English
Related papers

Related papers: SoftSAE: Dynamic Top-K Selection for Adaptive Spar…

200 papers

Sparse Autoencoders (SAEs) are a prominent tool in mechanistic interpretability (MI) for decomposing neural network activations into interpretable features. However, the aspiration to identify a canonical set of features is challenged by…

Machine Learning · Computer Science 2025-05-27 Xiangchen Song , Aashiq Muhamed , Yujia Zheng , Lingjing Kong , Zeyu Tang , Mona T. Diab , Virginia Smith , Kun Zhang

Sparse autoencoders (SAEs) are a promising unsupervised approach for identifying causally relevant and interpretable linear features in a language model's (LM) activations. To be useful for downstream tasks, SAEs need to decompose LM…

Machine Learning · Computer Science 2024-08-02 Senthooran Rajamanoharan , Tom Lieberum , Nicolas Sonnerat , Arthur Conmy , Vikrant Varma , János Kramár , Neel Nanda

Sparse autoencoders (SAEs) are a promising technique for decomposing language model activations into interpretable linear features. However, current SAEs fall short of completely explaining model performance, resulting in "dark matter":…

Machine Learning · Computer Science 2025-03-26 Joshua Engels , Logan Riggs , Max Tegmark

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…

Machine Learning · Computer Science 2026-04-15 Jiaqi Weng , Han Zheng , Hanyu Zhang , Ej Zhou , Qinqin He , Jialing Tao , Hui Xue , Zhixuan Chu , Xiting Wang

Sparse autoencoders (SAEs) have become a standard tool for mechanistic interpretability in autoregressive large language models (LLMs), enabling researchers to extract sparse, human-interpretable features and intervene on model behavior.…

Machine Learning · Computer Science 2026-02-06 Xu Wang , Bingqing Jiang , Yu Wan , Baosong Yang , Lingpeng Kong , Difan Zou

Sparse autoencoders (SAEs) have been successfully used to discover sparse and human-interpretable representations of the latent activations of LLMs. However, we would ultimately like to understand the computations performed by LLMs and not…

Machine Learning · Computer Science 2025-06-09 Lucy Farnik , Tim Lawson , Conor Houghton , Laurence Aitchison

We present the first systematic study of Sparse Autoencoders (SAEs) on video representations. Standard SAEs decompose video into interpretable, monosemantic features but destroy temporal coherence: hard TopK selection produces unstable…

Computer Vision and Pattern Recognition · Computer Science 2026-04-07 Atahan Dokme , Sriram Vishwanath

A new line of research for feature selection based on neural networks has recently emerged. Despite its superiority to classical methods, it requires many training iterations to converge and detect informative features. The computational…

Machine Learning · Computer Science 2022-11-29 Ghada Sokar , Zahra Atashgahi , Mykola Pechenizkiy , Decebal Constantin Mocanu

Sparse autoencoders (SAEs) extract millions of interpretable features from a language model, but flat feature inventories aren't very useful on their own. Domain concepts get mixed with generic and weakly grounded features, while related…

Artificial Intelligence · Computer Science 2026-04-29 John Winnicki , Abeynaya Gnanasekaran , Eric Darve

Sparse Autoencoders uncover thousands of features in vision models, yet explaining these features without requiring human intervention remains an open challenge. While previous work has proposed generating correlation-based explanations…

Computer Vision and Pattern Recognition · Computer Science 2026-03-26 Javier Ferrando , Enrique Lopez-Cuena , Pablo Agustin Martin-Torres , Daniel Hinjos , Anna Arias-Duart , Dario Garcia-Gasulla

Sparse Autoencoders (SAEs) can efficiently identify candidate monosemantic features from pretrained neural networks for galaxy morphology. We demonstrate this on Euclid Q1 images using both supervised (Zoobot) and new self-supervised (MAE)…

Instrumentation and Methods for Astrophysics · Physics 2025-11-13 John F. Wu , Michael Walmsley

Sparse autoencoders (SAEs) are an unsupervised method for learning a sparse decomposition of a neural network's latent representations into seemingly interpretable features. Despite recent excitement about their potential, research…

Sparse Autoencoders (SAEs) have shown promise in improving the interpretability of neural network activations, but can learn features that are not features of the input, limiting their effectiveness. We propose \textsc{Mutual Feature…

Machine Learning · Computer Science 2024-11-07 Luke Marks , Alasdair Paren , David Krueger , Fazl Barez

It is assumed that sparse autoencoders (SAEs) decompose polysemantic activations into interpretable linear directions, as long as the activations are composed of sparse linear combinations of underlying features. However, we find that if an…

Machine Learning · Computer Science 2025-09-29 David Chanin , Tomáš Dulka , Adrià Garriga-Alonso

Sparse autoencoders (SAEs) are a core interpretability tool for large language models, and progress on SAE architectures depends on benchmarks that reliably distinguish better SAEs from worse ones. We audit the SAE quality metrics in…

Machine Learning · Computer Science 2026-05-19 David Chanin

Sparse autoencoders (SAEs) have emerged as a promising approach for learning interpretable features from neural network activations. However, the optimization landscape for SAE training can be challenging due to correlations in the input…

Machine Learning · Computer Science 2025-11-19 Ashwin Saraswatula , David Klindt

A key challenge in interpretability is to decompose model activations into meaningful features. Sparse autoencoders (SAEs) have emerged as a promising tool for this task. However, a central problem in evaluating the quality of SAEs is the…

Machine Learning · Computer Science 2024-10-11 Constantin Venhoff , Anisoara Calinescu , Philip Torr , Christian Schroeder de Witt

Large-scale pre-trained vision-language models like CLIP demonstrate remarkable zero-shot performance across diverse tasks. However, fine-tuning these models to improve downstream performance often degrades robustness against distribution…

Computer Vision and Pattern Recognition · Computer Science 2026-05-18 Fabian Morelli , Arnas Uselis , Ankit Sonthalia , Seong Joon Oh

Sparse autoencoders (SAEs) improve interpretability in multimodal models, but it remains unclear whether SAE features form modular, composable units for reasoning-an assumption underlying many intervention-based steering methods. We test…

Artificial Intelligence · Computer Science 2026-03-27 Yunpeng Zhou

A common goal of mechanistic interpretability is to decompose the activations of neural networks into features: interpretable properties of the input computed by the model. Sparse autoencoders (SAEs) are a popular method for finding these…

Machine Learning · Computer Science 2025-02-10 Patrick Leask , Bart Bussmann , Michael Pearce , Joseph Bloom , Curt Tigges , Noura Al Moubayed , Lee Sharkey , Neel Nanda