Scalable Extraction of Training Data from (Production) Language Models

Nasr, Milad, Carlini, Nicholas, Hayase, Jonathan · arXiv · 2023