Locally reproducible Leaderboard (v2) scores

#946
by ekurtic - opened

Hi,
Do you plan to open-source the full evaluation pipeline such that it's locally reproducible?

Open LLM Leaderboard org

Hi!
It's already open source, you can follow the steps in our doc, reproducibility section :)
You'll also need to look at the normalization page to go from the raw to the normalised scores.

clefourrier changed discussion status to closed

Sign up or log in to comment