graph neural networks (GNNs) have become increasingly ubiquitous in numerous
applications and systems, necessitating explanations of their predictions,
especially when making critical decisions. However, explaining GNNs is
challenging due to the complexity of graph data and model execu