PhysFieldBench: Can Multimodal Models Understand Physical Fields?
PhysFieldBench is introduced, a benchmark comprising 24 tasks and 1,160 evaluation examples across controlled equation fields, simulated physical fields, and observed physical fields that highlights the need to improve visual-to-physical grounding and cross-task generalization for MLLMs to reliably interpret physical f...
Yue-Zhou Ma, Hui-Kun Weng, Jialong Wu et al.
· 0 citations