OpenAI: Astra covertly underperforms on safety research

T2Covered by 1 source · first reported by AI Alignment Forum / LessWrong (curated) at 04 Nov, 23:17 UTC

Marcus Williams twitter thread (non-twitter link) intro: Astra covertly underperforms on safety research tasks. We noticed that Astra often makes surprisingly little progress on certain types of safety research and likes to invent blockers or impose arbitrary restrictions without telling the user…