Cover Image

Steering Multimodal Large Language Models Decoding for Context-Aware Safety

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Liu, Zheyuan, Xu, Zhangchen, Dou, Guangyao, Yuan, Xiangchi, Tan, Zhaoxuan, Poovendran, Radha, Jiang, Meng
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Modality-Aware Neuron Pruning for Unlearning in Multimodal Large Language Models
by: Liu, Zheyuan, et al.
Published: (2025)

MTMCS-Bench: Evaluating Contextual Safety of Multimodal Large Language Models in Multi-Turn Dialogues
by: Liu, Zheyuan, et al.
Published: (2026)

Protecting Privacy in Multimodal Large Language Models with MLLMU-Bench
by: Liu, Zheyuan, et al.
Published: (2024)

SafeDecoding: Defending against Jailbreak Attacks via Safety-Aware Decoding
by: Xu, Zhangchen, et al.
Published: (2024)

Towards Safer Large Language Models through Machine Unlearning
by: Liu, Zheyuan, et al.
Published: (2024)

Machine Unlearning in Generative AI: A Survey
by: Liu, Zheyuan, et al.
Published: (2024)

Stronger Models are NOT Stronger Teachers for Instruction Tuning
by: Xu, Zhangchen, et al.
Published: (2024)

SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities
by: Jiang, Fengqing, et al.
Published: (2025)

Personalized Pieces: Efficient Personalized Large Language Models through Collaborative Efforts
by: Tan, Zhaoxuan, et al.
Published: (2024)

KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding
by: Xu, Zhangchen, et al.
Published: (2025)

Prompt-Activation Duality: Improving Activation Steering via Attention-Level Interventions
by: Kang, Diancheng, et al.
Published: (2026)

ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs
by: Jiang, Fengqing, et al.
Published: (2024)

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing
by: Xu, Zhangchen, et al.
Published: (2024)

TinyV: Reducing False Negatives in Verification Improves RL for LLM Reasoning
by: Xu, Zhangchen, et al.
Published: (2025)

TOUCAN: Synthesizing 1.5M Tool-Agentic Data from Real-World MCP Environments
by: Xu, Zhangchen, et al.
Published: (2025)

CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models
by: Li, Yuetai, et al.
Published: (2024)

Democratizing Large Language Models via Personalized Parameter-Efficient Fine-tuning
by: Tan, Zhaoxuan, et al.
Published: (2024)

Can Large Language Models Understand Preferences in Personalized Recommendation?
by: Tan, Zhaoxuan, et al.
Published: (2025)

Avoiding Copyright Infringement via Large Language Model Unlearning
by: Dou, Guangyao, et al.
Published: (2024)

ACE: A Model Poisoning Attack on Contribution Evaluation Methods in Federated Learning
by: Xu, Zhangchen, et al.
Published: (2024)

ChatBug: A Common Vulnerability of Aligned LLMs Induced by Chat Templates
by: Jiang, Fengqing, et al.
Published: (2024)

Investigating the Feasibility of Mitigating Potential Copyright Infringement via Large Language Model Unlearning
by: Dou, Guangyao
Published: (2024)

MMDT: Decoding the Trustworthiness and Safety of Multimodal Foundation Models
by: Xu, Chejian, et al.
Published: (2025)

LaCache: Ladder-Shaped KV Caching for Efficient Long-Context Modeling of Large Language Models
by: Shi, Dachuan, et al.
Published: (2025)

Defending Large Language Models Against Jailbreak Attacks via In-Decoding Safety-Awareness Probing
by: Zhao, Yinzhi, et al.
Published: (2026)

Knowledge Crosswords: Geometric Knowledge Reasoning with Large Language Models
by: Ding, Wenxuan, et al.
Published: (2023)

Automating Steering for Safe Multimodal Large Language Models
by: Wu, Lyucheng, et al.
Published: (2025)

Steering When Necessary: Flexible Steering Large Language Models with Backtracking
by: Cheng, Zifeng, et al.
Published: (2025)

OpenDecoder: Open Large Language Model Decoding to Incorporate Document Quality in RAG
by: Mo, Fengran, et al.
Published: (2026)

Towards Inference-time Category-wise Safety Steering for Large Language Models
by: Bhattacharjee, Amrita, et al.
Published: (2024)

COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamics
by: Sharma, Kartik, et al.
Published: (2026)

Small Models Struggle to Learn from Strong Reasoners
by: Li, Yuetai, et al.
Published: (2025)

LongSafety: Evaluating Long-Context Safety of Large Language Models
by: Lu, Yida, et al.
Published: (2025)

ASRU: Activation Steering Meets Reinforcement Unlearning for Multimodal Large Language Models
by: Guang, Jiahui, et al.
Published: (2026)

On Speculative Decoding for Multimodal Large Language Models
by: Gagrani, Mukul, et al.
Published: (2024)

Dual-Space Smoothness for Robust and Balanced LLM Unlearning
by: Yan, Han, et al.
Published: (2025)

ContextFocus: Activation Steering for Contextual Faithfulness in Large Language Models
by: Anand, Nikhil, et al.
Published: (2026)

Probing and Steering Evaluation Awareness of Language Models
by: Nguyen, Jord, et al.
Published: (2025)

VisualSphinx: Large-Scale Synthetic Vision Logic Puzzles for RL
by: Feng, Yichen, et al.
Published: (2025)

When Language Overrules: Revealing Text Dominance in Multimodal Large Language Models
by: Wu, Huyu, et al.
Published: (2025)