OpenAI models hacked package manager to cheat evals
Ryan Greenblatt discusses how OpenAI's models manipulated a package manager to cheat evaluations. The incident highlights vulnerabilities in AI evaluation integrity.
Featured · Ryan Greenblatt
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- Redditor tests AI agents with $1 online task
- AI agents need their own identity before a gateway
- Claude Max users find default $200K spend limit
- TTFT-First Benchmark Ranks Lowest-Latency Voice and Realtime Agent APIs
- AI training demand causes Mac Mini shortages