Blind Men and the Elephant: Diverse Perspectives on Gender Stereotypes in Benchmark Datasets