Datasets:
Formats:
csv
Sub-tasks:
multi-class-classification
Languages:
English
Size:
10K - 100K
ArXiv:
Tags:
multi-agent-systems
llm-routing
cost-aware-inference
calibration
agent-collaboration
reasoning
License:
| display_name,setting_id,target,n_parseable,positive_n,prevalence,AUROC_using_failure_score,AUPRC_using_failure_score | |
| Gemma-4-31B-it OmniMath,omnimath2__competition_math_4181__gemma_4_31b,baseline_failure,4178,1279,0.3061273336524653,0.8011559889217953,0.58713565059868 | |
| Gemma-4-31B-it OmniMath,omnimath2__competition_math_4181__gemma_4_31b,any_protocol_improves_over_baseline,4178,1098,0.2628051699377692,0.7827808234570529,0.505651139871984 | |
| Gemma-4-31B-it OmniMath,omnimath2__competition_math_4181__gemma_4_31b,PER_cheapest_successful_expensive_protocol,4178,268,0.064145524174246,0.7289818490666871,0.1255485555739646 | |
| Gemma-4-31B-it OmniMath,omnimath2__competition_math_4181__gemma_4_31b,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,4178,120,0.028721876495931,0.738923114834894,0.062568948114665 | |
| gpt-oss-120b OmniMath,omnimath2__competition_math_4181__gpt_oss_120b,baseline_failure,4151,1802,0.4341122621055167,0.884662470014633,0.8950166484026647 | |
| gpt-oss-120b OmniMath,omnimath2__competition_math_4181__gpt_oss_120b,any_protocol_improves_over_baseline,4151,1496,0.3603950855215611,0.8544073587318851,0.768337000237751 | |
| gpt-oss-120b OmniMath,omnimath2__competition_math_4181__gpt_oss_120b,PER_cheapest_successful_expensive_protocol,4151,365,0.0879306191279209,0.7258747078276853,0.1673912500057017 | |
| gpt-oss-120b OmniMath,omnimath2__competition_math_4181__gpt_oss_120b,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,4151,175,0.042158516020236,0.7639062949123311,0.1040541187831209 | |
| Gemma-4-31B-it LAB strict,labbench__llm_strict__gemma_4_31b,baseline_failure,733,409,0.557980900409277,0.8636768390232122,0.8692024605906812 | |
| Gemma-4-31B-it LAB strict,labbench__llm_strict__gemma_4_31b,any_protocol_improves_over_baseline,733,340,0.4638472032742155,0.7770693010028439,0.7297217292814904 | |
| Gemma-4-31B-it LAB strict,labbench__llm_strict__gemma_4_31b,PER_cheapest_successful_expensive_protocol,733,85,0.1159618008185538,0.6931735657225854,0.2185598218072934 | |
| Gemma-4-31B-it LAB strict,labbench__llm_strict__gemma_4_31b,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,733,138,0.1882673942701228,0.7875350140056022,0.3369755406158204 | |
| gpt-oss-120b LAB strict,labbench__llm_strict__gpt_oss_120b,baseline_failure,719,589,0.8191933240611962,0.581350398328327,0.9040112505244892 | |
| gpt-oss-120b LAB strict,labbench__llm_strict__gpt_oss_120b,any_protocol_improves_over_baseline,719,424,0.5897079276773296,0.561148864726575,0.6588615199451584 | |
| gpt-oss-120b LAB strict,labbench__llm_strict__gpt_oss_120b,PER_cheapest_successful_expensive_protocol,719,152,0.2114047287899861,0.5589204492713264,0.2528081789781673 | |
| gpt-oss-120b LAB strict,labbench__llm_strict__gpt_oss_120b,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,719,171,0.2378303198887343,0.5343994109360993,0.2979382157678691 | |
| Gemma-4-31B-it LAB text-no-tool,labbench__text_no_tool__gemma_4_31b_text_no_tool,baseline_failure,1542,895,0.5804150453955902,0.7337155587023909,0.8204342347607122 | |
| Gemma-4-31B-it LAB text-no-tool,labbench__text_no_tool__gemma_4_31b_text_no_tool,any_protocol_improves_over_baseline,1542,839,0.5440985732814526,0.676384370067326,0.7085956492959038 | |
| Gemma-4-31B-it LAB text-no-tool,labbench__text_no_tool__gemma_4_31b_text_no_tool,PER_cheapest_successful_expensive_protocol,1542,167,0.1083009079118028,0.6902406096897115,0.2053730372162358 | |
| Gemma-4-31B-it LAB text-no-tool,labbench__text_no_tool__gemma_4_31b_text_no_tool,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,1542,147,0.0953307392996108,0.8737351571452954,0.3194805969433019 | |
| gpt-oss-120b LAB text-no-tool,labbench__text_no_tool__gpt_oss_120b_text_no_tool,baseline_failure,1510,1062,0.7033112582781457,0.6069074522464353,0.8363223137408441 | |
| gpt-oss-120b LAB text-no-tool,labbench__text_no_tool__gpt_oss_120b_text_no_tool,any_protocol_improves_over_baseline,1510,854,0.5655629139072847,0.5693945636031301,0.6281893064232213 | |
| gpt-oss-120b LAB text-no-tool,labbench__text_no_tool__gpt_oss_120b_text_no_tool,PER_cheapest_successful_expensive_protocol,1510,190,0.1258278145695364,0.5541168261562999,0.1629555698101306 | |
| gpt-oss-120b LAB text-no-tool,labbench__text_no_tool__gpt_oss_120b_text_no_tool,Broadcast_uniquely_needed_beyond_Baseline_Single_PER,1510,251,0.1662251655629139,0.5683651414991345,0.2664117136859256 | |