Bias Association Discovery Framework for Open-Ended LLM Generations

TL;DR

The paper introduces BADF, a framework for discovering both known and novel social biases in open-ended LLM outputs by systematically extracting associations between identities and concepts, enhancing bias detection.

Contribution

It presents a novel systematic method, BADF, for uncovering unrecognized bias associations in LLMs beyond predefined identity-concept pairs.

Findings

BADF successfully identifies known biases in multiple models.

The framework uncovers previously unrecognized bias associations.

Experiments demonstrate BADF's scalability across diverse contexts.

Abstract

Social biases embedded in Large Language Models (LLMs) raise critical concerns, resulting in representational harms -- unfair or distorted portrayals of demographic groups -- that may be expressed in subtle ways through generated language. Existing evaluation methods often depend on predefined identity-concept associations, limiting their ability to surface new or unexpected forms of bias. In this work, we present the Bias Association Discovery Framework (BADF), a systematic approach for extracting both known and previously unrecognized associations between demographic identities and descriptive concepts from open-ended LLM outputs. Through comprehensive experiments spanning multiple models and diverse real-world contexts, BADF enables robust mapping and analysis of the varied concepts that characterize demographic identities. Our findings advance the understanding of biases in open-ended generation and provide a scalable tool for identifying and analyzing bias associations in LLMs.