1 paper · 1 filter
Manyu Li, Ruian He, Chenxi Ma +2
Multimodal large language models (MLLMs) show remarkable potential for scientific reasoning, yet their performance in specialized domains such as microscopy remains limited by the…