Question Answering on MKQA
47.6fEMLCRL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LCRLBase Model=Qwen3-8B2026.01 | 47.6 | 63.2 | 99.4 | |
| LCRLBase Model=Qwen2.5-3B-Instruct2026.01 | 41.2 | 57 | 99.1 | |
| LCRLBase Model=Qwen3-4B2026.01 | 40.6 | 54.9 | 98.9 | |
| mSearch-R1Base Model=Qwen3-8B2026.01 | 40.6 | 50.8 | 94.8 | |
| D-RAGBase Model=Qwen3-8B2026.01 | 38 | 46.7 | 94.1 | |
| mSearch-R1Base Model=Qwen2.5-3B-Instruct2026.01 | 37.9 | 53.2 | 95.6 | |
| D-RAGBase Model=Qwen2.5-3B-Instruct2026.01 | 37.4 | 43.3 | 90.2 | |
| SFTBase Model=Qwen2.5-3B-Instruct2026.01 | 35 | 43 | 92.9 | |
| D-RAGBase Model=Qwen3-4B2026.01 | 33 | 39.7 | 92.7 | |
| mSearch-R1Base Model=Qwen3-4B2026.01 | 30.3 | 40.7 | 95 | |
| SFTBase Model=Qwen3-8B2026.01 | 29.8 | 38.3 | 93.4 | |
| Search-R1Base Model=Qwen3-8B2026.01 | 29.1 | 37.4 | 86.4 | |
| IRCoTBase Model=Qwen3-8B2026.01 | 28.7 | 35.6 | 70.3 | |
| Search-o1Base Model=Qwen3-8B2026.01 | 28.6 | 35.8 | 73.8 | |
| RAGBase Model=Qwen2.5-3B-Instruct2026.01 | 28.2 | 34.6 | 66.3 | |
| Search-o1Base Model=Qwen2.5-3B-Instruct2026.01 | 27.7 | 35 | 70.2 | |
| RAGBase Model=Qwen3-8B2026.01 | 27.6 | 34.8 | 74.2 | |
| Direct InferenceBase Model=Qwen3-8B2026.01 | 26.2 | 24.3 | 66.8 | |
| Search-R1Base Model=Qwen3-4B2026.01 | 24.3 | 34.7 | 85.8 | |
| RAGBase Model=Qwen3-4B2026.01 | 22.9 | 29.7 | 71 | |
| Search-R1Base Model=Qwen2.5-3B-Instruct2026.01 | 22.6 | 34.8 | 83.6 | |
| IRCoTBase Model=Qwen2.5-3B-Instruct2026.01 | 22.4 | 34.9 | 65.3 | |
| Search-o1Base Model=Qwen3-4B2026.01 | 22.1 | 28.6 | 72.4 | |
| SFTBase Model=Qwen3-4B2026.01 | 21.7 | 30.6 | 92.1 | |
| Direct InferenceBase Model=Qwen3-4B2026.01 | 20.3 | 29.3 | 65.4 | |
| IRCoTBase Model=Qwen3-4B2026.01 | 20.3 | 28.8 | 68.2 | |
| Direct InferenceBase Model=Qwen2.5-3B-Instruct2026.01 | 19.7 | 27.4 | 72.3 |