Multi-hop Question Answering on 2Wiki (Score)
42.4ScoreOurs
Evaluation Results
| Method | Links | |
|---|---|---|
| OursBackbone=Qwen2.5-7B-Instruct2026.05 | 42.4 | |
| Dr.ZeroBackbone=Qwen2.5-7B-Instruct2026.05 | 41.4 | |
| OursBackbone=Qwen2.5-3B-Instruct2026.05 | 39.6 | |
| Dr.ZeroBackbone=Qwen2.5-3B-Instruct2026.05 | 37.5 | |
| SSPBackbone=Qwen2.5-7B-Instruct2026.05 | 36.6 | |
| Base ModelBackbone=Qwen2.5-7B-Instruct2026.05 | 31.8 | |
| Base ModelBackbone=Qwen2.5-3B-Instruct2026.05 | 24.2 | |
| SSPBackbone=Qwen2.5-3B-Instruct2026.05 | 20.2 |