Bias Unveiled: Investigating Social Bias in LLM-Generated Code

TL;DR

This paper introduces Solar, a framework to evaluate and reduce social biases in code generated by large language models, revealing significant biases and demonstrating effective mitigation strategies.

Contribution

The paper presents Solar, a novel fairness framework with a new bias dataset, and explores prompting strategies to significantly reduce social bias in LLM-generated code.

Findings

Severe social bias found in all tested LLMs' generated code.

Dialogue with Solar reduces social bias by up to 90%.

The framework and dataset are publicly available and extensible.

Abstract

Large language models (LLMs) have significantly advanced the field of automated code generation. However, a notable research gap exists in evaluating social biases that may be present in the code produced by LLMs. To solve this issue, we propose a novel fairness framework, i.e., Solar, to assess and mitigate the social biases of LLM-generated code. Specifically, Solar can automatically generate test cases for quantitatively uncovering social biases of the auto-generated code by LLMs. To quantify the severity of social biases in generated code, we develop a dataset that covers a diverse set of social problems. We applied Solar and the crafted dataset to four state-of-the-art LLMs for code generation. Our evaluation reveals severe bias in the LLM-generated code from all the subject LLMs. Furthermore, we explore several prompting strategies for mitigating bias, including Chain-of-Thought (CoT) prompting, combining positive role-playing with CoT prompting and dialogue with Solar. Our experiments show that dialogue with Solar can effectively reduce social bias in LLM-generated code by up to 90%. Last, we make the code and data publicly available is highly extensible to evaluate new social problems.