Hector

Hi! I'm Hector, a Ph.D. student working on Large Language Models.

My current research focuses on multi-turn instruction following and editing — understanding how LLMs drift when they iteratively revise their outputs across a conversation, and how to evaluate and mitigate it. More broadly, I care about rigorous evaluation of LLM behavior and building reliable, Python-based tooling for research.

This site hosts my research log & blog, where I write (mostly in Chinese) about papers I read, experiments I run, and the occasional DevOps adventure.

<!-- TODO: 加上实验室/导师信息,例如:I work at XXX Lab, advised by Prof. YYY. -->