English
Related papers

Related papers: Binary Autoencoder for Mechanistic Interpretabilit…

200 papers

Human face exhibits an inherent hierarchy in its representations (i.e., holistic facial expressions can be encoded via a set of facial action units (AUs) and their intensity). Variational (deep) auto-encoders (VAE) have shown great results…

Computer Vision and Pattern Recognition · Computer Science 2017-08-08 Dieu Linh Tran , Robert Walecki , Ognjen Rudovic , Stefanos Eleftheriadis , Bjørn Schuller , Maja Pantic

While sparse autoencoders (SAEs) successfully extract interpretable features from language models, applying them to audio generation faces unique challenges: audio's dense nature requires compression that obscures semantic meaning, and…

Machine Learning · Computer Science 2025-10-31 Nathan Paek , Yongyi Zang , Qihui Yang , Randal Leistikow

Semantic hashing has become a crucial component of fast similarity search in many large-scale information retrieval systems, in particular, for text data. Variational auto-encoders (VAEs) with binary latent variables as hashing codes…

Machine Learning · Computer Science 2020-05-22 Siamak Zamani Dadaneh , Shahin Boluki , Mingzhang Yin , Mingyuan Zhou , Xiaoning Qian

Mixture-of-Experts (MoE) architectures enhance the efficiency of large language models by activating only a subset of experts per token. However, standard MoE employs a fixed Top-K routing strategy, leading to redundant computation and…

Artificial Intelligence · Computer Science 2026-05-15 Juntong Wu , Jialiang Cheng , Qishen Yin , Yue Dai , Yuliang Yan , Fuyu Lv , Ou Dan , Li Yuan

Intermediate layers of large language models (LLMs) best predict human brain responses to language, one of the most robust findings in computational neurolinguistics, yet why remains mechanistically unexplained. We address this gap by…

Computation and Language · Computer Science 2026-05-25 Dongxin Guo , Jikun Wu , Siu Ming Yiu

Masked auto-encoder pre-training has emerged as a prevalent technique for initializing and enhancing dense retrieval systems. It generally utilizes additional Transformer decoder blocks to provide sustainable supervision signals and…

Information Retrieval · Computer Science 2024-04-23 Guangyuan Ma , Xing Wu , Zijia Lin , Songlin Hu

Sparse autoencoders (SAEs) are one of the main methods to interpret the inner workings of deep neural networks (DNNs), decomposing activations into higher-dimensional features. However, they exhibit critical shortcomings where a large…

Machine Learning · Computer Science 2026-05-19 Michał Brzozowski , Neo Christopher Chung

In this thesis, we develop methods to enhance the interpretability of recent representation learning techniques in natural language processing (NLP) while accounting for the unavailability of annotated data. We choose to leverage…

Computation and Language · Computer Science 2023-05-05 Ghazi Felhi

Recently, sparse autoencoders (SAEs) have emerged as a promising technique for interpreting activations in foundation models by disentangling features into a sparse set of concepts. However, identifying the optimal level of sparsity for…

Machine Learning · Computer Science 2026-04-17 Dongsheng Wang , Jinsen Zhang , Dawei Su , Hui Huang

Recent work leverages Vision Foundation Models as image encoders to boost the generative performance of latent diffusion models (LDMs), as their semantic feature distributions are easy to learn. However, such semantic features often lack…

Computer Vision and Pattern Recognition · Computer Science 2026-02-10 Siyu Liu , Chujie Qin , Hubery Yin , Qixin Yan , Zheng-Peng Duan , Chen Li , Jing Lyu , Chun-Le Guo , Chongyi Li

Sparse autoencoders (SAEs) have emerged as a powerful tool for interpreting language model activations by decomposing them into sparse, interpretable features. A popular approach is the TopK SAE, that uses a fixed number of the most active…

Machine Learning · Computer Science 2024-12-10 Bart Bussmann , Patrick Leask , Neel Nanda

Sparse Autoencoders (SAEs) are widely employed for mechanistic interpretability and model steering. Within this context, steering is by design performed by means of decoding altered SAE intermediate representations. This procedure…

Machine Learning · Computer Science 2025-12-08 Antonio Bărbălau , Cristian Daniel Păduraru , Teodor Poncu , Alexandru Tifrea , Elena Burceanu

Data-driven reduced-order models based on autoencoders generally lack interpretability compared to classical methods such as the proper orthogonal decomposition. More interpretability can be gained by disentangling the latent variables and…

Machine Learning · Computer Science 2025-02-21 Henning Schwarz , Pyei Phyo Lin , Jens-Peter M. Zemke , Thomas Rung

The choice of an appropriate bottleneck dimension and the application of effective regularization are both essential for Autoencoders to learn meaningful representations from unlabeled data. In this paper, we introduce a new class of…

Machine Learning · Computer Science 2025-03-26 Jad Mounayer , Sebastian Rodriguez , Chady Ghnatios , Charbel Farhat , Francisco Chinesta

We introduce a group of related methods for binary classification tasks using probes of the hidden state activations in large language models (LLMs). Performance is on par with the largest and most advanced LLMs currently available, but…

Machine Learning · Computer Science 2024-08-22 John Scoville , Shang Gao , Devanshu Agrawal , Javed Qadrud-Din

Unsupervised feature extractors are known to perform an efficient and discriminative representation of data. Insight into the mappings they perform and human ability to understand them, however, remain very limited. This is especially…

Machine Learning · Computer Science 2018-12-27 Babajide O. Ayinde , Jacek M. Zurada

The autoencoder is an unsupervised learning paradigm that aims to create a compact latent representation of data by minimizing the reconstruction loss. However, it tends to overlook the fact that most data (images) are embedded in a…

Machine Learning · Computer Science 2023-10-26 Alokendu Mazumder , Tirthajit Baruah , Bhartendu Kumar , Rishab Sharma , Vishwajeet Pattanaik , Punit Rathore

Translating the internal representations and computations of models into concepts that humans can understand is a key goal of interpretability. While recent dictionary learning methods such as Sparse Autoencoders (SAEs) provide a promising…

Computation and Language · Computer Science 2026-02-27 Usha Bhalla , Alex Oesterling , Claudio Mayrink Verdun , Himabindu Lakkaraju , Flavio P. Calmon

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…

Information Retrieval · Computer Science 2025-02-25 Hao Kang , Tevin Wang , Chenyan Xiong

Sparse Autoencoder (SAE) features have become essential tools for mechanistic interpretability research. SAE features are typically characterized by examining their activating examples, which are often "monosemantic" and align with human…

Artificial Intelligence · Computer Science 2025-09-30 Claire Tian , Katherine Tian , Nathan Hu