Tree SAE: Learning Hierarchical Feature Structures in Sparse Autoencoders

arXiv:2605.07922 · cs.LG · Submitted 2026-05-08 · Read on arXiv

cs.LG

Submitted: 2026-05-08

Updated: 2026-10-07

Code: https://github.com/tue147/tree-sae

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes
• Sparse Autoencoders (SAEs) — 1× in this paper · explained in 3 episodes

Sources

• Interpretability as Compression: Reconsidering SAE Explanations of Neural Activations with MDL-SAEs
• BatchTopK Sparse Autoencoders
• Learning Multi-Level Features with Matryoshka Sparse Autoencoders
• A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
• From Flat to Hierarchical: Extracting Sparse Representations with Matching Pursuit
• Sparse Autoencoders Find Highly Interpretable Features in Language Models
• The Pile: An 800GB Dataset of Diverse Text for Language Modeling
• Scaling and evaluating sparse autoencoders
• Finding Neurons in a Haystack: Case Studies with Sparse Probing
• SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
• Sparse Autoencoders Do Not Find Canonical Units of Analysis
• Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
• Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
• Overcoming Sparsity Artifacts in Crosscoders to Interpret Chat-Tuning
• The Geometry of Categorical and Hierarchical Concepts in Large Language Models
• The Linear Representation Hypothesis and the Geometry of Large Language Models
• Automatically Interpreting Millions of Features in Large Language Models
• Jumping Ahead: Improving Reconstruction Fidelity with JumpReLU Sparse Autoencoders
• Dense SAE Latents Are Features, Not Bugs

Related papers

• Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
• AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
• Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
• Convergence issues in Relational Concept Analysis based on AOC-posets
• Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
• Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks