Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
How did falcon become #1? And isn't it a bit overrated?
#72
by
Fredithefish
- opened
The leaderboard doesn't allow trust remote code did it get published manually?
In my tests falcon performs often not as good as it should be.
In logic tasks often the 7B RedPajama-INCITE-Chat model gave much better results.
Example is in the screenshots
My tests are based on the openchatkit hf space and the falcon-chat HF space
Yep, it got published manually!
Did you see our blog post on model scoring and MMLU? They could be interesting to you
clefourrier
changed discussion status to
closed