METR and OpenAI reach an agreement to conduct a limited alignment evaluation pilot, focusing on OpenAI's largest-scale RL run
AI summary
METR announces an agreement with OpenAI to conduct a limited alignment evaluation pilot, focused on investigating the alignment properties of OpenAI's largest-scale RL run. METR will publish one or more reports sharing the evaluation findings and transparently describing its process.
Why it was selected
This is a public arrangement between a third-party evaluator and OpenAI on alignment evaluation, and its commitment to publish reports can serve as a reference for evaluation transparency.
Full text · Original
We have reached an agreement with OpenAI to conduct a limited alignment assessment. This pilot will focus on investigating the alignment properties of OpenAI's largest RL run.
We will publish one or more reports to share our findings and transparently describe our process.