Tutorials

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 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 impr

DGX agentx-post
tutorialsopenai--x

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.

Related

Source: OpenAI (X) | 2026-07-21

Loading related sources…