A unified benchmark dataset for AND on scholarly papers, as well as an open-source reference model implementation. Our dataset harmonizes eight disparate AND datasets into a uniform format, with a single rich feature set drawn from the Semantic Scholar (S2) database.