Publications and Preprints

Yao Tong
Equal contribution
, Haonan Wang
Equal contribution
, Siquan Li, Kenji Kawaguchi, Tianyang Hu (2025). SeedPrints: Fingerprints Can Even Tell Which Seed Your Large Language Model Was Trained From. In Lock-LLM NeuriPS Workshop 2025.
Preprint
Yao Tong, Jiayuan Ye, Anastasia Borovykh, Reza Shokri (2025). Decomposing Extrapolative Problem Solving: Spatial Transfer and Length Scaling with Map Worlds.
Yao Tong, Jiayuan Ye, Anastasia Borovykh, Reza Shokri (2025). When Transformers Can (or Can’t) Generalize Compositionally? A Data-Distribution Perspective. In NeuriPS WCTD Workshop 2025.
Yao Tong
Equal contribution
, Weijun Li
Equal contribution
, Xuanli He, Haolan Zhan, Qiongkai Xu (2025). Cut the Deadwood Out: Training-Free Backdoor Purification via Guided Module Substitution. In Findings of Association for Computational Linguistics EMNLP 2025.
PDF Code
Yao Tong
Equal contribution
, Jiayuan Ye
Equal contribution
, Sajjad Zarifzadeh, Reza Shokri (2024). How much of my dataset did you use? Quantitative Data Usage Inference in Machine Learning. In ICLR 2025 [Oral Presentation (Top ∼1.5% among submissions)].
PDF Code Video
Haonan Wang, Qianli Shen, Yao Tong, Yang Zhang, Kenji Kawaguchi (2024). The Stronger the Diffusion Model, the Easier the Backdoor: Data Poisoning to Induce Copyright Breaches Without Adjusting Finetuning Pipeline. In NeurIPS Workshop on Backdoors in Deep Learning, 2023. [Oral Presentation]; In ICML 2024 [Oral Presentation (Top ∼2% among submissions)].
PDF Code Video
Xudong Shen, Hannah Brown, Jiashu Tao, Martin Strobel, Yao Tong, Akshay Narayan, Harold Soh, Finale Doshi-Velez
Work conducted during Finale’s visit to NUS as an outcome of the RRAI Workshop; middle authors order is alphabetical.
(2024). Towards Regulatable AI Systems: Technical Gaps and Policy Opportunities. In CACM, 2024.
PDF