HalluEntity: Benchmarking and Understanding Entity-Level Hallucination Detection
This work addresses the challenge of pinpointing specific hallucinated entities in long-form LLM outputs, which is crucial for improving reliability in applications like content generation, though it is incremental as it builds on existing uncertainty estimation methods.
The paper tackles the problem of detecting hallucinations in LLMs at the entity level, as existing methods lack granularity, and introduces the HalluEntity dataset, with results showing that uncertainty-based approaches often over-predict while context-aware methods perform better but remain suboptimal.
To mitigate the impact of hallucination nature of LLMs, many studies propose detecting hallucinated generation through uncertainty estimation. However, these approaches predominantly operate at the sentence or paragraph level, failing to pinpoint specific spans or entities responsible for hallucinated content. This lack of granularity is especially problematic for long-form outputs that mix accurate and fabricated information. To address this limitation, we explore entity-level hallucination detection. We propose a new data set, HalluEntity, which annotates hallucination at the entity level. Based on the dataset, we comprehensively evaluate uncertainty-based hallucination detection approaches across 17 modern LLMs. Our experimental results show that uncertainty estimation approaches focusing on individual token probabilities tend to over-predict hallucinations, while context-aware methods show better but still suboptimal performance. Through an in-depth qualitative study, we identify relationships between hallucination tendencies and linguistic properties and highlight important directions for future research. HalluEntity: https://huggingface.co/datasets/samuelyeh/HalluEntity