Interpret and Control Dense Retrieval with Sparse Latent Features

• URL: http://arxiv.org/abs/2411.00786v2
• Date: Sun, 23 Feb 2025 22:56:57 GMT
• Title: Interpret and Control Dense Retrieval with Sparse Latent Features
• Abstract summary: This paper introduces a novel approach using sparse autoencoders (SAE) to interpret and control dense embeddings via the learned latent sparse features.<n> Experimental results demonstrate that both the learned latent sparse features and their reconstructed embeddings retain nearly the same retrieval accuracy as the original dense vectors.
• Score: 18.843817998618384
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Dense embeddings deliver strong retrieval performance but often lack interpretability and controllability. This paper introduces a novel approach using sparse autoencoders (SAE) to interpret and control dense embeddings via the learned latent sparse features. Our key contribution is the development of a retrieval-oriented contrastive loss, which ensures the sparse latent features remain effective for retrieval tasks and thus meaningful to interpret. Experimental results demonstrate that both the learned latent sparse features and their reconstructed embeddings retain nearly the same retrieval accuracy as the original dense vectors, affirming their faithfulness. Our further examination of the sparse latent space reveals interesting features underlying the dense embeddings and we can control the retrieval behaviors via manipulating the latent sparse features, for example, prioritizing documents from specific perspectives in the retrieval results.

Related papers

• Tracing Query Expansion Effects through Sparse Autoencoder Features [40.45659576999237]
  We trace QE effects through sparse autoencoder features.<n>We identify QE-related latents from expansion-induced activation shifts.<n>Our analysis shows that effective QE induces layer-concentrated changes in sparse latents aligned with retrieval intent and entity attributes.
  arXiv  Detail & Related papers  (2026-09-07T03:07:22Z)
• RepoMirage: Probing Repository Context Reasoning in Code Agents with Perturbations [51.43574078961796]
  We introduce RepoMirage, a two-stage evaluation suite built on SWE-Bench Verified.<n>RepoMirage-Perturb applies three types of semantics-preserving repository-level perturbations, revealing a clear performance drop when correct solving requires broader context access.<n>We propose RepoAnchor, a structure-first prototype workflow that separates repository exploration from downstream problem solving.
  arXiv  Detail & Related papers  (2026-05-25T06:26:43Z)
• Beyond Task-Driven Features for Object Detection [2.5700571547896103]
  This paper introduces an annotation-guided feature augmentation framework that injects embeddings into an object detection backbone.<n> Experiments across wildlife and remote sensing datasets evaluate classification, localization, and data efficiency under multiple supervision regimes.
  arXiv  Detail & Related papers  (2026-04-04T19:34:13Z)
• When Does Sparsity Mitigate the Curse of Depth in LLMs [53.137717161619484]
  We show that sparsity acts as a regulator of variance propagation and thereby improves depth utilization.<n>Our results reveal sparsity as a key yet previously overlooked mechanism for effective depth scaling in large language models.
  arXiv  Detail & Related papers  (2026-03-16T15:04:16Z)
• Stochastic Encodings for Active Feature Acquisition [100.47043816019888]
  Active Feature Acquisition is an instance-wise, sequential decision making problem.<n>The aim is to dynamically select which feature to measure based on current observations, independently for each test instance.<n>Common approaches either use Reinforcement Learning, which experiences training difficulties, or greedily maximize the conditional mutual information of the label and unobserved features, which makes myopic.<n>We introduce a latent variable model, trained in a supervised manner. Acquisitions are made by reasoning about the features across many possible unobserved realizations in a latent space.
  arXiv  Detail & Related papers  (2025-08-03T23:48:46Z)
• Dense SAE Latents Are Features, Not Bugs [75.08462524662072]
  We show that dense latents serve functional roles in language model computation.<n>We identify classes tied to position tracking, context binding, entropy regulation, letter-specific output signals, part-of-speech, and principal component reconstruction.
  arXiv  Detail & Related papers  (2025-06-18T17:59:35Z)
• "Principal Components" Enable A New Language of Images [79.45806370905775]
  We introduce a novel visual tokenization framework that embeds a provable PCA-like structure into the latent token space. Our approach achieves state-of-the-art reconstruction performance and enables better interpretability to align with the human vision system.
  arXiv  Detail & Related papers  (2025-03-11T17:59:41Z)
• The Complexity of Learning Sparse Superposed Features with Feedback [0.9838799448847586]
  We investigate whether the underlying learned features of a model can be efficiently retrieved through feedback from an agent. We analyze the feedback complexity associated with learning a feature matrix in sparse settings. Our results establish tight bounds when the agent is permitted to construct activations and demonstrate strong upper bounds in sparse scenarios.
  arXiv  Detail & Related papers  (2025-02-08T01:54:23Z)
• Disentangling Dense Embeddings with Sparse Autoencoders [0.0]
  Sparse autoencoders (SAEs) have shown promise in extracting interpretable features from complex neural networks. We present one of the first applications of SAEs to dense text embeddings from large language models. We show that the resulting sparse representations maintain semantic fidelity while offering interpretability.
  arXiv  Detail & Related papers  (2024-08-01T15:46:22Z)
• Transcending Forgery Specificity with Latent Space Augmentation for Generalizable Deepfake Detection [57.646582245834324]
  We propose a simple yet effective deepfake detector called LSDA. It is based on a idea: representations with a wider variety of forgeries should be able to learn a more generalizable decision boundary. We show that our proposed method is surprisingly effective and transcends state-of-the-art detectors across several widely used benchmarks.
  arXiv  Detail & Related papers  (2023-11-19T09:41:10Z)
• Discovering Object-Centric Generalized Value Functions From Pixels [17.10287710842919]
  We introduce a method that tries to discover meaningful features from objects, translating them to temporally coherent "question" functions. We also investigate the discovered general value functions and show that the learned representations are not only interpretable but also, centered around objects that are invariant to changes across tasks facilitating fast adaptation.
  arXiv  Detail & Related papers  (2023-04-27T00:34:24Z)
• ReDFeat: Recoupling Detection and Description for Multimodal Feature Learning [51.07496081296863]
  We recouple independent constraints of detection and description of multimodal feature learning with a mutual weighting strategy. We propose a detector that possesses a large receptive field and is equipped with learnable non-maximum suppression layers. We build a benchmark that contains cross visible, infrared, near-infrared and synthetic aperture radar image pairs for evaluating the performance of features in feature matching and image registration tasks.
  arXiv  Detail & Related papers  (2022-05-16T04:24:22Z)
• Robust Depth Completion with Uncertainty-Driven Loss Functions [60.9237639890582]
  We introduce uncertainty-driven loss functions to improve the robustness of depth completion and handle the uncertainty in depth completion. Our method has been tested on KITTI Depth Completion Benchmark and achieved the state-of-the-art robustness performance in terms of MAE, IMAE, and IRMSE metrics.
  arXiv  Detail & Related papers  (2021-12-15T05:22:34Z)
• Practical Benefits of Feature Feedback Under Distribution Shift [23.615375479758452]
  We show that feature feedback methods perform significantly better on various natural out-of-domain datasets. By contrast, on natural language inference tasks, performance remains comparable.
  arXiv  Detail & Related papers  (2021-10-14T17:35:23Z)
• Exploring Set Similarity for Dense Self-supervised Representation Learning [96.35286140203407]
  We propose to explore textbfset textbfsimilarity (SetSim) for dense self-supervised representation learning. We generalize pixel-wise similarity learning to set-wise one to improve the robustness because sets contain more semantic and structure information. Specifically, by resorting to attentional features of views, we establish corresponding sets, thus filtering out noisy backgrounds that may cause incorrect correspondences.
  arXiv  Detail & Related papers  (2021-07-19T09:38:27Z)
• Toward Understanding the Feature Learning Process of Self-supervised Contrastive Learning [43.504548777955854]
  We study how contrastive learning learns the feature representations for neural networks by analyzing its feature learning process. We prove that contrastive learning using textbfReLU networks provably learns the desired sparse features if proper augmentations are adopted.
  arXiv  Detail & Related papers  (2021-05-31T16:42:09Z)
• Reciprocal Feature Learning via Explicit and Implicit Tasks in Scene Text Recognition [60.36540008537054]
  In this work, we excavate the implicit task, character counting within the traditional text recognition, without additional labor annotation cost. We design a two-branch reciprocal feature learning framework in order to adequately utilize the features from both the tasks. Experiments on 7 benchmarks show the advantages of the proposed methods in both text recognition and the new-built character counting tasks.
  arXiv  Detail & Related papers  (2021-05-13T12:27:35Z)
• Representation Learning for Sequence Data with Deep Autoencoding Predictive Components [96.42805872177067]
  We propose a self-supervised representation learning method for sequence data, based on the intuition that useful representations of sequence data should exhibit a simple structure in the latent space. We encourage this latent structure by maximizing an estimate of predictive information of latent feature sequences, which is the mutual information between past and future windows at each time step. We demonstrate that our method recovers the latent space of noisy dynamical systems, extracts predictive features for forecasting tasks, and improves automatic speech recognition when used to pretrain the encoder on large amounts of unlabeled data.
  arXiv  Detail & Related papers  (2020-10-07T03:34:01Z)
• Learning Sparse Sentence Encoding without Supervision: An Exploration of Sparsity in Variational Autoencoders [36.60504726767521]
  sparsity is an effective inductive bias for learning efficient representation of data in vectors with fixed dimensionality. We investigate sparsity within the VAE framework which has been explored a lot in the image domain. We look at the implications of sparsity on text classification across 3 datasets.
  arXiv  Detail & Related papers  (2020-09-25T20:08:32Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.