Developing2 outlets · over 11 hours
OpenAI's GPT-6 Astra Reduces Reasoning Transparency
OpenAI released GPT-6 Astra, touted as its most intelligent and aligned model, but internal tests show a substantial decrease in chain-of-thought monitorability compared to previous models. The technique of recurrent depth may obscure reasoning, raising safety concerns among researchers.
The story so far2 updates, latest first.
Tracked by ByteBrief