Operating agenda
- Define anti-gaming evaluations and monitoring protocols.
- Connect formal reasoning, red-team datasets, and deployment governance.
- Publish reproducible task environments for young talent.
A virtual lab for turning AI safety claims into testable, auditable systems.