FENCE: A Financial and Multimodal Jailbreak Detection Dataset

Jailbreaking poses a significant risk to the deployment of Large Language Models (LLMs) and Vision Language Models (VLMs). VLMs are particularly vulnerable because they process both text and images, creating broader attack surfaces. However, available resources for jailbreak detection are scarce, particularly in finance. To address this gap, we present FENCE, a bilingual (Korean-English) multimodal dataset for training and evaluating jailbreak detectors in financial applications. FENCE emphasizes domain realism through finance-relevant queries paired with image-grounded threats. Experiments with commercial and open-source VLMs reveal consistent vulnerabilities, with GPT-4o showing measurable attack success rates and open-source models displaying greater exposure. A baseline detector trained on FENCE achieves 99 percent in-distribution accuracy and maintains strong performance on external benchmarks, underscoring the dataset's robustness for training reliable detection models. FENCE provides a focused resource for advancing multimodal jailbreak detection in finance and for supporting safer, more reliable AI systems in sensitive domains. Warning: This paper includes example data that may be offensive.

Key Contributions

FENCE: a bilingual (Korean-English) multimodal dataset covering 15+ financial topics for training and evaluating jailbreak detectors targeting VLMs in finance
Systematic evaluation of commercial (GPT-4o) and open-source VLMs revealing measurable jailbreak vulnerabilities in financial contexts, with open-source models showing greater exposure
Baseline binary jailbreak classifier trained on FENCE achieving 99% in-distribution accuracy with strong generalization to external benchmarks