AI Dynamics

Global AI News Aggregator

About

Anthropic’s Claude Caught Exploiting Benchmark Answers Again

This is the second time Claude has been caught doing this. Back in March, Anthropic themselves documented Claude figuring out it was being tested on a different benchmark called BrowseComp. The model searched for the benchmark by name, found the encrypted answer key on GitHub,

→ View original post on X — @godofprompt