Papers by Masahiro Hamasaki
Evidential Semantic Entropy for LLM Uncertainty Quantification (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing methods for quantifying uncertainty in large language models do not account for the effects of the semantics of sampled answers. |
| Approach: | They propose to incorporate the semantics of sampled answers to estimate entropy by incorporating the semantic of sample answers into the estimation methods. |
| Outcome: | The proposed method significantly improves uncertainty quantification performance. |