Enabling efficient low-bit quantization based on matrix product operators for KV cache compression

Wang, Jia-Qi, Han, Xiao-Qi, Guo, Peng-Jie · Neural Networks · 2026