FactSelfCheck: Fact-Level Black-Box Hallucination Detection for LLMs
This addresses the challenge of ensuring factuality in LLM applications, offering a more granular detection approach than existing methods, though it is incremental in advancing detection techniques.
The paper tackles the problem of hallucinated content in Large Language Models by proposing FactSelfCheck, a black-box method for fact-level hallucination detection, which improves hallucination correction by achieving a 35.5% increase in factual content compared to baselines.
Large Language Models (LLMs) frequently generate hallucinated content, posing significant challenges for applications where factuality is crucial. While existing hallucination detection methods typically operate at the sentence level or passage level, we propose FactSelfCheck, a novel black-box sampling-based method that enables fine-grained fact-level detection. Our approach represents text as knowledge graphs consisting of facts in the form of triples. Through analyzing factual consistency across multiple LLM responses, we compute fine-grained hallucination scores without requiring external resources or training data. Our evaluation demonstrates that FactSelfCheck performs competitively with leading sentence-level sampling-based methods while providing more detailed insights. Most notably, our fact-level approach significantly improves hallucination correction, achieving a 35.5% increase in factual content compared to the baseline, while sentence-level SelfCheckGPT yields only a 10.6% improvement. The granular nature of our detection enables more precise identification and correction of hallucinated content. Additionally, we contribute a new dataset for evaluating sampling-based methods - FavaMultiSamples.