I wrote a post which tries to explain why Jonathan's paper on Catalangate is wrong and dangerous (and why I care about this).
medium.com/@maldr0id/misi…
New paper: "ICON: Intent-Context Coupling for Efficient Multi-Turn Jailbreak Attack"
We find LLM safety constraints are significantly relaxed when malicious intent is coupled with a semantically congruent context.
97.1% ASR across 8 LLMs.
arxiv.org/abs/2601.20903