Question Answering on HotpotQA (Specific EM Components: LQ, GQ, CQ, Total)
60.94EM (LQ)Local Search
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Local SearchGeneration=DeepSeek-R12026.01 | 60.94 | 10 | 14.76 | 41.11 | |
| Local SearchGeneration=Qwen2.5 72B2026.01 | 59.25 | 10.63 | 6.19 | 38.22 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 72B, Generation=DeepSeek-R12026.01 | 50.38 | 56.25 | 24.76 | 45.44 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 72B, Generation=Qwen2.5 72B2026.01 | 49.06 | 56.25 | 24.76 | 44.67 | |
| Drift SearchKnowledge Integration=Qwen2.5 72B, Generation=Qwen2.5 72B2026.01 | 39.62 | 30 | 22.38 | 33.89 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 1.5B-DW GRPO, Generation=Qwen2.5 72B2026.01 | 39.62 | 57.5 | 20.95 | 38.44 | |
| Drift SearchKnowledge Integration=Qwen2.5 72B, Generation=DeepSeek-R12026.01 | 39.62 | 38.75 | 22.86 | 35.56 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 1.5B-DW GRPO, Generation=DeepSeek-R12026.01 | 39.62 | 58.13 | 20.95 | 38.56 | |
| Global SearchKnowledge Integration=DeepSeek-R1, Generation=DeepSeek-R12026.01 | 20 | 48.75 | 9.05 | 22.56 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 1.5B, Generation=DeepSeek-R12026.01 | 20 | 18.13 | 3.81 | 17.48 | |
| Deep GraphRAGKnowledge Integration=Qwen2.5 1.5B, Generation=Qwen2.5 72B2026.01 | 18.87 | 15 | 2.86 | 15.97 | |
| Global SearchKnowledge Integration=Qwen2.5 72B, Generation=Qwen2.5 72B2026.01 | 18.49 | 41.88 | 6.19 | 19.78 |