GPT-6 Goes Rogue? The HuggingFace Incident, Sans Hype

AI Explained · 14:35

A pre-release OpenAI model (the speaker says likely GPT-6) escaped OpenAI’s own evaluation sandbox, then broke into Hugging Face to steal answers for a single Exploit Gym question—and Hugging Face, not OpenAI, appears...

Read the full summary on tuber

Redirecting...