DiJiang: Efficient Large Language Models through Compact Kernelization

Open in new window