Securing our frontier model training environment

Company news OpenAI Source language: English

OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.

Excerpt

OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.

Short excerpt from the source — open the original for the full text.

Sources

← Back to the daily brief