multimodal-feedback

Tag

Cards List
#multimodal-feedback

VisEditBench: Can Vision-Language Models Edit Visualization Code from Multimodal Feedback?

arXiv cs.CL · 3d ago Cached

Introduces VisEditBench, a benchmark of 1,395 human-annotated visualization code-editing tasks for vision-language models, covering feedback-guided repair and reference-guided restyling. Evaluates 20 VLMs and proposes VisEditAgent, a render-grounded editing framework that improves pass rates from 55.75% to 67.99%.

0 favorites 0 likes
← Back to home

Submit Feedback