English
Related papers

Related papers: Evolution of SAE Features Across Layers in LLMs

200 papers

Transformer models have become state-of-the-art in decoding stimuli and behavior from neural activity, significantly advancing neuroscience research. Yet greater transparency in their decision-making processes would substantially enhance…

Quantitative Methods · Quantitative Biology 2025-06-18 Laurence Freeman , Philip Shamash , Vinam Arora , Caswell Barry , Tiago Branco , Eva Dyer

Vision-language models (VLMs) have advanced rapidly and are increasingly deployed in real-world applications, especially with the rise of agent-based systems. However, their safety has received relatively limited attention. Even the latest…

Computer Vision and Pattern Recognition · Computer Science 2026-05-11 Hao Wang , Yiqun Sun , Pengfei Wei , Lawrence B. Hsieh , Daisuke Kawahara

Modern LLMs face inference efficiency challenges due to their scale. To address this, many compression methods have been proposed, such as pruning and quantization. However, the effect of compression on a model's interpretability remains…

Machine Learning · Computer Science 2025-07-23 Suchit Gupte , Vishnu Kabir Chhabra , Mohammad Mahdi Khalili

Recovering meaningful concepts from language model activations is a central aim of interpretability. While existing feature extraction methods aim to identify concepts that are independent directions, it is unclear if this assumption can…

Recent work on sparse autoencoders (SAEs) has shown promise in extracting interpretable features from neural networks and addressing challenges with polysemantic neurons caused by superposition. In this paper, we apply SAEs to the early…

Machine Learning · Computer Science 2024-09-10 Liv Gorton

Sparse autoencoders (SAEs) have recently emerged as a powerful tool for interpreting the features learned by large language models (LLMs). By reconstructing features with sparsely activated networks, SAEs aim to recover complex superposed…

Machine Learning · Computer Science 2026-03-05 Jingyi Cui , Qi Zhang , Yifei Wang , Yisen Wang

Modern text classification methods heavily rely on contextual embeddings from large language models (LLMs). Compared to human-engineered features, these embeddings provide automatic and effective representations for classification model…

Computation and Language · Computer Science 2025-07-29 Xuansheng Wu , Wenhao Yu , Xiaoming Zhai , Ninghao Liu

The success of deep networks is crucially attributed to their ability to capture latent features within a representation space. In this work, we investigate whether the underlying learned features of a model can be efficiently retrieved…

Machine Learning · Computer Science 2025-12-30 Akash Kumar

Sparse autoencoders (SAEs) are increasingly used for safety-relevant applications including alignment detection and model steering. These use cases require SAE latents to be as atomic as possible. Each latent should represent a single…

Machine Learning · Computer Science 2026-04-07 Matthew Levinson

Recurrent models for sequences have been recently successful at many tasks, especially for language modeling and machine translation. Nevertheless, it remains challenging to extract good representations from these models. For instance, even…

Machine Learning · Computer Science 2018-01-31 Łukasz Kaiser , Samy Bengio

Analyzing large-scale text corpora is a core challenge in machine learning, crucial for tasks like identifying undesirable model behaviors or biases in training data. Current methods often rely on costly LLM-based techniques (e.g.…

Artificial Intelligence · Computer Science 2025-12-12 Nick Jiang , Xiaoqing Sun , Lisa Dunlap , Lewis Smith , Neel Nanda

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

We study how reliably sparse autoencoders (SAEs) support claims about reasoning-related internal features in large language models. We first give a stylized analysis showing that sparsity-regularized decoding can preferentially retain…

Machine Learning · Computer Science 2026-05-19 George Ma , Zhongyuan Liang , Irene Y. Chen , Somayeh Sojoudi

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 are usually trained one layer at a time, even though transformer residual stream activations are strongly coupled across depth. This creates a practical problem for multi-layer interventions: different layerwise…

Machine Learning · Computer Science 2026-05-28 Prathyush Poduval , Calvin Yeung , Neel Desai , Mohsen Imani

Sparse autoencoders (SAEs) decompose large language model (LLM) activations into latent features that reveal mechanistic structure. Conventional SAEs train on broad data distributions, forcing a fixed latent budget to capture only…

Machine Learning · Computer Science 2025-08-14 Charles O'Neill , Mudith Jayasekara , Max Kirkby

We investigate feature universality in Gemma-2 language models (Gemma-2-2B and Gemma-2-9B), asking whether models with a four-fold difference in scale still converge on comparable internal concepts. Using the Sparse Autoencoder (SAE)…

Computation and Language · Computer Science 2025-08-01 Daniel Son , Sanjana Rathore , Andrew Rufail , Adrian Simon , Daniel Zhang , Soham Dave , Cole Blondin , Kevin Zhu , Sean O'Brien

Stack autoencoder (SAE), as a representative deep network, has unique and excellent performance in feature learning, and has received extensive attention from researchers. However, existing deep SAEs focus on original samples without…

Machine Learning · Computer Science 2022-10-28 Chuanyan Zhou , Jie Ma , Fan Li , Yongming Li , Pin Wang , Xiaoheng Zhang

We identify semantically coherent, context-consistent network components in large language models (LLMs) using coactivation of sparse autoencoder (SAE) features collected from just a handful of prompts. Focusing on concept-relation…

Computation and Language · Computer Science 2026-04-21 Ruixuan Deng , Xiaoyang Hu , Miles Gilberti , Shane Storks , Aman Taxali , Mike Angstadt , Chandra Sripada , Joyce Chai

Sparse autoencoders (SAEs) are a promising approach to extracting features from neural networks, enabling model interpretability as well as causal interventions on model internals. SAEs generate sparse feature representations using a…

Machine Learning · Computer Science 2024-11-11 Kola Ayonrinde