MUSE: A Comprehensive AI Framework for Evaluating Machine Learning Unlearning in Language Models
https://arxiv.org/abs/2407.06460 Language models (LMs) face significant privacy and copyright challenges as they are trained on vast amounts of text data. The inadvertent inclusion of private or copyrighted content in training…