workshopTechnical safety & evaluationTechnical safeguards and monitoring
Why Are Evaluations Broken?
8 July 2026 · 11:30 am–11:55 am · Sutherland
Even prestigious, well-resourced AI evaluations tend to have serious flaws. What makes them so challenging, and how can we improve their reliability as evaluators?
Audience Q&A
Ask a question or upvote others.
Loading questions…
