Cover Image

From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Mueller, Aaron, Lee, Andrew, Joshi, Shruti, Lubana, Ekdeep Singh, Sridhar, Dhanya, Reizinger, Patrik
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

Causality is Key for Interpretability Claims to Generalise
by: Joshi, Shruti, et al.
Published: (2026)

Sparse Shift Autoencoders for Identifying Concepts from Large Language Model Activations
by: Joshi, Shruti, et al.
Published: (2025)

Who Guards the Guardians? The Challenges of Evaluating Identifiability of Learned Representations
by: Joshi, Shruti, et al.
Published: (2026)

How Do LLMs Persuade? Linear Probes Can Uncover Persuasion Dynamics in Multi-Turn Conversations
by: Jaipersaud, Brandon, et al.
Published: (2025)

Are language models aware of the road not taken? Token-level uncertainty and hidden state dynamics
by: Zur, Amir, et al.
Published: (2025)

Towards Reliable Evaluation of Behavior Steering Interventions in LLMs
by: Pres, Itamar, et al.
Published: (2024)

Competition Dynamics Shape Algorithmic Phases of In-Context Learning
by: Park, Core Francisco, et al.
Published: (2024)

Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Space
by: Park, Core Francisco, et al.
Published: (2024)

ICLR: In-Context Learning of Representations
by: Park, Core Francisco, et al.
Published: (2024)

In-Context Learning Dynamics with Random Binary Sequences
by: Bigelow, Eric J., et al.
Published: (2023)

Missed Causes and Ambiguous Effects: Counterfactuals Pose Challenges for Interpreting Neural Networks
by: Mueller, Aaron
Published: (2024)

Projecting Assumptions: The Duality Between Sparse Autoencoders and Concept Geometry
by: Hindupur, Sai Sumedh R., et al.
Published: (2025)

Abrupt Learning in Transformers: A Case Study on Matrix Completion
by: Gopalani, Pulkit, et al.
Published: (2024)

Rule Extrapolation in Language Models: A Study of Compositional Generalization on OOD Prompts
by: Mészáros, Anna, et al.
Published: (2024)

Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering
by: Bigelow, Eric, et al.
Published: (2025)

Emergence of Hierarchical Emotion Organization in Large Language Models
by: Zhao, Bo, et al.
Published: (2025)

What Makes and Breaks Safety Fine-tuning? A Mechanistic Study
by: Jain, Samyak, et al.
Published: (2024)

Compositional Capabilities of Autoregressive Transformers: A Study on Synthetic, Interpretable Tasks
by: Ramesh, Rahul, et al.
Published: (2023)

The Shape of Beliefs: Geometry, Dynamics, and Interventions along Representation Manifolds of Language Models' Posteriors
by: Sarfati, Raphaël, et al.
Published: (2026)

Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts
by: Feucht, Sheridan, et al.
Published: (2026)

Stories in Space: In-Context Learning Trajectories in Conceptual Belief Space
by: Bigelow, Eric, et al.
Published: (2026)

Swing-by Dynamics in Concept Learning and Compositional Generalization
by: Yang, Yongyi, et al.
Published: (2024)

From Flat to Hierarchical: Extracting Sparse Representations with Matching Pursuit
by: Costa, Valérie, et al.
Published: (2025)

Evaluating Sparse Autoencoders: From Shallow Design to Matching Pursuit
by: Costa, Valérie, et al.
Published: (2025)

An Interventional Perspective on Identifiability in Gaussian LTI Systems with Independent Component Analysis
by: Rajendran, Goutham, et al.
Published: (2023)

Uncovering Conceptual Blindspots in Generative Image Models Using Sparse Autoencoders
by: Bohacek, Matyas, et al.
Published: (2025)

Analyzing (In)Abilities of SAEs via Formal Languages
by: Menon, Abhinav, et al.
Published: (2024)

Position: An Empirically Grounded Identifiability Theory Will Accelerate Self-Supervised Learning Research
by: Reizinger, Patrik, et al.
Published: (2025)

Large Language Models Share Representations of Latent Grammatical Concepts Across Typologically Diverse Languages
by: Brinkmann, Jannik, et al.
Published: (2025)

Do Sparse Autoencoders Capture Concept Manifolds?
by: Bhalla, Usha, et al.
Published: (2026)

When Interpretability Becomes a Liability: Adversarial Attacks on CBM Concept Layers
by: Sridhar, Aditya
Published: (2026)

Use Sparse Autoencoders to Discover Unknown Concepts, Not to Act on Known Concepts
by: Peng, Kenny, et al.
Published: (2025)

Concept-Based Interpretability for Toxicity Detection
by: Garg, Samarth, et al.
Published: (2025)

RAVEL: Evaluating Interpretability Methods on Disentangling Language Model Representations
by: Huang, Jing, et al.
Published: (2024)

A Percolation Model of Emergence: Analyzing Transformers Trained on a Formal Language
by: Lubana, Ekdeep Singh, et al.
Published: (2024)

Compositional Abilities Emerge Multiplicatively: Exploring Diffusion Models on a Synthetic Task
by: Okawa, Maya, et al.
Published: (2023)

Identifiable Exchangeable Mechanisms for Causal Structure and Representation Learning
by: Reizinger, Patrik, et al.
Published: (2024)

Into the Rabbit Hull: From Task-Relevant Concepts in DINO to Minkowski Geometry
by: Fel, Thomas, et al.
Published: (2025)

Sparse Graph Representations for Procedural Instructional Documents
by: Singh, Shruti, et al.
Published: (2024)

Evaluating Interventional Reasoning Capabilities of Large Language Models
by: Kasetty, Tejas, et al.
Published: (2024)