A Vision Check-up for Language Models
What does learning to model relationships between strings teach large language models (LLMs) about the visual world? We systematically evaluate LLMs' abilities to generate and recognize an assortment of visual concepts of increasing complexity and then demonstrate how a…
- Year
- 2024
- Hosting
- External sourcelicense unknown
Cite
Notes
Only stored in your browser.