DR.GAP: Mitigating Bias in Large Language Models using Gender-Aware Prompting with Demonstration and Reasoning

Large Language Models (LLMs) exhibit strong natural language processing capabilities but also inherit and amplify societal biases, including gender bias, raising fairness concerns. Existing debiasing methods face significant limitations: parameter tuning requires access to model weights, prompt-based approaches often degrade model utility, and optimization-based techniques lack generalizability. To address these challenges, we proposethis http URL(Demonstration and Reasoning for Gender-Aware Prompting), an automated and model-agnostic approach that mitigates gender bias while preserving model performance.this http URLselects bias-revealing examples and generates structured reasoning to guide models toward more impartial responses. Extensive experiments on coreference resolution and QA tasks across multiple LLMs (GPT-3.5, Llama3, and Llama2-Alpaca) demonstrate its effectiveness, generalization ability, and robustness.this http URLcan generalize to vision-language models (VLMs), achieving significant bias reduction.

@article{qiu2025_2502.11603,
  title={ DR.GAP: Mitigating Bias in Large Language Models using Gender-Aware Prompting with Demonstration and Reasoning },
  author={ Hongye Qiu and Yue Xu and Meikang Qiu and Wenjie Wang },
  journal={arXiv preprint arXiv:2502.11603},
  year={ 2025 }
}