I love AI evals and benchmarks, so cool to see interesting ways of capturing model progress

→ View original post on X — @officiallogank