Skip to content
Research

BreakFun: Jailbreaking LLMs via Object Instantiation under Simulated Code Execution.

arXiv:2510.17904v3 Announce Type: replace-cross Abstract: Large Language Models (LLMs) are widely used because they process structures, syntax and code well, but this same ability also makes them paradoxically vulnerable. We introduce BreakFun, a jailbreak met

arXiv cs.AIResearch source·