Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language Models

TL;DR

This paper introduces Direct Confidence Alignment (DCA), a method to improve the alignment between a large language model's verbalized confidence and its internal token-based confidence, aiming to enhance transparency and trustworthiness.

Contribution

The paper proposes DCA, a novel approach using Direct Preference Optimization to better align verbalized and internal confidence in LLMs, and introduces new calibration metrics for evaluation.

Findings

DCA improves confidence alignment on certain models.

It reduces inconsistencies in confidence expression.

Effectiveness varies across different model architectures.

Abstract

Producing trustworthy and reliable Large Language Models (LLMs) has become increasingly important as their usage becomes more widespread. Calibration seeks to achieve this by improving the alignment between the model's confidence and the actual likelihood of its responses being correct or desirable. However, it has been observed that the internal confidence of a model, derived from token probabilities, is not well aligned with its verbalized confidence, leading to misleading results with different calibration methods. In this paper, we propose Direct Confidence Alignment (DCA), a method using Direct Preference Optimization to align an LLM's verbalized confidence with its internal confidence rather than ground-truth accuracy, enhancing model transparency and reliability by ensuring closer alignment between the two confidence measures. We evaluate DCA across multiple open-weight LLMs on a wide range of datasets. To further assess this alignment, we also introduce three new calibration error-based metrics. Our results show that DCA improves alignment metrics on certain model architectures, reducing inconsistencies in a model's confidence expression. However, we also show that it can be ineffective on others, highlighting the need for more model-aware approaches in the pursuit of more interpretable and trustworthy LLMs.