Multi-document Question Answering on MDQA
72.6Recall@1SE
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SEBackbone=Qwen2025.12 | 72.6 | 54.4 | 55.8 | 52.4 | 56.4 | 58.3 | |
| SE w SIWBackbone=Qwen2025.12 | 72 | 57.2 | 56.6 | 52.4 | 55.2 | 58.7 | |
| SE w SIWBackbone=Vicuna2025.12 | 69.6 | 51.4 | 47.8 | 44.6 | 51.4 | 53 | |
| SEBackbone=Vicuna2025.12 | 69.2 | 49.8 | 43.8 | 43.8 | 52.2 | 51.8 | |
| SphsBackbone=Qwen2025.12 | 65.6 | 54 | 51.4 | 54.4 | 58.2 | 56.7 | |
| Sphs w SIWBackbone=Qwen2025.12 | 63.8 | 56 | 53.4 | 55.4 | 60.4 | 57.8 | |
| SphsBackbone=Vicuna2025.12 | 57.2 | 47.6 | 40.8 | 42.8 | 53.8 | 48.4 | |
| Sphs w SIWBackbone=Vicuna2025.12 | 56 | 49 | 42.6 | 44 | 54.4 | 49.2 | |
| Sphs w SIWBackbone=Tulu2025.12 | 55.4 | 47 | 44 | 46.2 | 55.6 | 49.6 | |
| SphsBackbone=Tulu2025.12 | 53.8 | 46.6 | 43 | 47 | 55.4 | 49.2 | |
| SE w SIWBackbone=Tulu2025.12 | 45 | 31.2 | 30.2 | 32.4 | 43.8 | 36.5 | |
| SEBackbone=Tulu2025.12 | 44.8 | 30 | 28.2 | 32.2 | 44.2 | 35.9 | |
| SE w SIWBackbone=Llama-7B2025.12 | 43.8 | 32.4 | 34.4 | 39.2 | 40.4 | 38 | |
| SEBackbone=Llama-7B2025.12 | 42.6 | 29.6 | 29.4 | 38.6 | 39.2 | 35.9 | |
| Sphs w SIWBackbone=Llama-7B2025.12 | 35.2 | 34.6 | 40.2 | 43.4 | 53.8 | 41.4 | |
| SphsBackbone=Llama-7B2025.12 | 33.6 | 34 | 40.6 | 43 | 51.8 | 40.6 |