1 paper
Caio F. Deberaldini Netto, Moshe Eliasof, Luana Ruiz
Graph Transformers produce, for each attention head, a dense n×n matrix of learned pairwise interactions. We ask a fundamental question: do these attention-induced graphs c…