Multi-hop Question Answering on KILT Average
71.7Average Judge ScoreLOGICALRAG
Evaluation results
| Method | Links | |
|---|---|---|
| LOGICALRAGCorpus=Large-scale (KILT)2026.05 | 71.7 | |
| Agentic HybridCorpus=Large-scale (KILT)2026.05 | 71.6 | |
| MA-RAGCorpus=Large-scale (KILT)2026.05 | 44.6 |