Decoding Dense Embeddings: Sparse Autoencoders for Interpreting and Discretizing Dense Retrieval

ACL ARR 2025 May Submission715 Authors

Abstract: Despite their strong performance, Dense Passage Retrieval (DPR) models suffer from a lack of interpretability. In this work, we propose a novel interpretability framework that leverages Sparse Autoencoders (SAEs) to decompose previously uninterpretable dense embeddings from DPR models into distinct, interpretable latent concepts. We generate natural language descriptions for each latent concept, enabling human interpretations of both the dense embeddings and the query-document similarity scores of DPR models. We further introduce Concept-Level Sparse Retrieval (CL-SR), a retrieval framework that directly utilizes the extracted latent concepts as indexing units. CL-SR effectively combines the semantic expressiveness of dense embeddings with the transparency and efficiency of sparse representations. We show that CL-SR achieves high computational and storage efficiency while maintaining robust performance across vocabulary and semantic mismatches.

Paper Type: Long

Research Area: Interpretability and Analysis of Models for NLP

Research Area Keywords: free-text/natural language explanations, passage retrieval, dense retrieval, document representation

Contribution Types: Model analysis & interpretability, NLP engineering experiment, Publicly available software and/or pre-trained models

Languages Studied: English

Submission Number: 715