We introduce ProText, a dataset for measuring gendering and misgendering in stylistically numerous long-form English texts. ProText spans three dimensions: Theme nouns (names, occupations, titles, kinship phrases), Theme class (stereotypically male, stereotypically feminine, gender-neutral/non-gendered), and Pronoun class (masculine, female, gender-neutral, none). The dataset is designed to probe (mis)gendering in textual content transformations comparable to summarization and rewrites utilizing state-of-the-art Massive Language Fashions, extending past conventional pronoun decision benchmarks and past the gender binary. We validated ProText by means of a mini case examine, exhibiting that even with simply two prompts and two fashions, we are able to draw nuanced insights relating to gender bias, stereotyping, misgendering, and gendering. We reveal systematic gender bias, significantly when inputs comprise no express gender cues or when fashions default to heteronormative assumptions.
