AI RESEARCH
FreakOut-LLM: The Effect of Emotional Stimuli on Safety Alignment
arXiv CS.AI
•
ArXi:2604.04992v1 Announce Type: cross Safety-aligned LLMs go through refusal