GANDissect provides tools to inspect the internal representations of GAN models, identifying how individual neurons align with human-interpretable concepts. It enables interactive visualization through GANPaint and generates static dissection summaries. Built on PyTorch, the project supports analyzing different GAN layers, diagnosing artifacts, and performing interventions like removing objects from generated images.