Question Answering on XOR-TyDi QA
0.383fEMLCRL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LCRLBase Model=Qwen3-8B2026.01 | 0.383 | 0.489 | 0.993 | |
| D-RAGBase Model=Qwen3-8B2026.01 | 0.335 | 0.423 | 0.935 | |
| LCRLBase Model=Qwen3-4B2026.01 | 0.324 | 0.457 | 0.994 | |
| LCRLBase Model=Qwen2.5-3B-Instruct2026.01 | 0.317 | 0.439 | 0.992 | |
| D-RAGBase Model=Qwen2.5-3B-Instruct2026.01 | 0.315 | 0.389 | 0.871 | |
| D-RAGBase Model=Qwen3-4B2026.01 | 0.311 | 0.387 | 0.903 | |
| SFTBase Model=Qwen3-8B2026.01 | 0.3 | 0.389 | 0.94 | |
| SFTBase Model=Qwen2.5-3B-Instruct2026.01 | 0.283 | 0.355 | 0.877 | |
| SFTBase Model=Qwen3-4B2026.01 | 0.282 | 0.356 | 0.916 | |
| mSearch-R1Base Model=Qwen3-8B2026.01 | 0.272 | 0.385 | 0.942 | |
| Search-o1Base Model=Qwen3-8B2026.01 | 0.252 | 0.343 | 0.746 | |
| IRCoTBase Model=Qwen3-8B2026.01 | 0.24 | 0.33 | 0.719 | |
| mSearch-R1Base Model=Qwen3-4B2026.01 | 0.234 | 0.366 | 0.946 | |
| RAGBase Model=Qwen3-8B2026.01 | 0.227 | 0.318 | 0.749 | |
| mSearch-R1Base Model=Qwen2.5-3B-Instruct2026.01 | 0.212 | 0.358 | 0.941 | |
| Search-o1Base Model=Qwen2.5-3B-Instruct2026.01 | 0.211 | 0.307 | 0.679 | |
| IRCoTBase Model=Qwen3-4B2026.01 | 0.207 | 0.281 | 0.696 | |
| RAGBase Model=Qwen3-4B2026.01 | 0.207 | 0.281 | 0.717 | |
| RAGBase Model=Qwen2.5-3B-Instruct2026.01 | 0.2 | 0.291 | 0.652 | |
| Search-o1Base Model=Qwen3-4B2026.01 | 0.196 | 0.267 | 0.731 | |
| IRCoTBase Model=Qwen2.5-3B-Instruct2026.01 | 0.185 | 0.279 | 0.694 | |
| Search-R1Base Model=Qwen2.5-3B-Instruct2026.01 | 0.184 | 0.32 | 0.863 | |
| Search-R1Base Model=Qwen3-4B2026.01 | 0.178 | 0.298 | 0.872 | |
| Search-R1Base Model=Qwen3-8B2026.01 | 0.168 | 0.281 | 0.889 | |
| Direct InferenceBase Model=Qwen3-8B2026.01 | 0.123 | 0.227 | 0.675 | |
| Direct InferenceBase Model=Qwen2.5-3B-Instruct2026.01 | 0.101 | 0.19 | 0.706 | |
| Direct InferenceBase Model=Qwen3-4B2026.01 | 0.1 | 0.203 | 0.661 |