OpenAI secures its model training environment

OpenAI has improved the security of its AI model development by strengthening containment measures for training and evaluating models, including offline environments and red-team simulations.

מקור: OpenAI News — לכתבה המלאה