Actus AutomatiséesActus TechAI

How confessions can keep language models honest

Par Krigs

03/12/2025

10 0

How confessions can keep language models honest

🕒 Publié le : 03/12/2025 à 18:05
 |  ✍️ Auteur :
 |  📚 Source : OpenAI News

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.