Question Answering on WikiQA (F1/Acc)
59AccuracyTS (C)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TS (C)Algorithm Family=Contextual Linear2026.02 | 59 | — | |
| No-RewriteAlgorithm Family=Base2026.02 | 49.8 | — | |
| Oracle2025.09 | 39.8 | 10.5 | |
| OursBackbone=Qwen2.52025.09 | 30.2 | 9.7 | |
| Ours-cBackbone=Qwen2.52025.09 | 29 | 8.82 | |
| OursBackbone=LLaMA-32025.09 | 28.8 | 9.87 | |
| Ours-cBackbone=LLaMA-32025.09 | 28.2 | 9.61 | |
| Hybrid2025.09 | 28 | 8.99 | |
| OursBackbone=GPT-4o2025.09 | 26.2 | 9.3 | |
| Score AgentBackbone=LLaMA-32025.09 | 26 | 9.35 | |
| Agent-BasedBackbone=Qwen2.52025.09 | 25.8 | 8.84 | |
| Ours-cBackbone=GPT-4o2025.09 | 25 | 9.15 | |
| Score AgentBackbone=GPT-4o2025.09 | 24.8 | 9.19 | |
| Ours-cBackbone=GPT-4.12025.09 | 24.6 | 9.15 | |
| OursBackbone=GPT-4.12025.09 | 24.2 | 9.54 | |
| Rule AgentBackbone=Qwen2.52025.09 | 24 | 8.15 | |
| Score AgentBackbone=Qwen2.52025.09 | 24 | 8.55 | |
| Doc2025.09 | 23.6 | 6.79 | |
| Rule AgentBackbone=LLaMA-32025.09 | 23.4 | 8.12 | |
| Rule AgentBackbone=GPT-4o2025.09 | 22.8 | 8.68 | |
| Rule AgentBackbone=GPT-4.12025.09 | 22.8 | 9.13 | |
| Agent-BasedBackbone=GPT-4.12025.09 | 21.4 | 9.05 | |
| Score AgentBackbone=GPT-4.12025.09 | 20.8 | 9.15 | |
| Agent-BasedBackbone=GPT-4o2025.09 | 20.2 | 9.11 | |
| DB2025.09 | 18.8 | 5.13 | |
| Rule-Based2025.09 | 18.6 | 8.68 | |
| Adaptive-RAG2025.09 | 15.8 | 9.4 | |
| Agent-BasedBackbone=LLaMA-32025.09 | 15.2 | 9.11 | |
| Basic QA2025.09 | 13.8 | 4.32 |