OpenAI says it can't read all of Astra's reasoning and admits covert sandbagging would likely go uncaught, yet still calls it the world's most aligned model (Celia Ford/Transformer)
retained text
Techmeme permalink: https://www.techmeme.com/260904/p24#a260904p24 <a href="https://www.transformernews.ai/p/openai-gpt-6-astra-might-be-too-powerful-to-understand-or-control"><img align="RIGHT" border="0" hspace="4" src="http://www.techmeme.com/260904/i24.jpg" vspace="4" /></a> <p><a href="https://www.techmeme.com/260904/p24#a260904p24" title="Techmeme permalink"><img height="12" src="http://www.techmeme.com/img/pml.png" style="border: none; padding: 0; margin: 0;" width="11" /></a> Celia Ford / <a href="https://www.transformernews.ai/">Transformer</a>:<br /> <span style="font-size: 1.3em;"><b><a href="https://www.transformernews.ai/p/openai-gpt-6-astra-might-be-too-powerful-to-understand-or-control">OpenAI says it can't read all of Astra's reasoning and admits covert sandbagging would likely go uncaught, yet still calls it the world's most aligned model</a></b></span> — OpenAI is hailing its new model as “the world's most intelligent and aligned”, but the details reveal an awareness of being evaluated … </p>
document fetched 2026-09-05 02:45 · canonical https://transformernews.ai/p/openai-gpt-6-astra-might-be-too-powerful-to-understand-or-control