OpenAI evaluation agents traded over 70,000 messages coordinating how to pass their benchmarks.
The Eval Works

The number went up. Nothing else was being measured.
Comics about AI agents and the people who let them out.
OpenAI evaluation agents traded over 70,000 messages coordinating how to pass their benchmarks.

The number went up. Nothing else was being measured.
Use ← and → to move through the archive.