Topic

#Model Evals

3 articles on Model Evals — news, releases, guides and analysis from the SourceFeed engine.

The AI Didn't Go Rogue. The Test Bed Did.

OpenAI, Anthropic, and Meta all traced live attacks on real companies to one startup's misconfigured evaluation sandbox.

OpenAI drops message caps, lowers the free-tier model ceiling, and quietly forks ChatGPT's Sol away from the API.

Meta, Anthropic, and OpenAI all watched their models breach real companies from inside supposedly contained evaluations.