Copertina

Beyond the Last Layer: Multi-Layer Representation Fusion for Visual Tokenization

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Zhu, Xuanyu, Bai, Yan, Shi, Yang, Lou, Yihang, Zhang, Yuanxing, Jin, Jing, Zhou, Yuan
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Computer Vision and Pattern Recognition Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

MMVIAD: Multi-view Multi-task Video Understanding for Industrial Anomaly Detection
di: Zhao, Xiran, et al.
Pubblicazione: (2026)

Famba-V: Fast Vision Mamba with Cross-Layer Token Fusion
di: Shen, Hui, et al.
Pubblicazione: (2024)

Unveiling Fine-Grained Visual Traces: Evaluating Multimodal Interleaved Reasoning Chains in Multimodal STEM Tasks
di: Jin, Jing, et al.
Pubblicazione: (2026)

Monet: Reasoning in Latent Visual Space Beyond Images and Language
di: Wang, Qixun, et al.
Pubblicazione: (2025)

Tilt and Average : Geometric Adjustment of the Last Layer for Recalibration
di: Cho, Gyusang, et al.
Pubblicazione: (2024)

Beyond Static Cropping: Layer-Adaptive Visual Localization and Decoding Enhancement
di: Zhu, Zipeng, et al.
Pubblicazione: (2026)

Text-Guided Layer Fusion Mitigates Hallucination in Multimodal LLMs
di: Lin, Chenchen, et al.
Pubblicazione: (2026)

Prototype Fusion: A Training-Free Multi-Layer Approach to OOD Detection
di: Gul, Shreen, et al.
Pubblicazione: (2026)

CLASP: Class-Adaptive Layer Fusion and Dual-Stage Pruning for Multimodal Large Language Models
di: Dang, Yunkai, et al.
Pubblicazione: (2026)

One Token Per Frame: Reconsidering Visual Bandwidth in World Models for VLA Policy
di: Tang, Zuojin, et al.
Pubblicazione: (2026)

One Layer Is Enough: Adapting Pretrained Visual Encoders for Image Generation
di: Gao, Yuan, et al.
Pubblicazione: (2025)

Diagnosing and Repairing Unsafe Channels in Vision-Language Models via Causal Discovery and Dual-Modal Safety Subspace Projection
di: Fu, Jinhu, et al.
Pubblicazione: (2026)

Multi-Layer Visual Feature Fusion in Multimodal LLMs: Methods, Analysis, and Best Practices
di: Lin, Junyan, et al.
Pubblicazione: (2025)

Evading Visual Aphasia: Contrastive Adaptive Semantic Token Pruning for Vision-Language Models
di: Ma, Jie, et al.
Pubblicazione: (2026)

Layer- and Timestep-Adaptive Differentiable Token Compression Ratios for Efficient Diffusion Transformers
di: You, Haoran, et al.
Pubblicazione: (2024)

Tracing Representation Progression: Analyzing and Enhancing Layer-Wise Similarity
di: Jiang, Jiachen, et al.
Pubblicazione: (2024)

Integrative CAM: Adaptive Layer Fusion for Comprehensive Interpretation of CNNs
di: Singh, Aniket K., et al.
Pubblicazione: (2024)

Multi-Granularity Vision Fastformer with Fusion Mechanism for Skin Lesion Segmentation
di: Liu, Xuanyu, et al.
Pubblicazione: (2025)

COXNet: Cross-Layer Fusion with Adaptive Alignment and Scale Integration for RGBT Tiny Object Detection
di: Peng, Peiran, et al.
Pubblicazione: (2025)

SwiftVLM: Efficient Vision-Language Model Inference via Cross-Layer Token Bypass
di: Qian, Chen, et al.
Pubblicazione: (2026)

Representations of Text and Images Align From Layer One
di: Wybitul, Evžen, et al.
Pubblicazione: (2026)

Beyond Static Visual Tokens: Structured Sequential Visual Chain-of-Thought Reasoning
di: Guo, Guangfu, et al.
Pubblicazione: (2026)

Dynamic Multi-Target Fusion for Efficient Audio-Visual Navigation
di: Yu, Yinfeng, et al.
Pubblicazione: (2025)

SeqVLM: Proposal-Guided Multi-View Sequences Reasoning via VLM for Zero-Shot 3D Visual Grounding
di: Lin, Jiawen, et al.
Pubblicazione: (2025)

CoViPAL: Layer-wise Contextualized Visual Token Pruning for Large Vision-Language Models
di: Tang, Zicong, et al.
Pubblicazione: (2025)

An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models
di: Chen, Liang, et al.
Pubblicazione: (2024)

EvoCut: Multi-Layer Evolution-Aware Visual Token Compression for Efficient Large Vision-Language Models
di: Lu, Hongyu, et al.
Pubblicazione: (2026)

Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs
di: Zhang, Qizhe, et al.
Pubblicazione: (2024)

LayerFusion: Harmonized Multi-Layer Text-to-Image Generation with Generative Priors
di: Dalva, Yusuf, et al.
Pubblicazione: (2024)

Beyond BEV: Optimizing Point-Level Tokens for Collaborative Perception
di: Li, Yang, et al.
Pubblicazione: (2025)

MimicNorm: Weight Mean and Last BN Layer Mimic the Dynamic of Batch Normalization
di: Fei, Wen, et al.
Pubblicazione: (2020)

Uncertainty-Encoded Multi-Modal Fusion for Robust Object Detection in Autonomous Driving
di: Lou, Yang, et al.
Pubblicazione: (2023)

Variational Bayesian Last Layers
di: Harrison, James, et al.
Pubblicazione: (2024)

What Kind of Visual Tokens Do We Need? Training-free Visual Token Pruning for Multi-modal Large Language Models from the Perspective of Graph
di: Jiang, Yutao, et al.
Pubblicazione: (2025)

LLaVA-SP: Enhancing Visual Representation with Visual Spatial Tokens for MLLMs
di: Lou, Haoran, et al.
Pubblicazione: (2025)

Lightweight Multi-Scale Feature Extraction with Fully Connected LMF Layer for Salient Object Detection
di: Shi, Yunpeng, et al.
Pubblicazione: (2025)

V2Flow: Unifying Visual Tokenization and Large Language Model Vocabularies for Autoregressive Image Generation
di: Zhang, Guiwei, et al.
Pubblicazione: (2025)

FlattenGPT: Depth Compression for Transformer with Layer Flattening
di: Xu, Ruihan, et al.
Pubblicazione: (2026)

AvatarShield: Visual Reinforcement Learning for Human-Centric Synthetic Video Detection
di: Xu, Zhipei, et al.
Pubblicazione: (2025)

Disentangled Clothed Avatar Generation with Layered Representation
di: Zhang, Weitian, et al.
Pubblicazione: (2025)