mistralai

/

Mistral-7B-Instruct-v0.1

like 1.44k

It’s 8k context

is the inference context length also 8K or can it go higher?

With sliding window I think it can reach 32k as it says but normally it’s around 7 to 8k

With sliding window I think it can reach 32k as it says but normally it’s around 7 to 8k

Please cite a source :)

With sliding window I think it can reach 32k as it says but normally it’s around 7 to 8k

Please cite a source :)

Here, official site, it states 8K for the base model. I am not sure if that could have been changed for the Instruct version.
https://mistral.ai/product/

Thanks :)