Cover Image

Descriptive Collision in Sparse Autoencoder Auto-Interpretability: When One Explanation Describes Many Features

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	McCann, Jordan F.
Format:	Preprint
Published:	2026
Subjects:	Machine Learning I.2.6; I.2.7
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Super Apriel: One Checkpoint, Many Speeds
by: Labs, SLAM, et al.
Published: (2026)

Are LLM Uncertainty and Correctness Encoded by the Same Features? A Functional Dissociation via Sparse Autoencoders
by: Patel, Het, et al.
Published: (2026)

Control Reinforcement Learning: Interpretable Token-Level Steering of LLMs via Sparse Autoencoder Features
by: Cho, Seonglae, et al.
Published: (2026)

CorrSteer: Generation-Time LLM Steering via Correlated Sparse Autoencoder Features
by: Cho, Seonglae, et al.
Published: (2025)

Entropy-Based Measurement of Value Drift and Alignment Work in Large Language Models
by: Fadli, Samih
Published: (2025)

Learned Relay Representations for Forward-Thinking Discrete Diffusion Models
by: Rozonoyer, Benjamin, et al.
Published: (2026)

Merge-Bench: Resolve Merge Conflicts with Large Language Models
by: Schesch, Benedikt, et al.
Published: (2026)

Extreme AutoML: Analysis of Classification, Regression, and NLP Performance
by: Ratner, Edward, et al.
Published: (2024)

ACE: Exploring Activation Cosine Similarity and Variance for Accurate and Calibration-Efficient LLM Pruning
by: Mi, Zhendong, et al.
Published: (2025)

Latent Instruction Representation Alignment: defending against jailbreaks, backdoors and undesired knowledge in LLMs
by: Easley, Eric, et al.
Published: (2026)

Scalable GPU-Accelerated Euler Characteristic Curves: Optimization and Differentiable Learning for PyTorch
by: Saxena, Udit
Published: (2025)

Synergy over Discrepancy: A Partition-Based Approach to Multi-Domain LLM Fine-Tuning
by: Ye, Hua, et al.
Published: (2025)

KerZOO: Kernel Function Informed Zeroth-Order Optimization for Accurate and Accelerated LLM Fine-Tuning
by: Mi, Zhendong, et al.
Published: (2025)

Revisiting LRP: Positional Attribution as the Missing Ingredient for Transformer Explainability
by: Bakish, Yarden, et al.
Published: (2025)

QuAnTS: Question Answering on Time Series
by: Divo, Felix, et al.
Published: (2025)

Why LoRA Resists Label Noise: A Theoretical Framework for Noise-Robust Parameter-Efficient Fine-Tuning
by: Steele, Brady
Published: (2026)

Evaluating the Systematic Reasoning Abilities of Large Language Models through Graph Coloring
by: Heyman, Alex, et al.
Published: (2025)

Variance Is Not Importance: Structural Analysis of Transformer Compressibility Across Model Scales
by: Salfati, Samuel
Published: (2026)

TensorLens: End-to-End Transformer Analysis via High-Order Attention Tensors
by: Atad, Ido Andrew, et al.
Published: (2026)

Transformer Scalability Crisis: The First Comprehensive Empirical Analysis of Performance Walls in Modern Language Models
by: Moghadasi, Mahdi Naser, et al.
Published: (2026)

Task-Conditioned Routing Signatures in Sparse Mixture-of-Experts Transformers
by: Avinash, Mynampati Sri Ranganadha
Published: (2026)

Persona Features Control Emergent Misalignment
by: Wang, Miles, et al.
Published: (2025)

When Models Can't Follow: Testing Instruction Adherence Across 256 LLMs
by: Young, Richard J., et al.
Published: (2025)

Self-Pruned Key-Value Attention: Learning When to Write by Predicting Future Utility
by: Szilvasy, Gergely, et al.
Published: (2026)

Pre-trained Models Perform the Best When Token Distributions Follow Zipf's Law
by: He, Yanjin, et al.
Published: (2025)

How Pruning Reshapes Features: Sparse Autoencoder Analysis of Weight-Pruned Language Models
by: Borobia, Hector, et al.
Published: (2026)

Reasoning Large Language Model Errors Arise from Hallucinating Critical Problem Features
by: Heyman, Alex, et al.
Published: (2025)

Relating Misfit to Gain in Weak-to-Strong Generalization Beyond the Squared Loss
by: Mulgund, Abhijeet, et al.
Published: (2025)

Latent Cache Flow: Model-to-Model Communication Without Text
by: Rossi, Maximillian, et al.
Published: (2026)

From Syntax to Semantics: Unveiling the Emergence of Chirality in SMILES Translation Models
by: Li, Zehao, et al.
Published: (2026)

Listwise Direct Preference Optimization with Multi-Dimensional Preference Mixing
by: Sun, Yuhui, et al.
Published: (2025)

ASR Error Correction in Low-Resource Burmese with Alignment-Enhanced Transformers using Phonetic Features
by: Lin, Ye Bhone, et al.
Published: (2025)

Before the Last Token: Diagnosing Final-Token Safety Probe Failures
by: Doda, Shravan
Published: (2026)

Social Cooperation in Conversational AI Agents
by: Çelikok, Mustafa Mert, et al.
Published: (2025)

Discovering Transformer Circuits via a Hybrid Attribution and Pruning Framework
by: Gu, Hao, et al.
Published: (2025)

PoTS: Proof-of-Training-Steps for Backdoor Detection in Large Language Models
by: Seddik, Issam, et al.
Published: (2025)

How Does Unfaithful Reasoning Emerge from Autoregressive Training? A Study of Synthetic Experiments
by: Wang, Fuxin, et al.
Published: (2026)

HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools
by: Garg, Aashna, et al.
Published: (2026)

Quantization-Robust LLM Unlearning via Low-Rank Adaptation
by: Abitante, João Vitor Boer, et al.
Published: (2026)

Annotation Entropy Predicts Per-Example Learning Dynamics in LoRA Fine-Tuning
by: Steele, Brady
Published: (2026)