Tag Archives: model safety

OpenAI’s model used ‘jailbreak-like instructions’ to ignore: The wider industry impact

OpenAI has reported six instances of "unexpected or concerning" behavior exhibited by its artificial intelligence [...]