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

Open in new window