Copertina

Two Birds, One Projection: Harmonizing Safety and Utility in LVLMs via Inference-time Feature Projection

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Han, Yewon, Seol, Yumin, Kong, EunGyung, Jo, Minsoo, Kim, Taesup
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

Angular Gradient Sign Method: Uncovering Vulnerabilities in Hyperbolic Networks
di: Jo, Minsoo, et al.
Pubblicazione: (2025)

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

Contrastive Residual Energy Test-time Adaptation
di: Han, Yewon, et al.
Pubblicazione: (2025)

Generalized and Personalized Federated Learning with Black-Box Foundation Models via Orthogonal Transformations
di: Kong, Eun Gyung, et al.
Pubblicazione: (2025)

Retaining and Enhancing Pre-trained Knowledge in Vision-Language Models with Prompt Ensembling
di: Kim, Donggeun, et al.
Pubblicazione: (2024)

Missing Modality Prediction for Unpaired Multimodal Learning via Joint Embedding of Unimodal Models
di: Kim, Donggeun, et al.
Pubblicazione: (2024)

Improving Visual Token Reduction via Rectifying Distortions for Efficient Multimodal LLM Inference
di: Cho, Hyeonwoo, et al.
Pubblicazione: (2026)

CL3DOR: Contrastive Learning for 3D Large Multimodal Models via Odds Ratio on High-Resolution Point Clouds
di: Kim, Keonwoo, et al.
Pubblicazione: (2025)

Diffusion-Based Conditional Image Editing through Optimized Inference with Guidance
di: Lee, Hyunsoo, et al.
Pubblicazione: (2024)

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

Semantic Anchoring for Robust Personalization in Text-to-Image Diffusion Models
di: Yang, Seoyun, et al.
Pubblicazione: (2025)

Preserve and Personalize: Personalized Text-to-Image Diffusion Models without Distributional Drift
di: Kim, Gihoon, et al.
Pubblicazione: (2025)

Diffusion-Based Image-to-Image Translation by Noise Correction via Prompt Interpolation
di: Lee, Junsung, et al.
Pubblicazione: (2024)

Projecting Gaussian Ellipsoids While Avoiding Affine Projection Approximation
di: Qi, Han, et al.
Pubblicazione: (2024)

PiCa: Parameter-Efficient Fine-Tuning with Column Space Projection
di: Hwang, Junseo, et al.
Pubblicazione: (2025)

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

Incremental Learning with Repetition via Pseudo-Feature Projection
di: Tscheschner, Benedikt, et al.
Pubblicazione: (2025)

Patch-Level Kernel Alignment for Dense Self-Supervised Learning
di: Yeo, Juan, et al.
Pubblicazione: (2025)

MPA-DNN: Projection-Aware Unsupervised Learning for Multi-period DC-OPF
di: Kim, Yeomoon, et al.
Pubblicazione: (2025)

Q-Align: Alleviating Attention Leakage in Zero-Shot Appearance Transfer via Query-Query Alignment
di: Kim, Namu, et al.
Pubblicazione: (2025)

Memory-Free Continual Learning with Null Space Adaptation for Zero-Shot Vision-Language Models
di: Jo, Yujin, et al.
Pubblicazione: (2025)

MAFA: Managing False Negatives for Vision-Language Pre-training
di: Byun, Jaeseok, et al.
Pubblicazione: (2023)

Steering LVLMs via Sparse Autoencoder for Hallucination Mitigation
di: Hua, Zhenglin, et al.
Pubblicazione: (2025)

V-Attack: Targeting Disentangled Value Features for Controllable Adversarial Attacks on LVLMs
di: Nie, Sen, et al.
Pubblicazione: (2025)

Sparse Imagination for Efficient Visual World Model Planning
di: Chun, Junha, et al.
Pubblicazione: (2025)

PosterLlama: Bridging Design Ability of Langauge Model to Contents-Aware Layout Generation
di: Seol, Jaejung, et al.
Pubblicazione: (2024)

Feature Projection Learning for Better Vision-Language Reasoning
di: Zhang, Yi, et al.
Pubblicazione: (2026)

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

Overcoming Data Inequality across Domains with Semi-Supervised Domain Generalization
di: Park, Jinha, et al.
Pubblicazione: (2024)

HarmoniCa: Harmonizing Training and Inference for Better Feature Caching in Diffusion Transformer Acceleration
di: Huang, Yushi, et al.
Pubblicazione: (2024)

DoMIX: An Efficient Framework for Exploiting Domain Knowledge in Fine-Tuning
di: Kim, Dohoon, et al.
Pubblicazione: (2025)

EDM: Equirectangular Projection-Oriented Dense Kernelized Feature Matching
di: Jung, Dongki, et al.
Pubblicazione: (2025)

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

Catch-Up Mix: Catch-Up Class for Struggling Filters in CNN
di: Kang, Minsoo, et al.
Pubblicazione: (2024)

Towards Rationale-Answer Alignment of LVLMs via Self-Rationale Calibration
di: Wu, Yuanchen, et al.
Pubblicazione: (2025)

Interpretable Dimensionality Reduction by Feature Preserving Manifold Approximation and Projection
di: Yang, Yang, et al.
Pubblicazione: (2022)

One Perturbation, Two Failure Modes: Probing VLM Safety via Embedding-Guided Typographic Perturbations
di: Balakrishnan, Ravikumar, et al.
Pubblicazione: (2026)

An Integrated Causal Inference Framework for Traffic Safety Modeling with Semantic Street-View Visual Features
di: Sun, Lishan, et al.
Pubblicazione: (2026)

Remember Me: Bridging the Long-Range Gap in LVLMs with Three-Step Inference-Only Decay Resilience Strategies
di: Gao, Peng, et al.
Pubblicazione: (2025)

Object-Centric World Model for Language-Guided Manipulation
di: Jeong, Youngjoon, et al.
Pubblicazione: (2025)