Cover Image

The Safety Reminder: A Soft Prompt to Reactivate Delayed Safety Awareness in Vision-Language Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Tang, Peiyuan, Xin, Haojie, Zhang, Xiaodong, Sun, Jun, Xia, Qin, Yang, Zijiang
Format:	Preprint
Published:	2025
Subjects:	Artificial Intelligence Computation and Language Cryptography and Security Computer Vision and Pattern Recognition Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Models
by: Ding, Yi, et al.
Published: (2025)

DAVSP: Safety Alignment for Large Vision-Language Models via Deep Aligned Visual Safety Prompt
by: Zhang, Yitong, et al.
Published: (2025)

Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety
by: Ma, Xingjun, et al.
Published: (2025)

Few-Shot Adversarial Prompt Learning on Vision-Language Models
by: Zhou, Yiwei, et al.
Published: (2024)

Evolving Contextual Safety in Multi-Modal Large Language Models via Inference-Time Self-Reflective Memory
by: Zhang, Ce, et al.
Published: (2026)

MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance
by: Pi, Renjie, et al.
Published: (2024)

PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of Visibility
by: Shahariar, G M, et al.
Published: (2026)

GradSafe: Detecting Jailbreak Prompts for LLMs via Safety-Critical Gradient Analysis
by: Xie, Yueqi, et al.
Published: (2024)

Self-Jailbreaking: Language Models Can Reason Themselves Out of Safety Alignment After Benign Reasoning Training
by: Yong, Zheng-Xin, et al.
Published: (2025)

Self-adaptive Dataset Construction for Real-World Multimodal Safety Scenarios
by: Qu, Jingen, et al.
Published: (2025)

Do Prompts Guarantee Safety? Mitigating Toxicity from LLM Generations through Subspace Intervention
by: Singh, Himanshu, et al.
Published: (2026)

Image-Based Geolocation Using Large Vision-Language Models
by: Liu, Yi, et al.
Published: (2024)

Contextual Image Attack: How Visual Context Exposes Multimodal Safety Vulnerabilities
by: Xiong, Yuan, et al.
Published: (2025)

DeepSight: An All-in-One LM Safety Toolkit
by: Zhang, Bo, et al.
Published: (2026)

BSPA: Exploring Black-box Stealthy Prompt Attacks against Image Generators
by: Tian, Yu, et al.
Published: (2024)

Safety is Not Only About Refusal: Reasoning-Enhanced Fine-tuning for Interpretable LLM Safety
by: Zhang, Yuyou, et al.
Published: (2025)

S-Eval: Towards Automated and Comprehensive Safety Evaluation for Large Language Models
by: Yuan, Xiaohan, et al.
Published: (2024)

SurrogatePrompt: Bypassing the Safety Filter of Text-to-Image Models via Substitution
by: Ba, Zhongjie, et al.
Published: (2023)

Effective and Efficient Adversarial Detection for Vision-Language Models via A Single Vector
by: Huang, Youcheng, et al.
Published: (2024)

ReasAlign: Reasoning Enhanced Safety Alignment against Prompt Injection Attack
by: Li, Hao, et al.
Published: (2026)

Understanding and Mitigating Over-refusal for Large Language Models via Safety Representation
by: Zhang, Junbo, et al.
Published: (2025)

RAG Safety: Exploring Knowledge Poisoning Attacks to Retrieval-Augmented Generation
by: Zhao, Tianzhe, et al.
Published: (2025)

Analysing the Safety Pitfalls of Steering Vectors
by: Li, Yuxiao, et al.
Published: (2026)

VLSBench: Unveiling Visual Leakage in Multimodal Safety
by: Hu, Xuhao, et al.
Published: (2024)

SOS! Soft Prompt Attack Against Open-Source Large Language Models
by: Yang, Ziqing, et al.
Published: (2024)

EnchTable: Unified Safety Alignment Transfer in Fine-tuned Large Language Models
by: Wu, Jialin, et al.
Published: (2025)

SGuard-v1: Safety Guardrail for Large Language Models
by: Lee, JoonHo, et al.
Published: (2025)

Overriding Safety protections of Open-source Models
by: Kumar, Sachin
Published: (2024)

ML-Bench&Guard: Policy-Grounded Multilingual Safety Benchmark and Guardrail for Large Language Models
by: Zhao, Yunhan, et al.
Published: (2026)

A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluations
by: Ye, Mang, et al.
Published: (2025)

VLM-Guard: Safeguarding Vision-Language Models via Fulfilling Safety Alignment Gap
by: Liu, Qin, et al.
Published: (2025)

Jailbreak Antidote: Runtime Safety-Utility Balance via Sparse Representation Adjustment in Large Language Models
by: Shen, Guobin, et al.
Published: (2024)

One Trigger Token Is Enough: A Defense Strategy for Balancing Safety and Usability in Large Language Models
by: Gu, Haoran, et al.
Published: (2025)

Bag of Tricks for Subverting Reasoning-based Safety Guardrails
by: Chen, Shuo, et al.
Published: (2025)

Prompt Stealing Attacks Against Large Language Models
by: Sha, Zeyang, et al.
Published: (2024)

Lifelong Safety Alignment for Language Models
by: Wang, Haoyu, et al.
Published: (2025)

Representation Bending for Large Language Model Safety
by: Yousefpour, Ashkan, et al.
Published: (2025)

SafeReview: Defending LLM-based Review Systems Against Adversarial Hidden Prompts
by: Xin, Yuan, et al.
Published: (2026)

Root Defence Strategies: Ensuring Safety of LLM at the Decoding Level
by: Zeng, Xinyi, et al.
Published: (2024)

Cross-Task Defense: Instruction-Tuning LLMs for Content Safety
by: Fu, Yu, et al.
Published: (2024)