Fanghui Liu

5

Papers

10

Total Citations

Papers (5)

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

Generalization of Scaled Deep ResNets in the Mean-Field Regime

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

High-Dimensional Kernel Methods under Covariate Shift: Data-Dependent Implicit Regularization

Revisiting Character-level Adversarial Attacks for Language Models