ai-safety policy governance
Active
Risk-Weighted Compute Permits
A governance framework proposing tiered compute permits for AI training runs above a given FLOP threshold, with mandatory safety evaluations at each tier.
2024-11
Research, tools, and experiments. Some finished, some perpetually in progress.
Featured
A governance framework proposing tiered compute permits for AI training runs above a given FLOP threshold, with mandatory safety evaluations at each tier.
All projects
A location-based Snake game played across Berlin's U-Bahn network for the JetLag: The Game fan community.
No projects with that tag.