作者位次 #3/6ACL 2026Findings· Findings of ACL
Understanding New-Knowledge-Induced Factual Hallucinations in LLM s: Analysis and Interpretation
Renfei Dang · Peng Hu · Zhejian Lai · Changjiang Gao · Min Zhang · Shujian Huang
摘要
Prior works have shown that fine-tuning on new knowledge can induce factual hallucinations in large language models (LLMs), leading to incorrect outputs when evaluated on previously known information. However, the specific manifestations of such hallucination and its underlying mechanisms remain...