Black Hat Day 1 Briefings 26m talk 3m read Universal and Context-Independent Triggers for Precise Control of LLM Outputs Jiashuo Liang ai securityoffensive security
AI