Functional-level Uncertainty Quantification for Calibrated Fine-tuning on LLMs

ArXi:2410.06431v5 Announce Type: replace Accurate uncertainty quantification in large language models (LLMs) is essential for reliable confidence estimation, yet fine-tuned LLMs often become overconfident under limited adaptation data. Existing uncertainty methods for PEFT-based LLMs are largely post hoc, estimating uncertainty after fine-tuning rather than improving how adapters specialize to task-specific input-output relationships.