Hi, I’m Tianyu. I am a Ph.D. student at Dartmouth College. I have the great honor of being advised by Prof. Yaoqing Yang.

I am passionate about model diagnostics and mechanistic interpretability . My current research is focused on

understanding the mechanisms, dynamics and generalization of LLMs from the perspective of random matrix theory, high-dimensional statistics and loss landscape;
leveraging model/data diagnostics and interpretability insights to improve the transparency, robustness and efficiency of (scientific) machine learning.

📖 Educations

2025.09 - present, Ph.D. in Machine Learning, Department of Computer Science, Dartmouth College.
2022.09 - 2025.06, M.S. in Mathematics, Department of Mathematics, Nanjing University.
2018.09 - 2022.06, B.S. in Statistics, Kuang Yaming Honors School, Nanjing University.

🔥 News

2026.02: Our work“Suspicious Alignment of SGD:A Fine-Grained Step Size Condition Analysis” was awarded the Best Student Paper Award at ALT 2026 .
2025.07: Our work“From Spikes to Heavy Tails: Unveiling the Spectral Evolution of Neural Networks” is accepted by TMLR.
2025.05: Our work “LIFT the Veil for the Truth: Principal Weights Emerge after Rank Reduction for Reasoning-Focused Supervised Fine-Tuning” is accepted by ICML 2025 as Poster!
2024.10: 😁 Completing a wonderful three-month visiting at Dartmouth College.
2024.09: Our work “Model Balancing Helps Low-data Training and Fine-tuning” is accepted by EMNLP 2024 as Oral Presentation!
2023.09: Our work “Temperature Balancing, Layer-wise Weight Analysis, and Neural Network Training” is accepted by NeurIPS 2023 as Spotlight!