Inducing Group Fairness in Prompt-Based Language Model Decisions

TL;DR

This paper explores methods to induce group fairness in prompt-based language model classifiers, adapting classical fairness techniques and developing new prompt-level approaches, with empirical evaluation on real datasets.

Contribution

It introduces adaptations of classical fairness methods to prompt-based models and proposes new prompt-level techniques to improve group fairness.

Findings

Classical fairness approaches are effective in LM-based classifiers.

Prompt-level methods show potential but need further development.

Room remains for improved fairness techniques leveraging LM structure.

Abstract

Classifiers are used throughout industry to enforce policies, ranging from the detection of toxic content to age-appropriate content filtering. While these classifiers serve important functions, it is also essential that they are built in ways that minimize unfair biases for users. One such fairness consideration is called group fairness, which desires that different sub-population of users receive equal treatment. This is a well-studied problem in the context of 'classical' classifiers. However, the emergence of prompt-based language model (LM) decision making has created new opportunities to solve text-based classification tasks, and the fairness properties of these new classifiers are not yet well understood. Further, the `remediation toolkit' is incomplete for LM-based decision makers and little is understood about how to improve decision maker group fairness while maintaining classifier performance. This work sets out to add more tools to that toolbox. We introduce adaptations of existing effective approaches from the classical classifier fairness to the prompt-based classifier space. We also devise simple methods that take advantage of the new structure of prompt-based decision makers and operate at the prompt level. We compare these approaches empirically on real data. Our results suggest that adaptations of approaches that are effective for classical classifiers remain effective in the LM-based classifier environment. However, there is room for further exploration of prompt-based remediation methods (and other remediation methods that take advantage of LM structure).