Question Answering on CWQ (EM, F1)
25Exact Match (EM)PRAG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PRAGBase Model=LLAMA3.2-1B2026.06 | 25 | 32.84 | |
| DMOEBase Model=LLAMA3.2-1B2026.06 | 24.67 | 34.79 | |
| FLAREBase Model=LLAMA3.2-1B2026.06 | 24 | 31.54 | |
| SFT-LORABase Model=LLAMA3.2-1B2026.06 | 21.67 | 30.92 | |
| PRAGBase Model=QWEN2.5-1.5B2026.06 | 21.67 | 30.5 | |
| DMOEBase Model=QWEN2.5-1.5B2026.06 | 21.67 | 32.85 | |
| SFT-LORABase Model=QWEN2.5-1.5B2026.06 | 19.67 | 30.92 | |
| BASIC-RAGBase Model=LLAMA3.2-1B2026.06 | 16.33 | 23.84 | |
| BASIC-RAGBase Model=QWEN2.5-1.5B2026.06 | 12.33 | 18.88 | |
| FLAREBase Model=QWEN2.5-1.5B2026.06 | 12 | 18.06 |