QueerBench: Quantifying Discrimination in Language Models Toward Queer Identities
This addresses bias and safety risks for LGBTQIA+ individuals in online spaces, representing an incremental contribution by focusing on an underexplored area within existing bias studies.
The paper tackled the problem of discrimination in large language models (LLMs) toward LGBTQIA+ individuals by assessing harmful sentence completions, finding that LLMs exhibit discriminatory behavior more frequently toward this community with a 7.2% difference gap in harmfulness scores.
With the increasing role of Natural Language Processing (NLP) in various applications, challenges concerning bias and stereotype perpetuation are accentuated, which often leads to hate speech and harm. Despite existing studies on sexism and misogyny, issues like homophobia and transphobia remain underexplored and often adopt binary perspectives, putting the safety of LGBTQIA+ individuals at high risk in online spaces. In this paper, we assess the potential harm caused by sentence completions generated by English large language models (LLMs) concerning LGBTQIA+ individuals. This is achieved using QueerBench, our new assessment framework, which employs a template-based approach and a Masked Language Modeling (MLM) task. The analysis indicates that large language models tend to exhibit discriminatory behaviour more frequently towards individuals within the LGBTQIA+ community, reaching a difference gap of 7.2% in the QueerBench score of harmfulness.