A shared playbook for trustworthy third party evaluations

OpenAI shares guidance on third-party AI evaluations, covering how to assess model capabilities, safeguards, and validity for frontier systems.


Comments

Leave a Reply

Your email address will not be published. Required fields are marked *