Character sentence as the quantitative metrics : leveraging large language models to measure literary characterization