1 paper
Yuefei Chen, Yihao Quan, Xiaodong Lin +1
LLMs frequently generate fictitious yet convincing citations, often expressing high confidence even when the underlying reference is wrong. We study this failure across 9 models an…