Cover Image

Towards Interpretability Without Sacrifice: Faithful Dense Layer Decomposition with Mixture of Decoders

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Oldfield, James, Im, Shawn, Li, Sharon, Nicolaou, Mihalis A., Patras, Ioannis, Chrysos, Grigorios G
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

Multilinear Mixture of Experts: Scalable Expert Specialization through Factorization
by: Oldfield, James, et al.
Published: (2024)

Certified Robustness Under Bounded Levenshtein Distance
by: Rocamora, Elias Abad, et al.
Published: (2025)

Leveraging the Context through Multi-Round Interactions for Jailbreaking Attacks
by: Cheng, Yixin, et al.
Published: (2024)

PolySAE: Modeling Feature Interactions in Sparse Autoencoders via Polynomial Decoding
by: Koromilas, Panagiotis, et al.
Published: (2026)

REST: Efficient and Accelerated EEG Seizure Analysis through Residual State Updates
by: Afzal, Arshia, et al.
Published: (2024)

fmxcoders: Factorized Masked Crosscoders for Cross-Layer Feature Discovery
by: Demou, Andreas D., et al.
Published: (2026)

Motor Imagery Decoding Using Ensemble Curriculum Learning and Collaborative Training
by: Zoumpourlis, Georgios, et al.
Published: (2022)

Robust NAS under adversarial training: benchmark, theory, and beyond
by: Wu, Yongtao, et al.
Published: (2024)

Understanding the Learning Dynamics of Alignment with Human Feedback
by: Im, Shawn, et al.
Published: (2024)

Revisiting Character-level Adversarial Attacks for Language Models
by: Rocamora, Elias Abad, et al.
Published: (2024)

Single-pass Detection of Jailbreaking Input in Large Language Models
by: Candogan, Leyla Naz, et al.
Published: (2025)

Efficient local linearity regularization to overcome catastrophic overfitting
by: Rocamora, Elias Abad, et al.
Published: (2024)

Why DDIM Hallucinates More Than DDPM: A Theoretical Analysis of Reverse Dynamics
by: Ashiq, Muhammad H., et al.
Published: (2026)

Neural Networks Decoded: Targeted and Robust Analysis of Neural Network Decisions via Causal Explanations and Reasoning
by: Diallo, Alec F., et al.
Published: (2024)

Faithful Interpretation for Graph Neural Networks
by: Hu, Lijie, et al.
Published: (2024)

Multimodal Machine Learning in Mental Health: A Survey of Data, Algorithms, and Challenges
by: Sahili, Zahraa Al, et al.
Published: (2024)

FairCoT: Enhancing Fairness in Text-to-Image Generation via Chain of Thought Reasoning with Multimodal Large Language Models
by: Sahili, Zahraa Al, et al.
Published: (2024)

Faithful and Stable Neuron Explanations for Trustworthy Mechanistic Interpretability
by: Yan, Ge, et al.
Published: (2025)

Zero-Sacrifice Persistent-Robustness Adversarial Defense for Pre-Trained Encoders
by: Lei, Zhuxin, et al.
Published: (2026)

Visual Instruction Bottleneck Tuning
by: Oh, Changdae, et al.
Published: (2025)

Dense Backpropagation Improves Training for Sparse Mixture-of-Experts
by: Panda, Ashwinee, et al.
Published: (2025)

FaithLM: Towards Faithful Explanations for Large Language Models
by: Chuang, Yu-Neng, et al.
Published: (2024)

Towards Metric-Faithful Neural Graph Matching
by: Shivottam, Jyotirmaya, et al.
Published: (2026)

Mixture of Layers with Hybrid Attention
by: Ternovtsii, Ivan, et al.
Published: (2026)

EUGens: Efficient, Unified, and General Dense Layers
by: Kim, Sang Min, et al.
Published: (2026)

Towards Faithful Explanations: Boosting Rationalization with Shortcuts Discovery
by: Yue, Linan, et al.
Published: (2024)

Improving Interpretation Faithfulness for Vision Transformers
by: Hu, Lijie, et al.
Published: (2023)

Mixture of Attentions For Speculative Decoding
by: Zimmer, Matthieu, et al.
Published: (2024)

Going beyond Compositions, DDPMs Can Produce Zero-Shot Interpolations
by: Deschenaux, Justin, et al.
Published: (2024)

Beyond Linear Probes: Dynamic Safety Monitoring for Language Models
by: Oldfield, James, et al.
Published: (2025)

Faithfulness Evaluation for Decoder-only LLM Attributions with Controlled Retained Information
by: Huang, Xin, et al.
Published: (2026)

Bilinear Convolution Decomposition for Causal RL Interpretability
by: Oozeer, Narmeen, et al.
Published: (2024)

Pruning and Distilling Mixture-of-Experts into Dense Language Models
by: Kim, Junhyuck, et al.
Published: (2026)

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

FashionSD-X: Multimodal Fashion Garment Synthesis using Latent Diffusion
by: Singh, Abhishek Kumar, et al.
Published: (2024)

Recursive Speculative Decoding: Accelerating LLM Inference via Sampling Without Replacement
by: Jeon, Wonseok, et al.
Published: (2024)

Understanding Multimodal LLMs Under Distribution Shifts: An Information-Theoretic Approach
by: Oh, Changdae, et al.
Published: (2025)

DecoKAN: Interpretable Decomposition for Forecasting Cryptocurrency Market Dynamics
by: Gao, Yuan, et al.
Published: (2025)

MLPMoE: Zero-Shot Architectural Metamorphosis of Dense LLM MLPs into Static Mixture-of-Experts
by: Novikov, Ivan
Published: (2025)

Interpretable Diffusion via Information Decomposition
by: Kong, Xianghao, et al.
Published: (2023)