paper-with-me

홈 › Papers

Beyond Output Faithfulness: Learning Attributions that Preserve Computational Pathways

2025-09-04 · Siyu Zhang, Kenneth Mcmillan arxiv

Faithfulness metrics such as insertion and deletion evaluate how feature removal affects model outputs but overlook whether explanations preserve the computational pathway the network actually uses. We show that external metrics can be maximized through alternative pathways -- perturbations that reroute computation via different feature detectors while preserving output behavior. To address this, we propose activation preservation as a tractable proxy for preserving computational pathways We introduce Faithfulness-guided Ensemble Interpretation (FEI), which jointly optimizes external faithfulness (via ensemble quantile optimization of insertion/deletion curves) and internal faithfulness (via selective gradient clipping). Across VGG and ResNet on ImageNet and CUB-200-2011, FEI achieves state-of-the-art insertion/deletion scores while maintaining significantly lower activation deviation, showing that both external and internal faithfulness are essential for reliable explanations.

📄 PDF Abstract BibTeX arXiv:2509.04588

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

AttributionLab: Faithfulness of Feature Attribution Under Controllable Environments

2023-10-10 · Yang Zhang, Yawei Li, Hannah Brown, Mina Rezaei 외

Feature attribution explains neural network outputs by identifying relevant input features. The attribution has to be faithful, meaning that the attributed features must mirror the input features that influence the outpu…

AIM: Adversarial Information Masking for Faithfulness Evaluation of Saliency Maps

2026-05-16 · Chia-Ying Hsieh, Hsin-Yuan Fang, Chun-Shu Wei arxiv

Post-hoc saliency methods are widely used to interpret deep neural networks, but their faithfulness is difficult to evaluate reliably. Existing evaluations mask features according to saliency-induced feature ordering and…

Explaining the Reasoning of Large Language Models Using Attribution Graphs

2025-12-17 · Chase Walker, Rickard Ewetz arxiv

Large language models (LLMs) exhibit remarkable capabilities, yet their reasoning remains opaque, raising safety and trust concerns. Attribution methods, which assign credit to input features, have proven effective for e…

Decision Making

Towards Faithful Natural Language Explanations: A Study Using Activation Patching in Large Language Models

2024-10-18 · Wei Jie Yeo, Ranjan Satapathy, Erik Cambria

Large Language Models (LLMs) are capable of generating persuasive Natural Language Explanations (NLEs) to justify their answers. However, the faithfulness of these explanations should not be readily trusted at face value…

What Helps---and What Hurts: Bidirectional Explanations for Vision Transformers

2026-03-02 · Qin Su, Tie Luo arxiv

Vision Transformers (ViTs) achieve strong performance in visual recognition, yet their decision-making remains difficult to interpret. We propose BiCAM, a bidirectional class activation mapping method that captures both …