Single-hop Question Answering on Natural Questions
47.8AccuracyAmber
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AmberRetrieval Strategy=Amber, Backbone=Qwen2-7b2025.02 | 47.8 | 49.84 | |
| AmberRetrieval Strategy=Amber, Backbone=GPT-3.52025.02 | 47.4 | 52.01 | |
| Adaptive-RAGRetrieval Strategy=Adaptive RAG, Backbone=Llama3-8b2025.02 | 44.6 | 47.3 | |
| AmberRetrieval Strategy=Amber, Backbone=Llama3-8b2025.02 | 44.2 | 50.49 | |
| Adaptive-NoteRetrieval Strategy=Adaptive RAG, Backbone=Llama3-8b2025.02 | 40 | 45.38 | |
| Self-RAGRetrieval Strategy=Adaptive RAG, Backbone=Llama3-8b2025.02 | 37.9 | 39.17 | |
| FLARERetrieval Strategy=Adaptive RAG, Backbone=Llama3-8b2025.02 | 36.4 | 38.94 | |
| VanillaRetrieval Strategy=Single-time RAG, Backbone=Qwen2-7b2025.02 | 36.2 | 24.62 | |
| VanillaRetrieval Strategy=Single-time RAG, Backbone=GPT-3.52025.02 | 35.9 | 38.43 | |
| Self-RefineRetrieval Strategy=Single-time RAG, Backbone=Llama3-8b2025.02 | 35.8 | 35.17 | |
| ReActRetrieval Strategy=Adaptive RAG, Backbone=Llama3-8b2025.02 | 35.4 | 38.37 | |
| Chain-of-noteRetrieval Strategy=Single-time RAG, Backbone=Llama3-8b2025.02 | 35.2 | 37.66 | |
| Self-RerankRetrieval Strategy=Single-time RAG, Backbone=Llama3-8b2025.02 | 34.3 | 39.05 | |
| VanillaRetrieval Strategy=Single-time RAG, Backbone=Llama3-8b2025.02 | 33.2 | 38.99 | |
| NORRetrieval Strategy=No Retrieval, Backbone=Llama3-8b2025.02 | 24 | 27.49 |