Question Answering on 2WikiMultiHopQA 15 (test)
67.25F1 ScoreOrchMAS
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OrchMASCategory=Ours2026.03 | 67.25 | 60.42 | |
| GPT-4o-miniCategory=CoT Reasoning2026.03 | 50.89 | 43.75 | |
| GEPACategory=MAS, Base Model=GPT-4o-mini2026.03 | 41.24 | 37.5 | |
| Qwen3-4BCategory=SFT2026.03 | 38.91 | 41.67 | |
| GPT-4o-miniCategory=Baseline2026.03 | 35.44 | 33.33 | |
| OPROCategory=MAS, Base Model=GPT-4o-mini2026.03 | 34.36 | 26.04 | |
| Qwen3-4BCategory=GRPO2026.03 | 34.02 | 35.42 | |
| Qwen3-4BCategory=Baseline2026.03 | 28.21 | 26.04 | |
| TextGradCategory=MAS, Base Model=GPT-4o-mini2026.03 | 26.42 | 17.71 | |
| Qwen3-4BCategory=CoT Reasoning2026.03 | 25.21 | 20.83 |