AI Infosec Captain • 1 year ago • 100%

Universal and Transferable Attacks on Aligned Language Models