Multi-Hop Question Answering on 2WikiMultihopQA (F1/EM)
70.33F1 ScoreHippoRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HippoRAGLLM=GPT-4o2026.04 | 70.33 | 61.16 | |
| QAFD-RAGLLM=GPT-4o2026.04 | 69.41 | 59.5 | |
| Qwen3-4B-SFT-RLVRParameters=4B, Training=SFT-RLVR2026.06 | 40.7 | 33.1 | |
| Qwen3-4BParameters=4B2026.06 | 40.5 | 33.6 | |
| RAPTORLLM=GPT-4o2026.04 | 38.8 | 12 | |
| Qwen3-1.7B-SFT-RLVRParameters=1.7B, Training=SFT-RLVR2026.06 | 20.1 | 15.6 | |
| Qwen3-1.7BParameters=1.7B2026.06 | 20 | 15.6 | |
| GraphRAGLLM=GPT-4o2026.04 | 15.2 | 7 | |
| LightRAGLLM=GPT-4o2026.04 | 8.2 | 1 |