SmartSight: Mitigating Hallucination in Video-LLMs Without Compromising Video Understanding via Temporal Attention Collapse

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Sun, Yiming, Zhang, Mi, Li, Feifei, Hong, Geng, Yang, Min
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

Mitigating Hallucination in VideoLLMs via Temporal-Aware Activation Engineering
by: Cai, Jianfeng, et al.
Published: (2025)

Hallucination Mitigation Prompts Long-term Video Understanding
by: Sun, Yiwei, et al.
Published: (2024)

Detect-and-Guide: Self-regulation of Diffusion Models for Safe Text-to-Image Generation via Guideline Token Optimization
by: Li, Feifei, et al.
Published: (2025)

MOOSE: Pay Attention to Temporal Dynamics for Video Understanding via Optical Flows
by: Nguyen, Hong, et al.
Published: (2025)

Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders
by: Rasekh, Ali, et al.
Published: (2025)

MASH-VLM: Mitigating Action-Scene Hallucination in Video-LLMs through Disentangled Spatial-Temporal Representations
by: Bae, Kyungho, et al.
Published: (2025)

Broken Memories: Detecting and Mitigating Memorization in Diffusion Models with Degraded Generations
by: Huang, Yuanmin, et al.
Published: (2026)

Tracing and Mitigating Hallucinations in Multimodal LLMs via Dynamic Attention Localization
by: Yang, Tiancheng, et al.
Published: (2025)

VideoHallu: Evaluating and Mitigating Multi-modal Hallucinations on Synthetic Video Understanding
by: Li, Zongxia, et al.
Published: (2025)

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

Exploring Hallucination of Large Multimodal Models in Video Understanding: Benchmark, Analysis and Mitigation
by: Gao, Hongcheng, et al.
Published: (2025)

Temporal2Seq: A Unified Framework for Temporal Video Understanding Tasks
by: Yang, Min, et al.
Published: (2024)

VideoINSTA: Zero-shot Long Video Understanding via Informative Spatial-Temporal Reasoning with LLMs
by: Liao, Ruotong, et al.
Published: (2024)

PaMi-VDPO: Mitigating Video Hallucinations by Prompt-Aware Multi-Instance Video Preference Learning
by: Ding, Xinpeng, et al.
Published: (2025)

Harnessing Synthetic Preference Data for Enhancing Temporal Understanding of Video-LLMs
by: Vani, Sameep, et al.
Published: (2025)

Mitigating Semantic Collapse in Partially Relevant Video Retrieval
by: Moon, WonJun, et al.
Published: (2025)

Adaptive Video Distillation: Mitigating Oversaturation and Temporal Collapse in Few-Step Generation
by: You, Yuyang, et al.
Published: (2026)

VidHalluc: Evaluating Temporal Hallucinations in Multimodal Large Language Models for Video Understanding
by: Li, Chaoyu, et al.
Published: (2024)

SafeRoPE: Risk-specific Head-wise Embedding Rotation for Safe Generation in Rectified Flow Transformers
by: Yang, Xiang, et al.
Published: (2026)

EventHallusion: Diagnosing Event Hallucinations in Video LLMs
by: Zhang, Jiacheng, et al.
Published: (2024)

Predictive Temporal Attention on Event-based Video Stream for Energy-efficient Situation Awareness
by: Bu, Yiming, et al.
Published: (2024)

ARGUS: Hallucination and Omission Evaluation in Video-LLMs
by: Rawal, Ruchit, et al.
Published: (2025)

LLaVA-MLB: Mitigating and Leveraging Attention Bias for Training-Free Video LLMs
by: Shen, Leqi, et al.
Published: (2025)

CounterVid: Counterfactual Video Generation for Mitigating Action and Temporal Hallucinations in Video-Language Models
by: Poppi, Tobia, et al.
Published: (2026)

TemporalVLM: Video LLMs for Temporal Reasoning in Long Videos
by: Fateh, Fawad Javed, et al.
Published: (2024)

Taming Hallucinations: Boosting MLLMs' Video Understanding via Counterfactual Video Generation
by: Huang, Zhe, et al.
Published: (2025)

TempCompass: Do Video LLMs Really Understand Videos?
by: Liu, Yuanxin, et al.
Published: (2024)

Motion-Adaptive Temporal Attention for Lightweight Video Generation with Stable Diffusion
by: Hong, Rui, et al.
Published: (2026)

INFACT: A Diagnostic Benchmark for Induced Faithfulness and Factuality Hallucinations in Video-LLMs
by: Yang, Junqi, et al.
Published: (2026)

Zoom-Zero: Reinforced Coarse-to-Fine Video Understanding via Temporal Zoom-in
by: Shen, Xiaoqian, et al.
Published: (2025)

VideoCompressa: Data-Efficient Video Understanding via Joint Temporal Compression and Spatial Reconstruction
by: Wang, Shaobo, et al.
Published: (2025)

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

MESH -- Understanding Videos Like Human: Measuring Hallucinations in Large Video Models
by: Yang, Garry, et al.
Published: (2025)

Benchmarking the Trustworthiness in Multimodal LLMs for Video Understanding
by: Wang, Youze, et al.
Published: (2025)

LLMs Can Check Their Own Results to Mitigate Hallucinations in Traffic Understanding Tasks
by: Dona, Malsha Ashani Mahawatta, et al.
Published: (2024)

VideoExpert: Augmented LLM for Temporal-Sensitive Video Understanding
by: Zhao, Henghao, et al.
Published: (2025)

V-CORE: Temporally Consistent Video Understanding for Video-LLM
by: Kang, Zhengjian, et al.
Published: (2026)

VideoCoF: Unified Video Editing with Temporal Reasoner
by: Yang, Xiangpeng, et al.
Published: (2025)

MotionSight: Boosting Fine-Grained Motion Understanding in Multimodal LLMs
by: Du, Yipeng, et al.
Published: (2025)

EGOILLUSION: Benchmarking Hallucinations in Egocentric Video Understanding
by: Seth, Ashish, et al.
Published: (2025)