Copertina

FADE: Why Bad Descriptions Happen to Good Features

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Puri, Bruno, Jain, Aakriti, Golimblevskaia, Elena, Kahardipraja, Patrick, Wiegand, Thomas, Samek, Wojciech, Lapuschkin, Sebastian
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Machine Learning Artificial Intelligence Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

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

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

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

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

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

A Close Look at Decomposition-based XAI-Methods for Transformer Language Models
di: Arras, Leila, et al.
Pubblicazione: (2025)

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

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

Attribution-Guided Decoding
di: Komorowski, Piotr, et al.
Pubblicazione: (2025)

From What to How: Attributing CLIP's Latent Components Reveals Unexpected Semantic Reliance
di: Dreyer, Maximilian, et al.
Pubblicazione: (2025)

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

Simulation, Modelling and Classification of Wiki Contributors: Spotting The Good, The Bad, and The Ugly
di: Méndez, Silvia García, et al.
Pubblicazione: (2024)

When Bad Data Leads to Good Models
di: Li, Kenneth, et al.
Pubblicazione: (2025)

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

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

The Good, The Bad, and Why: Unveiling Emotions in Generative AI
di: Li, Cheng, et al.
Pubblicazione: (2023)

AgentCollabBench: Diagnosing When Good Agents Make Bad Collaborators
di: Mazumder, Aritra, et al.
Pubblicazione: (2026)

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

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

Comparing Bad Apples to Good Oranges: Aligning Large Language Models via Joint Preference Optimization
di: Bansal, Hritik, et al.
Pubblicazione: (2024)

Position: Explain to Question not to Justify
di: Biecek, Przemyslaw, et al.
Pubblicazione: (2024)

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

Strategic Dishonesty Can Undermine AI Safety Evaluations of Frontier LLMs
di: Panfilov, Alexander, et al.
Pubblicazione: (2025)

SAEs Are Good for Steering -- If You Select the Right Features
di: Arad, Dana, et al.
Pubblicazione: (2025)