MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models
This addresses a critical safety gap for MLLM users by highlighting and benchmarking vulnerabilities to malicious image exploits, though it is incremental as it builds on existing safety concerns in LLMs.
The paper tackles the safety vulnerabilities of Multimodal Large Language Models (MLLMs) to image-based manipulations, finding that 12 state-of-the-art models are susceptible to breaches even with safety-aligned LLMs, and proposes a prompting strategy to enhance resilience.
The security concerns surrounding Large Language Models (LLMs) have been extensively explored, yet the safety of Multimodal Large Language Models (MLLMs) remains understudied. In this paper, we observe that Multimodal Large Language Models (MLLMs) can be easily compromised by query-relevant images, as if the text query itself were malicious. To address this, we introduce MM-SafetyBench, a comprehensive framework designed for conducting safety-critical evaluations of MLLMs against such image-based manipulations. We have compiled a dataset comprising 13 scenarios, resulting in a total of 5,040 text-image pairs. Our analysis across 12 state-of-the-art models reveals that MLLMs are susceptible to breaches instigated by our approach, even when the equipped LLMs have been safety-aligned. In response, we propose a straightforward yet effective prompting strategy to enhance the resilience of MLLMs against these types of attacks. Our work underscores the need for a concerted effort to strengthen and enhance the safety measures of open-source MLLMs against potential malicious exploits. The resource is available at https://github.com/isXinLiu/MM-SafetyBench