 0.4259,
"rewards/chosen": 0.24189604818820953,
"rewards/margins": 0.6399521082639694,
"rewards/rejected": -0.3980560600757599,
"step": 177
},
{
"epoch": 0.29978947368421055,
"grad_norm": 80.2865086271179,
"kl": 0.0,
"learning_rate": 4.430775480060972e-07,
"logits/chosen": -851462912.0,
"logits/rejected": -1203325952.0,
"logps/chosen": -203.14907836914062,
"logps/rejected": -263.4740905761719,
"loss": 0.3982,
"rewards/chosen": -1.356869101524353,
"rewards/margins": 0.5466030836105347,
"rewards/rejected": -1.9034721851348877,
"step": 178
},
{
"epoch": 0.30147368421052634,
"grad_norm": 84.09627558392572,
"kl": 0.0,
"learning_rate": 4.4213990360535267e-07,
"logits/chosen": -656672042.6666666,
"logits/rejected": -944876748.8,
"logps/chosen": -294.62876383463544,
"logps/rejected": -214.0318115234375,
"loss": 0.4823,
"rewards/chosen": -1.5061303774515789,
"rewards/margins": -1.1466392676035564,
"rewards/rejected": -0.35949110984802246,
"step": 179
},
{
"epoch": 0.3031578947368421,
"grad_norm": 50.8032139364716,
"kl": 0.0,
"learning_rate": 4.4119560902236174e-07,
"logits/chosen": -764519808.0,
"logits/rejected": -615565184.0,
"logps/chosen": -261.0996398925781,
"logps/rejected": -222.42861938476562,
"loss": 0.4579,
"rewards/chosen": -0.5169571042060852,
"rewards/margins": 1.855327069759369,
"rewards/rejected": -2.372284173965454,
"step": 180
},
{
"epoch": 0.3048421052631579,
"grad_norm": 60.14921204443068,
"kl": 0.0,
"learning_rate": 4.402446969402419e-07,
"logits/chosen": -1501441194.6666667,
"logits/rejected": -1463862681.6,
"logps/chosen": -328.10398356119794,
"logps/rejected": -221.4558349609375,
"loss": 0.4195,
"rewards/chosen": -3.202397664388021,
"rewards/margins": -1.721674950917562,
"rewards/rejected": -1.480722713470459,
"step": 181
},
{
"epoch": 0.3065263157894737,
"grad_norm": 57.771134136758484,
"kl": 0.0,
"learning_rate": 4.392872002711501e-07,
"logits/chosen": -459086745.6,
"logits/rejected": -1257842090.6666667,
"logps/chosen": -290.8982421875,
"logps/rejected": -208.20979817708334,
"loss": 0.4709,
"rewards/chosen": -1.0606164932250977,
"rewards/margins": 0.5609116554260254,
"rewards/rejected": -1.621528148651123,
"step": 182
},
{
"epoch": 0.3082105263157895,
"grad_norm": 237.395695237356,
"kl": 0.0,
"learning_rate": 4.3832315215514316e-07,
"logits/chosen": -752865587.2,
"logits/rejected": -800709973.3333334,
"logps/chosen": -222.4587646484375,
"logps/rejected": -164.1585693359375,
"loss": 0.3825,
"rewards/chosen": 0.5814694404602051,
"rewards/margins": 0.6762328982353211,
"rewards/rejected": -0.09476345777511597,
"step": 183
},
{
"epoch": 0.3098947368421053,
"grad_norm": 68.49160205355011,
"kl": 0.0,
"learning_rate": 4.3735258595903123e-07,
"logits/chosen": -1131358412.8,
"logits/rejected": -615695616.0,
"logps/chosen": -263.4253173828125,
"logps/rejected": -251.53666178385416,
"loss": 0.4091,
"rewards/chosen": -0.1255630373954773,
"rewards/margins": 2.537329304218292,
"rewards/rejected": -2.6628923416137695,
"step": 184
},
{
"epoch": 0.31157894736842107,
"grad_norm": 56.555913296788326,
"kl": 0.0,
"learning_rate": 4.3637553527522265e-07,
"logits/chosen": -537920128.0,
"logits/rejected": -931403434.6666666,
"logps/chosen": -261.276611328125,
"logps/rejected": -258.6404622395833,
"loss": 0.4377,
"rewards/chosen": -0.8533248901367188,
"rewards/margins": -0.1802154382069906,
"rewards/rejected": -0.6731094519297282,
"step": 185
},
{
"epoch": 0.31326315789473685,
"grad_norm": 62.41772660361089,
"kl": 0.0,
"learning_rate": 4.353920339205611e-07,
"logits/chosen": -722902016.0,
"logits/rejected": -948598144.0,
"logps/chosen": -224.0010223388672,
"logps/rejected": -272.22760009765625,
"loss": 0.4404,
"rewards/chosen": 0.367422878742218,
"rewards/margins": 0.9069034457206726,
"rewards/rejected": -0.5394805669784546,
"step": 186
},
{
"epoch": 0.31494736842105264,
"grad_norm": 64.75287382395764,
"kl": 0.0,
"learning_rate": 4.344021159351555e-07,
"logits/chosen": -1008326553.6,
"logits/rejected": -771843413.3333334,
"logps/chosen": -227.4381591796875,
"logps/rejected": -186.0328369140625,
"loss": 0.4424,
"rewards/chosen": -0.36620848178863524,
"rewards/margins": -0.2139639139175415,
"rewards/rejected": -0.15224456787109375,
"step": 187
},
{
"epoch": 0.31663157894736843,
"grad_norm": 70.9290525757904,
"kl": 0.0,
"learning_rate": 4.3340581558120195e-07,
"logits/chosen": -618546688.0,
"logits/rejected": -912823125.3333334,
"logps/chosen": -246.94599609375,
"logps/rejected": -296.3729654947917,
"loss": 0.3832,
"rewards/chosen": -0.18909727334976195,
"rewards/margins": 3.131844015916189,
"rewards/rejected": -3.3209412892659507,
"step": 188
},
{
"epoch": 0.3183157894736842,
"grad_norm": 70.41433062876469,
"kl": 0.0,
"learning_rate": 4.3240316734179705e-07,
"logits/chosen": -761420928.0,
"logits/rejected": -672102144.0,
"logps/chosen": -227.72821044921875,
"logps/rejected": -258.29815673828125,
"loss": 0.4255,
"rewards/chosen": 0.08762133121490479,
"rewards/margins": 3.5601266622543335,
"rewards/rejected": -3.4725053310394287,
"step": 189
},
{
"epoch": 0.32,
"grad_norm": 66.55630414028238,
"kl": 0.0,
"learning_rate": 4.313942059197456e-07,
"logits/chosen": -618079808.0,
"logits/rejected": -658955968.0,
"logps/chosen": -252.71575927734375,
"logps/rejected": -140.86489868164062,
"loss": 0.4927,
"rewards/chosen": -0.4330390989780426,
"rewards/margins": -0.2638881802558899,
"rewards/rejected": -0.1691509187221527,
"step": 190
},
{
"epoch": 0.3216842105263158,
"grad_norm": 94.67677667797632,
"kl": 0.0,
"learning_rate": 4.3037896623635864e-07,
"logits/chosen": -923324928.0,
"logits/rejected": -594331520.0,
"logps/chosen": -247.37379455566406,
"logps/rejected": -226.16342163085938,
"loss": 0.367,
"rewards/chosen": 0.9253654479980469,
"rewards/margins": 2.8820842504501343,
"rewards/rejected": -1.9567188024520874,
"step": 191
},
{
"epoch": 0.3233684210526316,
"grad_norm": 54.17392420976664,
"kl": 0.0,
"learning_rate": 4.29357483430245e-07,
"logits/chosen": -664615808.0,
"logits/rejected": -569389952.0,
"logps/chosen": -216.7957763671875,
"logps/rejected": -197.62344360351562,
"loss": 0.4826,
"rewards/chosen": -0.4946579138437907,
"rewards/margins": 0.1053409179051717,
"rewards/rejected": -0.5999988317489624,
"step": 192
},
{
"epoch": 0.32505263157894737,
"grad_norm": 75.00984735078153,
"kl": 1.3663177490234375,
"learning_rate": 4.2832979285609505e-07,
"logits/chosen": -614758997.3333334,
"logits/rejected": -654426368.0,
"logps/chosen": -235.02339680989584,
"logps/rejected": -149.56541442871094,
"loss": 0.3863,
"rewards/chosen": 0.690934975941976,
"rewards/margins": 0.510087807973226,
"rewards/rejected": 0.18084716796875,
"step": 193
},
{
"epoch": 0.32673684210526316,
"grad_norm": 60.51774121025042,
"kl": 0.0,
"learning_rate": 4.2729593008345734e-07,
"logits/chosen": -1201705642.6666667,
"logits/rejected": -741241139.2,
"logps/chosen": -172.75325520833334,
"logps/rejected": -184.46663818359374,
"loss": 0.3963,
"rewards/chosen": 1.0127630233764648,
"rewards/margins": 2.2689538955688477,
"rewards/rejected": -1.256190872192383,
"step": 194
},
{
"epoch": 0.32842105263157895,
"grad_norm": 90.59367998785133,
"kl": 0.0,
"learning_rate": 4.2625593089550717e-07,
"logits/chosen": -541644697.6,
"logits/rejected": -1247126954.6666667,
"logps/chosen": -258.279150390625,
"logps/rejected": -216.1505126953125,
"loss": 0.458,
"rewards/chosen": 0.21086244583129882,
"rewards/margins": 0.3597394009431203,
"rewards/rejected": -0.1488769551118215,
"step": 195
},
{
"epoch": 0.33010526315789473,
"grad_norm": 95.59189487331257,
"kl": 0.0,
"learning_rate": 4.2520983128780825e-07,
"logits/chosen": -635985920.0,
"logits/rejected": -922811136.0,
"logps/chosen": -279.5119323730469,
"logps/rejected": -208.4548543294271,
"loss": 0.3814,
"rewards/chosen": -0.1562713384628296,
"rewards/margins": 1.136791984240214,
"rewards/rejected": -1.2930633227030437,
"step": 196
},
{
"epoch": 0.3317894736842105,
"grad_norm": 54.38017510634436,
"kl": 0.0,
"learning_rate": 4.2415766746706674e-07,
"logits/chosen": -605963136.0,
"logits/rejected": -808883507.2,
"logps/chosen": -224.5291544596354,
"logps/rejected": -254.458984375,
"loss": 0.3932,
"rewards/chosen": -1.42087189356486,
"rewards/margins": 1.0473615964253742,
"rewards/rejected": -2.4682334899902343,
"step": 197
},
{
"epoch": 0.3334736842105263,
"grad_norm": 77.91234986865499,
"kl": 0.0,
"learning_rate": 4.230994758498782e-07,
"logits/chosen": -977752576.0,
"logits/rejected": -573459456.0,
"logps/chosen": -287.57672119140625,
"logps/rejected": -200.0487823486328,
"loss": 0.4666,
"rewards/chosen": -0.592963457107544,
"rewards/margins": -0.5604517459869385,
"rewards/rejected": -0.03251171112060547,
"step": 198
},
{
"epoch": 0.3351578947368421,
"grad_norm": 217.16301375592573,
"kl": 0.0,
"learning_rate": 4.220352930614672e-07,
"logits/chosen": -1347623082.6666667,
"logits/rejected": -1223722905.6,
"logps/chosen": -250.11665852864584,
"logps/rejected": -210.2994140625,
"loss": 0.3773,
"rewards/chosen": 0.5189147790273031,
"rewards/margins": 0.7931217749913534,
"rewards/rejected": -0.2742069959640503,
"step": 199
},
{
"epoch": 0.3368421052631579,
"grad_norm": 58.97358660347629,
"kl": 0.0,
"learning_rate": 4.209651559344194e-07,
"logits/chosen": -1134724096.0,
"logits/rejected": -713243093.3333334,
"logps/chosen": -202.700537109375,
"logps/rejected": -273.8854573567708,
"loss": 0.4269,
"rewards/chosen": 0.44961090087890626,
"rewards/margins": 0.7851913293202718,
"rewards/rejected": -0.33558042844136554,
"step": 200
},
{
"epoch": 0.3385263157894737,
"grad_norm": 72.17373452754133,
"kl": 0.0,
"learning_rate": 4.198891015074073e-07,
"logits/chosen": -646841984.0,
"logits/rejected": -952984896.0,
"logps/chosen": -190.56002807617188,
"logps/rejected": -252.1724090576172,
"loss": 0.4577,
"rewards/chosen": -0.3583839535713196,
"rewards/margins": -1.000853717327118,
"rewards/rejected": 0.6424697637557983,
"step": 201
},
{
"epoch": 0.34021052631578946,
"grad_norm": 80.45111424098052,
"kl": 0.0,
"learning_rate": 4.1880716702390763e-07,
"logits/chosen": -1190700544.0,
"logits/rejected": -969268480.0,
"logps/chosen": -192.81265258789062,
"logps/rejected": -200.2042236328125,
"loss": 0.414,
"rewards/chosen": 1.0531678199768066,
"rewards/margins": 3.0669215520222983,
"rewards/rejected": -2.0137537320454917,
"step": 202
},
{
"epoch": 0.34189473684210525,
"grad_norm": 94.02730380206307,
"kl": 0.0,
"learning_rate": 4.177193899309126e-07,
"logits/chosen": -503011993.6,
"logits/rejected": -704219050.6666666,
"logps/chosen": -218.402734375,
"logps/rejected": -179.79007975260416,
"loss": 0.4332,
"rewards/chosen": 0.8202567100524902,
"rewards/margins": 1.5902131398518882,
"rewards/rejected": -0.7699564297993978,
"step": 203
},
{
"epoch": 0.34357894736842104,
"grad_norm": 54.77503186544864,
"kl": 0.0,
"learning_rate": 4.166258078776341e-07,
"logits/chosen": -832879786.6666666,
"logits/rejected": -515199590.4,
"logps/chosen": -202.35528564453125,
"logps/rejected": -200.24393310546876,
"loss": 0.4434,
"rewards/chosen": 0.5717076857884725,
"rewards/margins": 2.901677473386129,
"rewards/rejected": -2.329969787597656,
"step": 204
},
{
"epoch": 0.3452631578947368,
"grad_norm": 148.39721947330267,
"kl": 0.0,
"learning_rate": 4.155264587142001e-07,
"logits/chosen": -1255011840.0,
"logits/rejected": -1037022464.0,
"logps/chosen": -242.1149139404297,
"logps/rejected": -220.04427083333334,
"loss": 0.5528,
"rewards/chosen": -0.22197723388671875,
"rewards/margins": -0.06859992941220602,
"rewards/rejected": -0.15337730447451273,
"step": 205
},
{
"epoch": 0.3469473684210526,
"grad_norm": 49.784955690964,
"kl": 0.0,
"learning_rate": 4.1442138049034494e-07,
"logits/chosen": -775684266.6666666,
"logits/rejected": -880259379.2,
"logps/chosen": -191.68634033203125,
"logps/rejected": -225.9909912109375,
"loss": 0.3432,
"rewards/chosen": 1.5534348487854004,
"rewards/margins": 3.249916934967041,
"rewards/rejected": -1.6964820861816405,
"step": 206
},
{
"epoch": 0.3486315789473684,
"grad_norm": 91.02392105011648,
"kl": 0.0,
"learning_rate": 4.133106114540923e-07,
"logits/chosen": -971414336.0,
"logits/rejected": -1156522368.0,
"logps/chosen": -213.8065185546875,
"logps/rejected": -225.66744995117188,
"loss": 0.3929,
"rewards/chosen": -0.12490347027778625,
"rewards/margins": 0.5315063297748566,
"rewards/rejected": -0.6564098000526428,
"step": 207
},
{
"epoch": 0.3503157894736842,
"grad_norm": 68.74268293327242,
"kl": 0.0,
"learning_rate": 4.1219419005043154e-07,
"logits/chosen": -916664064.0,
"logits/rejected": -622102208.0,
"logps/chosen": -237.611083984375,
"logps/rejected": -284.42645263671875,
"loss": 0.5005,
"rewards/chosen": 0.0357242226600647,
"rewards/margins": 0.9489158391952515,
"rewards/rejected": -0.9131916165351868,
"step": 208
},
{
"epoch": 0.352,
"grad_norm": 76.77975476469975,
"kl": 0.0,
"learning_rate": 4.110721549199866e-07,
"logits/chosen": -1072110400.0,
"logits/rejected": -753468608.0,
"logps/chosen": -184.95042419433594,
"logps/rejected": -249.1209716796875,
"loss": 0.3055,
"rewards/chosen": 0.7373764514923096,
"rewards/margins": 2.8142330646514893,
"rewards/rejected": -2.0768566131591797,
"step": 209
},
{
"epoch": 0.35368421052631577,
"grad_norm": 80.64382792874225,
"kl": 0.0,
"learning_rate": 4.0994454489767927e-07,
"logits/chosen": -689984256.0,
"logits/rejected": -1297480448.0,
"logps/chosen": -210.7186279296875,
"logps/rejected": -249.07969665527344,
"loss": 0.3988,
"rewards/chosen": 0.8691381613413492,
"rewards/margins": 4.103166739145915,
"rewards/rejected": -3.2340285778045654,
"step": 210
},
{
"epoch": 0.35536842105263156,
"grad_norm": 149.68736456789574,
"kl": 0.0,
"learning_rate": 4.088113990113846e-07,
"logits/chosen": -731458048.0,
"logits/rejected": -345419264.0,
"logps/chosen": -280.026025390625,
"logps/rejected": -250.71551513671875,
"loss": 0.3742,
"rewards/chosen": 0.011942720413208008,
"rewards/margins": 0.32606066862742106,
"rewards/rejected": -0.31411794821421307,
"step": 211
},
{
"epoch": 0.35705263157894734,
"grad_norm": 88.38448491891266,
"kl": 0.0,
"learning_rate": 4.076727564805802e-07,
"logits/chosen": -1213323571.2,
"logits/rejected": -650571477.3333334,
"logps/chosen": -225.9,
"logps/rejected": -215.1610107421875,
"loss": 0.3696,
"rewards/chosen": 1.0948437690734862,
"rewards/margins": 1.9371202627817787,
"rewards/rejected": -0.8422764937082926,
"step": 212
},
{
"epoch": 0.35873684210526313,
"grad_norm": 65.08222255296863,
"kl": 0.0,
"learning_rate": 4.0652865671498906e-07,
"logits/chosen": -1109334869.3333333,
"logits/rejected": -673303040.0,
"logps/chosen": -209.34431966145834,
"logps/rejected": -264.689892578125,
"loss": 0.4692,
"rewards/chosen": 0.025906383991241455,
"rewards/margins": 0.10383939743041992,
"rewards/rejected": -0.07793301343917847,
"step": 213
},
{
"epoch": 0.3604210526315789,
"grad_norm": 95.42507509421407,
"kl": 0.0,
"learning_rate": 4.053791393132149e-07,
"logits/chosen": -2192135936.0,
"logits/rejected": -679301717.3333334,
"logps/chosen": -280.882080078125,
"logps/rejected": -235.8901570638021,
"loss": 0.3862,
"rewards/chosen": 1.1142441034317017,
"rewards/margins": 1.286988377571106,
"rewards/rejected": -0.1727442741394043,
"step": 214
},
{
"epoch": 0.36210526315789476,
"grad_norm": 57.980133275698655,
"kl": 0.0,
"learning_rate": 4.0422424406137235e-07,
"logits/chosen": -533541728.0,
"logits/rejected": -664400128.0,
"logps/chosen": -241.3803253173828,
"logps/rejected": -255.393310546875,
"loss": 0.378,
"rewards/chosen": 0.4608551263809204,
"rewards/margins": 2.0191084146499634,
"rewards/rejected": -1.558253288269043,
"step": 215
},
{
"epoch": 0.36378947368421055,
"grad_norm": 62.774785038898976,
"kl": 0.0,
"learning_rate": 4.0306401093170956e-07,
"logits/chosen": -863296819.2,
"logits/rejected": -931824469.3333334,
"logps/chosen": -209.0201171875,
"logps/rejected": -258.80812581380206,
"loss": 0.4346,
"rewards/chosen": -0.07239959239959717,
"rewards/margins": -0.04090743859608969,
"rewards/rejected": -0.031492153803507485,
"step": 216
},
{
"epoch": 0.36547368421052634,
"grad_norm": 43.09358868475917,
"kl": 0.0,
"learning_rate": 4.0189848008122475e-07,
"logits/chosen": -832388096.0,
"logits/rejected": -767827328.0,
"logps/chosen": -231.43902587890625,
"logps/rejected": -311.52874755859375,
"loss": 0.3487,
"rewards/chosen": 0.1406116485595703,
"rewards/margins": 1.9062778949737549,
"rewards/rejected": -1.7656662464141846,
"step": 217
},
{
"epoch": 0.3671578947368421,
"grad_norm": 65.51992296550779,
"kl": 0.0,
"learning_rate": 4.007276918502763e-07,
"logits/chosen": -1086964224.0,
"logits/rejected": -893988352.0,
"logps/chosen": -246.42593383789062,
"logps/rejected": -244.95123291015625,
"loss": 0.3927,
"rewards/chosen": 0.23493842780590057,
"rewards/margins": 1.4225323647260666,
"rewards/rejected": -1.187593936920166,
"step": 218
},
{
"epoch": 0.3688421052631579,
"grad_norm": 88.60132651868206,
"kl": 0.0,
"learning_rate": 3.9955168676118645e-07,
"logits/chosen": -698185813.3333334,
"logits/rejected": -1118987673.6,
"logps/chosen": -195.71712239583334,
"logps/rejected": -205.612646484375,
"loss": 0.5847,
"rewards/chosen": 0.2993367513020833,
"rewards/margins": -0.6073127428690592,
"rewards/rejected": 0.9066494941711426,
"step": 219
},
{
"epoch": 0.3705263157894737,
"grad_norm": 49.2393342246424,
"kl": 0.0,
"learning_rate": 3.9837050551683904e-07,
"logits/chosen": -615184170.6666666,
"logits/rejected": -701940121.6,
"logps/chosen": -227.86104329427084,
"logps/rejected": -263.94287109375,
"loss": 0.3788,
"rewards/chosen": 0.2989359696706136,
"rewards/margins": 2.076623805363973,
"rewards/rejected": -1.7776878356933594,
"step": 220
},
{
"epoch": 0.3722105263157895,
"grad_norm": 186.40613554967234,
"kl": 0.0,
"learning_rate": 3.9718418899927056e-07,
"logits/chosen": -326911872.0,
"logits/rejected": -801610093.7142857,
"logps/chosen": -450.5743408203125,
"logps/rejected": -212.93069893973214,
"loss": 0.485,
"rewards/chosen": -2.387774705886841,
"rewards/margins": -2.170388392039708,
"rewards/rejected": -0.21738631384713308,
"step": 221
},
{
"epoch": 0.3738947368421053,
"grad_norm": 64.26673569575549,
"kl": 0.0,
"learning_rate": 3.959927782682551e-07,
"logits/chosen": -726571008.0,
"logits/rejected": -706536896.0,
"logps/chosen": -230.51875813802084,
"logps/rejected": -126.30707550048828,
"loss": 0.4306,
"rewards/chosen": -0.7994547684987386,
"rewards/margins": -1.1165289779504142,
"rewards/rejected": 0.3170742094516754,
"step": 222
},
{
"epoch": 0.37557894736842107,
"grad_norm": 73.20938961913613,
"kl": 0.0,
"learning_rate": 3.947963145598833e-07,
"logits/chosen": -1290946048.0,
"logits/rejected": -1101487957.3333333,
"logps/chosen": -192.78887939453125,
"logps/rejected": -246.4966023763021,
"loss": 0.4078,
"rewards/chosen": 1.7199859619140625,
"rewards/margins": 2.088035305341085,
"rewards/rejected": -0.3680493434270223,
"step": 223
},
{
"epoch": 0.37726315789473686,
"grad_norm": 47.763563197737376,
"kl": 0.0,
"learning_rate": 3.935948392851353e-07,
"logits/chosen": -2707193088.0,
"logits/rejected": -1093253997.7142856,
"logps/chosen": -202.72845458984375,
"logps/rejected": -189.14756556919642,
"loss": 0.3882,
"rewards/chosen": -2.7178618907928467,
"rewards/margins": -2.8144114358084544,
"rewards/rejected": 0.09654954501560756,
"step": 224
},
{
"epoch": 0.37894736842105264,
"grad_norm": 67.91419872933243,
"kl": 0.0,
"learning_rate": 3.923883940284472e-07,
"logits/chosen": -493479210.6666667,
"logits/rejected": -663955968.0,
"logps/chosen": -208.96756998697916,
"logps/rejected": -219.612744140625,
"loss": 0.4068,
"rewards/chosen": 0.5567906697591146,
"rewards/margins": 2.475810750325521,
"rewards/rejected": -1.9190200805664062,
"step": 225
},
{
"epoch": 0.38063157894736843,
"grad_norm": 45.896671613894135,
"kl": 0.0,
"learning_rate": 3.9117702054627164e-07,
"logits/chosen": -843394901.3333334,
"logits/rejected": -696314688.0,
"logps/chosen": -184.59212239583334,
"logps/rejected": -289.22589111328125,
"loss": 0.3531,
"rewards/chosen": 1.452480634053548,
"rewards/margins": 4.889726003011067,
"rewards/rejected": -3.4372453689575195,
"step": 226
},
{
"epoch": 0.3823157894736842,
"grad_norm": 117.9741866381254,
"kl": 0.0,
"learning_rate": 3.8996076076563333e-07,
"logits/chosen": -633168768.0,
"logits/rejected": -1228658005.3333333,
"logps/chosen": -255.6255645751953,
"logps/rejected": -198.0909423828125,
"loss": 0.5188,
"rewards/chosen": -0.49002379179000854,
"rewards/margins": -1.5519655744234722,
"rewards/rejected": 1.0619417826334636,
"step": 227
},
{
"epoch": 0.384,
"grad_norm": 59.87930468202007,
"kl": 0.0,
"learning_rate": 3.8873965678267686e-07,
"logits/chosen": -748600405.3333334,
"logits/rejected": -564317888.0,
"logps/chosen": -205.5136515299479,
"logps/rejected": -190.98033142089844,
"loss": 0.4398,
"rewards/chosen": 0.013506333033243815,
"rewards/margins": 0.8160205880800883,
"rewards/rejected": -0.8025142550468445,
"step": 228
},
{
"epoch": 0.3856842105263158,
"grad_norm": 55.75368595740401,
"kl": 0.0,
"learning_rate": 3.8751375086121027e-07,
"logits/chosen": -689110741.3333334,
"logits/rejected": -534848768.0,
"logps/chosen": -284.23940022786456,
"logps/rejected": -163.17276000976562,
"loss": 0.4007,
"rewards/chosen": 0.30055542786916095,
"rewards/margins": 0.026084144910176577,
"rewards/rejected": 0.2744712829589844,
"step": 229
},
{
"epoch": 0.3873684210526316,
"grad_norm": 85.00209812207287,
"kl": 0.0,
"learning_rate": 3.8628308543124264e-07,
"logits/chosen": -733420748.8,
"logits/rejected": -613972394.6666666,
"logps/chosen": -240.112109375,
"logps/rejected": -277.9342854817708,
"loss": 0.4103,
"rewards/chosen": 0.4192187309265137,
"rewards/margins": 0.19571206172307334,
"rewards/rejected": 0.22350666920344034,
"step": 230
},
{
"epoch": 0.38905263157894737,
"grad_norm": 67.82002108621575,
"kl": 0.0,
"learning_rate": 3.8504770308751465e-07,
"logits/chosen": -815699712.0,
"logits/rejected": -538265600.0,
"logps/chosen": -195.72711181640625,
"logps/rejected": -168.96177673339844,
"loss": 0.3659,
"rewards/chosen": 0.9801532030105591,
"rewards/margins": 0.6151527762413025,
"rewards/rejected": 0.3650004267692566,
"step": 231
},
{
"epoch": 0.39073684210526316,
"grad_norm": 77.76306759423939,
"kl": 0.0,
"learning_rate": 3.8380764658802476e-07,
"logits/chosen": -863703961.6,
"logits/rejected": -500137898.6666667,
"logps/chosen": -190.1537353515625,
"logps/rejected": -248.5384318033854,
"loss": 0.3315,
"rewards/chosen": 0.9144930839538574,
"rewards/margins": 1.550446311632792,
"rewards/rejected": -0.6359532276789347,
"step": 232
},
{
"epoch": 0.39242105263157895,
"grad_norm": 66.37493353593294,
"kl": 0.0,
"learning_rate": 3.8256295885254977e-07,
"logits/chosen": -929981747.2,
"logits/rejected": -953465258.6666666,
"logps/chosen": -216.4085205078125,
"logps/rejected": -182.04376220703125,
"loss": 0.3813,
"rewards/chosen": 1.277543067932129,
"rewards/margins": 1.0003930727640789,
"rewards/rejected": 0.2771499951680501,
"step": 233
},
{
"epoch": 0.39410526315789474,
"grad_norm": 75.1964810655982,
"kl": 0.0,
"learning_rate": 3.8131368296115823e-07,
"logits/chosen": -896362752.0,
"logits/rejected": -561620096.0,
"logps/chosen": -249.6129353841146,
"logps/rejected": -264.5303955078125,
"loss": 0.2827,
"rewards/chosen": 1.139868974685669,
"rewards/margins": 2.838852047920227,
"rewards/rejected": -1.698983073234558,
"step": 234
},
{
"epoch": 0.3957894736842105,
"grad_norm": 122.44554306763347,
"kl": 0.0,
"learning_rate": 3.800598621527205e-07,
"logits/chosen": -969984146.2857143,
"logits/rejected": -854562432.0,
"logps/chosen": -182.46547154017858,
"logps/rejected": -194.51100158691406,
"loss": 0.3864,
"rewards/chosen": 1.404153551374163,
"rewards/margins": 1.3077897163374084,
"rewards/rejected": 0.09636383503675461,
"step": 235
},
{
"epoch": 0.3974736842105263,
"grad_norm": 60.34196841703653,
"kl": 0.0,
"learning_rate": 3.7880153982341163e-07,
"logits/chosen": -823134464.0,
"logits/rejected": -873622101.3333334,
"logps/chosen": -235.61932373046875,
"logps/rejected": -271.49196370442706,
"loss": 0.3463,
"rewards/chosen": 2.0248725414276123,
"rewards/margins": 3.3654048442840576,
"rewards/rejected": -1.3405323028564453,
"step": 236
},
{
"epoch": 0.3991578947368421,
"grad_norm": 73.61126053660914,
"kl": 0.0,
"learning_rate": 3.7753875952520937e-07,
"logits/chosen": -791607424.0,
"logits/rejected": -736724992.0,
"logps/chosen": -260.60198974609375,
"logps/rejected": -270.4560852050781,
"loss": 0.3967,
"rewards/chosen": 0.33294641971588135,
"rewards/margins": 2.238220691680908,
"rewards/rejected": -1.9052742719650269,
"step": 237
},
{
"epoch": 0.4008421052631579,
"grad_norm": 44.26653587553288,
"kl": 0.0,
"learning_rate": 3.7627156496438685e-07,
"logits/chosen": -824610218.6666666,
"logits/rejected": -856471756.8,
"logps/chosen": -174.15934244791666,
"logps/rejected": -221.6721435546875,
"loss": 0.4355,
"rewards/chosen": 1.1449518998463948,
"rewards/margins": 0.593308385213216,
"rewards/rejected": 0.5516435146331787,
"step": 238
},
{
"epoch": 0.4025263157894737,
"grad_norm": 47.69085670063618,
"kl": 0.0,
"learning_rate": 3.75e-07,
"logits/chosen": -820081216.0,
"logits/rejected": -955681408.0,
"logps/chosen": -224.08152770996094,
"logps/rejected": -240.1038818359375,
"loss": 0.3694,
"rewards/chosen": 1.8589047193527222,
"rewards/margins": 3.0458600521087646,
"rewards/rejected": -1.1869553327560425,
"step": 239
},
{
"epoch": 0.40421052631578946,
"grad_norm": 48.281804889717954,
"kl": 0.0,
"learning_rate": 3.7372410864236947e-07,
"logits/chosen": -717564245.3333334,
"logits/rejected": -679164364.8,
"logps/chosen": -152.29527791341147,
"logps/rejected": -278.3427490234375,
"loss": 0.3254,
"rewards/chosen": 1.1597716808319092,
"rewards/margins": 3.844050931930542,
"rewards/rejected": -2.6842792510986326,
"step": 240
},
{
"epoch": 0.40589473684210525,
"grad_norm": 75.76751371273427,
"kl": 0.0,
"learning_rate": 3.724439350515571e-07,
"logits/chosen": -629566144.0,
"logits/rejected": -622408640.0,
"logps/chosen": -201.63623046875,
"logps/rejected": -248.00782775878906,
"loss": 0.3847,
"rewards/chosen": 0.5420559048652649,
"rewards/margins": 1.8232646584510803,
"rewards/rejected": -1.2812087535858154,
"step": 241
},
{
"epoch": 0.40757894736842104,
"grad_norm": 95.46822610468628,
"kl": 0.0,
"learning_rate": 3.7115952353583803e-07,
"logits/chosen": -702194048.0,
"logps/chosen": -212.0802459716797,
"loss": 0.3904,
"rewards/chosen": 1.2338132858276367,
"step": 242
},
{
"epoch": 0.40926315789473683,
"grad_norm": 86.1205385319129,
"kl": 0.0,
"learning_rate": 3.698709185501666e-07,
"logits/chosen": -567989094.4,
"logits/rejected": -984176554.6666666,
"logps/chosen": -259.487060546875,
"logps/rejected": -214.1506144205729,
"loss": 0.2951,
"rewards/chosen": 0.584656047821045,
"rewards/margins": 1.1261462450027466,
"rewards/rejected": -0.5414901971817017,
"step": 243
},
{
"epoch": 0.4109473684210526,
"grad_norm": 94.11033116313197,
"kl": 0.0,
"learning_rate": 3.6857816469463805e-07,
"logits/chosen": -414785728.0,
"logits/rejected": -834902308.5714285,
"logps/chosen": -117.85265350341797,
"logps/rejected": -237.00685337611608,
"loss": 0.4324,
"rewards/chosen": 1.1912513971328735,
"rewards/margins": 1.5301621811730521,
"rewards/rejected": -0.33891078404017855,
"step": 244
},
{
"epoch": 0.4126315789473684,
"grad_norm": 73.01801229445103,
"kl": 0.0,
"learning_rate": 3.6728130671294484e-07,
"logits/chosen": -480209600.0,
"logits/rejected": -1261644629.3333333,
"logps/chosen": -242.870361328125,
"logps/rejected": -297.1340738932292,
"loss": 0.3948,
"rewards/chosen": 1.906396508216858,
"rewards/margins": 3.281412720680237,
"rewards/rejected": -1.375016212463379,
"step": 245
},
{
"epoch": 0.4143157894736842,
"grad_norm": 50.21406501502345,
"kl": 0.0,
"learning_rate": 3.6598038949082773e-07,
"logits/chosen": -1082578688.0,
"logits/rejected": -1054340505.6,
"logps/chosen": -202.77239990234375,
"logps/rejected": -241.3439697265625,
"loss": 0.4015,
"rewards/chosen": 0.08610687653223674,
"rewards/margins": -0.06312040885289509,
"rewards/rejected": 0.14922728538513183,
"step": 246
},
{
"epoch": 0.416,
"grad_norm": 69.90136809562398,
"kl": 0.0,
"learning_rate": 3.646754580545226e-07,
"logits/chosen": -551080704.0,
"logits/rejected": -721525162.6666666,
"logps/chosen": -253.119091796875,
"logps/rejected": -209.84647623697916,
"loss": 0.4882,
"rewards/chosen": -0.7751388549804688,
"rewards/margins": -1.7328248023986816,
"rewards/rejected": 0.9576859474182129,
"step": 247
},
{
"epoch": 0.41768421052631577,
"grad_norm": 66.55136067755517,
"kl": 0.0,
"learning_rate": 3.633665575692019e-07,
"logits/chosen": -469412010.6666667,
"logits/rejected": -688726988.8,
"logps/chosen": -186.66743977864584,
"logps/rejected": -300.4759765625,
"loss": 0.3747,
"rewards/chosen": 0.8189966678619385,
"rewards/margins": 2.057415819168091,
"rewards/rejected": -1.2384191513061524,
"step": 248
},
{
"epoch": 0.41936842105263156,
"grad_norm": 91.34875948495586,
"kl": 0.0,
"learning_rate": 3.6205373333741137e-07,
"logits/chosen": -695206528.0,
"logits/rejected": -909713856.0,
"logps/chosen": -220.77886962890625,
"logps/rejected": -275.035400390625,
"loss": 0.4002,
"rewards/chosen": 0.9514415860176086,
"rewards/margins": 2.3362786173820496,
"rewards/rejected": -1.384837031364441,
"step": 249
},
{
"epoch": 0.42105263157894735,
"grad_norm": 41.18603961196617,
"kl": 0.0,
"learning_rate": 3.60737030797502e-07,
"logits/chosen": -578035840.0,
"logits/rejected": -939070378.6666666,
"logps/chosen": -261.39300537109375,
"logps/rejected": -181.3475341796875,
"loss": 0.4095,
"rewards/chosen": -1.6891616582870483,
"rewards/margins": -1.0875247716903687,
"rewards/rejected": -0.6016368865966797,
"step": 250
},
{
"epoch": 0.42273684210526313,
"grad_norm": 116.0424680075244,
"kl": 0.0,
"learning_rate": 3.5941649552205766e-07,
"logits/chosen": -1035228672.0,
"logits/rejected": -634597068.8,
"logps/chosen": -246.8405558268229,
"logps/rejected": -214.5775390625,
"loss": 0.4228,
"rewards/chosen": -0.7042236328125,
"rewards/margins": -1.805875873565674,
"rewards/rejected": 1.101652240753174,
"step": 251
},
{
"epoch": 0.4244210526315789,
"grad_norm": 57.738431491671584,
"kl": 0.0,
"learning_rate": 3.580921732163174e-07,
"logits/chosen": -844649472.0,
"logits/rejected": -548579840.0,
"logps/chosen": -214.6180419921875,
"logps/rejected": -231.5613250732422,
"loss": 0.3853,
"rewards/chosen": 0.8030085563659668,
"rewards/margins": 1.5388163328170776,
"rewards/rejected": -0.7358077764511108,
"step": 252
},
{
"epoch": 0.4261052631578947,
"grad_norm": 346.5318458002952,
"kl": 0.0,
"learning_rate": 3.56764109716594e-07,
"logits/chosen": -551624384.0,
"logits/rejected": -1136507611.4285715,
"logps/chosen": -178.53115844726562,
"logps/rejected": -271.80191476004467,
"loss": 0.3623,
"rewards/chosen": 3.3293793201446533,
"rewards/margins": 8.204052346093313,
"rewards/rejected": -4.874673025948661,
"step": 253
},
{
"epoch": 0.42778947368421055,
"grad_norm": 115.23971460689606,
"kl": 0.0,
"learning_rate": 3.55432350988687e-07,
"logits/chosen": -830976576.0,
"logits/rejected": -1067142144.0,
"logps/chosen": -153.7244415283203,
"logps/rejected": -251.48733520507812,
"loss": 0.4097,
"rewards/chosen": 0.4606727957725525,
"rewards/margins": 1.2488128542900085,
"rewards/rejected": -0.788140058517456,
"step": 254
},
{
"epoch": 0.42947368421052634,
"grad_norm": 54.92704059463097,
"kl": 0.0,
"learning_rate": 3.5409694312629193e-07,
"logits/chosen": -630669141.3333334,
"logits/rejected": -987313868.8,
"logps/chosen": -209.82574462890625,
"logps/rejected": -287.5651611328125,
"loss": 0.3034,
"rewards/chosen": 1.022467056910197,
"rewards/margins": 2.030047623316447,
"rewards/rejected": -1.00758056640625,
"step": 255
},
{
"epoch": 0.43115789473684213,
"grad_norm": 310.7398715383573,
"kl": 0.0,
"learning_rate": 3.5275793234940544e-07,
"logits/chosen": -917445760.0,
"logits/rejected": -839287808.0,
"logps/chosen": -376.3251953125,
"logps/rejected": -219.46243722098214,
"loss": 0.4546,
"rewards/chosen": -1.2787322998046875,
"rewards/margins": -1.270533732005528,
"rewards/rejected": -0.00819856779915946,
"step": 256
},
{
"epoch": 0.4328421052631579,
"grad_norm": 92.05663310439607,
"kl": 1.0852890014648438,
"learning_rate": 3.514153650027249e-07,
"logits/chosen": -917584298.6666666,
"logits/rejected": -720754227.2,
"logps/chosen": -277.08099365234375,
"logps/rejected": -215.7774169921875,
"loss": 0.4514,
"rewards/chosen": -1.6589380900065105,
"rewards/margins": -1.4071196715037029,
"rewards/rejected": -0.2518184185028076,
"step": 257
},
{
"epoch": 0.4345263157894737,
"grad_norm": 46.29681896211451,
"kl": 0.0,
"learning_rate": 3.5006928755404464e-07,
"logits/chosen": -601805260.8,
"logits/rejected": -705630848.0,
"logps/chosen": -302.690771484375,
"logps/rejected": -220.0159912109375,
"loss": 0.3499,
"rewards/chosen": 0.9885660171508789,
"rewards/margins": -0.45607048670450856,
"rewards/rejected": 1.4446365038553874,
"step": 258
},
{
"epoch": 0.4362105263157895,
"grad_norm": 60.3839696290393,
"kl": 0.0,
"learning_rate": 3.487197465926478e-07,
"logits/chosen": -877942016.0,
"logits/rejected": -762154956.8,
"logps/chosen": -241.57857259114584,
"logps/rejected": -244.444482421875,
"loss": 0.4307,
"rewards/chosen": 1.5284678141276042,
"rewards/margins": 1.983873716990153,
"rewards/rejected": -0.45540590286254884,
"step": 259
},
{
"epoch": 0.4378947368421053,
"grad_norm": 47.33258383634144,
"kl": 0.0,
"learning_rate": 3.473667888276935e-07,
"logits/chosen": -555102617.6,
"logits/rejected": -590850773.3333334,
"logps/chosen": -220.5026123046875,
"logps/rejected": -312.41845703125,
"loss": 0.3409,
"rewards/chosen": 1.8922588348388671,
"rewards/margins": 2.5513052860895793,
"rewards/rejected": -0.659046451250712,
"step": 260
},
{
"epoch": 0.43957894736842107,
"grad_norm": 108.21153319687332,
"kl": 0.0,
"learning_rate": 3.460104610866003e-07,
"logits/chosen": -1036087466.6666666,
"logits/rejected": -1177354444.8,
"logps/chosen": -258.5961100260417,
"logps/rejected": -187.52177734375,
"loss": 0.4231,
"rewards/chosen": 1.5323643684387207,
"rewards/margins": 2.1087326049804687,
"rewards/rejected": -0.5763682365417481,
"step": 261
},
{
"epoch": 0.44126315789473686,
"grad_norm": 119.88936087714019,
"kl": 0.0,
"learning_rate": 3.4465081031342583e-07,
"logits/chosen": -810871347.2,
"logits/rejected": -1137264128.0,
"logps/chosen": -274.263916015625,
"logps/rejected": -234.40445963541666,
"loss": 0.4701,
"rewards/chosen": 0.387792706489563,
"rewards/margins": -0.040941635767618834,
"rewards/rejected": 0.4287343422571818,
"step": 262
},
{
"epoch": 0.44294736842105265,
"grad_norm": 61.18523248696411,
"kl": 0.0,
"learning_rate": 3.432878835672413e-07,
"logits/chosen": -875720789.3333334,
"logits/rejected": -787683264.0,
"logps/chosen": -223.71048990885416,
"logps/rejected": -237.80686950683594,
"loss": 0.4187,
"rewards/chosen": 1.0855443477630615,
"rewards/margins": 0.4147740602493286,
"rewards/rejected": 0.6707702875137329,
"step": 263
},
{
"epoch": 0.44463157894736843,
"grad_norm": 85.74861272190243,
"kl": 0.0,
"learning_rate": 3.419217280205032e-07,
"logits/chosen": -493305770.6666667,
"logits/rejected": -738251980.8,
"logps/chosen": -250.9864501953125,
"logps/rejected": -200.50716552734374,
"loss": 0.4608,
"rewards/chosen": 0.6460779507954916,
"rewards/margins": 0.17245400746663414,
"rewards/rejected": 0.47362394332885743,
"step": 264
},
{
"epoch": 0.4463157894736842,
"grad_norm": 54.77540128655694,
"kl": 0.0,
"learning_rate": 3.405523909574206e-07,
"logits/chosen": -651585450.6666666,
"logits/rejected": -683791001.6,
"logps/chosen": -242.39778645833334,
"logps/rejected": -298.1034912109375,
"loss": 0.3501,
"rewards/chosen": 0.5337386926015218,
"rewards/margins": 0.8820349057515462,
"rewards/rejected": -0.3482962131500244,
"step": 265
},
{
"epoch": 0.448,
"grad_norm": 89.19503556391335,
"kl": 0.0,
"learning_rate": 3.391799197723185e-07,
"logits/chosen": -1162056499.2,
"logits/rejected": -1215143509.3333333,
"logps/chosen": -259.603076171875,
"logps/rejected": -309.9653727213542,
"loss": 0.4462,
"rewards/chosen": 0.42551660537719727,
"rewards/margins": 2.0183765093485517,
"rewards/rejected": -1.5928599039713542,
"step": 266
},
{
"epoch": 0.4496842105263158,
"grad_norm": 50.9461938511508,
"kl": 0.0,
"learning_rate": 3.378043619679973e-07,
"logits/chosen": -1112641536.0,
"logits/rejected": -554899541.3333334,
"logps/chosen": -208.4407470703125,
"logps/rejected": -223.9110107421875,
"loss": 0.37,
"rewards/chosen": 1.009368324279785,
"rewards/margins": 1.0311517755190531,
"rewards/rejected": -0.021783451239267986,
"step": 267
},
{
"epoch": 0.4513684210526316,
"grad_norm": 127.83348671346022,
"kl": 0.0,
"learning_rate": 3.3642576515408905e-07,
"logits/chosen": -642838357.3333334,
"logits/rejected": -1174076211.2,
"logps/chosen": -164.12750244140625,
"logps/rejected": -177.136376953125,
"loss": 0.4873,
"rewards/chosen": 0.642901619275411,
"rewards/margins": 1.1527833541234336,
"rewards/rejected": -0.5098817348480225,
"step": 268
},
{
"epoch": 0.4530526315789474,
"grad_norm": 106.6924383104548,
"kl": 0.0,
"learning_rate": 3.350441770454092e-07,
"logits/chosen": -881408614.4,
"logits/rejected": -748522496.0,
"logps/chosen": -271.259326171875,
"logps/rejected": -234.607421875,
"loss": 0.3324,
"rewards/chosen": 0.8681741714477539,
"rewards/margins": 1.9296697139739991,
"rewards/rejected": -1.0614955425262451,
"step": 269
},
{
"epoch": 0.45473684210526316,
"grad_norm": 51.51923388471137,
"kl": 0.0,
"learning_rate": 3.336596454603054e-07,
"logits/chosen": -1130812074.6666667,
"logits/rejected": -598989926.4,
"logps/chosen": -204.42683919270834,
"logps/rejected": -234.242333984375,
"loss": 0.3524,
"rewards/chosen": 0.7063151200612386,
"rewards/margins": 1.668326743443807,
"rewards/rejected": -0.9620116233825684,
"step": 270
},
{
"epoch": 0.45642105263157895,
"grad_norm": 56.40900184314831,
"kl": 0.0,
"learning_rate": 3.322722183190025e-07,
"logits/chosen": -734314154.6666666,
"logits/rejected": -522018406.4,
"logps/chosen": -165.00169881184897,
"logps/rejected": -248.694873046875,
"loss": 0.445,
"rewards/chosen": -1.2975707054138184,
"rewards/margins": -1.0448772430419921,
"rewards/rejected": -0.25269346237182616,
"step": 271
},
{
"epoch": 0.45810526315789474,
"grad_norm": 68.53879212164166,
"kl": 0.0,
"learning_rate": 3.308819436419437e-07,
"logits/chosen": -1095972181.3333333,
"logits/rejected": -1827424384.0,
"logps/chosen": -254.98893229166666,
"logps/rejected": -206.98135375976562,
"loss": 0.3692,
"rewards/chosen": 0.8086059888203939,
"rewards/margins": 4.705807526906331,
"rewards/rejected": -3.8972015380859375,
"step": 272
},
{
"epoch": 0.4597894736842105,
"grad_norm": 62.413627457878825,
"kl": 0.0,
"learning_rate": 3.294888695481287e-07,
"logits/chosen": -1661173418.6666667,
"logits/rejected": -1076040294.4,
"logps/chosen": -195.62664794921875,
"logps/rejected": -220.1724609375,
"loss": 0.4067,
"rewards/chosen": 1.7726329167683919,
"rewards/margins": 2.3173933347066242,
"rewards/rejected": -0.5447604179382324,
"step": 273
},
{
"epoch": 0.4614736842105263,
"grad_norm": 41.769964940057946,
"kl": 0.0,
"learning_rate": 3.2809304425344856e-07,
"logits/chosen": -1145147264.0,
"logits/rejected": -744639780.5714285,
"logps/chosen": -178.32012939453125,
"logps/rejected": -222.49560546875,
"loss": 0.4646,
"rewards/chosen": 1.3849213123321533,
"rewards/margins": 1.6690550361360823,
"rewards/rejected": -0.28413372380392893,
"step": 274
},
{
"epoch": 0.4631578947368421,
"grad_norm": 51.7666775396847,
"kl": 0.0,
"learning_rate": 3.266945160690159e-07,
"logits/chosen": -718938112.0,
"logits/rejected": -701548672.0,
"logps/chosen": -168.946533203125,
"logps/rejected": -342.5791015625,
"loss": 0.3497,
"rewards/chosen": 1.088832139968872,
"rewards/margins": 3.3126580715179443,
"rewards/rejected": -2.2238259315490723,
"step": 275
},
{
"epoch": 0.4648421052631579,
"grad_norm": 40.27881909450745,
"kl": 0.0,
"learning_rate": 3.252933333994942e-07,
"logits/chosen": -618843904.0,
"logits/rejected": -992992768.0,
"logps/chosen": -176.893798828125,
"logps/rejected": -240.34130859375,
"loss": 0.3206,
"rewards/chosen": 1.2537689208984375,
"rewards/margins": 2.309921582539876,
"rewards/rejected": -1.0561526616414387,
"step": 276
},
{
"epoch": 0.4665263157894737,
"grad_norm": 37.41871849188705,
"kl": 0.0,
"learning_rate": 3.23889544741421e-07,
"logits/chosen": -787857536.0,
"logits/rejected": -838022016.0,
"logps/chosen": -214.87179565429688,
"logps/rejected": -229.31048583984375,
"loss": 0.2983,
"rewards/chosen": 2.574376106262207,
"rewards/margins": 4.175034284591675,
"rewards/rejected": -1.6006581783294678,
"step": 277
},
{
"epoch": 0.46821052631578947,
"grad_norm": 46.77822204181921,
"kl": 0.0,
"learning_rate": 3.224831986815306e-07,
"logits/chosen": -1206424576.0,
"logits/rejected": -531875840.0,
"logps/chosen": -218.09373474121094,
"logps/rejected": -287.2392578125,
"loss": 0.3122,
"rewards/chosen": 1.3858611583709717,
"rewards/margins": 7.854579210281372,
"rewards/rejected": -6.4687180519104,
"step": 278
},
{
"epoch": 0.46989473684210525,
"grad_norm": 48.645481075648526,
"kl": 0.0,
"learning_rate": 3.2107434389507177e-07,
"logits/chosen": -829400217.6,
"logits/rejected": -402655232.0,
"logps/chosen": -180.144287109375,
"logps/rejected": -130.1499226888021,
"loss": 0.329,
"rewards/chosen": 1.9622688293457031,
"rewards/margins": 1.4281412760416665,
"rewards/rejected": 0.5341275533040365,
"step": 279
},
{
"epoch": 0.47157894736842104,
"grad_norm": 41.28195566807422,
"kl": 0.0,
"learning_rate": 3.19663029144123e-07,
"logits/chosen": -841884723.2,
"logits/rejected": -862919680.0,
"logps/chosen": -200.8014892578125,
"logps/rejected": -234.42822265625,
"loss": 0.3977,
"rewards/chosen": 0.43205790519714354,
"rewards/margins": 0.7687125682830811,
"rewards/rejected": -0.3366546630859375,
"step": 280
},
{
"epoch": 0.47326315789473683,
"grad_norm": 84.07688846755318,
"kl": 0.0,
"learning_rate": 3.1824930327590525e-07,
"logits/chosen": -710278314.6666666,
"logits/rejected": -995849472.0,
"logps/chosen": -224.30419921875,
"logps/rejected": -217.58340454101562,
"loss": 0.3496,
"rewards/chosen": 0.7394833564758301,
"rewards/margins": 0.8354802131652832,
"rewards/rejected": -0.09599685668945312,
"step": 281
},
{
"epoch": 0.4749473684210526,
"grad_norm": 112.7314053842234,
"kl": 0.0,
"learning_rate": 3.168332152210909e-07,
"logits/chosen": -497686954.6666667,
"logits/rejected": -851313049.6,
"logps/chosen": -303.6035970052083,
"logps/rejected": -268.5775390625,
"loss": 0.4291,
"rewards/chosen": -0.6700571378072103,
"rewards/margins": -1.136230500539144,
"rewards/rejected": 0.4661733627319336,
"step": 282
},
{
"epoch": 0.4766315789473684,
"grad_norm": 47.718340024126896,
"kl": 0.0,
"learning_rate": 3.154148139921102e-07,
"logits/chosen": -1520765056.0,
"logits/rejected": -917776018.2857143,
"logps/chosen": -182.46583557128906,
"logps/rejected": -224.23322405133928,
"loss": 0.3508,
"rewards/chosen": -1.4963241815567017,
"rewards/margins": -0.8272254637309483,
"rewards/rejected": -0.6690987178257534,
"step": 283
},
{
"epoch": 0.4783157894736842,
"grad_norm": 51.50174863063673,
"kl": 0.0,
"learning_rate": 3.1399414868145503e-07,
"logits/chosen": -879817557.3333334,
"logits/rejected": -509886771.2,
"logps/chosen": -196.01708984375,
"logps/rejected": -247.4649169921875,
"loss": 0.4224,
"rewards/chosen": -1.2518666585286458,
"rewards/margins": 2.6389341990152992,
"rewards/rejected": -3.8908008575439452,
"step": 284
},
{
"epoch": 0.48,
"grad_norm": 46.979767312290676,
"kl": 0.0,
"learning_rate": 3.1257126845997996e-07,
"logits/chosen": -605518720.0,
"logits/rejected": -689517056.0,
"logps/chosen": -217.33444213867188,
"logps/rejected": -329.7290954589844,
"loss": 0.2682,
"rewards/chosen": 0.8716354370117188,
"rewards/margins": 11.700860023498535,
"rewards/rejected": -10.829224586486816,
"step": 285
},
{
"epoch": 0.48168421052631577,
"grad_norm": 53.078466818339,
"kl": 0.0,
"learning_rate": 3.111462225752e-07,
"logits/chosen": -1177064704.0,
"logits/rejected": -1062161664.0,
"logps/chosen": -172.35658264160156,
"logps/rejected": -229.10531616210938,
"loss": 0.2759,
"rewards/chosen": 1.4827051162719727,
"rewards/margins": 3.3628339767456055,
"rewards/rejected": -1.8801288604736328,
"step": 286
},
{
"epoch": 0.48336842105263156,
"grad_norm": 73.05034651400905,
"kl": 0.0,
"learning_rate": 3.097190603495861e-07,
"logits/chosen": -1032015040.0,
"logits/rejected": -431953184.0,
"logps/chosen": -204.94126892089844,
"logps/rejected": -273.42724609375,
"loss": 0.4725,
"rewards/chosen": 0.42430874705314636,
"rewards/margins": 0.12238076329231262,
"rewards/rejected": 0.30192798376083374,
"step": 287
},
{
"epoch": 0.48505263157894735,
"grad_norm": 58.199485015649564,
"kl": 0.0,
"learning_rate": 3.0828983117885856e-07,
"logits/chosen": -947543360.0,
"logits/rejected": -1011533738.6666666,
"logps/chosen": -225.79100036621094,
"logps/rejected": -238.0352579752604,
"loss": 0.3671,
"rewards/chosen": 0.18504944443702698,
"rewards/margins": 2.1747958560784655,
"rewards/rejected": -1.9897464116414387,
"step": 288
},
{
"epoch": 0.48673684210526313,
"grad_norm": 38.613615018070185,
"kl": 0.0,
"learning_rate": 3.0685858453027665e-07,
"logits/chosen": -1476833689.6,
"logits/rejected": -866345984.0,
"logps/chosen": -210.87587890625,
"logps/rejected": -175.0899861653646,
"loss": 0.4495,
"rewards/chosen": 1.3699981689453125,
"rewards/margins": 0.4098469734191894,
"rewards/rejected": 0.960151195526123,
"step": 289
},
{
"epoch": 0.4884210526315789,
"grad_norm": 62.47965185776049,
"kl": 0.0,
"learning_rate": 3.0542536994092696e-07,
"logits/chosen": -1110704128.0,
"logits/rejected": -717989222.4,
"logps/chosen": -271.7138671875,
"logps/rejected": -226.588330078125,
"loss": 0.3672,
"rewards/chosen": 0.2757614056269328,
"rewards/margins": 2.1927677075068157,
"rewards/rejected": -1.9170063018798829,
"step": 290
},
{
"epoch": 0.4901052631578947,
"grad_norm": 44.32623223762354,
"kl": 0.0,
"learning_rate": 3.03990237016009e-07,
"logits/chosen": -798940876.8,
"logits/rejected": -814150997.3333334,
"logps/chosen": -253.6781005859375,
"logps/rejected": -307.87721761067706,
"loss": 0.405,
"rewards/chosen": 0.5707196235656739,
"rewards/margins": 3.1684674263000487,
"rewards/rejected": -2.597747802734375,
"step": 291
},
{
"epoch": 0.4917894736842105,
"grad_norm": 55.91294023967487,
"kl": 0.0,
"learning_rate": 3.025532354271178e-07,
"logits/chosen": -961314816.0,
"logits/rejected": -835593676.8,
"logps/chosen": -208.382568359375,
"logps/rejected": -190.52183837890624,
"loss": 0.4201,
"rewards/chosen": 0.48334558804829914,
"rewards/margins": 0.11008816560109452,
"rewards/rejected": 0.3732574224472046,
"step": 292
},
{
"epoch": 0.49347368421052634,
"grad_norm": 39.083753233538296,
"kl": 0.0,
"learning_rate": 3.0111441491052505e-07,
"logits/chosen": -898919116.8,
"logits/rejected": -784928000.0,
"logps/chosen": -233.350927734375,
"logps/rejected": -234.80265299479166,
"loss": 0.4615,
"rewards/chosen": 0.10661284923553467,
"rewards/margins": 1.890099851290385,
"rewards/rejected": -1.7834870020548503,
"step": 293
},
{
"epoch": 0.49515789473684213,
"grad_norm": 51.14880984410606,
"kl": 0.0,
"learning_rate": 2.996738252654577e-07,
"logits/chosen": -821468160.0,
"logits/rejected": -1598704128.0,
"logps/chosen": -196.221533203125,
"logps/rejected": -359.334228515625,
"loss": 0.4244,
"rewards/chosen": -0.5685586929321289,
"rewards/margins": 3.059499422709147,
"rewards/rejected": -3.628058115641276,
"step": 294
},
{
"epoch": 0.4968421052631579,
"grad_norm": 106.86865404878358,
"kl": 0.0,
"learning_rate": 2.982315163523742e-07,
"logits/chosen": -661637952.0,
"logits/rejected": -624379264.0,
"logps/chosen": -208.84210205078125,
"logps/rejected": -236.47945149739584,
"loss": 0.4826,
"rewards/chosen": 0.8578643798828125,
"rewards/margins": 1.4199414253234863,
"rewards/rejected": -0.5620770454406738,
"step": 295
},
{
"epoch": 0.4985263157894737,
"grad_norm": 60.74058573931991,
"kl": 0.0,
"learning_rate": 2.967875380912388e-07,
"logits/chosen": -1204205824.0,
"logits/rejected": -1308154026.6666667,
"logps/chosen": -226.80226135253906,
"logps/rejected": -224.9236857096354,
"loss": 0.506,
"rewards/chosen": 0.23896867036819458,
"rewards/margins": 0.5748037298520405,
"rewards/rejected": -0.335835059483846,
"step": 296
},
{
"epoch": 0.5002105263157894,
"grad_norm": 42.435672467475534,
"kl": 0.0,
"learning_rate": 2.9534194045979393e-07,
"logits/chosen": -706206506.6666666,
"logits/rejected": -569180224.0,
"logps/chosen": -252.4887898763021,
"logps/rejected": -193.00033569335938,
"loss": 0.372,
"rewards/chosen": 0.9058823585510254,
"rewards/margins": 3.202226400375366,
"rewards/rejected": -2.296344041824341,
"step": 297
},
{
"epoch": 0.5018947368421053,
"grad_norm": 56.50443377225829,
"kl": 0.0,
"learning_rate": 2.9389477349183015e-07,
"logits/chosen": -490414496.0,
"logits/rejected": -853274048.0,
"logps/chosen": -217.90597534179688,
"logps/rejected": -235.65103149414062,
"loss": 0.3792,
"rewards/chosen": 1.547549843788147,
"rewards/margins": 3.4790706634521484,
"rewards/rejected": -1.9315208196640015,
"step": 298
},
{
"epoch": 0.503578947368421,
"grad_norm": 61.19734777336964,
"kl": 0.0,
"learning_rate": 2.924460872754547e-07,
"logits/chosen": -481068714.6666667,
"logits/rejected": -665770188.8,
"logps/chosen": -227.10294596354166,
"logps/rejected": -185.52493896484376,
"loss": 0.3672,
"rewards/chosen": 1.1002014478047688,
"rewards/margins": 2.9621237119038897,
"rewards/rejected": -1.861922264099121,
"step": 299
},
{
"epoch": 0.5052631578947369,
"grad_norm": 104.48557270061485,
"kl": 0.0,
"learning_rate": 2.909959319513574e-07,
"logits/chosen": -709741760.0,
"logits/rejected": -1217407146.6666667,
"logps/chosen": -191.64903259277344,
"logps/rejected": -252.3373819986979,
"loss": 0.3704,
"rewards/chosen": -0.1919853240251541,
"rewards/margins": 0.7327715208133062,
"rewards/rejected": -0.9247568448384603,
"step": 300
},
{
"epoch": 0.5069473684210526,
"grad_norm": 43.12147455849188,
"kl": 0.0,
"learning_rate": 2.8954435771107597e-07,
"logits/chosen": -765106368.0,
"logits/rejected": -901597440.0,
"logps/chosen": -226.85153198242188,
"logps/rejected": -243.94345092773438,
"loss": 0.361,
"rewards/chosen": 1.015770435333252,
"rewards/margins": 2.2279579639434814,
"rewards/rejected": -1.2121875286102295,
"step": 301
},
{
"epoch": 0.5086315789473684,
"grad_norm": 68.53182250423035,
"kl": 0.0,
"learning_rate": 2.880914147952584e-07,
"logits/chosen": -892777728.0,
"logits/rejected": -893605683.2,
"logps/chosen": -231.4443359375,
"logps/rejected": -212.4324462890625,
"loss": 0.4275,
"rewards/chosen": 1.4770619074503581,
"rewards/margins": 0.6548136393229167,
"rewards/rejected": 0.8222482681274415,
"step": 302
},
{
"epoch": 0.5103157894736842,
"grad_norm": 57.399630263093485,
"kl": 0.0,
"learning_rate": 2.8663715349192383e-07,
"logits/chosen": -980971605.3333334,
"logits/rejected": -1283480320.0,
"logps/chosen": -271.35976155598956,
"logps/rejected": -338.93670654296875,
"loss": 0.3848,
"rewards/chosen": 1.7701203028361003,
"rewards/margins": 6.533517519632976,
"rewards/rejected": -4.763397216796875,
"step": 303
},
{
"epoch": 0.512,
"grad_norm": 50.30087172775045,
"kl": 0.0,
"learning_rate": 2.8518162413472263e-07,
"logits/chosen": -876101017.6,
"logits/rejected": -833215488.0,
"logps/chosen": -223.075244140625,
"logps/rejected": -205.84307861328125,
"loss": 0.4463,
"rewards/chosen": 1.0503494262695312,
"rewards/margins": 2.2924479166666667,
"rewards/rejected": -1.2420984903971355,
"step": 304
},
{
"epoch": 0.5136842105263157,
"grad_norm": 56.83106728944569,
"kl": 0.0,
"learning_rate": 2.837248771011937e-07,
"logits/chosen": -778440089.6,
"logits/rejected": -998400341.3333334,
"logps/chosen": -258.7751953125,
"logps/rejected": -218.8525594075521,
"loss": 0.4367,
"rewards/chosen": -0.8302111625671387,
"rewards/margins": -1.7733329931894937,
"rewards/rejected": 0.9431218306223551,
"step": 305
},
{
"epoch": 0.5153684210526316,
"grad_norm": 37.14827382589415,
"kl": 0.0,
"learning_rate": 2.822669628110213e-07,
"logits/chosen": -949675349.3333334,
"logits/rejected": -1104420966.4,
"logps/chosen": -197.35005696614584,
"logps/rejected": -236.111083984375,
"loss": 0.3883,
"rewards/chosen": 1.4387264251708984,
"rewards/margins": 1.295378041267395,
"rewards/rejected": 0.1433483839035034,
"step": 306
},
{
"epoch": 0.5170526315789473,
"grad_norm": 42.48400320185914,
"kl": 0.0,
"learning_rate": 2.808079317242896e-07,
"logits/chosen": -733285376.0,
"logits/rejected": -290018304.0,
"logps/chosen": -230.39531598772322,
"logps/rejected": -196.5876922607422,
"loss": 0.3848,
"rewards/chosen": -0.14083341189793178,
"rewards/margins": -1.530245338167463,
"rewards/rejected": 1.3894119262695312,
"step": 307
},
{
"epoch": 0.5187368421052632,
"grad_norm": 49.30251176881769,
"kl": 0.0,
"learning_rate": 2.793478343397367e-07,
"logits/chosen": -637947801.6,
"logits/rejected": -1100627200.0,
"logps/chosen": -236.856103515625,
"logps/rejected": -254.74955240885416,
"loss": 0.4514,
"rewards/chosen": 0.8377789497375489,
"rewards/margins": 0.31805851459503176,
"rewards/rejected": 0.5197204351425171,
"step": 308
},
{
"epoch": 0.5204210526315789,
"grad_norm": 38.15594290546514,
"kl": 0.0,
"learning_rate": 2.778867211930061e-07,
"logits/chosen": -946033956.5714285,
"logits/rejected": -322719744.0,
"logps/chosen": -188.54940359933036,
"logps/rejected": -337.9103088378906,
"loss": 0.3864,
"rewards/chosen": 0.2591233934674944,
"rewards/margins": 4.263719626835415,
"rewards/rejected": -4.00459623336792,
"step": 309
},
{
"epoch": 0.5221052631578947,
"grad_norm": 36.73206711146749,
"kl": 0.0,
"learning_rate": 2.7642464285489827e-07,
"logits/chosen": -1031071232.0,
"logits/rejected": -701534515.2,
"logps/chosen": -285.6693115234375,
"logps/rejected": -270.429541015625,
"loss": 0.3706,
"rewards/chosen": -0.7057943344116211,
"rewards/margins": 0.5104178428649901,
"rewards/rejected": -1.2162121772766112,
"step": 310
},
{
"epoch": 0.5237894736842105,
"grad_norm": 35.15174071748067,
"kl": 0.0,
"learning_rate": 2.749616499296199e-07,
"logits/chosen": -999771648.0,
"logits/rejected": -866819712.0,
"logps/chosen": -261.1120910644531,
"logps/rejected": -233.74642944335938,
"loss": 0.4233,
"rewards/chosen": -0.03837205469608307,
"rewards/margins": 2.5518483966588974,
"rewards/rejected": -2.5902204513549805,
"step": 311
},
{
"epoch": 0.5254736842105263,
"grad_norm": 40.26805134000718,
"kl": 0.0,
"learning_rate": 2.734977930530326e-07,
"logits/chosen": -521609728.0,
"logits/rejected": -1214733482.6666667,
"logps/chosen": -222.58474731445312,
"logps/rejected": -238.75984700520834,
"loss": 0.4109,
"rewards/chosen": 0.5281776785850525,
"rewards/margins": -0.3983253439267477,
"rewards/rejected": 0.9265030225118002,
"step": 312
},
{
"epoch": 0.5271578947368422,
"grad_norm": 41.427977504816596,
"kl": 0.0,
"learning_rate": 2.7203312289090044e-07,
"logits/chosen": -484565952.0,
"logits/rejected": -1246305024.0,
"logps/chosen": -273.0993347167969,
"logps/rejected": -276.9410400390625,
"loss": 0.4445,
"rewards/chosen": -0.8267297148704529,
"rewards/margins": 2.6424904465675354,
"rewards/rejected": -3.4692201614379883,
"step": 313
},
{
"epoch": 0.5288421052631579,
"grad_norm": 68.994308446122,
"kl": 0.0,
"learning_rate": 2.7056769013713623e-07,
"logits/chosen": -822423360.0,
"logits/rejected": -650067328.0,
"logps/chosen": -216.25306701660156,
"logps/rejected": -217.991943359375,
"loss": 0.3846,
"rewards/chosen": 1.5515110492706299,
"rewards/margins": 1.647828370332718,
"rewards/rejected": -0.09631732106208801,
"step": 314
},
{
"epoch": 0.5305263157894737,
"grad_norm": 57.35653485170902,
"kl": 0.0,
"learning_rate": 2.6910154551204676e-07,
"logits/chosen": -745939541.3333334,
"logits/rejected": -736545408.0,
"logps/chosen": -232.83365885416666,
"logps/rejected": -210.52879333496094,
"loss": 0.4926,
"rewards/chosen": 0.7454870541890463,
"rewards/margins": 0.6709303458531698,
"rewards/rejected": 0.07455670833587646,
"step": 315
},
{
"epoch": 0.5322105263157895,
"grad_norm": 81.62832673291715,
"kl": 0.0,
"learning_rate": 2.6763473976057773e-07,
"logits/chosen": -1231821397.3333333,
"logits/rejected": -768418918.4,
"logps/chosen": -211.95243326822916,
"logps/rejected": -236.2597900390625,
"loss": 0.351,
"rewards/chosen": 0.9875667095184326,
"rewards/margins": 3.251889371871948,
"rewards/rejected": -2.2643226623535155,
"step": 316
},
{
"epoch": 0.5338947368421053,
"grad_norm": 40.40366631438029,
"kl": 0.0,
"learning_rate": 2.6616732365055713e-07,
"logits/chosen": -878860083.2,
"logits/rejected": -1151675733.3333333,
"logps/chosen": -263.190625,
"logps/rejected": -241.0025634765625,
"loss": 0.3784,
"rewards/chosen": 1.9023317337036132,
"rewards/margins": 2.3599199295043944,
"rewards/rejected": -0.45758819580078125,
"step": 317
},
{
"epoch": 0.535578947368421,
"grad_norm": 37.906423958216756,
"kl": 0.0,
"learning_rate": 2.64699347970938e-07,
"logits/chosen": -643882432.0,
"logits/rejected": -725854293.3333334,
"logps/chosen": -210.35226440429688,
"logps/rejected": -264.7013346354167,
"loss": 0.39,
"rewards/chosen": 0.8370101451873779,
"rewards/margins": 0.3739887475967407,
"rewards/rejected": 0.4630213975906372,
"step": 318
},
{
"epoch": 0.5372631578947369,
"grad_norm": 41.62282940753045,
"kl": 0.0,
"learning_rate": 2.632308635300408e-07,
"logits/chosen": -1076104704.0,
"logits/rejected": -821822592.0,
"logps/chosen": -270.3343505859375,
"logps/rejected": -172.8352508544922,
"loss": 0.3992,
"rewards/chosen": -0.8739106059074402,
"rewards/margins": -0.6177101731300354,
"rewards/rejected": -0.2562004327774048,
"step": 319
},
{
"epoch": 0.5389473684210526,
"grad_norm": 32.33388694240573,
"kl": 0.0,
"learning_rate": 2.6176192115379494e-07,
"logits/chosen": -1163147776.0,
"logits/rejected": -671515520.0,
"logps/chosen": -232.58616638183594,
"logps/rejected": -390.2035319010417,
"loss": 0.4368,
"rewards/chosen": 3.005178928375244,
"rewards/margins": 4.992446422576904,
"rewards/rejected": -1.9872674942016602,
"step": 320
},
{
"epoch": 0.5406315789473685,
"grad_norm": 99.7767501397208,
"kl": 0.0,
"learning_rate": 2.6029257168397944e-07,
"logits/chosen": -1067986124.8,
"logits/rejected": -520208341.3333333,
"logps/chosen": -205.670068359375,
"logps/rejected": -209.88446044921875,
"loss": 0.365,
"rewards/chosen": 0.7348861694335938,
"rewards/margins": 0.3829381465911865,
"rewards/rejected": 0.3519480228424072,
"step": 321
},
{
"epoch": 0.5423157894736842,
"grad_norm": 43.79004322458768,
"kl": 0.0,
"learning_rate": 2.5882286597646313e-07,
"logits/chosen": -837222016.0,
"logits/rejected": -825369429.3333334,
"logps/chosen": -224.58160400390625,
"logps/rejected": -298.0353597005208,
"loss": 0.5333,
"rewards/chosen": 1.3384246826171875,
"rewards/margins": 2.8695327440897627,
"rewards/rejected": -1.531108061472575,
"step": 322
},
{
"epoch": 0.544,
"grad_norm": 45.949254583233454,
"kl": 0.0,
"learning_rate": 2.5735285489944485e-07,
"logits/chosen": -728969216.0,
"logits/rejected": -885923225.6,
"logps/chosen": -220.39337158203125,
"logps/rejected": -217.1091552734375,
"loss": 0.4497,
"rewards/chosen": 3.17012882232666,
"rewards/margins": 2.6761786460876467,
"rewards/rejected": 0.49395017623901366,
"step": 323
},
{
"epoch": 0.5456842105263158,
"grad_norm": 51.89181096647278,
"kl": 0.0,
"learning_rate": 2.5588258933169244e-07,
"logits/chosen": -918494272.0,
"logits/rejected": -728522112.0,
"logps/chosen": -144.35769653320312,
"logps/rejected": -252.4864501953125,
"loss": 0.4313,
"rewards/chosen": 0.6073247194290161,
"rewards/margins": 1.5144293308258057,
"rewards/rejected": -0.9071046113967896,
"step": 324
},
{
"epoch": 0.5473684210526316,
"grad_norm": 42.4363414069206,
"kl": 0.0,
"learning_rate": 2.544121201607822e-07,
"logits/chosen": -741229977.6,
"logits/rejected": -902070272.0,
"logps/chosen": -216.93974609375,
"logps/rejected": -230.6597900390625,
"loss": 0.4731,
"rewards/chosen": -0.7157364368438721,
"rewards/margins": -0.09154821236928312,
"rewards/rejected": -0.624188224474589,
"step": 325
},
{
"epoch": 0.5490526315789473,
"grad_norm": 49.977660612765106,
"kl": 0.0,
"learning_rate": 2.5294149828133704e-07,
"logits/chosen": -785820672.0,
"logits/rejected": -410074240.0,
"logps/chosen": -192.42119489397322,
"logps/rejected": -291.78887939453125,
"loss": 0.2975,
"rewards/chosen": 1.6415718623570033,
"rewards/margins": 2.1497139760426114,
"rewards/rejected": -0.5081421136856079,
"step": 326
},
{
"epoch": 0.5507368421052632,
"grad_norm": 40.20254402613094,
"kl": 0.0,
"learning_rate": 2.5147077459326555e-07,
"logits/chosen": -540746086.4,
"logits/rejected": -600018858.6666666,
"logps/chosen": -271.22294921875,
"logps/rejected": -190.82244873046875,
"loss": 0.3018,
"rewards/chosen": 1.12197265625,
"rewards/margins": 4.917408243815104,
"rewards/rejected": -3.795435587565104,
"step": 327
},
{
"epoch": 0.5524210526315789,
"grad_norm": 36.746566177202766,
"kl": 0.0,
"learning_rate": 2.5e-07,
"logits/chosen": -960942811.4285715,
"logits/rejected": -569875712.0,
"logps/chosen": -220.05186244419642,
"logps/rejected": -294.9991149902344,
"loss": 0.3669,
"rewards/chosen": 1.3416931969778878,
"rewards/margins": 3.6566162449972968,
"rewards/rejected": -2.314923048019409,
"step": 328
},
{
"epoch": 0.5541052631578948,
"grad_norm": 92.10516677186112,
"kl": 0.0,
"learning_rate": 2.485292254067345e-07,
"logits/chosen": -911412736.0,
"logits/rejected": -994976153.6,
"logps/chosen": -234.53729248046875,
"logps/rejected": -238.84951171875,
"loss": 0.3849,
"rewards/chosen": 0.3486948808034261,
"rewards/margins": 1.069074519475301,
"rewards/rejected": -0.720379638671875,
"step": 329
},
{
"epoch": 0.5557894736842105,
"grad_norm": 40.68592836090874,
"kl": 0.0,
"learning_rate": 2.4705850171866294e-07,
"logits/chosen": -916755840.0,
"logits/rejected": -721056704.0,
"logps/chosen": -226.9412841796875,
"logps/rejected": -197.04571533203125,
"loss": 0.3178,
"rewards/chosen": 1.0302788019180298,
"rewards/margins": 2.7032233476638794,
"rewards/rejected": -1.6729445457458496,
"step": 330
},
{
"epoch": 0.5574736842105263,
"grad_norm": 64.00175361910134,
"kl": 0.0,
"learning_rate": 2.4558787983921785e-07,
"logits/chosen": -1136037478.4,
"logits/rejected": -776420096.0,
"logps/chosen": -184.01240234375,
"logps/rejected": -270.2752278645833,
"loss": 0.354,
"rewards/chosen": 1.9217365264892579,
"rewards/margins": 2.225330440203349,
"rewards/rejected": -0.303593913714091,
"step": 331
},
{
"epoch": 0.5591578947368421,
"grad_norm": 53.20843592872294,
"kl": 0.0,
"learning_rate": 2.441174106683076e-07,
"logits/chosen": -987133542.4,
"logits/rejected": -757408768.0,
"logps/chosen": -270.960498046875,
"logps/rejected": -269.0520426432292,
"loss": 0.4314,
"rewards/chosen": 0.6490658760070801,
"rewards/margins": 1.0015769640604655,
"rewards/rejected": -0.35251108805338544,
"step": 332
},
{
"epoch": 0.5608421052631579,
"grad_norm": 39.62692923007265,
"kl": 0.0,
"learning_rate": 2.4264714510055513e-07,
"logits/chosen": -779204096.0,
"logits/rejected": -775985749.3333334,
"logps/chosen": -172.64938354492188,
"logps/rejected": -189.1490275065104,
"loss": 0.379,
"rewards/chosen": 0.8577301502227783,
"rewards/margins": 0.2588157653808594,
"rewards/rejected": 0.598914384841919,
"step": 333
},
{
"epoch": 0.5625263157894737,
"grad_norm": 189.00573139104904,
"kl": 0.0,
"learning_rate": 2.4117713402353685e-07,
"logits/chosen": -717026304.0,
"logits/rejected": -589375616.0,
"logps/chosen": -255.56656901041666,
"logps/rejected": -222.212158203125,
"loss": 0.4805,
"rewards/chosen": -0.3913642962773641,
"rewards/margins": -0.5911636402209599,
"rewards/rejected": 0.1997993439435959,
"step": 334
},
{
"epoch": 0.5642105263157895,
"grad_norm": 45.7662847906049,
"kl": 0.0,
"learning_rate": 2.397074283160206e-07,
"logits/chosen": -753268096.0,
"logits/rejected": -917319168.0,
"logps/chosen": -198.811279296875,
"logps/rejected": -218.18515014648438,
"loss": 0.3514,
"rewards/chosen": 1.425872802734375,
"rewards/margins": 2.8109121322631836,
"rewards/rejected": -1.3850393295288086,
"step": 335
},
{
"epoch": 0.5658947368421052,
"grad_norm": 39.80557628734281,
"kl": 0.0,
"learning_rate": 2.38238078846205e-07,
"logits/chosen": -806648490.6666666,
"logits/rejected": -1043137280.0,
"logps/chosen": -197.84334309895834,
"logps/rejected": -166.3507080078125,
"loss": 0.4203,
"rewards/chosen": 2.0990654627482095,
"rewards/margins": 0.3083709875742593,
"rewards/rejected": 1.7906944751739502,
"step": 336
},
{
"epoch": 0.5675789473684211,
"grad_norm": 55.26598793514629,
"kl": 0.0,
"learning_rate": 2.367691364699592e-07,
"logits/chosen": -990476117.3333334,
"logits/rejected": -843749478.4,
"logps/chosen": -211.03731282552084,
"logps/rejected": -200.39736328125,
"loss": 0.4159,
"rewards/chosen": 1.5658411979675293,
"rewards/margins": 2.0694480895996095,
"rewards/rejected": -0.5036068916320801,
"step": 337
},
{
"epoch": 0.5692631578947368,
"grad_norm": 43.55127827223215,
"kl": 0.0,
"learning_rate": 2.3530065202906208e-07,
"logits/chosen": -677904725.3333334,
"logits/rejected": -739208768.0,
"logps/chosen": -215.2685546875,
"logps/rejected": -168.70806884765625,
"loss": 0.365,
"rewards/chosen": -0.36894532044728595,
"rewards/margins": 3.7609441677729287,
"rewards/rejected": -4.129889488220215,
"step": 338
},
{
"epoch": 0.5709473684210526,
"grad_norm": 44.273467218054,
"kl": 0.0,
"learning_rate": 2.3383267634944288e-07,
"logits/chosen": -708441152.0,
"logits/rejected": -909734720.0,
"logps/chosen": -236.1248016357422,
"logps/rejected": -329.7475280761719,
"loss": 0.3328,
"rewards/chosen": 0.44537657499313354,
"rewards/margins": 2.3071479201316833,
"rewards/rejected": -1.8617713451385498,
"step": 339
},
{
"epoch": 0.5726315789473684,
"grad_norm": 44.104623030836734,
"kl": 0.0,
"learning_rate": 2.3236526023942222e-07,
"logits/chosen": -1229630873.6,
"logits/rejected": -615721045.3333334,
"logps/chosen": -201.258935546875,
"logps/rejected": -160.00018310546875,
"loss": 0.3583,
"rewards/chosen": 0.4522857666015625,
"rewards/margins": 3.613059679667155,
"rewards/rejected": -3.1607739130655923,
"step": 340
},
{
"epoch": 0.5743157894736842,
"grad_norm": 37.46747269379094,
"kl": 0.0,
"learning_rate": 2.3089845448795325e-07,
"logits/chosen": -972004778.6666666,
"logits/rejected": -684686336.0,
"logps/chosen": -186.35986328125,
"logps/rejected": -252.18388671875,
"loss": 0.3367,
"rewards/chosen": 2.4348222414652505,
"rewards/margins": 3.513483206431071,
"rewards/rejected": -1.0786609649658203,
"step": 341
},
{
"epoch": 0.576,
"grad_norm": 41.066699309390806,
"kl": 0.0,
"learning_rate": 2.2943230986286386e-07,
"logits/chosen": -776381866.6666666,
"logits/rejected": -411293952.0,
"logps/chosen": -237.33272298177084,
"logps/rejected": -230.616357421875,
"loss": 0.5505,
"rewards/chosen": -1.2874414920806885,
"rewards/margins": -2.0171528339385985,
"rewards/rejected": 0.7297113418579102,
"step": 342
},
{
"epoch": 0.5776842105263158,
"grad_norm": 34.78601277894388,
"kl": 0.0,
"learning_rate": 2.2796687710909961e-07,
"logits/chosen": -894285824.0,
"logits/rejected": -1057857365.3333334,
"logps/chosen": -193.6359619140625,
"logps/rejected": -198.5640665690104,
"loss": 0.4058,
"rewards/chosen": 0.9409942626953125,
"rewards/margins": -0.3375056584676106,
"rewards/rejected": 1.278499921162923,
"step": 343
},
{
"epoch": 0.5793684210526315,
"grad_norm": 43.656984836332974,
"kl": 0.0,
"learning_rate": 2.2650220694696746e-07,
"logits/chosen": -936853913.6,
"logits/rejected": -637026218.6666666,
"logps/chosen": -258.152392578125,
"logps/rejected": -172.32596842447916,
"loss": 0.3634,
"rewards/chosen": 0.7583493709564209,
"rewards/margins": 0.6318836331367492,
"rewards/rejected": 0.12646573781967163,
"step": 344
},
{
"epoch": 0.5810526315789474,
"grad_norm": 43.09840300795585,
"kl": 0.0,
"learning_rate": 2.2503835007038018e-07,
"logits/chosen": -1292558336.0,
"logits/rejected": -511415200.0,
"logps/chosen": -219.50491333007812,
"logps/rejected": -236.9748077392578,
"loss": 0.3775,
"rewards/chosen": 1.3266575336456299,
"rewards/margins": 2.1448956727981567,
"rewards/rejected": -0.8182381391525269,
"step": 345
},
{
"epoch": 0.5827368421052631,
"grad_norm": 38.706929628825115,
"kl": 0.0,
"learning_rate": 2.2357535714510179e-07,
"logits/chosen": -1250603776.0,
"logits/rejected": -919112908.8,
"logps/chosen": -281.291748046875,
"logps/rejected": -183.513916015625,
"loss": 0.3373,
"rewards/chosen": -0.5119430224100748,
"rewards/margins": 1.2942089398701988,
"rewards/rejected": -1.8061519622802735,
"step": 346
},
{
"epoch": 0.584421052631579,
"grad_norm": 54.06441877859246,
"kl": 0.0,
"learning_rate": 2.2211327880699389e-07,
"logits/chosen": -983545036.8,
"logits/rejected": -1222936832.0,
"logps/chosen": -196.76480712890626,
"logps/rejected": -226.22184244791666,
"loss": 0.4641,
"rewards/chosen": 0.9738927841186523,
"rewards/margins": -0.2919163544972737,
"rewards/rejected": 1.265809138615926,
"step": 347
},
{
"epoch": 0.5861052631578947,
"grad_norm": 31.758781014479883,
"kl": 0.0,
"learning_rate": 2.206521656602633e-07,
"logits/chosen": -929091264.0,
"logits/rejected": -505945888.0,
"logps/chosen": -176.41741943359375,
"logps/rejected": -188.78915405273438,
"loss": 0.3671,
"rewards/chosen": 1.3973584175109863,
"rewards/margins": 1.813145935535431,
"rewards/rejected": -0.4157875180244446,
"step": 348
},
{
"epoch": 0.5877894736842105,
"grad_norm": 37.86314064294317,
"kl": 0.0,
"learning_rate": 2.1919206827571034e-07,
"logits/chosen": -801264640.0,
"logits/rejected": -464600064.0,
"logps/chosen": -198.34344482421875,
"logps/rejected": -244.7729034423828,
"loss": 0.4117,
"rewards/chosen": 1.5484859943389893,
"rewards/margins": 3.387136936187744,
"rewards/rejected": -1.8386509418487549,
"step": 349
},
{
"epoch": 0.5894736842105263,
"grad_norm": 50.13743066285001,
"kl": 0.0,
"learning_rate": 2.177330371889787e-07,
"logits/chosen": -899666124.8,
"logits/rejected": -916757248.0,
"logps/chosen": -219.9362060546875,
"logps/rejected": -315.0555419921875,
"loss": 0.4165,
"rewards/chosen": 0.3476431369781494,
"rewards/margins": 0.9587682882944742,
"rewards/rejected": -0.6111251513163248,
"step": 350
},
{
"epoch": 0.5911578947368421,
"grad_norm": 51.02552672848151,
"kl": 0.0,
"learning_rate": 2.162751228988063e-07,
"logits/chosen": -853029760.0,
"logits/rejected": -762731136.0,
"logps/chosen": -260.21746826171875,
"logps/rejected": -254.34788513183594,
"loss": 0.3839,
"rewards/chosen": 0.05103302001953125,
"rewards/margins": 1.4993259906768799,
"rewards/rejected": -1.4482929706573486,
"step": 351
},
{
"epoch": 0.592842105263158,
"grad_norm": 34.82384239421157,
"kl": 0.0,
"learning_rate": 2.148183758652774e-07,
"logits/chosen": -740140032.0,
"logits/rejected": -741718784.0,
"logps/chosen": -147.6769561767578,
"logps/rejected": -276.81591796875,
"loss": 0.4139,
"rewards/chosen": 2.366483211517334,
"rewards/margins": 3.0736358165740967,
"rewards/rejected": -0.7071526050567627,
"step": 352
},
{
"epoch": 0.5945263157894737,
"grad_norm": 34.883343445970695,
"kl": 0.0,
"learning_rate": 2.1336284650807612e-07,
"logits/chosen": -1254429781.3333333,
"logits/rejected": -611748761.6,
"logps/chosen": -215.1737060546875,
"logps/rejected": -214.90478515625,
"loss": 0.3704,
"rewards/chosen": 2.0823793411254883,
"rewards/margins": 2.4012239456176756,
"rewards/rejected": -0.3188446044921875,
"step": 353
},
{
"epoch": 0.5962105263157895,
"grad_norm": 55.9131613356441,
"kl": 0.0,
"learning_rate": 2.1190858520474163e-07,
"logits/chosen": -1588147200.0,
"logits/rejected": -903177318.4,
"logps/chosen": -219.27388509114584,
"logps/rejected": -283.88779296875,
"loss": 0.3452,
"rewards/chosen": 2.384967645009359,
"rewards/margins": 2.509509845574697,
"rewards/rejected": -0.12454220056533813,
"step": 354
},
{
"epoch": 0.5978947368421053,
"grad_norm": 43.395183494537164,
"kl": 0.0,
"learning_rate": 2.10455642288924e-07,
"logits/chosen": -1037725632.0,
"logits/rejected": -652462464.0,
"logps/chosen": -192.9014892578125,
"logps/rejected": -218.373779296875,
"loss": 0.338,
"rewards/chosen": 1.9063019752502441,
"rewards/margins": 2.8689812421798706,
"rewards/rejected": -0.9626792669296265,
"step": 355
},
{
"epoch": 0.5995789473684211,
"grad_norm": 72.0621578330362,
"kl": 0.0,
"learning_rate": 2.090040680486426e-07,
"logits/chosen": -1054155008.0,
"logits/rejected": -724854912.0,
"logps/chosen": -225.2272491455078,
"logps/rejected": -268.515869140625,
"loss": 0.4356,
"rewards/chosen": -0.31277698278427124,
"rewards/margins": 1.6200557351112366,
"rewards/rejected": -1.9328327178955078,
"step": 356
},
{
"epoch": 0.6012631578947368,
"grad_norm": 46.80458240411995,
"kl": 0.0,
"learning_rate": 2.0755391272454537e-07,
"logits/chosen": -1309161600.0,
"logits/rejected": -773884096.0,
"logps/chosen": -221.47581481933594,
"logps/rejected": -243.53134155273438,
"loss": 0.3444,
"rewards/chosen": 0.6755388379096985,
"rewards/margins": 0.5105028748512268,
"rewards/rejected": 0.16503596305847168,
"step": 357
},
{
"epoch": 0.6029473684210527,
"grad_norm": 44.25731693592669,
"kl": 0.0,
"learning_rate": 2.0610522650816982e-07,
"logits/chosen": -781600384.0,
"logits/rejected": -991481728.0,
"logps/chosen": -269.00433349609375,
"logps/rejected": -233.53921508789062,
"loss": 0.3766,
"rewards/chosen": -0.19508057832717896,
"rewards/margins": 1.9026996493339539,
"rewards/rejected": -2.097780227661133,
"step": 358
},
{
"epoch": 0.6046315789473684,
"grad_norm": 51.072468854632184,
"kl": 0.0,
"learning_rate": 2.04658059540206e-07,
"logits/chosen": -955063210.6666666,
"logits/rejected": -842874060.8,
"logps/chosen": -206.79142252604166,
"logps/rejected": -263.4759521484375,
"loss": 0.3669,
"rewards/chosen": 1.126490831375122,
"rewards/margins": 1.8041922569274902,
"rewards/rejected": -0.6777014255523681,
"step": 359
},
{
"epoch": 0.6063157894736843,
"grad_norm": 43.797014090675,
"kl": 0.0,
"learning_rate": 2.0321246190876112e-07,
"logits/chosen": -813810773.3333334,
"logits/rejected": -735887360.0,
"logps/chosen": -223.61555989583334,
"logps/rejected": -310.8141357421875,
"loss": 0.2421,
"rewards/chosen": 2.7599121729532876,
"rewards/margins": 4.69793898264567,
"rewards/rejected": -1.9380268096923827,
"step": 360
},
{
"epoch": 0.608,
"grad_norm": 44.19945884028008,
"kl": 0.0,
"learning_rate": 2.0176848364762576e-07,
"logits/chosen": -1044594995.2,
"logits/rejected": -838090410.6666666,
"logps/chosen": -241.623193359375,
"logps/rejected": -180.17657470703125,
"loss": 0.4451,
"rewards/chosen": 0.6463582992553711,
"rewards/margins": 3.4257356007893884,
"rewards/rejected": -2.779377301534017,
"step": 361
},
{
"epoch": 0.6096842105263158,
"grad_norm": 84.65829093491772,
"kl": 0.0,
"learning_rate": 2.0032617473454225e-07,
"logits/chosen": -993763620.5714285,
"logits/rejected": -1168001152.0,
"logps/chosen": -222.610107421875,
"logps/rejected": -359.9516296386719,
"loss": 0.373,
"rewards/chosen": 0.029357058661324636,
"rewards/margins": 5.195747954504831,
"rewards/rejected": -5.166390895843506,
"step": 362
},
{
"epoch": 0.6113684210526316,
"grad_norm": 37.454198883632955,
"kl": 0.0,
"learning_rate": 1.9888558508947492e-07,
"logits/chosen": -656499840.0,
"logits/rejected": -800222016.0,
"logps/chosen": -248.09259033203125,
"logps/rejected": -252.30076599121094,
"loss": 0.3177,
"rewards/chosen": 1.0152060985565186,
"rewards/margins": 3.4270780086517334,
"rewards/rejected": -2.411871910095215,
"step": 363
},
{
"epoch": 0.6130526315789474,
"grad_norm": 44.05381109880352,
"kl": 0.0,
"learning_rate": 1.974467645728822e-07,
"logits/chosen": -734617600.0,
"logits/rejected": -1037793280.0,
"logps/chosen": -210.473876953125,
"logps/rejected": -227.2939453125,
"loss": 0.3765,
"rewards/chosen": 0.8343788385391235,
"rewards/margins": 2.4099492629369097,
"rewards/rejected": -1.5755704243977864,
"step": 364
},
{
"epoch": 0.6147368421052631,
"grad_norm": 48.57136273490337,
"kl": 0.0,
"learning_rate": 1.9600976298399108e-07,
"logits/chosen": -942329036.8,
"logits/rejected": -660537429.3333334,
"logps/chosen": -285.062890625,
"logps/rejected": -235.09102376302084,
"loss": 0.3531,
"rewards/chosen": -0.9990038871765137,
"rewards/margins": 0.40311670303344727,
"rewards/rejected": -1.402120590209961,
"step": 365
},
{
"epoch": 0.616421052631579,
"grad_norm": 37.488732058461395,
"kl": 0.0,
"learning_rate": 1.945746300590731e-07,
"logits/chosen": -483911520.0,
"logits/rejected": -1166225408.0,
"logps/chosen": -247.5726776123047,
"logps/rejected": -214.5377197265625,
"loss": 0.3618,
"rewards/chosen": 0.26900553703308105,
"rewards/margins": 0.6737111012140911,
"rewards/rejected": -0.40470556418100995,
"step": 366
},
{
"epoch": 0.6181052631578947,
"grad_norm": 44.38980472905698,
"kl": 0.0,
"learning_rate": 1.9314141546972344e-07,
"logits/chosen": -783111424.0,
"logits/rejected": -941412928.0,
"logps/chosen": -177.5583953857422,
"logps/rejected": -260.1361083984375,
"loss": 0.3547,
"rewards/chosen": -0.2703380584716797,
"rewards/margins": -0.19543498754501343,
"rewards/rejected": -0.07490307092666626,
"step": 367
},
{
"epoch": 0.6197894736842106,
"grad_norm": 37.93187426865055,
"kl": 0.0,
"learning_rate": 1.9171016882114152e-07,
"logits/chosen": -849338432.0,
"logits/rejected": -782459200.0,
"logps/chosen": -195.59857177734375,
"logps/rejected": -232.708984375,
"loss": 0.38,
"rewards/chosen": 1.189550757408142,
"rewards/margins": 3.122536540031433,
"rewards/rejected": -1.932985782623291,
"step": 368
},
{
"epoch": 0.6214736842105263,
"grad_norm": 38.196373434045015,
"kl": 0.0,
"learning_rate": 1.9028093965041391e-07,
"logits/chosen": -1023574976.0,
"logits/rejected": -983692288.0,
"logps/chosen": -205.63710021972656,
"logps/rejected": -268.6028747558594,
"loss": 0.3751,
"rewards/chosen": 0.3406059741973877,
"rewards/margins": 1.6527996063232422,
"rewards/rejected": -1.3121936321258545,
"step": 369
},
{
"epoch": 0.6231578947368421,
"grad_norm": 55.12173721267581,
"kl": 0.0,
"learning_rate": 1.888537774248e-07,
"logits/chosen": -741024597.3333334,
"logits/rejected": -600171980.8,
"logps/chosen": -179.9103800455729,
"logps/rejected": -237.39990234375,
"loss": 0.418,
"rewards/chosen": 1.0137186845143635,
"rewards/margins": 2.055362590154012,
"rewards/rejected": -1.0416439056396485,
"step": 370
},
{
"epoch": 0.6248421052631579,
"grad_norm": 34.16987497139576,
"kl": 0.0,
"learning_rate": 1.8742873154002004e-07,
"logits/chosen": -837830826.6666666,
"logits/rejected": -1010286720.0,
"logps/chosen": -276.2460530598958,
"logps/rejected": -183.48431396484375,
"loss": 0.4701,
"rewards/chosen": 0.2395154039065043,
"rewards/margins": -0.8625567158063253,
"rewards/rejected": 1.1020721197128296,
"step": 371
},
{
"epoch": 0.6265263157894737,
"grad_norm": 46.66359237058551,
"kl": 0.0,
"learning_rate": 1.86005851318545e-07,
"logits/chosen": -669656012.8,
"logits/rejected": -1303058517.3333333,
"logps/chosen": -244.3946044921875,
"logps/rejected": -313.6065266927083,
"loss": 0.3828,
"rewards/chosen": -0.9342806816101075,
"rewards/margins": 2.6151256879170734,
"rewards/rejected": -3.549406369527181,
"step": 372
},
{
"epoch": 0.6282105263157894,
"grad_norm": 35.82420878860377,
"kl": 0.0,
"learning_rate": 1.8458518600788987e-07,
"logits/chosen": -1138871722.6666667,
"logits/rejected": -761401753.6,
"logps/chosen": -230.2353312174479,
"logps/rejected": -246.2571044921875,
"loss": 0.3776,
"rewards/chosen": 0.059048473834991455,
"rewards/margins": 2.80713711977005,
"rewards/rejected": -2.7480886459350584,
"step": 373
},
{
"epoch": 0.6298947368421053,
"grad_norm": 47.50923482875504,
"kl": 0.0,
"learning_rate": 1.8316678477890913e-07,
"logits/chosen": -1124751530.6666667,
"logits/rejected": -855729868.8,
"logps/chosen": -216.44978841145834,
"logps/rejected": -233.7421875,
"loss": 0.422,
"rewards/chosen": 0.27029621601104736,
"rewards/margins": 3.8977864503860475,
"rewards/rejected": -3.627490234375,
"step": 374
},
{
"epoch": 0.631578947368421,
"grad_norm": 59.37971841261376,
"kl": 0.0,
"learning_rate": 1.8175069672409475e-07,
"logits/chosen": -564849280.0,
"logits/rejected": -890142080.0,
"logps/chosen": -229.4367218017578,
"logps/rejected": -251.96417236328125,
"loss": 0.4461,
"rewards/chosen": -2.096214532852173,
"rewards/margins": -0.3813058137893677,
"rewards/rejected": -1.7149087190628052,
"step": 375
},
{
"epoch": 0.6332631578947369,
"grad_norm": 42.11784105230497,
"kl": 0.0,
"learning_rate": 1.8033697085587696e-07,
"logits/chosen": -910276096.0,
"logits/rejected": -714325845.3333334,
"logps/chosen": -231.393359375,
"logps/rejected": -271.1864420572917,
"loss": 0.4749,
"rewards/chosen": -0.2165008544921875,
"rewards/margins": 1.6030135790506999,
"rewards/rejected": -1.8195144335428874,
"step": 376
},
{
"epoch": 0.6349473684210526,
"grad_norm": 38.23749676627054,
"kl": 0.0,
"learning_rate": 1.7892565610492826e-07,
"logits/chosen": -592622421.3333334,
"logits/rejected": -968836416.0,
"logps/chosen": -251.1493937174479,
"logps/rejected": -316.63690185546875,
"loss": 0.4272,
"rewards/chosen": 0.3538096348444621,
"rewards/margins": 1.5449076096216838,
"rewards/rejected": -1.1910979747772217,
"step": 377
},
{
"epoch": 0.6366315789473684,
"grad_norm": 33.65033547018444,
"kl": 0.0,
"learning_rate": 1.775168013184694e-07,
"logits/chosen": -1084484352.0,
"logits/rejected": -1348099840.0,
"logps/chosen": -294.9368896484375,
"logps/rejected": -199.22079467773438,
"loss": 0.3069,
"rewards/chosen": -0.5554192066192627,
"rewards/margins": 2.1953020095825195,
"rewards/rejected": -2.7507212162017822,
"step": 378
},
{
"epoch": 0.6383157894736842,
"grad_norm": 57.13484688248307,
"kl": 0.0,
"learning_rate": 1.7611045525857898e-07,
"logits/chosen": -878904466.2857143,
"logits/rejected": -939131328.0,
"logps/chosen": -230.76688058035714,
"logps/rejected": -233.318603515625,
"loss": 0.3319,
"rewards/chosen": 0.6585009438650948,
"rewards/margins": 0.9025790137904031,
"rewards/rejected": -0.24407806992530823,
"step": 379
},
{
"epoch": 0.64,
"grad_norm": 75.99248678640355,
"kl": 0.0,
"learning_rate": 1.7470666660050584e-07,
"logits/chosen": -838683050.6666666,
"logits/rejected": -667947072.0,
"logps/chosen": -205.16400146484375,
"logps/rejected": -321.7478332519531,
"loss": 0.3082,
"rewards/chosen": 1.483425776163737,
"rewards/margins": 1.0787352124849956,
"rewards/rejected": 0.40469056367874146,
"step": 380
},
{
"epoch": 0.6416842105263157,
"grad_norm": 41.23482970655455,
"kl": 0.0,
"learning_rate": 1.7330548393098403e-07,
"logits/chosen": -790441728.0,
"logits/rejected": -1023563200.0,
"logps/chosen": -277.3321533203125,
"logps/rejected": -204.98135375976562,
"loss": 0.3465,
"rewards/chosen": 0.8123489618301392,
"rewards/margins": 4.149082779884338,
"rewards/rejected": -3.336733818054199,
"step": 381
},
{
"epoch": 0.6433684210526316,
"grad_norm": 61.65012784187018,
"kl": 0.0,
"learning_rate": 1.7190695574655144e-07,
"logits/chosen": -850949056.0,
"logits/rejected": -890390656.0,
"logps/chosen": -191.81503295898438,
"logps/rejected": -262.1980285644531,
"loss": 0.4055,
"rewards/chosen": 2.1015682220458984,
"rewards/margins": 2.900312066078186,
"rewards/rejected": -0.7987438440322876,
"step": 382
},
{
"epoch": 0.6450526315789473,
"grad_norm": 40.46551574043127,
"kl": 0.0,
"learning_rate": 1.705111304518712e-07,
"logits/chosen": -800098986.6666666,
"logits/rejected": -1169969920.0,
"logps/chosen": -194.7923787434896,
"logps/rejected": -145.28848266601562,
"loss": 0.2469,
"rewards/chosen": 2.726773262023926,
"rewards/margins": 4.934856414794922,
"rewards/rejected": -2.208083152770996,
"step": 383
},
{
"epoch": 0.6467368421052632,
"grad_norm": 43.106097272592606,
"kl": 0.0,
"learning_rate": 1.6911805635805632e-07,
"logits/chosen": -827328576.0,
"logits/rejected": -977397696.0,
"logps/chosen": -231.96633911132812,
"logps/rejected": -252.32371520996094,
"loss": 0.3401,
"rewards/chosen": 0.41304701566696167,
"rewards/margins": 1.2065090537071228,
"rewards/rejected": -0.7934620380401611,
"step": 384
},
{
"epoch": 0.6484210526315789,
"grad_norm": 42.75933123532293,
"kl": 0.0,
"learning_rate": 1.677277816809975e-07,
"logits/chosen": -1216948864.0,
"logits/rejected": -623093888.0,
"logps/chosen": -240.3336181640625,
"logps/rejected": -257.3775939941406,
"loss": 0.3767,
"rewards/chosen": 1.1069988012313843,
"rewards/margins": 4.4674311876297,
"rewards/rejected": -3.3604323863983154,
"step": 385
},
{
"epoch": 0.6501052631578947,
"grad_norm": 41.09667706031157,
"kl": 0.0,
"learning_rate": 1.6634035453969457e-07,
"logits/chosen": -689569740.8,
"logits/rejected": -782888533.3333334,
"logps/chosen": -218.35732421875,
"logps/rejected": -258.15380859375,
"loss": 0.3759,
"rewards/chosen": 1.433096218109131,
"rewards/margins": 1.4153085629145306,
"rewards/rejected": 0.017787655194600422,
"step": 386
},
{
"epoch": 0.6517894736842105,
"grad_norm": 54.55329906409335,
"kl": 0.0,
"learning_rate": 1.6495582295459078e-07,
"logits/chosen": -605408320.0,
"logits/rejected": -774927232.0,
"logps/chosen": -204.93194580078125,
"logps/rejected": -248.58714294433594,
"loss": 0.3717,
"rewards/chosen": 0.8562226891517639,
"rewards/margins": 2.2260130047798157,
"rewards/rejected": -1.3697903156280518,
"step": 387
},
{
"epoch": 0.6534736842105263,
"grad_norm": 29.797849697928793,
"kl": 0.0,
"learning_rate": 1.6357423484591087e-07,
"logits/chosen": -766515712.0,
"logits/rejected": -736909653.3333334,
"logps/chosen": -210.6973388671875,
"logps/rejected": -224.88907877604166,
"loss": 0.312,
"rewards/chosen": 0.03613958358764648,
"rewards/margins": 0.5231950441996257,
"rewards/rejected": -0.4870554606119792,
"step": 388
},
{
"epoch": 0.655157894736842,
"grad_norm": 64.33167682170762,
"kl": 0.0,
"learning_rate": 1.621956380320027e-07,
"logits/rejected": -840579712.0,
"logps/rejected": -260.83721923828125,
"loss": 0.4427,
"rewards/rejected": -0.597730278968811,
"step": 389
},
{
"epoch": 0.6568421052631579,
"grad_norm": 42.22779651027316,
"kl": 0.0,
"learning_rate": 1.608200802276815e-07,
"logits/chosen": -596097536.0,
"logits/rejected": -820432298.6666666,
"logps/chosen": -243.89404296875,
"logps/rejected": -213.07832845052084,
"loss": 0.3828,
"rewards/chosen": 1.3395811080932618,
"rewards/margins": 0.6359276135762534,
"rewards/rejected": 0.7036534945170084,
"step": 390
},
{
"epoch": 0.6585263157894737,
"grad_norm": 38.06012726917251,
"kl": 0.0,
"learning_rate": 1.5944760904257942e-07,
"logits/chosen": -941375040.0,
"logits/rejected": -882885802.6666666,
"logps/chosen": -157.46560668945312,
"logps/rejected": -215.0748291015625,
"loss": 0.3784,
"rewards/chosen": 0.9262268543243408,
"rewards/margins": 0.7893735567728678,
"rewards/rejected": 0.136853297551473,
"step": 391
},
{
"epoch": 0.6602105263157895,
"grad_norm": 40.676015795529814,
"kl": 0.0,
"learning_rate": 1.5807827197949685e-07,
"logits/chosen": -708315968.0,
"logits/rejected": -1063172181.3333334,
"logps/chosen": -245.2415771484375,
"logps/rejected": -261.03733317057294,
"loss": 0.4446,
"rewards/chosen": 1.1207358837127686,
"rewards/margins": 0.9106475313504537,
"rewards/rejected": 0.21008835236231485,
"step": 392
},
{
"epoch": 0.6618947368421053,
"grad_norm": 120.64719239721539,
"kl": 0.0,
"learning_rate": 1.5671211643275875e-07,
"logits/chosen": -633163648.0,
"logits/rejected": -805495872.0,
"logps/chosen": -236.74386596679688,
"logps/rejected": -252.68606567382812,
"loss": 0.4374,
"rewards/chosen": 0.5052295923233032,
"rewards/margins": -0.2789260149002075,
"rewards/rejected": 0.7841556072235107,
"step": 393
},
{
"epoch": 0.663578947368421,
"grad_norm": 34.633986243971194,
"kl": 0.0,
"learning_rate": 1.553491896865742e-07,
"logits/chosen": -1026384000.0,
"logits/rejected": -597304448.0,
"logps/chosen": -198.23861694335938,
"logps/rejected": -220.9433135986328,
"loss": 0.3553,
"rewards/chosen": 2.9143166542053223,
"rewards/margins": 1.0535552501678467,
"rewards/rejected": 1.8607614040374756,
"step": 394
},
{
"epoch": 0.6652631578947369,
"grad_norm": 45.21200734298712,
"kl": 0.0,
"learning_rate": 1.5398953891339968e-07,
"logits/chosen": -1087951360.0,
"logits/rejected": -763728640.0,
"logps/chosen": -215.31341552734375,
"logps/rejected": -249.522705078125,
"loss": 0.3877,
"rewards/chosen": 0.6071091294288635,
"rewards/margins": 0.3096790115038554,
"rewards/rejected": 0.2974301179250081,
"step": 395
},
{
"epoch": 0.6669473684210526,
"grad_norm": 53.142787819676414,
"kl": 0.0,
"learning_rate": 1.5263321117230655e-07,
"logits/chosen": -674901376.0,
"logits/rejected": -1003873728.0,
"logps/chosen": -160.89675903320312,
"logps/rejected": -298.0811767578125,
"loss": 0.2986,
"rewards/chosen": 0.6914758682250977,
"rewards/margins": 3.5794568061828613,
"rewards/rejected": -2.8879809379577637,
"step": 396
},
{
"epoch": 0.6686315789473685,
"grad_norm": 48.05043349132663,
"kl": 2.46405029296875,
"learning_rate": 1.512802534073522e-07,
"logits/chosen": -578873856.0,
"logits/rejected": -697750656.0,
"logps/chosen": -232.189501953125,
"logps/rejected": -211.70320638020834,
"loss": 0.367,
"rewards/chosen": 1.3567285537719727,
"rewards/margins": 1.517380674680074,
"rewards/rejected": -0.1606521209081014,
"step": 397
},
{
"epoch": 0.6703157894736842,
"grad_norm": 62.66713109876527,
"kl": 0.0,
"learning_rate": 1.4993071244595534e-07,
"logits/chosen": -1050684544.0,
"logits/rejected": -831061440.0,
"logps/chosen": -178.69546508789062,
"logps/rejected": -267.62298583984375,
"loss": 0.4321,
"rewards/chosen": -0.0750776082277298,
"rewards/margins": 1.4986597746610641,
"rewards/rejected": -1.573737382888794,
"step": 398
},
{
"epoch": 0.672,
"grad_norm": 31.500731382305798,
"kl": 0.0,
"learning_rate": 1.4858463499727507e-07,
"logits/chosen": -985857280.0,
"logits/rejected": -838302720.0,
"logps/chosen": -201.73216247558594,
"logps/rejected": -263.0357666015625,
"loss": 0.3035,
"rewards/chosen": 1.0928525924682617,
"rewards/margins": 3.4981284141540527,
"rewards/rejected": -2.405275821685791,
"step": 399
},
{
"epoch": 0.6736842105263158,
"grad_norm": 56.975473997069834,
"kl": 0.0,
"learning_rate": 1.4724206765059454e-07,
"logits/chosen": -706666624.0,
"logits/rejected": -940830976.0,
"logps/chosen": -226.00015258789062,
"logps/rejected": -203.8423868815104,
"loss": 0.4627,
"rewards/chosen": 0.5999778509140015,
"rewards/margins": -0.3421023289362589,
"rewards/rejected": 0.9420801798502604,
"step": 400
},
{
"epoch": 0.6753684210526316,
"grad_norm": 38.51367993065121,
"kl": 0.0,
"learning_rate": 1.459030568737081e-07,
"logits/chosen": -688410496.0,
"logits/rejected": -619393088.0,
"logps/chosen": -228.07568359375,
"logps/rejected": -262.47265625,
"loss": 0.3821,
"rewards/chosen": -0.5633602738380432,
"rewards/margins": -0.32700619101524353,
"rewards/rejected": -0.23635408282279968,
"step": 401
},
{
"epoch": 0.6770526315789474,
"grad_norm": 48.63664615883306,
"kl": 0.0,
"learning_rate": 1.4456764901131308e-07,
"logits/chosen": -785745664.0,
"logits/rejected": -527159765.3333333,
"logps/chosen": -361.95355224609375,
"logps/rejected": -176.71563720703125,
"loss": 0.4017,
"rewards/chosen": 1.7527008056640625,
"rewards/margins": 5.218010902404785,
"rewards/rejected": -3.4653100967407227,
"step": 402
},
{
"epoch": 0.6787368421052632,
"grad_norm": 42.63233832130583,
"kl": 0.0,
"learning_rate": 1.4323589028340597e-07,
"logits/chosen": -542379605.3333334,
"logits/rejected": -1274020224.0,
"logps/chosen": -244.51863606770834,
"logps/rejected": -148.19590759277344,
"loss": 0.3506,
"rewards/chosen": 1.1265826225280762,
"rewards/margins": 0.32805508375167847,
"rewards/rejected": 0.7985275387763977,
"step": 403
},
{
"epoch": 0.6804210526315789,
"grad_norm": 61.18182583064187,
"kl": 0.0,
"learning_rate": 1.4190782678368257e-07,
"logits/chosen": -793320960.0,
"logits/rejected": -901524684.8,
"logps/chosen": -240.1772664388021,
"logps/rejected": -349.7687255859375,
"loss": 0.3981,
"rewards/chosen": 0.828582763671875,
"rewards/margins": 1.3413378715515136,
"rewards/rejected": -0.5127551078796386,
"step": 404
},
{
"epoch": 0.6821052631578948,
"grad_norm": 44.427082645231216,
"kl": 0.0,
"learning_rate": 1.4058350447794235e-07,
"logits/chosen": -734803200.0,
"logits/rejected": -947461184.0,
"logps/chosen": -209.5147705078125,
"logps/rejected": -400.060302734375,
"loss": 0.4264,
"rewards/chosen": 0.879941463470459,
"rewards/margins": 15.826271533966064,
"rewards/rejected": -14.946330070495605,
"step": 405
},
{
"epoch": 0.6837894736842105,
"grad_norm": 33.804563029238984,
"kl": 0.0,
"learning_rate": 1.3926296920249793e-07,
"logits/chosen": -536416288.0,
"logits/rejected": -671906624.0,
"logps/chosen": -211.46499633789062,
"logps/rejected": -246.7613067626953,
"loss": 0.3206,
"rewards/chosen": 1.6434662342071533,
"rewards/margins": 5.484874725341797,
"rewards/rejected": -3.8414084911346436,
"step": 406
},
{
"epoch": 0.6854736842105263,
"grad_norm": 41.68975540862706,
"kl": 0.0,
"learning_rate": 1.3794626666258866e-07,
"logits/chosen": -946841600.0,
"logits/rejected": -811282005.3333334,
"logps/chosen": -278.705859375,
"logps/rejected": -203.1152547200521,
"loss": 0.4434,
"rewards/chosen": -1.5871347427368163,
"rewards/margins": -2.073893650372823,
"rewards/rejected": 0.48675890763600665,
"step": 407
},
{
"epoch": 0.6871578947368421,
"grad_norm": 37.54233790104561,
"kl": 0.0,
"learning_rate": 1.3663344243079806e-07,
"logits/chosen": -1188021120.0,
"logits/rejected": -819954395.4285715,
"logps/chosen": -194.07516479492188,
"logps/rejected": -175.0321044921875,
"loss": 0.3924,
"rewards/chosen": 2.3552887439727783,
"rewards/margins": 1.9132578032357352,
"rewards/rejected": 0.44203094073704313,
"step": 408
},
{
"epoch": 0.6888421052631579,
"grad_norm": 49.55293923346674,
"kl": 0.0,
"learning_rate": 1.3532454194547732e-07,
"logits/chosen": -869699072.0,
"logits/rejected": -755703552.0,
"logps/chosen": -211.62156677246094,
"logps/rejected": -255.9964141845703,
"loss": 0.2897,
"rewards/chosen": 2.3169875144958496,
"rewards/margins": 3.9210219383239746,
"rewards/rejected": -1.604034423828125,
"step": 409
},
{
"epoch": 0.6905263157894737,
"grad_norm": 45.07517151760579,
"kl": 0.0,
"learning_rate": 1.3401961050917228e-07,
"logits/chosen": -1723337728.0,
"logits/rejected": -646925004.8,
"logps/chosen": -234.47102864583334,
"logps/rejected": -205.66240234375,
"loss": 0.3623,
"rewards/chosen": 0.6976583003997803,
"rewards/margins": 0.9717099666595459,
"rewards/rejected": -0.2740516662597656,
"step": 410
},
{
"epoch": 0.6922105263157895,
"grad_norm": 41.62482913947959,
"kl": 0.0,
"learning_rate": 1.3271869328705516e-07,
"logits/chosen": -539952554.6666666,
"logits/rejected": -892416409.6,
"logps/chosen": -198.6785888671875,
"logps/rejected": -201.31024169921875,
"loss": 0.3754,
"rewards/chosen": 1.5827067693074544,
"rewards/margins": 3.3067630132039385,
"rewards/rejected": -1.7240562438964844,
"step": 411
},
{
"epoch": 0.6938947368421052,
"grad_norm": 42.26855513911886,
"kl": 0.0,
"learning_rate": 1.314218353053619e-07,
"logits/chosen": -660784768.0,
"logits/rejected": -657162342.4,
"logps/chosen": -219.67671712239584,
"logps/rejected": -248.3271484375,
"loss": 0.3954,
"rewards/chosen": 0.8856130441029867,
"rewards/margins": 2.3070303757985435,
"rewards/rejected": -1.4214173316955567,
"step": 412
},
{
"epoch": 0.6955789473684211,
"grad_norm": 63.372635671964915,
"kl": 0.0,
"learning_rate": 1.301290814498335e-07,
"logits/chosen": -682650026.6666666,
"logits/rejected": -750546227.2,
"logps/chosen": -216.4493408203125,
"logps/rejected": -230.68525390625,
"loss": 0.3873,
"rewards/chosen": 2.892109235127767,
"rewards/margins": 3.49957750638326,
"rewards/rejected": -0.6074682712554932,
"step": 413
},
{
"epoch": 0.6972631578947368,
"grad_norm": 31.557969726141994,
"kl": 0.0,
"learning_rate": 1.2884047646416205e-07,
"logits/chosen": -455988778.6666667,
"logits/rejected": -1020206796.8,
"logps/chosen": -256.14182535807294,
"logps/rejected": -181.9263916015625,
"loss": 0.2895,
"rewards/chosen": 0.852165699005127,
"rewards/margins": 0.872660094499588,
"rewards/rejected": -0.02049439549446106,
"step": 414
},
{
"epoch": 0.6989473684210527,
"grad_norm": 44.99525550241327,
"kl": 0.0,
"learning_rate": 1.2755606494844294e-07,
"logits/chosen": -619836032.0,
"logits/rejected": -689233024.0,
"logps/chosen": -243.01300048828125,
"logps/rejected": -235.70297241210938,
"loss": 0.3066,
"rewards/chosen": 1.5389785766601562,
"rewards/margins": 2.8424285650253296,
"rewards/rejected": -1.3034499883651733,
"step": 415
},
{
"epoch": 0.7006315789473684,
"grad_norm": 46.608977034452934,
"kl": 0.0,
"learning_rate": 1.2627589135763066e-07,
"logits/chosen": -622393664.0,
"logits/rejected": -588082816.0,
"logps/chosen": -263.4278869628906,
"logps/rejected": -237.45460510253906,
"loss": 0.3805,
"rewards/chosen": 0.9929741621017456,
"rewards/margins": 2.5076760053634644,
"rewards/rejected": -1.5147018432617188,
"step": 416
},
{
"epoch": 0.7023157894736842,
"grad_norm": 45.15173225020037,
"kl": 0.0,
"learning_rate": 1.2500000000000005e-07,
"logits/chosen": -290351680.0,
"logits/rejected": -570943590.4,
"logps/chosen": -210.52325439453125,
"logps/rejected": -230.669921875,
"loss": 0.3545,
"rewards/chosen": 1.303624947865804,
"rewards/margins": 1.4630508740743,
"rewards/rejected": -0.1594259262084961,
"step": 417
},
{
"epoch": 0.704,
"grad_norm": 49.827214101384186,
"kl": 0.0,
"learning_rate": 1.2372843503561318e-07,
"logits/chosen": -860238028.8,
"logits/rejected": -1133508778.6666667,
"logps/chosen": -212.53994140625,
"logps/rejected": -191.24481201171875,
"loss": 0.3653,
"rewards/chosen": 1.4028396606445312,
"rewards/margins": 1.0293253660202026,
"rewards/rejected": 0.3735142946243286,
"step": 418
},
{
"epoch": 0.7056842105263158,
"grad_norm": 41.39161745285879,
"kl": 0.0,
"learning_rate": 1.2246124047479074e-07,
"logits/chosen": -1557299200.0,
"logits/rejected": -1170218752.0,
"logps/chosen": -264.32098388671875,
"logps/rejected": -288.19140625,
"loss": 0.2715,
"rewards/chosen": -0.8450756072998047,
"rewards/margins": 5.187082767486572,
"rewards/rejected": -6.032158374786377,
"step": 419
},
{
"epoch": 0.7073684210526315,
"grad_norm": 55.33615495444503,
"kl": 0.0,
"learning_rate": 1.211984601765884e-07,
"logits/chosen": -1000780970.6666666,
"logits/rejected": -1060321177.6,
"logps/chosen": -244.44415283203125,
"logps/rejected": -210.060546875,
"loss": 0.4162,
"rewards/chosen": 0.8905111153920492,
"rewards/margins": 1.1921974500020345,
"rewards/rejected": -0.30168633460998534,
"step": 420
},
{
"epoch": 0.7090526315789474,
"grad_norm": 48.50467903033437,
"kl": 0.0,
"learning_rate": 1.1994013784727946e-07,
"logits/chosen": -798935961.6,
"logits/rejected": -556915968.0,
"logps/chosen": -193.3767333984375,
"logps/rejected": -204.73103841145834,
"loss": 0.3481,
"rewards/chosen": 1.7989086151123046,
"rewards/margins": 0.6901429335276286,
"rewards/rejected": 1.108765681584676,
"step": 421
},
{
"epoch": 0.7107368421052631,
"grad_norm": 43.771933039570335,
"kl": 0.0,
"learning_rate": 1.1868631703884183e-07,
"logits/chosen": -770181836.8,
"logits/rejected": -566833578.6666666,
"logps/chosen": -239.477392578125,
"logps/rejected": -194.196533203125,
"loss": 0.3706,
"rewards/chosen": 0.2928351402282715,
"rewards/margins": -1.4519967397054034,
"rewards/rejected": 1.744831879933675,
"step": 422
},
{
"epoch": 0.712421052631579,
"grad_norm": 52.317703020939184,
"kl": 0.0,
"learning_rate": 1.1743704114745029e-07,
"logits/chosen": -1021569408.0,
"logits/rejected": -1079399082.6666667,
"logps/chosen": -271.7890625,
"logps/rejected": -211.10260009765625,
"loss": 0.3508,
"rewards/chosen": 1.666670322418213,
"rewards/margins": 3.195549964904785,
"rewards/rejected": -1.5288796424865723,
"step": 423
},
{
"epoch": 0.7141052631578947,
"grad_norm": 44.73164989384191,
"kl": 0.0,
"learning_rate": 1.1619235341197517e-07,
"logits/chosen": -833912896.0,
"logits/rejected": -611463552.0,
"logps/chosen": -261.9710693359375,
"logps/rejected": -261.09979248046875,
"loss": 0.4409,
"rewards/chosen": -0.13229060173034668,
"rewards/margins": 1.3757647673288982,
"rewards/rejected": -1.5080553690592449,
"step": 424
},
{
"epoch": 0.7157894736842105,
"grad_norm": 43.20707344363169,
"kl": 0.0,
"learning_rate": 1.1495229691248543e-07,
"logits/chosen": -771930453.3333334,
"logits/rejected": -296132800.0,
"logps/chosen": -194.66646321614584,
"logps/rejected": -196.60357666015625,
"loss": 0.3797,
"rewards/chosen": 0.4760231177012126,
"rewards/margins": -0.01816942294438678,
"rewards/rejected": 0.49419254064559937,
"step": 425
},
{
"epoch": 0.7174736842105263,
"grad_norm": 37.98854463213142,
"kl": 0.0,
"learning_rate": 1.1371691456875735e-07,
"logits/chosen": -1023635285.3333334,
"logits/rejected": -639436902.4,
"logps/chosen": -228.6927693684896,
"logps/rejected": -203.55322265625,
"loss": 0.4071,
"rewards/chosen": 3.219998041788737,
"rewards/margins": 3.7488079706827797,
"rewards/rejected": -0.528809928894043,
"step": 426
},
{
"epoch": 0.7191578947368421,
"grad_norm": 38.6260673498228,
"kl": 0.0,
"learning_rate": 1.1248624913878965e-07,
"logits/chosen": -868456618.6666666,
"logits/rejected": -591238860.8,
"logps/chosen": -246.4551798502604,
"logps/rejected": -202.9473876953125,
"loss": 0.4088,
"rewards/chosen": 2.625763416290283,
"rewards/margins": 2.438036632537842,
"rewards/rejected": 0.1877267837524414,
"step": 427
},
{
"epoch": 0.7208421052631578,
"grad_norm": 42.32410569595673,
"kl": 0.0,
"learning_rate": 1.1126034321732325e-07,
"logits/chosen": -789433344.0,
"logits/rejected": -534869043.2,
"logps/chosen": -241.09903971354166,
"logps/rejected": -267.5244384765625,
"loss": 0.3987,
"rewards/chosen": -0.5567977428436279,
"rewards/margins": -1.7331472873687743,
"rewards/rejected": 1.1763495445251464,
"step": 428
},
{
"epoch": 0.7225263157894737,
"grad_norm": 33.14828016562309,
"kl": 0.0,
"learning_rate": 1.100392392343667e-07,
"logits/chosen": -811491712.0,
"logits/rejected": -881575552.0,
"logps/chosen": -205.48550415039062,
"logps/rejected": -208.66879272460938,
"loss": 0.3453,
"rewards/chosen": 1.3217136859893799,
"rewards/margins": 3.224560260772705,
"rewards/rejected": -1.9028465747833252,
"step": 429
},
{
"epoch": 0.7242105263157895,
"grad_norm": 45.216833187005214,
"kl": 0.0,
"learning_rate": 1.0882297945372828e-07,
"logits/chosen": -879807385.6,
"logits/rejected": -1048175104.0,
"logps/chosen": -217.96474609375,
"logps/rejected": -228.4263916015625,
"loss": 0.2549,
"rewards/chosen": 2.451664924621582,
"rewards/margins": 4.334318161010742,
"rewards/rejected": -1.8826532363891602,
"step": 430
},
{
"epoch": 0.7258947368421053,
"grad_norm": 34.528325072499406,
"kl": 0.0,
"learning_rate": 1.0761160597155286e-07,
"logits/chosen": -767435264.0,
"logits/rejected": -604252245.3333334,
"logps/chosen": -238.663525390625,
"logps/rejected": -211.9125773111979,
"loss": 0.4655,
"rewards/chosen": 0.8303826332092286,
"rewards/margins": 3.884237257639567,
"rewards/rejected": -3.0538546244303384,
"step": 431
},
{
"epoch": 0.7275789473684211,
"grad_norm": 71.46182969123905,
"kl": 0.0,
"learning_rate": 1.0640516071486466e-07,
"logits/chosen": -645191936.0,
"logits/rejected": -718282432.0,
"logps/chosen": -243.2042236328125,
"logps/rejected": -196.67498779296875,
"loss": 0.484,
"rewards/chosen": 0.9913086295127869,
"rewards/margins": 0.5928112268447876,
"rewards/rejected": 0.39849740266799927,
"step": 432
},
{
"epoch": 0.7292631578947368,
"grad_norm": 37.01038450064629,
"kl": 0.0,
"learning_rate": 1.052036854401166e-07,
"logits/chosen": -916742485.3333334,
"logits/rejected": -1014475161.6,
"logps/chosen": -251.3818359375,
"logps/rejected": -294.1171630859375,
"loss": 0.3733,
"rewards/chosen": 2.161475658416748,
"rewards/margins": 5.233707332611084,
"rewards/rejected": -3.072231674194336,
"step": 433
},
{
"epoch": 0.7309473684210527,
"grad_norm": 47.994445942830765,
"kl": 0.0,
"learning_rate": 1.0400722173174489e-07,
"logits/chosen": -653794218.6666666,
"logits/rejected": -922273920.0,
"logps/chosen": -217.94051106770834,
"logps/rejected": -335.4775390625,
"loss": 0.3971,
"rewards/chosen": 1.1465115547180176,
"rewards/margins": 9.679410457611084,
"rewards/rejected": -8.532898902893066,
"step": 434
},
{
"epoch": 0.7326315789473684,
"grad_norm": 31.242483556300705,
"kl": 0.0,
"learning_rate": 1.0281581100072939e-07,
"logits/chosen": -1153430144.0,
"logits/rejected": -658751168.0,
"logps/chosen": -223.9492950439453,
"logps/rejected": -250.08029174804688,
"loss": 0.3316,
"rewards/chosen": -0.05509984493255615,
"rewards/margins": 2.957166314125061,
"rewards/rejected": -3.012266159057617,
"step": 435
},
{
"epoch": 0.7343157894736843,
"grad_norm": 44.882205337938544,
"kl": 0.0,
"learning_rate": 1.0162949448316088e-07,
"logits/chosen": -1179825561.6,
"logits/rejected": -795787349.3333334,
"logps/chosen": -184.4357421875,
"logps/rejected": -319.9234619140625,
"loss": 0.4252,
"rewards/chosen": 1.6069551467895509,
"rewards/margins": 3.3702004114786783,
"rewards/rejected": -1.7632452646891277,
"step": 436
},
{
"epoch": 0.736,
"grad_norm": 53.02185490371656,
"kl": 0.0,
"learning_rate": 1.0044831323881357e-07,
"logits/chosen": -786681753.6,
"logits/rejected": -730507520.0,
"logps/chosen": -197.95892333984375,
"logps/rejected": -195.9046630859375,
"loss": 0.3626,
"rewards/chosen": 0.8055712699890136,
"rewards/margins": 2.56037015914917,
"rewards/rejected": -1.7547988891601562,
"step": 437
},
{
"epoch": 0.7376842105263158,
"grad_norm": 41.26140691549541,
"kl": 0.0,
"learning_rate": 9.927230814972382e-08,
"logits/chosen": -1242336402.2857144,
"logits/rejected": -715042368.0,
"logps/chosen": -253.29446847098214,
"logps/rejected": -256.8112487792969,
"loss": 0.4456,
"rewards/chosen": 1.5902636391775948,
"rewards/margins": 5.37732127734593,
"rewards/rejected": -3.787057638168335,
"step": 438
},
{
"epoch": 0.7393684210526316,
"grad_norm": 39.11479149303325,
"kl": 0.0,
"learning_rate": 9.81015199187753e-08,
"logits/chosen": -1125344256.0,
"logits/rejected": -827617536.0,
"logps/chosen": -162.90849812825522,
"logps/rejected": -385.02578125,
"loss": 0.2965,
"rewards/chosen": 2.1140575408935547,
"rewards/margins": 4.867137718200683,
"rewards/rejected": -2.7530801773071287,
"step": 439
},
{
"epoch": 0.7410526315789474,
"grad_norm": 42.334848723588486,
"kl": 0.0,
"learning_rate": 9.693598906829046e-08,
"logits/chosen": -1135330304.0,
"logits/rejected": -764287616.0,
"logps/chosen": -208.2585652669271,
"logps/rejected": -316.421875,
"loss": 0.4066,
"rewards/chosen": 1.152496576309204,
"rewards/margins": 4.668377876281738,
"rewards/rejected": -3.515881299972534,
"step": 440
},
{
"epoch": 0.7427368421052631,
"grad_norm": 41.258031701194476,
"kl": 0.0,
"learning_rate": 9.577575593862775e-08,
"logits/chosen": -715377024.0,
"logits/rejected": -1129872384.0,
"logps/chosen": -234.7706298828125,
"logps/rejected": -306.43499755859375,
"loss": 0.3972,
"rewards/chosen": 1.0637226104736328,
"rewards/margins": 2.9347736835479736,
"rewards/rejected": -1.8710510730743408,
"step": 441
},
{
"epoch": 0.744421052631579,
"grad_norm": 54.07697190099627,
"kl": 0.0,
"learning_rate": 9.462086068678519e-08,
"logits/chosen": -766640000.0,
"logits/rejected": -959083136.0,
"logps/chosen": -211.8131103515625,
"logps/rejected": -283.1666259765625,
"loss": 0.401,
"rewards/chosen": -0.5408859252929688,
"rewards/margins": 1.790869951248169,
"rewards/rejected": -2.3317558765411377,
"step": 442
},
{
"epoch": 0.7461052631578947,
"grad_norm": 66.5169402578717,
"kl": 0.0,
"learning_rate": 9.347134328501097e-08,
"logits/chosen": -653176448.0,
"logits/rejected": -1222430976.0,
"logps/chosen": -176.54693603515625,
"logps/rejected": -252.2942657470703,
"loss": 0.3954,
"rewards/chosen": -0.43219953775405884,
"rewards/margins": 0.7258930802345276,
"rewards/rejected": -1.1580926179885864,
"step": 443
},
{
"epoch": 0.7477894736842106,
"grad_norm": 59.0005118657837,
"kl": 0.0,
"learning_rate": 9.232724351941978e-08,
"logits/chosen": -756462080.0,
"logits/rejected": -722251946.6666666,
"logps/chosen": -305.7142333984375,
"logps/rejected": -234.7782999674479,
"loss": 0.3794,
"rewards/chosen": 0.13024139404296875,
"rewards/margins": 0.5876843929290771,
"rewards/rejected": -0.4574429988861084,
"step": 444
},
{
"epoch": 0.7494736842105263,
"grad_norm": 55.62210680654247,
"kl": 0.0,
"learning_rate": 9.118860098861537e-08,
"logits/chosen": -813152153.6,
"logits/rejected": -290959360.0,
"logps/chosen": -197.15120849609374,
"logps/rejected": -282.9163004557292,
"loss": 0.4162,
"rewards/chosen": 0.6745959281921386,
"rewards/margins": 1.0792610804239908,
"rewards/rejected": -0.40466515223185223,
"step": 445
},
{
"epoch": 0.7511578947368421,
"grad_norm": 358.03078972226217,
"kl": 0.0,
"learning_rate": 9.005545510232068e-08,
"logits/chosen": -854770614.8571428,
"logits/rejected": -522669440.0,
"logps/chosen": -227.82121930803572,
"logps/rejected": -329.3843994140625,
"loss": 0.2992,
"rewards/chosen": 1.0909557342529297,
"rewards/margins": 7.157499313354492,
"rewards/rejected": -6.0665435791015625,
"step": 446
},
{
"epoch": 0.7528421052631579,
"grad_norm": 40.23280955297014,
"kl": 0.0,
"learning_rate": 8.892784508001341e-08,
"logits/chosen": -665233920.0,
"logits/rejected": -786437529.6,
"logps/chosen": -205.84651692708334,
"logps/rejected": -208.0626220703125,
"loss": 0.3587,
"rewards/chosen": 3.913280804951986,
"rewards/margins": 4.716941197713217,
"rewards/rejected": -0.8036603927612305,
"step": 447
},
{
"epoch": 0.7545263157894737,
"grad_norm": 331.3992420977293,
"kl": 0.0,
"learning_rate": 8.780580994956849e-08,
"logits/chosen": -991909824.0,
"logits/rejected": -1036429738.6666666,
"logps/chosen": -241.2684326171875,
"logps/rejected": -222.25858561197916,
"loss": 0.3923,
"rewards/chosen": 0.7948417663574219,
"rewards/margins": 0.6548532644907633,
"rewards/rejected": 0.13998850186665854,
"step": 448
},
{
"epoch": 0.7562105263157894,
"grad_norm": 33.19053977720104,
"kl": 0.0,
"learning_rate": 8.668938854590763e-08,
"logits/chosen": -644909875.2,
"logits/rejected": -695736832.0,
"logps/chosen": -230.962060546875,
"logps/rejected": -194.96405029296875,
"loss": 0.4064,
"rewards/chosen": 2.0331790924072264,
"rewards/margins": -0.12323621114095085,
"rewards/rejected": 2.1564153035481772,
"step": 449
},
{
"epoch": 0.7578947368421053,
"grad_norm": 45.92452873664749,
"kl": 0.0,
"learning_rate": 8.557861950965509e-08,
"logits/chosen": -1397461930.6666667,
"logits/rejected": -905682124.8,
"logps/chosen": -271.2666422526042,
"logps/rejected": -288.8114013671875,
"loss": 0.3724,
"rewards/chosen": -0.3505045572916667,
"rewards/margins": 2.9490411122639975,
"rewards/rejected": -3.299545669555664,
"step": 450
},
{
"epoch": 0.759578947368421,
"grad_norm": 40.248447099852605,
"kl": 0.0,
"learning_rate": 8.44735412857999e-08,
"logits/chosen": -1039405504.0,
"logits/rejected": -620940032.0,
"logps/chosen": -259.6157531738281,
"logps/rejected": -244.8777618408203,
"loss": 0.3763,
"rewards/chosen": 0.6175476312637329,
"rewards/margins": 3.04443895816803,
"rewards/rejected": -2.426891326904297,
"step": 451
},
{
"epoch": 0.7612631578947369,
"grad_norm": 37.694918257120385,
"kl": 0.0,
"learning_rate": 8.337419212236584e-08,
"logits/chosen": -523443744.0,
"logits/rejected": -701745856.0,
"logps/chosen": -225.3424072265625,
"logps/rejected": -293.00054931640625,
"loss": 0.3628,
"rewards/chosen": 2.430769443511963,
"rewards/margins": 3.699450373649597,
"rewards/rejected": -1.2686809301376343,
"step": 452
},
{
"epoch": 0.7629473684210526,
"grad_norm": 35.15022250675745,
"kl": 0.0,
"learning_rate": 8.228061006908738e-08,
"logits/chosen": -789152448.0,
"logits/rejected": -639658240.0,
"logps/chosen": -204.5469207763672,
"logps/rejected": -296.8457336425781,
"loss": 0.3558,
"rewards/chosen": 0.2700985074043274,
"rewards/margins": 10.158620655536652,
"rewards/rejected": -9.888522148132324,
"step": 453
},
{
"epoch": 0.7646315789473684,
"grad_norm": 47.5088310425883,
"kl": 0.0,
"learning_rate": 8.119283297609236e-08,
"logits/chosen": -1028607744.0,
"logits/rejected": -645622656.0,
"logps/chosen": -289.964111328125,
"logps/rejected": -224.54290771484375,
"loss": 0.3183,
"rewards/chosen": 0.8806011080741882,
"rewards/margins": 1.526337444782257,
"rewards/rejected": -0.6457363367080688,
"step": 454
},
{
"epoch": 0.7663157894736842,
"grad_norm": 34.441433884082336,
"kl": 0.0,
"learning_rate": 8.011089849259262e-08,
"logits/chosen": -749322752.0,
"logps/chosen": -202.4013671875,
"loss": 0.3446,
"rewards/chosen": 0.888677179813385,
"step": 455
},
{
"epoch": 0.768,
"grad_norm": 43.86036582680801,
"kl": 0.0,
"learning_rate": 7.903484406558053e-08,
"logits/chosen": -536868650.6666667,
"logits/rejected": -616575808.0,
"logps/chosen": -276.9130045572917,
"logps/rejected": -232.73355102539062,
"loss": 0.43,
"rewards/chosen": -0.4769521156946818,
"rewards/margins": 0.35738088687260944,
"rewards/rejected": -0.8343330025672913,
"step": 456
},
{
"epoch": 0.7696842105263157,
"grad_norm": 42.12234531239963,
"kl": 0.0,
"learning_rate": 7.79647069385328e-08,
"logits/chosen": -752578133.3333334,
"logits/rejected": -934050816.0,
"logps/chosen": -224.02742513020834,
"logps/rejected": -277.469970703125,
"loss": 0.4204,
"rewards/chosen": -0.5293665726979574,
"rewards/margins": -1.098210604985555,
"rewards/rejected": 0.5688440322875976,
"step": 457
},
{
"epoch": 0.7713684210526316,
"grad_norm": 42.59646325267393,
"kl": 0.0,
"learning_rate": 7.690052415012174e-08,
"logits/chosen": -938434252.8,
"logits/rejected": -1136305237.3333333,
"logps/chosen": -216.8803466796875,
"logps/rejected": -190.39237467447916,
"loss": 0.2499,
"rewards/chosen": 1.1629043579101563,
"rewards/margins": 4.905188814798991,
"rewards/rejected": -3.7422844568888345,
"step": 458
},
{
"epoch": 0.7730526315789473,
"grad_norm": 52.399177497226496,
"kl": 0.0,
"learning_rate": 7.584233253293326e-08,
"logits/chosen": -716863027.2,
"logits/rejected": -816465237.3333334,
"logps/chosen": -243.8283203125,
"logps/rejected": -196.03934733072916,
"loss": 0.2746,
"rewards/chosen": 1.0614694595336913,
"rewards/margins": 2.684300168355306,
"rewards/rejected": -1.6228307088216145,
"step": 459
},
{
"epoch": 0.7747368421052632,
"grad_norm": 52.891481922508966,
"kl": 0.0,
"learning_rate": 7.479016871219173e-08,
"logits/chosen": -917309440.0,
"logits/rejected": -898528870.4,
"logps/chosen": -282.1151936848958,
"logps/rejected": -213.7822998046875,
"loss": 0.4074,
"rewards/chosen": -0.2861933708190918,
"rewards/margins": 1.0193795204162597,
"rewards/rejected": -1.3055728912353515,
"step": 460
},
{
"epoch": 0.7764210526315789,
"grad_norm": 35.541679693506616,
"kl": 0.0,
"learning_rate": 7.374406910449276e-08,
"logits/chosen": -625601749.3333334,
"logits/rejected": -777565235.2,
"logps/chosen": -214.69976806640625,
"logps/rejected": -227.8533447265625,
"loss": 0.3537,
"rewards/chosen": 0.7599695523579916,
"rewards/margins": 2.2433162053426106,
"rewards/rejected": -1.483346652984619,
"step": 461
},
{
"epoch": 0.7781052631578947,
"grad_norm": 63.418725969793606,
"kl": 0.0,
"learning_rate": 7.270406991654273e-08,
"logits/chosen": -963097258.6666666,
"logits/rejected": -1547470848.0,
"logps/chosen": -263.2251383463542,
"logps/rejected": -274.23115234375,
"loss": 0.425,
"rewards/chosen": 0.2878738244374593,
"rewards/margins": 1.6172294457753498,
"rewards/rejected": -1.3293556213378905,
"step": 462
},
{
"epoch": 0.7797894736842105,
"grad_norm": 38.377982473660865,
"kl": 0.0,
"learning_rate": 7.167020714390501e-08,
"logits/chosen": -783421030.4,
"logits/rejected": -488972885.3333333,
"logps/chosen": -215.7775390625,
"logps/rejected": -179.737548828125,
"loss": 0.436,
"rewards/chosen": 0.9828885078430176,
"rewards/margins": 0.3364712874094645,
"rewards/rejected": 0.6464172204335531,
"step": 463
},
{
"epoch": 0.7814736842105263,
"grad_norm": 45.40428262520681,
"kl": 0.0,
"learning_rate": 7.064251656975503e-08,
"logits/chosen": -917309888.0,
"logits/rejected": -730922069.3333334,
"logps/chosen": -207.1556396484375,
"logps/rejected": -247.15775553385416,
"loss": 0.361,
"rewards/chosen": 1.675153374671936,
"rewards/margins": 3.9481043418248496,
"rewards/rejected": -2.2729509671529136,
"step": 464
},
{
"epoch": 0.783157894736842,
"grad_norm": 55.99199517489128,
"kl": 0.0,
"learning_rate": 6.96210337636414e-08,
"logits/chosen": -1108944512.0,
"logits/rejected": -576058944.0,
"logps/chosen": -289.2898864746094,
"logps/rejected": -202.31178283691406,
"loss": 0.3961,
"rewards/chosen": -0.21090269088745117,
"rewards/margins": -0.2718646228313446,
"rewards/rejected": 0.06096193194389343,
"step": 465
},
{
"epoch": 0.7848421052631579,
"grad_norm": 49.25923339661157,
"kl": 0.0,
"learning_rate": 6.860579408025435e-08,
"logits/chosen": -814812672.0,
"logits/rejected": -933722880.0,
"logps/chosen": -191.00594075520834,
"logps/rejected": -235.2879638671875,
"loss": 0.3267,
"rewards/chosen": 1.0561386744181316,
"rewards/margins": 2.257707277933757,
"rewards/rejected": -1.201568603515625,
"step": 466
},
{
"epoch": 0.7865263157894736,
"grad_norm": 52.73809432153172,
"kl": 0.0,
"learning_rate": 6.759683265820293e-08,
"logits/chosen": -749964800.0,
"logits/rejected": -1379509504.0,
"logps/chosen": -182.05026245117188,
"logps/rejected": -213.2086639404297,
"loss": 0.4184,
"rewards/chosen": 1.8726093769073486,
"rewards/margins": 3.3285834789276123,
"rewards/rejected": -1.4559741020202637,
"step": 467
},
{
"epoch": 0.7882105263157895,
"grad_norm": 55.06684904272287,
"kl": 0.0,
"learning_rate": 6.659418441879816e-08,
"logits/chosen": -886685760.0,
"logits/rejected": -867823936.0,
"logps/chosen": -242.12347412109375,
"logps/rejected": -234.9542236328125,
"loss": 0.4054,
"rewards/chosen": 0.7248154282569885,
"rewards/margins": 0.6286347508430481,
"rewards/rejected": 0.09618067741394043,
"step": 468
},
{
"epoch": 0.7898947368421053,
"grad_norm": 41.061950624386746,
"kl": 0.0,
"learning_rate": 6.559788406484445e-08,
"logits/chosen": -1067656874.6666666,
"logits/rejected": -635095756.8,
"logps/chosen": -228.18094889322916,
"logps/rejected": -251.1126708984375,
"loss": 0.3713,
"rewards/chosen": 2.0004892349243164,
"rewards/margins": 4.07933406829834,
"rewards/rejected": -2.0788448333740233,
"step": 469
},
{
"epoch": 0.791578947368421,
"grad_norm": 55.92018991052814,
"kl": 0.0,
"learning_rate": 6.460796607943888e-08,
"logits/chosen": -1008979114.6666666,
"logits/rejected": -580155904.0,
"logps/chosen": -220.10860188802084,
"logps/rejected": -200.3806640625,
"loss": 0.3572,
"rewards/chosen": 0.6659164428710938,
"rewards/margins": 2.2499670028686523,
"rewards/rejected": -1.5840505599975585,
"step": 470
},
{
"epoch": 0.7932631578947369,
"grad_norm": 44.05178707657523,
"kl": 0.0,
"learning_rate": 6.362446472477739e-08,
"logits/chosen": -886915754.6666666,
"logits/rejected": -697056153.6,
"logps/chosen": -161.08222452799478,
"logps/rejected": -195.1838623046875,
"loss": 0.3431,
"rewards/chosen": 2.043999671936035,
"rewards/margins": 1.8765502452850342,
"rewards/rejected": 0.167449426651001,
"step": 471
},
{
"epoch": 0.7949473684210526,
"grad_norm": 76.34832400410619,
"kl": 0.0,
"learning_rate": 6.264741404096873e-08,
"logits/chosen": -970343594.6666666,
"logits/rejected": -1028333772.8,
"logps/chosen": -240.02298990885416,
"logps/rejected": -181.0073486328125,
"loss": 0.4382,
"rewards/chosen": 1.713447093963623,
"rewards/margins": 0.9383099555969239,
"rewards/rejected": 0.7751371383666992,
"step": 472
},
{
"epoch": 0.7966315789473685,
"grad_norm": 53.36902208721296,
"kl": 0.0,
"learning_rate": 6.16768478448568e-08,
"logits/chosen": -770189056.0,
"logits/rejected": -784716697.6,
"logps/chosen": -198.6039021809896,
"logps/rejected": -230.421240234375,
"loss": 0.4316,
"rewards/chosen": 1.7407466570536296,
"rewards/margins": 0.8216514269510906,
"rewards/rejected": 0.919095230102539,
"step": 473
},
{
"epoch": 0.7983157894736842,
"grad_norm": 76.44940846297158,
"kl": 0.0,
"learning_rate": 6.071279972884996e-08,
"logits/chosen": -650762581.3333334,
"logits/rejected": -824322304.0,
"logps/chosen": -271.9395345052083,
"logps/rejected": -240.26007080078125,
"loss": 0.374,
"rewards/chosen": 0.2647084792455037,
"rewards/margins": 1.6960653861363728,
"rewards/rejected": -1.4313569068908691,
"step": 474
},
{
"epoch": 0.8,
"grad_norm": 38.84706884248548,
"kl": 0.0,
"learning_rate": 5.975530305975806e-08,
"logits/chosen": -953603481.6,
"logits/rejected": -874571178.6666666,
"logps/chosen": -224.2393310546875,
"logps/rejected": -248.4403076171875,
"loss": 0.3032,
"rewards/chosen": 0.2576324462890625,
"rewards/margins": 2.0944193522135417,
"rewards/rejected": -1.8367869059244792,
"step": 475
},
{
"epoch": 0.8016842105263158,
"grad_norm": 80.57439772482964,
"kl": 0.0,
"learning_rate": 5.88043909776382e-08,
"logits/chosen": -720325705.1428572,
"logits/rejected": -1631839360.0,
"logps/chosen": -212.89100864955358,
"logps/rejected": -245.09353637695312,
"loss": 0.3109,
"rewards/chosen": 0.8418189457484654,
"rewards/margins": 1.0114402196237018,
"rewards/rejected": -0.1696212738752365,
"step": 476
},
{
"epoch": 0.8033684210526316,
"grad_norm": 60.404391161081,
"kl": 0.0,
"learning_rate": 5.786009639464728e-08,
"logits/chosen": -754591168.0,
"logits/rejected": -984781056.0,
"logps/chosen": -213.29608154296875,
"logps/rejected": -191.29742431640625,
"loss": 0.3819,
"rewards/chosen": 1.0502116680145264,
"rewards/margins": 8.452488660812378,
"rewards/rejected": -7.402276992797852,
"step": 477
},
{
"epoch": 0.8050526315789474,
"grad_norm": 39.88427529417202,
"kl": 0.0,
"learning_rate": 5.69224519939028e-08,
"logits/chosen": -1256239872.0,
"logits/rejected": -680008704.0,
"logps/chosen": -207.2843017578125,
"logps/rejected": -281.371337890625,
"loss": 0.4186,
"rewards/chosen": 2.776592969894409,
"rewards/margins": 2.579076017652239,
"rewards/rejected": 0.19751695224217006,
"step": 478
},
{
"epoch": 0.8067368421052632,
"grad_norm": 46.39968561566463,
"kl": 0.0,
"learning_rate": 5.599149022835201e-08,
"logits/chosen": -757022515.2,
"logits/rejected": -568918784.0,
"logps/chosen": -190.7361572265625,
"logps/rejected": -234.748291015625,
"loss": 0.3371,
"rewards/chosen": 1.5194448471069335,
"rewards/margins": 3.035427951812744,
"rewards/rejected": -1.5159831047058105,
"step": 479
},
{
"epoch": 0.8084210526315789,
"grad_norm": 40.26692401415863,
"kl": 0.0,
"learning_rate": 5.5067243319648517e-08,
"logits/chosen": -628281651.2,
"logits/rejected": -631906560.0,
"logps/chosen": -231.4928466796875,
"logps/rejected": -164.00850423177084,
"loss": 0.3951,
"rewards/chosen": 0.1779022455215454,
"rewards/margins": 1.319037381807963,
"rewards/rejected": -1.1411351362864177,
"step": 480
},
{
"epoch": 0.8101052631578948,
"grad_norm": 58.215282154133604,
"kl": 0.0,
"learning_rate": 5.414974325703686e-08,
"logits/chosen": -691516928.0,
"logits/rejected": -898130624.0,
"logps/chosen": -199.2517852783203,
"logps/rejected": -193.02496337890625,
"loss": 0.4195,
"rewards/chosen": 0.9336433410644531,
"rewards/margins": 0.8423346877098083,
"rewards/rejected": 0.09130865335464478,
"step": 481
},
{
"epoch": 0.8117894736842105,
"grad_norm": 50.746031932726225,
"kl": 0.0,
"learning_rate": 5.3239021796245707e-08,
"logits/chosen": -643468480.0,
"logits/rejected": -744267861.3333334,
"logps/chosen": -202.1339111328125,
"logps/rejected": -236.04170735677084,
"loss": 0.3827,
"rewards/chosen": 0.5528717041015625,
"rewards/margins": 0.6056533952554067,
"rewards/rejected": -0.0527816911538442,
"step": 482
},
{
"epoch": 0.8134736842105263,
"grad_norm": 90.38953108333145,
"kl": 0.0,
"learning_rate": 5.2335110458388457e-08,
"logits/chosen": -861357653.3333334,
"logits/rejected": -1220613529.6,
"logps/chosen": -228.4935302734375,
"logps/rejected": -244.1078125,
"loss": 0.3977,
"rewards/chosen": 0.24794367949167886,
"rewards/margins": 1.8121562878290813,
"rewards/rejected": -1.5642126083374024,
"step": 483
},
{
"epoch": 0.8151578947368421,
"grad_norm": 40.4236540400168,
"kl": 0.0,
"learning_rate": 5.1438040528872265e-08,
"logits/chosen": -782614656.0,
"logits/rejected": -757872469.3333334,
"logps/chosen": -136.77349853515625,
"logps/rejected": -224.17281087239584,
"loss": 0.5002,
"rewards/chosen": 1.6036109924316406,
"rewards/margins": 0.9859415292739868,
"rewards/rejected": 0.6176694631576538,
"step": 484
},
{
"epoch": 0.8168421052631579,
"grad_norm": 121.37099361156412,
"kl": 0.0,
"learning_rate": 5.054784305631546e-08,
"logits/chosen": -910532864.0,
"logits/rejected": -786700117.3333334,
"logps/chosen": -177.5673065185547,
"logps/rejected": -266.01979573567706,
"loss": 0.2893,
"rewards/chosen": -0.057147204875946045,
"rewards/margins": 1.8595226407051086,
"rewards/rejected": -1.9166698455810547,
"step": 485
},
{
"epoch": 0.8185263157894737,
"grad_norm": 50.530115528782915,
"kl": 0.0,
"learning_rate": 4.96645488514727e-08,
"logits/chosen": -788890709.3333334,
"logits/rejected": -455711808.0,
"logps/chosen": -256.8352457682292,
"logps/rejected": -267.2051086425781,
"loss": 0.4482,
"rewards/chosen": -0.5332825183868408,
"rewards/margins": 0.43467551469802856,
"rewards/rejected": -0.9679580330848694,
"step": 486
},
{
"epoch": 0.8202105263157895,
"grad_norm": 70.07611011272452,
"kl": 0.0,
"learning_rate": 4.8788188486168616e-08,
"logits/chosen": -872919142.4,
"logits/rejected": -701736832.0,
"logps/chosen": -193.6914794921875,
"logps/rejected": -235.78462727864584,
"loss": 0.3179,
"rewards/chosen": 2.0400518417358398,
"rewards/margins": 2.919521633783976,
"rewards/rejected": -0.8794697920481364,
"step": 487
},
{
"epoch": 0.8218947368421052,
"grad_norm": 52.808019352043246,
"kl": 0.0,
"learning_rate": 4.7918792292239635e-08,
"logits/chosen": -670976153.6,
"logits/rejected": -896994133.3333334,
"logps/chosen": -245.185595703125,
"logps/rejected": -399.8438313802083,
"loss": 0.4311,
"rewards/chosen": 0.799212646484375,
"rewards/margins": 14.883302561442056,
"rewards/rejected": -14.084089914957682,
"step": 488
},
{
"epoch": 0.8235789473684211,
"grad_norm": 54.91548661991697,
"kl": 0.0,
"learning_rate": 4.705639036048439e-08,
"logits/chosen": -952622208.0,
"logits/rejected": -660675968.0,
"logps/chosen": -200.71133422851562,
"logps/rejected": -260.607666015625,
"loss": 0.3739,
"rewards/chosen": 1.0811817646026611,
"rewards/margins": 0.40599095821380615,
"rewards/rejected": 0.675190806388855,
"step": 489
},
{
"epoch": 0.8252631578947368,
"grad_norm": 62.12949703126072,
"kl": 0.0,
"learning_rate": 4.6201012539622054e-08,
"logits/chosen": -576909888.0,
"logits/rejected": -580735232.0,
"logps/chosen": -185.22006225585938,
"logps/rejected": -247.2104695638021,
"loss": 0.3748,
"rewards/chosen": 2.8498971462249756,
"rewards/margins": 3.258001764615377,
"rewards/rejected": -0.4081046183904012,
"step": 490
},
{
"epoch": 0.8269473684210527,
"grad_norm": 44.89977606216685,
"kl": 0.0,
"learning_rate": 4.535268843525908e-08,
"logits/chosen": -561914880.0,
"logits/rejected": -1542454400.0,
"logps/chosen": -279.1746520996094,
"logps/rejected": -252.60659790039062,
"loss": 0.4187,
"rewards/chosen": -0.5238960385322571,
"rewards/margins": -1.0888721942901611,
"rewards/rejected": 0.564976155757904,
"step": 491
},
{
"epoch": 0.8286315789473684,
"grad_norm": 33.62401765550933,
"kl": 0.0,
"learning_rate": 4.4511447408864975e-08,
"logits/chosen": -705677184.0,
"logits/rejected": -642696576.0,
"logps/chosen": -219.29757690429688,
"logps/rejected": -180.13229370117188,
"loss": 0.3157,
"rewards/chosen": 0.919715940952301,
"rewards/margins": 3.393587648868561,
"rewards/rejected": -2.4738717079162598,
"step": 492
},
{
"epoch": 0.8303157894736842,
"grad_norm": 44.80536332639876,
"kl": 0.0,
"learning_rate": 4.367731857675569e-08,
"logits/chosen": -1012552362.6666666,
"logits/rejected": -1070745088.0,
"logps/chosen": -210.34134928385416,
"logps/rejected": -197.56212158203124,
"loss": 0.4666,
"rewards/chosen": 2.4052974383036294,
"rewards/margins": 1.760869280497233,
"rewards/rejected": 0.6444281578063965,
"step": 493
},
{
"epoch": 0.832,
"grad_norm": 69.77716049036187,
"kl": 0.0,
"learning_rate": 4.2850330809085874e-08,
"logits/chosen": -630044373.3333334,
"logits/rejected": -1326352512.0,
"logps/chosen": -251.13602701822916,
"logps/rejected": -142.5716094970703,
"loss": 0.3156,
"rewards/chosen": 0.20519916216532388,
"rewards/margins": -1.417292316754659,
"rewards/rejected": 1.622491478919983,
"step": 494
},
{
"epoch": 0.8336842105263158,
"grad_norm": 36.11198104501435,
"kl": 0.0,
"learning_rate": 4.2030512728849945e-08,
"logits/chosen": -921905493.3333334,
"logits/rejected": -791465369.6,
"logps/chosen": -237.568115234375,
"logps/rejected": -210.0384521484375,
"loss": 0.4533,
"rewards/chosen": 0.15185137589772543,
"rewards/margins": 1.321685512860616,
"rewards/rejected": -1.1698341369628906,
"step": 495
},
{
"epoch": 0.8353684210526315,
"grad_norm": 47.31830867232736,
"kl": 0.0,
"learning_rate": 4.1217892710891126e-08,
"logits/chosen": -688228992.0,
"logits/rejected": -928642474.6666666,
"logps/chosen": -180.07191467285156,
"logps/rejected": -247.9576416015625,
"loss": 0.3971,
"rewards/chosen": 2.725332736968994,
"rewards/margins": 4.285257339477539,
"rewards/rejected": -1.559924602508545,
"step": 496
},
{
"epoch": 0.8370526315789474,
"grad_norm": 58.64030492711,
"kl": 0.0,
"learning_rate": 4.041249888091941e-08,
"logits/chosen": -930103466.6666666,
"logits/rejected": -520454144.0,
"logps/chosen": -223.98262532552084,
"logps/rejected": -291.0794982910156,
"loss": 0.3902,
"rewards/chosen": 0.41166090965270996,
"rewards/margins": 2.5037872791290283,
"rewards/rejected": -2.0921263694763184,
"step": 497
},
{
"epoch": 0.8387368421052631,
"grad_norm": 42.51804510909749,
"kl": 0.0,
"learning_rate": 3.9614359114538196e-08,
"logits/chosen": -594655701.3333334,
"logits/rejected": -1102049280.0,
"logps/chosen": -239.44209798177084,
"logps/rejected": -235.8392578125,
"loss": 0.4429,
"rewards/chosen": -0.807283083597819,
"rewards/margins": -1.4231882254282633,
"rewards/rejected": 0.6159051418304443,
"step": 498
},
{
"epoch": 0.840421052631579,
"grad_norm": 60.0326086702577,
"kl": 0.0,
"learning_rate": 3.8823501036279514e-08,
"logits/chosen": -403930848.0,
"logits/rejected": -417444640.0,
"logps/chosen": -302.0504150390625,
"logps/rejected": -168.2909698486328,
"loss": 0.3504,
"rewards/chosen": 0.14431113004684448,
"rewards/margins": 1.6669214367866516,
"rewards/rejected": -1.5226103067398071,
"step": 499
},
{
"epoch": 0.8421052631578947,
"grad_norm": 44.106324622609606,
"kl": 0.0,
"learning_rate": 3.803995201864762e-08,
"logits/chosen": -775272089.6,
"logits/rejected": -703777536.0,
"logps/chosen": -191.228076171875,
"logps/rejected": -175.0321248372396,
"loss": 0.3079,
"rewards/chosen": 0.6324722290039062,
"rewards/margins": 2.6547363917032873,
"rewards/rejected": -2.0222641626993814,
"step": 500
},
{
"epoch": 0.8421052631578947,
"eval_logits/chosen": -922359676.4770643,
"eval_logits/rejected": -826945158.35461,
"eval_logps/chosen": -217.7674706135321,
"eval_logps/rejected": -228.38810394503545,
"eval_loss": 0.378118097782135,
"eval_rewards/chosen": 0.8047234771448538,
"eval_rewards/margins": 1.8294198813516913,
"eval_rewards/rejected": -1.0246964042068374,
"eval_runtime": 1555.2949,
"eval_samples_per_second": 0.643,
"eval_steps_per_second": 0.161,
"kl": 0.0,
"step": 500
},
{
"epoch": 0.8437894736842105,
"grad_norm": 55.94728204942989,
"kl": 0.0,
"learning_rate": 3.726373918117196e-08,
"logits/chosen": -971137472.0,
"logits/rejected": -1374881280.0,
"logps/chosen": -255.43304443359375,
"logps/rejected": -218.9608612060547,
"loss": 0.4355,
"rewards/chosen": -0.6300251483917236,
"rewards/margins": -1.8687539100646973,
"rewards/rejected": 1.2387287616729736,
"step": 501
},
{
"epoch": 0.8454736842105263,
"grad_norm": 47.235552657321186,
"kl": 0.0,
"learning_rate": 3.6494889389468434e-08,
"logits/chosen": -602287744.0,
"logits/rejected": -920995072.0,
"logps/chosen": -210.18441772460938,
"logps/rejected": -227.16696166992188,
"loss": 0.3343,
"rewards/chosen": 0.2986224889755249,
"rewards/margins": 1.3155208826065063,
"rewards/rejected": -1.0168983936309814,
"step": 502
},
{
"epoch": 0.8471578947368421,
"grad_norm": 49.23690109646782,
"kl": 0.0,
"learning_rate": 3.573342925430925e-08,
"logits/chosen": -778606080.0,
"logits/rejected": -654501376.0,
"logps/chosen": -210.22725423177084,
"logps/rejected": -293.99200439453125,
"loss": 0.373,
"rewards/chosen": 0.8925824165344238,
"rewards/margins": 2.6825954914093018,
"rewards/rejected": -1.790013074874878,
"step": 503
},
{
"epoch": 0.8488421052631578,
"grad_norm": 47.196350435325876,
"kl": 0.0,
"learning_rate": 3.497938513070234e-08,
"logits/chosen": -747158118.4,
"logits/rejected": -715651498.6666666,
"logps/chosen": -196.84251708984374,
"logps/rejected": -247.5506388346354,
"loss": 0.3714,
"rewards/chosen": 1.6987966537475585,
"rewards/margins": 2.460467561086019,
"rewards/rejected": -0.7616709073384603,
"step": 504
},
{
"epoch": 0.8505263157894737,
"grad_norm": 67.84038276109705,
"kl": 0.0,
"learning_rate": 3.423278311697897e-08,
"logits/chosen": -1138272768.0,
"logits/rejected": -622846156.8,
"logps/chosen": -215.9085896809896,
"logps/rejected": -196.0285888671875,
"loss": 0.3555,
"rewards/chosen": 2.7768580118815103,
"rewards/margins": 3.196271006266276,
"rewards/rejected": -0.41941299438476565,
"step": 505
},
{
"epoch": 0.8522105263157894,
"grad_norm": 42.41057229807221,
"kl": 0.0,
"learning_rate": 3.349364905389032e-08,
"logits/chosen": -1031722581.3333334,
"logits/rejected": -827191808.0,
"logps/chosen": -204.5508015950521,
"logps/rejected": -485.5096130371094,
"loss": 0.3319,
"rewards/chosen": 1.5830011367797852,
"rewards/margins": 23.542262077331543,
"rewards/rejected": -21.959260940551758,
"step": 506
},
{
"epoch": 0.8538947368421053,
"grad_norm": 40.22957411853309,
"kl": 0.0,
"learning_rate": 3.2762008523713386e-08,
"logits/chosen": -694314709.3333334,
"logits/rejected": -912536064.0,
"logps/chosen": -231.19136555989584,
"logps/rejected": -210.5927978515625,
"loss": 0.4044,
"rewards/chosen": -1.8982666333516438,
"rewards/margins": -1.2363632520039876,
"rewards/rejected": -0.6619033813476562,
"step": 507
},
{
"epoch": 0.8555789473684211,
"grad_norm": 56.68241962573519,
"kl": 0.0,
"learning_rate": 3.203788684936534e-08,
"logits/chosen": -1072353133.7142857,
"logits/rejected": -857103872.0,
"logps/chosen": -251.75599888392858,
"logps/rejected": -181.1580810546875,
"loss": 0.4473,
"rewards/chosen": -0.4974602290562221,
"rewards/margins": -0.4106224605015346,
"rewards/rejected": -0.0868377685546875,
"step": 508
},
{
"epoch": 0.8572631578947368,
"grad_norm": 47.126301959598706,
"kl": 0.0,
"learning_rate": 3.132130909352709e-08,
"logits/chosen": -802446774.8571428,
"logits/rejected": -754396800.0,
"logps/chosen": -243.97293526785714,
"logps/rejected": -171.71823120117188,
"loss": 0.4578,
"rewards/chosen": 0.9422939164297921,
"rewards/margins": 0.3230052930968148,
"rewards/rejected": 0.6192886233329773,
"step": 509
},
{
"epoch": 0.8589473684210527,
"grad_norm": 67.1017756333287,
"kl": 0.0,
"learning_rate": 3.061230005777593e-08,
"logits/chosen": -716743552.0,
"logits/rejected": -928516315.4285715,
"logps/chosen": -392.4697265625,
"logps/rejected": -267.4026576450893,
"loss": 0.3928,
"rewards/chosen": -4.59370756149292,
"rewards/margins": -1.8247188159397671,
"rewards/rejected": -2.7689887455531528,
"step": 510
},
{
"epoch": 0.8606315789473684,
"grad_norm": 37.48246855118462,
"kl": 0.0,
"learning_rate": 2.991088428172722e-08,
"logits/chosen": -882642329.6,
"logits/rejected": -579270997.3333334,
"logps/chosen": -216.027587890625,
"logps/rejected": -224.06245930989584,
"loss": 0.3588,
"rewards/chosen": 1.8644773483276367,
"rewards/margins": 4.355569076538086,
"rewards/rejected": -2.491091728210449,
"step": 511
},
{
"epoch": 0.8623157894736843,
"grad_norm": 46.701572144028056,
"kl": 0.0,
"learning_rate": 2.9217086042184534e-08,
"logits/chosen": -951701888.0,
"logits/rejected": -1123305728.0,
"logps/chosen": -232.40447998046875,
"logps/rejected": -214.9777374267578,
"loss": 0.3516,
"rewards/chosen": 0.27192726731300354,
"rewards/margins": 5.410427957773209,
"rewards/rejected": -5.138500690460205,
"step": 512
},
{
"epoch": 0.864,
"grad_norm": 47.82460894266106,
"kl": 0.0,
"learning_rate": 2.8530929352300087e-08,
"logits/chosen": -747589632.0,
"logits/rejected": -987404083.2,
"logps/chosen": -221.045166015625,
"logps/rejected": -198.009423828125,
"loss": 0.3888,
"rewards/chosen": 0.2495401899019877,
"rewards/margins": 2.997531179587046,
"rewards/rejected": -2.7479909896850585,
"step": 513
},
{
"epoch": 0.8656842105263158,
"grad_norm": 50.95999181346899,
"kl": 0.0,
"learning_rate": 2.7852437960743326e-08,
"logits/chosen": -816897996.8,
"logits/rejected": -827434154.6666666,
"logps/chosen": -209.32421875,
"logps/rejected": -241.70332845052084,
"loss": 0.3541,
"rewards/chosen": 0.15461671352386475,
"rewards/margins": 2.3556827306747437,
"rewards/rejected": -2.201066017150879,
"step": 514
},
{
"epoch": 0.8673684210526316,
"grad_norm": 70.2346543078714,
"kl": 0.0,
"learning_rate": 2.718163535087864e-08,
"logits/chosen": -805316096.0,
"logits/rejected": -1582949376.0,
"logps/chosen": -179.23245239257812,
"logps/rejected": -174.32058715820312,
"loss": 0.3397,
"rewards/chosen": -0.28184014558792114,
"rewards/margins": 1.7464155554771423,
"rewards/rejected": -2.0282557010650635,
"step": 515
},
{
"epoch": 0.8690526315789474,
"grad_norm": 44.88916114173372,
"kl": 0.0,
"learning_rate": 2.6518544739953187e-08,
"logits/chosen": -884145472.0,
"logits/rejected": -1249547136.0,
"logps/chosen": -265.9125671386719,
"logps/rejected": -233.0666961669922,
"loss": 0.3362,
"rewards/chosen": 0.07709883153438568,
"rewards/margins": 0.21887779235839844,
"rewards/rejected": -0.14177896082401276,
"step": 516
},
{
"epoch": 0.8707368421052631,
"grad_norm": 46.28385480900765,
"kl": 0.0,
"learning_rate": 2.586318907829291e-08,
"logits/chosen": -802108842.6666666,
"logits/rejected": -925224960.0,
"logps/chosen": -203.9495646158854,
"logps/rejected": -222.9334716796875,
"loss": 0.3661,
"rewards/chosen": 1.3522602717081706,
"rewards/margins": 0.6993664304415386,
"rewards/rejected": 0.6528938412666321,
"step": 517
},
{
"epoch": 0.872421052631579,
"grad_norm": 46.93723588820055,
"kl": 0.0,
"learning_rate": 2.5215591048508152e-08,
"logits/chosen": -1172611498.6666667,
"logits/rejected": -601648947.2,
"logps/chosen": -245.01177978515625,
"logps/rejected": -234.228076171875,
"loss": 0.4201,
"rewards/chosen": -0.7683883508046468,
"rewards/margins": 0.7730680624643961,
"rewards/rejected": -1.541456413269043,
"step": 518
},
{
"epoch": 0.8741052631578947,
"grad_norm": 37.74401010007106,
"kl": 0.0,
"learning_rate": 2.4575773064708898e-08,
"logits/chosen": -844627520.0,
"logits/rejected": -680875776.0,
"logps/chosen": -191.596923828125,
"logps/rejected": -210.30389404296875,
"loss": 0.3464,
"rewards/chosen": 1.64203679561615,
"rewards/margins": 2.2509559392929077,
"rewards/rejected": -0.6089191436767578,
"step": 519
},
{
"epoch": 0.8757894736842106,
"grad_norm": 38.54143890619264,
"kl": 0.0,
"learning_rate": 2.3943757271728816e-08,
"logits/chosen": -821064064.0,
"logits/rejected": -720391232.0,
"logps/chosen": -230.83758544921875,
"logps/rejected": -239.45230102539062,
"loss": 0.2565,
"rewards/chosen": 2.128859043121338,
"rewards/margins": 4.223225116729736,
"rewards/rejected": -2.0943660736083984,
"step": 520
},
{
"epoch": 0.8774736842105263,
"grad_norm": 46.42071867786955,
"kl": 0.0,
"learning_rate": 2.3319565544358628e-08,
"logits/chosen": -550017280.0,
"logits/rejected": -889491840.0,
"logps/chosen": -228.99757385253906,
"logps/rejected": -294.79571533203125,
"loss": 0.4034,
"rewards/chosen": 1.803693413734436,
"rewards/margins": 10.621366143226624,
"rewards/rejected": -8.817672729492188,
"step": 521
},
{
"epoch": 0.8791578947368421,
"grad_norm": 165.5741913948296,
"kl": 0.0,
"learning_rate": 2.270321948658943e-08,
"logits/chosen": -1067505152.0,
"logits/rejected": -1063348160.0,
"logps/chosen": -216.38136291503906,
"logps/rejected": -210.46290588378906,
"loss": 0.4473,
"rewards/chosen": 0.3500576317310333,
"rewards/margins": 0.7664283812046051,
"rewards/rejected": -0.4163707494735718,
"step": 522
},
{
"epoch": 0.8808421052631579,
"grad_norm": 53.27002172942906,
"kl": 0.0,
"learning_rate": 2.2094740430864567e-08,
"logits/chosen": -694395392.0,
"logits/rejected": -859106099.2,
"logps/chosen": -148.91910807291666,
"logps/rejected": -242.549169921875,
"loss": 0.4046,
"rewards/chosen": 1.4957799911499023,
"rewards/margins": 2.2706830501556396,
"rewards/rejected": -0.7749030590057373,
"step": 523
},
{
"epoch": 0.8825263157894737,
"grad_norm": 62.65204336260102,
"kl": 0.0,
"learning_rate": 2.1494149437341373e-08,
"logits/chosen": -1015188992.0,
"logits/rejected": -330501728.0,
"logps/chosen": -239.76932779947916,
"logps/rejected": -281.90582275390625,
"loss": 0.3585,
"rewards/chosen": 1.6041777928670247,
"rewards/margins": 3.236354072888692,
"rewards/rejected": -1.6321762800216675,
"step": 524
},
{
"epoch": 0.8842105263157894,
"grad_norm": 59.41452814443212,
"kl": 0.0,
"learning_rate": 2.0901467293162444e-08,
"logits/chosen": -898306944.0,
"logits/rejected": -934284970.6666666,
"logps/chosen": -239.46893310546875,
"logps/rejected": -347.1363525390625,
"loss": 0.4087,
"rewards/chosen": -0.3174598515033722,
"rewards/margins": 1.7363191147645316,
"rewards/rejected": -2.053778966267904,
"step": 525
},
{
"epoch": 0.8858947368421053,
"grad_norm": 28.835874353397827,
"kl": 0.0,
"learning_rate": 2.0316714511735998e-08,
"logits/chosen": -985848448.0,
"logits/rejected": -997492416.0,
"logps/chosen": -201.36636352539062,
"logps/rejected": -174.35333251953125,
"loss": 0.259,
"rewards/chosen": 2.0680689811706543,
"rewards/margins": 4.799753904342651,
"rewards/rejected": -2.731684923171997,
"step": 526
},
{
"epoch": 0.887578947368421,
"grad_norm": 54.565049728962386,
"kl": 0.0,
"learning_rate": 1.9739911332025793e-08,
"logits/chosen": -839355136.0,
"logits/rejected": -553939626.6666666,
"logps/chosen": -199.08831787109375,
"logps/rejected": -176.72452799479166,
"loss": 0.34,
"rewards/chosen": 2.425278663635254,
"rewards/margins": 2.334609886010488,
"rewards/rejected": 0.09066877762476604,
"step": 527
},
{
"epoch": 0.8892631578947369,
"grad_norm": 46.20449911567436,
"kl": 0.0,
"learning_rate": 1.9171077717850955e-08,
"logits/chosen": -786516309.3333334,
"logits/rejected": -727864640.0,
"logps/chosen": -230.94038899739584,
"logps/rejected": -252.71160888671875,
"loss": 0.3229,
"rewards/chosen": 0.6411038637161255,
"rewards/margins": 6.392639756202698,
"rewards/rejected": -5.751535892486572,
"step": 528
},
{
"epoch": 0.8909473684210526,
"grad_norm": 57.38593942004128,
"kl": 0.0,
"learning_rate": 1.8610233357194747e-08,
"logits/chosen": -670264618.6666666,
"logits/rejected": -815440179.2,
"logps/chosen": -183.37162272135416,
"logps/rejected": -351.1501708984375,
"loss": 0.3711,
"rewards/chosen": 1.5523789723714192,
"rewards/margins": 9.300235112508139,
"rewards/rejected": -7.747856140136719,
"step": 529
},
{
"epoch": 0.8926315789473684,
"grad_norm": 68.33693509306372,
"kl": 0.0,
"learning_rate": 1.805739766152309e-08,
"logits/chosen": -777834240.0,
"logits/rejected": -670208640.0,
"logps/chosen": -191.6290283203125,
"logps/rejected": -211.78524780273438,
"loss": 0.3876,
"rewards/chosen": 1.8237395286560059,
"rewards/margins": 1.157799780368805,
"rewards/rejected": 0.6659397482872009,
"step": 530
},
{
"epoch": 0.8943157894736842,
"grad_norm": 34.749494357287055,
"kl": 0.0,
"learning_rate": 1.7512589765112994e-08,
"logits/chosen": -896000804.5714285,
"logits/rejected": -150322144.0,
"logps/chosen": -230.301513671875,
"logps/rejected": -289.5316467285156,
"loss": 0.3033,
"rewards/chosen": 1.0773948260716029,
"rewards/margins": 1.1100638934544154,
"rewards/rejected": -0.0326690673828125,
"step": 531
},
{
"epoch": 0.896,
"grad_norm": 52.937522598027066,
"kl": 0.0,
"learning_rate": 1.697582852439011e-08,
"logits/chosen": -1075272448.0,
"logits/rejected": -1150560512.0,
"logps/chosen": -186.83663940429688,
"logps/rejected": -228.51455688476562,
"loss": 0.3423,
"rewards/chosen": 1.7373024225234985,
"rewards/margins": 1.5837997794151306,
"rewards/rejected": 0.15350264310836792,
"step": 532
},
{
"epoch": 0.8976842105263158,
"grad_norm": 30.475440103504955,
"kl": 0.0,
"learning_rate": 1.6447132517276004e-08,
"logits/chosen": -637437952.0,
"logits/rejected": -962437324.8,
"logps/chosen": -198.0897216796875,
"logps/rejected": -254.03193359375,
"loss": 0.3967,
"rewards/chosen": 1.2216943899790447,
"rewards/margins": 0.7942300478617351,
"rewards/rejected": 0.42746434211730955,
"step": 533
},
{
"epoch": 0.8993684210526316,
"grad_norm": 43.2574141571032,
"kl": 0.0,
"learning_rate": 1.5926520042545383e-08,
"logits/chosen": -1247182848.0,
"logits/rejected": -590226432.0,
"logps/chosen": -222.28861490885416,
"logps/rejected": -237.892724609375,
"loss": 0.3006,
"rewards/chosen": 0.06360169251759847,
"rewards/margins": 1.3943130572636921,
"rewards/rejected": -1.3307113647460938,
"step": 534
},
{
"epoch": 0.9010526315789473,
"grad_norm": 53.71591400639584,
"kl": 0.0,
"learning_rate": 1.5414009119192635e-08,
"logits/chosen": -915837120.0,
"logits/rejected": -808716458.6666666,
"logps/chosen": -179.90530395507812,
"logps/rejected": -195.09493001302084,
"loss": 0.4779,
"rewards/chosen": 0.5700050592422485,
"rewards/margins": -0.024598677953084347,
"rewards/rejected": 0.5946037371953329,
"step": 535
},
{
"epoch": 0.9027368421052632,
"grad_norm": 61.45180482744315,
"kl": 0.0,
"learning_rate": 1.4909617485808073e-08,
"logits/chosen": -1060230348.8,
"logits/rejected": -902803541.3333334,
"logps/chosen": -229.759912109375,
"logps/rejected": -223.30476888020834,
"loss": 0.3913,
"rewards/chosen": 1.0647006988525392,
"rewards/margins": 0.6491708119710287,
"rewards/rejected": 0.41552988688151044,
"step": 536
},
{
"epoch": 0.9044210526315789,
"grad_norm": 101.47573520441375,
"kl": 0.0,
"learning_rate": 1.4413362599964118e-08,
"logits/chosen": -850445994.6666666,
"logits/rejected": -538516864.0,
"logps/chosen": -285.9638264973958,
"logps/rejected": -175.34397888183594,
"loss": 0.4247,
"rewards/chosen": -0.0778544048468272,
"rewards/margins": -3.1665999988714852,
"rewards/rejected": 3.088745594024658,
"step": 537
},
{
"epoch": 0.9061052631578947,
"grad_norm": 45.25790194746145,
"kl": 0.0,
"learning_rate": 1.3925261637611068e-08,
"logits/chosen": -965904588.8,
"logits/rejected": -1320554325.3333333,
"logps/chosen": -204.27237548828126,
"logps/rejected": -217.86629231770834,
"loss": 0.4245,
"rewards/chosen": 1.1585922241210938,
"rewards/margins": 4.713977813720703,
"rewards/rejected": -3.5553855895996094,
"step": 538
},
{
"epoch": 0.9077894736842105,
"grad_norm": 31.80801282899897,
"kl": 0.0,
"learning_rate": 1.3445331492482614e-08,
"logits/chosen": -1130592682.6666667,
"logits/rejected": -715334297.6,
"logps/chosen": -185.36031087239584,
"logps/rejected": -224.770458984375,
"loss": 0.3553,
"rewards/chosen": 0.9161597887674967,
"rewards/margins": 2.255999151865641,
"rewards/rejected": -1.3398393630981444,
"step": 539
},
{
"epoch": 0.9094736842105263,
"grad_norm": 39.75979427316365,
"kl": 0.0,
"learning_rate": 1.2973588775511024e-08,
"logits/chosen": -829291840.0,
"logits/rejected": -1113173418.6666667,
"logps/chosen": -194.16685485839844,
"logps/rejected": -230.64664713541666,
"loss": 0.325,
"rewards/chosen": 2.4532127380371094,
"rewards/margins": 2.8305301666259766,
"rewards/rejected": -0.3773174285888672,
"step": 540
},
{
"epoch": 0.9111578947368421,
"grad_norm": 49.13868786217956,
"kl": 0.0,
"learning_rate": 1.2510049814252299e-08,
"logits/chosen": -1220529152.0,
"logits/rejected": -959601868.8,
"logps/chosen": -205.97456868489584,
"logps/rejected": -292.898876953125,
"loss": 0.4062,
"rewards/chosen": 0.8909465471903483,
"rewards/margins": 2.5595290819803873,
"rewards/rejected": -1.6685825347900392,
"step": 541
},
{
"epoch": 0.9128421052631579,
"grad_norm": 46.27018727053655,
"kl": 0.0,
"learning_rate": 1.2054730652321127e-08,
"logits/chosen": -801094336.0,
"logits/rejected": -564379648.0,
"logps/chosen": -316.7994384765625,
"logps/rejected": -242.9945831298828,
"loss": 0.4345,
"rewards/chosen": -1.3522491455078125,
"rewards/margins": -0.6712818145751953,
"rewards/rejected": -0.6809673309326172,
"step": 542
},
{
"epoch": 0.9145263157894736,
"grad_norm": 40.13535818390141,
"kl": 0.0,
"learning_rate": 1.1607647048835462e-08,
"logits/chosen": -819266389.3333334,
"logits/rejected": -731518822.4,
"logps/chosen": -188.42167154947916,
"logps/rejected": -174.18707275390625,
"loss": 0.3224,
"rewards/chosen": 1.533565839131673,
"rewards/margins": 3.497674496968587,
"rewards/rejected": -1.964108657836914,
"step": 543
},
{
"epoch": 0.9162105263157895,
"grad_norm": 58.704500927311344,
"kl": 0.0,
"learning_rate": 1.116881447787113e-08,
"logits/chosen": -644635840.0,
"logits/rejected": -1043029376.0,
"logps/chosen": -210.66639709472656,
"logps/rejected": -211.0460968017578,
"loss": 0.357,
"rewards/chosen": 0.7050827145576477,
"rewards/margins": 0.4989025592803955,
"rewards/rejected": 0.2061801552772522,
"step": 544
},
{
"epoch": 0.9178947368421052,
"grad_norm": 38.106880910390885,
"kl": 0.0,
"learning_rate": 1.0738248127926342e-08,
"logits/chosen": -678777770.6666666,
"logits/rejected": -1563541504.0,
"logps/chosen": -219.55989583333334,
"logps/rejected": -179.98016357421875,
"loss": 0.3392,
"rewards/chosen": 0.5457992553710938,
"rewards/margins": -1.0953590869903564,
"rewards/rejected": 1.6411583423614502,
"step": 545
},
{
"epoch": 0.919578947368421,
"grad_norm": 58.09818425552494,
"kl": 0.0,
"learning_rate": 1.0315962901395802e-08,
"logits/chosen": -517518624.0,
"logits/rejected": -936108714.6666666,
"logps/chosen": -197.85256958007812,
"logps/rejected": -324.84869384765625,
"loss": 0.3334,
"rewards/chosen": 4.236547470092773,
"rewards/margins": 13.935715993245443,
"rewards/rejected": -9.69916852315267,
"step": 546
},
{
"epoch": 0.9212631578947369,
"grad_norm": 87.5472135904328,
"kl": 0.0,
"learning_rate": 9.901973414055186e-09,
"logits/chosen": -972204544.0,
"logits/rejected": -999611456.0,
"logps/chosen": -240.75531005859375,
"logps/rejected": -262.8578186035156,
"loss": 0.4516,
"rewards/chosen": 0.7045586109161377,
"rewards/margins": 3.072298526763916,
"rewards/rejected": -2.3677399158477783,
"step": 547
},
{
"epoch": 0.9229473684210526,
"grad_norm": 61.8636347167815,
"kl": 0.0,
"learning_rate": 9.496293994555066e-09,
"logits/chosen": -967555788.8,
"logits/rejected": -894452138.6666666,
"logps/chosen": -237.103955078125,
"logps/rejected": -238.02288818359375,
"loss": 0.4624,
"rewards/chosen": -0.08766875267028809,
"rewards/margins": 3.0026972929636635,
"rewards/rejected": -3.0903660456339517,
"step": 548
},
{
"epoch": 0.9246315789473685,
"grad_norm": 76.28246005925376,
"kl": 0.0,
"learning_rate": 9.098938683924972e-09,
"logits/chosen": -610943360.0,
"logits/rejected": -872966400.0,
"logps/chosen": -243.48297119140625,
"logps/rejected": -183.1815185546875,
"loss": 0.4655,
"rewards/chosen": 0.6167190869649252,
"rewards/margins": 0.13095130523045861,
"rewards/rejected": 0.48576778173446655,
"step": 549
},
{
"epoch": 0.9263157894736842,
"grad_norm": 39.99028773675479,
"kl": 0.0,
"learning_rate": 8.709921235087597e-09,
"logits/chosen": -773891264.0,
"logits/rejected": -839812096.0,
"logps/chosen": -239.84304809570312,
"logps/rejected": -216.54505920410156,
"loss": 0.3937,
"rewards/chosen": -0.7027072906494141,
"rewards/margins": 0.8752220869064331,
"rewards/rejected": -1.5779293775558472,
"step": 550
},
{
"epoch": 0.928,
"grad_norm": 176.8083268495628,
"kl": 0.0,
"learning_rate": 8.329255112382665e-09,
"logits/chosen": -674264934.4,
"logits/rejected": -985622101.3333334,
"logps/chosen": -153.71854248046876,
"logps/rejected": -222.71295166015625,
"loss": 0.3363,
"rewards/chosen": 1.8852203369140625,
"rewards/margins": 4.823611068725586,
"rewards/rejected": -2.9383907318115234,
"step": 551
},
{
"epoch": 0.9296842105263158,
"grad_norm": 47.516751657127465,
"kl": 0.0,
"learning_rate": 7.956953491100871e-09,
"logits/chosen": -841082624.0,
"logits/rejected": -614038528.0,
"logps/chosen": -277.8362731933594,
"logps/rejected": -308.65185546875,
"loss": 0.3715,
"rewards/chosen": 0.449728399515152,
"rewards/margins": 3.049138320343835,
"rewards/rejected": -2.599409920828683,
"step": 552
},
{
"epoch": 0.9313684210526316,
"grad_norm": 92.74872502810405,
"kl": 0.0,
"learning_rate": 7.593029257027956e-09,
"logits/chosen": -639304106.6666666,
"logits/rejected": -1038373478.4,
"logps/chosen": -212.13045247395834,
"logps/rejected": -236.4179931640625,
"loss": 0.3602,
"rewards/chosen": -0.6566390991210938,
"rewards/margins": 1.910294532775879,
"rewards/rejected": -2.5669336318969727,
"step": 553
},
{
"epoch": 0.9330526315789474,
"grad_norm": 49.323568144147856,
"kl": 0.0,
"learning_rate": 7.23749500599874e-09,
"logits/chosen": -1499314858.6666667,
"logits/rejected": -812018636.8,
"logps/chosen": -252.00286865234375,
"logps/rejected": -275.40224609375,
"loss": 0.3429,
"rewards/chosen": -0.8385945955912272,
"rewards/margins": 0.727512296040853,
"rewards/rejected": -1.5661068916320802,
"step": 554
},
{
"epoch": 0.9347368421052632,
"grad_norm": 48.76961282183268,
"kl": 0.0,
"learning_rate": 6.89036304346105e-09,
"logits/chosen": -709829034.6666666,
"logits/rejected": -862226176.0,
"logps/chosen": -221.70589192708334,
"logps/rejected": -222.41966247558594,
"loss": 0.3214,
"rewards/chosen": 0.810302734375,
"rewards/margins": 1.5894142389297485,
"rewards/rejected": -0.7791115045547485,
"step": 555
},
{
"epoch": 0.9364210526315789,
"grad_norm": 51.49949200706721,
"kl": 0.0,
"learning_rate": 6.551645384049897e-09,
"logits/chosen": -880212906.6666666,
"logits/rejected": -670162816.0,
"logps/chosen": -202.01947021484375,
"logps/rejected": -240.55711364746094,
"loss": 0.3524,
"rewards/chosen": 0.031015316645304363,
"rewards/margins": 1.1750317017237346,
"rewards/rejected": -1.1440163850784302,
"step": 556
},
{
"epoch": 0.9381052631578948,
"grad_norm": 48.535906716358035,
"kl": 0.0,
"learning_rate": 6.221353751171665e-09,
"logits/chosen": -726684928.0,
"logits/rejected": -747469531.4285715,
"logps/chosen": -186.82913208007812,
"logps/rejected": -262.0355922154018,
"loss": 0.3274,
"rewards/chosen": 4.104483127593994,
"rewards/margins": 5.761045115334647,
"rewards/rejected": -1.656561987740653,
"step": 557
},
{
"epoch": 0.9397894736842105,
"grad_norm": 44.18140203861248,
"kl": 0.0,
"learning_rate": 5.899499576598216e-09,
"logits/chosen": -919181721.6,
"logits/rejected": -773407402.6666666,
"logps/chosen": -205.78271484375,
"logps/rejected": -287.0757242838542,
"loss": 0.36,
"rewards/chosen": 1.863102340698242,
"rewards/margins": 3.7114522298177084,
"rewards/rejected": -1.848349889119466,
"step": 558
},
{
"epoch": 0.9414736842105264,
"grad_norm": 34.08124127057782,
"kl": 0.0,
"learning_rate": 5.586094000071401e-09,
"logits/chosen": -597630336.0,
"logits/rejected": -568648192.0,
"logps/chosen": -231.41092936197916,
"logps/rejected": -189.46743774414062,
"loss": 0.4097,
"rewards/chosen": 1.2908663749694824,
"rewards/margins": 0.8174479305744171,
"rewards/rejected": 0.4734184443950653,
"step": 559
},
{
"epoch": 0.9431578947368421,
"grad_norm": 36.1106808708771,
"kl": 0.0,
"learning_rate": 5.2811478689173686e-09,
"logits/chosen": -405102976.0,
"logits/rejected": -868101461.3333334,
"logps/chosen": -254.44976806640625,
"logps/rejected": -242.90071614583334,
"loss": 0.3742,
"rewards/chosen": 1.0309677124023438,
"rewards/margins": 2.66564400990804,
"rewards/rejected": -1.6346762975056965,
"step": 560
},
{
"epoch": 0.9448421052631579,
"grad_norm": 45.82329638700707,
"kl": 0.0,
"learning_rate": 4.984671737671142e-09,
"logits/chosen": -742133077.3333334,
"logits/rejected": -740478361.6,
"logps/chosen": -258.3249918619792,
"logps/rejected": -183.15859375,
"loss": 0.4038,
"rewards/chosen": -0.7480514844258627,
"rewards/margins": -0.37228356202443447,
"rewards/rejected": -0.3757679224014282,
"step": 561
},
{
"epoch": 0.9465263157894737,
"grad_norm": 39.6743299184901,
"kl": 0.0,
"learning_rate": 4.696675867711386e-09,
"logits/chosen": -535260979.2,
"logits/rejected": -851761066.6666666,
"logps/chosen": -218.016943359375,
"logps/rejected": -338.2209065755208,
"loss": 0.4127,
"rewards/chosen": 1.0443305015563964,
"rewards/margins": 3.631992403666178,
"rewards/rejected": -2.5876619021097818,
"step": 562
},
{
"epoch": 0.9482105263157895,
"grad_norm": 25.945047187533934,
"kl": 0.0,
"learning_rate": 4.417170226905187e-09,
"logits/chosen": -638596096.0,
"logits/rejected": -722204416.0,
"logps/chosen": -176.8044230143229,
"logps/rejected": -205.5111572265625,
"loss": 0.3022,
"rewards/chosen": 1.1705973148345947,
"rewards/margins": 1.9772413730621339,
"rewards/rejected": -0.806644058227539,
"step": 563
},
{
"epoch": 0.9498947368421052,
"grad_norm": 56.04530688467595,
"kl": 0.0,
"learning_rate": 4.146164489263054e-09,
"logits/chosen": -803748693.3333334,
"logits/rejected": -645856576.0,
"logps/chosen": -237.97408040364584,
"logps/rejected": -261.6069030761719,
"loss": 0.4415,
"rewards/chosen": 0.1846969723701477,
"rewards/margins": 2.00448077917099,
"rewards/rejected": -1.8197838068008423,
"step": 564
},
{
"epoch": 0.9515789473684211,
"grad_norm": 45.941589484462845,
"kl": 0.0,
"learning_rate": 3.88366803460416e-09,
"logits/chosen": -1046677888.0,
"logits/rejected": -937844800.0,
"logps/chosen": -216.30455017089844,
"logps/rejected": -152.97848510742188,
"loss": 0.3607,
"rewards/chosen": 1.318871259689331,
"rewards/margins": 1.2597434520721436,
"rewards/rejected": 0.0591278076171875,
"step": 565
},
{
"epoch": 0.9532631578947368,
"grad_norm": 61.47288256142775,
"kl": 0.0,
"learning_rate": 3.6296899482316236e-09,
"logits/chosen": -661587904.0,
"logits/rejected": -835168896.0,
"logps/chosen": -245.52484130859375,
"logps/rejected": -229.88687133789062,
"loss": 0.4462,
"rewards/chosen": 1.543544054031372,
"rewards/margins": 1.7612977027893066,
"rewards/rejected": -0.21775364875793457,
"step": 566
},
{
"epoch": 0.9549473684210527,
"grad_norm": 47.39961532223011,
"kl": 0.0,
"learning_rate": 3.384239020618018e-09,
"logits/chosen": -1413642368.0,
"logits/rejected": -993228288.0,
"logps/chosen": -251.99951171875,
"logps/rejected": -258.76263427734375,
"loss": 0.3247,
"rewards/chosen": 0.9964095950126648,
"rewards/margins": 2.568497121334076,
"rewards/rejected": -1.5720875263214111,
"step": 567
},
{
"epoch": 0.9566315789473684,
"grad_norm": 42.291585331791644,
"kl": 0.0,
"learning_rate": 3.1473237471012214e-09,
"logits/chosen": -733431910.4,
"logits/rejected": -799960064.0,
"logps/chosen": -177.69178466796876,
"logps/rejected": -213.2462158203125,
"loss": 0.3644,
"rewards/chosen": 0.27896909713745116,
"rewards/margins": 2.7935458501180013,
"rewards/rejected": -2.5145767529805503,
"step": 568
},
{
"epoch": 0.9583157894736842,
"grad_norm": 40.929057929848376,
"kl": 0.0,
"learning_rate": 2.9189523275903736e-09,
"logits/chosen": -978359091.2,
"logits/rejected": -652150016.0,
"logps/chosen": -271.872216796875,
"logps/rejected": -174.5239461263021,
"loss": 0.3284,
"rewards/chosen": -0.7042415618896485,
"rewards/margins": 0.11138847668965657,
"rewards/rejected": -0.815630038579305,
"step": 569
},
{
"epoch": 0.96,
"grad_norm": 40.061462493884235,
"kl": 0.0,
"learning_rate": 2.6991326662819667e-09,
"logits/chosen": -729751552.0,
"logits/rejected": -752365772.8,
"logps/chosen": -195.5203857421875,
"logps/rejected": -209.675634765625,
"loss": 0.3478,
"rewards/chosen": 1.747754414876302,
"rewards/margins": 3.8098199208577475,
"rewards/rejected": -2.0620655059814452,
"step": 570
},
{
"epoch": 0.9616842105263158,
"grad_norm": 37.651701881002374,
"kl": 0.0,
"learning_rate": 2.4878723713864234e-09,
"logits/chosen": -627430656.0,
"logits/rejected": -945733734.4,
"logps/chosen": -186.00541178385416,
"logps/rejected": -253.2623046875,
"loss": 0.3783,
"rewards/chosen": 1.6760249137878418,
"rewards/margins": 4.277830410003662,
"rewards/rejected": -2.60180549621582,
"step": 571
},
{
"epoch": 0.9633684210526315,
"grad_norm": 70.57315668742685,
"kl": 0.0,
"learning_rate": 2.285178754864614e-09,
"logits/chosen": -987434325.3333334,
"logits/rejected": -616783462.4,
"logps/chosen": -218.68375651041666,
"logps/rejected": -293.998388671875,
"loss": 0.4376,
"rewards/chosen": 1.6107417742411296,
"rewards/margins": 7.283069960276286,
"rewards/rejected": -5.672328186035156,
"step": 572
},
{
"epoch": 0.9650526315789474,
"grad_norm": 69.90429499041211,
"kl": 0.0,
"learning_rate": 2.091058832174891e-09,
"logits/chosen": -730186432.0,
"logits/rejected": -509659392.0,
"logps/chosen": -196.79669189453125,
"logps/rejected": -303.237060546875,
"loss": 0.317,
"rewards/chosen": 0.6219147443771362,
"rewards/margins": 11.697727084159851,
"rewards/rejected": -11.075812339782715,
"step": 573
},
{
"epoch": 0.9667368421052631,
"grad_norm": 38.30542476552404,
"kl": 0.0,
"learning_rate": 1.905519322030258e-09,
"logits/chosen": -938985881.6,
"logits/rejected": -598871637.3333334,
"logps/chosen": -183.621826171875,
"logps/rejected": -228.38956705729166,
"loss": 0.428,
"rewards/chosen": 0.9899630546569824,
"rewards/margins": 1.8534589608510337,
"rewards/rejected": -0.8634959061940511,
"step": 574
},
{
"epoch": 0.968421052631579,
"grad_norm": 80.5219219878979,
"kl": 0.0,
"learning_rate": 1.7285666461657467e-09,
"logits/chosen": -844385426.2857143,
"logits/rejected": -1569694336.0,
"logps/chosen": -221.71473911830358,
"logps/rejected": -172.7010498046875,
"loss": 0.4116,
"rewards/chosen": -0.09411777768816267,
"rewards/margins": -0.8716171554156712,
"rewards/rejected": 0.7774993777275085,
"step": 575
},
{
"epoch": 0.9701052631578947,
"grad_norm": 89.03407725109749,
"kl": 0.0,
"learning_rate": 1.5602069291162368e-09,
"logits/chosen": -925635904.0,
"logits/rejected": -844432000.0,
"logps/chosen": -256.534912109375,
"logps/rejected": -220.78697204589844,
"loss": 0.2972,
"rewards/chosen": -0.9312546849250793,
"rewards/margins": 2.055956780910492,
"rewards/rejected": -2.9872114658355713,
"step": 576
},
{
"epoch": 0.9717894736842105,
"grad_norm": 33.63530023405151,
"kl": 0.0,
"learning_rate": 1.4004459980045124e-09,
"logits/chosen": -958183424.0,
"logits/rejected": -700009045.3333334,
"logps/chosen": -239.659912109375,
"logps/rejected": -270.008544921875,
"loss": 0.4455,
"rewards/chosen": -0.18169254064559937,
"rewards/margins": 2.001987040042877,
"rewards/rejected": -2.1836795806884766,
"step": 577
},
{
"epoch": 0.9734736842105263,
"grad_norm": 34.10535983128858,
"kl": 0.0,
"learning_rate": 1.2492893823394246e-09,
"logits/chosen": -1030501120.0,
"logps/chosen": -254.66343688964844,
"loss": 0.3929,
"rewards/chosen": -0.19029521942138672,
"step": 578
},
{
"epoch": 0.9751578947368421,
"grad_norm": 130.35662841898622,
"kl": 0.0,
"learning_rate": 1.10674231382471e-09,
"logits/chosen": -418937888.0,
"logits/rejected": -518623584.0,
"logps/chosen": -232.24008178710938,
"logps/rejected": -243.5550994873047,
"loss": 0.3157,
"rewards/chosen": 0.7040752172470093,
"rewards/margins": 1.6154407858848572,
"rewards/rejected": -0.9113655686378479,
"step": 579
},
{
"epoch": 0.9768421052631578,
"grad_norm": 42.15293348378943,
"kl": 0.0,
"learning_rate": 9.7280972617772e-10,
"logits/chosen": -1371472384.0,
"logits/rejected": -1195231402.6666667,
"logps/chosen": -254.03028869628906,
"logps/rejected": -272.2608642578125,
"loss": 0.3864,
"rewards/chosen": -2.0985825061798096,
"rewards/margins": -0.8987653255462646,
"rewards/rejected": -1.199817180633545,
"step": 580
},
{
"epoch": 0.9785263157894737,
"grad_norm": 82.82639202731109,
"kl": 0.0,
"learning_rate": 8.474962549588349e-10,
"logits/chosen": -553952512.0,
"logits/rejected": -1753602389.3333333,
"logps/chosen": -212.57861328125,
"logps/rejected": -235.5814208984375,
"loss": 0.3613,
"rewards/chosen": 1.6823585510253907,
"rewards/margins": 2.548850742975871,
"rewards/rejected": -0.8664921919504801,
"step": 581
},
{
"epoch": 0.9802105263157894,
"grad_norm": 39.562779110049625,
"kl": 0.0,
"learning_rate": 7.308062374108692e-10,
"logits/chosen": -836093888.0,
"logits/rejected": -541282218.6666666,
"logps/chosen": -245.46795654296875,
"logps/rejected": -265.79986572265625,
"loss": 0.3856,
"rewards/chosen": 1.5786278247833252,
"rewards/margins": 3.83101216952006,
"rewards/rejected": -2.252384344736735,
"step": 582
},
{
"epoch": 0.9818947368421053,
"grad_norm": 49.619503065339295,
"kl": 0.0,
"learning_rate": 6.227437123090539e-10,
"logits/chosen": -1412194432.0,
"logits/rejected": -683454902.8571428,
"logps/chosen": -247.62109375,
"logps/rejected": -278.0474853515625,
"loss": 0.2775,
"rewards/chosen": -2.1934449672698975,
"rewards/margins": 0.6917969499315535,
"rewards/rejected": -2.885241917201451,
"step": 583
},
{
"epoch": 0.983578947368421,
"grad_norm": 41.65613682455663,
"kl": 0.0,
"learning_rate": 5.233124198212035e-10,
"logits/chosen": -1068381312.0,
"logits/rejected": -620764608.0,
"logps/chosen": -203.81106567382812,
"logps/rejected": -199.25198364257812,
"loss": 0.3426,
"rewards/chosen": 0.44310152530670166,
"rewards/margins": 1.8038746118545532,
"rewards/rejected": -1.3607730865478516,
"step": 584
},
{
"epoch": 0.9852631578947368,
"grad_norm": 74.23272136164195,
"kl": 0.0,
"learning_rate": 4.3251580137831924e-10,
"logits/chosen": -815114496.0,
"logits/rejected": -602160384.0,
"logps/chosen": -206.22401428222656,
"logps/rejected": -258.4097900390625,
"loss": 0.2408,
"rewards/chosen": 1.1723206043243408,
"rewards/margins": 2.8920695781707764,
"rewards/rejected": -1.7197489738464355,
"step": 585
},
{
"epoch": 0.9869473684210527,
"grad_norm": 37.528975070907734,
"kl": 0.0,
"learning_rate": 3.5035699955540675e-10,
"logits/chosen": -1021857088.0,
"logits/rejected": -1152371072.0,
"logps/chosen": -198.44970703125,
"logps/rejected": -276.451904296875,
"loss": 0.2932,
"rewards/chosen": 1.2920650243759155,
"rewards/margins": 4.100528836250305,
"rewards/rejected": -2.8084638118743896,
"step": 586
},
{
"epoch": 0.9886315789473684,
"grad_norm": 66.83382682650414,
"kl": 0.0,
"learning_rate": 2.768388579627301e-10,
"logits/chosen": -696945877.3333334,
"logits/rejected": -635326361.6,
"logps/chosen": -188.5633748372396,
"logps/rejected": -227.542333984375,
"loss": 0.3505,
"rewards/chosen": -1.1209592819213867,
"rewards/margins": 2.0206165313720703,
"rewards/rejected": -3.141575813293457,
"step": 587
},
{
"epoch": 0.9903157894736843,
"grad_norm": 38.439783990376164,
"kl": 0.0,
"learning_rate": 2.1196392114744556e-10,
"logits/chosen": -732253184.0,
"logits/rejected": -485836117.3333333,
"logps/chosen": -229.332763671875,
"logps/rejected": -178.27132161458334,
"loss": 0.3191,
"rewards/chosen": 2.1135427474975588,
"rewards/margins": 2.325268598397573,
"rewards/rejected": -0.21172585090001425,
"step": 588
},
{
"epoch": 0.992,
"grad_norm": 84.00543667151456,
"kl": 0.0,
"learning_rate": 1.5573443450545009e-10,
"logits/chosen": -681184563.2,
"logits/rejected": -883190613.3333334,
"logps/chosen": -231.067626953125,
"logps/rejected": -222.4321492513021,
"loss": 0.3878,
"rewards/chosen": 0.3701077699661255,
"rewards/margins": 1.0806505282719931,
"rewards/rejected": -0.7105427583058676,
"step": 589
},
{
"epoch": 0.9936842105263158,
"grad_norm": 65.06311520099162,
"kl": 0.0,
"learning_rate": 1.0815234420369357e-10,
"logits/chosen": -837959372.8,
"logits/rejected": -1049096704.0,
"logps/chosen": -237.297705078125,
"logps/rejected": -172.59944661458334,
"loss": 0.3653,
"rewards/chosen": 1.0379376411437988,
"rewards/margins": -0.5275972684224446,
"rewards/rejected": 1.5655349095662434,
"step": 590
},
{
"epoch": 0.9953684210526316,
"grad_norm": 131.03458960479955,
"kl": 0.0,
"learning_rate": 6.921929711287134e-11,
"logits/chosen": -793252928.0,
"logits/rejected": -1009639296.0,
"logps/chosen": -249.7359619140625,
"logps/rejected": -276.18560791015625,
"loss": 0.4221,
"rewards/chosen": 1.2013773918151855,
"rewards/margins": 2.3829771280288696,
"rewards/rejected": -1.181599736213684,
"step": 591
},
{
"epoch": 0.9970526315789474,
"grad_norm": 59.274731916237144,
"kl": 0.0,
"learning_rate": 3.893664075035885e-11,
"logits/chosen": -958084010.6666666,
"logits/rejected": -776463616.0,
"logps/chosen": -219.0360107421875,
"logps/rejected": -177.7460174560547,
"loss": 0.4209,
"rewards/chosen": 0.09142877658208211,
"rewards/margins": -0.03964764873186748,
"rewards/rejected": 0.13107642531394958,
"step": 592
},
{
"epoch": 0.9987368421052631,
"grad_norm": 35.198610073696365,
"kl": 0.0,
"learning_rate": 1.730542323355455e-11,
"logits/chosen": -1364308699.4285715,
"logits/rejected": -682940672.0,
"logps/chosen": -254.698486328125,
"logps/rejected": -193.01634216308594,
"loss": 0.3481,
"rewards/chosen": -0.044670513698032925,
"rewards/margins": -1.6197544847215926,
"rewards/rejected": 1.5750839710235596,
"step": 593
},
{
"epoch": 1.0,
"grad_norm": 35.198610073696365,
"kl": 0.0,
"learning_rate": 4.326393243742066e-12,
"logits/chosen": -436645888.0,
"logits/rejected": -1217262336.0,
"logps/chosen": -184.77606201171875,
"logps/rejected": -178.923583984375,
"loss": 0.342,
"rewards/chosen": -0.0059494078159332275,
"rewards/margins": 1.3890884816646576,
"rewards/rejected": -1.3950378894805908,
"step": 594
},
{
"epoch": 1.0,
"step": 594,
"total_flos": 2.3597531637009613e+17,
"train_loss": 0.4100885201333348,
"train_runtime": 53730.3673,
"train_samples_per_second": 0.354,
"train_steps_per_second": 0.011
}
],
"logging_steps": 1,
"max_steps": 594,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 1000,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": true
},
"attributes": {}
}
},
"total_flos": 2.3597531637009613e+17,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}
Reference in New Issue Copy Permalink