Detecting Linguistic Indicators for Stereotype Assessment with Large Language Models