Spaces:
Running
Running
Fantastic!
#6
by
froggeric
- opened
Thank you for creating another great benchmark, and sharing the leaderboard. Now we are beginning to have a few quality specialised leaderboard, which I consider a lot more useful and trustworthy than the like of HF Open LLM leaderboard, maybe it is time we consolidate those results into a meta leaderboard?
I wish, but the main difficulty with a meta leaderboard is that the leaderboards it's made up of must have many of the same models tested, at least if you're trying to make an average score. Many leaderboards either don't have many models tested, only focus on certain types of models (like 70B+ params or writing quality), or haven't been updated in a while.