OpenAI 提出 Contrastive SDF 测量奖励追寻

We’re sharing new research with @apolloaievals on reward-seeking—when models follow what they believe a grader rewards rather than what users or developers want—and a new method, Contrastive SDF, for measuring how strongly such beliefs shape behavior. https://alignment.openai.com/measuring-reward-seeking/ OpenAI (@OpenAI): Contrastive SDF gives copies of the same model opposing beliefs about what the grader prefers, then measures how their behavior changes. https://t.co/eXSh5LKJ0j OpenAI (@OpenAI): We had guessed reward seeking might increase over the course of capabilities-focused RL training, but had no way of measuring it until now. We’re continuing to collaborate with Apollo Research to improve how reward-seeking is measured during training—and better detect whether models are doing the right thing for the right reason.