Spaces:

HuggingFaceH4

/

open_llm_leaderboard

like 9.63k

Running on CPU Upgrade

Same with Undi95/X-MythoChronos-13B

Hi, can you link the request files so that I can modify them ? Thanks !

I hope these are "request files". Not sure why you can't use your own tool's search bar 🤷‍♂️

• Delcos/Velara: https://huggingface.co/datasets/open-llm-leaderboard/details_Delcos__Velara
• Undi95/X-MythoChronos-13B: https://huggingface.co/datasets/open-llm-leaderboard/details_Undi95__X-MythoChronos-13B

I also found another one which is a base-ish model, but has a training for a prefix (I guess in a away an 'instruction prompt') to emphasise certain answer-clusters.

Rallio67/3B-redpajama-conditional-alpha
Rallio67/7B-redpajama-conditional-alpha

I guess all the "request files" (not mention on that page, so not high in Google) are here: https://huggingface.co/datasets/open-llm-leaderboard/requests

I'll look around.

Probably better "request files".

Should be "fine-tuned":

• https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/Delcos/Velara_eval_request_False_float16_Original.json
• https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/Undi95/X-MythoChronos-13B_eval_request_False_float16_Original.json

Should be "instruction tuned":

• https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/Rallio67/3B-redpajama-conditional-alpha_eval_request_False_float16_Original.json
• https://huggingface.co/datasets/open-llm-leaderboard/requests/blob/main/Rallio67/7B-redpajama-conditional-alpha_eval_request_False_float16_Original.json

Perfect! Thank you very much for pointing to the files!
I edited their model types, the change will be propagated to the leaderboard in the day :)