Cover Image

Dense SAE Latents Are Features, Not Bugs

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Sun, Xiaoqing, Stolfo, Alessandro, Engels, Joshua, Wu, Ben, Rajamanoharan, Senthooran, Sachan, Mrinmaya, Tegmark, Max
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence Computation and Language
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Are Sparse Autoencoders Useful? A Case Study in Sparse Probing
by: Kantamneni, Subhash, et al.
Published: (2025)

Probing for Arithmetic Errors in Language Models
by: Sun, Yucheng, et al.
Published: (2025)

Confidence Regulation Neurons in Language Models
by: Stolfo, Alessandro, et al.
Published: (2024)

Do I Know This Entity? Knowledge Awareness and Hallucinations in Language Models
by: Ferrando, Javier, et al.
Published: (2024)

Do Language Models Exhibit the Same Cognitive Biases in Problem Solving as Human Learners?
by: Opedal, Andreas, et al.
Published: (2024)

Thought Branches: Interpreting LLM Reasoning Requires Resampling
by: Macar, Uzay, et al.
Published: (2025)

Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning
by: Casademunt, Helena, et al.
Published: (2025)

Chain-of-Thought Reasoning In The Wild Is Not Always Faithful
by: Arcuschin, Iván, et al.
Published: (2025)

Simple Mechanistic Explanations for Out-Of-Context Reasoning
by: Wang, Atticus, et al.
Published: (2025)

Simulating Students or Sycophantic Problem Solving? On Misconception Faithfulness of LLM Simulators
by: Do, Heejin, et al.
Published: (2026)

Language Models Use Trigonometry to Do Addition
by: Kantamneni, Subhash, et al.
Published: (2025)

Language Models Represent Space and Time
by: Gurnee, Wes, et al.
Published: (2023)

On the Emergence of Induction Heads for In-Context Learning
by: Musat, Tiberiu, et al.
Published: (2025)

Improving Large Language Model Safety with Contrastive Representation Learning
by: Simko, Samuel, et al.
Published: (2025)

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

Towards Aligning Language Models with Textual Feedback
by: Lloret, Saüc Abadal, et al.
Published: (2024)

Scaling Laws For Scalable Oversight
by: Engels, Joshua, et al.
Published: (2025)

Stepwise Verification and Remediation of Student Reasoning Errors with Large Language Model Tutors
by: Daheim, Nico, et al.
Published: (2024)

MathGAP: Out-of-Distribution Evaluation on Problems with Arbitrarily Complex Proofs
by: Opedal, Andreas, et al.
Published: (2024)

MathTutorBench: A Benchmark for Measuring Open-ended Pedagogical Capabilities of LLM Tutors
by: Macina, Jakub, et al.
Published: (2025)

Pointwise Mutual Information as a Performance Gauge for Retrieval-Augmented Generation
by: Liu, Tianyu, et al.
Published: (2024)

Improving Instruction-Following in Language Models through Activation Steering
by: Stolfo, Alessandro, et al.
Published: (2024)

Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
by: Lieberum, Tom, et al.
Published: (2024)

Learning to Reason Efficiently with A* Post-Training
by: Opedal, Andreas, et al.
Published: (2026)

The Remarkable Robustness of LLMs: Stages of Inference?
by: Lad, Vedang, et al.
Published: (2024)

Emergent Misalignment is Easy, Narrow Misalignment is Hard
by: Soligo, Anna, et al.
Published: (2026)

Convergent Linear Representations of Emergent Misalignment
by: Soligo, Anna, et al.
Published: (2025)

Groundedness in Retrieval-augmented Long-form Generation: An Empirical Study
by: Stolfo, Alessandro
Published: (2024)

Can Large Language Models Infer Causation from Correlation?
by: Jin, Zhijing, et al.
Published: (2023)

Autoformalizing Natural Language to First-Order Logic: A Case Study in Logical Fallacy Detection
by: Lalwani, Abhinav, et al.
Published: (2024)

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

CLadder: Assessing Causal Reasoning in Language Models
by: Jin, Zhijing, et al.
Published: (2023)

Compositional Literary Primitives in Instruction-Tuned LLMs: Cross-Architectural SAE Features for Self, Style, and Affect
by: Presa, Joao Paulo Cavalcante, et al.
Published: (2026)

Quriosity: Analyzing Human Questioning Behavior and Causal Inquiry through Curiosity-Driven Queries
by: Ceraolo, Roberto, et al.
Published: (2024)

Model Organisms for Emergent Misalignment
by: Turner, Edward, et al.
Published: (2025)

SAE-V: Interpreting Multimodal Models for Enhanced Alignment
by: Lou, Hantao, et al.
Published: (2025)

Interpretable Safety Alignment via SAE-Constructed Low-Rank Subspace Adaptation
by: Wang, Dianyun, et al.
Published: (2025)

Fantastic Bugs and Where to Find Them in AI Benchmarks
by: Truong, Sang, et al.
Published: (2025)

CausalCite: A Causal Formulation of Paper Citations
by: Kumar, Ishan, et al.
Published: (2023)

Are Language Models Efficient Reasoners? A Perspective from Logic Programming
by: Opedal, Andreas, et al.
Published: (2025)