Artifacts for paper titled "CAT-LM: Training Language Models on Aligned Code And Tests"

Includes the model weights for CAT-LM, code and datasets for training and evaluating CAT-LM, results of additional experiments and comparison with TeCo and CodeGen.