Related papers: SPARC: Concept-Aligned Sparse Autoencoders for Cro…
In current multimodal tasks, models typically freeze the encoder and decoder while adapting intermediate layers to task-specific goals, such as region captioning. Region-level visual understanding presents significant challenges for…
Despite the central role of action in embodied intelligence, learning transferable action representations from visual transitions remains a fundamental challenge, particularly when world models must generalize across embodiments under…
Multimodal learning aims to capture both shared and private information from multiple modalities. However, existing methods that project all modalities into a single latent space for fusion often overlook the asynchronous, multi-level…
We consider the design and analysis of spatially coupled sparse regression codes (SC-SPARCs), which were recently introduced by Barbier et al. for efficient communication over the additive white Gaussian noise channel. SC-SPARCs can be…
Self-supervised learning (SSL) faces a fundamental conflict between semantic understanding and image reconstruction. High-level semantic SSL (e.g., DINO) relies on global tokens that are forced to be location-invariant for augmentation…
Sparse coding is an unsupervised learning algorithm that learns a succinct high-level representation of the inputs given only unlabeled data; it represents each input as a sparse linear combination of a set of basis functions. Originally…
Prototypical part learning is emerging as a promising approach for making semantic segmentation interpretable. The model selects real patches seen during training as prototypes and constructs the dense prediction map based on the similarity…
Accurate fault detection in high-dimensional industrial environments remains a major challenge due to the inherent complexity, noise, and redundancy in sensor data. This paper introduces CLAIRE, i.e., a hybrid end-to-end learning framework…
High-dimensional data sets are often analyzed and explored via the construction of a latent low-dimensional space which enables convenient visualization and efficient predictive modeling or clustering. For complex data structures, linear…
Sparse autoencoders (SAEs) trained on large language model activations output thousands of features that enable mapping to human-interpretable concepts. The current practice for analyzing these features primarily relies on inspecting…
Interpretability is crucial for building safe, reliable, and controllable language models, yet existing interpretability pipelines remain costly and difficult to scale. Interpreting a new model typically requires training model-specific…
In recent years, a large amount of multi-disciplinary research has been conducted on sparse models and their applications. In statistics and machine learning, the sparsity principle is used to perform model selection---that is,…
For large language models (LLMs), sparse autoencoders (SAEs) have been shown to decompose intermediate representations that often are not interpretable directly into sparse sums of interpretable features, facilitating better control and…
Context compression is an advanced technique that accelerates large language model (LLM) inference by converting long inputs into compact representations. Existing methods primarily rely on autoencoding tasks to train special compression…
Sparse autoencoders (SAEs) have received considerable recent attention as tools for mechanistic interpretability, showing success at extracting interpretable features even from very large LLMs. However, this research has been largely…
Feature extraction from infrared (IR) images remains a challenging task. Learning based methods that can work on raw imagery/patches have therefore assumed significance. We propose a novel multi-task extension of the widely used…
Sparse Autoencoder (SAE) has emerged as a powerful tool for mechanistic interpretability of large language models. Recent works apply SAE to protein language models (PLMs), aiming to extract and analyze biologically meaningful features from…
Sparse autoencoders (SAEs) enable interpretability research by decomposing entangled model activations into monosemantic features. However, under what circumstances SAEs derive most fine-grained latent features for safety, a low-frequency…
A fundamental challenge in embodied AI is verifying if agents build internal models of spatial structure or merely learn to mimic task-specific expert trajectories. This is critical as foundational approaches rooted in action-centric tasks…
Sparse autoencoders (SAEs) are used to decompose neural network activations into sparsely activating features, but many SAE features are only interpretable at high activation strengths. To address this issue we propose to use binary sparse…