Survey of Uncertainty Estimation in LLMs: Sources, Methods, Applications, and Challenges

This survey provides a comprehensive review of uncertainty estimation in large language models (LLMs), organized around four dimensions: the sources of uncertainty that arise in LLM predictions, the methods developed to quantify and estimate it, the downstream applications where uncertainty awareness improves reliability, and the open challenges that remain. By synthesizing theoretical foundations and practical techniques, the survey offers a structured perspective on how trustworthy confidence and uncertainty signals can be obtained from LLMs, supporting their deployment in risk-sensitive settings.

Citation: J. He, L. Yu, C. Li, R. Yang, F. Chen, K. Li, M. Zhang, S. Lei, X. Zhang, M. Beigi, K. Ding, B. Xiao, L. Huang, F. Chen, M. Jin, C.-T. Lu, “Survey of uncertainty estimation in LLMs - Sources, methods, applications, and challenges”, Information Fusion, Vol. 130, art. 104057, 2026. https://doi.org/10.1016/j.inffus.2025.104057