gradientai

/

Llama-3-8B-Instruct-Gradient-1048k

like 607

Hello,
I have tryed llama 3 and it does not works as good as mistral so far . Ironically enough , when I asked who are your authors , it answered "ChatGPT" . Also it have problems with math representations ... Is it possible to build extended versions of mistral instruct v0.2 with context 128k , 256 k , 512 k , 1024 k ? If yes please try to make some versions of it. Lots of peoples will love it ! Thank you very much !