Seeing vs. Believing: Evaluating the Language Bias of Open-Source MLLMs in Counter-Intuitive Scenes 事件

Seeing vs. Believing: Evaluating the Language Bias of Open-Source MLLMs in Counter-Intuitive Scenes arXiv:2601.07737v2 Announce Type: replace Abstract: Multimodal Large Language Models (MLLMs) have demonstrated remarkable performance in mainstream visual understanding tasks, but their ability to process action scenes that contradict everyday common sense remains undertested. To address this gap, we introduce CAIT, a benchmark comprising 400 high-fidelity synthetic scenes focused on counter-intu