The evidence that Astra is more aligned than prior AIs seems dubious to me. Evidence appears consistent with the AI being as or more interested in score-seeking at the expense of user intent, but having beliefs+instincts that the scorer will catch a broader range of cheating behaviors. At a more basic level, the AI is extremely evaluation-aware and much less monitorable than prior AIs, making detecting misalignment much more difficult. If you train against specific reward hacks you ended up…

OpenAI@OpenAISep 3

Astra is our most aligned model, with substantial improvements in understanding user intent. https://t.co/rY0An7o3rq