Datasetpublic

anirudhb11/rebase_gpt-oss-20b_rg_cognition_ns8_md4_bt0_1_seed73_rg_cognition__v0

Aggregated Metrics Weighted mean of per-shard wandb summary values (weights = shard rows). See _meta/aggregated_metrics.json for raw values and _meta/splits.json for the per-shard wandb run names. Aggregated from 10 shards. Metric Value avg_response_tokens 5418.58 generation_phase_time_s 257.32 judge_output_tokens_level_1 26540.6 judge_output_tokens_level_2 8524.6 judge_output_tokens_level_3 4980.7 judge_output_tokens_level_4 3824.9… See the full description on the dataset page: https://huggingface.co/datasets/anirudhb11/rebase_gpt-oss-20b_rg_cognition_ns8_md4_bt0_1_seed73_rg_cognition__v0.

sourceHugging Faceupdated 5mo agoView on Hugging Face

1likes4downloads

Dataset Card

Aggregated Metrics

Weighted mean of per-shard wandb summary values (weights = shard rows). See `meta/aggregatedmetrics.json` for raw values and `meta/splits.json` for the per-shard wandb run names._

Aggregated from 10 shards.

Metric	Value
`avg_response_tokens`	5418.58
`generation_phase_time_s`	257.32
`judge_output_tokens_level_1`	26540.6
`judge_output_tokens_level_2`	8524.6
`judge_output_tokens_level_3`	4980.7
`judge_output_tokens_level_4`	3824.9
`judge_output_tokens_search`	43870.8
`judge_skipped_completed_level_1`	47.6
`judge_skipped_completed_level_2`	15.3
`judge_skipped_completed_level_3`	5.5
`judge_skipped_completed_level_4`	3.8
`judge_token_cdf_level_1`	1
`judge_token_cdf_level_2`	1
`judge_token_cdf_level_3`	1
`judge_token_cdf_level_4`	1
`judge_token_count_level_1`	5015
`judge_token_count_level_2`	2060
`judge_token_count_level_3`	1030
`judge_token_count_level_4`	1583.33
`maj@1`	0.493272
`maj@1_std`	0.0621241
`maj@2`	0.501822
`maj@2_std`	0.0654355
`maj@4`	0.533923
`maj@4_std`	0.0416096
`maj@8`	0.538972
`maj@8_std`	0.0375325
`median_response_tokens`	2838.1
`num_unique_answers@1`	0.931
`num_unique_answers@2`	1.353
`num_unique_answers@4`	2.087
`num_unique_answers@8`	3.27
`pass@1`	0.4625
`pass@2`	0.529643
`pass@4`	0.558286
`pass@8`	0.58
`policy_output_tokens_final_completion`	54925.4
`policy_output_tokens_level_1`	186264
`policy_output_tokens_level_2`	97098.8
`policy_output_tokens_level_3`	55764.4
`policy_output_tokens_level_4`	39511.5
`policy_output_tokens_search_phases`	378638
`policy_output_tokens_total_so_far`	433564
`token_cdf`	1
`token_count`	26450
`total_generated_output_tokens`	517760
`total_judge_output_tokens`	84196.5
`total_judge_output_tokens_search`	43870.8
`total_judge_output_tokens_weighted_mv`	40325.7
`total_policy_output_tokens`	433564
`total_time_s`	292.748
`truncated_count_level_1`	0
`truncated_count_level_2`	0
`truncated_count_level_3`	0
`truncated_count_level_4`	0
`truncated_rate_level_1`	0
`truncated_rate_level_2`	0
`truncated_rate_level_3`	0
`truncated_rate_level_4`	0
`w_best@1`	0.504398
`w_best@1_std`	0.0707404
`w_best@2`	0.485355
`w_best@2_std`	0.0635852
`w_best@4`	0.467005
`w_best@4_std`	0.0704265
`w_best@8`	0.47229
`w_best@8_std`	0.0643953
`w_maj@1`	0.493272
`w_maj@1_std`	0.0621241
`w_maj@2`	0.497393
`w_maj@2_std`	0.0618034
`w_maj@4`	0.531034
`w_maj@4_std`	0.045131
`w_maj@8`	0.546511
`w_maj@8_std`	0.0345272
`w_num_unique_answers@1`	0.931
`w_num_unique_answers@2`	1.353
`w_num_unique_answers@4`	2.087
`w_num_unique_answers@8`	3.27