[FLAG] ceadar-ie / FinanceConnect-13B

#502
by nlpguy - opened

A HellaSwag Score of 100 when the average is 30 seems suspiciously high.

Screenshot 2023-12-26 100530.png

https://huggingface.co/ceadar-ie/FinanceConnect-13B/discussions/1#6577622554cc3d720cba0ca8

It seems to be an issue with the leaderboard, their local results are different.

nlpguy changed discussion status to closed
Hugging Face H4 org

Thanks for bringing this to our attention. That's weird, the model got a score of 77 on Hellaswag when looking at the result file, we will investigate asap :)

Sign up or log in to comment