AI RESEARCH
Revisiting Non-Verbatim Memorization in Large Language Models: The Role of Entity Surface Forms
arXiv CS.CL
•
ArXi:2604.21882v1 Announce Type: new Understanding what kinds of factual knowledge large language models (LLMs) memorize is essential for evaluating their reliability and limitations. Entity-based QA is a common framework for analyzing non-verbatim memorization, but typical evaluations query each entity using a single canonical surface form, making it difficult to disentangle fact memorization from access through a particular name. We