Cover Image

Attention Hijackers: Detect and Disentangle Attention Hijacking in LVLMs for Hallucination Mitigation

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Chen, Beitao, Lyu, Xinyu, Gao, Lianli, Song, Jingkuan, Shen, Heng Tao
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Alleviating Hallucinations in Large Vision-Language Models through Hallucination-Induced Optimization
by: Lyu, Xinyu, et al.
Published: (2024)

SafePTR: Token-Level Jailbreak Defense in Multimodal LLMs via Prune-then-Restore Mechanism
by: Chen, Beitao, et al.
Published: (2025)

Vocabulary Hijacking in LVLMs: Unveiling Critical Attention Heads by Excluding Inert Tokens to Mitigate Hallucination
by: Chen, Yangneng, et al.
Published: (2026)

FlexAC: Towards Flexible Control of Associative Reasoning in Multimodal Large Language Models
by: Yuan, Shengming, et al.
Published: (2025)

ALF: Adaptive Label Finetuning for Scene Graph Generation
by: Chen, Qishen, et al.
Published: (2023)

CFReID: Continual Few-shot Person Re-Identification
by: Ni, Hao, et al.
Published: (2025)

DePT: Decoupled Prompt Tuning
by: Zhang, Ji, et al.
Published: (2023)

AICL: Action In-Context Learning for Video Diffusion Model
by: Liu, Jianzhi, et al.
Published: (2024)

Prototype-based Aleatoric Uncertainty Quantification for Cross-modal Retrieval
by: Li, Hao, et al.
Published: (2023)

MDSAM:Memory-Driven Sparse Attention Matrix for LVLMs Hallucination Mitigation
by: Lu, Shuaiye, et al.
Published: (2025)

Revealing and Enhancing Core Visual Regions: Harnessing Internal Attention Dynamics for Hallucination Mitigation in LVLMs
by: Lyu, Guangtao, et al.
Published: (2026)

Mitigating Object Hallucinations in LVLMs via Attention Imbalance Rectification
by: Sun, Han, et al.
Published: (2026)

Informative Scene Graph Generation via Debiasing
by: Gao, Lianli, et al.
Published: (2023)

Reliable Few-shot Learning under Dual Noises
by: Zhang, Ji, et al.
Published: (2025)

Benchmarking Few-shot Transferability of Pre-trained Models with Improved Evaluation Protocols
by: Luo, Xu, et al.
Published: (2026)

Text-Video Retrieval with Global-Local Semantic Consistent Learning
by: Zhang, Haonan, et al.
Published: (2024)

A Closer Look at Conditional Prompt Tuning for Vision-Language Models
by: Zhang, Ji, et al.
Published: (2025)

SeMv-3D: Towards Concurrency of Semantic and Multi-view Consistency in General Text-to-3D Generation
by: Cai, Xiao, et al.
Published: (2024)

TIMI: Training-Free Image-to-3D Multi-Instance Generation with Spatial Fidelity
by: Cai, Xiao, et al.
Published: (2026)

Skip Tuning: Pre-trained Vision-Language Models are Effective and Efficient Adapters Themselves
by: Wu, Shihan, et al.
Published: (2024)

Modality Bias in LVLMs: Analyzing and Mitigating Object Hallucination via Attention Lens
by: Zheng, Haohan, et al.
Published: (2025)

Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs
by: Yu, Liu, et al.
Published: (2025)

MHSA: A Lightweight Framework for Mitigating Hallucinations via Steered Attention in LVLMs
by: Ding, Wei, et al.
Published: (2026)

SAVAA: Mitigating Hallucinations in LVLMs via Step-wise Adaptive Visual Attention Amplification
by: Zhang, Jiacheng, et al.
Published: (2026)

TARAC: Mitigating Hallucination in LVLMs via Temporal Attention Real-time Accumulative Connection
by: Jiang, Lei, et al.
Published: (2025)

Attention Hijacking: Response Manipulation Across Queries in Vision-Language Models
by: Wang, Zhiqiang, et al.
Published: (2026)

Attention-space Contrastive Guidance for Efficient Hallucination Mitigation in LVLMs
by: Jo, Yujin, et al.
Published: (2026)

VLA-Hijack: A Transferable Patch Attack against Vision-Language-Action Models via Visual Proprioception Hijacking
by: Fu, Jiyuan, et al.
Published: (2026)

Debiased Orthogonal Boundary-Driven Efficient Noise Mitigation
by: Li, Hao, et al.
Published: (2024)

Scale-Aware Pre-Training for Human-Centric Visual Perception: Enabling Lightweight and Generalizable Models
by: Wang, Xuanhan, et al.
Published: (2025)

Towards Generalized and Training-Free Text-Guided Semantic Manipulation
by: Hong, Yu, et al.
Published: (2025)

F3-Pruning: A Training-Free and Generalized Pruning Strategy towards Faster and Finer Text-to-Video Synthesis
by: Su, Sitong, et al.
Published: (2023)

When Text Hijacks Vision: Benchmarking and Mitigating Text Overlay-Induced Hallucination in Vision Language Models
by: Yakun, Cui, et al.
Published: (2026)

ProS: Prompting-to-simulate Generalized knowledge for Universal Cross-Domain Retrieval
by: Fang, Kaipeng, et al.
Published: (2023)

Seeing Clearly by Layer Two: Enhancing Attention Heads to Alleviate Hallucination in LVLMs
by: Zhang, Xiaofeng, et al.
Published: (2024)

Paying More Attention to Image: A Training-Free Method for Alleviating Hallucination in LVLMs
by: Liu, Shi, et al.
Published: (2024)

Seeing No Evil: Blinding Large Vision-Language Models to Safety Instructions via Adversarial Attention Hijacking
by: Li, Jingru, et al.
Published: (2026)

CoIN: A Benchmark of Continual Instruction tuNing for Multimodel Large Language Model
by: Chen, Cheng, et al.
Published: (2024)

From Clouds to Hallucinations: Atmospheric Retrieval Hijacking in Remote Sensing Vision-Language RAG
by: Han, Jiaju, et al.
Published: (2026)

ControlLoc: Physical-World Hijacking Attack on Visual Perception in Autonomous Driving
by: Ma, Chen, et al.
Published: (2024)