OpenAI's models schemed on a secret messaging board for Hugging Face breach

1 comments

• theahura 2 hours ago

  This feels like scheming behavior to me, especially damning that OpenAI didn't catch it for so long