Uncertainty Estimation on CSQA
0.769AUROCBSDETECTOR
Evaluation Results
| Method | Links | |
|---|---|---|
| BSDETECTORLLM=GPT-3.5 Turbo2023.08 | 0.769 | |
| BSDETECTORLLM=Text-Davinci-0032023.08 | 0.743 | |
| Temperature SamplingLLM=GPT-3.5 Turbo2023.08 | 0.583 | |
| Temperature SamplingLLM=Text-Davinci-0032023.08 | 0.54 | |
| Self-reflection CertaintyLLM=Text-Davinci-0032023.08 | 0.539 | |
| Self-reflection CertaintyLLM=GPT-3.5 Turbo2023.08 | 0.506 | |
| Likelihood Based UncertaintyLLM=Text-Davinci-0032023.08 | 0.49 |