Evaluating Offensive Security Capabilities of Large Language Modelsgoogleprojectzero.blogspot.com2 pointsunderanalyzer2 years ago