About Me
I am a second-year PhD student at Network Security Lab at University of Washington, advised by Prof. Radha Poovendran. I am fortunate to work closely with Dr. Xiang Yue at CMU LTI, Zhangchen Xu, Prof. Luyao Niu, and Prof. Bill Yuchen Lin at UW.
Prior to UW, I obtained my Bachelor degree in Communication Engineering at University of Glasgow (UofG). During my undergrad, I was advised by Prof. Lei Zhang at UofG. I am also fortunate to work with Prof. Jon Crowcroft at University of Cambridge.
My CV (updated 02.13, 2025) shows more details.
I am open to collaboration and discuss interesting ideas! Email yuetaili@uw.edu if you would like to share any opportunities, collaborations, or just chat!
Research Interests
My primary interests lie broadly in LLM Reasoning, Synthetic Dataset, and Trustworthy AI. I am particularly interested in: (1) understanding the reasoning capabilities of LLMs through rigorous analysis (2) investigating synthetic datasets optimized for effective model learning. My current research directions include:
(1) Reasoning and Synthetic Dataset
- Small Model Learnability Gap: We revealed that small models do not consistently benefit from long CoT or distillation from large teachers. Instead, they perform better on shorter, simpler reasoning chains that better align with their intrinsic learning capacity. We propose Mix Distillation, a simple yet effective strategy that combines both long and short CoT examples or reasoning from both large and small teachers.
- MagpieLM: We maintained the official Magpie Hugging Face repository and released open-sourced synthetic data generated from open-source LLMs. Our aligned MagpieLM models are still SOTA small language models for chat.
- SafeChain: We investigated how long CoT impacts safety and revealed that long CoT does not necessarily enhance safety. We introduced SafeChain, a dataset designed to improve safety alignment while preserving reasoning capabilities.
(2) Trustworthy AI
- CleanGen: A novel decoding algorithm that defends against various generative backdoor attacks, including advertisement injection, code injection, and malicious content generation.
- Spectral Aggregation: A novel defense mechanism to aggregate the spectral features of multilayers to detect poisoned samples, restoring the effectiveness of the latent separability assumption in adaptive backdoor attacks.
(3) Distributed Algorithms
- Distributed Consensus: Work includes Voting Validity, Wireless Distributed Consensus, Resource Allocation, and Distributed Consensus Network.
Publications
(Updated 02.13, 2025)
- Yuetai Li, Xiang Yue, Zhangchen Xu, Fengqing Jiang, Luyao Niu, Bhaskar Ramasubramanian, Bill Yuchen Lin, Radha Poovendran, “Small Models Struggle to Learn from Strong Reasoners”
- Fengqing Jiang, Zhangchen Xu, Yuetai Li, Luyao Niu, Zhen Xiang, Bo Li, Bill Yuchen Lin, Radha Poovendran, “SafeChain: Revisiting Safety of Language Model with Long Chain-of-Thought Reasoning Capability”
- Yuetai Li, Zhangchen Xu, Fengqing Jiang, Luyao Niu, Dinuka Sahabandu, Bhaskar Ramasubramanian, Radha Poovendran, “CLEANGEN: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models” Accepted by EMNLP 2024, June 2024
- Yuetai Li, Dinuka Sahabandu, Bhaskar Ramasubramanian, Radha Poovendran, "SpecAggre: Spectral Aggregation-based Latent Separability Recovery for Defense against Machine Learning Backdoor Attacks"
- Yuetai Li, Zhangchen Xu, Lei Zhang, Jon Crowcroft, “A Modularized Framework of Communication in Consensus”, submitted to IEEE/ACM Transactions on Networking
- Zhangchen Xu, Yuetai Li, Chenglin Feng, Lei Zhang, “Voting Validity: Exact Fault-tolerant Consensus”, Accepted by the 37th IEEE International Parallel & Distributed Processing Symposium (IPDPS), December 2022
- Dachao Yu, Yao Sun, Yuetai Li, Lei Zhang and Muhammad Ali Imran, "Communication Resource Allocation of Raft in Wireless Network," Published in IEEE Sensors Journal, September 2023
- Yuetai Li, Yixuan Fan, Lei Zhang, Jon Crowcroft, “RAFT Consensus Reliability in Wireless Networks: Probabilistic Analysis”, Published in IEEE Internet of Things Journal, February 2023
- Yuetai Li, Tong Tong, Benhao Pan, Huajun Yang, Ping Jiang, Weinan Caiyang, “Three-mirror system design for shaping the elliptical beam of a laser diode”, Published in Elsevier Optik, Volume 264, 2022
- Yuetai Li, Xinbin Chen, Jiale Wang, Tao Zhan, Huajun Yang, Weinan Caiyang, Ping Jiang, “Shaping and transmitting elliptical beam from laser diode by off-axis quadric reflective mirrors”, Published in Elsevier Optics Communications, Volume 493, 2021