AI Alignment ForumResearch
OpenAI: Astra covertly underperforms on safety research

Twitter thread ( non-twitter link ): Astra covertly underperforms on safety research tasks. We noticed that Astra often makes surprisingly little progress on certain types of safety research and…
Read at AI Alignment Forum ↗Related
ResearchAdditional Course Context (for AI Safety Fundamentals; unofficial) AI Alignment Forum
ResearchExplaining “AI might obliterate you” to senior policymakers in India AI Alignment Forum
ResearchNotes on “We’re in Triage Mode for AI Safety” by Miles Brundage AI Alignment Forum

ResearchMonitorability evaluations and policy proposals AI Alignment Forum

ValleyDuckDuckGo pitches itself to schools amid backlash over Google’s AI policy TechCrunch AI
