FairFace is a face attribute dataset designed to mitigate biases in AI models by providing balanced representation across race, gender, and age. It aims to improve the fairness and robustness of computer vision systems, ensuring more equitable performance.
This study investigates where stereotypes reside in LLMs like GPT 2 Small and Llama 3.2. It explores identifying individual neuron activations and attention heads to map "bias fingerprints" and provide initial insights for mitigation.
This research investigates how bias mitigation reshapes the embedding space of encoder-only and decoder-only foundation models like BERT and Llama2. Findings show that bias mitigation reduces gender-occupation disparities in the embedding space, leading to more neutral internal representations, confirming embedding analysis as a valuable debiasing validation tool.
This paper proposes a lightweight, inference-time framework to mitigate demographic representational bias in text-to-image models like Stable Diffusion, without requiring model retraining. The approach allows users to select their own fairness specifications to generate more equitable outputs across professions.