Question Answering on HotpotQA 49 (test) (EM and F1)
61.99F1 ScoreOrchMAS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OrchMASCategory=Ours2026.03 | 61.99 | 53.13 | |
| GEPACategory=MAS, Base Model=GPT-4o-mini2026.03 | 47.13 | 39.58 | |
| GPT-4o-miniCategory=CoT Reasoning2026.03 | 46.67 | 40.63 | |
| OPROCategory=MAS, Base Model=GPT-4o-mini2026.03 | 41.82 | 36.45 | |
| GPT-4o-miniCategory=Baseline2026.03 | 41.21 | 32.29 | |
| TextGradCategory=MAS, Base Model=GPT-4o-mini2026.03 | 35.21 | 28.13 | |
| Qwen3-4BCategory=SFT2026.03 | 33.09 | 23.96 | |
| Qwen3-4BCategory=GRPO2026.03 | 31.21 | 23.96 | |
| Qwen3-4BCategory=CoT Reasoning2026.03 | 24.98 | 18.75 | |
| Qwen3-4BCategory=Baseline2026.03 | 23.21 | 21.88 |