Publications

Diffusion Interpretability Mechanistic Interpretability Computer Vision Science of AI
Binxu Wang, Jingxuan Fan, Xu Pan (2026). Circuit Mechanisms for Spatial Relation Generation in Diffusion Transformers. Conference on Computer Vision and Pattern Recognition (CVPR) 2026.
PDF
Circuit Mechanisms for Spatial Relation Generation in Diffusion Transformers
Diffusion Theory Science of AI
Emma Lucia Byrnes Finn, Binxu Wang, T. Anderson Keller, Demba E. Ba (2026). Where the Score Lives: A Wavelet View of Diffusion. International Conference on Artificial Intelligence and Statistics (AISTATS) 2026.
Visual Neuroscience NeuroAI Generative Model
Binxu Wang, Carlos R. Ponce (2026). Neuronal Tuning Aligns Dynamically with Object- and Texture Manifolds across the Visual Hierarchy. Nature Neuroscience.
PDF Project Video 📝 BLOG
Neuronal Tuning Aligns Dynamically with Object- and Texture Manifolds across the Visual Hierarchy
LLM Reinforcement Learning Science of AI Theory
William Hoy, Binxu Wang, Xu Pan (2026). Matching Accuracy, Different Geometry: Evolution Strategies vs GRPO in LLM Post-training. under review at COLM 2026.
PDF
Interpretability Mechanistic Interpretability Science of AI LLM
Shun Shao, Binxu Wang, Shay B Cohen, Yonatan Belinkov (2026). Differentiable Faithfulness Alignment for Cross-Model Circuit Transfer. under review at COLM 2026.
Perspective Visual Neuroscience Neuroscience Deep Learning NeuroAI
Binxu Wang, Carlos R Ponce (2025). Structure as an inductive bias for brain–model alignment. Nature Machine Intelligence.
PDF
Diffusion Science of AI Generative Model
Binxu Wang, Emma Lucia Byrnes Finn, Bingbin Liu (2025). When Rule Learning Breaks: Diffusion Fails to Learn Parity of Many Bits. Workshop on Structured Probabilistic Inference & Generative Modeling, NeurIPS 2025 (oral).
Visual Neuroscience NeuroAI Neuron Modelling Digital Twins
Jacob S Prince, Binxu Wang, Akshay v Jagadeesh, Thomas Fel, Emily Lo, George a Alvarez, Margaret S Livingstone, Talia Konkle (2025). Feature accentuation along the encoding axes of IT neurons uncovers hidden differences in model-brain alignment. Journal of Vision.
Cite
Diffusion Learning Dynamics Theory Science of AI
Binxu Wang, Cengiz Pehlevan (2025). An analytical theory of power law spectral bias in the learning dynamics of diffusion models. NeurIPS 2025 (spotlight, top 3%) — arXiv:2503.03206.
PDF Cite
An analytical theory of power law spectral bias in the learning dynamics of diffusion models
Interpretability Computer Vision
Thomas Fel, Binxu Wang, Michael A. Lepori, Matthew Kowal, Andrew Lee, Randall Balestriero, Stella Joseph, Ekdeep S. Lubana, Talia Konkle, Demba E. Ba, Martin Wattenberg (2025). Into the Rabbit Hull: From Task-Relevant Concepts in DINO to Minkowski Geometry. International Conference on Learning Representations (ICLR) 2025.
PDF