{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.25, "eval_steps": 500, "global_step": 441, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3204.0, "completions/mean_length": 1194.96875, "completions/mean_terminated_length": 894.862060546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.504598390311003, "epoch": 0.0005668934240362812, "frac_reward_zero_std": 0.75, "grad_norm": 0.189453125, "kl": 0.000330764095451741, "learning_rate": 1e-06, "loss": 0.0204, "num_tokens": 86206.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 1, "step_time": 290.4004911398515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3540.0, "completions/mean_length": 1322.1875, "completions/mean_terminated_length": 746.4906005859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.345731921494007, "epoch": 0.0011337868480725624, "frac_reward_zero_std": 0.5, "grad_norm": 0.2314453125, "kl": 0.0003442415154495393, "learning_rate": 9.994331065759637e-07, "loss": 0.0463, "num_tokens": 183818.0, "reward": 0.203125, "reward_std": 0.5471335649490356, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 2, "step_time": 303.91933149658144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3854.0, "completions/mean_length": 2008.53125, "completions/mean_terminated_length": 1369.5101318359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 5.243577856570482, "epoch": 0.0017006802721088435, "frac_reward_zero_std": 0.875, "grad_norm": 0.10302734375, "kl": 0.00029097342212480726, "learning_rate": 9.988662131519275e-07, "loss": 0.0205, "num_tokens": 323068.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 3, "step_time": 299.3002619827166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3719.0, "completions/mean_length": 1336.40625, "completions/mean_terminated_length": 699.5769653320312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.6891971342265606, "epoch": 0.0022675736961451248, "frac_reward_zero_std": 0.625, "grad_norm": 0.2158203125, "kl": 0.00041494359993521357, "learning_rate": 9.982993197278912e-07, "loss": 0.0298, "num_tokens": 420150.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 4, "step_time": 294.2055949298665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3626.0, "completions/mean_length": 1425.28125, "completions/mean_terminated_length": 930.7037353515625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.3760005459189415, "epoch": 0.002834467120181406, "frac_reward_zero_std": 0.875, "grad_norm": 0.1953125, "kl": 0.0003999843729616259, "learning_rate": 9.977324263038547e-07, "loss": 0.0157, "num_tokens": 522424.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 5, "step_time": 293.4846060844138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3887.0, "completions/mean_length": 1280.796875, "completions/mean_terminated_length": 989.5689697265625, "completions/min_length": 82.0, "completions/min_terminated_length": 82.0, "entropy": 4.2538034208118916, "epoch": 0.003401360544217687, "frac_reward_zero_std": 1.0, "grad_norm": 6.29425048828125e-05, "kl": 0.0003539047747835866, "learning_rate": 9.971655328798185e-07, "loss": 0.0, "num_tokens": 616499.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 6, "step_time": 296.5241157421842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3576.0, "completions/mean_length": 1286.109375, "completions/mean_terminated_length": 765.75927734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.3659513257443905, "epoch": 0.003968253968253968, "frac_reward_zero_std": 0.625, "grad_norm": 0.2255859375, "kl": 0.0003789493484873674, "learning_rate": 9.965986394557822e-07, "loss": 0.0023, "num_tokens": 712794.0, "reward": 0.2734375, "reward_std": 0.7449815273284912, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 7, "step_time": 309.2328038820997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1192.53125, "completions/mean_terminated_length": 654.8518676757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 1.9319973811507225, "epoch": 0.0045351473922902496, "frac_reward_zero_std": 0.75, "grad_norm": 0.1875, "kl": 0.0004903619992546737, "learning_rate": 9.96031746031746e-07, "loss": 0.0217, "num_tokens": 802204.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 8, "step_time": 297.1033435128629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3889.0, "completions/mean_length": 1329.078125, "completions/mean_terminated_length": 754.8113403320312, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 2.8315951824188232, "epoch": 0.00510204081632653, "frac_reward_zero_std": 0.75, "grad_norm": 0.1494140625, "kl": 0.00039701382593193557, "learning_rate": 9.954648526077097e-07, "loss": -0.0048, "num_tokens": 899905.0, "reward": 0.1796875, "reward_std": 0.5866720080375671, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 9, "step_time": 299.20237035956234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3610.0, "completions/mean_length": 1129.09375, "completions/mean_terminated_length": 764.7368774414062, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "entropy": 2.454593241214752, "epoch": 0.005668934240362812, "frac_reward_zero_std": 0.75, "grad_norm": 0.255859375, "kl": 0.00046382144591916585, "learning_rate": 9.948979591836735e-07, "loss": -0.0261, "num_tokens": 982783.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 10, "step_time": 295.87636501993984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3606.0, "completions/mean_length": 1484.53125, "completions/mean_terminated_length": 1000.9259033203125, "completions/min_length": 109.0, "completions/min_terminated_length": 109.0, "entropy": 4.456293493509293, "epoch": 0.006235827664399093, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "kl": 0.0003735653808689676, "learning_rate": 9.94331065759637e-07, "loss": 0.0021, "num_tokens": 1088801.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 11, "step_time": 296.7332943305373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3438.0, "completions/mean_length": 1243.078125, "completions/mean_terminated_length": 892.7192993164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 3.2929567843675613, "epoch": 0.006802721088435374, "frac_reward_zero_std": 0.75, "grad_norm": 0.2041015625, "kl": 0.0003951908856834052, "learning_rate": 9.937641723356008e-07, "loss": 0.0524, "num_tokens": 1179622.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 12, "step_time": 294.44508679863065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3944.0, "completions/mean_length": 1431.640625, "completions/mean_terminated_length": 995.654541015625, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "entropy": 3.8930897265672684, "epoch": 0.007369614512471655, "frac_reward_zero_std": 0.75, "grad_norm": 0.2265625, "kl": 0.00038424173817475094, "learning_rate": 9.931972789115645e-07, "loss": -0.0218, "num_tokens": 1283623.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 13, "step_time": 294.92571747582406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4029.0, "completions/mean_length": 1600.640625, "completions/mean_terminated_length": 964.5686645507812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.1041195038706064, "epoch": 0.007936507936507936, "frac_reward_zero_std": 0.875, "grad_norm": 0.11669921875, "kl": 0.00040818309116730234, "learning_rate": 9.926303854875283e-07, "loss": -0.0074, "num_tokens": 1398664.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 14, "step_time": 297.8464042544365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3815.0, "completions/mean_length": 1455.21875, "completions/mean_terminated_length": 845.8077392578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.581351801753044, "epoch": 0.008503401360544218, "frac_reward_zero_std": 0.875, "grad_norm": 0.1083984375, "kl": 0.0003867176119456417, "learning_rate": 9.92063492063492e-07, "loss": -0.0003, "num_tokens": 1503838.0, "reward": 0.15625, "reward_std": 0.5833333730697632, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 15, "step_time": 294.84605468437076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3655.0, "completions/mean_length": 1439.03125, "completions/mean_terminated_length": 887.5848999023438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.61439785733819, "epoch": 0.009070294784580499, "frac_reward_zero_std": 0.75, "grad_norm": 0.1962890625, "kl": 0.00041348650029249256, "learning_rate": 9.914965986394558e-07, "loss": 0.028, "num_tokens": 1607136.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 16, "step_time": 295.82693486101925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3806.0, "completions/mean_length": 1300.1875, "completions/mean_terminated_length": 782.4444580078125, "completions/min_length": 88.0, "completions/min_terminated_length": 88.0, "entropy": 3.3572088442742825, "epoch": 0.00963718820861678, "frac_reward_zero_std": 0.875, "grad_norm": 0.16015625, "kl": 0.00045007884546066634, "learning_rate": 9.909297052154193e-07, "loss": 0.011, "num_tokens": 1702620.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 17, "step_time": 294.88669333048165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3298.0, "completions/mean_length": 1205.390625, "completions/mean_terminated_length": 850.4035034179688, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 3.502040285617113, "epoch": 0.01020408163265306, "frac_reward_zero_std": 1.0, "grad_norm": 7.43865966796875e-05, "kl": 0.0004136692868996761, "learning_rate": 9.90362811791383e-07, "loss": 0.0, "num_tokens": 1790741.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 18, "step_time": 295.91272540576756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3705.0, "completions/mean_length": 978.734375, "completions/mean_terminated_length": 714.559326171875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.546327754855156, "epoch": 0.010770975056689343, "frac_reward_zero_std": 0.75, "grad_norm": 0.32421875, "kl": 0.0004484470018724096, "learning_rate": 9.897959183673468e-07, "loss": 0.0622, "num_tokens": 1864268.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 19, "step_time": 292.4661794034764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3801.0, "completions/mean_length": 1386.65625, "completions/mean_terminated_length": 999.607177734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.70910806581378, "epoch": 0.011337868480725623, "frac_reward_zero_std": 0.625, "grad_norm": 0.22265625, "kl": 0.00038008083629392786, "learning_rate": 9.892290249433106e-07, "loss": 0.0765, "num_tokens": 1964438.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 20, "step_time": 301.6167815560475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3725.0, "completions/mean_length": 1592.34375, "completions/mean_terminated_length": 954.1569213867188, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "entropy": 3.0310422237962484, "epoch": 0.011904761904761904, "frac_reward_zero_std": 0.75, "grad_norm": 0.1845703125, "kl": 0.00045945892634335905, "learning_rate": 9.886621315192743e-07, "loss": -0.0155, "num_tokens": 2080588.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 21, "step_time": 304.0985494265333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3757.0, "completions/mean_length": 1292.734375, "completions/mean_terminated_length": 773.6111450195312, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.643630724400282, "epoch": 0.012471655328798186, "frac_reward_zero_std": 0.75, "grad_norm": 0.1728515625, "kl": 0.00043581062300290796, "learning_rate": 9.88095238095238e-07, "loss": 0.0503, "num_tokens": 2183251.0, "reward": 0.1484375, "reward_std": 0.5246290564537048, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 22, "step_time": 343.65668355021626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3839.0, "completions/mean_length": 1119.25, "completions/mean_terminated_length": 811.3103637695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.981692060828209, "epoch": 0.013038548752834467, "frac_reward_zero_std": 0.75, "grad_norm": 0.18359375, "kl": 0.0004370460965219536, "learning_rate": 9.875283446712018e-07, "loss": -0.0105, "num_tokens": 2267099.0, "reward": 0.109375, "reward_std": 0.41277948021888733, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 23, "step_time": 296.2689810972661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3834.0, "completions/mean_length": 1517.109375, "completions/mean_terminated_length": 1095.1090087890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.816467195749283, "epoch": 0.013605442176870748, "frac_reward_zero_std": 0.75, "grad_norm": 0.1767578125, "kl": 0.00040354001430387143, "learning_rate": 9.869614512471656e-07, "loss": 0.0066, "num_tokens": 2376002.0, "reward": 0.140625, "reward_std": 0.5525464415550232, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 24, "step_time": 295.8765085712075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3787.0, "completions/mean_length": 1134.984375, "completions/mean_terminated_length": 884.0508422851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.6541102938354015, "epoch": 0.01417233560090703, "frac_reward_zero_std": 0.875, "grad_norm": 0.1875, "kl": 0.0004305867323637358, "learning_rate": 9.86394557823129e-07, "loss": 0.0042, "num_tokens": 2460153.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 25, "step_time": 298.3754459451884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4090.0, "completions/mean_length": 1289.859375, "completions/mean_terminated_length": 945.24560546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4574208334088326, "epoch": 0.01473922902494331, "frac_reward_zero_std": 0.75, "grad_norm": 0.294921875, "kl": 0.0004790655384567799, "learning_rate": 9.858276643990928e-07, "loss": -0.0492, "num_tokens": 2553296.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 26, "step_time": 293.37725717946887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3986.0, "completions/mean_length": 1333.1875, "completions/mean_terminated_length": 993.8947143554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.69367503374815, "epoch": 0.015306122448979591, "frac_reward_zero_std": 0.625, "grad_norm": 0.275390625, "kl": 0.0004016220236735535, "learning_rate": 9.852607709750566e-07, "loss": 0.0141, "num_tokens": 2650476.0, "reward": 0.109375, "reward_std": 0.46691107749938965, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 27, "step_time": 294.41278407908976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3878.0, "completions/mean_length": 1388.0, "completions/mean_terminated_length": 825.9622802734375, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "entropy": 4.408259846270084, "epoch": 0.015873015873015872, "frac_reward_zero_std": 0.5, "grad_norm": 0.212890625, "kl": 0.0004581758648782852, "learning_rate": 9.846938775510203e-07, "loss": 0.0412, "num_tokens": 2750436.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 28, "step_time": 294.78673951141536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3728.0, "completions/mean_length": 1695.734375, "completions/mean_terminated_length": 1197.5660400390625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.804402470588684, "epoch": 0.016439909297052153, "frac_reward_zero_std": 0.5, "grad_norm": 0.255859375, "kl": 0.00035922019833378727, "learning_rate": 9.84126984126984e-07, "loss": -0.001, "num_tokens": 2873219.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 29, "step_time": 300.3509062519297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4084.0, "completions/mean_length": 1472.375, "completions/mean_terminated_length": 866.923095703125, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 2.803918108344078, "epoch": 0.017006802721088437, "frac_reward_zero_std": 0.75, "grad_norm": 0.1962890625, "kl": 0.0004838097711399314, "learning_rate": 9.835600907029478e-07, "loss": 0.0102, "num_tokens": 2980107.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 30, "step_time": 306.3423152305186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4080.0, "completions/mean_length": 1367.78125, "completions/mean_terminated_length": 862.5555419921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.843935079872608, "epoch": 0.017573696145124718, "frac_reward_zero_std": 0.875, "grad_norm": 0.15234375, "kl": 0.00045960527859278955, "learning_rate": 9.829931972789116e-07, "loss": 0.0181, "num_tokens": 3078581.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 31, "step_time": 295.59960942994803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3777.0, "completions/mean_length": 1544.828125, "completions/mean_terminated_length": 956.09619140625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.920840870589018, "epoch": 0.018140589569160998, "frac_reward_zero_std": 0.75, "grad_norm": 0.1630859375, "kl": 0.0004275208129911334, "learning_rate": 9.824263038548753e-07, "loss": -0.0099, "num_tokens": 3191306.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 32, "step_time": 305.35615600366145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3834.0, "completions/mean_length": 1434.1875, "completions/mean_terminated_length": 688.8800048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.467706087976694, "epoch": 0.01870748299319728, "frac_reward_zero_std": 0.5, "grad_norm": 0.2353515625, "kl": 0.0004915133140457328, "learning_rate": 9.818594104308389e-07, "loss": 0.0415, "num_tokens": 3293022.0, "reward": 0.265625, "reward_std": 0.6664806604385376, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 33, "step_time": 294.9509489592165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3522.0, "completions/mean_length": 1028.40625, "completions/mean_terminated_length": 711.0689697265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.900582242757082, "epoch": 0.01927437641723356, "frac_reward_zero_std": 0.75, "grad_norm": 0.236328125, "kl": 0.0004916769539704546, "learning_rate": 9.812925170068026e-07, "loss": 0.0127, "num_tokens": 3370272.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 34, "step_time": 294.09036393929273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4024.0, "completions/mean_length": 1438.984375, "completions/mean_terminated_length": 1059.4107666015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.51833862811327, "epoch": 0.01984126984126984, "frac_reward_zero_std": 0.75, "grad_norm": 0.2138671875, "kl": 0.0004253339166098158, "learning_rate": 9.807256235827664e-07, "loss": 0.0187, "num_tokens": 3475223.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 35, "step_time": 302.7437364095822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3704.0, "completions/mean_length": 1525.28125, "completions/mean_terminated_length": 932.0385131835938, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 3.4461216628551483, "epoch": 0.02040816326530612, "frac_reward_zero_std": 0.875, "grad_norm": 0.142578125, "kl": 0.0004800664210051764, "learning_rate": 9.801587301587301e-07, "loss": -0.0016, "num_tokens": 3584969.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 36, "step_time": 298.9434612961486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4052.0, "completions/mean_length": 1216.53125, "completions/mean_terminated_length": 862.9122924804688, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 3.0853926092386246, "epoch": 0.020975056689342405, "frac_reward_zero_std": 0.75, "grad_norm": 0.197265625, "kl": 0.0004732625448014005, "learning_rate": 9.795918367346939e-07, "loss": 0.0305, "num_tokens": 3675523.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 37, "step_time": 303.82939208392054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3573.0, "completions/mean_length": 1365.390625, "completions/mean_terminated_length": 859.7222290039062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.5588138587772846, "epoch": 0.021541950113378686, "frac_reward_zero_std": 0.625, "grad_norm": 0.26171875, "kl": 0.0005098498504594318, "learning_rate": 9.790249433106576e-07, "loss": 0.0071, "num_tokens": 3773716.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 38, "step_time": 293.0893633402884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3852.0, "completions/mean_length": 1856.796875, "completions/mean_terminated_length": 1229.8199462890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 4.590333938598633, "epoch": 0.022108843537414966, "frac_reward_zero_std": 0.875, "grad_norm": 0.1279296875, "kl": 0.00041188216528098565, "learning_rate": 9.784580498866214e-07, "loss": 0.0116, "num_tokens": 3904815.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 39, "step_time": 296.1281976234168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4055.0, "completions/mean_length": 1368.21875, "completions/mean_terminated_length": 1137.0509033203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 5.263170003890991, "epoch": 0.022675736961451247, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "kl": 0.00041604944817663636, "learning_rate": 9.778911564625851e-07, "loss": 0.0112, "num_tokens": 4008293.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 40, "step_time": 314.7039760276675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3685.0, "completions/mean_length": 1472.890625, "completions/mean_terminated_length": 738.4199829101562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.3532409109175205, "epoch": 0.023242630385487528, "frac_reward_zero_std": 0.5, "grad_norm": 0.259765625, "kl": 0.0005240712735030684, "learning_rate": 9.773242630385487e-07, "loss": -0.0396, "num_tokens": 4114230.0, "reward": 0.21875, "reward_std": 0.5900968313217163, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 41, "step_time": 294.0841481247917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4092.0, "completions/mean_length": 1695.96875, "completions/mean_terminated_length": 1303.236328125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "entropy": 4.116798646748066, "epoch": 0.023809523809523808, "frac_reward_zero_std": 0.875, "grad_norm": 0.1650390625, "kl": 0.0003809477675531525, "learning_rate": 9.767573696145124e-07, "loss": -0.0104, "num_tokens": 4235436.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 42, "step_time": 305.93450375366956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4051.0, "completions/mean_length": 1133.921875, "completions/mean_terminated_length": 770.1578979492188, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 3.4529144503176212, "epoch": 0.024376417233560092, "frac_reward_zero_std": 0.375, "grad_norm": 0.3203125, "kl": 0.0005080552914478176, "learning_rate": 9.761904761904762e-07, "loss": 0.0488, "num_tokens": 4319519.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 43, "step_time": 294.02621566783637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3383.0, "completions/mean_length": 1556.59375, "completions/mean_terminated_length": 970.5769653320312, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 2.7311813980340958, "epoch": 0.024943310657596373, "frac_reward_zero_std": 0.75, "grad_norm": 0.1767578125, "kl": 0.00046972488871688256, "learning_rate": 9.7562358276644e-07, "loss": 0.0289, "num_tokens": 4430557.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 44, "step_time": 296.04308369010687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 4096.0, "completions/max_terminated_length": 4089.0, "completions/mean_length": 1701.234375, "completions/mean_terminated_length": 902.9791870117188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.2801453731954098, "epoch": 0.025510204081632654, "frac_reward_zero_std": 0.625, "grad_norm": 0.212890625, "kl": 0.0005189456205698662, "learning_rate": 9.750566893424036e-07, "loss": 0.0432, "num_tokens": 4552132.0, "reward": 0.203125, "reward_std": 0.6154024600982666, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 45, "step_time": 297.76370439492166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3815.0, "completions/mean_length": 1415.671875, "completions/mean_terminated_length": 919.3148193359375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.2611497938632965, "epoch": 0.026077097505668934, "frac_reward_zero_std": 0.875, "grad_norm": 0.138671875, "kl": 0.0004771515623360756, "learning_rate": 9.744897959183674e-07, "loss": 0.0514, "num_tokens": 4654695.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 46, "step_time": 297.30547140631825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3786.0, "completions/mean_length": 1615.46875, "completions/mean_terminated_length": 856.1224365234375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 2.364711243659258, "epoch": 0.026643990929705215, "frac_reward_zero_std": 0.625, "grad_norm": 0.1806640625, "kl": 0.0005641347979690181, "learning_rate": 9.739229024943311e-07, "loss": 0.0202, "num_tokens": 4772365.0, "reward": 0.1171875, "reward_std": 0.46922925114631653, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 47, "step_time": 309.7107340320945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4046.0, "completions/mean_length": 1357.828125, "completions/mean_terminated_length": 1021.5614013671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.842697562649846, "epoch": 0.027210884353741496, "frac_reward_zero_std": 0.875, "grad_norm": 0.1708984375, "kl": 0.0004335203693699441, "learning_rate": 9.733560090702947e-07, "loss": 0.0217, "num_tokens": 4869858.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 48, "step_time": 293.2417637016624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3148.0, "completions/mean_length": 1485.171875, "completions/mean_terminated_length": 1001.6851806640625, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "entropy": 4.688731737434864, "epoch": 0.027777777777777776, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "kl": 0.0004430160397532745, "learning_rate": 9.727891156462584e-07, "loss": 0.0057, "num_tokens": 4975557.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 49, "step_time": 295.9001120068133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3707.0, "completions/mean_length": 1632.328125, "completions/mean_terminated_length": 878.142822265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.1174243800342083, "epoch": 0.02834467120181406, "frac_reward_zero_std": 0.875, "grad_norm": 0.1318359375, "kl": 0.00046842659867252223, "learning_rate": 9.722222222222222e-07, "loss": -0.0031, "num_tokens": 5091690.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 50, "step_time": 294.6722645731643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4058.0, "completions/mean_length": 1200.78125, "completions/mean_terminated_length": 901.27587890625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 4.219451703131199, "epoch": 0.02891156462585034, "frac_reward_zero_std": 0.75, "grad_norm": 0.2158203125, "kl": 0.00045792363744112663, "learning_rate": 9.71655328798186e-07, "loss": 0.0414, "num_tokens": 5183484.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 51, "step_time": 341.01204745844007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3549.0, "completions/mean_length": 1386.125, "completions/mean_terminated_length": 999.0000610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.865479376167059, "epoch": 0.02947845804988662, "frac_reward_zero_std": 0.75, "grad_norm": 0.177734375, "kl": 0.00048682725173421204, "learning_rate": 9.710884353741497e-07, "loss": 0.0312, "num_tokens": 5282884.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 52, "step_time": 293.0214045289904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3679.0, "completions/mean_length": 1611.1875, "completions/mean_terminated_length": 850.5305786132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.6968845427036285, "epoch": 0.030045351473922902, "frac_reward_zero_std": 0.75, "grad_norm": 0.1611328125, "kl": 0.00048145676919375546, "learning_rate": 9.705215419501134e-07, "loss": 0.0033, "num_tokens": 5397376.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 53, "step_time": 295.37289924081415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3980.0, "completions/mean_length": 1000.234375, "completions/mean_terminated_length": 847.9835815429688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.3745584413409233, "epoch": 0.030612244897959183, "frac_reward_zero_std": 0.875, "grad_norm": 0.1259765625, "kl": 0.000502369279274717, "learning_rate": 9.69954648526077e-07, "loss": 0.003, "num_tokens": 5473719.0, "reward": 0.1875, "reward_std": 0.613990306854248, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 54, "step_time": 296.9595435075462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3687.0, "completions/mean_length": 1366.578125, "completions/mean_terminated_length": 919.9454345703125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 4.502534218132496, "epoch": 0.031179138321995464, "frac_reward_zero_std": 0.75, "grad_norm": 0.1884765625, "kl": 0.00041477979357296135, "learning_rate": 9.693877551020407e-07, "loss": 0.0287, "num_tokens": 5573684.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 55, "step_time": 296.263409120962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3797.0, "completions/mean_length": 1568.328125, "completions/mean_terminated_length": 1100.24072265625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.88920546323061, "epoch": 0.031746031746031744, "frac_reward_zero_std": 0.875, "grad_norm": 0.12109375, "kl": 0.0004977654971298762, "learning_rate": 9.688208616780045e-07, "loss": 0.0257, "num_tokens": 5687057.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 56, "step_time": 297.1691315295175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3718.0, "completions/mean_length": 1286.59375, "completions/mean_terminated_length": 941.5789794921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.255432682111859, "epoch": 0.03231292517006803, "frac_reward_zero_std": 0.875, "grad_norm": 0.16796875, "kl": 0.0004884136069449596, "learning_rate": 9.682539682539682e-07, "loss": 0.0592, "num_tokens": 5783583.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 57, "step_time": 309.3456033077091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3141.0, "completions/mean_length": 871.546875, "completions/mean_terminated_length": 598.2881469726562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.336066670715809, "epoch": 0.032879818594104306, "frac_reward_zero_std": 0.375, "grad_norm": 0.427734375, "kl": 0.0005954704392934218, "learning_rate": 9.67687074829932e-07, "loss": 0.0393, "num_tokens": 5849834.0, "reward": 0.234375, "reward_std": 0.5907269716262817, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 58, "step_time": 292.01666195411235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3611.0, "completions/mean_length": 983.6875, "completions/mean_terminated_length": 776.2000122070312, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 2.50449425727129, "epoch": 0.03344671201814059, "frac_reward_zero_std": 0.625, "grad_norm": 0.443359375, "kl": 0.0006079958839109167, "learning_rate": 9.671201814058957e-07, "loss": 0.0248, "num_tokens": 5924790.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 59, "step_time": 295.90404415596277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3673.0, "completions/mean_length": 1256.65625, "completions/mean_terminated_length": 792.0363159179688, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.7395902425050735, "epoch": 0.034013605442176874, "frac_reward_zero_std": 1.0, "grad_norm": 0.00010395050048828125, "kl": 0.0005482348942678072, "learning_rate": 9.665532879818595e-07, "loss": 0.0, "num_tokens": 6017456.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 60, "step_time": 299.53357178531587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4025.0, "completions/mean_length": 1479.203125, "completions/mean_terminated_length": 1105.375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "entropy": 3.7954771779477596, "epoch": 0.03458049886621315, "frac_reward_zero_std": 0.625, "grad_norm": 0.2421875, "kl": 0.0004666336762966239, "learning_rate": 9.65986394557823e-07, "loss": 0.0083, "num_tokens": 6124485.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 61, "step_time": 303.2139684362337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4038.0, "completions/mean_length": 1400.875, "completions/mean_terminated_length": 959.8544921875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.448891811072826, "epoch": 0.035147392290249435, "frac_reward_zero_std": 0.875, "grad_norm": 0.138671875, "kl": 0.0005504604605448549, "learning_rate": 9.654195011337868e-07, "loss": 0.0034, "num_tokens": 6224125.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 62, "step_time": 294.8489862466231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4077.0, "completions/mean_length": 1326.484375, "completions/mean_terminated_length": 930.83935546875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.4797469154000282, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.75, "grad_norm": 0.2236328125, "kl": 0.0005405677475209814, "learning_rate": 9.648526077097505e-07, "loss": 0.0177, "num_tokens": 6320292.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 63, "step_time": 294.98920669872314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3665.0, "completions/mean_length": 1381.421875, "completions/mean_terminated_length": 1048.0526123046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.080914679914713, "epoch": 0.036281179138321996, "frac_reward_zero_std": 0.75, "grad_norm": 0.28125, "kl": 0.0004902618484265986, "learning_rate": 9.642857142857142e-07, "loss": -0.0307, "num_tokens": 6422399.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 64, "step_time": 300.90544505324215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3904.0, "completions/mean_length": 1388.59375, "completions/mean_terminated_length": 887.2222290039062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.027115806937218, "epoch": 0.036848072562358274, "frac_reward_zero_std": 0.625, "grad_norm": 0.2255859375, "kl": 0.0006420780437110807, "learning_rate": 9.63718820861678e-07, "loss": 0.012, "num_tokens": 6523989.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 65, "step_time": 298.16601424105465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4046.0, "completions/mean_length": 1429.3125, "completions/mean_terminated_length": 992.9454345703125, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "entropy": 3.2920237369835377, "epoch": 0.03741496598639456, "frac_reward_zero_std": 0.875, "grad_norm": 0.142578125, "kl": 0.0005064156493972405, "learning_rate": 9.631519274376417e-07, "loss": 0.0187, "num_tokens": 6629153.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 66, "step_time": 300.80143326800317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3894.0, "completions/mean_length": 1645.015625, "completions/mean_terminated_length": 958.739990234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.213624369353056, "epoch": 0.03798185941043084, "frac_reward_zero_std": 0.5, "grad_norm": 0.2158203125, "kl": 0.0004907570573777775, "learning_rate": 9.625850340136053e-07, "loss": 0.0149, "num_tokens": 6746194.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 67, "step_time": 295.89689590316266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3982.0, "completions/mean_length": 1378.15625, "completions/mean_terminated_length": 1044.385986328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.088643457740545, "epoch": 0.03854875283446712, "frac_reward_zero_std": 0.75, "grad_norm": 0.1591796875, "kl": 0.00045124743974156445, "learning_rate": 9.62018140589569e-07, "loss": -0.0338, "num_tokens": 6845908.0, "reward": 0.15625, "reward_std": 0.5409793853759766, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 68, "step_time": 296.33633978664875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3903.0, "completions/mean_length": 1253.875, "completions/mean_terminated_length": 847.857177734375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.7947971746325493, "epoch": 0.0391156462585034, "frac_reward_zero_std": 0.625, "grad_norm": 0.2421875, "kl": 0.0005858802542206831, "learning_rate": 9.614512471655328e-07, "loss": 0.0518, "num_tokens": 6939060.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 69, "step_time": 297.3353471541777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3946.0, "completions/mean_length": 1359.484375, "completions/mean_terminated_length": 852.7222290039062, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "entropy": 3.448988128453493, "epoch": 0.03968253968253968, "frac_reward_zero_std": 0.875, "grad_norm": 0.1123046875, "kl": 0.0007160816503528622, "learning_rate": 9.608843537414965e-07, "loss": 0.0037, "num_tokens": 7037675.0, "reward": 0.0625, "reward_std": 0.32732683420181274, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 70, "step_time": 293.6382048111409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.0, "completions/mean_length": 1022.046875, "completions/mean_terminated_length": 817.11669921875, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "entropy": 3.8905701339244843, "epoch": 0.040249433106575964, "frac_reward_zero_std": 0.625, "grad_norm": 0.287109375, "kl": 0.0005872798137716018, "learning_rate": 9.603174603174603e-07, "loss": 0.0793, "num_tokens": 7114054.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 71, "step_time": 293.8014563927427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2784.0, "completions/mean_length": 938.9375, "completions/mean_terminated_length": 487.9285888671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 1.9663578495383263, "epoch": 0.04081632653061224, "frac_reward_zero_std": 0.75, "grad_norm": 0.259765625, "kl": 0.0007659744123884593, "learning_rate": 9.59750566893424e-07, "loss": 0.019, "num_tokens": 7183450.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 72, "step_time": 289.2986155534163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4064.0, "completions/mean_length": 1038.734375, "completions/mean_terminated_length": 722.4655151367188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.3653258718550205, "epoch": 0.041383219954648526, "frac_reward_zero_std": 0.625, "grad_norm": 0.306640625, "kl": 0.0006311132347036619, "learning_rate": 9.591836734693876e-07, "loss": 0.0313, "num_tokens": 7269185.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 73, "step_time": 375.2334147216752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3538.0, "completions/mean_length": 881.609375, "completions/mean_terminated_length": 667.3167114257812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.874113205820322, "epoch": 0.04195011337868481, "frac_reward_zero_std": 0.875, "grad_norm": 0.2412109375, "kl": 0.0006657410795014584, "learning_rate": 9.586167800453513e-07, "loss": 0.0335, "num_tokens": 7336264.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 74, "step_time": 290.8858574004844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3541.0, "completions/mean_length": 1720.453125, "completions/mean_terminated_length": 1114.921630859375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.012122828513384, "epoch": 0.04251700680272109, "frac_reward_zero_std": 0.75, "grad_norm": 0.1669921875, "kl": 0.0004141690405958798, "learning_rate": 9.58049886621315e-07, "loss": 0.0161, "num_tokens": 7457877.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 75, "step_time": 296.16961095202714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3963.0, "completions/mean_length": 1003.28125, "completions/mean_terminated_length": 683.3448486328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.2306234762072563, "epoch": 0.04308390022675737, "frac_reward_zero_std": 0.5, "grad_norm": 0.3125, "kl": 0.000566702869036817, "learning_rate": 9.574829931972788e-07, "loss": 0.0374, "num_tokens": 7532191.0, "reward": 0.328125, "reward_std": 0.7303985953330994, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 76, "step_time": 290.138320222497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3892.0, "completions/mean_length": 1390.09375, "completions/mean_terminated_length": 1003.5357666015625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.7910470850765705, "epoch": 0.04365079365079365, "frac_reward_zero_std": 0.875, "grad_norm": 0.173828125, "kl": 0.0005802920577480108, "learning_rate": 9.569160997732426e-07, "loss": -0.0138, "num_tokens": 7632333.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 77, "step_time": 293.7238145461306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3881.0, "completions/mean_length": 1424.296875, "completions/mean_terminated_length": 1096.1929931640625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.150249805301428, "epoch": 0.04421768707482993, "frac_reward_zero_std": 0.625, "grad_norm": 0.236328125, "kl": 0.0005237193527136697, "learning_rate": 9.563492063492063e-07, "loss": 0.0141, "num_tokens": 7736216.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 78, "step_time": 299.65833119209856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4074.0, "completions/mean_length": 1354.09375, "completions/mean_terminated_length": 962.3928833007812, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 4.0368705578148365, "epoch": 0.04478458049886622, "frac_reward_zero_std": 0.75, "grad_norm": 0.1884765625, "kl": 0.0005636081277771154, "learning_rate": 9.5578231292517e-07, "loss": 0.0239, "num_tokens": 7834734.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 79, "step_time": 296.3580552712083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3816.0, "completions/mean_length": 1199.765625, "completions/mean_terminated_length": 844.0877075195312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.758032266050577, "epoch": 0.045351473922902494, "frac_reward_zero_std": 0.75, "grad_norm": 0.2392578125, "kl": 0.000635372903161624, "learning_rate": 9.552154195011338e-07, "loss": 0.0212, "num_tokens": 7923047.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 80, "step_time": 294.42624179366976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3742.0, "completions/mean_length": 1513.125, "completions/mean_terminated_length": 1090.47265625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "entropy": 3.8117652274668217, "epoch": 0.04591836734693878, "frac_reward_zero_std": 1.0, "grad_norm": 0.00011539459228515625, "kl": 0.0005118364797453978, "learning_rate": 9.546485260770974e-07, "loss": 0.0, "num_tokens": 8030687.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 81, "step_time": 291.95394702255726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3812.0, "completions/mean_length": 1353.421875, "completions/mean_terminated_length": 784.2075805664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.888309068977833, "epoch": 0.046485260770975055, "frac_reward_zero_std": 0.875, "grad_norm": 0.13671875, "kl": 0.0005780409746876103, "learning_rate": 9.54081632653061e-07, "loss": 0.0015, "num_tokens": 8133042.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 82, "step_time": 329.26528967916965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4070.0, "completions/mean_length": 1369.96875, "completions/mean_terminated_length": 1035.1929931640625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.8123384825885296, "epoch": 0.04705215419501134, "frac_reward_zero_std": 0.625, "grad_norm": 0.328125, "kl": 0.00047865481610642746, "learning_rate": 9.53514739229025e-07, "loss": -0.1455, "num_tokens": 8231048.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 83, "step_time": 293.00953847169876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3999.0, "completions/mean_length": 1308.75, "completions/mean_terminated_length": 852.654541015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.287667583674192, "epoch": 0.047619047619047616, "frac_reward_zero_std": 0.75, "grad_norm": 0.21484375, "kl": 0.000524261982718599, "learning_rate": 9.529478458049887e-07, "loss": 0.0314, "num_tokens": 8326752.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 84, "step_time": 294.7698496468365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3809.0, "completions/mean_length": 1655.015625, "completions/mean_terminated_length": 1091.7115478515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.714879844337702, "epoch": 0.0481859410430839, "frac_reward_zero_std": 0.625, "grad_norm": 0.2197265625, "kl": 0.0005542064991459483, "learning_rate": 9.523809523809522e-07, "loss": 0.0609, "num_tokens": 8445641.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 85, "step_time": 295.54682763200253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4049.0, "completions/mean_length": 1364.390625, "completions/mean_terminated_length": 974.1607666015625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.585228767246008, "epoch": 0.048752834467120185, "frac_reward_zero_std": 0.875, "grad_norm": 0.1513671875, "kl": 0.0005338712680895696, "learning_rate": 9.51814058956916e-07, "loss": -0.0084, "num_tokens": 8545090.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 86, "step_time": 297.61800406873226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3979.0, "completions/mean_length": 1530.359375, "completions/mean_terminated_length": 1055.24072265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.526728063821793, "epoch": 0.04931972789115646, "frac_reward_zero_std": 0.75, "grad_norm": 0.22265625, "kl": 0.0005592237139353529, "learning_rate": 9.512471655328797e-07, "loss": 0.0105, "num_tokens": 8657049.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 87, "step_time": 302.05309406947345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3760.0, "completions/mean_length": 1434.265625, "completions/mean_terminated_length": 1054.0179443359375, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "entropy": 4.485512547194958, "epoch": 0.049886621315192746, "frac_reward_zero_std": 0.75, "grad_norm": 0.2197265625, "kl": 0.000496680951982853, "learning_rate": 9.506802721088435e-07, "loss": 0.036, "num_tokens": 8769642.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 88, "step_time": 371.294574201107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.328125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3597.0, "completions/mean_length": 1918.34375, "completions/mean_terminated_length": 854.8372192382812, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "entropy": 2.781709646806121, "epoch": 0.05045351473922902, "frac_reward_zero_std": 0.875, "grad_norm": 0.1103515625, "kl": 0.000628817981123575, "learning_rate": 9.501133786848072e-07, "loss": 0.0114, "num_tokens": 8904432.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 89, "step_time": 301.35453115589917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3035.0, "completions/mean_length": 1303.71875, "completions/mean_terminated_length": 846.7999877929688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.7423397190868855, "epoch": 0.05102040816326531, "frac_reward_zero_std": 0.625, "grad_norm": 0.251953125, "kl": 0.0005851384448760655, "learning_rate": 9.49546485260771e-07, "loss": 0.0313, "num_tokens": 9000566.0, "reward": 0.2421875, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 90, "step_time": 295.8532868111506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3730.0, "completions/mean_length": 1143.890625, "completions/mean_terminated_length": 781.3508911132812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.400463607162237, "epoch": 0.051587301587301584, "frac_reward_zero_std": 0.875, "grad_norm": 0.1884765625, "kl": 0.0006745926693838555, "learning_rate": 9.489795918367347e-07, "loss": 0.0016, "num_tokens": 9087631.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 91, "step_time": 310.4487501978874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3783.0, "completions/mean_length": 1199.046875, "completions/mean_terminated_length": 953.5423583984375, "completions/min_length": 61.0, "completions/min_terminated_length": 61.0, "entropy": 3.66338399797678, "epoch": 0.05215419501133787, "frac_reward_zero_std": 0.625, "grad_norm": 0.265625, "kl": 0.0005909999781579245, "learning_rate": 9.484126984126984e-07, "loss": 0.0785, "num_tokens": 9175770.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 92, "step_time": 296.0928402291611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4031.0, "completions/mean_length": 1609.140625, "completions/mean_terminated_length": 1148.611083984375, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "entropy": 5.085915923118591, "epoch": 0.05272108843537415, "frac_reward_zero_std": 0.875, "grad_norm": 0.1484375, "kl": 0.00045808015329384943, "learning_rate": 9.47845804988662e-07, "loss": 0.0198, "num_tokens": 9291971.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 93, "step_time": 305.8303446471691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3274.0, "completions/mean_length": 1377.671875, "completions/mean_terminated_length": 750.3654174804688, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 3.596812978386879, "epoch": 0.05328798185941043, "frac_reward_zero_std": 0.625, "grad_norm": 0.2294921875, "kl": 0.0006529647143906914, "learning_rate": 9.472789115646258e-07, "loss": -0.0227, "num_tokens": 9392806.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 94, "step_time": 298.4613697146997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2999.0, "completions/mean_length": 987.5, "completions/mean_terminated_length": 780.2667236328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.4541202299296856, "epoch": 0.053854875283446714, "frac_reward_zero_std": 0.75, "grad_norm": 0.298828125, "kl": 0.0006040437892806949, "learning_rate": 9.467120181405895e-07, "loss": -0.0633, "num_tokens": 9467174.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 95, "step_time": 294.7388898720965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4009.0, "completions/mean_length": 1225.59375, "completions/mean_terminated_length": 928.6551513671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.783916585147381, "epoch": 0.05442176870748299, "frac_reward_zero_std": 0.75, "grad_norm": 0.197265625, "kl": 0.000592722135479562, "learning_rate": 9.461451247165533e-07, "loss": 0.0735, "num_tokens": 9557764.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 96, "step_time": 302.38117443956435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4062.0, "completions/mean_length": 1765.25, "completions/mean_terminated_length": 1227.3846435546875, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "entropy": 4.494186479598284, "epoch": 0.054988662131519275, "frac_reward_zero_std": 0.875, "grad_norm": 0.10888671875, "kl": 0.00047747676944709383, "learning_rate": 9.45578231292517e-07, "loss": 0.0183, "num_tokens": 9683260.0, "reward": 0.0390625, "reward_std": 0.3125, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 97, "step_time": 299.90580482501537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3940.0, "completions/mean_length": 1477.703125, "completions/mean_terminated_length": 1049.2545166015625, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "entropy": 3.7976103238761425, "epoch": 0.05555555555555555, "frac_reward_zero_std": 0.75, "grad_norm": 0.244140625, "kl": 0.0005473416022141464, "learning_rate": 9.450113378684807e-07, "loss": -0.0045, "num_tokens": 9790001.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 98, "step_time": 294.92261796910316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3028.0, "completions/mean_length": 1345.3125, "completions/mean_terminated_length": 835.9259033203125, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.06356118991971, "epoch": 0.05612244897959184, "frac_reward_zero_std": 0.625, "grad_norm": 0.2099609375, "kl": 0.0006883190781081794, "learning_rate": 9.444444444444444e-07, "loss": 0.0357, "num_tokens": 9890381.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 99, "step_time": 321.9701617071405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 4096.0, "completions/max_terminated_length": 4009.0, "completions/mean_length": 1932.703125, "completions/mean_terminated_length": 1211.604248046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.165743313729763, "epoch": 0.05668934240362812, "frac_reward_zero_std": 0.75, "grad_norm": 0.15234375, "kl": 0.0005034771747887135, "learning_rate": 9.438775510204082e-07, "loss": 0.0204, "num_tokens": 10026154.0, "reward": 0.078125, "reward_std": 0.4005824029445648, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 100, "step_time": 297.60639240499586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3944.0, "completions/mean_length": 1256.65625, "completions/mean_terminated_length": 792.0363159179688, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "entropy": 3.866720013320446, "epoch": 0.0572562358276644, "frac_reward_zero_std": 0.875, "grad_norm": 0.15234375, "kl": 0.0006163184216347872, "learning_rate": 9.433106575963718e-07, "loss": 0.0163, "num_tokens": 10117540.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 101, "step_time": 294.1511483108625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4069.0, "completions/mean_length": 938.515625, "completions/mean_terminated_length": 836.6612548828125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.3969429675489664, "epoch": 0.05782312925170068, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "kl": 0.0005263586681394372, "learning_rate": 9.427437641723356e-07, "loss": -0.0032, "num_tokens": 10188845.0, "reward": 0.0859375, "reward_std": 0.3398725986480713, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 102, "step_time": 293.2026412822306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3905.0, "completions/mean_length": 1319.109375, "completions/mean_terminated_length": 804.870361328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.2274572625756264, "epoch": 0.05839002267573696, "frac_reward_zero_std": 0.875, "grad_norm": 0.1220703125, "kl": 0.0006206169309734832, "learning_rate": 9.421768707482993e-07, "loss": -0.0151, "num_tokens": 10284436.0, "reward": 0.0546875, "reward_std": 0.32265254855155945, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 103, "step_time": 294.4274990260601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4060.0, "completions/mean_length": 1213.171875, "completions/mean_terminated_length": 914.9482421875, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "entropy": 3.7609485015273094, "epoch": 0.05895691609977324, "frac_reward_zero_std": 0.625, "grad_norm": 0.2119140625, "kl": 0.0006420473982871044, "learning_rate": 9.41609977324263e-07, "loss": 0.0083, "num_tokens": 10375711.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 104, "step_time": 301.6561495643109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3825.0, "completions/mean_length": 1011.78125, "completions/mean_terminated_length": 692.72412109375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.2102625109255314, "epoch": 0.05952380952380952, "frac_reward_zero_std": 0.5, "grad_norm": 0.35546875, "kl": 0.0006728825683239847, "learning_rate": 9.410430839002267e-07, "loss": 0.0437, "num_tokens": 10451249.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 105, "step_time": 293.18639434315264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3894.0, "completions/mean_length": 1096.640625, "completions/mean_terminated_length": 842.4576416015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.8951598592102528, "epoch": 0.060090702947845805, "frac_reward_zero_std": 0.75, "grad_norm": 0.23828125, "kl": 0.0006520739998450154, "learning_rate": 9.404761904761904e-07, "loss": 0.0211, "num_tokens": 10532410.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 106, "step_time": 295.260443424806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3860.0, "completions/mean_length": 1784.734375, "completions/mean_terminated_length": 1195.5882568359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 4.265113152563572, "epoch": 0.06065759637188209, "frac_reward_zero_std": 0.75, "grad_norm": 0.1552734375, "kl": 0.00046251650746853556, "learning_rate": 9.399092970521542e-07, "loss": -0.0044, "num_tokens": 10659921.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 107, "step_time": 303.3586766477674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3901.0, "completions/mean_length": 1430.84375, "completions/mean_terminated_length": 1050.107177734375, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "entropy": 3.967142190784216, "epoch": 0.061224489795918366, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "kl": 0.0005185274985706201, "learning_rate": 9.393424036281179e-07, "loss": 0.0245, "num_tokens": 10762111.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 108, "step_time": 294.1047155074775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3891.0, "completions/mean_length": 1249.171875, "completions/mean_terminated_length": 899.5614013671875, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "entropy": 3.897375989705324, "epoch": 0.06179138321995465, "frac_reward_zero_std": 0.75, "grad_norm": 0.2060546875, "kl": 0.0006039557038093335, "learning_rate": 9.387755102040816e-07, "loss": 0.0343, "num_tokens": 10852498.0, "reward": 0.1484375, "reward_std": 0.5820031762123108, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 109, "step_time": 294.5808366378769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3963.0, "completions/mean_length": 1259.546875, "completions/mean_terminated_length": 795.3999633789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.542051460593939, "epoch": 0.06235827664399093, "frac_reward_zero_std": 0.875, "grad_norm": 0.2353515625, "kl": 0.0007039975898806006, "learning_rate": 9.382086167800452e-07, "loss": 0.0688, "num_tokens": 10944333.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 110, "step_time": 293.68635532446206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4064.0, "completions/mean_length": 1230.65625, "completions/mean_terminated_length": 878.7719116210938, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "entropy": 3.576985750347376, "epoch": 0.06292517006802721, "frac_reward_zero_std": 0.875, "grad_norm": 0.1572265625, "kl": 0.0005771926189481746, "learning_rate": 9.37641723356009e-07, "loss": 0.0235, "num_tokens": 11036399.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 111, "step_time": 306.3100119102746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3691.0, "completions/mean_length": 1242.515625, "completions/mean_terminated_length": 650.2830200195312, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.644986890256405, "epoch": 0.06349206349206349, "frac_reward_zero_std": 0.875, "grad_norm": 0.140625, "kl": 0.0007677831636101473, "learning_rate": 9.370748299319727e-07, "loss": 0.0263, "num_tokens": 11127544.0, "reward": 0.140625, "reward_std": 0.5379911661148071, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 112, "step_time": 296.9050910929218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4023.0, "completions/mean_length": 1472.0, "completions/mean_terminated_length": 986.0740966796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 4.75571757927537, "epoch": 0.06405895691609978, "frac_reward_zero_std": 0.875, "grad_norm": 0.10009765625, "kl": 0.0005265878962745774, "learning_rate": 9.365079365079365e-07, "loss": 0.0034, "num_tokens": 11234496.0, "reward": 0.1484375, "reward_std": 0.5395448207855225, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 113, "step_time": 312.21911957859993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3857.0, "completions/mean_length": 1244.5, "completions/mean_terminated_length": 716.4444580078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.6412566378712654, "epoch": 0.06462585034013606, "frac_reward_zero_std": 0.75, "grad_norm": 0.150390625, "kl": 0.0007886644025347778, "learning_rate": 9.359410430839002e-07, "loss": -0.0186, "num_tokens": 11327856.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 114, "step_time": 313.32709429971874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3941.0, "completions/mean_length": 1519.6875, "completions/mean_terminated_length": 798.3200073242188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.032174177467823, "epoch": 0.06519274376417233, "frac_reward_zero_std": 0.5, "grad_norm": 0.263671875, "kl": 0.0008032457553781569, "learning_rate": 9.35374149659864e-07, "loss": 0.0317, "num_tokens": 11436668.0, "reward": 0.25, "reward_std": 0.6299408078193665, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 115, "step_time": 296.70230418629944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3928.0, "completions/mean_length": 818.59375, "completions/mean_terminated_length": 657.4097900390625, "completions/min_length": 93.0, "completions/min_terminated_length": 93.0, "entropy": 2.9340442791581154, "epoch": 0.06575963718820861, "frac_reward_zero_std": 0.75, "grad_norm": 0.3359375, "kl": 0.000689146967488341, "learning_rate": 9.348072562358276e-07, "loss": 0.0312, "num_tokens": 11506282.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 116, "step_time": 338.6807353729382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3776.0, "completions/mean_length": 1385.96875, "completions/mean_terminated_length": 998.8214721679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.467217389494181, "epoch": 0.0663265306122449, "frac_reward_zero_std": 0.875, "grad_norm": 0.166015625, "kl": 0.00047883675961202243, "learning_rate": 9.342403628117913e-07, "loss": 0.0138, "num_tokens": 11609216.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 117, "step_time": 313.09567215200514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3807.0, "completions/mean_length": 826.984375, "completions/mean_terminated_length": 609.050048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.981639474630356, "epoch": 0.06689342403628118, "frac_reward_zero_std": 0.75, "grad_norm": 0.20703125, "kl": 0.0006725377279508393, "learning_rate": 9.33673469387755e-07, "loss": -0.0067, "num_tokens": 11675175.0, "reward": 0.2109375, "reward_std": 0.653468668460846, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 118, "step_time": 309.6762728523463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3890.0, "completions/mean_length": 1198.265625, "completions/mean_terminated_length": 842.4035034179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.0181843265891075, "epoch": 0.06746031746031746, "frac_reward_zero_std": 0.75, "grad_norm": 0.2177734375, "kl": 0.0006909497224114602, "learning_rate": 9.331065759637188e-07, "loss": 0.0041, "num_tokens": 11761832.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 119, "step_time": 291.38728551380336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4093.0, "completions/mean_length": 1715.453125, "completions/mean_terminated_length": 986.7142944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1485245525836945, "epoch": 0.06802721088435375, "frac_reward_zero_std": 0.75, "grad_norm": 0.1513671875, "kl": 0.0005977986420475645, "learning_rate": 9.325396825396825e-07, "loss": 0.0032, "num_tokens": 11883405.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 120, "step_time": 296.01865003444254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3722.0, "completions/mean_length": 1502.234375, "completions/mean_terminated_length": 841.0784912109375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.063643492758274, "epoch": 0.06859410430839002, "frac_reward_zero_std": 0.75, "grad_norm": 0.1640625, "kl": 0.0007376213197858306, "learning_rate": 9.319727891156463e-07, "loss": 0.0133, "num_tokens": 11992156.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 121, "step_time": 302.0782301174477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3696.0, "completions/mean_length": 1317.6875, "completions/mean_terminated_length": 863.0545043945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.4850979894399643, "epoch": 0.0691609977324263, "frac_reward_zero_std": 0.75, "grad_norm": 0.244140625, "kl": 0.0006848964658274781, "learning_rate": 9.3140589569161e-07, "loss": 0.0451, "num_tokens": 12088696.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 122, "step_time": 300.00162667408586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3851.0, "completions/mean_length": 1209.234375, "completions/mean_terminated_length": 796.8392944335938, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.3270882070064545, "epoch": 0.06972789115646258, "frac_reward_zero_std": 0.5, "grad_norm": 0.263671875, "kl": 0.0006902841369083035, "learning_rate": 9.308390022675737e-07, "loss": 0.0542, "num_tokens": 12180911.0, "reward": 0.21875, "reward_std": 0.5900968313217163, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 123, "step_time": 306.8877000166103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4032.0, "completions/mean_length": 1511.90625, "completions/mean_terminated_length": 1033.370361328125, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 5.586065195500851, "epoch": 0.07029478458049887, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "kl": 0.0005649285230902024, "learning_rate": 9.302721088435374e-07, "loss": 0.061, "num_tokens": 12288641.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 124, "step_time": 296.0335800014436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4025.0, "completions/mean_length": 1398.34375, "completions/mean_terminated_length": 1012.96435546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.474613543599844, "epoch": 0.07086167800453515, "frac_reward_zero_std": 1.0, "grad_norm": 0.00015354156494140625, "kl": 0.0006384559055732097, "learning_rate": 9.29705215419501e-07, "loss": 0.0, "num_tokens": 12391743.0, "reward": 0.0625, "reward_std": 0.1666666716337204, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 125, "step_time": 300.94666452147067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4047.0, "completions/mean_length": 1286.171875, "completions/mean_terminated_length": 884.7678833007812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.363495498895645, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.625, "grad_norm": 0.296875, "kl": 0.0007903738214736222, "learning_rate": 9.291383219954648e-07, "loss": -0.0225, "num_tokens": 12486010.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 126, "step_time": 298.92847953084856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3481.0, "completions/mean_length": 1562.15625, "completions/mean_terminated_length": 977.423095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.3883329667150974, "epoch": 0.07199546485260772, "frac_reward_zero_std": 0.75, "grad_norm": 0.173828125, "kl": 0.0006784117667848477, "learning_rate": 9.285714285714285e-07, "loss": 0.0166, "num_tokens": 12598372.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 127, "step_time": 296.65200543869287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2759.0, "completions/mean_length": 989.515625, "completions/mean_terminated_length": 668.1551513671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.646180473268032, "epoch": 0.07256235827664399, "frac_reward_zero_std": 0.75, "grad_norm": 0.2431640625, "kl": 0.0008213972214434762, "learning_rate": 9.280045351473923e-07, "loss": 0.0103, "num_tokens": 12673509.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 128, "step_time": 295.8863944094628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4078.0, "completions/mean_length": 1332.421875, "completions/mean_terminated_length": 820.6481323242188, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "entropy": 2.9267875254154205, "epoch": 0.07312925170068027, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "kl": 0.000708395169567666, "learning_rate": 9.274376417233559e-07, "loss": 0.0156, "num_tokens": 12769648.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 129, "step_time": 294.252164516598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3993.0, "completions/mean_length": 1347.3125, "completions/mean_terminated_length": 1062.9654541015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.332023099064827, "epoch": 0.07369614512471655, "frac_reward_zero_std": 0.75, "grad_norm": 0.1875, "kl": 0.000581516707825358, "learning_rate": 9.268707482993197e-07, "loss": 0.0189, "num_tokens": 12867388.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 130, "step_time": 296.34269182849675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2999.0, "completions/mean_length": 1607.640625, "completions/mean_terminated_length": 845.89794921875, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "entropy": 4.000800345093012, "epoch": 0.07426303854875284, "frac_reward_zero_std": 0.75, "grad_norm": 0.166015625, "kl": 0.0009954783781722654, "learning_rate": 9.263038548752834e-07, "loss": 0.0278, "num_tokens": 12982205.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 131, "step_time": 296.1071062963456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3894.0, "completions/mean_length": 1362.4375, "completions/mean_terminated_length": 856.2222290039062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.674266640096903, "epoch": 0.07482993197278912, "frac_reward_zero_std": 0.875, "grad_norm": 0.12451171875, "kl": 0.0006412754437405965, "learning_rate": 9.257369614512472e-07, "loss": -0.0001, "num_tokens": 13080969.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 132, "step_time": 303.2742968611419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3891.0, "completions/mean_length": 1077.09375, "completions/mean_terminated_length": 764.7930908203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4341964684426785, "epoch": 0.07539682539682539, "frac_reward_zero_std": 0.5, "grad_norm": 0.478515625, "kl": 0.0007655090921616647, "learning_rate": 9.251700680272108e-07, "loss": -0.0576, "num_tokens": 13161559.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 133, "step_time": 296.6053422614932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4039.0, "completions/mean_length": 1340.3125, "completions/mean_terminated_length": 946.6428833007812, "completions/min_length": 107.0, "completions/min_terminated_length": 107.0, "entropy": 3.7123965956270695, "epoch": 0.07596371882086168, "frac_reward_zero_std": 0.625, "grad_norm": 0.2177734375, "kl": 0.0006033002264302922, "learning_rate": 9.246031746031746e-07, "loss": 0.0028, "num_tokens": 13259563.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 134, "step_time": 297.93012297712266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3623.0, "completions/mean_length": 1054.921875, "completions/mean_terminated_length": 620.482177734375, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 2.5744658894836903, "epoch": 0.07653061224489796, "frac_reward_zero_std": 0.5, "grad_norm": 0.287109375, "kl": 0.0008069383547990583, "learning_rate": 9.240362811791382e-07, "loss": 0.0362, "num_tokens": 13337550.0, "reward": 0.2734375, "reward_std": 0.6294977068901062, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 135, "step_time": 292.86695772688836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4013.0, "completions/mean_length": 1441.484375, "completions/mean_terminated_length": 1115.4912109375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.732803553342819, "epoch": 0.07709750566893424, "frac_reward_zero_std": 0.625, "grad_norm": 0.255859375, "kl": 0.00047232820179488044, "learning_rate": 9.23469387755102e-07, "loss": -0.0178, "num_tokens": 13440885.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 136, "step_time": 293.1682682046667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2807.0, "completions/mean_length": 700.015625, "completions/mean_terminated_length": 533.0, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.496282309293747, "epoch": 0.07766439909297052, "frac_reward_zero_std": 0.75, "grad_norm": 0.302734375, "kl": 0.0009696580927993637, "learning_rate": 9.229024943310657e-07, "loss": -0.002, "num_tokens": 13497606.0, "reward": 0.265625, "reward_std": 0.7451896071434021, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 137, "step_time": 294.9174865875393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3858.0, "completions/mean_length": 1188.640625, "completions/mean_terminated_length": 994.8167114257812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.9258103258907795, "epoch": 0.0782312925170068, "frac_reward_zero_std": 0.5, "grad_norm": 0.341796875, "kl": 0.0006540652311741724, "learning_rate": 9.223356009070295e-07, "loss": 0.0428, "num_tokens": 13588127.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 138, "step_time": 316.7908872719854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3090.0, "completions/mean_length": 1407.359375, "completions/mean_terminated_length": 967.3999633789062, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "entropy": 3.2234241291880608, "epoch": 0.07879818594104308, "frac_reward_zero_std": 0.625, "grad_norm": 0.25, "kl": 0.0007198409548436757, "learning_rate": 9.217687074829932e-07, "loss": 0.0617, "num_tokens": 13691222.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 139, "step_time": 297.9994641896337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3727.0, "completions/mean_length": 1078.921875, "completions/mean_terminated_length": 823.2373046875, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.5499210990965366, "epoch": 0.07936507936507936, "frac_reward_zero_std": 0.375, "grad_norm": 0.37109375, "kl": 0.00083128790538467, "learning_rate": 9.21201814058957e-07, "loss": 0.0825, "num_tokens": 13770089.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 140, "step_time": 290.72720505669713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4028.0, "completions/mean_length": 1039.96875, "completions/mean_terminated_length": 723.8275756835938, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "entropy": 3.0028465539216995, "epoch": 0.07993197278911565, "frac_reward_zero_std": 0.625, "grad_norm": 0.34375, "kl": 0.0008300166300614364, "learning_rate": 9.206349206349205e-07, "loss": -0.0147, "num_tokens": 13847463.0, "reward": 0.1953125, "reward_std": 0.5747131109237671, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 141, "step_time": 292.8372485693544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4023.0, "completions/mean_length": 886.84375, "completions/mean_terminated_length": 672.9000244140625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.4044470489025116, "epoch": 0.08049886621315193, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "kl": 0.0011324633942422224, "learning_rate": 9.200680272108843e-07, "loss": 0.0173, "num_tokens": 13915485.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 142, "step_time": 294.4719723528251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4092.0, "completions/mean_length": 1612.421875, "completions/mean_terminated_length": 1152.5, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "entropy": 4.717786908149719, "epoch": 0.0810657596371882, "frac_reward_zero_std": 0.75, "grad_norm": 0.2158203125, "kl": 0.0006212099324329756, "learning_rate": 9.19501133786848e-07, "loss": -0.0116, "num_tokens": 14030184.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 143, "step_time": 296.9507727250457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3716.0, "completions/mean_length": 1006.296875, "completions/mean_terminated_length": 744.4576416015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.8041425682604313, "epoch": 0.08163265306122448, "frac_reward_zero_std": 0.75, "grad_norm": 0.236328125, "kl": 0.0008463433550787158, "learning_rate": 9.189342403628118e-07, "loss": 0.0045, "num_tokens": 14106299.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 144, "step_time": 294.0263333497569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3609.0, "completions/mean_length": 1302.328125, "completions/mean_terminated_length": 722.5094604492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.2059508860111237, "epoch": 0.08219954648526077, "frac_reward_zero_std": 0.75, "grad_norm": 0.18359375, "kl": 0.0009646604430599837, "learning_rate": 9.183673469387755e-07, "loss": -0.0243, "num_tokens": 14202584.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 145, "step_time": 301.0547637110576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3827.0, "completions/mean_length": 938.734375, "completions/mean_terminated_length": 728.2500610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.7331823147833347, "epoch": 0.08276643990929705, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "kl": 0.000867834141899948, "learning_rate": 9.178004535147393e-07, "loss": 0.0108, "num_tokens": 14277135.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 146, "step_time": 311.6211075047031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.0, "completions/mean_length": 1596.484375, "completions/mean_terminated_length": 1077.717041015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.653958112001419, "epoch": 0.08333333333333333, "frac_reward_zero_std": 0.875, "grad_norm": 0.1552734375, "kl": 0.00059996831441822, "learning_rate": 9.172335600907029e-07, "loss": -0.0023, "num_tokens": 14390822.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 147, "step_time": 294.62713649868965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3383.0, "completions/mean_length": 1330.984375, "completions/mean_terminated_length": 935.982177734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.549279760569334, "epoch": 0.08390022675736962, "frac_reward_zero_std": 0.75, "grad_norm": 0.236328125, "kl": 0.0007655783811060246, "learning_rate": 9.166666666666665e-07, "loss": 0.0043, "num_tokens": 14487165.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 148, "step_time": 292.3098398465663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3942.0, "completions/mean_length": 1446.515625, "completions/mean_terminated_length": 896.6226806640625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.011128805577755, "epoch": 0.0844671201814059, "frac_reward_zero_std": 0.75, "grad_norm": 0.1669921875, "kl": 0.0007254601096065016, "learning_rate": 9.160997732426303e-07, "loss": 0.0292, "num_tokens": 14591454.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 149, "step_time": 295.3695561401546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3775.0, "completions/mean_length": 1427.84375, "completions/mean_terminated_length": 991.236328125, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.2528973296284676, "epoch": 0.08503401360544217, "frac_reward_zero_std": 0.75, "grad_norm": 0.1787109375, "kl": 0.0006524867385451216, "learning_rate": 9.15532879818594e-07, "loss": 0.0224, "num_tokens": 14694284.0, "reward": 0.125, "reward_std": 0.48795002698898315, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 150, "step_time": 295.72957531083375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3824.0, "completions/mean_length": 1812.015625, "completions/mean_terminated_length": 1229.8236083984375, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 4.662757508456707, "epoch": 0.08560090702947845, "frac_reward_zero_std": 0.875, "grad_norm": 0.107421875, "kl": 0.0005463318366309977, "learning_rate": 9.149659863945578e-07, "loss": -0.006, "num_tokens": 14821861.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 151, "step_time": 295.2445678981021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3852.0, "completions/mean_length": 1448.96875, "completions/mean_terminated_length": 707.7999877929688, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 2.8441838547587395, "epoch": 0.08616780045351474, "frac_reward_zero_std": 0.875, "grad_norm": 0.1318359375, "kl": 0.0010018190896516899, "learning_rate": 9.143990929705215e-07, "loss": -0.0178, "num_tokens": 14926715.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 152, "step_time": 296.1488225897774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4022.0, "completions/mean_length": 1175.703125, "completions/mean_terminated_length": 1032.0819091796875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.440500654280186, "epoch": 0.08673469387755102, "frac_reward_zero_std": 0.625, "grad_norm": 0.26171875, "kl": 0.0006847782442491734, "learning_rate": 9.138321995464853e-07, "loss": 0.0343, "num_tokens": 15013960.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 153, "step_time": 294.9221243271604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4052.0, "completions/mean_length": 1557.359375, "completions/mean_terminated_length": 1087.24072265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.584035370498896, "epoch": 0.0873015873015873, "frac_reward_zero_std": 0.875, "grad_norm": 0.1259765625, "kl": 0.0006917003101989394, "learning_rate": 9.132653061224489e-07, "loss": 0.0154, "num_tokens": 15124903.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 154, "step_time": 296.17265504971147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3368.0, "completions/mean_length": 1180.875, "completions/mean_terminated_length": 703.8545532226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.2370591796934605, "epoch": 0.08786848072562359, "frac_reward_zero_std": 0.625, "grad_norm": 0.2451171875, "kl": 0.0010188462001679, "learning_rate": 9.126984126984127e-07, "loss": 0.0303, "num_tokens": 15216191.0, "reward": 0.1328125, "reward_std": 0.4734388589859009, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 155, "step_time": 313.3282301053405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3849.0, "completions/mean_length": 1148.734375, "completions/mean_terminated_length": 843.8448486328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.053316328674555, "epoch": 0.08843537414965986, "frac_reward_zero_std": 0.5, "grad_norm": 0.328125, "kl": 0.0006997470491114655, "learning_rate": 9.121315192743763e-07, "loss": -0.0515, "num_tokens": 15301862.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 156, "step_time": 299.08300923835486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3414.0, "completions/mean_length": 1203.828125, "completions/mean_terminated_length": 904.637939453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.421551700681448, "epoch": 0.08900226757369614, "frac_reward_zero_std": 0.375, "grad_norm": 0.373046875, "kl": 0.0007786054011376109, "learning_rate": 9.115646258503401e-07, "loss": 0.0071, "num_tokens": 15390323.0, "reward": 0.2265625, "reward_std": 0.6294977068901062, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 157, "step_time": 294.19915232714266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3767.0, "completions/mean_length": 969.390625, "completions/mean_terminated_length": 704.4237060546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.2527791261672974, "epoch": 0.08956916099773243, "frac_reward_zero_std": 0.625, "grad_norm": 0.35546875, "kl": 0.0009187185551127186, "learning_rate": 9.109977324263038e-07, "loss": 0.0473, "num_tokens": 15462524.0, "reward": 0.25, "reward_std": 0.7014724016189575, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 158, "step_time": 290.991029872559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3310.0, "completions/mean_length": 1289.6875, "completions/mean_terminated_length": 770.0, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.842462159693241, "epoch": 0.09013605442176871, "frac_reward_zero_std": 0.625, "grad_norm": 0.296875, "kl": 0.0010018895627581514, "learning_rate": 9.104308390022676e-07, "loss": -0.0134, "num_tokens": 15557616.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 159, "step_time": 298.8811679929495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4079.0, "completions/mean_length": 1188.234375, "completions/mean_terminated_length": 994.3833618164062, "completions/min_length": 32.0, "completions/min_terminated_length": 32.0, "entropy": 4.564862992614508, "epoch": 0.09070294784580499, "frac_reward_zero_std": 0.75, "grad_norm": 0.2138671875, "kl": 0.0007492936820199247, "learning_rate": 9.098639455782312e-07, "loss": 0.0066, "num_tokens": 15644159.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 160, "step_time": 291.18783896882087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3548.0, "completions/mean_length": 1427.71875, "completions/mean_terminated_length": 933.5925903320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1501113399863243, "epoch": 0.09126984126984126, "frac_reward_zero_std": 0.875, "grad_norm": 0.109375, "kl": 0.0007935078247101046, "learning_rate": 9.09297052154195e-07, "loss": 0.0065, "num_tokens": 15748317.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 161, "step_time": 298.3245401792228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3652.0, "completions/mean_length": 1111.59375, "completions/mean_terminated_length": 858.677978515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.9220812749117613, "epoch": 0.09183673469387756, "frac_reward_zero_std": 0.625, "grad_norm": 0.333984375, "kl": 0.0007548401827079942, "learning_rate": 9.087301587301587e-07, "loss": 0.0032, "num_tokens": 15830043.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 162, "step_time": 292.2724645556882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4003.0, "completions/mean_length": 1288.25, "completions/mean_terminated_length": 1050.3050537109375, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "entropy": 3.6803053356707096, "epoch": 0.09240362811791383, "frac_reward_zero_std": 0.875, "grad_norm": 0.1591796875, "kl": 0.001038718895870261, "learning_rate": 9.081632653061225e-07, "loss": 0.0187, "num_tokens": 15924699.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 163, "step_time": 295.85355997178704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3796.0, "completions/mean_length": 1493.40625, "completions/mean_terminated_length": 1121.607177734375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.4482744075357914, "epoch": 0.09297052154195011, "frac_reward_zero_std": 0.75, "grad_norm": 0.296875, "kl": 0.0007815584158379352, "learning_rate": 9.075963718820861e-07, "loss": -0.0798, "num_tokens": 16033693.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 164, "step_time": 298.8449371922761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4059.0, "completions/mean_length": 1206.765625, "completions/mean_terminated_length": 907.8793334960938, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 4.567217536270618, "epoch": 0.0935374149659864, "frac_reward_zero_std": 0.75, "grad_norm": 0.2578125, "kl": 0.0009052703826455399, "learning_rate": 9.070294784580499e-07, "loss": -0.0011, "num_tokens": 16121230.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 165, "step_time": 292.9822875568643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3983.0, "completions/mean_length": 1307.03125, "completions/mean_terminated_length": 850.654541015625, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 3.0286555476486683, "epoch": 0.09410430839002268, "frac_reward_zero_std": 0.875, "grad_norm": 0.1689453125, "kl": 0.001030411476676818, "learning_rate": 9.064625850340135e-07, "loss": 0.0178, "num_tokens": 16217136.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 166, "step_time": 304.10015850979835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3961.0, "completions/mean_length": 1006.484375, "completions/mean_terminated_length": 686.8793334960938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.243671216070652, "epoch": 0.09467120181405896, "frac_reward_zero_std": 0.75, "grad_norm": 0.2109375, "kl": 0.0010857144989131484, "learning_rate": 9.058956916099772e-07, "loss": 0.0292, "num_tokens": 16292823.0, "reward": 0.203125, "reward_std": 0.6529466509819031, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 167, "step_time": 297.90582262165844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3689.0, "completions/mean_length": 1202.625, "completions/mean_terminated_length": 729.1636352539062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.780062321573496, "epoch": 0.09523809523809523, "frac_reward_zero_std": 0.75, "grad_norm": 0.166015625, "kl": 0.0010662161021173233, "learning_rate": 9.05328798185941e-07, "loss": 0.0149, "num_tokens": 16382351.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 168, "step_time": 298.7541268877685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3988.0, "completions/mean_length": 1284.109375, "completions/mean_terminated_length": 1096.6500244140625, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 4.45481001585722, "epoch": 0.09580498866213152, "frac_reward_zero_std": 0.875, "grad_norm": 0.1748046875, "kl": 0.0007455445957020856, "learning_rate": 9.047619047619047e-07, "loss": 0.0211, "num_tokens": 16477158.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 169, "step_time": 296.70281307492405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4081.0, "completions/mean_length": 1465.640625, "completions/mean_terminated_length": 919.7169799804688, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 3.220314707607031, "epoch": 0.0963718820861678, "frac_reward_zero_std": 0.875, "grad_norm": 0.09765625, "kl": 0.0007882528479967732, "learning_rate": 9.041950113378685e-07, "loss": -0.0049, "num_tokens": 16585927.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 170, "step_time": 312.9412234034389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3918.0, "completions/mean_length": 1084.859375, "completions/mean_terminated_length": 829.677978515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.584567127749324, "epoch": 0.09693877551020408, "frac_reward_zero_std": 0.625, "grad_norm": 0.365234375, "kl": 0.001218626624904573, "learning_rate": 9.036281179138322e-07, "loss": 0.0048, "num_tokens": 16665574.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 171, "step_time": 291.5702755851671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3636.0, "completions/mean_length": 1258.171875, "completions/mean_terminated_length": 909.6666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.920179232954979, "epoch": 0.09750566893424037, "frac_reward_zero_std": 0.75, "grad_norm": 0.1728515625, "kl": 0.0009037836898642126, "learning_rate": 9.030612244897958e-07, "loss": -0.0098, "num_tokens": 16757289.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 172, "step_time": 296.3245643340051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4070.0, "completions/mean_length": 1590.71875, "completions/mean_terminated_length": 1070.7547607421875, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "entropy": 4.072601463645697, "epoch": 0.09807256235827665, "frac_reward_zero_std": 0.625, "grad_norm": 0.2392578125, "kl": 0.0006884984322823584, "learning_rate": 9.024943310657595e-07, "loss": -0.0396, "num_tokens": 16870271.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 173, "step_time": 292.68767904955894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3377.0, "completions/mean_length": 1331.140625, "completions/mean_terminated_length": 878.7090454101562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.180808439850807, "epoch": 0.09863945578231292, "frac_reward_zero_std": 0.875, "grad_norm": 0.1689453125, "kl": 0.0011846703018818516, "learning_rate": 9.019274376417233e-07, "loss": 0.0019, "num_tokens": 16967256.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 174, "step_time": 296.2339540896937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3701.0, "completions/mean_length": 1500.0625, "completions/mean_terminated_length": 901.0000610351562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.574791617691517, "epoch": 0.0992063492063492, "frac_reward_zero_std": 0.625, "grad_norm": 0.2236328125, "kl": 0.0007171590723373811, "learning_rate": 9.01360544217687e-07, "loss": 0.0315, "num_tokens": 17077364.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 175, "step_time": 315.0234984504059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3318.0, "completions/mean_length": 1103.359375, "completions/mean_terminated_length": 793.77587890625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.082523837685585, "epoch": 0.09977324263038549, "frac_reward_zero_std": 0.75, "grad_norm": 0.275390625, "kl": 0.0009175605064228876, "learning_rate": 9.007936507936508e-07, "loss": -0.031, "num_tokens": 17159283.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 176, "step_time": 294.17216272838414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3481.0, "completions/mean_length": 1434.796875, "completions/mean_terminated_length": 882.4717407226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4342534486204386, "epoch": 0.10034013605442177, "frac_reward_zero_std": 0.75, "grad_norm": 0.1396484375, "kl": 0.0009482698242209153, "learning_rate": 9.002267573696145e-07, "loss": 0.006, "num_tokens": 17262086.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 177, "step_time": 294.06646401435137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3024.0, "completions/mean_length": 1340.296875, "completions/mean_terminated_length": 768.3585205078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.5993550419807434, "epoch": 0.10090702947845805, "frac_reward_zero_std": 0.875, "grad_norm": 0.177734375, "kl": 0.0009339848129457096, "learning_rate": 8.996598639455783e-07, "loss": 0.0196, "num_tokens": 17358305.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 178, "step_time": 295.2236841134727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3958.0, "completions/mean_length": 1307.96875, "completions/mean_terminated_length": 1019.5516967773438, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.9585630483925343, "epoch": 0.10147392290249434, "frac_reward_zero_std": 0.875, "grad_norm": 0.1259765625, "kl": 0.0008951632571552182, "learning_rate": 8.990929705215419e-07, "loss": 0.0021, "num_tokens": 17454799.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 179, "step_time": 304.4246026193723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3604.0, "completions/mean_length": 1124.375, "completions/mean_terminated_length": 759.4385986328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.9999672435224056, "epoch": 0.10204081632653061, "frac_reward_zero_std": 0.75, "grad_norm": 0.2119140625, "kl": 0.0012945063681399915, "learning_rate": 8.985260770975056e-07, "loss": 0.0298, "num_tokens": 17540831.0, "reward": 0.0859375, "reward_std": 0.28942590951919556, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 180, "step_time": 312.59750212263316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3975.0, "completions/mean_length": 1310.15625, "completions/mean_terminated_length": 854.2908935546875, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.9158868305385113, "epoch": 0.10260770975056689, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "kl": 0.0009765556005731924, "learning_rate": 8.979591836734693e-07, "loss": 0.0723, "num_tokens": 17637145.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 181, "step_time": 298.92864027246833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3849.0, "completions/mean_length": 1212.890625, "completions/mean_terminated_length": 858.8245849609375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.62555655837059, "epoch": 0.10317460317460317, "frac_reward_zero_std": 0.625, "grad_norm": 0.2890625, "kl": 0.0009739097476995084, "learning_rate": 8.973922902494331e-07, "loss": 0.0806, "num_tokens": 17725850.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 182, "step_time": 293.8928272379562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4081.0, "completions/mean_length": 1504.3125, "completions/mean_terminated_length": 1080.2181396484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.831495199352503, "epoch": 0.10374149659863946, "frac_reward_zero_std": 0.875, "grad_norm": 0.1708984375, "kl": 0.0009495031881670002, "learning_rate": 8.968253968253968e-07, "loss": -0.0043, "num_tokens": 17833318.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 183, "step_time": 295.0573527580127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4065.0, "completions/mean_length": 1112.140625, "completions/mean_terminated_length": 745.7017822265625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 2.254999879747629, "epoch": 0.10430839002267574, "frac_reward_zero_std": 0.625, "grad_norm": 0.404296875, "kl": 0.0011347034851496574, "learning_rate": 8.962585034013606e-07, "loss": -0.1194, "num_tokens": 17915535.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 184, "step_time": 292.29882525559515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3741.0, "completions/mean_length": 1387.078125, "completions/mean_terminated_length": 943.7999877929688, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "entropy": 3.113091289997101, "epoch": 0.10487528344671201, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "kl": 0.000856978060255642, "learning_rate": 8.956916099773242e-07, "loss": 0.001, "num_tokens": 18015060.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 185, "step_time": 295.9827573513612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3272.0, "completions/mean_length": 1304.546875, "completions/mean_terminated_length": 787.6111450195312, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.156414870172739, "epoch": 0.1054421768707483, "frac_reward_zero_std": 1.0, "grad_norm": 0.00018310546875, "kl": 0.0009036011124408105, "learning_rate": 8.95124716553288e-07, "loss": 0.0, "num_tokens": 18111255.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 186, "step_time": 300.288443938829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1590.90625, "completions/mean_terminated_length": 1180.9818115234375, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "entropy": 3.8824569322168827, "epoch": 0.10600907029478458, "frac_reward_zero_std": 0.625, "grad_norm": 0.1875, "kl": 0.0007456481725967024, "learning_rate": 8.945578231292517e-07, "loss": 0.0122, "num_tokens": 18224001.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 187, "step_time": 292.1030384255573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3740.0, "completions/mean_length": 1235.4375, "completions/mean_terminated_length": 826.7857666015625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.6696518138051033, "epoch": 0.10657596371882086, "frac_reward_zero_std": 0.75, "grad_norm": 0.2001953125, "kl": 0.0008291571648442186, "learning_rate": 8.939909297052153e-07, "loss": 0.0292, "num_tokens": 18314429.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 188, "step_time": 293.7575126942247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3709.0, "completions/mean_length": 1473.890625, "completions/mean_terminated_length": 988.3148193359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.7852463982999325, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.875, "grad_norm": 0.111328125, "kl": 0.0007867878121032845, "learning_rate": 8.934240362811791e-07, "loss": 0.0176, "num_tokens": 18421734.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 189, "step_time": 298.22808313928545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3318.0, "completions/mean_length": 1484.609375, "completions/mean_terminated_length": 942.6226806640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.4584941156208515, "epoch": 0.10770975056689343, "frac_reward_zero_std": 0.625, "grad_norm": 0.2021484375, "kl": 0.0011089006511610933, "learning_rate": 8.928571428571428e-07, "loss": 0.0267, "num_tokens": 18527917.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 190, "step_time": 296.6196203548461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3976.0, "completions/mean_length": 1231.890625, "completions/mean_terminated_length": 822.732177734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.165639989078045, "epoch": 0.1082766439909297, "frac_reward_zero_std": 0.5, "grad_norm": 0.34375, "kl": 0.0011495280305098277, "learning_rate": 8.922902494331065e-07, "loss": 0.0808, "num_tokens": 18620350.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 191, "step_time": 300.34041826426983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4074.0, "completions/mean_length": 1468.296875, "completions/mean_terminated_length": 1092.9107666015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 4.020199842751026, "epoch": 0.10884353741496598, "frac_reward_zero_std": 0.875, "grad_norm": 0.09033203125, "kl": 0.0010856879234779626, "learning_rate": 8.917233560090702e-07, "loss": -0.0153, "num_tokens": 18725657.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 192, "step_time": 295.3089788649231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3248.0, "completions/mean_length": 949.515625, "completions/mean_terminated_length": 739.7500610351562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.225232597440481, "epoch": 0.10941043083900227, "frac_reward_zero_std": 0.75, "grad_norm": 0.28515625, "kl": 0.0010666561047401046, "learning_rate": 8.91156462585034e-07, "loss": -0.0041, "num_tokens": 18797514.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 193, "step_time": 296.91239412501454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2699.0, "completions/mean_length": 1026.953125, "completions/mean_terminated_length": 766.8643798828125, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "entropy": 1.918291261419654, "epoch": 0.10997732426303855, "frac_reward_zero_std": 0.875, "grad_norm": 0.2080078125, "kl": 0.000961602780080284, "learning_rate": 8.905895691609977e-07, "loss": -0.0118, "num_tokens": 18873423.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 194, "step_time": 292.62212283257395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3505.0, "completions/mean_length": 1220.703125, "completions/mean_terminated_length": 1029.0167236328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.112483162432909, "epoch": 0.11054421768707483, "frac_reward_zero_std": 0.875, "grad_norm": 0.2177734375, "kl": 0.0008761044791754102, "learning_rate": 8.900226757369615e-07, "loss": 0.0674, "num_tokens": 18962124.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 195, "step_time": 294.0724186487496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3981.0, "completions/mean_length": 1406.90625, "completions/mean_terminated_length": 966.8726806640625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.886863924562931, "epoch": 0.1111111111111111, "frac_reward_zero_std": 0.5, "grad_norm": 0.349609375, "kl": 0.0009999317062465707, "learning_rate": 8.894557823129251e-07, "loss": 0.0176, "num_tokens": 19067294.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 196, "step_time": 326.1539095537737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1156.734375, "completions/mean_terminated_length": 736.8392944335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.956690076738596, "epoch": 0.1116780045351474, "frac_reward_zero_std": 0.75, "grad_norm": 0.251953125, "kl": 0.0012019378827972105, "learning_rate": 8.888888888888888e-07, "loss": 0.0566, "num_tokens": 19154893.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 197, "step_time": 301.8200301043689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3912.0, "completions/mean_length": 1050.71875, "completions/mean_terminated_length": 735.6896362304688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.1997528932988644, "epoch": 0.11224489795918367, "frac_reward_zero_std": 0.75, "grad_norm": 0.1826171875, "kl": 0.0012252896412974223, "learning_rate": 8.883219954648525e-07, "loss": 0.0173, "num_tokens": 19234467.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 198, "step_time": 302.63945909496397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4051.0, "completions/mean_length": 1176.40625, "completions/mean_terminated_length": 874.3793334960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.246236026287079, "epoch": 0.11281179138321995, "frac_reward_zero_std": 0.5, "grad_norm": 0.349609375, "kl": 0.0011566742414288456, "learning_rate": 8.877551020408163e-07, "loss": 0.0341, "num_tokens": 19320229.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 199, "step_time": 294.37082410696894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4041.0, "completions/mean_length": 1260.75, "completions/mean_terminated_length": 796.7999877929688, "completions/min_length": 75.0, "completions/min_terminated_length": 75.0, "entropy": 3.445892170071602, "epoch": 0.11337868480725624, "frac_reward_zero_std": 0.625, "grad_norm": 0.310546875, "kl": 0.0012227596816956066, "learning_rate": 8.8718820861678e-07, "loss": -0.0124, "num_tokens": 19412173.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 200, "step_time": 293.2042292887345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3265.0, "completions/mean_length": 1124.421875, "completions/mean_terminated_length": 759.4912109375, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "entropy": 2.0993868820369244, "epoch": 0.11394557823129252, "frac_reward_zero_std": 0.625, "grad_norm": 0.2294921875, "kl": 0.0008618845185992541, "learning_rate": 8.866213151927438e-07, "loss": 0.0038, "num_tokens": 19495376.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 201, "step_time": 296.0682121384889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4012.0, "completions/mean_length": 1554.3125, "completions/mean_terminated_length": 906.431396484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.8276742175221443, "epoch": 0.1145124716553288, "frac_reward_zero_std": 0.75, "grad_norm": 0.13671875, "kl": 0.0011864844018418808, "learning_rate": 8.860544217687075e-07, "loss": -0.0066, "num_tokens": 19608212.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 202, "step_time": 312.8659586533904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3812.0, "completions/mean_length": 1436.359375, "completions/mean_terminated_length": 1001.1454467773438, "completions/min_length": 51.0, "completions/min_terminated_length": 51.0, "entropy": 4.753630865365267, "epoch": 0.11507936507936507, "frac_reward_zero_std": 0.875, "grad_norm": 0.154296875, "kl": 0.0015447768164449371, "learning_rate": 8.854875283446712e-07, "loss": 0.0476, "num_tokens": 19710979.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 203, "step_time": 294.42433897126466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3775.0, "completions/mean_length": 1548.28125, "completions/mean_terminated_length": 1076.4814453125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.073138736188412, "epoch": 0.11564625850340136, "frac_reward_zero_std": 0.625, "grad_norm": 0.2138671875, "kl": 0.0007871567868278362, "learning_rate": 8.849206349206348e-07, "loss": 0.0537, "num_tokens": 19821317.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 204, "step_time": 295.367243854329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3729.0, "completions/mean_length": 1307.0, "completions/mean_terminated_length": 908.5714721679688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.909986387938261, "epoch": 0.11621315192743764, "frac_reward_zero_std": 0.75, "grad_norm": 0.1982421875, "kl": 0.0007760782882542117, "learning_rate": 8.843537414965985e-07, "loss": 0.0162, "num_tokens": 19915189.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 205, "step_time": 291.66046211402863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3515.0, "completions/mean_length": 1305.734375, "completions/mean_terminated_length": 789.0185546875, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "entropy": 2.36519917473197, "epoch": 0.11678004535147392, "frac_reward_zero_std": 0.875, "grad_norm": 0.1796875, "kl": 0.0011316735999571392, "learning_rate": 8.837868480725623e-07, "loss": 0.0065, "num_tokens": 20011828.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 206, "step_time": 299.7828513002023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4080.0, "completions/mean_length": 881.109375, "completions/mean_terminated_length": 830.0794067382812, "completions/min_length": 85.0, "completions/min_terminated_length": 85.0, "entropy": 3.779766697436571, "epoch": 0.11734693877551021, "frac_reward_zero_std": 0.375, "grad_norm": 0.4921875, "kl": 0.0010524408498895355, "learning_rate": 8.83219954648526e-07, "loss": 0.0453, "num_tokens": 20080899.0, "reward": 0.4296875, "reward_std": 0.8538762331008911, "rewards/accuracy_reward_func/mean": 0.375, "rewards/accuracy_reward_func/std": 0.7867957949638367, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 207, "step_time": 308.67669108323753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3160.0, "completions/mean_length": 1045.453125, "completions/mean_terminated_length": 842.0833740234375, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "entropy": 3.642435859888792, "epoch": 0.11791383219954649, "frac_reward_zero_std": 0.75, "grad_norm": 0.30078125, "kl": 0.0010334904181945603, "learning_rate": 8.826530612244898e-07, "loss": 0.0998, "num_tokens": 20158808.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 208, "step_time": 293.9838742958382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3693.0, "completions/mean_length": 1332.265625, "completions/mean_terminated_length": 880.0181274414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.808041248470545, "epoch": 0.11848072562358276, "frac_reward_zero_std": 0.875, "grad_norm": 0.15234375, "kl": 0.0011558761889318703, "learning_rate": 8.820861678004535e-07, "loss": 0.0039, "num_tokens": 20254625.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 209, "step_time": 291.6761906463653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4050.0, "completions/mean_length": 1218.421875, "completions/mean_terminated_length": 974.559326171875, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.6133025735616684, "epoch": 0.11904761904761904, "frac_reward_zero_std": 0.75, "grad_norm": 0.224609375, "kl": 0.0008826196917652851, "learning_rate": 8.815192743764172e-07, "loss": -0.0015, "num_tokens": 20343972.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 210, "step_time": 292.4662137590349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3527.0, "completions/mean_length": 1256.0, "completions/mean_terminated_length": 532.0784301757812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.478566935285926, "epoch": 0.11961451247165533, "frac_reward_zero_std": 0.5, "grad_norm": 0.294921875, "kl": 0.0017366717056574998, "learning_rate": 8.809523809523809e-07, "loss": 0.0029, "num_tokens": 20435484.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 211, "step_time": 297.2094409316778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3968.0, "completions/mean_length": 1577.015625, "completions/mean_terminated_length": 1110.5369873046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.70728861913085, "epoch": 0.12018140589569161, "frac_reward_zero_std": 0.875, "grad_norm": 0.146484375, "kl": 0.0008267538469226565, "learning_rate": 8.803854875283446e-07, "loss": 0.028, "num_tokens": 20547469.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 212, "step_time": 295.8759975992143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3732.0, "completions/mean_length": 1148.28125, "completions/mean_terminated_length": 843.3448486328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.0519875772297382, "epoch": 0.12074829931972789, "frac_reward_zero_std": 0.75, "grad_norm": 0.234375, "kl": 0.0012846419031120604, "learning_rate": 8.798185941043083e-07, "loss": 0.0341, "num_tokens": 20633791.0, "reward": 0.1171875, "reward_std": 0.41539984941482544, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 213, "step_time": 298.65141285955906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4010.0, "completions/mean_length": 1745.609375, "completions/mean_terminated_length": 1087.5, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.023875005543232, "epoch": 0.12131519274376418, "frac_reward_zero_std": 0.75, "grad_norm": 0.138671875, "kl": 0.0007914034013083437, "learning_rate": 8.792517006802721e-07, "loss": 0.0095, "num_tokens": 20757390.0, "reward": 0.171875, "reward_std": 0.5579067468643188, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 214, "step_time": 299.87372230086476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3233.0, "completions/mean_length": 976.15625, "completions/mean_terminated_length": 711.7626953125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 2.71496070548892, "epoch": 0.12188208616780045, "frac_reward_zero_std": 0.75, "grad_norm": 0.2451171875, "kl": 0.0010727268054324668, "learning_rate": 8.786848072562358e-07, "loss": 0.0041, "num_tokens": 20831504.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 215, "step_time": 298.7091171545908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3868.0, "completions/mean_length": 1516.609375, "completions/mean_terminated_length": 794.3800048828125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.812237862497568, "epoch": 0.12244897959183673, "frac_reward_zero_std": 0.75, "grad_norm": 0.203125, "kl": 0.0013310160175024066, "learning_rate": 8.781179138321995e-07, "loss": -0.019, "num_tokens": 20944471.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 216, "step_time": 313.4162803757936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3581.0, "completions/mean_length": 1257.0625, "completions/mean_terminated_length": 1117.4425048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.202047243714333, "epoch": 0.12301587301587301, "frac_reward_zero_std": 0.625, "grad_norm": 0.28125, "kl": 0.0007741837871435564, "learning_rate": 8.775510204081632e-07, "loss": 0.0311, "num_tokens": 21036235.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 217, "step_time": 300.48187935724854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3732.0, "completions/mean_length": 1682.34375, "completions/mean_terminated_length": 1287.3818359375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.5055681094527245, "epoch": 0.1235827664399093, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "kl": 0.000785970254582935, "learning_rate": 8.76984126984127e-07, "loss": 0.0282, "num_tokens": 21155937.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 218, "step_time": 299.3899869751185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4067.0, "completions/mean_length": 1030.34375, "completions/mean_terminated_length": 825.9667358398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4379561692476273, "epoch": 0.12414965986394558, "frac_reward_zero_std": 0.75, "grad_norm": 0.2412109375, "kl": 0.0011134496562590357, "learning_rate": 8.764172335600907e-07, "loss": 0.0166, "num_tokens": 21233175.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 219, "step_time": 292.3021672721952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3961.0, "completions/mean_length": 1548.75, "completions/mean_terminated_length": 1020.0755004882812, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 3.782156143337488, "epoch": 0.12471655328798185, "frac_reward_zero_std": 0.875, "grad_norm": 0.1396484375, "kl": 0.001276508145565458, "learning_rate": 8.758503401360544e-07, "loss": -0.0023, "num_tokens": 21344287.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 220, "step_time": 298.39621212985367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4026.0, "completions/mean_length": 1192.40625, "completions/mean_terminated_length": 998.8333740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.345248498022556, "epoch": 0.12528344671201813, "frac_reward_zero_std": 0.875, "grad_norm": 0.177734375, "kl": 0.0008278152436105302, "learning_rate": 8.752834467120181e-07, "loss": 0.0474, "num_tokens": 21432529.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 221, "step_time": 294.27885660901666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3505.0, "completions/mean_length": 1050.15625, "completions/mean_terminated_length": 792.0338745117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.82575561478734, "epoch": 0.12585034013605442, "frac_reward_zero_std": 0.625, "grad_norm": 0.333984375, "kl": 0.0010389726776338648, "learning_rate": 8.747165532879818e-07, "loss": -0.0022, "num_tokens": 21513307.0, "reward": 0.1796875, "reward_std": 0.5298036932945251, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 222, "step_time": 312.0486962515861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3608.0, "completions/mean_length": 1200.875, "completions/mean_terminated_length": 845.3333129882812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.7446229569613934, "epoch": 0.1264172335600907, "frac_reward_zero_std": 0.5, "grad_norm": 0.310546875, "kl": 0.001183695454528788, "learning_rate": 8.741496598639455e-07, "loss": 0.0173, "num_tokens": 21601027.0, "reward": 0.234375, "reward_std": 0.5907269716262817, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 223, "step_time": 293.27236039657146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4052.0, "completions/mean_length": 1081.3125, "completions/mean_terminated_length": 825.8305053710938, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 2.898800101131201, "epoch": 0.12698412698412698, "frac_reward_zero_std": 0.75, "grad_norm": 0.2412109375, "kl": 0.0013538178318412974, "learning_rate": 8.735827664399093e-07, "loss": 0.0675, "num_tokens": 21681167.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 224, "step_time": 292.89520471170545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3207.0, "completions/mean_length": 1052.921875, "completions/mean_terminated_length": 738.1206665039062, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 3.210470274090767, "epoch": 0.12755102040816327, "frac_reward_zero_std": 0.625, "grad_norm": 0.3125, "kl": 0.0011044722068618285, "learning_rate": 8.73015873015873e-07, "loss": 0.0513, "num_tokens": 21765786.0, "reward": 0.3359375, "reward_std": 0.8020543456077576, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 225, "step_time": 337.17833896446973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4095.0, "completions/mean_length": 924.1875, "completions/mean_terminated_length": 712.7333984375, "completions/min_length": 68.0, "completions/min_terminated_length": 68.0, "entropy": 2.6770739629864693, "epoch": 0.12811791383219956, "frac_reward_zero_std": 0.5, "grad_norm": 0.341796875, "kl": 0.0011462787570053479, "learning_rate": 8.724489795918368e-07, "loss": 0.0218, "num_tokens": 21835750.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 226, "step_time": 294.12168524228036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3575.0, "completions/mean_length": 1203.609375, "completions/mean_terminated_length": 790.4107666015625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 3.102189105004072, "epoch": 0.12868480725623582, "frac_reward_zero_std": 0.875, "grad_norm": 0.169921875, "kl": 0.001319015835179016, "learning_rate": 8.718820861678005e-07, "loss": 0.0073, "num_tokens": 21923997.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 227, "step_time": 296.67299195192754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4084.0, "completions/mean_length": 1021.75, "completions/mean_terminated_length": 870.5573120117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.400593936443329, "epoch": 0.1292517006802721, "frac_reward_zero_std": 0.75, "grad_norm": 0.224609375, "kl": 0.001097561624192167, "learning_rate": 8.71315192743764e-07, "loss": -0.002, "num_tokens": 22000061.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 228, "step_time": 292.2873935243115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3552.0, "completions/mean_length": 1350.84375, "completions/mean_terminated_length": 958.6785888671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.85386623442173, "epoch": 0.12981859410430838, "frac_reward_zero_std": 0.625, "grad_norm": 0.25, "kl": 0.0013140782157279318, "learning_rate": 8.707482993197278e-07, "loss": 0.0063, "num_tokens": 22097731.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 229, "step_time": 295.80912455078214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3852.0, "completions/mean_length": 1410.8125, "completions/mean_terminated_length": 913.5555419921875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.6953197680413723, "epoch": 0.13038548752834467, "frac_reward_zero_std": 0.75, "grad_norm": 0.154296875, "kl": 0.0011070836553699337, "learning_rate": 8.701814058956915e-07, "loss": 0.0022, "num_tokens": 22199487.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 230, "step_time": 296.0781445270404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4094.0, "completions/mean_length": 1486.34375, "completions/mean_terminated_length": 1003.0740966796875, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 3.8559432066977024, "epoch": 0.13095238095238096, "frac_reward_zero_std": 0.75, "grad_norm": 0.1884765625, "kl": 0.001084855654880812, "learning_rate": 8.696145124716553e-07, "loss": 0.0235, "num_tokens": 22304917.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 231, "step_time": 294.1224720198661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3586.0, "completions/mean_length": 1244.890625, "completions/mean_terminated_length": 894.75439453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.2090540044009686, "epoch": 0.13151927437641722, "frac_reward_zero_std": 0.75, "grad_norm": 0.298828125, "kl": 0.0009592715159669751, "learning_rate": 8.69047619047619e-07, "loss": 0.0003, "num_tokens": 22396678.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 232, "step_time": 299.46676432620734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3794.0, "completions/mean_length": 976.46875, "completions/mean_terminated_length": 653.7586059570312, "completions/min_length": 67.0, "completions/min_terminated_length": 67.0, "entropy": 3.0212653540074825, "epoch": 0.1320861678004535, "frac_reward_zero_std": 0.875, "grad_norm": 0.1943359375, "kl": 0.0018015931818808895, "learning_rate": 8.684807256235828e-07, "loss": -0.0004, "num_tokens": 22470708.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 233, "step_time": 294.826646476984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4034.0, "completions/mean_length": 1293.921875, "completions/mean_terminated_length": 1004.0516967773438, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "entropy": 3.842056378722191, "epoch": 0.1326530612244898, "frac_reward_zero_std": 0.75, "grad_norm": 0.25390625, "kl": 0.0010123011907126056, "learning_rate": 8.679138321995464e-07, "loss": -0.0169, "num_tokens": 22566031.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 234, "step_time": 304.18481398466974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3868.0, "completions/mean_length": 1329.65625, "completions/mean_terminated_length": 817.370361328125, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 3.2029078900814056, "epoch": 0.13321995464852607, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "kl": 0.0013659713295055553, "learning_rate": 8.673469387755102e-07, "loss": 0.0055, "num_tokens": 22667241.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 235, "step_time": 315.4635110665113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3624.0, "completions/mean_length": 1535.328125, "completions/mean_terminated_length": 1061.129638671875, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "entropy": 2.929383337497711, "epoch": 0.13378684807256236, "frac_reward_zero_std": 0.75, "grad_norm": 0.1787109375, "kl": 0.0010227088023384567, "learning_rate": 8.667800453514738e-07, "loss": 0.0159, "num_tokens": 22780798.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 236, "step_time": 308.49995029717684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3946.0, "completions/mean_length": 1041.796875, "completions/mean_terminated_length": 838.183349609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.236738458275795, "epoch": 0.13435374149659865, "frac_reward_zero_std": 0.75, "grad_norm": 0.2734375, "kl": 0.0012075366357748862, "learning_rate": 8.662131519274376e-07, "loss": 0.0624, "num_tokens": 22859321.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 237, "step_time": 300.28507947269827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4058.0, "completions/mean_length": 1361.09375, "completions/mean_terminated_length": 913.5635986328125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.180329939350486, "epoch": 0.1349206349206349, "frac_reward_zero_std": 0.625, "grad_norm": 0.2265625, "kl": 0.0014698277136631077, "learning_rate": 8.656462585034013e-07, "loss": 0.0046, "num_tokens": 22957831.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 238, "step_time": 294.9036829229444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3610.0, "completions/mean_length": 1153.0, "completions/mean_terminated_length": 903.5932006835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.6870043724775314, "epoch": 0.1354875283446712, "frac_reward_zero_std": 0.875, "grad_norm": 0.1591796875, "kl": 0.0015672321533202194, "learning_rate": 8.650793650793651e-07, "loss": 0.0013, "num_tokens": 23042711.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 239, "step_time": 295.62402374856174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3883.0, "completions/mean_length": 1244.921875, "completions/mean_terminated_length": 837.6250610351562, "completions/min_length": 86.0, "completions/min_terminated_length": 86.0, "entropy": 3.7111910469830036, "epoch": 0.1360544217687075, "frac_reward_zero_std": 0.875, "grad_norm": 0.154296875, "kl": 0.0011819430092145922, "learning_rate": 8.645124716553288e-07, "loss": 0.0069, "num_tokens": 23133466.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 240, "step_time": 295.11580411344767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4088.0, "completions/mean_length": 1177.9375, "completions/mean_terminated_length": 761.0714721679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.964115008711815, "epoch": 0.13662131519274376, "frac_reward_zero_std": 0.75, "grad_norm": 0.2236328125, "kl": 0.0012315598942223005, "learning_rate": 8.639455782312925e-07, "loss": -0.0057, "num_tokens": 23220238.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 241, "step_time": 296.22310270369053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3904.0, "completions/mean_length": 1137.546875, "completions/mean_terminated_length": 831.5, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.6043158210814, "epoch": 0.13718820861678005, "frac_reward_zero_std": 0.875, "grad_norm": 0.11962890625, "kl": 0.0013280221428431105, "learning_rate": 8.633786848072562e-07, "loss": -0.0033, "num_tokens": 23304257.0, "reward": 0.1171875, "reward_std": 0.4519989788532257, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 242, "step_time": 296.49121211096644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2450.0, "completions/mean_length": 911.875, "completions/mean_terminated_length": 582.4827270507812, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "entropy": 2.447322765365243, "epoch": 0.1377551020408163, "frac_reward_zero_std": 0.75, "grad_norm": 0.375, "kl": 0.002247496593554388, "learning_rate": 8.6281179138322e-07, "loss": -0.0561, "num_tokens": 23374905.0, "reward": 0.140625, "reward_std": 0.5076302886009216, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 243, "step_time": 301.67275542300195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3944.0, "completions/mean_length": 1379.09375, "completions/mean_terminated_length": 990.96435546875, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.6930803544819355, "epoch": 0.1383219954648526, "frac_reward_zero_std": 0.625, "grad_norm": 0.2138671875, "kl": 0.0011230875043111155, "learning_rate": 8.622448979591836e-07, "loss": 0.0147, "num_tokens": 23474367.0, "reward": 0.203125, "reward_std": 0.5890451669692993, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 244, "step_time": 293.819984880276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3930.0, "completions/mean_length": 1300.984375, "completions/mean_terminated_length": 843.6181640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.570481050759554, "epoch": 0.1388888888888889, "frac_reward_zero_std": 0.625, "grad_norm": 0.251953125, "kl": 0.0012484762719395803, "learning_rate": 8.616780045351474e-07, "loss": 0.0091, "num_tokens": 23571702.0, "reward": 0.1484375, "reward_std": 0.47709178924560547, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 245, "step_time": 321.6608476731926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3975.0, "completions/mean_length": 1374.109375, "completions/mean_terminated_length": 928.7090454101562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.8169440925121307, "epoch": 0.13945578231292516, "frac_reward_zero_std": 0.75, "grad_norm": 0.197265625, "kl": 0.001357796514639631, "learning_rate": 8.611111111111111e-07, "loss": -0.016, "num_tokens": 23673061.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 246, "step_time": 299.5307418331504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4082.0, "completions/mean_length": 1398.5, "completions/mean_terminated_length": 898.9629516601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.31706964597106, "epoch": 0.14002267573696145, "frac_reward_zero_std": 0.625, "grad_norm": 0.2734375, "kl": 0.0010165272087760968, "learning_rate": 8.605442176870747e-07, "loss": -0.0157, "num_tokens": 23778381.0, "reward": 0.203125, "reward_std": 0.6281666159629822, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 247, "step_time": 323.87792470585555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3988.0, "completions/mean_length": 858.96875, "completions/mean_terminated_length": 754.54833984375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 2.975732732564211, "epoch": 0.14058956916099774, "frac_reward_zero_std": 0.75, "grad_norm": 0.2890625, "kl": 0.0014932136873540003, "learning_rate": 8.599773242630385e-07, "loss": 0.082, "num_tokens": 23843507.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 248, "step_time": 296.37086434010416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4060.0, "completions/mean_length": 1291.453125, "completions/mean_terminated_length": 1104.4833984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.400741457939148, "epoch": 0.141156462585034, "frac_reward_zero_std": 0.875, "grad_norm": 0.16796875, "kl": 0.0010249239840049995, "learning_rate": 8.594104308390022e-07, "loss": -0.0014, "num_tokens": 23939296.0, "reward": 0.0625, "reward_std": 0.32732683420181274, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 249, "step_time": 302.395604101941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3825.0, "completions/mean_length": 1397.15625, "completions/mean_terminated_length": 1011.607177734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 4.738048255443573, "epoch": 0.1417233560090703, "frac_reward_zero_std": 0.75, "grad_norm": 0.1904296875, "kl": 0.001050947124895174, "learning_rate": 8.58843537414966e-07, "loss": 0.0179, "num_tokens": 24040082.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 250, "step_time": 295.45087849814445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.0, "completions/mean_length": 1354.984375, "completions/mean_terminated_length": 963.4107666015625, "completions/min_length": 76.0, "completions/min_terminated_length": 76.0, "entropy": 4.9353371784091, "epoch": 0.14229024943310659, "frac_reward_zero_std": 0.875, "grad_norm": 0.1611328125, "kl": 0.0012083352667104919, "learning_rate": 8.582766439909297e-07, "loss": 0.019, "num_tokens": 24139025.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 251, "step_time": 298.3130410639569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3563.0, "completions/mean_length": 1210.84375, "completions/mean_terminated_length": 676.5555419921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.7371104396879673, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.75, "grad_norm": 0.2333984375, "kl": 0.0015491756021219771, "learning_rate": 8.577097505668934e-07, "loss": 0.0058, "num_tokens": 24228535.0, "reward": 0.125, "reward_std": 0.48795002698898315, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 252, "step_time": 294.7551212403923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4043.0, "completions/mean_length": 804.984375, "completions/mean_terminated_length": 752.74609375, "completions/min_length": 61.0, "completions/min_terminated_length": 61.0, "entropy": 2.979743916541338, "epoch": 0.14342403628117914, "frac_reward_zero_std": 0.625, "grad_norm": 0.400390625, "kl": 0.0013989633553137537, "learning_rate": 8.57142857142857e-07, "loss": 0.057, "num_tokens": 24290910.0, "reward": 0.203125, "reward_std": 0.6154024600982666, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 253, "step_time": 293.3673425382003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3405.0, "completions/mean_length": 1081.078125, "completions/mean_terminated_length": 769.1896362304688, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 2.687511883676052, "epoch": 0.14399092970521543, "frac_reward_zero_std": 0.75, "grad_norm": 0.283203125, "kl": 0.0014965555710659828, "learning_rate": 8.565759637188208e-07, "loss": 0.0324, "num_tokens": 24372307.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 254, "step_time": 297.3848839784041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3124.0, "completions/mean_length": 1213.203125, "completions/mean_terminated_length": 614.8867797851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.7044095527380705, "epoch": 0.1445578231292517, "frac_reward_zero_std": 0.875, "grad_norm": 0.166015625, "kl": 0.0013742621849814896, "learning_rate": 8.560090702947845e-07, "loss": -0.0254, "num_tokens": 24461304.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 255, "step_time": 294.57220361568034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3713.0, "completions/mean_length": 1023.65625, "completions/mean_terminated_length": 705.8275756835938, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.3083490654826164, "epoch": 0.14512471655328799, "frac_reward_zero_std": 0.875, "grad_norm": 0.2197265625, "kl": 0.0014701982800033875, "learning_rate": 8.554421768707483e-07, "loss": 0.0396, "num_tokens": 24540786.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 256, "step_time": 314.96845386736095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2654.0, "completions/mean_length": 1142.859375, "completions/mean_terminated_length": 659.6181640625, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "entropy": 2.601846519857645, "epoch": 0.14569160997732428, "frac_reward_zero_std": 0.625, "grad_norm": 0.26171875, "kl": 0.0011016668959200615, "learning_rate": 8.54875283446712e-07, "loss": 0.0042, "num_tokens": 24627689.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 257, "step_time": 312.0238794479519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3674.0, "completions/mean_length": 1136.515625, "completions/mean_terminated_length": 990.9671630859375, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 4.248551677912474, "epoch": 0.14625850340136054, "frac_reward_zero_std": 0.75, "grad_norm": 0.31640625, "kl": 0.0011931121989618987, "learning_rate": 8.543083900226758e-07, "loss": 0.0525, "num_tokens": 24710786.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 258, "step_time": 291.4342953702435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3763.0, "completions/mean_length": 1476.3125, "completions/mean_terminated_length": 808.549072265625, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "entropy": 2.9006321597844362, "epoch": 0.14682539682539683, "frac_reward_zero_std": 0.75, "grad_norm": 0.1923828125, "kl": 0.0013805588678224012, "learning_rate": 8.537414965986394e-07, "loss": 0.008, "num_tokens": 24817006.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 259, "step_time": 302.66142302937806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3869.0, "completions/mean_length": 1068.75, "completions/mean_terminated_length": 812.203369140625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.3819946944713593, "epoch": 0.1473922902494331, "frac_reward_zero_std": 0.5, "grad_norm": 0.373046875, "kl": 0.0014789492888667155, "learning_rate": 8.531746031746031e-07, "loss": 0.0723, "num_tokens": 24895974.0, "reward": 0.3125, "reward_std": 0.6986381411552429, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 260, "step_time": 296.31661470234394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3818.0, "completions/mean_length": 957.65625, "completions/mean_terminated_length": 803.3114013671875, "completions/min_length": 60.0, "completions/min_terminated_length": 60.0, "entropy": 3.298921812325716, "epoch": 0.14795918367346939, "frac_reward_zero_std": 0.625, "grad_norm": 0.30859375, "kl": 0.001135968590460834, "learning_rate": 8.526077097505668e-07, "loss": 0.0214, "num_tokens": 24969304.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 261, "step_time": 301.2923899097368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3981.0, "completions/mean_length": 1562.875, "completions/mean_terminated_length": 1251.7894287109375, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "entropy": 3.8599331825971603, "epoch": 0.14852607709750568, "frac_reward_zero_std": 0.875, "grad_norm": 0.1708984375, "kl": 0.0009275648299080785, "learning_rate": 8.520408163265306e-07, "loss": 0.0194, "num_tokens": 25082000.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 262, "step_time": 300.24944328609854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3120.0, "completions/mean_length": 1391.6875, "completions/mean_terminated_length": 767.6154174804688, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.442591927945614, "epoch": 0.14909297052154194, "frac_reward_zero_std": 0.875, "grad_norm": 0.15234375, "kl": 0.0016389939119108021, "learning_rate": 8.514739229024943e-07, "loss": 0.028, "num_tokens": 25182972.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 263, "step_time": 293.67484748456627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3836.0, "completions/mean_length": 1541.796875, "completions/mean_terminated_length": 826.6199951171875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.9540892839431763, "epoch": 0.14965986394557823, "frac_reward_zero_std": 0.5, "grad_norm": 0.267578125, "kl": 0.001244919574673986, "learning_rate": 8.509070294784581e-07, "loss": -0.0043, "num_tokens": 25293119.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 264, "step_time": 294.5663408571854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3745.0, "completions/mean_length": 1096.5625, "completions/mean_terminated_length": 728.2105102539062, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 3.3549214117228985, "epoch": 0.15022675736961452, "frac_reward_zero_std": 0.75, "grad_norm": 0.2392578125, "kl": 0.0013892023780499585, "learning_rate": 8.503401360544217e-07, "loss": 0.0056, "num_tokens": 25375267.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 265, "step_time": 298.0915329474956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3938.0, "completions/mean_length": 1631.65625, "completions/mean_terminated_length": 1062.9615478515625, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "entropy": 5.047824844717979, "epoch": 0.15079365079365079, "frac_reward_zero_std": 0.875, "grad_norm": 0.1591796875, "kl": 0.001154161127487896, "learning_rate": 8.497732426303855e-07, "loss": 0.0211, "num_tokens": 25490301.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 266, "step_time": 294.235875419341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3978.0, "completions/mean_length": 1112.96875, "completions/mean_terminated_length": 804.3793334960938, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "entropy": 3.4297971054911613, "epoch": 0.15136054421768708, "frac_reward_zero_std": 0.25, "grad_norm": 0.416015625, "kl": 0.001714525820716517, "learning_rate": 8.492063492063492e-07, "loss": 0.0191, "num_tokens": 25573715.0, "reward": 0.34375, "reward_std": 0.7175520658493042, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 267, "step_time": 300.0728285070509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3800.0, "completions/mean_length": 1327.265625, "completions/mean_terminated_length": 752.6226806640625, "completions/min_length": 74.0, "completions/min_terminated_length": 74.0, "entropy": 2.689064610749483, "epoch": 0.15192743764172337, "frac_reward_zero_std": 0.75, "grad_norm": 0.1953125, "kl": 0.0015337007698690286, "learning_rate": 8.486394557823128e-07, "loss": 0.0092, "num_tokens": 25673220.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 268, "step_time": 310.35281988978386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2822.0, "completions/mean_length": 844.140625, "completions/mean_terminated_length": 568.559326171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.2734266445040703, "epoch": 0.15249433106575963, "frac_reward_zero_std": 0.5, "grad_norm": 0.396484375, "kl": 0.0019259779910498764, "learning_rate": 8.480725623582766e-07, "loss": -0.0005, "num_tokens": 25740221.0, "reward": 0.203125, "reward_std": 0.5471335649490356, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 269, "step_time": 308.26017513033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2713.0, "completions/mean_length": 999.296875, "completions/mean_terminated_length": 619.0, "completions/min_length": 60.0, "completions/min_terminated_length": 60.0, "entropy": 2.6519531086087227, "epoch": 0.15306122448979592, "frac_reward_zero_std": 0.625, "grad_norm": 0.29296875, "kl": 0.0014964106158004142, "learning_rate": 8.475056689342403e-07, "loss": 0.0044, "num_tokens": 25816448.0, "reward": 0.171875, "reward_std": 0.528690755367279, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 270, "step_time": 305.5210933703929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3468.0, "completions/mean_length": 1258.546875, "completions/mean_terminated_length": 910.0877075195312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.859191494062543, "epoch": 0.1536281179138322, "frac_reward_zero_std": 0.875, "grad_norm": 0.10693359375, "kl": 0.0013441895043797558, "learning_rate": 8.469387755102041e-07, "loss": 0.0013, "num_tokens": 25908467.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 271, "step_time": 296.35443599242717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4000.0, "completions/mean_length": 1257.25, "completions/mean_terminated_length": 963.586181640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.312869865447283, "epoch": 0.15419501133786848, "frac_reward_zero_std": 0.625, "grad_norm": 0.275390625, "kl": 0.00116521892050514, "learning_rate": 8.463718820861677e-07, "loss": 0.0534, "num_tokens": 26000899.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 272, "step_time": 293.82065302692354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4063.0, "completions/mean_length": 1249.640625, "completions/mean_terminated_length": 1008.4237060546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.00705923140049, "epoch": 0.15476190476190477, "frac_reward_zero_std": 0.5, "grad_norm": 0.380859375, "kl": 0.0010208657022303669, "learning_rate": 8.458049886621315e-07, "loss": -0.0283, "num_tokens": 26092732.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 273, "step_time": 300.6625357680023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3971.0, "completions/mean_length": 1185.390625, "completions/mean_terminated_length": 769.5892944335938, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.417567305266857, "epoch": 0.15532879818594103, "frac_reward_zero_std": 0.625, "grad_norm": 0.298828125, "kl": 0.001261524190340424, "learning_rate": 8.452380952380952e-07, "loss": 0.0086, "num_tokens": 26179717.0, "reward": 0.21875, "reward_std": 0.5900968313217163, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 274, "step_time": 298.452397396788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3687.0, "completions/mean_length": 1130.609375, "completions/mean_terminated_length": 706.982177734375, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 2.985208410769701, "epoch": 0.15589569160997732, "frac_reward_zero_std": 0.75, "grad_norm": 0.201171875, "kl": 0.001736221991450293, "learning_rate": 8.44671201814059e-07, "loss": -0.0012, "num_tokens": 26262564.0, "reward": 0.265625, "reward_std": 0.6782841682434082, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 275, "step_time": 294.14131032209843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4013.0, "completions/mean_length": 1411.953125, "completions/mean_terminated_length": 914.9074096679688, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "entropy": 3.070179019123316, "epoch": 0.1564625850340136, "frac_reward_zero_std": 0.75, "grad_norm": 0.189453125, "kl": 0.0013637927659146953, "learning_rate": 8.441043083900226e-07, "loss": -0.0005, "num_tokens": 26365065.0, "reward": 0.0703125, "reward_std": 0.39708447456359863, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 276, "step_time": 296.15611359570175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4049.0, "completions/mean_length": 1408.421875, "completions/mean_terminated_length": 1078.368408203125, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "entropy": 4.443922260776162, "epoch": 0.15702947845804988, "frac_reward_zero_std": 0.5, "grad_norm": 0.2333984375, "kl": 0.001101307847420685, "learning_rate": 8.435374149659864e-07, "loss": 0.0635, "num_tokens": 26467676.0, "reward": 0.265625, "reward_std": 0.6664806604385376, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 277, "step_time": 296.64144099131227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3967.0, "completions/mean_length": 1329.296875, "completions/mean_terminated_length": 1094.83056640625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "entropy": 3.343450289219618, "epoch": 0.15759637188208617, "frac_reward_zero_std": 0.75, "grad_norm": 0.193359375, "kl": 0.0009303500173700741, "learning_rate": 8.4297052154195e-07, "loss": 0.0212, "num_tokens": 26565863.0, "reward": 0.1171875, "reward_std": 0.41539984941482544, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 278, "step_time": 306.69521214906126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3969.0, "completions/mean_length": 1104.734375, "completions/mean_terminated_length": 795.2930908203125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.88244404271245, "epoch": 0.15816326530612246, "frac_reward_zero_std": 0.75, "grad_norm": 0.291015625, "kl": 0.0014865316625218838, "learning_rate": 8.424036281179138e-07, "loss": -0.0034, "num_tokens": 26647478.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 279, "step_time": 296.1632002191618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3919.0, "completions/mean_length": 1359.15625, "completions/mean_terminated_length": 968.1785888671875, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "entropy": 4.191066320985556, "epoch": 0.15873015873015872, "frac_reward_zero_std": 0.75, "grad_norm": 0.2314453125, "kl": 0.0012526897280622507, "learning_rate": 8.418367346938775e-07, "loss": 0.0037, "num_tokens": 26745072.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 280, "step_time": 291.89493420906365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3779.0, "completions/mean_length": 1178.59375, "completions/mean_terminated_length": 931.35595703125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.5426559150218964, "epoch": 0.159297052154195, "frac_reward_zero_std": 0.625, "grad_norm": 0.5, "kl": 0.0011631767283688532, "learning_rate": 8.412698412698413e-07, "loss": -0.1123, "num_tokens": 26832014.0, "reward": 0.1171875, "reward_std": 0.46922925114631653, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 281, "step_time": 300.498527049087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3771.0, "completions/mean_length": 1375.5625, "completions/mean_terminated_length": 986.9285888671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.4970968775451183, "epoch": 0.1598639455782313, "frac_reward_zero_std": 0.625, "grad_norm": 0.2275390625, "kl": 0.0015308281963370973, "learning_rate": 8.40702947845805e-07, "loss": 0.033, "num_tokens": 26931434.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 282, "step_time": 294.31987982895225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3254.0, "completions/mean_length": 1049.328125, "completions/mean_terminated_length": 614.0892944335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.0270057693123817, "epoch": 0.16043083900226757, "frac_reward_zero_std": 0.875, "grad_norm": 0.1015625, "kl": 0.0014629809629695956, "learning_rate": 8.401360544217688e-07, "loss": -0.0042, "num_tokens": 27012255.0, "reward": 0.1796875, "reward_std": 0.6131313443183899, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 283, "step_time": 303.08907681982964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3777.0, "completions/mean_length": 1197.109375, "completions/mean_terminated_length": 897.22412109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.958734929561615, "epoch": 0.16099773242630386, "frac_reward_zero_std": 0.5, "grad_norm": 0.30078125, "kl": 0.0010162165599467698, "learning_rate": 8.395691609977323e-07, "loss": 0.0478, "num_tokens": 27099862.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 284, "step_time": 296.52208467107266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3984.0, "completions/mean_length": 1357.4375, "completions/mean_terminated_length": 850.2963256835938, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "entropy": 3.6308336928486824, "epoch": 0.16156462585034015, "frac_reward_zero_std": 0.625, "grad_norm": 0.2197265625, "kl": 0.0008931458942242898, "learning_rate": 8.39002267573696e-07, "loss": 0.0107, "num_tokens": 27198162.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 285, "step_time": 299.7050953200087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3696.0, "completions/mean_length": 1512.46875, "completions/mean_terminated_length": 1089.7091064453125, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "entropy": 4.266504943370819, "epoch": 0.1621315192743764, "frac_reward_zero_std": 0.875, "grad_norm": 0.125, "kl": 0.0014134251723589841, "learning_rate": 8.384353741496598e-07, "loss": 0.0412, "num_tokens": 27306912.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 286, "step_time": 298.54193136747926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3498.0, "completions/mean_length": 1383.46875, "completions/mean_terminated_length": 995.96435546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.3081691190600395, "epoch": 0.1626984126984127, "frac_reward_zero_std": 0.75, "grad_norm": 0.177734375, "kl": 0.0010290788322890876, "learning_rate": 8.378684807256235e-07, "loss": 0.0324, "num_tokens": 27407398.0, "reward": 0.109375, "reward_std": 0.41277948021888733, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 287, "step_time": 297.7804460711777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3605.0, "completions/mean_length": 1589.84375, "completions/mean_terminated_length": 888.1199951171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.3641773611307144, "epoch": 0.16326530612244897, "frac_reward_zero_std": 0.875, "grad_norm": 0.1630859375, "kl": 0.001085398758732481, "learning_rate": 8.373015873015873e-07, "loss": -0.0172, "num_tokens": 27521996.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 288, "step_time": 298.2156117670238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3755.0, "completions/mean_length": 1014.3125, "completions/mean_terminated_length": 695.5172119140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.941107463091612, "epoch": 0.16383219954648526, "frac_reward_zero_std": 0.625, "grad_norm": 0.36328125, "kl": 0.0012372971614240669, "learning_rate": 8.36734693877551e-07, "loss": 0.0165, "num_tokens": 27598744.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 289, "step_time": 294.38263658620417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3727.0, "completions/mean_length": 1210.796875, "completions/mean_terminated_length": 798.6250610351562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.147449314594269, "epoch": 0.16439909297052155, "frac_reward_zero_std": 0.75, "grad_norm": 0.2255859375, "kl": 0.0011653340752673103, "learning_rate": 8.361678004535147e-07, "loss": -0.0157, "num_tokens": 27687795.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 290, "step_time": 296.3157152188942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4025.0, "completions/mean_length": 1743.390625, "completions/mean_terminated_length": 1143.7059326171875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "entropy": 3.946227416396141, "epoch": 0.1649659863945578, "frac_reward_zero_std": 0.875, "grad_norm": 0.1328125, "kl": 0.0010434207342768786, "learning_rate": 8.356009070294783e-07, "loss": 0.028, "num_tokens": 27812332.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 291, "step_time": 309.8255847459659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3957.0, "completions/mean_length": 1133.796875, "completions/mean_terminated_length": 827.362060546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.052217453718185, "epoch": 0.1655328798185941, "frac_reward_zero_std": 0.875, "grad_norm": 0.1083984375, "kl": 0.0013799356638628524, "learning_rate": 8.350340136054421e-07, "loss": 0.0006, "num_tokens": 27900503.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 292, "step_time": 313.85175424348563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2972.0, "completions/mean_length": 1030.1875, "completions/mean_terminated_length": 713.0344848632812, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 3.4501704461872578, "epoch": 0.1660997732426304, "frac_reward_zero_std": 0.75, "grad_norm": 0.2080078125, "kl": 0.0015028614325274248, "learning_rate": 8.344671201814058e-07, "loss": 0.0217, "num_tokens": 27978755.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 293, "step_time": 302.06553180422634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3667.0, "completions/mean_length": 1097.328125, "completions/mean_terminated_length": 949.8524169921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.379862379282713, "epoch": 0.16666666666666666, "frac_reward_zero_std": 0.75, "grad_norm": 0.1943359375, "kl": 0.002225910628112615, "learning_rate": 8.339002267573696e-07, "loss": 0.0001, "num_tokens": 28063072.0, "reward": 0.15625, "reward_std": 0.5554563403129578, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 294, "step_time": 305.31014814879745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3855.0, "completions/mean_length": 1184.140625, "completions/mean_terminated_length": 768.1607666015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.5274267680943012, "epoch": 0.16723356009070295, "frac_reward_zero_std": 0.5, "grad_norm": 0.41796875, "kl": 0.0014518765474349493, "learning_rate": 8.333333333333333e-07, "loss": -0.1013, "num_tokens": 28150961.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 295, "step_time": 300.1005868501961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3222.0, "completions/mean_length": 1218.96875, "completions/mean_terminated_length": 865.6491088867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.963806878775358, "epoch": 0.16780045351473924, "frac_reward_zero_std": 0.375, "grad_norm": 0.435546875, "kl": 0.0015549707040918292, "learning_rate": 8.327664399092971e-07, "loss": -0.092, "num_tokens": 28240447.0, "reward": 0.203125, "reward_std": 0.5890451669692993, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 296, "step_time": 296.3705997876823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3883.0, "completions/mean_length": 1019.421875, "completions/mean_terminated_length": 814.3167114257812, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "entropy": 3.574819404631853, "epoch": 0.1683673469387755, "frac_reward_zero_std": 0.75, "grad_norm": 0.26953125, "kl": 0.0015025965749373427, "learning_rate": 8.321995464852607e-07, "loss": 0.1117, "num_tokens": 28317514.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 297, "step_time": 294.27986282110214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3922.0, "completions/mean_length": 1468.890625, "completions/mean_terminated_length": 923.6415405273438, "completions/min_length": 94.0, "completions/min_terminated_length": 94.0, "entropy": 3.008502796292305, "epoch": 0.1689342403628118, "frac_reward_zero_std": 0.875, "grad_norm": 0.166015625, "kl": 0.001358457573587657, "learning_rate": 8.316326530612245e-07, "loss": 0.0075, "num_tokens": 28423035.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 298, "step_time": 293.9276619926095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3777.0, "completions/mean_length": 1020.203125, "completions/mean_terminated_length": 642.4736938476562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.857605691999197, "epoch": 0.16950113378684808, "frac_reward_zero_std": 0.875, "grad_norm": 0.2109375, "kl": 0.0017286588372371625, "learning_rate": 8.310657596371881e-07, "loss": 0.0104, "num_tokens": 28500880.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 299, "step_time": 302.13770243152976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3957.0, "completions/mean_length": 1521.671875, "completions/mean_terminated_length": 1100.4180908203125, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "entropy": 4.238505080342293, "epoch": 0.17006802721088435, "frac_reward_zero_std": 0.875, "grad_norm": 0.140625, "kl": 0.001234137591382023, "learning_rate": 8.304988662131519e-07, "loss": 0.0313, "num_tokens": 28611323.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 300, "step_time": 302.78664544876665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3890.0, "completions/mean_length": 1194.28125, "completions/mean_terminated_length": 779.7500610351562, "completions/min_length": 69.0, "completions/min_terminated_length": 69.0, "entropy": 2.7570454962551594, "epoch": 0.17063492063492064, "frac_reward_zero_std": 0.625, "grad_norm": 0.361328125, "kl": 0.0012360909440758405, "learning_rate": 8.299319727891156e-07, "loss": -0.0846, "num_tokens": 28700133.0, "reward": 0.15625, "reward_std": 0.4787135720252991, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 301, "step_time": 297.30019121896476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3754.0, "completions/mean_length": 1068.0625, "completions/mean_terminated_length": 754.8275756835938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.0655543729662895, "epoch": 0.1712018140589569, "frac_reward_zero_std": 0.625, "grad_norm": 0.328125, "kl": 0.0014381433647940867, "learning_rate": 8.293650793650794e-07, "loss": 0.0318, "num_tokens": 28780617.0, "reward": 0.171875, "reward_std": 0.5719553828239441, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 302, "step_time": 294.7732902197167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3913.0, "completions/mean_length": 1324.046875, "completions/mean_terminated_length": 748.73583984375, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 2.493778932839632, "epoch": 0.1717687074829932, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "kl": 0.0013097982155159116, "learning_rate": 8.28798185941043e-07, "loss": 0.0305, "num_tokens": 28876980.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 303, "step_time": 295.6098387800157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4044.0, "completions/mean_length": 1439.65625, "completions/mean_terminated_length": 1060.1785888671875, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "entropy": 5.044021558016539, "epoch": 0.17233560090702948, "frac_reward_zero_std": 0.75, "grad_norm": 0.228515625, "kl": 0.0010676935780793428, "learning_rate": 8.282312925170068e-07, "loss": 0.0166, "num_tokens": 28978598.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 304, "step_time": 290.7202107226476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3698.0, "completions/mean_length": 1337.046875, "completions/mean_terminated_length": 764.4339599609375, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 3.115324232727289, "epoch": 0.17290249433106575, "frac_reward_zero_std": 0.75, "grad_norm": 0.248046875, "kl": 0.00165786925208522, "learning_rate": 8.276643990929705e-07, "loss": 0.0409, "num_tokens": 29076329.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 305, "step_time": 305.97279978822917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3989.0, "completions/mean_length": 1449.640625, "completions/mean_terminated_length": 959.5740966796875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.806913925334811, "epoch": 0.17346938775510204, "frac_reward_zero_std": 0.75, "grad_norm": 0.140625, "kl": 0.0014475634743575938, "learning_rate": 8.270975056689343e-07, "loss": -0.0037, "num_tokens": 29182658.0, "reward": 0.234375, "reward_std": 0.7125226259231567, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 306, "step_time": 304.92379070911556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3885.0, "completions/mean_length": 1247.125, "completions/mean_terminated_length": 1057.2000732421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.925425849854946, "epoch": 0.17403628117913833, "frac_reward_zero_std": 0.75, "grad_norm": 0.279296875, "kl": 0.0012668101971939905, "learning_rate": 8.26530612244898e-07, "loss": -0.0205, "num_tokens": 29274002.0, "reward": 0.15625, "reward_std": 0.5409793853759766, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 307, "step_time": 301.54330573696643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1325.75, "completions/mean_terminated_length": 1039.17236328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.7136062681674957, "epoch": 0.1746031746031746, "frac_reward_zero_std": 0.75, "grad_norm": 0.1337890625, "kl": 0.0014722158321092138, "learning_rate": 8.259637188208616e-07, "loss": 0.0019, "num_tokens": 29376322.0, "reward": 0.1796875, "reward_std": 0.5866720080375671, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 308, "step_time": 326.7536249374971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4049.0, "completions/mean_length": 1406.890625, "completions/mean_terminated_length": 908.9074096679688, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "entropy": 3.7738971263170242, "epoch": 0.17517006802721088, "frac_reward_zero_std": 0.625, "grad_norm": 0.2294921875, "kl": 0.0013544335824917653, "learning_rate": 8.253968253968253e-07, "loss": 0.0256, "num_tokens": 29478883.0, "reward": 0.109375, "reward_std": 0.46691107749938965, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 309, "step_time": 296.07146343030035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3803.0, "completions/mean_length": 1261.296875, "completions/mean_terminated_length": 1072.3167724609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.0862605310976505, "epoch": 0.17573696145124718, "frac_reward_zero_std": 0.625, "grad_norm": 0.359375, "kl": 0.0015789931640028954, "learning_rate": 8.24829931972789e-07, "loss": 0.0932, "num_tokens": 29571286.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 310, "step_time": 298.1583162872121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4042.0, "completions/mean_length": 920.015625, "completions/mean_terminated_length": 817.5645141601562, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 3.515986368060112, "epoch": 0.17630385487528344, "frac_reward_zero_std": 0.75, "grad_norm": 0.53125, "kl": 0.0011895883180841338, "learning_rate": 8.242630385487528e-07, "loss": -0.0069, "num_tokens": 29641503.0, "reward": 0.0859375, "reward_std": 0.28942590951919556, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 311, "step_time": 295.0998282721266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3785.0, "completions/mean_length": 1333.203125, "completions/mean_terminated_length": 821.5740966796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.287606105208397, "epoch": 0.17687074829931973, "frac_reward_zero_std": 0.5, "grad_norm": 0.32421875, "kl": 0.001888270318886498, "learning_rate": 8.236961451247165e-07, "loss": -0.0212, "num_tokens": 29738348.0, "reward": 0.296875, "reward_std": 0.69987952709198, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 312, "step_time": 297.20515647530556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3803.0, "completions/mean_length": 1015.25, "completions/mean_terminated_length": 754.1694946289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.613906405866146, "epoch": 0.17743764172335602, "frac_reward_zero_std": 0.5, "grad_norm": 0.376953125, "kl": 0.001490443704824429, "learning_rate": 8.231292517006803e-07, "loss": 0.0632, "num_tokens": 29814492.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 313, "step_time": 292.44761195778847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3079.0, "completions/mean_length": 1275.890625, "completions/mean_terminated_length": 753.6481323242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.8377242982387543, "epoch": 0.17800453514739228, "frac_reward_zero_std": 0.75, "grad_norm": 0.23828125, "kl": 0.0013854403405275661, "learning_rate": 8.22562358276644e-07, "loss": 0.0308, "num_tokens": 29907221.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 314, "step_time": 291.80689421482384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3869.0, "completions/mean_length": 1183.765625, "completions/mean_terminated_length": 826.122802734375, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 4.167344301939011, "epoch": 0.17857142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 0.2578125, "kl": 0.0011127866673632525, "learning_rate": 8.219954648526076e-07, "loss": 0.077, "num_tokens": 29995166.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 315, "step_time": 305.0239759385586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3820.0, "completions/mean_length": 1142.765625, "completions/mean_terminated_length": 780.0877075195312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 1.9886627234518528, "epoch": 0.17913832199546487, "frac_reward_zero_std": 0.875, "grad_norm": 0.240234375, "kl": 0.001561849749123212, "learning_rate": 8.214285714285713e-07, "loss": -0.0347, "num_tokens": 30081383.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 316, "step_time": 302.644008141011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3916.0, "completions/mean_length": 1109.390625, "completions/mean_terminated_length": 962.5081176757812, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 4.265066012740135, "epoch": 0.17970521541950113, "frac_reward_zero_std": 0.5, "grad_norm": 0.35546875, "kl": 0.0012646747200051323, "learning_rate": 8.208616780045351e-07, "loss": 0.0464, "num_tokens": 30164376.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 317, "step_time": 296.1302981870249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3752.0, "completions/mean_length": 800.109375, "completions/mean_terminated_length": 693.790283203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.901810046285391, "epoch": 0.18027210884353742, "frac_reward_zero_std": 0.625, "grad_norm": 0.5078125, "kl": 0.0015241189685184509, "learning_rate": 8.202947845804988e-07, "loss": -0.1508, "num_tokens": 30227879.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 318, "step_time": 300.9886961122975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3451.0, "completions/mean_length": 1135.765625, "completions/mean_terminated_length": 772.2280883789062, "completions/min_length": 88.0, "completions/min_terminated_length": 88.0, "entropy": 3.722517892718315, "epoch": 0.18083900226757368, "frac_reward_zero_std": 0.75, "grad_norm": 0.2109375, "kl": 0.0013234901507530594, "learning_rate": 8.197278911564626e-07, "loss": 0.0294, "num_tokens": 30311352.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 319, "step_time": 291.4153741868213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4090.0, "completions/mean_length": 1477.09375, "completions/mean_terminated_length": 992.1111450195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.208458241075277, "epoch": 0.18140589569160998, "frac_reward_zero_std": 0.875, "grad_norm": 0.11181640625, "kl": 0.0012379130166664254, "learning_rate": 8.191609977324263e-07, "loss": 0.0162, "num_tokens": 30417158.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 320, "step_time": 293.3395473388955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3407.0, "completions/mean_length": 969.90625, "completions/mean_terminated_length": 704.9830322265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.662312377244234, "epoch": 0.18197278911564627, "frac_reward_zero_std": 0.625, "grad_norm": 0.283203125, "kl": 0.001847629977419274, "learning_rate": 8.1859410430839e-07, "loss": 0.0102, "num_tokens": 30491000.0, "reward": 0.3046875, "reward_std": 0.6993034482002258, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 321, "step_time": 296.2531007146463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4051.0, "completions/mean_length": 795.203125, "completions/mean_terminated_length": 688.7257690429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.8032184205949306, "epoch": 0.18253968253968253, "frac_reward_zero_std": 0.625, "grad_norm": 0.37109375, "kl": 0.0014939455886633368, "learning_rate": 8.180272108843537e-07, "loss": 0.0062, "num_tokens": 30553837.0, "reward": 0.15625, "reward_std": 0.5409793853759766, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 322, "step_time": 300.56049086526036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2537.0, "completions/mean_length": 988.140625, "completions/mean_terminated_length": 666.637939453125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 1.7907675132155418, "epoch": 0.18310657596371882, "frac_reward_zero_std": 0.75, "grad_norm": 0.310546875, "kl": 0.00149022775440244, "learning_rate": 8.174603174603174e-07, "loss": 0.0241, "num_tokens": 30627958.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 323, "step_time": 292.3509076265618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4004.0, "completions/mean_length": 1345.296875, "completions/mean_terminated_length": 952.33935546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.6916390731930733, "epoch": 0.1836734693877551, "frac_reward_zero_std": 0.75, "grad_norm": 0.2275390625, "kl": 0.001267036413992173, "learning_rate": 8.168934240362811e-07, "loss": 0.0096, "num_tokens": 30724769.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 324, "step_time": 293.23362885788083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3688.0, "completions/mean_length": 1114.75, "completions/mean_terminated_length": 806.3448486328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.472703419625759, "epoch": 0.18424036281179137, "frac_reward_zero_std": 0.5, "grad_norm": 0.33984375, "kl": 0.0011231746357225347, "learning_rate": 8.163265306122449e-07, "loss": 0.0674, "num_tokens": 30808553.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 325, "step_time": 299.0592782208696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4014.0, "completions/mean_length": 1138.828125, "completions/mean_terminated_length": 775.6666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.44533871114254, "epoch": 0.18480725623582767, "frac_reward_zero_std": 0.75, "grad_norm": 0.341796875, "kl": 0.001377893022436183, "learning_rate": 8.157596371882086e-07, "loss": -0.0992, "num_tokens": 30893550.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 326, "step_time": 296.04104791674763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4055.0, "completions/mean_length": 1251.875, "completions/mean_terminated_length": 1010.8474731445312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.31509043648839, "epoch": 0.18537414965986396, "frac_reward_zero_std": 0.75, "grad_norm": 0.2353515625, "kl": 0.0012605721276486292, "learning_rate": 8.151927437641724e-07, "loss": 0.0133, "num_tokens": 30986430.0, "reward": 0.21875, "reward_std": 0.5482656359672546, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 327, "step_time": 305.77011601254344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3758.0, "completions/mean_length": 1622.890625, "completions/mean_terminated_length": 1319.1754150390625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.529617294669151, "epoch": 0.18594104308390022, "frac_reward_zero_std": 0.875, "grad_norm": 0.1474609375, "kl": 0.0010493327517906437, "learning_rate": 8.14625850340136e-07, "loss": 0.0473, "num_tokens": 31101519.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 328, "step_time": 292.69737570080906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4076.0, "completions/mean_length": 1463.6875, "completions/mean_terminated_length": 976.2222290039062, "completions/min_length": 137.0, "completions/min_terminated_length": 137.0, "entropy": 3.6420319005846977, "epoch": 0.1865079365079365, "frac_reward_zero_std": 0.75, "grad_norm": 0.2001953125, "kl": 0.001224763143909513, "learning_rate": 8.140589569160997e-07, "loss": 0.0244, "num_tokens": 31207643.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 329, "step_time": 296.2488160226494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4065.0, "completions/mean_length": 1161.46875, "completions/mean_terminated_length": 857.8965454101562, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.393191885203123, "epoch": 0.1870748299319728, "frac_reward_zero_std": 0.625, "grad_norm": 0.267578125, "kl": 0.0013154964290151838, "learning_rate": 8.134920634920635e-07, "loss": 0.0648, "num_tokens": 31293169.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 330, "step_time": 293.66058210935444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3771.0, "completions/mean_length": 1535.578125, "completions/mean_terminated_length": 1004.1697998046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 5.231051322072744, "epoch": 0.18764172335600907, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "kl": 0.001104443465010263, "learning_rate": 8.129251700680271e-07, "loss": 0.0174, "num_tokens": 31402286.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 331, "step_time": 292.3300448982045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3969.0, "completions/mean_length": 1272.453125, "completions/mean_terminated_length": 810.4181518554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.7793682869523764, "epoch": 0.18820861678004536, "frac_reward_zero_std": 0.875, "grad_norm": 0.04931640625, "kl": 0.0014106326598266605, "learning_rate": 8.123582766439909e-07, "loss": -0.0106, "num_tokens": 31498211.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 332, "step_time": 307.6453470662236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3822.0, "completions/mean_length": 1324.375, "completions/mean_terminated_length": 928.4285888671875, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "entropy": 2.3627333492040634, "epoch": 0.18877551020408162, "frac_reward_zero_std": 0.75, "grad_norm": 0.2421875, "kl": 0.001209826299600536, "learning_rate": 8.117913832199546e-07, "loss": -0.0133, "num_tokens": 31594515.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 333, "step_time": 295.49152721557766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3381.0, "completions/mean_length": 983.875, "completions/mean_terminated_length": 776.4000244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.2466306276619434, "epoch": 0.1893424036281179, "frac_reward_zero_std": 0.75, "grad_norm": 0.25390625, "kl": 0.0015662444275221787, "learning_rate": 8.112244897959183e-07, "loss": -0.0071, "num_tokens": 31668547.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 334, "step_time": 292.95015636552125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3820.0, "completions/mean_length": 1548.234375, "completions/mean_terminated_length": 960.2885131835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.4881539437919855, "epoch": 0.1899092970521542, "frac_reward_zero_std": 0.875, "grad_norm": 0.140625, "kl": 0.0013060569126537303, "learning_rate": 8.10657596371882e-07, "loss": 0.0143, "num_tokens": 31779666.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 335, "step_time": 293.80418958794326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3865.0, "completions/mean_length": 1332.578125, "completions/mean_terminated_length": 880.3817749023438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.352604014798999, "epoch": 0.19047619047619047, "frac_reward_zero_std": 0.75, "grad_norm": 0.189453125, "kl": 0.00127808340221236, "learning_rate": 8.100907029478458e-07, "loss": 0.0419, "num_tokens": 31875927.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 336, "step_time": 293.79132070112973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3297.0, "completions/mean_length": 1400.984375, "completions/mean_terminated_length": 901.9074096679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.137748945504427, "epoch": 0.19104308390022676, "frac_reward_zero_std": 0.75, "grad_norm": 0.166015625, "kl": 0.0011957738133787643, "learning_rate": 8.095238095238095e-07, "loss": 0.0034, "num_tokens": 31977318.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 337, "step_time": 300.70368077233434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3987.0, "completions/mean_length": 1183.015625, "completions/mean_terminated_length": 766.8750610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.303982373327017, "epoch": 0.19160997732426305, "frac_reward_zero_std": 0.625, "grad_norm": 0.265625, "kl": 0.0018087913922499865, "learning_rate": 8.089569160997733e-07, "loss": 0.0331, "num_tokens": 32066511.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 338, "step_time": 298.1635686745867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2989.0, "completions/mean_length": 872.140625, "completions/mean_terminated_length": 657.2166748046875, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "entropy": 2.553455151617527, "epoch": 0.1921768707482993, "frac_reward_zero_std": 0.75, "grad_norm": 0.353515625, "kl": 0.0014670600940007716, "learning_rate": 8.083900226757369e-07, "loss": 0.0204, "num_tokens": 32135224.0, "reward": 0.0859375, "reward_std": 0.28942590951919556, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 339, "step_time": 305.7137638395652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3732.0, "completions/mean_length": 1285.421875, "completions/mean_terminated_length": 702.0943603515625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.686339009553194, "epoch": 0.1927437641723356, "frac_reward_zero_std": 0.875, "grad_norm": 0.11181640625, "kl": 0.0014709026254422497, "learning_rate": 8.078231292517006e-07, "loss": 0.0062, "num_tokens": 32229171.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 340, "step_time": 294.1925411038101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3938.0, "completions/mean_length": 1565.140625, "completions/mean_terminated_length": 920.0196533203125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.788554474711418, "epoch": 0.1933106575963719, "frac_reward_zero_std": 0.875, "grad_norm": 0.1015625, "kl": 0.0017092519883590285, "learning_rate": 8.072562358276643e-07, "loss": 0.0042, "num_tokens": 32343316.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 341, "step_time": 300.09620239492506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2418.0, "completions/mean_length": 1192.125, "completions/mean_terminated_length": 716.9454345703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 1.80996223539114, "epoch": 0.19387755102040816, "frac_reward_zero_std": 0.75, "grad_norm": 0.2216796875, "kl": 0.001654357838560827, "learning_rate": 8.066893424036281e-07, "loss": 0.0122, "num_tokens": 32430700.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 342, "step_time": 296.0148100089282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3730.0, "completions/mean_length": 1481.6875, "completions/mean_terminated_length": 878.3846435546875, "completions/min_length": 99.0, "completions/min_terminated_length": 99.0, "entropy": 2.1769518852233887, "epoch": 0.19444444444444445, "frac_reward_zero_std": 0.875, "grad_norm": 0.14453125, "kl": 0.00137901808557217, "learning_rate": 8.061224489795918e-07, "loss": -0.0012, "num_tokens": 32537680.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 343, "step_time": 296.50169356074184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4093.0, "completions/mean_length": 1020.703125, "completions/mean_terminated_length": 760.084716796875, "completions/min_length": 65.0, "completions/min_terminated_length": 65.0, "entropy": 2.139278154820204, "epoch": 0.19501133786848074, "frac_reward_zero_std": 0.625, "grad_norm": 0.337890625, "kl": 0.0013905934702052036, "learning_rate": 8.055555555555556e-07, "loss": 0.0395, "num_tokens": 32614197.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 344, "step_time": 295.34965351130813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3910.0, "completions/mean_length": 1866.703125, "completions/mean_terminated_length": 1548.232177734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 6.025232031941414, "epoch": 0.195578231292517, "frac_reward_zero_std": 0.875, "grad_norm": 0.107421875, "kl": 0.0009480748649366433, "learning_rate": 8.049886621315193e-07, "loss": -0.0161, "num_tokens": 32747778.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 345, "step_time": 301.9893612070009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3226.0, "completions/mean_length": 1214.71875, "completions/mean_terminated_length": 681.1481323242188, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.4195906836539507, "epoch": 0.1961451247165533, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "kl": 0.0014571854117093608, "learning_rate": 8.04421768707483e-07, "loss": 0.0527, "num_tokens": 32837936.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 346, "step_time": 296.6905417237431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3731.0, "completions/mean_length": 1326.546875, "completions/mean_terminated_length": 930.9107666015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.9649287778884172, "epoch": 0.19671201814058956, "frac_reward_zero_std": 0.625, "grad_norm": 0.25, "kl": 0.0010013702285505133, "learning_rate": 8.038548752834466e-07, "loss": 0.0383, "num_tokens": 32932835.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 347, "step_time": 290.2288642572239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3716.0, "completions/mean_length": 1499.609375, "completions/mean_terminated_length": 1018.7963256835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.093114998191595, "epoch": 0.19727891156462585, "frac_reward_zero_std": 0.875, "grad_norm": 0.14453125, "kl": 0.0013866425415471895, "learning_rate": 8.032879818594103e-07, "loss": 0.0199, "num_tokens": 33039826.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 348, "step_time": 292.7541776187718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4085.0, "completions/mean_length": 1253.203125, "completions/mean_terminated_length": 847.0892944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.132527057081461, "epoch": 0.19784580498866214, "frac_reward_zero_std": 0.875, "grad_norm": 0.1337890625, "kl": 0.001471931976993801, "learning_rate": 8.027210884353741e-07, "loss": 0.0275, "num_tokens": 33132607.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 349, "step_time": 302.33250446710736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3581.0, "completions/mean_length": 1014.09375, "completions/mean_terminated_length": 808.6333618164062, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.1049369759857655, "epoch": 0.1984126984126984, "frac_reward_zero_std": 0.625, "grad_norm": 0.3203125, "kl": 0.0014407995804504026, "learning_rate": 8.021541950113378e-07, "loss": 0.0057, "num_tokens": 33214365.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 350, "step_time": 343.75238325074315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3666.0, "completions/mean_length": 1014.90625, "completions/mean_terminated_length": 696.1724243164062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.3992355689406395, "epoch": 0.1989795918367347, "frac_reward_zero_std": 0.5, "grad_norm": 0.5, "kl": 0.0015600469305354636, "learning_rate": 8.015873015873016e-07, "loss": -0.0205, "num_tokens": 33291855.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 351, "step_time": 309.2162681380287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3978.0, "completions/mean_length": 938.9375, "completions/mean_terminated_length": 837.0967407226562, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.9943065233528614, "epoch": 0.19954648526077098, "frac_reward_zero_std": 0.75, "grad_norm": 0.2412109375, "kl": 0.0011394561715860618, "learning_rate": 8.010204081632652e-07, "loss": 0.0025, "num_tokens": 33362971.0, "reward": 0.15625, "reward_std": 0.5409793853759766, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 352, "step_time": 294.9366546738893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4009.0, "completions/mean_length": 1198.15625, "completions/mean_terminated_length": 898.3793334960938, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.235837150365114, "epoch": 0.20011337868480725, "frac_reward_zero_std": 0.875, "grad_norm": 0.171875, "kl": 0.0013127148977218894, "learning_rate": 8.00453514739229e-07, "loss": 0.0103, "num_tokens": 33451005.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 353, "step_time": 300.2821004744619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4045.0, "completions/mean_length": 1028.609375, "completions/mean_terminated_length": 768.6610107421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.8050526566803455, "epoch": 0.20068027210884354, "frac_reward_zero_std": 0.625, "grad_norm": 0.224609375, "kl": 0.0013078464253339916, "learning_rate": 7.998866213151927e-07, "loss": 0.0162, "num_tokens": 33528148.0, "reward": 0.203125, "reward_std": 0.5754139423370361, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 354, "step_time": 295.9015074791387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3844.0, "completions/mean_length": 1436.625, "completions/mean_terminated_length": 822.923095703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.596452224999666, "epoch": 0.20124716553287983, "frac_reward_zero_std": 0.875, "grad_norm": 0.142578125, "kl": 0.0013235371661721729, "learning_rate": 7.993197278911564e-07, "loss": 0.0163, "num_tokens": 33630372.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 355, "step_time": 291.1839153263718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3704.0, "completions/mean_length": 1521.9375, "completions/mean_terminated_length": 1045.25927734375, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.9435150884091854, "epoch": 0.2018140589569161, "frac_reward_zero_std": 0.75, "grad_norm": 0.1943359375, "kl": 0.0012276593606657116, "learning_rate": 7.987528344671201e-07, "loss": 0.0343, "num_tokens": 33743432.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 356, "step_time": 320.05601400416344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3949.0, "completions/mean_length": 1252.609375, "completions/mean_terminated_length": 958.4655151367188, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 3.4948107935488224, "epoch": 0.20238095238095238, "frac_reward_zero_std": 0.625, "grad_norm": 0.26953125, "kl": 0.0015216540377878118, "learning_rate": 7.981859410430839e-07, "loss": 0.0455, "num_tokens": 33833911.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 357, "step_time": 293.9998562326655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3068.0, "completions/mean_length": 1035.0, "completions/mean_terminated_length": 775.5932006835938, "completions/min_length": 104.0, "completions/min_terminated_length": 104.0, "entropy": 2.5392488501966, "epoch": 0.20294784580498867, "frac_reward_zero_std": 0.5, "grad_norm": 0.255859375, "kl": 0.00149408013021457, "learning_rate": 7.976190476190476e-07, "loss": -0.0005, "num_tokens": 33912631.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 358, "step_time": 306.4622711185366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3355.0, "completions/mean_length": 815.796875, "completions/mean_terminated_length": 709.9838256835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.474315457046032, "epoch": 0.20351473922902494, "frac_reward_zero_std": 0.5, "grad_norm": 0.474609375, "kl": 0.0015659075288567692, "learning_rate": 7.970521541950113e-07, "loss": 0.1007, "num_tokens": 33976834.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 359, "step_time": 299.45716447010636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4089.0, "completions/mean_length": 1094.6875, "completions/mean_terminated_length": 894.6000366210938, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "entropy": 3.4159761611372232, "epoch": 0.20408163265306123, "frac_reward_zero_std": 0.625, "grad_norm": 0.283203125, "kl": 0.0012001038285234245, "learning_rate": 7.96485260770975e-07, "loss": 0.0838, "num_tokens": 34057582.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 360, "step_time": 295.3705814424902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3910.0, "completions/mean_length": 1578.796875, "completions/mean_terminated_length": 1112.648193359375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.302098356187344, "epoch": 0.2046485260770975, "frac_reward_zero_std": 0.875, "grad_norm": 0.09814453125, "kl": 0.0013481216483341996, "learning_rate": 7.959183673469388e-07, "loss": 0.01, "num_tokens": 34170953.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 361, "step_time": 299.9199784323573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4005.0, "completions/mean_length": 1218.203125, "completions/mean_terminated_length": 920.5, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "entropy": 3.2134172283113003, "epoch": 0.20521541950113378, "frac_reward_zero_std": 0.625, "grad_norm": 0.279296875, "kl": 0.0012985770554223564, "learning_rate": 7.953514739229025e-07, "loss": 0.0405, "num_tokens": 34260014.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 362, "step_time": 292.89271545317024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4052.0, "completions/mean_length": 1267.3125, "completions/mean_terminated_length": 804.4363403320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.3208404891192913, "epoch": 0.20578231292517007, "frac_reward_zero_std": 0.5, "grad_norm": 0.28515625, "kl": 0.0013819267478538677, "learning_rate": 7.947845804988662e-07, "loss": 0.0248, "num_tokens": 34352338.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 363, "step_time": 294.02128049358726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3690.0, "completions/mean_length": 1085.859375, "completions/mean_terminated_length": 716.1929931640625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.4948936328291893, "epoch": 0.20634920634920634, "frac_reward_zero_std": 0.5, "grad_norm": 0.302734375, "kl": 0.001578690076712519, "learning_rate": 7.942176870748299e-07, "loss": 0.0191, "num_tokens": 34435649.0, "reward": 0.1953125, "reward_std": 0.5883605480194092, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 364, "step_time": 304.2129575004801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4053.0, "completions/mean_length": 1091.1875, "completions/mean_terminated_length": 780.3448486328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.3134211003780365, "epoch": 0.20691609977324263, "frac_reward_zero_std": 0.5, "grad_norm": 0.341796875, "kl": 0.001404934359015897, "learning_rate": 7.936507936507936e-07, "loss": 0.0356, "num_tokens": 34515757.0, "reward": 0.2734375, "reward_std": 0.6662480235099792, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 365, "step_time": 292.800573060289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3431.0, "completions/mean_length": 963.0, "completions/mean_terminated_length": 754.1333618164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.9491483457386494, "epoch": 0.20748299319727892, "frac_reward_zero_std": 0.5, "grad_norm": 0.412109375, "kl": 0.0019139847026963253, "learning_rate": 7.930839002267573e-07, "loss": 0.0073, "num_tokens": 34591013.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 366, "step_time": 306.374175994657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3739.0, "completions/mean_length": 1556.3125, "completions/mean_terminated_length": 970.2308349609375, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "entropy": 3.646450661122799, "epoch": 0.20804988662131518, "frac_reward_zero_std": 0.875, "grad_norm": 0.1650390625, "kl": 0.001971228062757291, "learning_rate": 7.92517006802721e-07, "loss": -0.0172, "num_tokens": 34702361.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 367, "step_time": 296.2610024763271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3948.0, "completions/mean_length": 1494.671875, "completions/mean_terminated_length": 831.5882568359375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.0113005340099335, "epoch": 0.20861678004535147, "frac_reward_zero_std": 0.75, "grad_norm": 0.1630859375, "kl": 0.001619795017177239, "learning_rate": 7.919501133786848e-07, "loss": 0.0211, "num_tokens": 34810428.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 368, "step_time": 297.6598177989945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4086.0, "completions/mean_length": 1478.078125, "completions/mean_terminated_length": 1104.08935546875, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "entropy": 3.442749384790659, "epoch": 0.20918367346938777, "frac_reward_zero_std": 0.75, "grad_norm": 0.1806640625, "kl": 0.0011501683402457274, "learning_rate": 7.913832199546485e-07, "loss": 0.0666, "num_tokens": 34916737.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 369, "step_time": 296.4353799261153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3747.0, "completions/mean_length": 1408.828125, "completions/mean_terminated_length": 969.1090698242188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.96785619109869, "epoch": 0.20975056689342403, "frac_reward_zero_std": 0.625, "grad_norm": 0.267578125, "kl": 0.0015126020189200062, "learning_rate": 7.908163265306123e-07, "loss": 0.0936, "num_tokens": 35019102.0, "reward": 0.359375, "reward_std": 0.8090203404426575, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 370, "step_time": 295.2090236181393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3804.0, "completions/mean_length": 830.734375, "completions/mean_terminated_length": 613.050048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1338421292603016, "epoch": 0.21031746031746032, "frac_reward_zero_std": 0.875, "grad_norm": 0.27734375, "kl": 0.0019984587888757233, "learning_rate": 7.902494331065758e-07, "loss": 0.0544, "num_tokens": 35084013.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 371, "step_time": 294.6177222598344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3749.0, "completions/mean_length": 1224.578125, "completions/mean_terminated_length": 754.7090454101562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.497921671718359, "epoch": 0.2108843537414966, "frac_reward_zero_std": 0.625, "grad_norm": 0.25390625, "kl": 0.0022498509297292912, "learning_rate": 7.896825396825396e-07, "loss": 0.0223, "num_tokens": 35172866.0, "reward": 0.15625, "reward_std": 0.5261042714118958, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 372, "step_time": 292.2975083729252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3942.0, "completions/mean_length": 1154.328125, "completions/mean_terminated_length": 734.0892944335938, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.794364996254444, "epoch": 0.21145124716553287, "frac_reward_zero_std": 0.625, "grad_norm": 0.248046875, "kl": 0.001845422304540989, "learning_rate": 7.891156462585033e-07, "loss": -0.0219, "num_tokens": 35257919.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 373, "step_time": 293.63794895634055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3577.0, "completions/mean_length": 1372.796875, "completions/mean_terminated_length": 744.3654174804688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4559489972889423, "epoch": 0.21201814058956917, "frac_reward_zero_std": 0.75, "grad_norm": 0.21484375, "kl": 0.001547759529785253, "learning_rate": 7.885487528344671e-07, "loss": 0.0102, "num_tokens": 35357138.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 374, "step_time": 296.37745548412204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3983.0, "completions/mean_length": 1502.6875, "completions/mean_terminated_length": 964.4528198242188, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "entropy": 3.559647735208273, "epoch": 0.21258503401360543, "frac_reward_zero_std": 0.875, "grad_norm": 0.1298828125, "kl": 0.0020263616788724903, "learning_rate": 7.879818594104308e-07, "loss": 0.0046, "num_tokens": 35464750.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 375, "step_time": 294.14539925567806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3554.0, "completions/mean_length": 1158.1875, "completions/mean_terminated_length": 738.5000610351562, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "entropy": 2.1302271597087383, "epoch": 0.21315192743764172, "frac_reward_zero_std": 0.5, "grad_norm": 0.328125, "kl": 0.0016697567079972941, "learning_rate": 7.874149659863946e-07, "loss": 0.0032, "num_tokens": 35550202.0, "reward": 0.1953125, "reward_std": 0.5883605480194092, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 376, "step_time": 294.99223101790994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3483.0, "completions/mean_length": 1143.125, "completions/mean_terminated_length": 892.88134765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.006809718906879, "epoch": 0.213718820861678, "frac_reward_zero_std": 0.75, "grad_norm": 0.1943359375, "kl": 0.0018984173857461428, "learning_rate": 7.868480725623582e-07, "loss": 0.0186, "num_tokens": 35633106.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 377, "step_time": 290.6259912196547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2856.0, "completions/mean_length": 1040.265625, "completions/mean_terminated_length": 724.1551513671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1822089068591595, "epoch": 0.21428571428571427, "frac_reward_zero_std": 0.5, "grad_norm": 0.353515625, "kl": 0.0018458862286934163, "learning_rate": 7.86281179138322e-07, "loss": 0.0425, "num_tokens": 35711123.0, "reward": 0.1875, "reward_std": 0.5455447435379028, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 378, "step_time": 294.7133179754019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4025.0, "completions/mean_length": 1367.15625, "completions/mean_terminated_length": 920.6181640625, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.002276659011841, "epoch": 0.21485260770975056, "frac_reward_zero_std": 0.625, "grad_norm": 0.3046875, "kl": 0.0013462407514452934, "learning_rate": 7.857142857142856e-07, "loss": 0.0192, "num_tokens": 35811501.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 379, "step_time": 299.2446444295347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4073.0, "completions/mean_length": 1600.015625, "completions/mean_terminated_length": 963.7843627929688, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 3.1706656143069267, "epoch": 0.21541950113378686, "frac_reward_zero_std": 0.625, "grad_norm": 0.208984375, "kl": 0.0026700084927142598, "learning_rate": 7.851473922902494e-07, "loss": -0.0208, "num_tokens": 35929654.0, "reward": 0.296875, "reward_std": 0.6649904847145081, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 380, "step_time": 308.2830661116168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4096.0, "completions/mean_length": 1219.4375, "completions/mean_terminated_length": 921.862060546875, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "entropy": 3.471789874136448, "epoch": 0.21598639455782312, "frac_reward_zero_std": 0.625, "grad_norm": 0.27734375, "kl": 0.0013272836004034616, "learning_rate": 7.845804988662131e-07, "loss": 0.0727, "num_tokens": 36019194.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 381, "step_time": 307.0177404200658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4085.0, "completions/mean_length": 809.609375, "completions/mean_terminated_length": 647.9835815429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.630410388112068, "epoch": 0.2165532879818594, "frac_reward_zero_std": 0.5, "grad_norm": 0.453125, "kl": 0.0020975194602215197, "learning_rate": 7.840136054421769e-07, "loss": 0.1147, "num_tokens": 36082337.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 382, "step_time": 295.45017794612795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3947.0, "completions/mean_length": 1296.328125, "completions/mean_terminated_length": 952.5087890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 4.066781774163246, "epoch": 0.2171201814058957, "frac_reward_zero_std": 0.625, "grad_norm": 0.2236328125, "kl": 0.0011764016635424923, "learning_rate": 7.834467120181405e-07, "loss": 0.0375, "num_tokens": 36177758.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 383, "step_time": 306.97376155201346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3918.0, "completions/mean_length": 1130.859375, "completions/mean_terminated_length": 766.7192993164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 2.8013731352984905, "epoch": 0.21768707482993196, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "kl": 0.001920601582241943, "learning_rate": 7.828798185941043e-07, "loss": -0.0019, "num_tokens": 36263821.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 384, "step_time": 302.87433452066034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3958.0, "completions/mean_length": 1097.515625, "completions/mean_terminated_length": 787.3275756835938, "completions/min_length": 105.0, "completions/min_terminated_length": 105.0, "entropy": 3.3111561834812164, "epoch": 0.21825396825396826, "frac_reward_zero_std": 0.625, "grad_norm": 0.291015625, "kl": 0.0017533823083795141, "learning_rate": 7.82312925170068e-07, "loss": 0.0183, "num_tokens": 36345678.0, "reward": 0.3359375, "reward_std": 0.7295067310333252, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 385, "step_time": 301.5573451537639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3591.0, "completions/mean_length": 1211.484375, "completions/mean_terminated_length": 612.8113403320312, "completions/min_length": 116.0, "completions/min_terminated_length": 116.0, "entropy": 2.12896141782403, "epoch": 0.21882086167800455, "frac_reward_zero_std": 0.75, "grad_norm": 0.1845703125, "kl": 0.002257652839034563, "learning_rate": 7.817460317460318e-07, "loss": 0.0186, "num_tokens": 36439149.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 386, "step_time": 322.2181938337162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3369.0, "completions/mean_length": 1139.0, "completions/mean_terminated_length": 888.4067993164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.556562624871731, "epoch": 0.2193877551020408, "frac_reward_zero_std": 0.625, "grad_norm": 0.2890625, "kl": 0.002096428152071894, "learning_rate": 7.811791383219954e-07, "loss": 0.0817, "num_tokens": 36525397.0, "reward": 0.1953125, "reward_std": 0.5883605480194092, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 387, "step_time": 301.1546203037724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4021.0, "completions/mean_length": 1397.546875, "completions/mean_terminated_length": 1066.157958984375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 3.776941567659378, "epoch": 0.2199546485260771, "frac_reward_zero_std": 0.625, "grad_norm": 0.236328125, "kl": 0.0012676407204708084, "learning_rate": 7.806122448979591e-07, "loss": 0.0388, "num_tokens": 36626040.0, "reward": 0.2734375, "reward_std": 0.6662480235099792, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 388, "step_time": 295.3469817014411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3844.0, "completions/mean_length": 1362.78125, "completions/mean_terminated_length": 795.5094604492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.865013089030981, "epoch": 0.2205215419501134, "frac_reward_zero_std": 0.625, "grad_norm": 0.2265625, "kl": 0.00144556881787139, "learning_rate": 7.800453514739229e-07, "loss": 0.0219, "num_tokens": 36726586.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 389, "step_time": 296.1360622383654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3881.0, "completions/mean_length": 1110.765625, "completions/mean_terminated_length": 911.7500610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1485432758927345, "epoch": 0.22108843537414966, "frac_reward_zero_std": 0.75, "grad_norm": 0.21875, "kl": 0.0012700370098173153, "learning_rate": 7.794784580498865e-07, "loss": -0.0158, "num_tokens": 36808931.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 390, "step_time": 293.5197555180639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4037.0, "completions/mean_length": 798.921875, "completions/mean_terminated_length": 579.11669921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.0194799602031708, "epoch": 0.22165532879818595, "frac_reward_zero_std": 0.5, "grad_norm": 0.404296875, "kl": 0.0020651695485867094, "learning_rate": 7.789115646258503e-07, "loss": 0.0445, "num_tokens": 36871334.0, "reward": 0.25, "reward_std": 0.6299408078193665, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 391, "step_time": 292.6065103439614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3644.0, "completions/mean_length": 887.375, "completions/mean_terminated_length": 729.57373046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.4883970953524113, "epoch": 0.2222222222222222, "frac_reward_zero_std": 0.75, "grad_norm": 0.287109375, "kl": 0.0020892868415103294, "learning_rate": 7.78344671201814e-07, "loss": 0.0372, "num_tokens": 36939526.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 392, "step_time": 292.00554332882166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4071.0, "completions/mean_length": 1419.390625, "completions/mean_terminated_length": 1037.0179443359375, "completions/min_length": 116.0, "completions/min_terminated_length": 116.0, "entropy": 4.893246464431286, "epoch": 0.2227891156462585, "frac_reward_zero_std": 0.625, "grad_norm": 0.30859375, "kl": 0.001233271841556416, "learning_rate": 7.777777777777778e-07, "loss": -0.0503, "num_tokens": 37041447.0, "reward": 0.1875, "reward_std": 0.6267831921577454, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 393, "step_time": 295.95994062069803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3812.0, "completions/mean_length": 1084.953125, "completions/mean_terminated_length": 936.8687744140625, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 3.491551138460636, "epoch": 0.2233560090702948, "frac_reward_zero_std": 0.5, "grad_norm": 0.39453125, "kl": 0.0019169310544384643, "learning_rate": 7.772108843537415e-07, "loss": 0.0128, "num_tokens": 37123420.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 394, "step_time": 300.3087148098275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3717.0, "completions/mean_length": 1581.21875, "completions/mean_terminated_length": 940.1961059570312, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 4.595789223909378, "epoch": 0.22392290249433106, "frac_reward_zero_std": 0.75, "grad_norm": 0.140625, "kl": 0.0017633351071708603, "learning_rate": 7.766439909297052e-07, "loss": 0.012, "num_tokens": 37236626.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 395, "step_time": 300.34654100984335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3981.0, "completions/mean_length": 1304.671875, "completions/mean_terminated_length": 1015.913818359375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 4.356452081352472, "epoch": 0.22448979591836735, "frac_reward_zero_std": 0.5, "grad_norm": 0.271484375, "kl": 0.0014264194014685927, "learning_rate": 7.760770975056688e-07, "loss": 0.0583, "num_tokens": 37332269.0, "reward": 0.265625, "reward_std": 0.6782841682434082, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 396, "step_time": 303.81499348487705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3467.0, "completions/mean_length": 720.453125, "completions/mean_terminated_length": 666.873046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.8045205511152744, "epoch": 0.22505668934240364, "frac_reward_zero_std": 0.75, "grad_norm": 0.23828125, "kl": 0.0016409798354288796, "learning_rate": 7.755102040816326e-07, "loss": 0.0154, "num_tokens": 37388594.0, "reward": 0.140625, "reward_std": 0.4753340482711792, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 397, "step_time": 293.10425563156605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4065.0, "completions/mean_length": 1258.9375, "completions/mean_terminated_length": 853.6428833007812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.4828469827771187, "epoch": 0.2256235827664399, "frac_reward_zero_std": 0.625, "grad_norm": 0.306640625, "kl": 0.0014735936601937283, "learning_rate": 7.749433106575963e-07, "loss": 0.0701, "num_tokens": 37482614.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 398, "step_time": 299.26312240678817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3801.0, "completions/mean_length": 1303.6875, "completions/mean_terminated_length": 1014.8275756835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.619399968534708, "epoch": 0.2261904761904762, "frac_reward_zero_std": 0.625, "grad_norm": 0.2890625, "kl": 0.001796510121494066, "learning_rate": 7.743764172335601e-07, "loss": -0.0496, "num_tokens": 37577410.0, "reward": 0.2109375, "reward_std": 0.6412084698677063, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 399, "step_time": 295.9423144198954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3935.0, "completions/mean_length": 941.234375, "completions/mean_terminated_length": 786.0819091796875, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 2.890203081071377, "epoch": 0.22675736961451248, "frac_reward_zero_std": 0.75, "grad_norm": 0.29296875, "kl": 0.0014154092605167534, "learning_rate": 7.738095238095238e-07, "loss": 0.0658, "num_tokens": 37647273.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 400, "step_time": 288.85523932240903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2998.0, "completions/mean_length": 1129.484375, "completions/mean_terminated_length": 822.6034545898438, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.657008931040764, "epoch": 0.22732426303854875, "frac_reward_zero_std": 0.75, "grad_norm": 0.205078125, "kl": 0.0014798003430769313, "learning_rate": 7.732426303854876e-07, "loss": 0.012, "num_tokens": 37732472.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 401, "step_time": 298.95484398212284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3958.0, "completions/mean_length": 1498.96875, "completions/mean_terminated_length": 1018.0370483398438, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 3.3186546824872494, "epoch": 0.22789115646258504, "frac_reward_zero_std": 0.5, "grad_norm": 0.283203125, "kl": 0.0013445516051433515, "learning_rate": 7.726757369614512e-07, "loss": 0.0539, "num_tokens": 37842662.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 402, "step_time": 302.0707771368325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3225.0, "completions/mean_length": 1102.53125, "completions/mean_terminated_length": 612.69091796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.0880488380789757, "epoch": 0.22845804988662133, "frac_reward_zero_std": 0.75, "grad_norm": 0.22265625, "kl": 0.0020504709355009254, "learning_rate": 7.721088435374149e-07, "loss": 0.0082, "num_tokens": 37923744.0, "reward": 0.203125, "reward_std": 0.5890451669692993, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 403, "step_time": 292.8356831539422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3901.0, "completions/mean_length": 1096.15625, "completions/mean_terminated_length": 896.1666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1105887927114964, "epoch": 0.2290249433106576, "frac_reward_zero_std": 0.375, "grad_norm": 0.423828125, "kl": 0.001272784222237533, "learning_rate": 7.715419501133786e-07, "loss": 0.0298, "num_tokens": 38006754.0, "reward": 0.265625, "reward_std": 0.6544642448425293, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 404, "step_time": 304.97734472341835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3833.0, "completions/mean_length": 1515.21875, "completions/mean_terminated_length": 857.37255859375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "entropy": 4.143358360975981, "epoch": 0.22959183673469388, "frac_reward_zero_std": 0.75, "grad_norm": 0.1767578125, "kl": 0.002094690022204304, "learning_rate": 7.709750566893424e-07, "loss": 0.0119, "num_tokens": 38117344.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 405, "step_time": 297.9922364205122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3902.0, "completions/mean_length": 974.28125, "completions/mean_terminated_length": 873.5806274414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.750041037797928, "epoch": 0.23015873015873015, "frac_reward_zero_std": 0.625, "grad_norm": 0.353515625, "kl": 0.0013144648346496979, "learning_rate": 7.704081632653061e-07, "loss": 0.01, "num_tokens": 38189906.0, "reward": 0.2578125, "reward_std": 0.6784212589263916, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 406, "step_time": 294.31997813936323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4061.0, "completions/mean_length": 1544.671875, "completions/mean_terminated_length": 1015.1509399414062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.97406118363142, "epoch": 0.23072562358276644, "frac_reward_zero_std": 0.625, "grad_norm": 0.22265625, "kl": 0.001828713609938859, "learning_rate": 7.698412698412699e-07, "loss": 0.0481, "num_tokens": 38301269.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 407, "step_time": 298.18998036999255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4067.0, "completions/mean_length": 1311.453125, "completions/mean_terminated_length": 913.6607666015625, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "entropy": 3.1563703641295433, "epoch": 0.23129251700680273, "frac_reward_zero_std": 0.625, "grad_norm": 0.248046875, "kl": 0.001512056820502039, "learning_rate": 7.692743764172335e-07, "loss": 0.0162, "num_tokens": 38395658.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 408, "step_time": 292.98286969866604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3982.0, "completions/mean_length": 1415.984375, "completions/mean_terminated_length": 1033.125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.2954092267900705, "epoch": 0.231859410430839, "frac_reward_zero_std": 0.625, "grad_norm": 0.298828125, "kl": 0.0015296718975150725, "learning_rate": 7.687074829931972e-07, "loss": -0.076, "num_tokens": 38498889.0, "reward": 0.234375, "reward_std": 0.6664806604385376, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 409, "step_time": 296.84177739638835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3918.0, "completions/mean_length": 1194.25, "completions/mean_terminated_length": 1000.800048828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.125131338834763, "epoch": 0.23242630385487528, "frac_reward_zero_std": 0.5, "grad_norm": 0.328125, "kl": 0.0014890253532939823, "learning_rate": 7.68140589569161e-07, "loss": 0.0111, "num_tokens": 38589425.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 410, "step_time": 318.92869498115033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3369.0, "completions/mean_length": 933.71875, "completions/mean_terminated_length": 665.7288208007812, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.0793898813426495, "epoch": 0.23299319727891157, "frac_reward_zero_std": 0.75, "grad_norm": 0.322265625, "kl": 0.001860350399510935, "learning_rate": 7.675736961451246e-07, "loss": 0.0908, "num_tokens": 38661255.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 411, "step_time": 300.06950910016894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3551.0, "completions/mean_length": 1163.859375, "completions/mean_terminated_length": 803.7719116210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.3974614329636097, "epoch": 0.23356009070294784, "frac_reward_zero_std": 0.75, "grad_norm": 0.2421875, "kl": 0.002429354473861167, "learning_rate": 7.670068027210884e-07, "loss": -0.0233, "num_tokens": 38746166.0, "reward": 0.2890625, "reward_std": 0.7549029588699341, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 412, "step_time": 293.9550670655444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4083.0, "completions/mean_length": 1285.171875, "completions/mean_terminated_length": 825.2181396484375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.517598506063223, "epoch": 0.23412698412698413, "frac_reward_zero_std": 0.625, "grad_norm": 0.2578125, "kl": 0.0017115061400545528, "learning_rate": 7.664399092970521e-07, "loss": 0.0288, "num_tokens": 38838617.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 413, "step_time": 290.1487058866769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3951.0, "completions/mean_length": 1313.15625, "completions/mean_terminated_length": 857.7817993164062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.6312940698117018, "epoch": 0.23469387755102042, "frac_reward_zero_std": 0.75, "grad_norm": 0.23828125, "kl": 0.0016931954687606776, "learning_rate": 7.658730158730159e-07, "loss": -0.0279, "num_tokens": 38934019.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 414, "step_time": 296.73214874975383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3842.0, "completions/mean_length": 1059.296875, "completions/mean_terminated_length": 686.368408203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.0324822701513767, "epoch": 0.23526077097505668, "frac_reward_zero_std": 0.75, "grad_norm": 0.2109375, "kl": 0.0022498586713481927, "learning_rate": 7.653061224489795e-07, "loss": 0.0614, "num_tokens": 39013902.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 415, "step_time": 297.1698560230434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3978.0, "completions/mean_length": 1172.1875, "completions/mean_terminated_length": 924.4067993164062, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "entropy": 3.493779791519046, "epoch": 0.23582766439909297, "frac_reward_zero_std": 0.75, "grad_norm": 0.17578125, "kl": 0.0016595613633398898, "learning_rate": 7.647392290249433e-07, "loss": 0.0295, "num_tokens": 39102850.0, "reward": 0.3671875, "reward_std": 0.8648432493209839, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 416, "step_time": 302.1476899283007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4090.0, "completions/mean_length": 1529.375, "completions/mean_terminated_length": 996.6792602539062, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "entropy": 4.744919311255217, "epoch": 0.23639455782312926, "frac_reward_zero_std": 0.875, "grad_norm": 0.134765625, "kl": 0.0017254153863177635, "learning_rate": 7.64172335600907e-07, "loss": 0.0176, "num_tokens": 39213930.0, "reward": 0.125, "reward_std": 0.4714045226573944, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 417, "step_time": 297.82898204773664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3593.0, "completions/mean_length": 982.546875, "completions/mean_terminated_length": 718.6949462890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.9274300076067448, "epoch": 0.23696145124716553, "frac_reward_zero_std": 0.625, "grad_norm": 0.349609375, "kl": 0.0019668883469421417, "learning_rate": 7.636054421768708e-07, "loss": 0.0132, "num_tokens": 39289413.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 418, "step_time": 301.8002496920526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3302.0, "completions/mean_length": 1272.078125, "completions/mean_terminated_length": 685.9811401367188, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "entropy": 2.2984911371022463, "epoch": 0.23752834467120182, "frac_reward_zero_std": 0.75, "grad_norm": 0.2412109375, "kl": 0.001998718649701914, "learning_rate": 7.630385487528344e-07, "loss": -0.0016, "num_tokens": 39382370.0, "reward": 0.109375, "reward_std": 0.41277948021888733, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 419, "step_time": 294.1703432239592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3649.0, "completions/mean_length": 1323.515625, "completions/mean_terminated_length": 983.0350952148438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.132300119847059, "epoch": 0.23809523809523808, "frac_reward_zero_std": 0.625, "grad_norm": 0.322265625, "kl": 0.0013739411515416577, "learning_rate": 7.624716553287982e-07, "loss": 0.0231, "num_tokens": 39479235.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 420, "step_time": 295.6117933075875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3990.0, "completions/mean_length": 1386.859375, "completions/mean_terminated_length": 885.1666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.570806212723255, "epoch": 0.23866213151927437, "frac_reward_zero_std": 0.75, "grad_norm": 0.2412109375, "kl": 0.0014071986370254308, "learning_rate": 7.619047619047618e-07, "loss": 0.0175, "num_tokens": 39579754.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 421, "step_time": 294.0098496126011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3329.0, "completions/mean_length": 1107.296875, "completions/mean_terminated_length": 798.1206665039062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.599139340221882, "epoch": 0.23922902494331066, "frac_reward_zero_std": 0.625, "grad_norm": 0.34375, "kl": 0.001797916227587848, "learning_rate": 7.613378684807256e-07, "loss": -0.0228, "num_tokens": 39661421.0, "reward": 0.1640625, "reward_std": 0.5274578928947449, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 422, "step_time": 293.1747271474451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3955.0, "completions/mean_length": 1037.703125, "completions/mean_terminated_length": 778.5254516601562, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "entropy": 3.3564374186098576, "epoch": 0.23979591836734693, "frac_reward_zero_std": 0.875, "grad_norm": 0.21875, "kl": 0.0013273896947794128, "learning_rate": 7.607709750566893e-07, "loss": -0.0055, "num_tokens": 39739746.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 423, "step_time": 295.4561696825549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3359.0, "completions/mean_length": 1113.25, "completions/mean_terminated_length": 860.4745483398438, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 3.37971917912364, "epoch": 0.24036281179138322, "frac_reward_zero_std": 0.5, "grad_norm": 0.3984375, "kl": 0.001881603437141166, "learning_rate": 7.602040816326531e-07, "loss": 0.0556, "num_tokens": 39822450.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 424, "step_time": 293.7390189040452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3140.0, "completions/mean_length": 1238.84375, "completions/mean_terminated_length": 771.30908203125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.6694610193371773, "epoch": 0.2409297052154195, "frac_reward_zero_std": 0.75, "grad_norm": 0.2294921875, "kl": 0.002417201538264635, "learning_rate": 7.596371882086168e-07, "loss": 0.0354, "num_tokens": 39914328.0, "reward": 0.2421875, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 425, "step_time": 297.5649088798091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3497.0, "completions/mean_length": 1007.765625, "completions/mean_terminated_length": 688.2930908203125, "completions/min_length": 65.0, "completions/min_terminated_length": 65.0, "entropy": 3.1933873035013676, "epoch": 0.24149659863945577, "frac_reward_zero_std": 0.75, "grad_norm": 0.2138671875, "kl": 0.0018550817931100028, "learning_rate": 7.590702947845806e-07, "loss": -0.0315, "num_tokens": 39993041.0, "reward": 0.2421875, "reward_std": 0.666620135307312, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 426, "step_time": 319.5642735818401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4004.0, "completions/mean_length": 1179.046875, "completions/mean_terminated_length": 820.8245849609375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.247973505407572, "epoch": 0.24206349206349206, "frac_reward_zero_std": 0.5, "grad_norm": 0.318359375, "kl": 0.0020758294776896946, "learning_rate": 7.585034013605441e-07, "loss": 0.0782, "num_tokens": 40081708.0, "reward": 0.25, "reward_std": 0.7014724016189575, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 427, "step_time": 304.9623340750113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3750.0, "completions/mean_length": 1334.765625, "completions/mean_terminated_length": 940.3035888671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.298788692802191, "epoch": 0.24263038548752835, "frac_reward_zero_std": 0.875, "grad_norm": 0.1748046875, "kl": 0.0012791893022949807, "learning_rate": 7.579365079365078e-07, "loss": 0.0129, "num_tokens": 40179269.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 428, "step_time": 299.1383136184886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3310.0, "completions/mean_length": 1314.578125, "completions/mean_terminated_length": 799.5, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.217808473855257, "epoch": 0.24319727891156462, "frac_reward_zero_std": 0.75, "grad_norm": 0.2265625, "kl": 0.0015218931566778338, "learning_rate": 7.573696145124716e-07, "loss": 0.026, "num_tokens": 40274234.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 429, "step_time": 292.306060356088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4059.0, "completions/mean_length": 1422.6875, "completions/mean_terminated_length": 927.629638671875, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.6152460537850857, "epoch": 0.2437641723356009, "frac_reward_zero_std": 0.875, "grad_norm": 0.16015625, "kl": 0.0012690617950283922, "learning_rate": 7.568027210884353e-07, "loss": -0.0062, "num_tokens": 40379126.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 430, "step_time": 299.54230482690036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3615.0, "completions/mean_length": 932.734375, "completions/mean_terminated_length": 721.8500366210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.814725812524557, "epoch": 0.2443310657596372, "frac_reward_zero_std": 0.5, "grad_norm": 0.419921875, "kl": 0.0026313944181310944, "learning_rate": 7.562358276643991e-07, "loss": 0.0156, "num_tokens": 40449701.0, "reward": 0.234375, "reward_std": 0.5907269716262817, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 431, "step_time": 294.5451189512387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3927.0, "completions/mean_length": 1063.9375, "completions/mean_terminated_length": 914.8196411132812, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.978424996137619, "epoch": 0.24489795918367346, "frac_reward_zero_std": 0.375, "grad_norm": 0.419921875, "kl": 0.0019791342401731526, "learning_rate": 7.556689342403628e-07, "loss": 0.1045, "num_tokens": 40528257.0, "reward": 0.2890625, "reward_std": 0.6655030846595764, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 432, "step_time": 293.1097595356405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2882.0, "completions/mean_length": 771.5625, "completions/mean_terminated_length": 608.0655517578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 1.9814389497041702, "epoch": 0.24546485260770975, "frac_reward_zero_std": 0.5, "grad_norm": 0.478515625, "kl": 0.0018759423110168427, "learning_rate": 7.551020408163265e-07, "loss": -0.033, "num_tokens": 40590053.0, "reward": 0.4296875, "reward_std": 0.7759642601013184, "rewards/accuracy_reward_func/mean": 0.375, "rewards/accuracy_reward_func/std": 0.7867957949638367, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 433, "step_time": 310.92460263706744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3897.0, "completions/mean_length": 1108.859375, "completions/mean_terminated_length": 799.8448486328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.3193601071834564, "epoch": 0.24603174603174602, "frac_reward_zero_std": 0.5, "grad_norm": 0.453125, "kl": 0.0019465246878098696, "learning_rate": 7.545351473922901e-07, "loss": -0.1088, "num_tokens": 40680636.0, "reward": 0.2578125, "reward_std": 0.666620135307312, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 434, "step_time": 346.41499474365264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3182.0, "completions/mean_length": 1367.8125, "completions/mean_terminated_length": 978.0714721679688, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 3.699978746473789, "epoch": 0.2465986394557823, "frac_reward_zero_std": 0.75, "grad_norm": 0.1884765625, "kl": 0.0017686764949758071, "learning_rate": 7.539682539682539e-07, "loss": 0.0306, "num_tokens": 40778840.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 435, "step_time": 298.42550599202514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3836.0, "completions/mean_length": 1504.03125, "completions/mean_terminated_length": 1185.7193603515625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 3.9608581140637398, "epoch": 0.2471655328798186, "frac_reward_zero_std": 0.75, "grad_norm": 0.1806640625, "kl": 0.001418873338479898, "learning_rate": 7.534013605442176e-07, "loss": 0.0507, "num_tokens": 40885506.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 436, "step_time": 294.686938627623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3131.0, "completions/mean_length": 1098.296875, "completions/mean_terminated_length": 730.1578979492188, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.6076142862439156, "epoch": 0.24773242630385486, "frac_reward_zero_std": 0.75, "grad_norm": 0.26953125, "kl": 0.0018644657066033687, "learning_rate": 7.528344671201814e-07, "loss": -0.0304, "num_tokens": 40966829.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 437, "step_time": 301.3715253164992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3701.0, "completions/mean_length": 1595.5625, "completions/mean_terminated_length": 1076.603759765625, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "entropy": 2.996962059289217, "epoch": 0.24829931972789115, "frac_reward_zero_std": 0.5, "grad_norm": 0.2255859375, "kl": 0.0010735964206105564, "learning_rate": 7.522675736961451e-07, "loss": 0.0582, "num_tokens": 41080169.0, "reward": 0.15625, "reward_std": 0.5261042714118958, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 438, "step_time": 292.6421411121264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4054.0, "completions/mean_length": 812.8125, "completions/mean_terminated_length": 760.698486328125, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "entropy": 2.873863724991679, "epoch": 0.24886621315192745, "frac_reward_zero_std": 0.625, "grad_norm": 0.333984375, "kl": 0.0016787676504463889, "learning_rate": 7.517006802721088e-07, "loss": -0.0004, "num_tokens": 41144797.0, "reward": 0.265625, "reward_std": 0.6297439336776733, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 439, "step_time": 295.166032015346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3724.0, "completions/mean_length": 1005.078125, "completions/mean_terminated_length": 743.1356201171875, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "entropy": 3.237634778022766, "epoch": 0.2494331065759637, "frac_reward_zero_std": 0.75, "grad_norm": 0.271484375, "kl": 0.0020494155378401047, "learning_rate": 7.511337868480725e-07, "loss": 0.0711, "num_tokens": 41220354.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 440, "step_time": 298.98595599643886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3787.0, "completions/mean_length": 1482.34375, "completions/mean_terminated_length": 1108.96435546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.5558236204087734, "epoch": 0.25, "frac_reward_zero_std": 0.875, "grad_norm": 0.1796875, "kl": 0.001472385167289758, "learning_rate": 7.505668934240363e-07, "loss": 0.0118, "num_tokens": 41327408.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 441, "step_time": 299.76319946255535 } ], "logging_steps": 1, "max_steps": 1764, "num_input_tokens_seen": 41327408, "num_train_epochs": 1, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }