1 paper
Chih-Ting Liao, Xi Xiao, Chunlei Meng +6
Multimodal large language models (MLLMs) have advanced static visual--spatial reasoning, yet they often fail to preserve long-horizon spatial coherence in embodied settings where b…