Image de couverture de livre

AbsTopK: Rethinking Sparse Autoencoders For Bidirectional Features

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhu, Xudong, Khalili, Mohammad Mahdi, Zhu, Zhihui
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning Artificial Intelligence Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

From Emergence to Control: Probing and Modulating Self-Reflection in Language Models
par: Zhu, Xudong, et autres
Publié: (2025)

Sparse Autoencoder Features for Classifications and Transferability
par: Gallifant, Jack, et autres
Publié: (2025)

On the transferability of Sparse Autoencoders for interpreting compressed models
par: Gupte, Suchit, et autres
Publié: (2025)

Rethinking Evaluation of Sparse Autoencoders through the Representation of Polysemous Words
par: Minegishi, Gouki, et autres
Publié: (2025)

Feature Hedging: Correlated Features Break Narrow Sparse Autoencoders
par: Chanin, David, et autres
Publié: (2025)

Improving Steering Vectors by Targeting Sparse Autoencoder Features
par: Chalnev, Sviatoslav, et autres
Publié: (2024)

Sparse but Wrong: Incorrect L0 Leads to Incorrect Features in Sparse Autoencoders
par: Chanin, David, et autres
Publié: (2025)

BatchTopK Sparse Autoencoders
par: Bussmann, Bart, et autres
Publié: (2024)

Time-Aware Feature Selection: Adaptive Temporal Masking for Stable Sparse Autoencoder Training
par: Li, T. Ed, et autres
Publié: (2025)

Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders
par: Lan, Michael, et autres
Publié: (2024)

FaithfulSAE: Towards Capturing Faithful Features with Sparse Autoencoders without External Dataset Dependencies
par: Cho, Seonglae, et autres
Publié: (2025)

Dissecting Chronos: Sparse Autoencoders Reveal Causal Feature Hierarchies in Time Series Foundation Models
par: Mishra, Anurag
Publié: (2026)

Sparse Autoencoder Decomposition of Clinical Sequence Model Representations: Feature Complexity, Task Specialisation, and Mortality Prediction
par: Sainsbury, Chris, et autres
Publié: (2026)

Incorporating Hierarchical Semantics in Sparse Autoencoder Architectures
par: Muchane, Mark, et autres
Publié: (2025)

Jacobian Sparse Autoencoders: Sparsify Computations, Not Just Activations
par: Farnik, Lucy, et autres
Publié: (2025)

Evaluating Adversarial Robustness of Concept Representations in Sparse Autoencoders
par: Li, Aaron J., et autres
Publié: (2025)

Temporal Sparse Autoencoders: Leveraging the Sequential Nature of Language for Interpretability
par: Bhalla, Usha, et autres
Publié: (2025)

SAEMark: Steering Personalized Multilingual LLM Watermarks with Sparse Autoencoders
par: Yu, Zhuohao, et autres
Publié: (2025)

Beyond Input Activations: Identifying Influential Latents by Gradient Sparse Autoencoders
par: Shu, Dong, et autres
Publié: (2025)

Steering LLMs? Actually, Sparse Autoencoders can outperform simple baselines
par: Jørgensen, Mikkel Godsk, et autres
Publié: (2026)

From Compression to Expression: A Layerwise Analysis of In-Context Learning
par: Jiang, Jiachen, et autres
Publié: (2025)

Accelerating Prefilling for Long-Context LLMs via Sparse Pattern Sharing
par: Peng, Dan, et autres
Publié: (2025)

A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models
par: Shu, Dong, et autres
Publié: (2025)

Diversity-driven Data Selection for Language Model Tuning through Sparse Autoencoder
par: Yang, Xianjun, et autres
Publié: (2025)

Sparse Shift Autoencoders for Identifying Concepts from Large Language Model Activations
par: Joshi, Shruti, et autres
Publié: (2025)

Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
par: Lieberum, Tom, et autres
Publié: (2024)

DLM-Scope: Mechanistic Interpretability of Diffusion Language Models via Sparse Autoencoders
par: Wang, Xu, et autres
Publié: (2026)

Towards Understanding the Robustness of Sparse Autoencoders
par: Saiyed, Ahson, et autres
Publié: (2026)

Post-processing for Fair Regression via Explainable SVD
par: Zuo, Zhiqun, et autres
Publié: (2025)

An Efficient Training Algorithm for Models with Block-wise Sparsity
par: Zhu, Ding, et autres
Publié: (2025)

Does higher interpretability imply better utility? A Pairwise Analysis on Sparse Autoencoders
par: Wang, Xu, et autres
Publié: (2025)

SAIF: A Sparse Autoencoder Framework for Interpreting and Steering Instruction Following of Language Models
par: He, Zirui, et autres
Publié: (2025)

Decoding Dark Matter: Specialized Sparse Autoencoders for Interpreting Rare Concepts in Foundation Models
par: Muhamed, Aashiq, et autres
Publié: (2024)

Guiding LLM Post-training Data Engineering with Model Internals from Sparse Autoencoders
par: Jing, Yi, et autres
Publié: (2026)

Gap-K%: Measuring Top-1 Prediction Gap for Detecting Pretraining Data
par: Kwak, Minseo, et autres
Publié: (2026)

Dense Training, Sparse Inference: Rethinking Training of Mixture-of-Experts Language Models
par: Pan, Bowen, et autres
Publié: (2024)

Use Sparse Autoencoders to Discover Unknown Concepts, Not to Act on Known Concepts
par: Peng, Kenny, et autres
Publié: (2025)

Neuroplasticity and Corruption in Model Mechanisms: A Case Study Of Indirect Object Identification
par: Chhabra, Vishnu Kabir, et autres
Publié: (2025)

Attention Mechanisms Don't Learn Additive Models: Rethinking Feature Importance for Transformers
par: Leemann, Tobias, et autres
Publié: (2024)

Safe-SAIL: Towards a Fine-grained Safety Landscape of Large Language Models via Sparse Autoencoder Interpretation Framework
par: Weng, Jiaqi, et autres
Publié: (2025)