690d83983a63aa1818423fd6edd3bfdb-AuthorFeedback.pdf

We thank the reviewers for their time and valuable feedback. We conduct new experiments showing that edge pruning is a necessary compo-26 nent of GNNGUARD. Further, edge pruning28 get even more important when we useGNNGUARD on graphs with heterophily because pruning of adversarial29 edges has a direct effect on the choice of structural similarity between nodes (e.g., graphlet degree similarity).30 We note that if an edge is pruned in one layer, it will32 get pruned inallsubsequent layers. R1 raises an important point on examining defense performance with varying47 budgets.