arXiv cs.AI· Ponhvoan Srey, Quang Minh Nguyen, Xiaobao Wu, Anh Tuan Luu·· 12 小时前AI 评分31
面向大语言模型可靠真值对齐不确定性估计的 Truth AnChoring（TAC）方法
Towards Reliable Truth-Aligned Uncertainty Estimation in Large Language Models
AI 导读
研究提出 Truth AnChoring（TAC），一种轻量级事后校准方法，将 LLM 的原始不确定性估计（UE）分数映射为与事实正确性对齐的概率，在监督信号含噪且样本极少时仍能持续改善校准。
正文 · 原文
Abstract:Uncertainty estimation (UE) aims to quantify uncertainty in large language model (LLM) outputs to flag potential hallucinations and support reliable deployment. However, UE metrics often exhibit unstable performance across models and datasets, limiting their practical utility. We formalise this phenomenon as proxy failure, arising because UE metrics are derived from model behaviour rather than being explicitly grounded in factual correctness. We show that when mutual information between a UE score and correctness is low, the score's discriminability must be close to chance. To address this, we propose Truth AnChoring (TAC), a lightweight post-hoc calibration method that maps raw UE scores to truth-aligned correctness probabilities. Even with noisy and few-shot supervision, TAC continues to improve calibration, providing a practical calibration protocol. Our findings highlight the limitations of treating heuristic UE metrics as direct indicators of uncertainty about factual correctness, and position our TAC as a principled step toward more reliable uncertainty estimation for LLMs. The code repository is available at this https URL.
| Comments: | Accepted to EMNLP Findings 2026 |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2604.00445 [cs.AI] |
| (or arXiv:2604.00445v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2604.00445 arXiv-issued DOI via DataCite |
Submission history
From: Ponhvoan Srey [view email]
[v1]
Wed, 1 Apr 2026 03:42:16 UTC (6,423 KB)
[v2]
Fri, 16 Oct 2026 14:40:44 UTC (6,432 KB)
来源：arXiv cs.AI · arxiv.org