相关论文: Kronecker Factorization Improves Efficiency and In…

Sparse autoencoders (SAEs) offer a natural path toward comparable explanations across different representation spaces. However, current SAEs are trained per modality, producing dictionaries whose features are not directly understandable and…

Protein language models have revolutionized structure prediction, but their nonlinear nature obscures how sequence representations inform structure prediction. While sparse autoencoders (SAEs) offer a path to interpretability here by…

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…

Sparse Autoencoders (SAEs) are an interpretability technique aimed at decomposing neural network activations into interpretable units. However, a major bottleneck for SAE development has been the lack of high-quality performance metrics,…

Developing human understandable interpretation of large language models (LLMs) becomes increasingly critical for their deployment in essential domains. Mechanistic interpretability seeks to mitigate the issues through extracts…

Sparse autoencoders (SAEs) have been used widely to decompose and interpret neural network activations, especially those of transformer language models. One key issue with SAEs is their inability to directly model multidimensional features.…

The Linear Representation Hypothesis asserts that the embeddings learned by neural networks can be understood as linear combinations of features corresponding to high-level concepts. Based on this ansatz, sparse autoencoders (SAEs) have…

Scientific archives now contain hundreds of petabytes of data across genomics, ecology, climate, and molecular biology that could reveal undiscovered patterns if systematically analyzed at scale. Large-scale, weakly-supervised datasets in…

Sparse Autoencoders (SAEs) can extract interpretable features from large language models (LLMs) without supervision. However, their effectiveness in downstream steering tasks is limited by the requirement for contrastive datasets or large…

Sparse Autoencoders (SAEs) are widely used to interpret neural networks by identifying meaningful concepts from their representations. However, do SAEs truly uncover all concepts a model relies on, or are they inherently biased toward…

Autoencoders have been used for finding interpretable and disentangled features underlying neural network representations in both image and text domains. While the efficacy and pitfalls of such methods are well-studied in vision, there is a…

Sparse Autoencoders (SAEs) have emerged as a powerful unsupervised method for extracting sparse representations from language models, yet scalable training remains a significant challenge. We introduce a suite of 256 SAEs, trained on each…

Large Reasoning Models (LRMs) exhibit human-like cognitive reasoning strategies (e.g. backtracking, cross-verification) during reasoning process, which improves their performance on complex tasks. Currently, reasoning strategies are…

This paper studies the problem of Kronecker-structured sparse vector recovery from an underdetermined linear system with a Kronecker-structured dictionary. Such a problem arises in many real-world applications such as the sparse channel…

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…

Pathology plays an important role in disease diagnosis, treatment decision-making and drug development. Previous works on interpretability for machine learning models on pathology images have revolved around methods such as attention value…

Ultra-high-dimensional tensor predictors are increasingly common in neuroimaging and other biomedical studies, yet existing methods rarely integrate continuous, count, and binary responses in a single coherent model. We present a Bayesian…

RFdiffusion is a popular and well-established model for generation of protein structures. However, this generative process offers limited insight into its internal representations and how they contribute to the final protein structure.…

Adapting foundation models for specific purposes has become a standard approach to build machine learning systems for downstream applications. Yet, it is an open question which mechanisms take place during adaptation. Here we develop a new…

Dictionary learning and component analysis models are fundamental for learning compact representations that are relevant to a given task (feature extraction, dimensionality reduction, denoising, etc.). The model complexity is encoded by…