Search papers, labs, and topics across Lattice.
2
0
4
No existing model can effectively ground the spatial structure of student reasoning in multi-page handwritten homework, revealing a significant gap in automated assessment capabilities.
Multimodal perception is no longer just an add-on: GLM-5V-Turbo bakes it directly into the core of reasoning, planning, and action.
