LGMay 13

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

arXiv:2410.0643181.72 citationsh-index: 8
Predicted impact top 22% in LG · last 90 daysOriginality Incremental advance
AI Analysis

For practitioners fine-tuning LLMs with limited data, this method provides better-calibrated confidence estimates, reducing overconfidence.

UQ4CT reduces Expected Calibration Error by over 25% on multiple-choice and open-ended QA benchmarks while preserving accuracy, improving calibration and generalization for fine-tuned LLMs.

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. We propose Functional-Level Uncertainty Quantification for Calibrated Fine-Tuning (UQ4CT), which calibrates uncertainty over the functional space induced by prompt-dependent mixtures of LoRA experts. UQ4CT implements this perspective through a mixture-of-experts fine-tuning framework, where a calibration loss aligns functional-level confidence with predictive correctness during training. Across four multiple-choice benchmarks and two open-ended generative QA tasks, UQ4CT reduces Expected Calibration Error (ECE) by over $25\%$ while preserving high accuracy. Under distribution shift, UQ4CT maintains superior calibration and competitive accuracy, demonstrating improved reliability and generalization for fine-tuned LLMs.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes