Retrieval-Augmented Generation evaluation on RAGChecker
66.3Factual RecallRioRAG
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| RioRAGCategory=RL-based Methods2025.05 | 66.3 | 224.6 | 27.8 | 4.3 | 3.6 | 20.9 | 5 | 58.2 | |
| GopherCiteCategory=Supervised Fine-tuning based Methods2025.05 | 62.6 | 209.9 | 26 | 5.1 | 4.3 | 29.2 | 10.8 | 52.5 | |
| DPOCategory=RL-based Methods2025.05 | 61.2 | 149.6 | 26 | 5.2 | 6 | 27.8 | 8 | 53.1 | |
| Chain-of-NoteCategory=Supervised Fine-tuning based Methods2025.05 | 54.2 | 190.2 | 22.7 | 4.3 | 3.7 | 22.6 | 7.8 | 30.2 | |
| GopherCiteCategory=Prompt-based Methods2025.05 | 51.4 | 138.5 | 26 | 5.1 | 4.3 | 29.2 | 10.8 | 47.5 | |
| Chain-of-ThoughtCategory=Prompt-based Methods2025.05 | 50.4 | 146.5 | 24.3 | 4.6 | 4.2 | 30.2 | 9.7 | 48 | |
| Chain-of-NoteCategory=Prompt-based Methods2025.05 | 38.7 | 144.3 | 18.3 | 6.8 | 5.1 | 53 | 6.9 | 35.7 | |
| Direct-RAGCategory=Prompt-based Methods2025.05 | 38.3 | 91.6 | 22.6 | 8.2 | 7.5 | 37 | 8.1 | 45.3 |