RECURRENT READING FOR SINGLE-VECTOR RETRIEVAL

In reasoning-intensive retrieval, a relevant document may share an underlying principle with the query rather than its words, requiring several points of contact between query and document. Late-interaction models provide this capacity but incur additional cost at retrieval time. We ask whether a single vector representation can instead be trained to encode this information. In our aligned-score configurations with untied output projections, we observe slot locking: corresponding query and document reads attain the highest similarity, and the additional reads partly compensate for a weakened base embedding. We introduce Recurrent Reading for Single-Vector Retrieval (RLI), which reads token states with a recurrent head and matches each query read to its best document read through a projection shared across reads. We find that late interaction is more useful as a training signal than as a scoring rule. At retrieval time, RLI ranks documents using one inner product between the packed query and document vectors. Across four backbones from 0.6B to 4B and three model families, removing the reads from the trained model costs $3.9$ to $5.2$ nDCG@10. On BRIGHT, using original queries without query-time generation or reranking, a 4B RLI averages $42.42\pm 0.88$ nDCG@10 over three seeds (best $43.41$), $4.1$ points above the strongest published retriever-only baseline in our comparison and $3.6$ above the same recipe without reads ($38.80\pm 1.06$).