Multi-document Multi-entity Question Answering on Loong (All sets)
85.06Spotlight Locating Avg ScoreDocSage
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| DocSageMethod=DocSage (Ours)2026.03 | 85.06 | 84 | 73.28 | 43 | 64.43 | 42 | 73.67 | 57 | 68.29 | 53 | |
| GPT-4oMethod=GPT-4o2026.03 | 76.79 | 65 | 50.98 | 29 | 45.04 | 10 | 57.46 | 27 | 54.17 | 26 | |
| StructRAGMethod=StructRAG2026.03 | 68.07 | 40 | 63.36 | 36 | 60.71 | 14 | 53.27 | 18 | 60.56 | 23 | |
| GPT-4o + RAGMethod=GPT-4o + RAG2026.03 | 64.04 | 44 | 41.85 | 26 | 35.37 | 3 | 41.62 | 19 | 43.05 | 18 | |
| GraphRAGMethod=GraphRAG2026.03 | 22.49 | 0 | 22.91 | 1 | 37.52 | 3 | 45.34 | 23 | 33.44 | 7 |