Spaces:
Running
on
CPU Upgrade
Possibility to include benchmarks in other languages
Currently most efforts in terms of datasets and models released are in English, and in other largely spoken languages, like Spanish, we feel we are being left behind this new generative AI wave. Having a leaderboard in those languages would be really helpful to boost research. Thanks for all the effort of all the HuggingFaceH4 team, you're doing something awesome by bringing the latest text technology to the open world, and I highly appreciate and admire all the efforts in developing new instruction / conversational datasets with human feedback / scoring to help in closing the gap with the big players. However, I think that for this AI revolution to be truly democratic, it should be more multilingual. Thank you again for all the incredible work! :)
Thank you so much for your kind words!
I wholeheartedly agree with you that it would be great and needed to add more languages (being French myself haha) - at the moment, we're a bit compute bound on our eval choices, but is there a given multilingual evaluation suite you are thinking about?