Cover Image

Aligned Training: A Parameter-Free Method to Improve Feature Quality and Stability of Sparse Autoencoders (SAE)

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Brzozowski, Michał, Chung, Neo Christopher
Format:	Preprint
Published:	2026
Subjects:	Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Ablating Archetypes: The Stability of Archetypal SAEs is an Artifact of Initialization and Metric Design
by: Brzozowski, Michał, et al.
Published: (2026)

GPart: End-to-End Isometric Fine-Tuning via Global Parameter Partitioning
by: Mandica, Paolo, et al.
Published: (2026)

AlignSAE: Concept-Aligned Sparse Autoencoders
by: Yang, Minglai, et al.
Published: (2025)

Reading the Finetuning Prior: Verbatim Content Recovery via Contrastive Decoding Diffing
by: Brzozowski, Michał, et al.
Published: (2026)

Tree SAE: Learning Hierarchical Feature Structures in Sparse Autoencoders
by: Cao, Tue M., et al.
Published: (2026)

OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features
by: Korznikov, Anton, et al.
Published: (2025)

SAE-FD: Sparse Autoencoder Feature Distillation for Continual Learning of Large Language Models
by: Zhang, Mingxu, et al.
Published: (2026)

PolySAE: Modeling Feature Interactions in Sparse Autoencoders via Polynomial Decoding
by: Koromilas, Panagiotis, et al.
Published: (2026)

Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders
by: Marks, Luke, et al.
Published: (2024)

FaithfulSAE: Towards Capturing Faithful Features with Sparse Autoencoders without External Dataset Dependencies
by: Cho, Seonglae, et al.
Published: (2025)

SAE-FiRE: Enhancing Earnings Surprise Predictions Through Sparse Autoencoder Feature Selection
by: Zhang, Huopu, et al.
Published: (2025)

ReSAE: Residualized Sparse Autoencoders for Multi-Layer Transformer Interventions
by: Poduval, Prathyush, et al.
Published: (2026)

Sparse Autoencoders Trained on the Same Data Learn Different Features
by: Paulo, Gonçalo, et al.
Published: (2025)

SoftSAE: Dynamic Top-K Selection for Adaptive Sparse Autoencoders
by: Stępień, Jakub, et al.
Published: (2026)

Improving Steering Vectors by Targeting Sparse Autoencoder Features
by: Chalnev, Sviatoslav, et al.
Published: (2024)

Regularizing Attention Scores with Bootstrapping
by: Chung, Neo Christopher, et al.
Published: (2026)

Improving Sparse Autoencoder with Dynamic Attention
by: Wang, Dongsheng, et al.
Published: (2026)

Evolution of SAE Features Across Layers in LLMs
by: Balcells, Daniel, et al.
Published: (2024)

Sparse Autoencoder Features for Classifications and Transferability
by: Gallifant, Jack, et al.
Published: (2025)

Do Sparse Autoencoders Identify Reasoning Features in Language Models?
by: Ma, George, et al.
Published: (2026)

On the Relationship Between Activation Outliers and Feature Death in Sparse Autoencoders
by: Simon, Elana, et al.
Published: (2026)

SplInterp: Improving our Understanding and Training of Sparse Autoencoders
by: Budd, Jeremy, et al.
Published: (2025)

Adaptive Sparse Allocation with Mutual Choice & Feature Choice Sparse Autoencoders
by: Ayonrinde, Kola
Published: (2024)

Training Superior Sparse Autoencoders for Instruct Models
by: Li, Jiaming, et al.
Published: (2025)

Improving Dictionary Learning with Gated Sparse Autoencoders
by: Rajamanoharan, Senthooran, et al.
Published: (2024)

Data Whitening Improves Sparse Autoencoder Learning
by: Saraswatula, Ashwin, et al.
Published: (2025)

MoRFI: Monotonic Sparse Autoencoder Feature Identification
by: Dimakopoulos, Dimitris, et al.
Published: (2026)

Learning Multi-Level Features with Matryoshka Sparse Autoencoders
by: Bussmann, Bart, et al.
Published: (2025)

Time-Aware Feature Selection: Adaptive Temporal Masking for Stable Sparse Autoencoder Training
by: Li, T. Ed, et al.
Published: (2025)

Feature Starvation as Geometric Instability in Sparse Autoencoders
by: Chaudhry, Faris, et al.
Published: (2026)

The Geometry of Concepts: Sparse Autoencoder Feature Structure
by: Li, Yuxiao, et al.
Published: (2024)

RandAlign: A Parameter-Free Method for Regularizing Graph Convolutional Networks
by: Zhang, Haimin, et al.
Published: (2024)

GeoSAE: Geometric Prior-Guided Layer-Wise Sparse Autoencoder Annotation of Brain MRI Foundation Models
by: Nerrise, Favour, et al.
Published: (2026)

Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders
by: Chanin, David, et al.
Published: (2025)

Representation-based Broad Hallucination Detectors Fail to Generalize Out of Distribution
by: Dubanowska, Zuzanna, et al.
Published: (2025)

Improving Robustness In Sparse Autoencoders via Masked Regularization
by: Narayanaswamy, Vivek, et al.
Published: (2026)

Kronecker Factorization Improves Efficiency and Interpretability of Sparse Autoencoders
by: Kurochkin, Vadim, et al.
Published: (2025)

Dense SAE Latents Are Features, Not Bugs
by: Sun, Xiaoqing, et al.
Published: (2025)

Class-Discriminative Attention Maps for Vision Transformers
by: Brocki, Lennart, et al.
Published: (2023)

Visual Exploration of Feature Relationships in Sparse Autoencoders with Curated Concepts
by: Yan, Xinyuan, et al.
Published: (2025)