Haihan Zhang

Strong Correlations Induce Cause Only Predictions in Transformer Training

Reliable Fine-Grained Evaluation of Natural Language Math Proofs

Scaling Law for SGD in Quadratically Parameterized Linear Regression

Generalization Analysis of SGD in Linear Regression under Covariate Shift: A View from Preconditioning

Learning Curves of Stochastic Gradient Descent in Kernel Regression

Tool Decoding: A Plug-and-Play Approach to Enhancing Language Models for Tool Usage