Bookmark · 8 February 2025, 08:52 Consistent Jailbreaks in GPT-4, o1, and o3 - General Analysis ml-security