#artificial intelligence
Oct 2025
On the Faithfulness of Visual Thinking: Measurement and Enhancement
This paper first probes MCoT faithfulness by measuring how much the prediction changes when its visual and textual thoughts are intervened, and proposes Sufficient-Component Cause Model (SCCM) learning, which encourages the MCoT to generate sufficient yet minimal visual components that are independently capable of lead...
Zujing Liu, Jun-Wen Pan, Qi She et al.
· arXiv.org · 15 citations
· ⚡2