Question Answering on ASDiv
66.91AUROCTOKENSAR + CoT-UQ
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| TOKENSAR + CoT-UQBackbone=Llama 3.1-8B, CoT-UQ=true2025.02 | 66.91 | 75.43 | |
| Probas-mean + CoT-UQBackbone=Llama 3.1-8B, CoT-UQ=true2025.02 | 64.52 | 73.95 | |
| TOKENSARBackbone=Llama 3.1-8B, CoT-UQ=false2025.02 | 58.71 | 68.6 | |
| Probas-meanBackbone=Llama 3.1-8B, CoT-UQ=false2025.02 | 58.34 | 69.03 | |
| P(True) + CoT-UQBackbone=Llama 3.1-8B, CoT-UQ=true2025.02 | 53.2 | 64.15 | |
| P(True)Backbone=Llama 3.1-8B, CoT-UQ=false2025.02 | 47.23 | 60.28 |