Binxu Wang ☕️

About Me

Binxu Wang is a research fellow at Kempner Institute. Her research interests include visual neuroscience, generative models, and science of AI. She uses tools from AI to study neural system, and uses method of neuroscience to dissect AI system.

Download CV
Interests
  • Visual Neuroscience
  • Computational Neuroscience
  • Generative Model
  • Science of AI
Education
  • PhD in Neuroscience

    Washington University | Harvard Medical School

  • BSc Physics

    Peking University

My research: 🧠 → 🖼️

Generative Models x Visual Neuroscience

I’m a research fellow at Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University,

I work at the intersection of generative model and visual neuroscience – generating images from the brain, and dissecting how generative models actually work. I combine theory and experiments, with a toolkit spanning geometry, interpretability, and optimization.

I blog about deep learning, math and neuroscience.

Feel free to reach out for questions and collaborations 😃

Selected works
Recent Publications
(2025). Feature accentuation along the encoding axes of IT neurons uncovers hidden differences in model-brain alignment. Journal of Vision.
(2025). Archetypal SAE: Adaptive and Stable Dictionary Learning for Concept Extraction in Large Vision Models. ICML 2025 (arXiv:2502.12892).
(2024). Diverse capability and scaling of diffusion and auto-regressive models when learning abstract rules. The First Workshop on System-2 Reasoning at Scale, NeurIPS'24.
Recent & Upcoming Talks