Searching for Privacy Risks in LLM Agents via Simulation, Preprint, 2025
Attacking Vision-Language Computer Agents via Pop-ups, ACL, 2025
Distilling an End-to-End Voice Assistant from Speech Recognition Data, ACL, 2025
Sketch2Code: Evaluating Vision-Language Models for Interactive Web Design Prototyping, NAACL, 2025
Design2Code: How Far Are We From Automating Front-End Engineering?, NAACL, 2025
Dynamic LLM-Agent Network: An LLM-agent Collaboration Framework with Agent Evaluation, COLM, 2024
Auditing Gender Presentation Differences in Text-to-Image Models, EAAMO, 2024
Enhanced Visual Instruction Tuning for Text-rich Image Understanding, NeurIPS Workshop on Instruction Tuning and Instruction Following, 2023; Improved version, CVPR 2024
Robustness of Demonstration-based Learning Under Limited Data Scenario, EMNLP, 2022
Continual Sequence Generation with Adaptive Compositional Modules, ACL, 2022
Continual Learning for Text Classification with Information Disentanglement Based Regularization, NAACL, 2021
Research Experience
Interned at Adobe Research (2022-2023) under Ruiyi Zhang; currently visiting Stanford NLP.
Education
Received a bachelor's degree from Zhejiang University in 2021; currently a fourth-year Ph.D. student in Computer Science at Georgia Tech, advised by Diyi Yang.
Background
Research Interests: Continual learning, robustness, fairness, safety, and enabling AI to benefit from and for other AI and humans.
Miscellany
Also known as Steven; hobbies not specified. Participated in projects like SWE-Smith and Computer Agent Arena. Reviewer for ARR, ACL, NAACL, EMNLP, EACL, COLM, CoLLAs, ICLR. Website code is from Jon Barron.