Multi-hop Question Answering on MuSiQue fullwiki (val)
41.4Exact Match (EM)SDP
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SDPLLM=GPT-4o2026.05 | 41.4 | 51.9 | |
| IRCoTLLM=GPT-32026.05 | 34.2 | 43.8 | |
| PRISMLLM=GPT-4o2026.05 | 31.2 | 41.8 | |
| SETR-CoT & IRILLM=GPT-4o2026.05 | 12.3 | 16.9 | |
| RankGPTLLM=GPT-4o2026.05 | 9.5 | 13.5 | |
| RankZephyr + CoTLLM=GPT-4o2026.05 | 9.4 | 13.3 | |
| RankZephyrLLM=GPT-4o2026.05 | 8.6 | 12.8 |