Question Answering on 2Wiki Normal
23.63F1 ScoreOpenDecoder
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| OpenDecoderBackbone=Qwen-2.5-3B-Instruct2026.01 | 23.63 | 22.53 | |
| RbFTBackbone=Qwen-2.5-3B-Instruct2026.01 | 23.02 | 21.9 | |
| No RAGBackbone=Qwen-2.5-3B-Instruct2026.01 | 22.38 | — | |
| AstuteRAGBackbone=Qwen-2.5-3B-Instruct2026.01 | 20.79 | 21.1 | |
| Vanilla SFTBackbone=Qwen-2.5-3B-Instruct2026.01 | 19.91 | 20.33 | |
| InstructRAGBackbone=Qwen-2.5-3B-Instruct2026.01 | 17.37 | 20.87 | |
| Vanilla RAGBackbone=Qwen-2.5-3B-Instruct2026.01 | 11.23 | 20.93 | |
| RobustRAGBackbone=Qwen-2.5-3B-Instruct2026.01 | 5.84 | 9.07 |