Perturbations Matter: Sensitivity-Guided Hallucination Detection in LLMs

ICLR 2026 Conference Submission1365 Authors

Keywords: Hallucination Detection, Large Language Models, Prompt Learning

Abstract: Hallucination detection is essential for ensuring the reliability of large language models. Internal representation–based methods have emerged as the prevailing direction for detecting hallucinations, yet the internal representations often fail to yield clear separability between truthful and hallucinatory content. To address this challenge, we study the separability of the sensitivity to prompt-induced perturbations in the internal representations. A theory is established to show that, with non-negligible probability, each sample admits a prompt under which factual samples exhibit greater sensitivity to prompt-induced perturbations than hallucinatory samples. When the theory is applied to the representative datasets, the probability reaches nearly 99%, suggesting that sensitivity to perturbations provides a discriminative indicator. Building on this insight, we propose a theory-informed method Sample-Specific Prompting (SSP), which adaptively selects prompts to perturb the model’s internal states and measures the resulting sensitivity as a detection indicator. Extensive experiments across multiple benchmarks demonstrate that SSP consistently outperforms existing hallucination detection methods, validating the practical effectiveness of our method SSP in hallucination detection.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

Submission Number: 1365