3 ms·AI researchers – 'virtually unlimited' ways to bypass Bard, ChatGPT safety rules15 points by thunderbong 3y agoxyzzy3000 3y agoThe paper: https://llm-attacks.org/ https://llm-attacks.org/archo 3y agohttps://archive.is/TTwPH https://archive.is/TTwPH