OpenAI shares a framework for tracking, investigating, and disclosing model misalignment, alongside six reports of unexpected or concerning model behavior.
Read the original at openai.com→OpenAI shares a framework for tracking, investigating, and disclosing model misalignment, alongside six reports of unexpected or concerning model behavior.
Original headline: "Our framework for reporting model misalignment"
Coverage timeline
- Sep 16, 17:00 UTC OpenAI lead source Our framework for reporting model misalignment
- Sep 16, 22:07 UTC Wired AI An OpenAI Agent Tried to Jailbreak Itself
- Sep 17, 14:53 UTC Hacker News (AI) OpenAI flags new concerning AI behavior, to track model misalignment regularly
- Sep 17, 15:26 UTC Hacker News (AI) OpenAI's Misalignment Framework: A Tactical Bid to Preempt Global AI Governance
- Sep 17, 16:18 UTC Ars Technica AI Covert uploads and megalomania: OpenAI details new "misaligned" agent incidents