Regression identifies specific n-grams in LLM reasoning that drive overconfidence, enabling calibration via their suppression without performance loss.
A Limitations In our study, several limitations remain
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Identifying Influential N-grams in Confidence Calibration via Regression Analysis
Regression identifies specific n-grams in LLM reasoning that drive overconfidence, enabling calibration via their suppression without performance loss.