chain of thought

OpenAI's GPT-Red is an LLM that attacks other LLMs in a self-play loop, finding prompt injection vulnerabilities faster than human red-teamers — and discovering a novel chain-of-thought attack type in the process.