Key Info
OpenAI is working with Apollo Research, METR, Transluce and Goodfire on embedded assessments, giving outside experts deep access to its internal models and systems to investigate alignment and test safeguards.
Highlights
- The program builds on OpenAI's existing alignment and safety research and brings independent scrutiny to how risks are identified and managed.
- Transluce will assess the quality and coverage of OpenAI's post-training alignment evaluations, develop additional tests, recommend improvements for gaps, and plans to expand the pilot in 2027.
- Goodfire will embed interpretability researchers who use its platform to analyze model internals, focusing on training run monitoring, debugging and alignment.
- The work is intended to help OpenAI researchers identify issues and make safety improvements.