Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
FairVisgithub.comFairVis is a visual analytics system designed to audit machine learning classifiers for intersectional bias. It lets users generate subgroups from their data and explore whether a model underperforms for particular populations by comparing distributions and performance metrics across groups. The project showcases live demos and a research paper (illustrating biases such as false positive rate disparities on datasets like COMPAS) to demonstrate how bias can emerge beyond base rates.
Parse Score
Sources
arxiv.org shapes more of what AI says about FairVis than any other source, at 50% of its citations.
pmc.ncbi.nlm.nih.gov
