Multi-hop Question Answering on MuSiQue 500 samples (test)
73.9F1 ScoreLLM-Wiki
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| LLM-Wiki2026.05 | 73.9 | 63.4 | |
| LightRAG2026.05 | 65.9 | 55 | |
| HippoRAG 22026.05 | 62.4 | 51.4 | |
| Vanilla RAGretrieval=Dense2026.05 | 61.1 | 50 | |
| GraphRAG2026.05 | 58.2 | 48.2 | |
| Vanilla RAGretrieval=BM252026.05 | 54.5 | 44.2 | |
| RAPTOR2026.05 | 52.2 | 44.2 | |
| Nonemode=Closed-book2026.05 | 45.6 | 37.2 |