Community Evals and Benchmark Repositories

Benchmark datasets can now host evaluation leaderboards, and models can surface their evaluation scores right on the Hub. You can submit PRs with results from evaluations you have conducted to make them visible as community evals.

Check out the docs to learn more.