Automating Steering for Safe Multimodal Large Language Models

• URL: http://arxiv.org/abs/2507.13255v1
• Date: Thu, 17 Jul 2025 16:04:55 GMT
• Title: Automating Steering for Safe Multimodal Large Language Models
• Authors: Lyucheng Wu, Mengru Wang, Ziwen Xu, Tri Cao, Nay Oo, Bryan Hooi, Shumin Deng,
• Abstract summary: We introduce a modular and adaptive inference-time intervention technology, AutoSteer, without requiring any fine-tuning of the underlying model.<n>AutoSteer incorporates three core components: (1) a novel Safety Awareness Score (SAS) that automatically identifies the most safety-relevant distinctions among the model's internal layers; (2) an adaptive safety prober trained to estimate the likelihood of toxic outputs from intermediate representations; and (3) a lightweight Refusal Head that selectively intervenes to modulate generation when safety risks are detected.
• Score: 36.99946524593795
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Recent progress in Multimodal Large Language Models (MLLMs) has unlocked powerful cross-modal reasoning abilities, but also raised new safety concerns, particularly when faced with adversarial multimodal inputs. To improve the safety of MLLMs during inference, we introduce a modular and adaptive inference-time intervention technology, AutoSteer, without requiring any fine-tuning of the underlying model. AutoSteer incorporates three core components: (1) a novel Safety Awareness Score (SAS) that automatically identifies the most safety-relevant distinctions among the model's internal layers; (2) an adaptive safety prober trained to estimate the likelihood of toxic outputs from intermediate representations; and (3) a lightweight Refusal Head that selectively intervenes to modulate generation when safety risks are detected. Experiments on LLaVA-OV and Chameleon across diverse safety-critical benchmarks demonstrate that AutoSteer significantly reduces the Attack Success Rate (ASR) for textual, visual, and cross-modal threats, while maintaining general abilities. These findings position AutoSteer as a practical, interpretable, and effective framework for safer deployment of multimodal AI systems.

Related papers

• Align is not Enough: Multimodal Universal Jailbreak Attack against Multimodal Large Language Models [83.80177564873094]
  We propose a unified multimodal universal jailbreak attack framework.<n>We evaluate the undesirable context generation of MLLMs like LLaVA, Yi-VL, MiniGPT4, MiniGPT-v2, and InstructBLIP.<n>This study underscores the urgent need for robust safety measures in MLLMs.
  arXiv  Detail & Related papers  (2025-06-02T04:33:56Z)
• DREAM: Disentangling Risks to Enhance Safety Alignment in Multimodal Large Language Models [37.104276926258095]
  Multimodal Large Language Models (MLLMs) pose unique safety challenges due to their integration of visual and textual data.<n>We introduce textbfDREAM (textittextbfDisentangling textbfRisks to textbfEnhance Safety textbfAlignment in textbfMLLMs), a novel approach that enhances safety alignment in MLLMs through supervised fine-tuning and iterative Reinforcement Learning from AI Feedback.
  arXiv  Detail & Related papers  (2025-04-25T03:54:24Z)
• SafeMLRM: Demystifying Safety in Multi-modal Large Reasoning Models [50.34706204154244]
  Acquiring reasoning capabilities catastrophically degrades inherited safety alignment.<n>Certain scenarios suffer 25 times higher attack rates.<n>Despite tight reasoning-answer safety coupling, MLRMs demonstrate nascent self-correction.
  arXiv  Detail & Related papers  (2025-04-09T06:53:23Z)
• Safe RLHF-V: Safe Reinforcement Learning from Multi-modal Human Feedback [34.01716144973483]
  Multimodal large language models (MLLMs) are essential for building general-purpose AI assistants.<n>How can we ensure safety alignment of MLLMs to prevent undesired behaviors?<n>In this work, we present the first exploration of the Safe RLHF-V -- the first multimodal safety alignment framework.
  arXiv  Detail & Related papers  (2025-03-22T07:40:20Z)
• SafeVLA: Towards Safety Alignment of Vision-Language-Action Model via Constrained Learning [10.844235123282056]
  Vision-language-action models (VLAs) show potential as generalist robot policies.<n>These models pose extreme safety challenges during real-world deployment, including the risk of harm to the environment, the robot itself, and humans.<n>We address this by exploring an integrated safety approach (ISA), systematically modeling safety requirements, then actively eliciting diverse unsafe behaviors.
  arXiv  Detail & Related papers  (2025-03-05T13:16:55Z)
• SafeAuto: Knowledge-Enhanced Safe Autonomous Driving with Multimodal Foundation Models [63.71984266104757]
  We propose SafeAuto, a framework that enhances MLLM-based autonomous driving by incorporating both unstructured and structured knowledge.<n>To explicitly integrate safety knowledge, we develop a reasoning component that translates traffic rules into first-order logic.<n>Our Multimodal Retrieval-Augmented Generation model leverages video, control signals, and environmental attributes to learn from past driving experiences.
  arXiv  Detail & Related papers  (2025-02-28T21:53:47Z)
• SafeDrive: Knowledge- and Data-Driven Risk-Sensitive Decision-Making for Autonomous Vehicles with Large Language Models [14.790308656087316]
  SafeDrive is a knowledge- and data-driven risk-sensitive decision-making framework to enhance autonomous driving safety and adaptability.<n>By integrating knowledge-driven insights with adaptive learning mechanisms, the framework ensures robust decision-making under uncertain conditions.
  arXiv  Detail & Related papers  (2024-12-17T16:45:27Z)
• Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training [67.30423823744506]
  We introduce a novel approach, Decoupled Refusal Training (DeRTa), designed to empower LLMs to refuse compliance to harmful prompts at any response position.<n>DeRTa incorporates two novel components: (1) Maximum Likelihood Estimation with Harmful Response Prefix, which trains models to recognize and avoid unsafe content by appending a segment of harmful response to the beginning of a safe response, and (2) Reinforced Transition Optimization (RTO), which equips models with the ability to transition from potential harm to safety refusal consistently throughout the harmful response sequence.
  arXiv  Detail & Related papers  (2024-07-12T09:36:33Z)
• Empowering Autonomous Driving with Large Language Models: A Safety Perspective [82.90376711290808]
  This paper explores the integration of Large Language Models (LLMs) into Autonomous Driving systems. LLMs are intelligent decision-makers in behavioral planning, augmented with a safety verifier shield for contextual safety learning. We present two key studies in a simulated environment: an adaptive LLM-conditioned Model Predictive Control (MPC) and an LLM-enabled interactive behavior planning scheme with a state machine.
  arXiv  Detail & Related papers  (2023-11-28T03:13:09Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.