Datasetpublic

anirudhb11/rebase_gemma-4-E4B-it_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 2748.36 generation_phase_time_s 297.389 judge_output_tokens_level_1 18502.7 judge_output_tokens_level_2 3549.6 judge_output_tokens_level_3 496 judge_output_tokens_level_4 678 judge_output_tokens_search… See the full description on the dataset page: https://huggingface.co/datasets/anirudhb11/rebase_gemma-4-E4B-it_rg_cognition_ns8_md4_bt0_1_seed73_rg_cognition__v0.

sourceHugging Faceupdated 5mo agoView on Hugging Face

0likes6downloads

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`	2748.36
`generation_phase_time_s`	297.389
`judge_output_tokens_level_1`	18502.7
`judge_output_tokens_level_2`	3549.6
`judge_output_tokens_level_3`	496
`judge_output_tokens_level_4`	678
`judge_output_tokens_search`	22467.3
`judge_skipped_completed_level_1`	62.4
`judge_skipped_completed_level_2`	14.3
`judge_skipped_completed_level_3`	4.14286
`judge_skipped_completed_level_4`	3
`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`	3415
`judge_token_count_level_2`	1778.57
`judge_token_count_level_3`	1150
`judge_token_count_level_4`	700
`maj@1`	0.472284
`maj@1_std`	0.0646094
`maj@2`	0.473413
`maj@2_std`	0.0680784
`maj@4`	0.498676
`maj@4_std`	0.0530126
`maj@8`	0.511256
`maj@8_std`	0.0482403
`median_response_tokens`	2088.15
`num_unique_answers@1`	0.946
`num_unique_answers@2`	1.361
`num_unique_answers@4`	2.182
`num_unique_answers@8`	3.53
`pass@1`	0.44875
`pass@2`	0.505714
`pass@4`	0.544
`pass@8`	0.58
`policy_output_tokens_final_completion`	11.6
`policy_output_tokens_level_1`	184310
`policy_output_tokens_level_2`	29116.7
`policy_output_tokens_level_3`	8197.29
`policy_output_tokens_level_4`	6968
`policy_output_tokens_search_phases`	219862
`policy_output_tokens_total_so_far`	219873
`token_cdf`	1
`token_count`	9900
`total_generated_output_tokens`	279616
`total_judge_output_tokens`	59742.3
`total_judge_output_tokens_search`	22467.3
`total_judge_output_tokens_weighted_mv`	37275
`total_policy_output_tokens`	219873
`total_time_s`	351.386
`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`	0.461682
`w_best@1_std`	0.0649131
`w_best@2`	0.478806
`w_best@2_std`	0.0631255
`w_best@4`	0.47344
`w_best@4_std`	0.0638913
`w_best@8`	0.488858
`w_best@8_std`	0.0561242
`w_maj@1`	0.472284
`w_maj@1_std`	0.0655185
`w_maj@2`	0.473253
`w_maj@2_std`	0.066409
`w_maj@4`	0.498566
`w_maj@4_std`	0.0526795
`w_maj@8`	0.514936
`w_maj@8_std`	0.0482455
`w_num_unique_answers@1`	0.946
`w_num_unique_answers@2`	1.361
`w_num_unique_answers@4`	2.182
`w_num_unique_answers@8`	3.53