Probing LLM Hallucination from Within: Perturbation-Driven Approach via Internal Knowledge

作者：Seongmin Lee, Hsiang Hsu, Chun-Fu Chen, Chau, Duen Horng · 发表于：arXiv (Cornell University) · 年份：2024 · DOI：10.48550/arxiv.2411.09689 · 被引用次数：3 · 研究领域：Cryptography and Residue Arithmetic、Logic, Reasoning, and Knowledge、Logic, programming, and type systems

LLM hallucination, where unfaithful text is generated, presents a critical challenge for LLMs' practical applications. Current detection methods often resort to external knowledge, LLM fine-tuning, or supervised training with large hallucination-labeled datasets. Moreover, these approaches do not distinguish between different types of hallucinations, which is crucial for enhancing detection performance. To address such limitations, we introduce hallucination probing, a new task that classifies LLM-generated text into three categories: aligned, misaligned, and fabricated. Driven by our novel discovery that perturbing key entities in prompts affects LLM's generation of these three types of text differently, we propose SHINE, a novel hallucination probing method that does not require external knowledge, supervised training, or LLM fine-tuning. SHINE is effective in hallucination probing across three modern LLMs, and achieves state-of-the-art performance in hallucination detection, outperforming seven competing methods across four datasets and four LLMs, underscoring the importance of probing for accurate detection.