Image de couverture de livre

EntropyScan: Towards Model-level Backdoor Detection in LVLMs via Visual Attention Entropy

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Ge, Xuanyu, Wang, Zhongqi, Zhang, Jie, Shan, Shiguang, Chen, Xilin
Format:	Preprint
Publié:	2026
Sujets:	Computer Vision and Pattern Recognition
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Dynamic Attention Analysis for Backdoor Detection in Text-to-Image Diffusion Models
par: Wang, Zhongqi, et autres
Publié: (2025)

Assimilation Matters: Model-level Backdoor Detection in Vision-Language Pretrained Models
par: Wang, Zhongqi, et autres
Publié: (2025)

Trigger without Trace: Towards Stealthy Backdoor Attack on Text-to-Image Diffusion Models
par: Zhang, Jie, et autres
Publié: (2025)

T2IShield: Defending Against Backdoors on Text-to-Image Diffusion Models
par: Wang, Zhongqi, et autres
Publié: (2024)

Dysca: A Dynamic and Scalable Benchmark for Evaluating Perception Ability of LVLMs
par: Zhang, Jie, et autres
Publié: (2024)

Neural Gate: Mitigating Privacy Risks in LVLMs via Neuron-Level Gradient Gating
par: Cao, Xiangkui, et autres
Publié: (2026)

What Makes VLMs Robust? Towards Reconciling Robustness and Accuracy in Vision-Language Models
par: Nie, Sen, et autres
Publié: (2026)

V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs
par: Nie, Sen, et autres
Publié: (2025)

Towards Robust Semantic Segmentation against Patch-based Attack via Attention Refinement
par: Yuan, Zheng, et autres
Publié: (2024)

Rethinking the Evaluation of Out-of-Distribution Detection: A Sorites Paradox
par: Long, Xingming, et autres
Publié: (2024)

Semantic or Covariate? A Study on the Intractable Case of Out-of-Distribution Detection
par: Long, Xingming, et autres
Publié: (2024)

VOPE: Revisiting Hallucination of Vision-Language Models in Voluntary Imagination Task
par: Long, Xingming, et autres
Publié: (2025)

GLip: A Global-Local Integrated Progressive Framework for Robust Visual Speech Recognition
par: Wang, Tianyue, et autres
Publié: (2025)

Dual Attention Guided Defense Against Malicious Edits
par: Zhang, Jie, et autres
Publié: (2025)

FullLoRA: Efficiently Boosting the Robustness of Pretrained Vision Transformers
par: Yuan, Zheng, et autres
Publié: (2024)

ACT Now: Preempting LVLM Hallucinations via Adaptive Context Integration
par: Yan, Bei, et autres
Publié: (2026)

SHALE: A Scalable Benchmark for Fine-grained Hallucination Evaluation in LVLMs
par: Yan, Bei, et autres
Publié: (2025)

REVAL: A Comprehension Evaluation on Reliability and Values of Large Vision-Language Models
par: Zhang, Jie, et autres
Publié: (2025)

Towards Transferable Defense Against Malicious Image Edits
par: Zhang, Jie, et autres
Publié: (2025)

Contrastive Spectral Rectification: Test-Time Defense towards Zero-shot Adversarial Robustness of CLIP
par: Nie, Sen, et autres
Publié: (2026)

Jodi: Unification of Visual Generation and Understanding via Joint Modeling
par: Xu, Yifeng, et autres
Publié: (2025)

MM-MoralBench: A MultiModal Moral Evaluation Benchmark for Large Vision-Language Models
par: Yan, Bei, et autres
Publié: (2024)

Generalized Face Liveness Detection via De-fake Face Generator
par: Long, Xingming, et autres
Publié: (2024)

Measuring the Measurers: Quality Evaluation of Hallucination Benchmarks for Large Vision-Language Models
par: Yan, Bei, et autres
Publié: (2024)

Component-Based Out-of-Distribution Detection
par: Liu, Wenrui, et autres
Publié: (2026)

T2VAttack: Adversarial Attack on Text-to-Video Diffusion Models
par: Li, Changzhen, et autres
Publié: (2025)

un$^2$CLIP: Improving CLIP's Visual Detail Capturing Ability via Inverting unCLIP
par: Li, Yinqi, et autres
Publié: (2025)

HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention
par: Tang, Xiaolong, et autres
Publié: (2024)

CtrLoRA: An Extensible and Efficient Framework for Controllable Image Generation
par: Xu, Yifeng, et autres
Publié: (2024)

INFACT: A Diagnostic Benchmark for Induced Faithfulness and Factuality Hallucinations in Video-LLMs
par: Yang, Junqi, et autres
Publié: (2026)

EfficientMT: Efficient Temporal Adaptation for Motion Transfer in Text-to-Video Diffusion Models
par: Cai, Yufei, et autres
Publié: (2025)

Revisiting Logit Distributions for Reliable Out-of-Distribution Detection
par: Liang, Jiachen, et autres
Publié: (2025)

DIVE: Inverting Conditional Diffusion Models for Discriminative Tasks
par: Li, Yinqi, et autres
Publié: (2025)

VLBiasBench: A Comprehensive Benchmark for Evaluating Bias in Large Vision-Language Model
par: Wang, Sibo, et autres
Publié: (2024)

Attention Hijackers: Detect and Disentangle Attention Hijacking in LVLMs for Hallucination Mitigation
par: Chen, Beitao, et autres
Publié: (2025)

Task-adaptive Q-Face
par: Sun, Haomiao, et autres
Publié: (2024)

SAVAA: Mitigating Hallucinations in LVLMs via Step-wise Adaptive Visual Attention Amplification
par: Zhang, Jiacheng, et autres
Publié: (2026)

UMFC: Unsupervised Multi-Domain Feature Calibration for Vision-Language Models
par: Liang, Jiachen, et autres
Publié: (2024)

Pre-trained Model Guided Fine-Tuning for Zero-Shot Adversarial Robustness
par: Wang, Sibo, et autres
Publié: (2024)

Steering Vision-Language Pre-trained Models for Incremental Face Presentation Attack Detection
par: Li, Haoze, et autres
Publié: (2025)