Ze-Xi Chen

Can Multi-modal (reasoning) LLMs detect document manipulation?

This study investigates the efficacy of state-of-the-art multi-modal large language models-including OpenAI O1, OpenAI 4o, Gemini Flash (thinking), Deepseek Janus, Grok, Llama 3.2 and 4, Qwen 2 and 2.5 VL, Mistral Pixtral, and Claude detecting fraudulent documents, and reveals that top-performing multi-modal LLMs demon...