Position: Lecturer, School of Information Science and Engineering
Affiliation: East China University of Science and Technology
E-mail: "biaoyi" + "@" + "ecust.edu.cn"
Address: 130 Meilong Road, Xuhui District, Shanghai 200237, China
My research is on the safety and trustworthiness of large language models. Two questions run through it: how do you keep a model from being turned harmful — by a jailbreak, a planted backdoor, or fine-tuning on harmful data — and how do you verify what a model gives back, by reading its internal states, tracing corrupted knowledge to its source, and marking what belongs to whom. Earlier work in linguistic steganography and steganalysis set the recurring theme: what can be hidden inside text, and how it can be found. I am building a group around these questions at ECUST and am looking for students and interns to join it.
† corresponding author · * equal contribution
Hiding in Plain Sight: A Steganographic Approach to Stealthy LLM Jailbreaks [Paper] [Code]
Jianing Geng*, Biao Yi*, Zekun Fei, Ruiqi He, Lihai Nie, Tong Li, Zheli Liu
EMNLP 2026 (Findings)
CTRAP: Embedding Collapse Trap to Safeguard Large Language Models from Harmful Fine-Tuning [Paper] [Code]
Biao Yi, Tiansheng Huang, Baolei Zhang, Tong Li, Lihai Nie, Zheli Liu, Li Shen
ACL 2026
SEAD: A Surrogate-free Label-only Membership Inference Attack against Pre-trained LLMs with Semantic-Aware Density [Paper] [Code]
Biao Yi, Jiahao Li, Yiming Li, Yu He, Baolei Zhang, Zheli Liu, Dacheng Tao
ACL 2026 (Findings)
Who Taught the Lie? Responsibility Attribution for Poisoned Knowledge in Retrieval-Augmented Generation [Paper] [Code]
Baolei Zhang, Haoran Xin, Yuxi Chen, Zhuqing Liu, Biao Yi, Tong Li, Lihai Nie, Zheli Liu, Minghong Fang
IEEE S&P 2026
BadReasoner: Planting Tunable Overthinking Backdoors into Large Reasoning Models for Fun or Profit [Paper] [Code]
Biao Yi, Zekun Fei, Jianing Geng, Tong Li, Lihai Nie, Zheli Liu, Yiming Li
ICASSP 2026 (Oral)
SafeGrad: Gradient Surgery for Safe LLM Fine-Tuning [Paper] [Code]
Biao Yi, Jiahao Li, Baolei Zhang, Lihai Nie, Tong Li, Tiansheng Huang, Zheli Liu
ICASSP 2026
Practical Framework for Privacy-Preserving and Byzantine-Robust Federated Learning [Paper] [Code]
Baolei Zhang, Minghong Fang, Zhuqing Liu, Biao Yi, Peng Zhou, Yan Wang, Tong Li, Zheli Liu
IEEE TIFS 2026
Your Semantic-Independent Watermark is Fragile: A Semantic Perturbation Attack against EaaS Watermark [Paper] [Code]
Zekun Fei*, Biao Yi*, Jianing Geng, Ruiqi He, Lihai Nie, Zheli Liu
EMNLP 2025 (Findings)
Prompt-Guided Internal States for Hallucination Detection of Large Language Models [Paper] [Code]
Fujie Zhang, Peiqi Yu, Biao Yi†, Baolei Zhang, Tong Li, Zheli Liu
ACL 2025
Traceback of Poisoning Attacks to Retrieval-Augmented Generation [Paper] [Code]
Baolei Zhang, Haoran Xin, Minghong Fang, Zhuqing Liu, Biao Yi, Tong Li, Zheli Liu
WWW 2025
Probe before You Talk: Towards Black-box Defense against Backdoor Unalignment for Large Language Models [Paper] [Code]
Biao Yi, Tiansheng Huang, Sishuo Chen, Tong Li, Zheli Liu, Zhixuan Chu, Yiming Li
ICLR 2025
BadActs: A Universal Backdoor Defense in the Activation Space [Paper] [Code]
Biao Yi, Sishuo Chen, Yiming Li, Tong Li, Baolei Zhang, Zheli Liu
ACL 2024 (Findings)
Semantic-preserving Linguistic Steganography by Pivot Translation and Semantic-aware Bins Coding [Paper] [Code]
Tianyu Yang, Hanzhou Wu, Biao Yi, Guorui Feng, Xinpeng Zhang
IEEE TDSC 2024
Exploiting Language Model for Efficient Linguistic Steganalysis [Paper] [Code]
Biao Yi, Hanzhou Wu, Guorui Feng, Xinpeng Zhang
ICASSP 2022
ALiSa: Acrostic Linguistic Steganography based on BERT and Gibbs Sampling [Paper] [Code]
Biao Yi, Hanzhou Wu, Guorui Feng, Xinpeng Zhang
IEEE SPL 2022
Linguistic Steganalysis with Graph Neural Networks [Paper] [Code]
Hanzhou Wu, Biao Yi, Feng Ding, Guorui Feng, Xinpeng Zhang
IEEE SPL 2021
Nankai University, China
Ph.D. • Sept. 2022 - June 2026
Advisor: Prof. Zheli Liu, Changjiang Scholar and Dean of the College of Computer Science
Shanghai University, China
M.Eng. • Sept. 2019 - June 2022
Advisor: Assoc. Prof. Hanzhou Wu, in the group of Prof. Xinpeng Zhang (NSFC Distinguished Young Scholar)
University of South China, China
B.Eng. • Sept. 2015 - June 2019
East China University of Science and Technology, China
Lecturer, School of Information Science and Engineering • July 2026 - present
Program Committee: USENIX Security, NeurIPS, ICLR, AAAI, WWW, SIGIR
Journal Reviewer: IEEE Transactions on Information Forensics and Security, IEEE Signal Processing Letters
Last update: Sept. 2026