StereoMap: Quantifying the Awareness of Human-like Stereotypes in Large Language Models