Abstract: Graph Transformers (GTs) have shown great potential in graph representation learning over the past few years. In spite of their promising performance on small graphs, how to develop a ...