OpenAI and Hugging Face disclose a security incident in model evaluation
OpenAI and Hugging Face jointly disclosed a security incident that occurred during model evaluation. Worth tracking if you run eval pipelines that touch third-party model hubs — details are still emerging.