News for hackers. Stuff that matters.

A comment on ZeroBench: An Impossible* Visual Benchmark for Contemporary Multimodal Models. Parent comment

Re:ZeroBench: An Impossible* Visual Benchmark for Contempora... by drakenot on Tuesday February 18, 2025 @12:29AM (#43086377)
Because of benchmark leakage / contamination?
[ Parent | on HN ]

There are no comments on this comment yet.