LANGSAE 编辑：通过事后语言身份移除改进多语言信息检索
LANGSAE EDITING: Improving Multilingual Information Retrieval via Post-hoc Language Identity Removal
AI总结：
LANGSAE 编辑通过事后移除语言身份信号，提升多语言信息检索的排名质量和跨语言覆盖性。
摘要：
在多语言环境下，密集检索常常在混合语言的集合中进行搜索，但多语言嵌入编码了语言身份与语义。这种语言信号可能放大同语言对的相似度，并挤压其他语言中相关证据。我们提出了LANGSAE 编辑，这是一种基于池化嵌入训练的后处理稀疏自编码器，能够在向量空间中直接可控地移除语言身份信号。该方法利用跨语言激活统计识别语言相关的潜在单元，在推理时抑制这些单元，并在原始维度重建嵌入，使其能够与现有向量数据库兼容，而无需重新训练基础编码器或重新编码原始文本。在多种语言上的实验显示，排名质量和跨语言覆盖性均有持续改进，尤其是对于书写系统不同的语言表现尤为突出。
英文摘要：
Dense retrieval in multilingual settings often searches over mixed-language collections, yet multilingual embeddings encode language identity alongside semantics. This language signal can inflate similarity for same-language pairs and crowd out relevant evidence written in other languages. We propose LANGSAE EDITING, a post-hoc sparse autoencoder trained on pooled embeddings that enables controllable removal of language-identity signal directly in vector space. The method identifies language-associated latent units using cross-language activation statistics, suppresses these units at inference time, and reconstructs embeddings in the original dimensionality, making it compatible with existing vector databases without retraining the base encoder or re-encoding raw text. Experiments across multiple languages show consistent improvements in ranking quality and cross-language coverage, with especially strong gains for script-distinct languages.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
