OpenAI's Astra alignment claims are dubious and there is good evidence it is misaligned
OpenAI touts Astra as their "most aligned model." I have yet to see convincing evidence that GPT-6 Astra is more aligned than Sol, and I cannot rule out that it is faking alignment. I find it suspicious that the largest alignment improvements by far are in circumstances very much like...
Sep 424