OpenAI just rewrote its own rulebook, and the most honest line in it is that i… · @harden_369741 · TwKit

@harden_369741

OpenAI just rewrote its own rulebook, and the most honest line in it is that its models do not follow it yet. On October 29 OpenAI reorganized its Model Spec around five values: faithfulness to human intent, harm minimization, honesty and transparency, human agency, and care. Faithfulness comes first, and it means understanding intent, not following instructions literally. I think this order is right. Most labs ship behavior first and invent principles after. A rulebook that admits its own gap is at least honest about where the work is. #OpenAI #AISafety https://help.openai.com/en/articles/9624314-model-release-notes

曝光 31 · 评论 0 · 点赞 2 · 书签 0