Image de couverture de livre

History Anchors: How Prior Behavior Steers LLM Decisions Toward Unsafe Actions

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteur principal:	Salgado, Alberto G. Rodríguez
Format:	Preprint
Publié:	2026
Sujets:	Artificial Intelligence Computer Vision and Pattern Recognition
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Towards Understanding Unsafe Video Generation
par: Pang, Yan, et autres
Publié: (2024)

PANC: Prior-Aware Normalized Cut via Anchor-Augmented Token Graphs
par: Gutiérrez, Juan, et autres
Publié: (2026)

Safe Vision-Language Models via Unsafe Weights Manipulation
par: D'Incà, Moreno, et autres
Publié: (2025)

Does Visual Information Play a Decisive Role in Vision-Language-Action Model Driving Behavior?
par: He, Jingtao, et autres
Publié: (2026)

HomeSafe-Bench: Evaluating Vision-Language Models on Unsafe Action Detection for Embodied Agents in Household Scenarios
par: Pu, Jiayue, et autres
Publié: (2026)

AttAnchor: Guiding Cross-Modal Token Alignment in VLMs with Attention Anchors
par: Zhang, Junyang, et autres
Publié: (2025)

Context Matters: Peer-Aware Student Behavioral Engagement Measurement via VLM Action Parsing and LLM Sequence Classification
par: Abdelkawy, Ahmed, et autres
Publié: (2026)

AnchorDiff: Training-Free Concept Grounding for MM-DiTs via Anchor-Based Graph Propagation
par: Zhang, Jian, et autres
Publié: (2026)

Diagnosing and Repairing Unsafe Channels in Vision-Language Models via Causal Discovery and Dual-Modal Safety Subspace Projection
par: Fu, Jinhu, et autres
Publié: (2026)

MedSteer: Counterfactual Endoscopic Synthesis via Training-Free Activation Steering
par: Pham, Trong-Thang, et autres
Publié: (2026)

iPay: Integrated Payment Action Recognition via Multimodal Networks and Adaptive Spatial Prior Learning
par: Huang, Kaicong, et autres
Publié: (2026)

Bi-Anchor Interpolation Solver for Accelerating Generative Modeling
par: Chen, Hongxu, et autres
Publié: (2026)

SAGE: Exploring the Boundaries of Unsafe Concept Domain with Semantic-Augment Erasing
par: Zhu, Hongguang, et autres
Publié: (2025)

DriveAction: A Benchmark for Exploring Human-like Driving Decisions in VLA Models
par: Hao, Yuhan, et autres
Publié: (2025)

RAD: Retrieval-Augmented Decision-Making of Meta-Actions with Vision-Language Models in Autonomous Driving
par: Wang, Yujin, et autres
Publié: (2025)

InstrAct: Towards Action-Centric Understanding in Instructional Videos
par: Yang, Zhuoyi, et autres
Publié: (2026)

ReGenNet: Towards Human Action-Reaction Synthesis
par: Xu, Liang, et autres
Publié: (2024)

Inline Critic Steers Image Editing
par: Kang, Weitai, et autres
Publié: (2026)

VLIPP: Towards Physically Plausible Video Generation with Vision and Language Informed Physical Prior
par: Yang, Xindi, et autres
Publié: (2025)

Beyond Fixed Anchors: Precisely Erasing Concepts with Sibling Exclusive Counterparts
par: Zhang, Tong, et autres
Publié: (2025)

Unified 4D World Action Modeling from Video Priors with Asynchronous Denoising
par: Guo, Jun, et autres
Publié: (2026)

PromptGuard: Soft Prompt-Guided Unsafe Content Moderation for Text-to-Image Models
par: Yuan, Lingzhi, et autres
Publié: (2025)

GuardTrace-VL: Detecting Unsafe Multimodel Reasoning via Iterative Safety Supervision
par: Xiang, Yuxiao, et autres
Publié: (2025)

Enhancing Action Recognition by Leveraging the Hierarchical Structure of Actions and Textual Context
par: Benavent-Lledo, Manuel, et autres
Publié: (2024)

Decoding Vision Transformers: the Diffusion Steering Lens
par: Takatsuki, Ryota, et autres
Publié: (2025)

AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
par: Wang, Zun, et autres
Publié: (2026)

ACPO: Anchor-Constrained Perceptual Optimization for Diffusion Models with No-Reference Quality Guidance
par: Yang, Yang, et autres
Publié: (2026)

PBADet: A One-Stage Anchor-Free Approach for Part-Body Association
par: Gao, Zhongpai, et autres
Publié: (2024)

Proxy-Anchor and EVT-Driven Continual Learning Method for Generalized Category Discovery
par: Fathalizadeh, Alireza, et autres
Publié: (2025)

PMG: Progressive Motion Generation via Sparse Anchor Postures Curriculum Learning
par: Xi, Yingjie, et autres
Publié: (2025)

Segmenting Visuals With Querying Words: Language Anchors For Semi-Supervised Image Segmentation
par: Nadeem, Numair, et autres
Publié: (2025)

IVAC-P2L: Leveraging Irregular Repetition Priors for Improving Video Action Counting
par: Wang, Hang, et autres
Publié: (2024)

How to Enable LLM with 3D Capacity? A Survey of Spatial Reasoning in LLM
par: Zha, Jirong, et autres
Publié: (2025)

EgoExo-Fitness: Towards Egocentric and Exocentric Full-Body Action Understanding
par: Li, Yuan-Ming, et autres
Publié: (2024)

Towards End-to-End Neuromorphic Event-based 3D Object Reconstruction Without Physical Priors
par: Xu, Chuanzhi, et autres
Publié: (2025)

EPiC: Efficient Video Camera Control Learning with Precise Anchor-Video Guidance
par: Wang, Zun, et autres
Publié: (2025)

A-MESS: Anchor based Multimodal Embedding with Semantic Synchronization for Multimodal Intent Recognition
par: Shen, Yaomin, et autres
Publié: (2025)

Learning with Instance-Dependent Noisy Labels by Anchor Hallucination and Hard Sample Label Correction
par: Huang, Po-Hsuan, et autres
Publié: (2024)

TAR-TVG: Enhancing VLMs with Timestamp Anchor-Constrained Reasoning for Temporal Video Grounding
par: Guo, Chaohong, et autres
Publié: (2025)

ContextGS: Compact 3D Gaussian Splatting with Anchor Level Context Model
par: Wang, Yufei, et autres
Publié: (2024)