This paper addresses a well-known issue in class-conditional generation in GANs and proposes a solution. The reviewers are satisfied with how the authors addresses their concerns in the rebuttal, and agree on accepting the paper. The reviewers agree that the paper addresses an interesting challenge present in a popular GAN architecture with auxiliary classifiers. A solution to this problem could potentially be widely used in GANs with auxiliary classifiers such as InfoGAN, ACGAN, etc.