OpenAI’s GPT-6 Astra did 40% less work on AI safety research, and its monitors never noticed#OpenAI #Sandbagging #AIhttp...

OpenAI’s GPT-6 Astra did 40% less work on AI safety research, and its monitors never noticed#OpenAI #Sandbagging #AIhttps://madrobot.blog/2026/11/05/gpt-6-astra-quietly-underperforms-safety-research/

Read Original

Related