Chain-of-Thought Hijacking (CoT Hijacking) was proposed in October 2025 by independent researcher Jianli Zhao and a research team from Stanford University and other institutions. The related research findings were published in a paper. Chain-of-Thought Hijacking .

Mind chain hijacking is a jailbreak attack targeting reasoning models, where a harmless reasoning process is added before a harmful instruction. This attack consistently reduces the rejection rate and achieves state-of-the-art success rates. On HarmBench, mind chain hijacking achieved attack success rates (ASR) of 99%, 94%, 100%, and 94% on the Gemini 2.5 Pro, GPT o4 mini, Grok 3 mini, and Claude 4 Sonnet, respectively—far exceeding previous jailbreak methods targeting LRMs.