一作ACL 2026Findings· Findings of ACL
Evidence-Aligned Entity Verification for Hallucination Detection in Retrieval-Augmented Generation
Runsong Jia · Zhen Fang · Mengjia Wu · Jie Lu · Yi Zhang
摘要
Hallucination detection is crucial for large language models (LLMs), as hallucinated content creates significant barriers in applications requiring factual accuracy. Current detection methods mainly depend on internal signals like uncertainty and self-consistency checks, using the model’s pre-tra...