Teaching experience 1
Undergraduate course, University 1, Department, 2014
Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Teaching experience 2
Workshop, University 1, Department, 2015
Page Not Found
Page not found. Your pixels are in another canvas.
Zixuan Wang
About me
Page not in menu
This is a page not in th emain menu
Posts
Future Blog Post
Published:
This post will show up by default. To disable scheduling of future posts, edit config.yml and set future: false.
Why Are “Asymmetric” Power Laws Necessary for Reasoning?
Published:
Uniform sampling sounds obviously right for long-tail learning: rare skills need more examples. But for compositional reasoning, that intuition can be exactly wrong. Uniform data fails due to worse landscape, but power-law data breaks the symmetry and creates a slope.
Blog Post number 4
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 3
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 2
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
Blog Post number 1
Published:
This is a sample blog post. Lorem ipsum I can’t remember the rest of lorem ipsum and don’t have an internet connection right now. Testing testing testing this blog post. Blog posts are cool.
publications
Analyzing Sharpness along GD Trajectory: Progressive Sharpening and Edge of Stability
Published in 36th Conference on Neural Information Processing Systems (NeurIPS), 2022
Recommended citation: Li, Z., Wang, Z., & Li, J. (2022). Analyzing sharpness along gd trajectory: Progressive sharpening and edge of stability. arXiv preprint arXiv:2207.12678. http://wangzx19.github.io/files/analyzing_sharpness_along_gd_t.pdf
Understanding Edge-of-Stability Training Dynamics With a Minimalist Example
Published in arXiv:2210.03294, 2022
Recommended citation: Zhu, X., Wang, Z., Wang, X., Zhou, M., & Ge, R. (2022). Understanding Edge-of-Stability Training Dynamics with a Minimalist Example. arXiv preprint arXiv:2210.03294. http://wangzx19.github.io/files/EoS_tex_ICLRv2.pdf
Learning Compositional Functions with Transformers from Easy-to-Hard Data
Published in , 2024
Recommended citation: Wang, Z., Wei, S., Hsu, D., & Lee, J. D. (2024). Transformers provably learn sparse token selection while fully-connected nets cannot. arXiv preprint arXiv:2406.06893.
Transformers Provably Learn Sparse Token Selection While Fully-Connected Nets Cannot
Published in arXiv:2406.06893, 2024
Recommended citation: Wang, Z., Wei, S., Hsu, D., & Lee, J. D. (2024). Transformers provably learn sparse token selection while fully-connected nets cannot. arXiv preprint arXiv:2406.06893. https://arxiv.org/pdf/2406.06893
Learning and Transferring Sparse Contextual Bigrams with Linear Transformers
Published in arXiv:2410.23438, 2024
Recommended citation: Ren, Y., Wang, Z., & Lee, J. D. (2024). Learning and Transferring Sparse Contextual Bigrams with Linear Transformers. arXiv preprint arXiv:2410.23438. https://arxiv.org/pdf/2410.23438
Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought
Published in arXiv:2502.21212, 2025
Recommended citation: Huang, J., Wang, Z., & Lee, J. D. (2025). Transformers Learn to Implement Multi-step Gradient Descent with Chain of Thought. arXiv preprint arXiv:2502.21212. https://arxiv.org/pdf/2502.21212
What Makes a Reward Model a Good Teacher? An Optimization Perspective
Published in arXiv:2503.15477, 2025
Recommended citation: Razin, N., Wang, Z., Strauss, H., Wei, S., Lee, J. D., & Arora, S. (2025). What Makes a Reward Model a Good Teacher? An Optimization Perspective. arXiv preprint arXiv:2503.15477. https://arxiv.org/abs/2503.15477
The Power of Power Law: Asymmetry Enables Compositional Reasoning
Published in ICML 2026, Spotlight, 2026
Recommended citation: Wang, Z., Dang, X., Lee, J. D., & Lyu, K. (2026). The Power of Power Law: Asymmetry Enables Compositional Reasoning. arXiv preprint arXiv:2604.22951. https://arxiv.org/abs/2604.22951
talks
Analyzing Sharpness along GD Trajectory: Progressive Sharpening and Edge of Stability
Published:
Here’s the link of this virtual talk.
Understanding Edge-of-Stability Training Dynamics With a Minimalist Example
Published:
Slides can be accessed here.