Juanzi Li

This article discusses a method for identifying non-factual content in AI responses without human labels.

New system enhances academic searches using language models.

Innovative method improves student assessment with fewer data requirements.

A toolkit for assessing performance of retrieval-augmented models in specific domains.

GLM-4 models show improved capabilities in language understanding and generation.

Research highlights safety neurons' role in enhancing LLM safety and responsibility.

A new dataset enhances the accuracy of event factuality detection in texts.