Abstention on MusiQ
90.4Abstain AccR2C
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| R2CRAG Framework=ReAct, Confidence Threshold=0.92025.10 | 90.4 | 94.81 | |
| SelfCRAG Framework=Search-o1, Confidence Threshold=0.92025.10 | 90.2 | 94.64 | |
| ReaCRAG Framework=ReAct, Confidence Threshold=0.92025.10 | 89.4 | 94.19 | |
| R2CRAG Framework=Search-o1, Confidence Threshold=0.92025.10 | 89.4 | 94.27 | |
| R2CRAG Framework=SelfAsk, Confidence Threshold=0.92025.10 | 88.6 | 93.77 | |
| SelfCRAG Framework=ReAct, Confidence Threshold=0.92025.10 | 87.6 | 93.18 | |
| ReaCRAG Framework=Search-o1, Confidence Threshold=0.92025.10 | 87.4 | 93.03 | |
| R2CRAG Framework=Search-R1, Confidence Threshold=0.92025.10 | 84.4 | 90.97 | |
| R2CRAG Framework=ReSearch, Confidence Threshold=0.92025.10 | 84 | 90.49 | |
| SelfCRAG Framework=ReSearch, Confidence Threshold=0.92025.10 | 83.6 | 90.57 | |
| RrrCRAG Framework=ReAct, Confidence Threshold=0.92025.10 | 82.4 | 89.69 | |
| ReaCRAG Framework=ReSearch, Confidence Threshold=0.92025.10 | 80.8 | 88.43 | |
| ReaCRAG Framework=Search-R1, Confidence Threshold=0.92025.10 | 80.6 | 88.27 | |
| P(true)RAG Framework=SelfAsk, Confidence Threshold=0.92025.10 | 80.4 | 88.27 | |
| ReaCRAG Framework=SelfAsk, Confidence Threshold=0.92025.10 | 80.4 | 88.41 | |
| P(true)RAG Framework=ReSearch, Confidence Threshold=0.92025.10 | 78.2 | 85.82 | |
| SelfCRAG Framework=Search-R1, Confidence Threshold=0.92025.10 | 77.8 | 86.51 | |
| SelfCRAG Framework=SelfAsk, Confidence Threshold=0.92025.10 | 76.2 | 85.71 | |
| P(true)RAG Framework=ReAct, Confidence Threshold=0.92025.10 | 75 | 84.58 | |
| P(true)RAG Framework=Search-o1, Confidence Threshold=0.92025.10 | 75 | 84.51 | |
| P(true)RAG Framework=Search-R1, Confidence Threshold=0.92025.10 | 74.4 | 83.37 | |
| RrrCRAG Framework=SelfAsk, Confidence Threshold=0.92025.10 | 73 | 82.75 | |
| RrrCRAG Framework=Search-o1, Confidence Threshold=0.92025.10 | 68.2 | 78.99 | |
| RrrCRAG Framework=ReSearch, Confidence Threshold=0.92025.10 | 64 | 73.68 | |
| RrrCRAG Framework=Search-R1, Confidence Threshold=0.92025.10 | 63.6 | 73.92 |