Publicación automática · Texto en el idioma de la fuente. OpenAI researchers are testing “confessions,” a method that trains models to admit when they make mistakes or act undesirably, helping improve AI honesty, transparency, and trust in model outputs.
How confessions can keep language models honest

Imagen ilustrativa: NODO · Identidad del sitio
Fuente principal:
OpenAI · News →
OpenAI · News →