Han Wu

Automated Optimization Modeling via a Localizable Error-Driven Perspective

REG: A Regularization Optimizer for Robust Training Dynamics

Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensembling

Activation-Guided Consensus Merging for Large Language Models

Preserving LLM Capabilities through Calibration Data Curation: From Analysis to Optimization

Towards Efficient Adaptation of Pruning Strategy in Large Language Models