Anthropic published an article giving 10 recommendations on the design, running and reporting of safety evaluations
AI summary
Anthropic published an article on its engineering blog giving 10 recommendations on how to design, run and report safety evaluations. The article says safety evaluations are critical to judging whether a model can be deployed, which guardrails should be built, and where training and research should focus. Original source link: https://www.anthropic.com/research/building-strong-safety-evaluations
Full text · Original
Safety evaluations are critical for understanding whether a model is ready to deploy, which safeguards to build for it, and where to focus training and research.
On our Engineering Blog, we offer 10 recommendations for how to design, run and report them.
https://www.anthropic.com/research/building-strong-safety-evaluations