Do Neural Retrievers Prefer Certain Documents? Evidence of Learned Relevance Priors

ArXi:2606.02814v1 Announce Type: cross Neural retrievers are trained to estimate query-document relevance from annotated query-document pairs. Yet annotation protocols may not purely reflect relevance: they select only a subset of documents for labeling, and this selection can favor certain document types over others. We investigate whether supervised bi-encoder retrievers implicitly on annotated data. We estimate this prior by