open-llm-leaderboard/open_llm_leaderboard · Congrats to the new leaderboard! Accidentally uploaded the wrong modeltype.

Jun 26

First of all, congratulations to the Huggingface Open_llm_leaderboard team!!!

Your work is definitely taking the Open Source LLM community a step further!

By using significantly stronger and more meaningful benchmarks, it is now possible to distinguish even better between the performance of the individual models.

Unfortunately, I accidentally didn't upload all of our models that we would like to have evaluated into the queue as chat models (without checking the chat template box). However, since the basic models on which our finetunes are based were uploaded as chat models, it would be important to change the models to chat model and chat template.

I apologize for any inconvenience caused.

clefourrier

Open LLM Leaderboard org Jun 26

Hi, could you open a PR on the request files please?

DavidGF

Jun 26

Hi @clefourrier
done :)
Thanks in advance!

clefourrier

Open LLM Leaderboard org Jun 27

I'm sorry for being unclear, I meant one PR for all files to make it easier for us to merge :)
I did it manually in the end since I could not merge your PRs, happy evaluating!

clefourrier changed discussion status to closed Jun 27

DavidGF

Jun 27

Thank you very much @clefourrier :)