Chenyan Xiong

Evaluating how well language models perform research surveys across various academic fields.

Exploring how in-context probing and influence functions enhance data selection for models.

A new method improves search clarity and user control.

RAGViz reveals how AI generates answers, making its workings more transparent.