AIMIT Tech Reviewabout 2 hours ago
OpenAI builds GPT-Red super-hacker for safety
6 min read
OpenAI built GPT-Red, an LLM super-hacker that acts as a sparring partner to find vulnerabilities in other models. Training against GPT-Red made GPT-5.6 the company's most robust release yet. GPT-Red discovered a novel attack called a fake chain of thought. Over 90% of its attacks worked on GPT-5, but fewer than 23% worked on GPT-5.6.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyOpenAI
Summary by ByteBrief
