An Efficient Pruner for Large Language Model with Theoretical Guarantee | Read Paper on Bytez