Yu Leon Liu

I am a researcher working on machine learning, with a focus on reasoning, interpretability, and alignment. I recently completed a PhD in mathematics at Harvard, where my research was in mathematical physics and topology. During that time, I also worked on quantitative research at a stealth high-frequency trading startup and on agentic AI systems as an engineering intern.

Email / GitHub / Google Scholar / CV (PDF)

Projects

GRPO training dynamics: accuracy up, response length down then up

LLM Post-Training with GRPO and SFT (Reinforcement Learning)
Trained and evaluated Qwen2.5-3B on Countdown with GRPO/RLVR and SFT. GRPO raised add/subtract coverage from 54% to 94% (pass@10), with reasoning length tracing a three-phase arc as responses became coherent: rambling at 580 characters, down to a terse 172, then settling at 287. SFT was then used to enhance multiplication, and an SFT-then-GRPO pipeline raised add/subtract pass@1 to 71% (more confident single-shot, though lower at pass@k). Multiplication set to 0%.
[writeup] [models] [data]

Selected Publications

I have written around ten papers in mathematics; a few are below, and the full list is on the papers page.

The Smith Fiber Sequence of Invertible Field Theories
A. Debray, S. K. Devalapurkar, C. Krulewski, Y. L. Liu, N. Pacheco-Tallaj, R. Thorngren
Communications in Mathematical Physics, 2026 [arXiv]

A Long Exact Sequence in Symmetry Breaking
A. Debray, S. K. Devalapurkar, C. Krulewski, Y. L. Liu, N. Pacheco-Tallaj, R. Thorngren
Journal of High Energy Physics, 2025 [arXiv]

A Braided Monoidal (∞,2)-Category of Soergel Bimodules
Y. L. Liu, A. Mazel-Gee, D. Reutter, C. Stroppel, P. Wedrich
2024 (presented in an ICM plenary talk) [arXiv]

Constructing the Virasoro Groups from Differential Cohomology
A. Debray, Y. L. Liu, C. Weis
International Mathematics Research Notices, 2023 [arXiv]