2024/11/06 by Song, Dingjie, Lai, Sicheng, Wang, Mingxuan +3 · 6 citations
#Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Multimedia (cs.MM)
paper · doi:10.48550/arxiv.2411.03823
The rapid advancement of multimodal large language models (MLLMs) has significantly enhanced performance across benchmarks. However, data contamination-unintentional memorization of benchmark data during model training-poses critical challenges for fair evaluation. Existing detection methods for unimodal large language models (LLMs) are inadequate for MLLMs due to multimodal data complexity and multi-phase training. We systematically analyze multimodal data contamination using our analytical framework, MM-Detect, which defines two contamination categories-unimodal and cross-modal-and effectively quantifies contamination severity across multiple-choice and caption-based Visual Question Answering tasks. Evaluations on twelve MLLMs and five benchmarks reveal significant contamination, particularly in proprietary models and older benchmarks. Crucially, contamination sometimes originates during unimodal pre-training rather than solely from multimodal fine-tuning. Our insights refine contamination understanding, guiding evaluation practices and improving multimodal model reliability.