BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation

TL;DR

This paper introduces BOLD, a large dataset and new metrics to systematically measure social biases in open-ended language generation models, revealing that current models exhibit significant biases compared to human text.

Contribution

The paper presents BOLD, a comprehensive dataset and automated metrics for benchmarking biases in open-ended language generation across multiple social domains.

Findings

Language models show larger biases than human text across domains.

New metrics effectively measure toxicity, psycholinguistic norms, and gender polarity.

Bias benchmarking highlights the need for bias mitigation in language models.

Abstract

Recent advances in deep learning techniques have enabled machines to generate cohesive open-ended text when prompted with a sequence of words as context. While these models now empower many downstream applications from conversation bots to automatic storytelling, they have been shown to generate texts that exhibit social biases. To systematically study and benchmark social biases in open-ended language generation, we introduce the Bias in Open-Ended Language Generation Dataset (BOLD), a large-scale dataset that consists of 23,679 English text generation prompts for bias benchmarking across five domains: profession, gender, race, religion, and political ideology. We also propose new automated metrics for toxicity, psycholinguistic norms, and text gender polarity to measure social biases in open-ended text generation from multiple angles. An examination of text generated from three popular language models reveals that the majority of these models exhibit a larger social bias than human-written Wikipedia text across all domains. With these results we highlight the need to benchmark biases in open-ended language generation and caution users of language generation models on downstream tasks to be cognizant of these embedded prejudices.