Preprint Open access
ThinkingGuard: Decoding Implicit Hazards via Step-by-Step Risk Attribution in Multimodal Large Language Models
While Multimodal Large Language Models (MLLMs) are increasingly deployed in safety-critical domains, their reliability is threatened by multimodal implicit risks. Unlike explicit threats, these hazards emerge when individually benign text and neutral visual entities logically converge to induce unsafe outputs. Current …