Multi-Hop Question Answering on HotpotQA (Exact Match (EM))
44.7Exact Match (EM)SKILLGRAPH
Evaluation results
| Method | Links | |
|---|---|---|
| SKILLGRAPHEvaluation Protocol=In-domain2026.05 | 44.7 | |
| SkillRLEvaluation Protocol=In-domain2026.05 | 43.2 | |
| EvolveREvaluation Protocol=In-domain2026.05 | 38.2 | |
| Search-R1Evaluation Protocol=In-domain2026.05 | 37 | |
| ZeroSearchEvaluation Protocol=In-domain2026.05 | 34.6 | |
| RAGEvaluation Protocol=In-domain2026.05 | 25.8 | |
| Search-o1Evaluation Protocol=In-domain2026.05 | 17 | |
| Qwen2.5-7B-InstructEvaluation Protocol=In-domain2026.05 | 16.4 | |
| CoTEvaluation Protocol=In-domain2026.05 | 16.2 |