Multi-LLM response agreement plus focal loss in an auxiliary BERT confidence model improves LLM calibration across 12 models and four prompt styles.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Influences on LLM Calibration: A Study of Response Agreement, Loss Functions, and Prompt Styles
Multi-LLM response agreement plus focal loss in an auxiliary BERT confidence model improves LLM calibration across 12 models and four prompt styles.