AlpacaEval
#1
by
jhartman
- opened
Hi, this model looks pretty incredible - closes all the gaps in the bagel models. Thanks for releasing it!
Have you considered adding it to the LMSYS leaderboard? It is the best approach to evaluation vs GPT-4 and via human annotation. It would be great to see if we have another ChatGPT level open model along with Mixtral.
Have you seen this:
https://huggingface.co/abacusai/MetaMath-Bagel-DPO-34B
Its the improved version of this model. We have an additional improvement that we have published but have not yet finished documenting fully.