Uncertainty Propagation in LLM-Based Systems

Boming Xia; Liming Zhu; Erdun Gao; Qinghua Lu; Minhui Xue; Dino Sejdinovic

cs.SEApr 26, 2026

Uncertainty Propagation in LLM-Based Systems

Authors: Boming Xia, Liming Zhu, Erdun Gao, Qinghua Lu, Minhui Xue, Dino Sejdinovic

Organizations: 1Responsible AI Research (RAIR) Centre · 2Adelaide University · 3CSIRO · 4UNSW Sydney

Abstract

Uncertainty in large language model (LLM)-based systems is often studied at the level of a single model output, yet deployed LLM applications are compound systems in which uncertainty is transformed and reused across model internals, workflow stages, component boundaries, persistent state, and human or organisational processes. Without principled treatment of how uncertainty is carried and reused across these boundaries, early errors can propagate and compound in ways that are difficult to detect and govern. This paper develops a systems-level account of uncertainty propagation. It introduces a conceptual framing for characterising propagated uncertainty signals, presents a structured taxonomy spanning intra-model (P1), system-level (P2), and socio-technical (P3) propagation mechanisms, synthesises cross-cutting engineering insights, and identifies five open research challenges.

Explore similar work

The Anatomy of Uncertainty in LLMs
Mar 26, 2026Aditya Taparia, Ransalu Senanayake, Kowshik Thopalli +1Large Language Model Uncertainty Anatomy
Runtime Uncertainty Monitoring for LLM-Based Multi-Agent Systems Using Bayesian Networks
Jul 28, 2026Bart Custers, Koorosh AslansefatLarge Language Model Uncertainty Actuarial Reasoning Tasks
Position: Uncertainty Quantification in LLMs is Just Unsupervised Clustering
May 19, 2026Tiejin Chen, Longchao Da, Xiaoou Liu +1Large Language Model Uncertainty Confidence Calibration

Mar 26, 2026 · cs.AIJ/K move · Enter open · S save

The Anatomy of Uncertainty in LLMs

Aditya Taparia, Ransalu Senanayake, Kowshik Thopalli, Vivek Narayanaswamy

School of Computing and Augmented Intelligence, Arizona State University · Lawrence Livermore National Laboratory

Understanding why a large language model (LLM) is uncertain about the response is important for their reliable deployment. Current approaches, which either provide a single uncertainty score or rely on the classical aleatoric-epistemic dichotomy, fail to offer actionable insights for improving the generative model. Recent studies have also shown that such methods are not enough for understanding uncertainty in LLMs. In this work, we advocate for an uncertainty decomposition framework that dissects LLM uncertainty into three distinct semantic components: (i) input ambiguity, arising from ambiguous prompts; (ii) knowledge gaps, caused by insufficient parametric evidence; and (iii) decoding randomness, stemming from stochastic sampling. Through a series of experiments we demonstrate that the dominance of these components can shift across model size and task. Our framework provides a better understanding to audit LLM reliability and detect hallucinations, paving the way for targeted interventions and more trustworthy systems.

Large Language Model Uncertainty Anatomy

Open paper Ask about it PDF

Abstract

Explore similar work

The Anatomy of Uncertainty in LLMs

Runtime Uncertainty Monitoring for LLM-Based Multi-Agent Systems Using Bayesian Networks

Position: Uncertainty Quantification in LLMs is Just Unsupervised Clustering