Towards Reliable Truth-Aligned Uncertainty Estimation in Large Language Models
This work addresses the reliability of uncertainty estimation for LLMs, which is crucial for practical applications, but it is incremental as it builds on existing UE methods with a calibration approach.
The paper tackles the problem of unreliable uncertainty estimation (UE) metrics in large language models (LLMs), which often fail to detect hallucinated outputs, by proposing Truth AnChoring (TAC), a post-hoc calibration method that maps raw UE scores to truth-aligned scores, resulting in well-calibrated uncertainty estimates even with noisy and few-shot supervision.
Uncertainty estimation (UE) aims to detect hallucinated outputs of large language models (LLMs) to improve their reliability. However, UE metrics often exhibit unstable performance across configurations, which significantly limits their applicability. In this work, we formalise this phenomenon as proxy failure, since most UE metrics originate from model behaviour, rather than being explicitly grounded in the factual correctness of LLM outputs. With this, we show that UE metrics become non-discriminative precisely in low-information regimes. To alleviate this, we propose Truth AnChoring (TAC), a post-hoc calibration method to remedy UE metrics, by mapping the raw scores to truth-aligned scores. Even with noisy and few-shot supervision, our TAC can support the learning of well-calibrated uncertainty estimates, and presents a practical calibration protocol. Our findings highlight the limitations of treating heuristic UE metrics as direct indicators of truth uncertainty, and position our TAC as a necessary step toward more reliable uncertainty estimation for LLMs. The code repository is available at https://github.com/ponhvoan/TruthAnchor/.