The artifact for evaluating CIT4DNN, a method for generating diverse and rare inputs for testing deep neural networks, is publicly available on Github and Software Heritage and it is reusable. The artifact includes Python scripts, C++ code, pretrained deep neural network and variational autoencoder models, and configuration files for running the experiments from the paper. The artifact also includes a Shell script for running different modules of CIT4DNN to replicate the research results. Users should be familiar with running Python scripts in Linux environment.