Biography
Advised by Professor Daphne Ippolito, my Ph.D. thesis research currently explores the topics of LLM Security and Adapters.
Previously at Cornell, advised by Professor Thomas Ristenpart, I concentrated my master thesis on End-to-End Encryption applications in security research. My research aims to enhance online community governance and privacy while balancing anti-abuse.
In my NLP work during undergraduate and master degree, guided by Professor Claire Cardie, my focus was on improving Information Extraction and model interpretability, particularly when inputs are at document level.
Education
- Ph.D. Student in C.S. - Carnegie Mellon University
- M.S. in C.S., 2024 - Cornell University
- B.A. in C.S., 2022 - Cornell University
Publications
For the most up-to-date publications list, see Google Scholar.
Command-V: Training-Free Representation Finetuning Transfer
LLMs unlock new paths to monetizing exploits
NoveltyBench: Evaluating Language Models for Humanlike Diversity
Are Triggers Needed for Document-Level Event Extraction?
IQA-EVAL: Automatic Evaluation of Human-Model Interactive Question Answering
Private Hierarchical Governance for Encrypted Messaging
Probing Representations for Document-level Event Extraction
Automatic Error Analysis for Document-level Information Extraction
Experience
Teaching Assistant
Cornell University • Ithaca, New York
August 2020 - May 2024
- CS4410 Operating System (Spring 2022, Spring 2024)
- CS5430 System Security (Fall 2023)
- CS3410 Computer System Organization and Programming (Spring 2021, Fall 2021, Fall 2022)
- CS2112 Object-Oriented Design and Data Structures (Honors) (Fall 2020)
Software Engineering Intern (Research)
Roblox • San Mateo, California
May 2023 - August 2023
- Machine Learning research @ Roblox Security