Fanghui Liu

DAG-Math: Graph-of-Thought Guided Mathematical Reasoning in LLMs

The $\varphi$ Curve: The Shape of Generalization through the Lens of Norm-based Capacity Control

LoRA-One: One-Step Full Gradient Could Suffice for Fine-Tuning Large Language Models, Provably and Efficiently

How Gradient descent balances features: A dynamical analysis for two-layer neural networks

Physics-Informed Interpolator Generalizes Well in Fixed Dimension: Inductive Bias and Benign Overfitting