Learning to See Before Seeing: Demystifying LLM Visual Priors from Language Pre-training
Paper
•
2509.26625
•
Published
•
43
Zihan Li
zl111
AI & ML interests
My research focuses on the Foundation Model including the multimodal large language model and Large-scale Data Harmonization.
Recent Activity
upvoted
a
paper
about 1 month ago
Learning to See Before Seeing: Demystifying LLM Visual Priors from
Language Pre-training
authored
a paper
3 months ago
VisionUnite: A Vision-Language Foundation Model for Ophthalmology
Enhanced with Clinical Knowledge
Organizations
upvoted
a
paper
about 1 month ago