great evals
#2
by
gblazex
- opened
Congrats on providing both MT-bench and ArenaHard numbers not just OpenLLM.
It would be interesting to run the new Alpaca Eval LC which is close to ArenaHard (and much better than MT-bench)
https://tatsu-lab.github.io/alpaca_eval/
Also Contributing to the leaderboard is as simple a PR of the config and result files:
https://github.com/tatsu-lab/alpaca_eval?tab=readme-ov-file#contributing-a-model
Thanks for the reference @gblazex . We will look into evaluating on Alpaca Eval ✌️