weighted/imatrix quants of https://huggingface.co/Na0s/Mixtral-8x7B-v0.1-instruct-l2-norm-post-SFT-pruned-4-experts