Academic
About Me#
我是中国科学技术大学（USTC）少年班学院 2024 级人工智能本科生。我的研究兴趣集中在可解释 AI 系统，主要关注 LLM 推理与内部机制、Agent Memory，以及生成式推荐。
我尤其关注具有清晰数学、几何或结构化解释的方法，希望借此理解经验现象、支持验证，并提升 AI 系统的可靠性。
Research Interests#
LLM Reasoning & Interpretability
理解并塑造 LLM 推理与行为背后的内部机制，寻找跨任务、可验证和可泛化的表征规律。
Agent Memory
研究具有显式结构、可验证检索路径与可审计修正机制的长期记忆系统。
Generative Recommendation
研究推理何时有助于推荐，以及生成式推荐中的信号、表征与决策机制。
Selected Research#
SHELF: From Similarity Retrieval to Path-Aware Auditable Memory for LLM Agents
Under Review at NeurIPS 2026Research Intern, USTC — advised by Prof. Xiang Wang
SHELF is a path-aware auditable memory framework that writes facts into explicit structural addresses and retrieves evidence through query-conditioned paths. It supports variable-level diagnosis and targeted revision of memory-access decisions.
Adaptive-Thinking for Generative Recommendation
Ongoing ResearchResearch Intern, Alpha-Lab, USTC — advised by Prof. An Zhang
Investigating when explicit reasoning helps or hurts generative recommendation, and whether think/not-think inference relies on different recommendation signals.
Selected Project#
CoSwipe: AI Companion for Short-Video Feed Interaction
Third Prize, Track 1, Douyin AI Innovator Plan 2026 Hackathon, USTC Station
CoSwipe is an AI companion embedded in the short-video feed. It responds to recent browsing behavior in a friend-like way, making passive scrolling more interactive and personally connected.
DemoCV#
More details about my education, research experience, projects, awards, and technical background are available in my CV.
Download CV