Baihan Zhang

I'm an M.S. candidate in Computer Science at Stanford University. I work with the Stanford NeuroAI Lab, advised by Prof. Dan Yamins. I also have the fortune to work with Prof. Dan Boneh on AI safety. I completed my undergraduate studies at Stanford and was a Research Intern at Snowflake.

Email  /  CV  /  Scholar  /  Github  /  Twitter

profile photo

Research

I'm interested in deep learning applied to computer vision, robotics, and the security of AI systems.

Ψ-0.6: Adaptive Streams of Thought in a Unified Vision-Language World Model
Khai Loong Aw*, Klemen Kotar*, Baihan Zhang*, Yifan Liu*, Lilian Naing Chen,
Atlas Kazemian, Rahul Mysore Venkatesh, Daniel L. K. Yamins
* denotes equal contribution
arXiv, 2026 (under review at ICLR 2027)
project page / arXiv / code

Chain-of-thought transformed AI reasoning, but it is limited to words. What's the physical-world analogue? We introduce Ψ-0.6, one model that reasons in streams of thought: words, depth, motion, camera pose, and imagined scenes – and learns which to use for each problem. Ψ-0.6 surpasses leading open-weight models zero-shot on challenging spatial reasoning and makes two breakthroughs. First, it thinks beyond words, in streams of thought. Second, it learns how to think: which representation to use, and when.


Website template borrowed from Jon Barron. Source code is available here.