Cover Image

Steering LVLMs via Sparse Autoencoder for Hallucination Mitigation

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Hua, Zhenglin, He, Jinghan, Yao, Zijun, Han, Tianxu, Guo, Haiyun, Jia, Yuheng, Fang, Junfeng
Format:	Preprint
Published:	2025
Subjects:	Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Cracking the Code of Hallucination in LVLMs with Vision-aware Head Divergence
by: He, Jinghan, et al.
Published: (2024)

Finding the Correct Visual Evidence Without Forgetting: Mitigating Hallucination in LVLMs via Inter-Layer Visual Attention Discrepancy
by: Xie, Yutong, et al.
Published: (2026)

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

Towards Interpretable Hallucination Analysis and Mitigation in LVLMs via Contrastive Neuron Steering
by: Lyu, Guangtao, et al.
Published: (2026)

Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimization
by: Zhao, Zhiyuan, et al.
Published: (2023)

HELPD: Mitigating Hallucination of LVLMs by Hierarchical Feedback Learning with Vision-enhanced Penalty Decoding
by: Yuan, Fan, et al.
Published: (2024)

Mitigating Hallucinations in Large Vision-Language Models (LVLMs) via Language-Contrastive Decoding (LCD)
by: Manevich, Avshalom, et al.
Published: (2024)

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

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

GLSim: Detecting Object Hallucinations in LVLMs via Global-Local Similarity
by: Park, Seongheon, et al.
Published: (2025)

Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs
by: Fang, Hao, et al.
Published: (2025)

VEGAS: Mitigating Hallucinations in Large Vision-Language Models via Vision-Encoder Attention Guided Adaptive Steering
by: Wang, Zihu, et al.
Published: (2025)

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

Active Zero: Self-Evolving Vision-Language Models through Active Environment Exploration
by: He, Jinghan, et al.
Published: (2026)

HIME: Mitigating Object Hallucinations in LVLMs via Hallucination Insensitivity Model Editing
by: Akl, Ahmed, et al.
Published: (2026)

Visual Description Grounding Reduces Hallucinations and Boosts Reasoning in LVLMs
by: Ghosh, Sreyan, et al.
Published: (2024)

Why LVLMs Are More Prone to Hallucinations in Longer Responses: The Role of Context
by: Zheng, Ge, et al.
Published: (2025)

When RAG Hurts: Diagnosing and Mitigating Attention Distraction in Retrieval-Augmented LVLMs
by: Zhao, Beidi, et al.
Published: (2026)

Revealing Perception and Generation Dynamics in LVLMs: Mitigating Hallucinations via Validated Dominance Correction
by: Lyu, Guangtao, et al.
Published: (2025)

Mitigating Object Hallucination via Concentric Causal Attention
by: Xing, Yun, et al.
Published: (2024)

Improving Alignment in LVLMs with Debiased Self-Judgment
by: Yang, Sihan, et al.
Published: (2025)

Analyzing and Mitigating Object Hallucination: A Training Bias Perspective
by: Li, Yifan, et al.
Published: (2025)

Systematic Reward Gap Optimization for Mitigating VLM Hallucinations
by: He, Lehan, et al.
Published: (2024)

Make LVLMs Focus: Context-Aware Attention Modulation for Better Multimodal In-Context Learning
by: Li, Yanshu, et al.
Published: (2025)

When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs
by: Khayatan, Pegah, et al.
Published: (2026)

Mitigating Multimodal Hallucination via Phase-wise Self-reward
by: Zhang, Yu, et al.
Published: (2026)

Mitigating Multimodal Hallucinations via Gradient-based Self-Reflection
by: Wang, Shan, et al.
Published: (2025)

Mitigating Object Hallucination via Robust Local Perception Search
by: Gao, Zixian, et al.
Published: (2025)

Mitigating Action-Relation Hallucinations in LVLMs via Relation-aware Visual Enhancement
by: Qin, Zhenxin, et al.
Published: (2026)

Decoding by Perturbation: Mitigating MLLM Hallucinations via Dynamic Textual Perturbation
by: Jia, Sihang, et al.
Published: (2026)

Mitigating Hallucinations in Large Vision-Language Models by Self-Injecting Hallucinations
by: Lu, Yifan, et al.
Published: (2025)

Video-SafetyBench: A Benchmark for Safety Evaluation of Video LVLMs
by: Liu, Xuannan, et al.
Published: (2025)

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

Do Vision Encoders Truly Explain Object Hallucination?: Mitigating Object Hallucination via Simple Fine-Grained CLIPScore
by: Oh, Hongseok, et al.
Published: (2025)

Optimizing LVLMs with On-Policy Data for Effective Hallucination Mitigation
by: Yu, Chengzhi, et al.
Published: (2025)

See Different, Think Better: Visual Variations Mitigating Hallucinations in LVLMs
by: Dai, Ziyun, et al.
Published: (2025)

Attention Hijackers: Detect and Disentangle Attention Hijacking in LVLMs for Hallucination Mitigation
by: Chen, Beitao, et al.
Published: (2025)

TARAC: Mitigating Hallucination in LVLMs via Temporal Attention Real-time Accumulative Connection
by: Jiang, Lei, 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)

Steering the Verifiability of Multimodal AI Hallucinations
by: Pang, Jianhong, et al.
Published: (2026)