Xiaoyu Shen

StableMask enhances attention distribution for better language model performance.

Exploring in-context learning and its implications for multilingual AI performance.

InternLM-Law enhances responses to diverse Chinese legal questions with advanced training.

This study evaluates how well large language models use external information.

A new approach to training reward models that aligns with human preferences.

This study assesses large language models as judges in math reasoning tasks.