Fetching the paper…

Rethinking Channel Dimensions to Isolate Outliers for Low-bit Weight Quantization of Large Language Models · Around