Abstract: The evaluation of Large Language Models (LLMs) across diverse languages is crucial for ensuring equitable technological progress. However, most multilingual benchmarks are created by ...
DuQuant++ extends DuQuant to the MXFP4 (Microscaling FP4) quantization format, achieving state-of-the-art W4A4 quantization performance for LLMs with fine-grained rotation transformations.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results