SeeGULL: A Stereotype Benchmark with Broad Geo-Cultural Coverage Leveraging Generative Models

TL;DR

SeeGULL is a comprehensive, globally diverse stereotype dataset created using large language models and diverse raters, addressing the Western-centric bias of existing datasets in NLP.

Contribution

The paper introduces SeeGULL, a broad-coverage stereotype benchmark with global representation, generated with LLMs and validated by diverse raters, filling a major gap in stereotype datasets.

Findings

SeeGULL covers stereotypes from 178 countries across 8 regions.

Global disparities in stereotypes are demonstrated through offensive scores.

Regional differences in stereotypes are identified between local and North American annotators.

Abstract

Stereotype benchmark datasets are crucial to detect and mitigate social stereotypes about groups of people in NLP models. However, existing datasets are limited in size and coverage, and are largely restricted to stereotypes prevalent in the Western society. This is especially problematic as language technologies gain hold across the globe. To address this gap, we present SeeGULL, a broad-coverage stereotype dataset, built by utilizing generative capabilities of large language models such as PaLM, and GPT-3, and leveraging a globally diverse rater pool to validate the prevalence of those stereotypes in society. SeeGULL is in English, and contains stereotypes about identity groups spanning 178 countries across 8 different geo-political regions across 6 continents, as well as state-level identities within the US and India. We also include fine-grained offensiveness scores for different stereotypes and demonstrate their global disparities. Furthermore, we include comparative annotations about the same groups by annotators living in the region vs. those that are based in North America, and demonstrate that within-region stereotypes about groups differ from those prevalent in North America. CONTENT WARNING: This paper contains stereotype examples that may be offensive.