Researcher at METR
Based on 3 scored articles
Referenced in coverage
“Spent three weeks red teaming OpenAI's internal agent monitoring; produced 61-page report on vulnerabilities.”
“Spent three weeks stress-testing OpenAI's internal misalignment monitoring system, identifying multiple vulnerabilities and workarounds.”
“METR staff member who noted that measurements are extremely noisy and cautioned against taking latest Claude Opus results as gospel.”