小模型实现顶尖文档重排序，关键在查询与文档的深度交互。
jina-reranker-v3: Last but Not Late Interaction for Listwise Document Reranking
- 查询与文档共用上下文窗口，通过因果注意力实现双向互动。
- 在BEIR基准上达61.94 nDCG@10，参数仅0.6B，远小于同类模型。
- 适合追求高精度与低资源部署的多语言检索场景。
jina-reranker-v3 是一个 0.6B 参数的多语言列表级重排序模型，提出一种新颖的「最后但不迟」交互机制。不同于 ColBERT 等延迟交互模型在独立编码文档后再进行多向量匹配，本方法将查询与所有候选文档置于同一上下文窗口，应用因果注意力实现丰富交互，再从各文档的最终标记中提取上下文嵌入。该模型在 BEIR 基准上取得 61.94 nDCG@10 的最佳性能，同时显著小于其他具备相似表现的模型。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。