{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.11295180722891567, "eval_steps": 500, "global_step": 150, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1461.0, "completions/mean_length": 539.939453125, "completions/mean_terminated_length": 534.0687866210938, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.0007530120481927711, "frac_reward_zero_std": 0.0, "grad_norm": 0.2677452564239502, "learning_rate": 1e-06, "loss": -0.0491, "num_tokens": 322497.0, "reward": 1.8222503662109375, "reward_std": 0.2692643404006958, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.8222503662109375, "rewards/icrm_reward/std": 0.35891062021255493, "step": 1 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1361.0, "completions/mean_length": 541.125, "completions/mean_terminated_length": 535.2612915039062, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.0015060240963855422, "frac_reward_zero_std": 0.0, "grad_norm": 0.2547134757041931, "learning_rate": 1e-06, "loss": -0.0405, "num_tokens": 643217.0, "reward": 1.8759613037109375, "reward_std": 0.26730749011039734, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.8759613037109375, "rewards/icrm_reward/std": 0.35533255338668823, "step": 2 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1268.0, "completions/max_terminated_length": 1268.0, "completions/mean_length": 503.001953125, "completions/mean_terminated_length": 503.001953125, "completions/min_length": 107.0, "completions/min_terminated_length": 107.0, "epoch": 0.002259036144578313, "frac_reward_zero_std": 0.0, "grad_norm": 0.28800466656684875, "learning_rate": 1e-06, "loss": -0.0702, "num_tokens": 946338.0, "reward": 1.807281494140625, "reward_std": 0.2657621502876282, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.807281494140625, "rewards/icrm_reward/std": 0.36616575717926025, "step": 3 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 482.392578125, "completions/mean_terminated_length": 482.392578125, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.0030120481927710845, "frac_reward_zero_std": 0.0, "grad_norm": 0.2880045473575592, "learning_rate": 1e-06, "loss": -0.0543, "num_tokens": 1238651.0, "reward": 1.8897247314453125, "reward_std": 0.24010181427001953, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.8897247314453125, "rewards/icrm_reward/std": 0.2975654602050781, "step": 4 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1447.0, "completions/mean_length": 548.681640625, "completions/mean_terminated_length": 546.74951171875, "completions/min_length": 95.0, "completions/min_terminated_length": 95.0, "epoch": 0.0037650602409638554, "frac_reward_zero_std": 0.0, "grad_norm": 0.26260432600975037, "learning_rate": 1e-06, "loss": -0.0915, "num_tokens": 1564648.0, "reward": 1.8785552978515625, "reward_std": 0.2939695119857788, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.8785552978515625, "rewards/icrm_reward/std": 0.3403075635433197, "step": 5 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1248.0, "completions/mean_length": 465.359375, "completions/mean_terminated_length": 461.1607971191406, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 0.004518072289156626, "frac_reward_zero_std": 0.0, "grad_norm": 0.2783867418766022, "learning_rate": 1e-06, "loss": -0.0511, "num_tokens": 1845008.0, "reward": 1.886444091796875, "reward_std": 0.24519476294517517, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.886444091796875, "rewards/icrm_reward/std": 0.32863083481788635, "step": 6 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1419.0, "completions/mean_length": 558.427734375, "completions/mean_terminated_length": 556.5146484375, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.005271084337349397, "frac_reward_zero_std": 0.0, "grad_norm": 0.2398781031370163, "learning_rate": 1e-06, "loss": -0.0528, "num_tokens": 2178827.0, "reward": 1.89752197265625, "reward_std": 0.2825944125652313, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.89752197265625, "rewards/icrm_reward/std": 0.3314611613750458, "step": 7 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1315.0, "completions/max_terminated_length": 1315.0, "completions/mean_length": 516.62890625, "completions/mean_terminated_length": 516.62890625, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.006024096385542169, "frac_reward_zero_std": 0.0, "grad_norm": 0.26950332522392273, "learning_rate": 1e-06, "loss": -0.0515, "num_tokens": 2490557.0, "reward": 1.880584716796875, "reward_std": 0.24953413009643555, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.880584716796875, "rewards/icrm_reward/std": 0.31805965304374695, "step": 8 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1486.0, "completions/mean_length": 540.916015625, "completions/mean_terminated_length": 538.9686889648438, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.00677710843373494, "frac_reward_zero_std": 0.0, "grad_norm": 0.27774062752723694, "learning_rate": 1e-06, "loss": -0.0696, "num_tokens": 2810898.0, "reward": 1.8914337158203125, "reward_std": 0.2737649083137512, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.8914337158203125, "rewards/icrm_reward/std": 0.3299334645271301, "step": 9 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1449.0, "completions/mean_length": 489.474609375, "completions/mean_terminated_length": 487.4266052246094, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.007530120481927711, "frac_reward_zero_std": 0.0, "grad_norm": 0.29050636291503906, "learning_rate": 1e-06, "loss": -0.0795, "num_tokens": 3108869.0, "reward": 1.938812255859375, "reward_std": 0.2572469711303711, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.938812255859375, "rewards/icrm_reward/std": 0.338980495929718, "step": 10 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1475.0, "completions/max_terminated_length": 1475.0, "completions/mean_length": 520.849609375, "completions/mean_terminated_length": 520.849609375, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "epoch": 0.008283132530120483, "frac_reward_zero_std": 0.0, "grad_norm": 0.2942567765712738, "learning_rate": 1e-06, "loss": -0.0717, "num_tokens": 3421848.0, "reward": 1.9404754638671875, "reward_std": 0.2595525085926056, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9404754638671875, "rewards/icrm_reward/std": 0.3023451566696167, "step": 11 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1396.0, "completions/mean_length": 519.873046875, "completions/mean_terminated_length": 517.884521484375, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.009036144578313253, "frac_reward_zero_std": 0.0, "grad_norm": 0.24970945715904236, "learning_rate": 1e-06, "loss": -0.0503, "num_tokens": 3729655.0, "reward": 1.9094085693359375, "reward_std": 0.25953686237335205, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9094085693359375, "rewards/icrm_reward/std": 0.331910640001297, "step": 12 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1369.0, "completions/mean_length": 547.90625, "completions/mean_terminated_length": 545.9725952148438, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.009789156626506024, "frac_reward_zero_std": 0.0, "grad_norm": 0.28067365288734436, "learning_rate": 1e-06, "loss": -0.0752, "num_tokens": 4053863.0, "reward": 1.9482421875, "reward_std": 0.26520416140556335, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9482421875, "rewards/icrm_reward/std": 0.3480120599269867, "step": 13 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1511.0, "completions/mean_length": 548.40234375, "completions/mean_terminated_length": 546.4696655273438, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.010542168674698794, "frac_reward_zero_std": 0.0, "grad_norm": 0.25670015811920166, "learning_rate": 1e-06, "loss": -0.0691, "num_tokens": 4378933.0, "reward": 1.9578399658203125, "reward_std": 0.27680855989456177, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9578399658203125, "rewards/icrm_reward/std": 0.3370186686515808, "step": 14 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1521.0, "completions/max_terminated_length": 1521.0, "completions/mean_length": 524.80859375, "completions/mean_terminated_length": 524.80859375, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.011295180722891566, "frac_reward_zero_std": 0.0, "grad_norm": 0.2583891451358795, "learning_rate": 1e-06, "loss": -0.0585, "num_tokens": 4691523.0, "reward": 1.9633941650390625, "reward_std": 0.2548086643218994, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9633941650390625, "rewards/icrm_reward/std": 0.31804537773132324, "step": 15 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1247.0, "completions/mean_length": 560.337890625, "completions/mean_terminated_length": 556.5117797851562, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "epoch": 0.012048192771084338, "frac_reward_zero_std": 0.0, "grad_norm": 0.24440975487232208, "learning_rate": 1e-06, "loss": -0.0549, "num_tokens": 5027184.0, "reward": 1.9822998046875, "reward_std": 0.268485963344574, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 1.9822998046875, "rewards/icrm_reward/std": 0.3278394639492035, "step": 16 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1356.0, "completions/mean_length": 591.25390625, "completions/mean_terminated_length": 589.4050903320312, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.012801204819277108, "frac_reward_zero_std": 0.0, "grad_norm": 0.22858311235904694, "learning_rate": 1e-06, "loss": -0.049, "num_tokens": 5376418.0, "reward": 2.090728759765625, "reward_std": 0.2703300714492798, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.090728759765625, "rewards/icrm_reward/std": 0.3051627278327942, "step": 17 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1427.0, "completions/mean_length": 560.716796875, "completions/mean_terminated_length": 556.8922119140625, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "epoch": 0.01355421686746988, "frac_reward_zero_std": 0.0, "grad_norm": 0.24253730475902557, "learning_rate": 1e-06, "loss": -0.0476, "num_tokens": 5708881.0, "reward": 2.023834228515625, "reward_std": 0.22830849885940552, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.023834228515625, "rewards/icrm_reward/std": 0.3195403516292572, "step": 18 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1396.0, "completions/max_terminated_length": 1396.0, "completions/mean_length": 554.669921875, "completions/mean_terminated_length": 554.669921875, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.01430722891566265, "frac_reward_zero_std": 0.0, "grad_norm": 0.22568698227405548, "learning_rate": 1e-06, "loss": -0.0628, "num_tokens": 6034520.0, "reward": 2.0350494384765625, "reward_std": 0.23877058923244476, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.0350494384765625, "rewards/icrm_reward/std": 0.2821340560913086, "step": 19 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1408.0, "completions/mean_length": 618.142578125, "completions/mean_terminated_length": 612.7328491210938, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.015060240963855422, "frac_reward_zero_std": 0.0, "grad_norm": 0.21298259496688843, "learning_rate": 1e-06, "loss": -0.0312, "num_tokens": 6399457.0, "reward": 2.043853759765625, "reward_std": 0.2564227879047394, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.043853759765625, "rewards/icrm_reward/std": 0.32547545433044434, "step": 20 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1500.0, "completions/mean_length": 647.5078125, "completions/mean_terminated_length": 636.9723510742188, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.01581325301204819, "frac_reward_zero_std": 0.0, "grad_norm": 0.21974076330661774, "learning_rate": 1e-06, "loss": -0.0194, "num_tokens": 6773909.0, "reward": 2.0822906494140625, "reward_std": 0.2559783458709717, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.0822906494140625, "rewards/icrm_reward/std": 0.3165861964225769, "step": 21 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 629.095703125, "completions/mean_terminated_length": 627.3209228515625, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.016566265060240965, "frac_reward_zero_std": 0.0, "grad_norm": 0.2140408158302307, "learning_rate": 1e-06, "loss": -0.0158, "num_tokens": 7137062.0, "reward": 2.0489654541015625, "reward_std": 0.22745975852012634, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.0489654541015625, "rewards/icrm_reward/std": 0.28583163022994995, "step": 22 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1474.0, "completions/mean_length": 668.0, "completions/mean_terminated_length": 666.3013916015625, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 0.017319277108433735, "frac_reward_zero_std": 0.0, "grad_norm": 0.19740059971809387, "learning_rate": 1e-06, "loss": -0.0199, "num_tokens": 7525974.0, "reward": 2.114471435546875, "reward_std": 0.22889575362205505, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.114471435546875, "rewards/icrm_reward/std": 0.2792186439037323, "step": 23 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.033203125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1507.0, "completions/mean_length": 688.009765625, "completions/mean_terminated_length": 658.8869018554688, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.018072289156626505, "frac_reward_zero_std": 0.0, "grad_norm": 0.19638773798942566, "learning_rate": 1e-06, "loss": 0.0102, "num_tokens": 7924795.0, "reward": 2.080596923828125, "reward_std": 0.23665004968643188, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.080596923828125, "rewards/icrm_reward/std": 0.3111322522163391, "step": 24 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1517.0, "completions/max_terminated_length": 1517.0, "completions/mean_length": 625.5546875, "completions/mean_terminated_length": 625.5546875, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.01882530120481928, "frac_reward_zero_std": 0.0, "grad_norm": 0.22388488054275513, "learning_rate": 1e-06, "loss": -0.0295, "num_tokens": 8289031.0, "reward": 2.1454315185546875, "reward_std": 0.22431066632270813, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1454315185546875, "rewards/icrm_reward/std": 0.2752014100551605, "step": 25 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.025390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1518.0, "completions/mean_length": 620.7578125, "completions/mean_terminated_length": 596.913818359375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.01957831325301205, "frac_reward_zero_std": 0.0, "grad_norm": 0.21399712562561035, "learning_rate": 1e-06, "loss": -0.0254, "num_tokens": 8656715.0, "reward": 2.0743865966796875, "reward_std": 0.2408333122730255, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.0743865966796875, "rewards/icrm_reward/std": 0.3148554563522339, "step": 26 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1519.0, "completions/max_terminated_length": 1519.0, "completions/mean_length": 640.244140625, "completions/mean_terminated_length": 640.244140625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.02033132530120482, "frac_reward_zero_std": 0.0, "grad_norm": 0.2085241973400116, "learning_rate": 1e-06, "loss": -0.0445, "num_tokens": 9027544.0, "reward": 2.1477813720703125, "reward_std": 0.22815221548080444, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1477813720703125, "rewards/icrm_reward/std": 0.3022710084915161, "step": 27 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1444.0, "completions/max_terminated_length": 1444.0, "completions/mean_length": 574.10546875, "completions/mean_terminated_length": 574.10546875, "completions/min_length": 118.0, "completions/min_terminated_length": 118.0, "epoch": 0.02108433734939759, "frac_reward_zero_std": 0.0, "grad_norm": 0.22086061537265778, "learning_rate": 1e-06, "loss": -0.0415, "num_tokens": 9368830.0, "reward": 2.0839691162109375, "reward_std": 0.22651252150535583, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.0839691162109375, "rewards/icrm_reward/std": 0.2819969654083252, "step": 28 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1455.0, "completions/mean_length": 608.298828125, "completions/mean_terminated_length": 602.8310546875, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.021837349397590362, "frac_reward_zero_std": 0.0, "grad_norm": 0.21593217551708221, "learning_rate": 1e-06, "loss": -0.0165, "num_tokens": 9728471.0, "reward": 2.12347412109375, "reward_std": 0.2261788547039032, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.12347412109375, "rewards/icrm_reward/std": 0.2729581892490387, "step": 29 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1490.0, "completions/mean_length": 602.5390625, "completions/mean_terminated_length": 600.7123413085938, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.022590361445783132, "frac_reward_zero_std": 0.0, "grad_norm": 0.21327072381973267, "learning_rate": 1e-06, "loss": -0.0265, "num_tokens": 10078123.0, "reward": 2.134918212890625, "reward_std": 0.22259432077407837, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.134918212890625, "rewards/icrm_reward/std": 0.2794201970100403, "step": 30 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1390.0, "completions/mean_length": 624.75390625, "completions/mean_terminated_length": 622.9706420898438, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.023343373493975902, "frac_reward_zero_std": 0.0, "grad_norm": 0.21840304136276245, "learning_rate": 1e-06, "loss": -0.0354, "num_tokens": 10439629.0, "reward": 2.1316375732421875, "reward_std": 0.2339249849319458, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1316375732421875, "rewards/icrm_reward/std": 0.2949000895023346, "step": 31 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1455.0, "completions/mean_length": 618.02734375, "completions/mean_terminated_length": 610.7991943359375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.024096385542168676, "frac_reward_zero_std": 0.0, "grad_norm": 0.21134598553180695, "learning_rate": 1e-06, "loss": -0.0239, "num_tokens": 10801515.0, "reward": 2.1789169311523438, "reward_std": 0.2280135154724121, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1789169311523438, "rewards/icrm_reward/std": 0.3063806891441345, "step": 32 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1394.0, "completions/max_terminated_length": 1394.0, "completions/mean_length": 579.873046875, "completions/mean_terminated_length": 579.873046875, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.024849397590361446, "frac_reward_zero_std": 0.0, "grad_norm": 0.23027564585208893, "learning_rate": 1e-06, "loss": -0.0456, "num_tokens": 11142458.0, "reward": 2.1141510009765625, "reward_std": 0.22557678818702698, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1141510009765625, "rewards/icrm_reward/std": 0.2943926155567169, "step": 33 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1466.0, "completions/mean_length": 569.83203125, "completions/mean_terminated_length": 567.9412841796875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.025602409638554216, "frac_reward_zero_std": 0.0, "grad_norm": 0.2484128624200821, "learning_rate": 1e-06, "loss": -0.0377, "num_tokens": 11478676.0, "reward": 2.138275146484375, "reward_std": 0.2324836403131485, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.138275146484375, "rewards/icrm_reward/std": 0.2837066352367401, "step": 34 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1524.0, "completions/mean_length": 649.62109375, "completions/mean_terminated_length": 644.3968505859375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.02635542168674699, "frac_reward_zero_std": 0.0, "grad_norm": 0.22154265642166138, "learning_rate": 1e-06, "loss": -0.0408, "num_tokens": 11857058.0, "reward": 2.131011962890625, "reward_std": 0.2266952395439148, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.131011962890625, "rewards/icrm_reward/std": 0.31931453943252563, "step": 35 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1451.0, "completions/mean_length": 637.251953125, "completions/mean_terminated_length": 628.3885498046875, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.02710843373493976, "frac_reward_zero_std": 0.0, "grad_norm": 0.20767764747142792, "learning_rate": 1e-06, "loss": 0.0058, "num_tokens": 12229043.0, "reward": 2.1409454345703125, "reward_std": 0.22632841765880585, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1409454345703125, "rewards/icrm_reward/std": 0.2695969045162201, "step": 36 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.013671875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1456.0, "completions/mean_length": 659.4453125, "completions/mean_terminated_length": 647.2950439453125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.02786144578313253, "frac_reward_zero_std": 0.0, "grad_norm": 0.2131509631872177, "learning_rate": 1e-06, "loss": -0.0088, "num_tokens": 12614919.0, "reward": 2.208251953125, "reward_std": 0.24248948693275452, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.208251953125, "rewards/icrm_reward/std": 0.3120294213294983, "step": 37 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1361.0, "completions/mean_length": 630.41015625, "completions/mean_terminated_length": 626.85888671875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.0286144578313253, "frac_reward_zero_std": 0.0, "grad_norm": 0.2243937849998474, "learning_rate": 1e-06, "loss": -0.0279, "num_tokens": 12980249.0, "reward": 2.111053466796875, "reward_std": 0.22191520035266876, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.111053466796875, "rewards/icrm_reward/std": 0.2775036096572876, "step": 38 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1459.0, "completions/mean_length": 675.94140625, "completions/mean_terminated_length": 665.7431030273438, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.029367469879518073, "frac_reward_zero_std": 0.0, "grad_norm": 0.20395302772521973, "learning_rate": 1e-06, "loss": -0.0178, "num_tokens": 13371051.0, "reward": 2.2367706298828125, "reward_std": 0.24021776020526886, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2367706298828125, "rewards/icrm_reward/std": 0.30560144782066345, "step": 39 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1495.0, "completions/mean_length": 665.2265625, "completions/mean_terminated_length": 658.3700561523438, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.030120481927710843, "frac_reward_zero_std": 0.0, "grad_norm": 0.21775901317596436, "learning_rate": 1e-06, "loss": -0.0184, "num_tokens": 13762655.0, "reward": 2.22381591796875, "reward_std": 0.23306353390216827, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.22381591796875, "rewards/icrm_reward/std": 0.2940692901611328, "step": 40 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.017578125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1489.0, "completions/mean_length": 711.74609375, "completions/mean_terminated_length": 696.9979858398438, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.030873493975903613, "frac_reward_zero_std": 0.0, "grad_norm": 0.20008227229118347, "learning_rate": 1e-06, "loss": -0.0203, "num_tokens": 14174493.0, "reward": 2.1737060546875, "reward_std": 0.23047159612178802, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1737060546875, "rewards/icrm_reward/std": 0.2827721834182739, "step": 41 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1519.0, "completions/mean_length": 628.2109375, "completions/mean_terminated_length": 622.8605346679688, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.03162650602409638, "frac_reward_zero_std": 0.0, "grad_norm": 0.23574763536453247, "learning_rate": 1e-06, "loss": -0.024, "num_tokens": 14538633.0, "reward": 2.2064056396484375, "reward_std": 0.23086489737033844, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2064056396484375, "rewards/icrm_reward/std": 0.30988746881484985, "step": 42 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1509.0, "completions/mean_length": 651.89453125, "completions/mean_terminated_length": 648.427490234375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.03237951807228916, "frac_reward_zero_std": 0.0, "grad_norm": 0.22128161787986755, "learning_rate": 1e-06, "loss": -0.0133, "num_tokens": 14917715.0, "reward": 2.1989898681640625, "reward_std": 0.22341391444206238, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1989898681640625, "rewards/icrm_reward/std": 0.2708163857460022, "step": 43 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 752.138671875, "completions/mean_terminated_length": 730.1023559570312, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.03313253012048193, "frac_reward_zero_std": 0.0, "grad_norm": 0.18046841025352478, "learning_rate": 1e-06, "loss": -0.0114, "num_tokens": 15353770.0, "reward": 2.202545166015625, "reward_std": 0.22210244834423065, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.202545166015625, "rewards/icrm_reward/std": 0.30588632822036743, "step": 44 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1520.0, "completions/mean_length": 648.603515625, "completions/mean_terminated_length": 619.977783203125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.0338855421686747, "frac_reward_zero_std": 0.0, "grad_norm": 0.21456339955329895, "learning_rate": 1e-06, "loss": -0.0282, "num_tokens": 15731183.0, "reward": 2.1501007080078125, "reward_std": 0.24410219490528107, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1501007080078125, "rewards/icrm_reward/std": 0.3128403425216675, "step": 45 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 664.4765625, "completions/mean_terminated_length": 662.7710571289062, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.03463855421686747, "frac_reward_zero_std": 0.0, "grad_norm": 0.23055920004844666, "learning_rate": 1e-06, "loss": -0.0041, "num_tokens": 16113635.0, "reward": 2.2421112060546875, "reward_std": 0.23560726642608643, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2421112060546875, "rewards/icrm_reward/std": 0.29645806550979614, "step": 46 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.013671875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 679.16796875, "completions/mean_terminated_length": 667.2911376953125, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.035391566265060244, "frac_reward_zero_std": 0.0, "grad_norm": 0.20098020136356354, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 16510985.0, "reward": 2.1724700927734375, "reward_std": 0.24453820288181305, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.1724700927734375, "rewards/icrm_reward/std": 0.3007848858833313, "step": 47 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1473.0, "completions/mean_length": 629.65625, "completions/mean_terminated_length": 626.1019897460938, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.03614457831325301, "frac_reward_zero_std": 0.0, "grad_norm": 0.21083861589431763, "learning_rate": 1e-06, "loss": -0.0175, "num_tokens": 16878169.0, "reward": 2.199188232421875, "reward_std": 0.21884912252426147, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.199188232421875, "rewards/icrm_reward/std": 0.27998584508895874, "step": 48 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.013671875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 673.232421875, "completions/mean_terminated_length": 661.2733154296875, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.036897590361445784, "frac_reward_zero_std": 0.0, "grad_norm": 0.22509372234344482, "learning_rate": 1e-06, "loss": -0.0074, "num_tokens": 17269376.0, "reward": 2.2296295166015625, "reward_std": 0.2356814444065094, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2296295166015625, "rewards/icrm_reward/std": 0.2954954504966736, "step": 49 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1424.0, "completions/mean_length": 636.708984375, "completions/mean_terminated_length": 629.6279296875, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.03765060240963856, "frac_reward_zero_std": 0.0, "grad_norm": 0.24557623267173767, "learning_rate": 1e-06, "loss": -0.0308, "num_tokens": 17640139.0, "reward": 2.254730224609375, "reward_std": 0.24260641634464264, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.254730224609375, "rewards/icrm_reward/std": 0.2840177118778229, "step": 50 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 600.349609375, "completions/mean_terminated_length": 598.5186157226562, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.038403614457831324, "frac_reward_zero_std": 0.0, "grad_norm": 0.24236264824867249, "learning_rate": 1e-06, "loss": -0.0534, "num_tokens": 17994222.0, "reward": 2.2987060546875, "reward_std": 0.24154716730117798, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2987060546875, "rewards/icrm_reward/std": 0.3029851019382477, "step": 51 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 686.248046875, "completions/mean_terminated_length": 677.8678588867188, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.0391566265060241, "frac_reward_zero_std": 0.0, "grad_norm": 0.21713781356811523, "learning_rate": 1e-06, "loss": -0.0079, "num_tokens": 18393245.0, "reward": 2.2697906494140625, "reward_std": 0.2312566339969635, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.2697906494140625, "rewards/icrm_reward/std": 0.3033978343009949, "step": 52 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 734.234375, "completions/mean_terminated_length": 729.5088500976562, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.039909638554216864, "frac_reward_zero_std": 0.0, "grad_norm": 0.20006483793258667, "learning_rate": 1e-06, "loss": 0.0053, "num_tokens": 18813301.0, "reward": 2.4068450927734375, "reward_std": 0.24239230155944824, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4068450927734375, "rewards/icrm_reward/std": 0.2884853184223175, "step": 53 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1436.0, "completions/mean_length": 682.37109375, "completions/mean_terminated_length": 673.9526977539062, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.04066265060240964, "frac_reward_zero_std": 0.0, "grad_norm": 0.22565065324306488, "learning_rate": 1e-06, "loss": 0.0167, "num_tokens": 19205987.0, "reward": 2.411712646484375, "reward_std": 0.2517935633659363, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.411712646484375, "rewards/icrm_reward/std": 0.29691916704177856, "step": 54 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 686.0625, "completions/mean_terminated_length": 675.9841918945312, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.04141566265060241, "frac_reward_zero_std": 0.0, "grad_norm": 0.25505706667900085, "learning_rate": 1e-06, "loss": 0.0116, "num_tokens": 19603747.0, "reward": 2.3581085205078125, "reward_std": 0.25292664766311646, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.3581085205078125, "rewards/icrm_reward/std": 0.3019520044326782, "step": 55 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.021484375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1487.0, "completions/mean_length": 738.88671875, "completions/mean_terminated_length": 721.38525390625, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.04216867469879518, "frac_reward_zero_std": 0.0, "grad_norm": 0.19398604333400726, "learning_rate": 1e-06, "loss": 0.0298, "num_tokens": 20029593.0, "reward": 2.420806884765625, "reward_std": 0.2693680226802826, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.420806884765625, "rewards/icrm_reward/std": 0.33825576305389404, "step": 56 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.017578125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1482.0, "completions/mean_length": 693.65625, "completions/mean_terminated_length": 678.58447265625, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.04292168674698795, "frac_reward_zero_std": 0.0, "grad_norm": 0.23698727786540985, "learning_rate": 1e-06, "loss": 0.0359, "num_tokens": 20431033.0, "reward": 2.45440673828125, "reward_std": 0.2709549069404602, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.45440673828125, "rewards/icrm_reward/std": 0.3200661540031433, "step": 57 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1520.0, "completions/max_terminated_length": 1520.0, "completions/mean_length": 651.880859375, "completions/mean_terminated_length": 651.880859375, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.043674698795180725, "frac_reward_zero_std": 0.0, "grad_norm": 0.22394704818725586, "learning_rate": 1e-06, "loss": 0.0063, "num_tokens": 20809052.0, "reward": 2.4409332275390625, "reward_std": 0.24943535029888153, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4409332275390625, "rewards/icrm_reward/std": 0.29196539521217346, "step": 58 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.029296875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1531.0, "completions/mean_length": 728.9296875, "completions/mean_terminated_length": 704.5714111328125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.04442771084337349, "frac_reward_zero_std": 0.0, "grad_norm": 0.24043507874011993, "learning_rate": 1e-06, "loss": 0.0294, "num_tokens": 21230728.0, "reward": 2.4530029296875, "reward_std": 0.29039081931114197, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4530029296875, "rewards/icrm_reward/std": 0.3333180546760559, "step": 59 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.017578125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1522.0, "completions/mean_length": 668.4765625, "completions/mean_terminated_length": 652.9542236328125, "completions/min_length": 88.0, "completions/min_terminated_length": 88.0, "epoch": 0.045180722891566265, "frac_reward_zero_std": 0.0, "grad_norm": 0.24050478637218475, "learning_rate": 1e-06, "loss": -0.0008, "num_tokens": 21620748.0, "reward": 2.416748046875, "reward_std": 0.29372185468673706, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.416748046875, "rewards/icrm_reward/std": 0.335464745759964, "step": 60 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1520.0, "completions/mean_length": 722.58203125, "completions/mean_terminated_length": 682.5778198242188, "completions/min_length": 103.0, "completions/min_terminated_length": 103.0, "epoch": 0.04593373493975904, "frac_reward_zero_std": 0.0, "grad_norm": 0.22767792642116547, "learning_rate": 1e-06, "loss": 0.0407, "num_tokens": 22032694.0, "reward": 2.4194793701171875, "reward_std": 0.2609785497188568, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4194793701171875, "rewards/icrm_reward/std": 0.35149145126342773, "step": 61 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1526.0, "completions/max_terminated_length": 1526.0, "completions/mean_length": 627.765625, "completions/mean_terminated_length": 627.765625, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.046686746987951805, "frac_reward_zero_std": 0.0, "grad_norm": 0.24805991351604462, "learning_rate": 1e-06, "loss": 0.0129, "num_tokens": 22402686.0, "reward": 2.4908294677734375, "reward_std": 0.2669978737831116, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4908294677734375, "rewards/icrm_reward/std": 0.3251093924045563, "step": 62 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.033203125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1490.0, "completions/mean_length": 766.275390625, "completions/mean_terminated_length": 739.8404541015625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.04743975903614458, "frac_reward_zero_std": 0.0, "grad_norm": 0.23006977140903473, "learning_rate": 1e-06, "loss": 0.0456, "num_tokens": 22844203.0, "reward": 2.4503021240234375, "reward_std": 0.2922241687774658, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4503021240234375, "rewards/icrm_reward/std": 0.33814340829849243, "step": 63 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.021484375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1507.0, "completions/mean_length": 725.68359375, "completions/mean_terminated_length": 707.8922119140625, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.04819277108433735, "frac_reward_zero_std": 0.0, "grad_norm": 0.22646617889404297, "learning_rate": 1e-06, "loss": 0.0394, "num_tokens": 23258473.0, "reward": 2.477783203125, "reward_std": 0.26488780975341797, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.477783203125, "rewards/icrm_reward/std": 0.33223187923431396, "step": 64 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.033203125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1503.0, "completions/mean_length": 754.828125, "completions/mean_terminated_length": 728.0000610351562, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.04894578313253012, "frac_reward_zero_std": 0.0, "grad_norm": 0.2258726954460144, "learning_rate": 1e-06, "loss": 0.0125, "num_tokens": 23690801.0, "reward": 2.45458984375, "reward_std": 0.27226293087005615, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.45458984375, "rewards/icrm_reward/std": 0.3482906222343445, "step": 65 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1522.0, "completions/mean_length": 698.974609375, "completions/mean_terminated_length": 689.0494384765625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.04969879518072289, "frac_reward_zero_std": 0.0, "grad_norm": 0.2544415593147278, "learning_rate": 1e-06, "loss": 0.0095, "num_tokens": 24092564.0, "reward": 2.50372314453125, "reward_std": 0.2619969844818115, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.50372314453125, "rewards/icrm_reward/std": 0.32770606875419617, "step": 66 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1461.0, "completions/max_terminated_length": 1461.0, "completions/mean_length": 648.00390625, "completions/mean_terminated_length": 648.00390625, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "epoch": 0.050451807228915665, "frac_reward_zero_std": 0.0, "grad_norm": 0.2574257552623749, "learning_rate": 1e-06, "loss": 0.0096, "num_tokens": 24467686.0, "reward": 2.5450897216796875, "reward_std": 0.2660745680332184, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.5450897216796875, "rewards/icrm_reward/std": 0.3114686608314514, "step": 67 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 744.58203125, "completions/mean_terminated_length": 738.3504028320312, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.05120481927710843, "frac_reward_zero_std": 0.0, "grad_norm": 0.23591530323028564, "learning_rate": 1e-06, "loss": 0.0251, "num_tokens": 24894160.0, "reward": 2.490386962890625, "reward_std": 0.25365495681762695, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.490386962890625, "rewards/icrm_reward/std": 0.3168938159942627, "step": 68 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.025390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1526.0, "completions/mean_length": 769.4921875, "completions/mean_terminated_length": 749.5230712890625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.051957831325301206, "frac_reward_zero_std": 0.0, "grad_norm": 0.20792824029922485, "learning_rate": 1e-06, "loss": 0.0559, "num_tokens": 25328156.0, "reward": 2.5142974853515625, "reward_std": 0.26831918954849243, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.5142974853515625, "rewards/icrm_reward/std": 0.33375680446624756, "step": 69 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.03515625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 776.34765625, "completions/mean_terminated_length": 748.6680297851562, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.05271084337349398, "frac_reward_zero_std": 0.0, "grad_norm": 0.209881991147995, "learning_rate": 1e-06, "loss": 0.0291, "num_tokens": 25767454.0, "reward": 2.4936981201171875, "reward_std": 0.2859433889389038, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.4936981201171875, "rewards/icrm_reward/std": 0.31998544931411743, "step": 70 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1534.0, "completions/mean_length": 756.26953125, "completions/mean_terminated_length": 737.5560302734375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.053463855421686746, "frac_reward_zero_std": 0.0, "grad_norm": 0.22324053943157196, "learning_rate": 1e-06, "loss": 0.0431, "num_tokens": 26204136.0, "reward": 2.529388427734375, "reward_std": 0.287563294172287, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.529388427734375, "rewards/icrm_reward/std": 0.3346646726131439, "step": 71 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 739.04296875, "completions/mean_terminated_length": 729.5928955078125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.05421686746987952, "frac_reward_zero_std": 0.0, "grad_norm": 0.23168617486953735, "learning_rate": 1e-06, "loss": 0.0376, "num_tokens": 26628030.0, "reward": 2.5133209228515625, "reward_std": 0.2752232849597931, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.5133209228515625, "rewards/icrm_reward/std": 0.32146480679512024, "step": 72 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 801.69140625, "completions/mean_terminated_length": 784.0680541992188, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.054969879518072286, "frac_reward_zero_std": 0.0, "grad_norm": 0.21675075590610504, "learning_rate": 1e-06, "loss": 0.0333, "num_tokens": 27084016.0, "reward": 2.6229248046875, "reward_std": 0.29856833815574646, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.6229248046875, "rewards/icrm_reward/std": 0.3686384856700897, "step": 73 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1500.0, "completions/mean_length": 859.42578125, "completions/mean_terminated_length": 840.4055786132812, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.05572289156626506, "frac_reward_zero_std": 0.0, "grad_norm": 0.21499691903591156, "learning_rate": 1e-06, "loss": 0.0493, "num_tokens": 27565834.0, "reward": 2.593292236328125, "reward_std": 0.2805570960044861, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.593292236328125, "rewards/icrm_reward/std": 0.36220526695251465, "step": 74 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1518.0, "completions/mean_length": 773.33984375, "completions/mean_terminated_length": 764.2964477539062, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.05647590361445783, "frac_reward_zero_std": 0.0, "grad_norm": 0.2344229519367218, "learning_rate": 1e-06, "loss": 0.0423, "num_tokens": 28007656.0, "reward": 2.5948028564453125, "reward_std": 0.293756902217865, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.5948028564453125, "rewards/icrm_reward/std": 0.3333764374256134, "step": 75 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.04296875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1534.0, "completions/mean_length": 817.533203125, "completions/mean_terminated_length": 785.2754516601562, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.0572289156626506, "frac_reward_zero_std": 0.0, "grad_norm": 0.22105208039283752, "learning_rate": 1e-06, "loss": 0.0472, "num_tokens": 28470185.0, "reward": 2.571807861328125, "reward_std": 0.3125259280204773, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.571807861328125, "rewards/icrm_reward/std": 0.39160090684890747, "step": 76 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.04296875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 840.224609375, "completions/mean_terminated_length": 808.9856567382812, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.05798192771084337, "frac_reward_zero_std": 0.0, "grad_norm": 0.2159413844347, "learning_rate": 1e-06, "loss": 0.0593, "num_tokens": 28947228.0, "reward": 2.6000213623046875, "reward_std": 0.31856799125671387, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.6000213623046875, "rewards/icrm_reward/std": 0.3859856426715851, "step": 77 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1521.0, "completions/mean_length": 812.81640625, "completions/mean_terminated_length": 792.4859008789062, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.058734939759036146, "frac_reward_zero_std": 0.0, "grad_norm": 0.24093309044837952, "learning_rate": 1e-06, "loss": 0.0314, "num_tokens": 29408302.0, "reward": 2.6257476806640625, "reward_std": 0.29263830184936523, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.6257476806640625, "rewards/icrm_reward/std": 0.36981090903282166, "step": 78 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.076171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 920.380859375, "completions/mean_terminated_length": 869.62158203125, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.05948795180722891, "frac_reward_zero_std": 0.0, "grad_norm": 0.22334524989128113, "learning_rate": 1e-06, "loss": 0.0437, "num_tokens": 29929569.0, "reward": 2.54388427734375, "reward_std": 0.32523876428604126, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.54388427734375, "rewards/icrm_reward/std": 0.43486225605010986, "step": 79 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.041015625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 916.134765625, "completions/mean_terminated_length": 889.623291015625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.060240963855421686, "frac_reward_zero_std": 0.0, "grad_norm": 0.230862557888031, "learning_rate": 1e-06, "loss": 0.0692, "num_tokens": 30444198.0, "reward": 2.6412506103515625, "reward_std": 0.3409099280834198, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.6412506103515625, "rewards/icrm_reward/std": 0.39923959970474243, "step": 80 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.021484375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 770.642578125, "completions/mean_terminated_length": 753.8383178710938, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.06099397590361446, "frac_reward_zero_std": 0.0, "grad_norm": 0.26526880264282227, "learning_rate": 1e-06, "loss": 0.0325, "num_tokens": 30882223.0, "reward": 2.6500091552734375, "reward_std": 0.31351447105407715, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.6500091552734375, "rewards/icrm_reward/std": 0.37751612067222595, "step": 81 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1524.0, "completions/mean_length": 793.802734375, "completions/mean_terminated_length": 789.4283447265625, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.061746987951807226, "frac_reward_zero_std": 0.0, "grad_norm": 0.2738642990589142, "learning_rate": 1e-06, "loss": 0.0415, "num_tokens": 31337226.0, "reward": 2.7096099853515625, "reward_std": 0.32019662857055664, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.7096099853515625, "rewards/icrm_reward/std": 0.3571399450302124, "step": 82 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1526.0, "completions/mean_length": 802.03125, "completions/mean_terminated_length": 784.416015625, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.0625, "frac_reward_zero_std": 0.0, "grad_norm": 0.29248520731925964, "learning_rate": 1e-06, "loss": 0.0371, "num_tokens": 31799898.0, "reward": 2.713287353515625, "reward_std": 0.30860912799835205, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.713287353515625, "rewards/icrm_reward/std": 0.38378050923347473, "step": 83 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 831.005859375, "completions/mean_terminated_length": 808.2640991210938, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.06325301204819277, "frac_reward_zero_std": 0.0, "grad_norm": 0.2955339550971985, "learning_rate": 1e-06, "loss": 0.0373, "num_tokens": 32271437.0, "reward": 2.73077392578125, "reward_std": 0.34596526622772217, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.73077392578125, "rewards/icrm_reward/std": 0.3958691656589508, "step": 84 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.013671875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1454.0, "completions/mean_length": 784.931640625, "completions/mean_terminated_length": 774.5208129882812, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.06400602409638555, "frac_reward_zero_std": 0.0, "grad_norm": 0.2862742841243744, "learning_rate": 1e-06, "loss": 0.0388, "num_tokens": 32717626.0, "reward": 2.7470855712890625, "reward_std": 0.3193947374820709, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.7470855712890625, "rewards/icrm_reward/std": 0.377297043800354, "step": 85 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1518.0, "completions/mean_length": 779.779296875, "completions/mean_terminated_length": 770.812255859375, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.06475903614457831, "frac_reward_zero_std": 0.0, "grad_norm": 0.27621668577194214, "learning_rate": 1e-06, "loss": 0.0273, "num_tokens": 33162073.0, "reward": 2.8101959228515625, "reward_std": 0.3282198905944824, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.8101959228515625, "rewards/icrm_reward/std": 0.38994351029396057, "step": 86 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 819.9453125, "completions/mean_terminated_length": 815.7249755859375, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.06551204819277108, "frac_reward_zero_std": 0.0, "grad_norm": 0.2889441251754761, "learning_rate": 1e-06, "loss": 0.0111, "num_tokens": 33630589.0, "reward": 2.8496246337890625, "reward_std": 0.31779032945632935, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.8496246337890625, "rewards/icrm_reward/std": 0.3903903365135193, "step": 87 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 808.7109375, "completions/mean_terminated_length": 779.1463012695312, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "epoch": 0.06626506024096386, "frac_reward_zero_std": 0.0, "grad_norm": 0.32299426198005676, "learning_rate": 1e-06, "loss": 0.0374, "num_tokens": 34088185.0, "reward": 2.8147430419921875, "reward_std": 0.3786037564277649, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.8147430419921875, "rewards/icrm_reward/std": 0.45253899693489075, "step": 88 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1422.0, "completions/mean_length": 845.720703125, "completions/mean_terminated_length": 817.6605224609375, "completions/min_length": 100.0, "completions/min_terminated_length": 100.0, "epoch": 0.06701807228915663, "frac_reward_zero_std": 0.0, "grad_norm": 0.31927403807640076, "learning_rate": 1e-06, "loss": 0.0419, "num_tokens": 34564490.0, "reward": 2.8685455322265625, "reward_std": 0.4015999734401703, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.8685455322265625, "rewards/icrm_reward/std": 0.462121844291687, "step": 89 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.021484375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1517.0, "completions/mean_length": 885.173828125, "completions/mean_terminated_length": 870.8842163085938, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.0677710843373494, "frac_reward_zero_std": 0.0, "grad_norm": 0.30388039350509644, "learning_rate": 1e-06, "loss": 0.0421, "num_tokens": 35065059.0, "reward": 2.9256591796875, "reward_std": 0.381799578666687, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.9256591796875, "rewards/icrm_reward/std": 0.4587096571922302, "step": 90 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.04296875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1520.0, "completions/mean_length": 860.423828125, "completions/mean_terminated_length": 830.091796875, "completions/min_length": 93.0, "completions/min_terminated_length": 93.0, "epoch": 0.06852409638554217, "frac_reward_zero_std": 0.0, "grad_norm": 0.31677141785621643, "learning_rate": 1e-06, "loss": 0.0362, "num_tokens": 35548764.0, "reward": 2.9433441162109375, "reward_std": 0.4057391285896301, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.9433441162109375, "rewards/icrm_reward/std": 0.4758445918560028, "step": 91 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1499.0, "completions/mean_length": 880.548828125, "completions/mean_terminated_length": 864.8180541992188, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.06927710843373494, "frac_reward_zero_std": 0.0, "grad_norm": 0.2733721435070038, "learning_rate": 1e-06, "loss": 0.0376, "num_tokens": 36041157.0, "reward": 2.95556640625, "reward_std": 0.3688337206840515, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.95556640625, "rewards/icrm_reward/std": 0.42398521304130554, "step": 92 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.017578125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 882.771484375, "completions/mean_terminated_length": 871.0834350585938, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.07003012048192771, "frac_reward_zero_std": 0.0, "grad_norm": 0.30676305294036865, "learning_rate": 1e-06, "loss": 0.0245, "num_tokens": 36536928.0, "reward": 2.928802490234375, "reward_std": 0.37581419944763184, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 2.928802490234375, "rewards/icrm_reward/std": 0.4320792555809021, "step": 93 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 912.806640625, "completions/mean_terminated_length": 897.8500366210938, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.07078313253012049, "frac_reward_zero_std": 0.0, "grad_norm": 0.2867027521133423, "learning_rate": 1e-06, "loss": 0.0418, "num_tokens": 37049405.0, "reward": 3.07061767578125, "reward_std": 0.38050028681755066, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.07061767578125, "rewards/icrm_reward/std": 0.4632287621498108, "step": 94 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 887.109375, "completions/mean_terminated_length": 874.1832885742188, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.07153614457831325, "frac_reward_zero_std": 0.0, "grad_norm": 0.3053477108478546, "learning_rate": 1e-06, "loss": 0.0042, "num_tokens": 37546421.0, "reward": 3.01019287109375, "reward_std": 0.4029240608215332, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.01019287109375, "rewards/icrm_reward/std": 0.46473872661590576, "step": 95 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 895.69921875, "completions/mean_terminated_length": 880.33203125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.07228915662650602, "frac_reward_zero_std": 0.0, "grad_norm": 0.3111552894115448, "learning_rate": 1e-06, "loss": 0.0166, "num_tokens": 38050571.0, "reward": 3.07513427734375, "reward_std": 0.3884848952293396, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.07513427734375, "rewards/icrm_reward/std": 0.47221118211746216, "step": 96 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1512.0, "completions/mean_length": 902.5625, "completions/mean_terminated_length": 882.1290283203125, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.0730421686746988, "frac_reward_zero_std": 0.0, "grad_norm": 0.272835373878479, "learning_rate": 1e-06, "loss": 0.0133, "num_tokens": 38557883.0, "reward": 3.056396484375, "reward_std": 0.4365222454071045, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.056396484375, "rewards/icrm_reward/std": 0.5650331974029541, "step": 97 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.005859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1490.0, "completions/mean_length": 836.984375, "completions/mean_terminated_length": 832.8644409179688, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.07379518072289157, "frac_reward_zero_std": 0.0, "grad_norm": 0.32397598028182983, "learning_rate": 1e-06, "loss": 0.0132, "num_tokens": 39035331.0, "reward": 3.0126800537109375, "reward_std": 0.3690912127494812, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.0126800537109375, "rewards/icrm_reward/std": 0.45690035820007324, "step": 98 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1495.0, "completions/mean_length": 909.12890625, "completions/mean_terminated_length": 904.1929321289062, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.07454819277108433, "frac_reward_zero_std": 0.0, "grad_norm": 0.2951310873031616, "learning_rate": 1e-06, "loss": -0.0122, "num_tokens": 39545477.0, "reward": 3.144927978515625, "reward_std": 0.367910236120224, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.144927978515625, "rewards/icrm_reward/std": 0.42396417260169983, "step": 99 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.025390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 900.78515625, "completions/mean_terminated_length": 884.2365112304688, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.07530120481927711, "frac_reward_zero_std": 0.0, "grad_norm": 0.28665030002593994, "learning_rate": 1e-06, "loss": 0.0229, "num_tokens": 40053143.0, "reward": 3.1378173828125, "reward_std": 0.40286144614219666, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.1378173828125, "rewards/icrm_reward/std": 0.4643743336200714, "step": 100 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.03515625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1519.0, "completions/mean_length": 945.23046875, "completions/mean_terminated_length": 923.7044677734375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.07605421686746988, "frac_reward_zero_std": 0.0, "grad_norm": 0.2595527172088623, "learning_rate": 1e-06, "loss": 0.0348, "num_tokens": 40579549.0, "reward": 3.1948089599609375, "reward_std": 0.4271482229232788, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.1948089599609375, "rewards/icrm_reward/std": 0.49850770831108093, "step": 101 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1531.0, "completions/mean_length": 933.57421875, "completions/mean_terminated_length": 903.9466552734375, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.07680722891566265, "frac_reward_zero_std": 0.0, "grad_norm": 0.2670970857143402, "learning_rate": 1e-06, "loss": 0.0484, "num_tokens": 41107283.0, "reward": 3.1592559814453125, "reward_std": 0.4245299994945526, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.1592559814453125, "rewards/icrm_reward/std": 0.5013904571533203, "step": 102 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.037109375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 933.482421875, "completions/mean_terminated_length": 910.2616577148438, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.07756024096385543, "frac_reward_zero_std": 0.0, "grad_norm": 0.2757066488265991, "learning_rate": 1e-06, "loss": 0.0084, "num_tokens": 41630394.0, "reward": 3.312530517578125, "reward_std": 0.44548290967941284, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.312530517578125, "rewards/icrm_reward/std": 0.5510988831520081, "step": 103 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.017578125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1523.0, "completions/mean_length": 947.587890625, "completions/mean_terminated_length": 937.0595703125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.0783132530120482, "frac_reward_zero_std": 0.0, "grad_norm": 0.26052749156951904, "learning_rate": 1e-06, "loss": 0.0167, "num_tokens": 42162615.0, "reward": 3.2794189453125, "reward_std": 0.38833147287368774, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.2794189453125, "rewards/icrm_reward/std": 0.44736889004707336, "step": 104 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.025390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 961.701171875, "completions/mean_terminated_length": 946.739501953125, "completions/min_length": 457.0, "completions/min_terminated_length": 457.0, "epoch": 0.07906626506024096, "frac_reward_zero_std": 0.0, "grad_norm": 0.25815528631210327, "learning_rate": 1e-06, "loss": 0.0105, "num_tokens": 42698590.0, "reward": 3.326141357421875, "reward_std": 0.4088899791240692, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.326141357421875, "rewards/icrm_reward/std": 0.49012550711631775, "step": 105 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.05859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 915.328125, "completions/mean_terminated_length": 876.6971435546875, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.07981927710843373, "frac_reward_zero_std": 0.0, "grad_norm": 0.3288217782974243, "learning_rate": 1e-06, "loss": 0.0242, "num_tokens": 43214550.0, "reward": 3.265228271484375, "reward_std": 0.3938843011856079, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.265228271484375, "rewards/icrm_reward/std": 0.5416343808174133, "step": 106 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.060546875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1525.0, "completions/mean_length": 932.228515625, "completions/mean_terminated_length": 893.3160400390625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.08057228915662651, "frac_reward_zero_std": 0.0, "grad_norm": 0.2760145366191864, "learning_rate": 1e-06, "loss": 0.0482, "num_tokens": 43737771.0, "reward": 3.3134918212890625, "reward_std": 0.4481176435947418, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.3134918212890625, "rewards/icrm_reward/std": 0.5695783495903015, "step": 107 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.095703125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1043.150390625, "completions/mean_terminated_length": 990.9913330078125, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.08132530120481928, "frac_reward_zero_std": 0.0, "grad_norm": 0.2719205319881439, "learning_rate": 1e-06, "loss": 0.0224, "num_tokens": 44318760.0, "reward": 3.3242950439453125, "reward_std": 0.46499037742614746, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.3242950439453125, "rewards/icrm_reward/std": 0.6479662656784058, "step": 108 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 1041.3828125, "completions/mean_terminated_length": 999.4661254882812, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.08207831325301204, "frac_reward_zero_std": 0.0, "grad_norm": 0.2970697581768036, "learning_rate": 1e-06, "loss": 0.0232, "num_tokens": 44896012.0, "reward": 3.4527130126953125, "reward_std": 0.4472125172615051, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.4527130126953125, "rewards/icrm_reward/std": 0.6081138849258423, "step": 109 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.052734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1054.5625, "completions/mean_terminated_length": 1027.7608642578125, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.08283132530120482, "frac_reward_zero_std": 0.0, "grad_norm": 0.29613715410232544, "learning_rate": 1e-06, "loss": 0.0159, "num_tokens": 45483756.0, "reward": 3.5418701171875, "reward_std": 0.5048837661743164, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.5418701171875, "rewards/icrm_reward/std": 0.6320242285728455, "step": 110 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.05078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1517.0, "completions/mean_length": 1061.486328125, "completions/mean_terminated_length": 1036.100830078125, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.08358433734939759, "frac_reward_zero_std": 0.0, "grad_norm": 0.31563860177993774, "learning_rate": 1e-06, "loss": 0.0197, "num_tokens": 46069909.0, "reward": 3.607818603515625, "reward_std": 0.48346221446990967, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.607818603515625, "rewards/icrm_reward/std": 0.6252943873405457, "step": 111 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.072265625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 1062.2109375, "completions/mean_terminated_length": 1025.30517578125, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.08433734939759036, "frac_reward_zero_std": 0.0, "grad_norm": 0.3350493311882019, "learning_rate": 1e-06, "loss": 0.0268, "num_tokens": 46662017.0, "reward": 3.5582427978515625, "reward_std": 0.49701404571533203, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.5582427978515625, "rewards/icrm_reward/std": 0.6976381540298462, "step": 112 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1031.458984375, "completions/mean_terminated_length": 1019.3500366210938, "completions/min_length": 515.0, "completions/min_terminated_length": 515.0, "epoch": 0.08509036144578314, "frac_reward_zero_std": 0.0, "grad_norm": 0.334583580493927, "learning_rate": 1e-06, "loss": 0.016, "num_tokens": 47235772.0, "reward": 3.6624908447265625, "reward_std": 0.45720618963241577, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.6624908447265625, "rewards/icrm_reward/std": 0.6205590963363647, "step": 113 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.080078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1115.578125, "completions/mean_terminated_length": 1078.98095703125, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.0858433734939759, "frac_reward_zero_std": 0.0, "grad_norm": 0.3087984323501587, "learning_rate": 1e-06, "loss": 0.0323, "num_tokens": 47852820.0, "reward": 3.5841064453125, "reward_std": 0.5725352764129639, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.5841064453125, "rewards/icrm_reward/std": 0.7000143527984619, "step": 114 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.07421875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 1063.62109375, "completions/mean_terminated_length": 1025.7509765625, "completions/min_length": 530.0, "completions/min_terminated_length": 530.0, "epoch": 0.08659638554216867, "frac_reward_zero_std": 0.0, "grad_norm": 0.33008214831352234, "learning_rate": 1e-06, "loss": 0.0192, "num_tokens": 48446802.0, "reward": 3.625518798828125, "reward_std": 0.49653565883636475, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.625518798828125, "rewards/icrm_reward/std": 0.6842305064201355, "step": 115 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0546875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1078.359375, "completions/mean_terminated_length": 1051.88427734375, "completions/min_length": 604.0, "completions/min_terminated_length": 604.0, "epoch": 0.08734939759036145, "frac_reward_zero_std": 0.0, "grad_norm": 0.35016050934791565, "learning_rate": 1e-06, "loss": 0.0386, "num_tokens": 49046842.0, "reward": 3.6656494140625, "reward_std": 0.5081939101219177, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.6656494140625, "rewards/icrm_reward/std": 0.6160948276519775, "step": 116 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.068359375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 1129.044921875, "completions/mean_terminated_length": 1099.1844482421875, "completions/min_length": 514.0, "completions/min_terminated_length": 514.0, "epoch": 0.08810240963855422, "frac_reward_zero_std": 0.0, "grad_norm": 0.34213417768478394, "learning_rate": 1e-06, "loss": 0.043, "num_tokens": 49670449.0, "reward": 3.842926025390625, "reward_std": 0.5554987192153931, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.842926025390625, "rewards/icrm_reward/std": 0.6801345348358154, "step": 117 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1073.904296875, "completions/mean_terminated_length": 1026.101318359375, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.08885542168674698, "frac_reward_zero_std": 0.0, "grad_norm": 0.35978832840919495, "learning_rate": 1e-06, "loss": 0.0266, "num_tokens": 50263392.0, "reward": 3.68927001953125, "reward_std": 0.6048953533172607, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.68927001953125, "rewards/icrm_reward/std": 0.7318205833435059, "step": 118 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.041015625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 1037.619140625, "completions/mean_terminated_length": 1016.3035278320312, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.08960843373493976, "frac_reward_zero_std": 0.0, "grad_norm": 0.40669694542884827, "learning_rate": 1e-06, "loss": 0.0281, "num_tokens": 50840125.0, "reward": 3.836822509765625, "reward_std": 0.5344001054763794, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.836822509765625, "rewards/icrm_reward/std": 0.7598003149032593, "step": 119 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.033203125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 1001.3125, "completions/mean_terminated_length": 982.9495239257812, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.09036144578313253, "frac_reward_zero_std": 0.0, "grad_norm": 0.3870098888874054, "learning_rate": 1e-06, "loss": 0.0093, "num_tokens": 51397917.0, "reward": 3.87884521484375, "reward_std": 0.5162875652313232, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.87884521484375, "rewards/icrm_reward/std": 0.6479455828666687, "step": 120 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1528.0, "completions/mean_length": 1010.32421875, "completions/mean_terminated_length": 1006.18505859375, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.0911144578313253, "frac_reward_zero_std": 0.0, "grad_norm": 0.39014145731925964, "learning_rate": 1e-06, "loss": -0.0001, "num_tokens": 51959539.0, "reward": 3.9730072021484375, "reward_std": 0.4872249364852905, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.9730072021484375, "rewards/icrm_reward/std": 0.6410709023475647, "step": 121 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 1034.78515625, "completions/mean_terminated_length": 1001.3709106445312, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.09186746987951808, "frac_reward_zero_std": 0.0, "grad_norm": 0.4095689058303833, "learning_rate": 1e-06, "loss": 0.0201, "num_tokens": 52537701.0, "reward": 3.8944854736328125, "reward_std": 0.5506025552749634, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 3.8944854736328125, "rewards/icrm_reward/std": 0.675002932548523, "step": 122 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.087890625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1102.3671875, "completions/mean_terminated_length": 1060.5823974609375, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.09262048192771084, "frac_reward_zero_std": 0.0, "grad_norm": 0.39829981327056885, "learning_rate": 1e-06, "loss": 0.0222, "num_tokens": 53145905.0, "reward": 4.00750732421875, "reward_std": 0.6116676330566406, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.00750732421875, "rewards/icrm_reward/std": 0.8072155117988586, "step": 123 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.025390625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1062.55078125, "completions/mean_terminated_length": 1050.2164306640625, "completions/min_length": 484.0, "completions/min_terminated_length": 484.0, "epoch": 0.09337349397590361, "frac_reward_zero_std": 0.0, "grad_norm": 0.430706262588501, "learning_rate": 1e-06, "loss": 0.0132, "num_tokens": 53736795.0, "reward": 4.0050048828125, "reward_std": 0.5629526972770691, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.0050048828125, "rewards/icrm_reward/std": 0.7918898463249207, "step": 124 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1077.34765625, "completions/mean_terminated_length": 1066.340087890625, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.09412650602409639, "frac_reward_zero_std": 0.0, "grad_norm": 0.45189860463142395, "learning_rate": 1e-06, "loss": -0.0027, "num_tokens": 54334269.0, "reward": 4.1801300048828125, "reward_std": 0.5930963754653931, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.1801300048828125, "rewards/icrm_reward/std": 0.6826416850090027, "step": 125 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.01953125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1522.0, "completions/mean_length": 1056.853515625, "completions/mean_terminated_length": 1047.308837890625, "completions/min_length": 554.0, "completions/min_terminated_length": 554.0, "epoch": 0.09487951807228916, "frac_reward_zero_std": 0.0, "grad_norm": 0.45464563369750977, "learning_rate": 1e-06, "loss": 0.0094, "num_tokens": 54921730.0, "reward": 4.1747894287109375, "reward_std": 0.5486907958984375, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.1747894287109375, "rewards/icrm_reward/std": 0.7279157042503357, "step": 126 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.056640625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 1100.255859375, "completions/mean_terminated_length": 1074.0931396484375, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.09563253012048192, "frac_reward_zero_std": 0.0, "grad_norm": 0.43557268381118774, "learning_rate": 1e-06, "loss": 0.0199, "num_tokens": 55532117.0, "reward": 4.1165618896484375, "reward_std": 0.652687668800354, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.1165618896484375, "rewards/icrm_reward/std": 0.8171107172966003, "step": 127 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.037109375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1531.0, "completions/mean_length": 1049.5, "completions/mean_terminated_length": 1030.75048828125, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.0963855421686747, "frac_reward_zero_std": 0.0, "grad_norm": 1.2762600183486938, "learning_rate": 1e-06, "loss": 0.0118, "num_tokens": 56113925.0, "reward": 4.0705718994140625, "reward_std": 0.5888340473175049, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.0705718994140625, "rewards/icrm_reward/std": 0.7753996849060059, "step": 128 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0234375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1075.615234375, "completions/mean_terminated_length": 1064.5660400390625, "completions/min_length": 498.0, "completions/min_terminated_length": 498.0, "epoch": 0.09713855421686747, "frac_reward_zero_std": 0.0, "grad_norm": 0.5985656976699829, "learning_rate": 1e-06, "loss": 0.0204, "num_tokens": 56707232.0, "reward": 4.26837158203125, "reward_std": 0.6652629971504211, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.26837158203125, "rewards/icrm_reward/std": 0.7837300896644592, "step": 129 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0546875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1517.0, "completions/mean_length": 1126.310546875, "completions/mean_terminated_length": 1102.6094970703125, "completions/min_length": 487.0, "completions/min_terminated_length": 487.0, "epoch": 0.09789156626506024, "frac_reward_zero_std": 0.0, "grad_norm": 0.6995136737823486, "learning_rate": 1e-06, "loss": 0.0255, "num_tokens": 57327535.0, "reward": 4.19671630859375, "reward_std": 0.7013879418373108, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.19671630859375, "rewards/icrm_reward/std": 0.8030691742897034, "step": 130 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.037109375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1534.0, "completions/mean_length": 1108.6328125, "completions/mean_terminated_length": 1092.1622314453125, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.09864457831325302, "frac_reward_zero_std": 0.0, "grad_norm": 0.5040700435638428, "learning_rate": 1e-06, "loss": 0.0236, "num_tokens": 57940419.0, "reward": 4.236175537109375, "reward_std": 0.7264065742492676, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.236175537109375, "rewards/icrm_reward/std": 0.8259881138801575, "step": 131 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.052734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1150.2265625, "completions/mean_terminated_length": 1128.7506103515625, "completions/min_length": 550.0, "completions/min_terminated_length": 550.0, "epoch": 0.09939759036144578, "frac_reward_zero_std": 0.0, "grad_norm": 0.5671800971031189, "learning_rate": 1e-06, "loss": 0.0306, "num_tokens": 58572871.0, "reward": 4.3385772705078125, "reward_std": 0.7649765610694885, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.3385772705078125, "rewards/icrm_reward/std": 0.9581444263458252, "step": 132 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.091796875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1531.0, "completions/mean_length": 1201.853515625, "completions/mean_terminated_length": 1168.07958984375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.10015060240963855, "frac_reward_zero_std": 0.0, "grad_norm": 0.5382966995239258, "learning_rate": 1e-06, "loss": 0.0238, "num_tokens": 59231148.0, "reward": 4.30035400390625, "reward_std": 0.7863367199897766, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.30035400390625, "rewards/icrm_reward/std": 0.9911512732505798, "step": 133 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.1328125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1527.0, "completions/mean_length": 1154.716796875, "completions/mean_terminated_length": 1096.3221435546875, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.10090361445783133, "frac_reward_zero_std": 0.0, "grad_norm": 0.5057360529899597, "learning_rate": 1e-06, "loss": 0.0231, "num_tokens": 59869675.0, "reward": 4.17413330078125, "reward_std": 0.7724943161010742, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.17413330078125, "rewards/icrm_reward/std": 1.1001423597335815, "step": 134 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.037109375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1109.478515625, "completions/mean_terminated_length": 1093.04052734375, "completions/min_length": 680.0, "completions/min_terminated_length": 680.0, "epoch": 0.1016566265060241, "frac_reward_zero_std": 0.0, "grad_norm": 0.4980197846889496, "learning_rate": 1e-06, "loss": 0.0189, "num_tokens": 60486016.0, "reward": 4.4792022705078125, "reward_std": 0.6606467962265015, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.4792022705078125, "rewards/icrm_reward/std": 0.7966200113296509, "step": 135 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 1091.37890625, "completions/mean_terminated_length": 1078.8795166015625, "completions/min_length": 513.0, "completions/min_terminated_length": 513.0, "epoch": 0.10240963855421686, "frac_reward_zero_std": 0.0, "grad_norm": 0.5023417472839355, "learning_rate": 1e-06, "loss": 0.0115, "num_tokens": 61090290.0, "reward": 4.5316009521484375, "reward_std": 0.6734751462936401, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.5316009521484375, "rewards/icrm_reward/std": 0.8923114538192749, "step": 136 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.044921875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1130.16015625, "completions/mean_terminated_length": 1111.071533203125, "completions/min_length": 592.0, "completions/min_terminated_length": 592.0, "epoch": 0.10316265060240964, "frac_reward_zero_std": 0.0, "grad_norm": 0.5186536312103271, "learning_rate": 1e-06, "loss": 0.0244, "num_tokens": 61718724.0, "reward": 4.3646087646484375, "reward_std": 0.6948167085647583, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.3646087646484375, "rewards/icrm_reward/std": 0.826988160610199, "step": 137 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.076171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1533.0, "completions/mean_length": 1148.951171875, "completions/mean_terminated_length": 1117.0380859375, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.10391566265060241, "frac_reward_zero_std": 0.0, "grad_norm": 0.5542064905166626, "learning_rate": 1e-06, "loss": 0.0254, "num_tokens": 62349675.0, "reward": 4.36920166015625, "reward_std": 0.7360410690307617, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.36920166015625, "rewards/icrm_reward/std": 0.9057928323745728, "step": 138 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0703125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1157.771484375, "completions/mean_terminated_length": 1129.166015625, "completions/min_length": 699.0, "completions/min_terminated_length": 699.0, "epoch": 0.10466867469879518, "frac_reward_zero_std": 0.0, "grad_norm": 0.4985194802284241, "learning_rate": 1e-06, "loss": 0.0401, "num_tokens": 62992966.0, "reward": 4.42987060546875, "reward_std": 0.7369166612625122, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.42987060546875, "rewards/icrm_reward/std": 0.9678148031234741, "step": 139 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.05859375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1534.0, "completions/mean_length": 1177.798828125, "completions/mean_terminated_length": 1155.504150390625, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.10542168674698796, "frac_reward_zero_std": 0.0, "grad_norm": 0.48446330428123474, "learning_rate": 1e-06, "loss": 0.0295, "num_tokens": 63639679.0, "reward": 4.651458740234375, "reward_std": 0.7585673332214355, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.651458740234375, "rewards/icrm_reward/std": 0.9639466404914856, "step": 140 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.07421875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1178.580078125, "completions/mean_terminated_length": 1149.9261474609375, "completions/min_length": 594.0, "completions/min_terminated_length": 594.0, "epoch": 0.10617469879518072, "frac_reward_zero_std": 0.0, "grad_norm": 0.5235411524772644, "learning_rate": 1e-06, "loss": 0.0409, "num_tokens": 64287064.0, "reward": 4.468170166015625, "reward_std": 0.8118422031402588, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.468170166015625, "rewards/icrm_reward/std": 0.9429880380630493, "step": 141 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.12890625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1532.0, "completions/mean_length": 1191.279296875, "completions/mean_terminated_length": 1140.266845703125, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.10692771084337349, "frac_reward_zero_std": 0.0, "grad_norm": 0.5035867094993591, "learning_rate": 1e-06, "loss": 0.021, "num_tokens": 64939303.0, "reward": 4.414886474609375, "reward_std": 0.7666824460029602, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.414886474609375, "rewards/icrm_reward/std": 1.1432216167449951, "step": 142 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 1205.037109375, "completions/mean_terminated_length": 1164.392578125, "completions/min_length": 658.0, "completions/min_terminated_length": 658.0, "epoch": 0.10768072289156627, "frac_reward_zero_std": 0.0, "grad_norm": 0.4995163679122925, "learning_rate": 1e-06, "loss": 0.0249, "num_tokens": 65599322.0, "reward": 4.5460968017578125, "reward_std": 0.8546550273895264, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.5460968017578125, "rewards/icrm_reward/std": 1.0646564960479736, "step": 143 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.123046875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1522.0, "completions/mean_length": 1184.833984375, "completions/mean_terminated_length": 1135.561279296875, "completions/min_length": 679.0, "completions/min_terminated_length": 679.0, "epoch": 0.10843373493975904, "frac_reward_zero_std": 0.0, "grad_norm": 0.5785022377967834, "learning_rate": 1e-06, "loss": 0.0241, "num_tokens": 66252757.0, "reward": 4.5584564208984375, "reward_std": 0.9171419143676758, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.5584564208984375, "rewards/icrm_reward/std": 1.2026164531707764, "step": 144 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.044921875, "completions/max_length": 1536.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1152.01171875, "completions/mean_terminated_length": 1133.950927734375, "completions/min_length": 670.0, "completions/min_terminated_length": 670.0, "epoch": 0.1091867469879518, "frac_reward_zero_std": 0.0, "grad_norm": 0.5453264713287354, "learning_rate": 1e-06, "loss": 0.0126, "num_tokens": 66890171.0, "reward": 4.862518310546875, "reward_std": 0.7816354632377625, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.862518310546875, "rewards/icrm_reward/std": 0.9664557576179504, "step": 145 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.083984375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1525.0, "completions/mean_length": 1212.98828125, "completions/mean_terminated_length": 1183.3731689453125, "completions/min_length": 746.0, "completions/min_terminated_length": 746.0, "epoch": 0.10993975903614457, "frac_reward_zero_std": 0.0, "grad_norm": 0.5098794102668762, "learning_rate": 1e-06, "loss": 0.0254, "num_tokens": 67554341.0, "reward": 4.792633056640625, "reward_std": 0.8540831804275513, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.792633056640625, "rewards/icrm_reward/std": 1.080270528793335, "step": 146 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.103515625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1214.3984375, "completions/mean_terminated_length": 1177.2635498046875, "completions/min_length": 718.0, "completions/min_terminated_length": 718.0, "epoch": 0.11069277108433735, "frac_reward_zero_std": 0.0, "grad_norm": 0.4801645576953888, "learning_rate": 1e-06, "loss": 0.0272, "num_tokens": 68224849.0, "reward": 4.768585205078125, "reward_std": 0.8221277594566345, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.768585205078125, "rewards/icrm_reward/std": 1.0661983489990234, "step": 147 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.119140625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1154.353515625, "completions/mean_terminated_length": 1102.73388671875, "completions/min_length": 624.0, "completions/min_terminated_length": 624.0, "epoch": 0.11144578313253012, "frac_reward_zero_std": 0.0, "grad_norm": 0.5634762048721313, "learning_rate": 1e-06, "loss": 0.0147, "num_tokens": 68857798.0, "reward": 4.6792144775390625, "reward_std": 0.8073952198028564, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.6792144775390625, "rewards/icrm_reward/std": 1.1656705141067505, "step": 148 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1530.0, "completions/mean_length": 1139.81640625, "completions/mean_terminated_length": 1113.4041748046875, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.11219879518072289, "frac_reward_zero_std": 0.0, "grad_norm": 0.46944481134414673, "learning_rate": 1e-06, "loss": 0.0192, "num_tokens": 69488120.0, "reward": 5.034210205078125, "reward_std": 0.7564300298690796, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 5.034210205078125, "rewards/icrm_reward/std": 1.0557838678359985, "step": 149 }, { "clip_ratio": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 1536.0, "completions/max_terminated_length": 1535.0, "completions/mean_length": 1128.408203125, "completions/mean_terminated_length": 1086.2435302734375, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.11295180722891567, "frac_reward_zero_std": 0.0, "grad_norm": 0.4718353748321533, "learning_rate": 1e-06, "loss": 0.0156, "num_tokens": 70109049.0, "reward": 4.9261627197265625, "reward_std": 0.740686297416687, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "rewards/icrm_reward/mean": 4.9261627197265625, "rewards/icrm_reward/std": 1.173972725868225, "step": 150 } ], "logging_steps": 1, "max_steps": 1328, "num_input_tokens_seen": 70109049, "num_train_epochs": 1, "save_steps": 50, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 64, "trial_name": null, "trial_params": null }