Multi-hop Question Answering on MuSiQue (pass@1, pass@2)
69.4Pass@1EXG-SE-Rev
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EXG-SE-RevModel=Qwen-Plus2026.05 | 69.4 | 99 | |
| EXG-ReflexionModel=Qwen-Plus2026.05 | 65.2 | 94.6 | |
| EXG-SEModel=Qwen-Plus2026.05 | 65.2 | 85.4 | |
| EXGModel=Qwen-Plus2026.05 | 60.8 | 67.2 | |
| SE-Agent-RevModel=Qwen-Plus2026.05 | 56.8 | 97 | |
| EXG-SE-RevModel=Qwen3-14B2026.05 | 56 | 89.6 | |
| SE-AgentModel=Qwen-Plus2026.05 | 55.2 | 84 | |
| ReflexionModel=Qwen-Plus2026.05 | 54.2 | 94 | |
| EXG-SE-RevModel=Qwen-Max2026.05 | 51.8 | 92.4 | |
| EXG-SEModel=Qwen3-14B2026.05 | 49 | 73.6 | |
| EXG-SEModel=Qwen-Max2026.05 | 48.6 | 92.6 | |
| EXG-ReflexionModel=Qwen3-14B2026.05 | 48 | 74.8 | |
| EXG-ReflexionModel=Qwen-Max2026.05 | 47.8 | 76.2 | |
| EXGModel=Qwen-Max2026.05 | 44.6 | 51.6 | |
| SE-Agent-RevModel=Qwen3-14B2026.05 | 38.2 | 83.6 | |
| SE-Agent-RevModel=Qwen-Max2026.05 | 37 | 88.2 | |
| SE-AgentModel=Qwen3-14B2026.05 | 35.6 | 70 | |
| SE-AgentModel=Qwen-Max2026.05 | 34.8 | 74.6 | |
| ReflexionModel=Qwen-Max2026.05 | 34.6 | 69.2 | |
| EXGModel=Qwen3-14B2026.05 | 32 | 40.6 | |
| ReflexionModel=Qwen3-14B2026.05 | 17.4 | 32.6 |