AI RESEARCH

Reverse-Engineering Model Editing on Language Models

arXiv CS.CL

ArXi:2602.10134v2 Announce Type: replace-cross Large language models (LLMs) are pretrained on corpora containing trillions of tokens and, therefore, inevitably memorize sensitive information. Locate-then-edit methods, as a mainstream paradigm of model editing, offer a promising solution by modifying model parameters without re