English

Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventions

Computation and Language 2023-06-08 v1 Machine Learning

Abstract

Societal biases present in pre-trained large language models are a critical issue as these models have been shown to propagate biases in countless downstream applications, rendering them unfair towards specific groups of people. Since large-scale retraining of these models from scratch is both time and compute-expensive, a variety of approaches have been previously proposed that de-bias a pre-trained model. While the majority of current state-of-the-art debiasing methods focus on changes to the training regime, in this paper, we propose data intervention strategies as a powerful yet simple technique to reduce gender bias in pre-trained models. Specifically, we empirically show that by fine-tuning a pre-trained model on only 10 de-biased (intervened) training examples, the tendency to favor any gender is significantly reduced. Since our proposed method only needs a few training examples, our few-shot debiasing approach is highly feasible and practical. Through extensive experimentation, we show that our debiasing technique performs better than competitive state-of-the-art baselines with minimal loss in language modeling ability.

Keywords

Cite

@article{arxiv.2306.04597,
  title  = {Language Models Get a Gender Makeover: Mitigating Gender Bias with Few-Shot Data Interventions},
  author = {Himanshu Thakur and Atishay Jain and Praneetha Vaddamanu and Paul Pu Liang and Louis-Philippe Morency},
  journal= {arXiv preprint arXiv:2306.04597},
  year   = {2023}
}

Comments

Accepted to ACL 2023 Main Conference

R2 v1 2026-06-28T10:59:06.755Z