Cover Image

From What to How: Attributing CLIP's Latent Components Reveals Unexpected Semantic Reliance

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Dreyer, Maximilian, Hufe, Lorenz, Berend, Jim, Wiegand, Thomas, Lapuschkin, Sebastian, Samek, Wojciech
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Dyslexify: A Mechanistic Defense Against Typographic Attacks in CLIP
by: Hufe, Lorenz, et al.
Published: (2025)

Mechanistic understanding and validation of large AI models with SemanticLens
by: Dreyer, Maximilian, et al.
Published: (2025)

From Attribution Maps to Human-Understandable Explanations through Concept Relevance Propagation
by: Achtibat, Reduan, et al.
Published: (2022)

Pruning By Explaining Revisited: Optimizing Attribution Methods to Prune CNNs and Transformers
by: Hatefi, Sayed Mohammad Vakilzadeh, et al.
Published: (2024)

Atlas-Alignment: Making Interpretability Transferable Across Language Models
by: Puri, Bruno, et al.
Published: (2025)

Sparse, Efficient and Explainable Data Attribution with DualXDA
by: Yolcu, Galip Ümit, et al.
Published: (2024)

Understanding the (Extra-)Ordinary: Validating Deep Model Decisions with Prototypical Concept-based Explanations
by: Dreyer, Maximilian, et al.
Published: (2023)

From Attribution to Action: A Human-Centered Application of Activation Steering
by: Labarta, Tobias, et al.
Published: (2026)

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data
by: Pahde, Frederik, et al.
Published: (2025)

Contrastive Semantic Projection: Faithful Neuron Labeling with Contrastive Examples
by: Bouanani, Oussama, et al.
Published: (2026)

Attribution-Guided Pruning for Insight and Control: Circuit Discovery and Targeted Correction in Small-scale LLMs
by: Hatefi, Sayed Mohammad Vakilzadeh, et al.
Published: (2025)

Efficient and Flexible Neural Network Training through Layer-wise Feedback Propagation
by: Weber, Leander, et al.
Published: (2023)

ECQ$^{\text{x}}$: Explainability-Driven Quantization for Low-Bit and Sparse DNNs
by: Becking, Daniel, et al.
Published: (2021)

PURE: Turning Polysemantic Neurons Into Pure Features by Identifying Relevant Circuits
by: Dreyer, Maximilian, et al.
Published: (2024)

Reactive Model Correction: Mitigating Harm to Task-Relevant Features via Conditional Bias Suppression
by: Bareeva, Dilyara, et al.
Published: (2024)

Quanda: An Interpretability Toolkit for Training Data Attribution Evaluation and Beyond
by: Bareeva, Dilyara, et al.
Published: (2024)

AttnLRP: Attention-Aware Layer-Wise Relevance Propagation for Transformers
by: Achtibat, Reduan, et al.
Published: (2024)

Navigating Neural Space: Revisiting Concept Activation Vectors to Overcome Directional Divergence
by: Pahde, Frederik, et al.
Published: (2022)

Iterative Inference in a Chess-Playing Neural Network
by: Sandmann, Elias, et al.
Published: (2025)

X-SYS: A Reference Architecture for Interactive Explanation Systems
by: Labarta, Tobias, et al.
Published: (2026)

Post-Hoc Concept Disentanglement: From Correlated to Isolated Concept Representations
by: Erogullari, Eren, et al.
Published: (2025)

FADE: Why Bad Descriptions Happen to Good Features
by: Puri, Bruno, et al.
Published: (2025)

Explaining Predictive Uncertainty by Exposing Second-Order Effects
by: Bley, Florian, et al.
Published: (2024)

Steering CLIP's vision transformer with sparse autoencoders
by: Joseph, Sonia, et al.
Published: (2025)

Playing the network backward: A Game Theoretic Attribution Framework
by: Zimmermann, Jakob Paul, et al.
Published: (2026)

LieSolver: A PDE-constrained solver for IBVPs using Lie symmetries
by: Klausen, René P., et al.
Published: (2025)

Structural Compactness as a Complementary Criterion for Explanation Quality
by: Mesgari, Mohammad Mahdi, et al.
Published: (2026)

PINNfluence: Influence Functions for Physics-Informed Neural Networks
by: Naujoks, Jonas R., et al.
Published: (2024)

Attribution-Guided Decoding
by: Komorowski, Piotr, et al.
Published: (2025)

Explainable concept mappings of MRI: Revealing the mechanisms underlying deep learning-based brain disease classification
by: Tinauer, Christian, et al.
Published: (2024)

Human-Centered Evaluation of XAI Methods
by: Dawoud, Karam, et al.
Published: (2023)

The Atlas of In-Context Learning: How Attention Heads Shape In-Context Retrieval Augmentation
by: Kahardipraja, Patrick, et al.
Published: (2025)

Relevance-driven Input Dropout: an Explanation-guided Regularization Technique
by: Gururaj, Shreyas, et al.
Published: (2025)

Leveraging Influence Functions for Resampling Data in Physics-Informed Neural Networks
by: Naujoks, Jonas R., et al.
Published: (2025)

Beyond Scalars: Concept-Based Alignment Analysis in Vision Transformers
by: Vielhaben, Johanna, et al.
Published: (2024)

Circuit Insights: Towards Interpretability Beyond Activations
by: Golimblevskaia, Elena, et al.
Published: (2025)

Building Trust in PINNs: Error Estimation through Finite Difference Methods
by: Krasowski, Aleksander, et al.
Published: (2026)

See What I Mean? CUE: A Cognitive Model of Understanding Explanations
by: Labarta, Tobias, et al.
Published: (2025)

Model Science: getting serious about verification, explanation and control of AI systems
by: Biecek, Przemyslaw, et al.
Published: (2025)

XAI-guided Insulator Anomaly Detection for Imbalanced Datasets
by: Hoefler, Maximilian Andreas, et al.
Published: (2024)