Question Answering on AmbigQA (EM and F1)
61.3EMRAGShaper
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RAGShaperprotocol=Learning-Based, training_data_size=4.5k2026.01 | 61.3 | 71.4 | |
| RAGShaperprotocol=Learning-Based, training_data_size=6.5k2026.01 | 61.1 | 71.1 | |
| HL-Dataprotocol=Learning-Based, training_data_size=4.5k2026.01 | 52.1 | 69 | |
| DecEx-RAGprotocol=Learning-Based2026.01 | 49.5 | 59.5 | |
| Search-R1protocol=Learning-Based2026.01 | 49.2 | 60.4 | |
| Iter-RetGenprotocol=Prompt-Based2026.01 | 47 | 58.8 | |
| IKEAprotocol=Learning-Based2026.01 | 47 | 57.9 | |
| Search-o1protocol=Prompt-Based2026.01 | 42.5 | 53.4 | |
| FLAREprotocol=Prompt-Based2026.01 | 40.6 | 50.1 | |
| ReasonRAGprotocol=Learning-Based2026.01 | 39.7 | 51.9 | |
| IR-COTprotocol=Prompt-Based2026.01 | 24.5 | 40.6 |