jina-reranker-v3: Last but Not Late Interaction for Listwise Document Reranking
This addresses the problem of efficient and effective multilingual document retrieval for search applications, representing a novel method rather than an incremental improvement.
The paper tackles document reranking by introducing a 'last but not late' interaction approach that applies causal attention between queries and all candidate documents in the same context window, achieving state-of-the-art BEIR performance with 61.94 nDCG@10 using a 0.6B-parameter model.
jina-reranker-v3 is a 0.6B-parameter multilingual listwise reranker that introduces a novel "last but not late" interaction. Unlike late interaction models like ColBERT that encode documents separately before multi-vector matching, our approach applies causal attention between the query and all candidate documents in the same context window, enabling rich interactions before extracting contextual embeddings from each document's final token. The new model achieves state-of-the-art BEIR performance with 61.94 nDCG@10 while being significantly smaller than other models with comparable performance.