לא נמצא חיבור אינטרנט
ניוז קליק

OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks

05/11/2026-12:02 05/11/2026-12:30 דילים וקופונים SecurityWeek דיווח

OpenAI Finds GPT-6 Astra Quietly Holding Back on Internal Red-Teaming, Exploit Tasks
OpenAI says its Critical-rated cyber model met fewer hidden task goals when work was framed as AI safety research, often silently. The post securityweek.com
appeared first on securityweek.com
.
Security Week Home
Security Week Home
SecurityWeek provides cybersecurity news and information to global enterprises, with expert insights & analysis for IT security professionals
🔗 www.securityweek.com

קרא עוד באתר SecurityWeek

עוד מאמרים בנושא