CORTEXA
← Browse

Hong Bu

1 paper indexed

arxivcs.CV2026-07-23

Do Pathology Vision-Language Models Truly See Pathology?

Chengyang Zhang, Wenchuan Zhang, Bo Li, Xinyu Liu, Jiaming Yang, Mengran Li, et al.

Pathology vision-language models (VLMs) have recently progressed rapidly and are commonly evaluated by answer accuracy on pathology VQA benchmarks. However, we dig into current evaluations and identify three overlooked issues: 1) Visual evidence is not always necessary. For insta…

View free PDFSource page