Image de couverture de livre

When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Khayatan, Pegah, Parekh, Jayneel, Dapogny, Arnaud, Shukor, Mustafa, Newson, Alasdair, Cord, Matthieu
Format:	Preprint
Publié:	2026
Sujets:	Computer Vision and Pattern Recognition Artificial Intelligence Computation and Language Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Learning to Steer: Input-dependent Steering for Multimodal LLMs
par: Parekh, Jayneel, et autres
Publié: (2025)

Analyzing Finetuning Representation Shift for Multimodal LLMs Steering
par: Khayatan, Pegah, et autres
Publié: (2025)

A Concept-Based Explainability Framework for Large Multimodal Models
par: Parekh, Jayneel, et autres
Publié: (2024)

Implicit Multimodal Alignment: On the Generalization of Frozen LLMs to Multimodal Inputs
par: Shukor, Mustafa, et autres
Publié: (2024)

Restyling Unsupervised Concept Based Interpretable Networks with Generative Models
par: Parekh, Jayneel, et autres
Publié: (2024)

Skipping Computations in Multimodal LLMs
par: Shukor, Mustafa, et autres
Publié: (2024)

What Makes Multimodal In-Context Learning Work?
par: Baldassini, Folco Bertini, et autres
Publié: (2024)

Improved Baselines for Data-efficient Perceptual Augmentation of LLMs
par: Vallaeys, Théophane, et autres
Publié: (2024)

ReGentS: Real-World Safety-Critical Driving Scenario Generation Made Stable
par: Yin, Yuan, et autres
Publié: (2024)

Mitigating Hallucinations in Large Vision-Language Models (LVLMs) via Language-Contrastive Decoding (LCD)
par: Manevich, Avshalom, et autres
Publié: (2024)

Visual Description Grounding Reduces Hallucinations and Boosts Reasoning in LVLMs
par: Ghosh, Sreyan, et autres
Publié: (2024)

Beyond Task Performance: Evaluating and Reducing the Flaws of Large Multimodal Models with In-Context Learning
par: Shukor, Mustafa, et autres
Publié: (2023)

Cracking the Code of Hallucination in LVLMs with Vision-aware Head Divergence
par: He, Jinghan, et autres
Publié: (2024)

Black-Box Visual Prompt Engineering for Mitigating Object Hallucination in Large Vision Language Models
par: Woo, Sangmin, et autres
Publié: (2025)

GLSim: Detecting Object Hallucinations in LVLMs via Global-Local Similarity
par: Park, Seongheon, et autres
Publié: (2025)

Why LVLMs Are More Prone to Hallucinations in Longer Responses: The Role of Context
par: Zheng, Ge, et autres
Publié: (2025)

Steering LVLMs via Sparse Autoencoder for Hallucination Mitigation
par: Hua, Zhenglin, et autres
Publié: (2025)

Mitigating Hallucinations in Multimodal Spatial Relations through Constraint-Aware Prompting
par: Wu, Jiarui, et autres
Publié: (2025)

HELPD: Mitigating Hallucination of LVLMs by Hierarchical Feedback Learning with Vision-enhanced Penalty Decoding
par: Yuan, Fan, et autres
Publié: (2024)

When RAG Hurts: Diagnosing and Mitigating Attention Distraction in Retrieval-Augmented LVLMs
par: Zhao, Beidi, et autres
Publié: (2026)

FrescoDiffusion: 4K Image-to-Video with Prior-Regularized Tiled Diffusion
par: Caselles-Dupré, Hugo, et autres
Publié: (2026)

DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut
par: Couairon, Paul, et autres
Publié: (2024)

Beyond Hallucinations: Enhancing LVLMs through Hallucination-Aware Direct Preference Optimization
par: Zhao, Zhiyuan, et autres
Publié: (2023)

Magnifier Prompt: Tackling Multimodal Hallucination via Extremely Simple Instructions
par: Fu, Yuhan, et autres
Publié: (2024)

Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs
par: Fang, Hao, et autres
Publié: (2025)

Historical Test-time Prompt Tuning for Vision Foundation Models
par: Zhang, Jingyi, et autres
Publié: (2024)

Bayesian Principles Improve Prompt Learning In Vision-Language Models
par: Kim, Mingyu, et autres
Publié: (2025)

When Seeing Overrides Knowing: Disentangling Knowledge Conflicts in Vision-Language Models
par: Ortu, Francesco, et autres
Publié: (2025)

Mechanisms of Prompt-Induced Hallucination in Vision-Language Models
par: Rudman, William, et autres
Publié: (2026)

Why Only Text: Empowering Vision-and-Language Navigation with Multi-modal Prompts
par: Hong, Haodong, et autres
Publié: (2024)

Modality Bias in LVLMs: Analyzing and Mitigating Object Hallucination via Attention Lens
par: Zheng, Haohan, et autres
Publié: (2025)

Zero-Shot Refinement of Buildings' Segmentation Models using SAM
par: Mayladan, Ali, et autres
Publié: (2023)

FreeSeg-Diff: Training-Free Open-Vocabulary Segmentation with Diffusion Models
par: Corradini, Barbara Toniella, et autres
Publié: (2024)

Differentiable Prompt Learning for Vision Language Models
par: Huang, Zhenhan, et autres
Publié: (2024)

Musketeer: Joint Training for Multi-task Vision Language Model with Task Explanation Prompts
par: Zhang, Zhaoyang, et autres
Publié: (2023)

Towards Generalizable Trajectory Prediction Using Dual-Level Representation Learning And Adaptive Prompting
par: Messaoud, Kaouther, et autres
Publié: (2025)

Prompt-Induced Score Variance in Zero-Shot Binary Vision-Language Safety Classification
par: Weng, Charles, et autres
Publié: (2026)

Multi-Object Hallucination in Vision-Language Models
par: Chen, Xuweiyi, et autres
Publié: (2024)

Mitigating Multilingual Hallucination in Large Vision-Language Models
par: Qu, Xiaoye, et autres
Publié: (2024)

Benchmarking Deflection and Hallucination in Large Vision-Language Models
par: Moratelli, Nicholas, et autres
Publié: (2026)