NEXUS: Network Exploration for eXploiting Unsafe Sequences in Multi-Turn LLM Jailbreaks

Explore NEXUS, a modular framework for systematically exploring adversarial query spaces in multi-turn LLM jailbreaks using semantic networks and gradient-ba...

Level: advanced

By Javad Rafiei Asl, Sidhant Narula, Mohammad Ghasemigol, Eduardo Blanco, Daniel Takabi

Category: discussion