GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning Paper • 2505.11049 • Published May 16, 2025 • 60