Researcher
Co-authored research on Render-of-Thought for visual latent reasoning achieving 3-4x token compression over explicit CoT.
How media typically covers Yifan Li
Referenced in coverage
Render-of-Thought (RoT) converts textual chain-of-thought reasoning into images, achieving 3-4x token compression and inference acceleration while maintaining competitive performance on mathematical and logical reasoning tasks.
“Co-authored research on Render-of-Thought for visual latent reasoning achieving 3-4x token compression over explicit CoT.”