Harnessing Textual Refusal Directions for Multimodal Safety

arXiv:2606.31876 · cs.AI, cs.CV, cs.LG · Submitted 2026-06-30 · Read on arXiv

cs.AI, cs.CV, cs.LG

Submitted: 2026-06-30

Updated: 2026-10-16

Terminology

• Steering — 3× in this paper · explained in 2 episodes
• MLLMs — 2× in this paper · explained in 2 episodes
• Benchmark — 1× in this paper · explained in 2 episodes
• Cross-modal Alignment — 1× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 8 episodes
• Modality-Agnostic Refusal Steering — 1× in this paper
• Multimodal Inputs — 1× in this paper · explained in 2 episodes

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems