Jingtong Su 苏敬童

Ph.D. Student, Center for Data Science, New York University Email: js12196 [@] nyu [DOT] edu

About me

Hi! I'm a Ph.D. student at NYU Center for Data Science, working with Prof. Julia Kempe. Currently, I am interested in the reliability of foundation models, spanning robustness, safety, and interpretability. I spent two wonderful years at Meta AI, FAIR as a visiting researcher, mentored by Karen Ullrich. I have collaborated with experts from Center for Computational Astrophysics, Flatiron Institute. I obtained my Bachelor's degree in Data Science and Big Data Technology at Yuanpei College, Peking University. At that time, I did research with Prof. Jason D. Lee and Prof. Yisen Wang.

Publications

• A Single Character can Make or Break Your LLM Evals.
  Jingtong Su, Jianyu Zhang, Karen Ullrich, Léon Bottou, and Mark Ibrahim.
  

• From Concepts to Components: Concept-Agnostic Attention Module Discovery in Transformers.
  Jingtong Su, Julia Kempe, and Karen Ullrich.
  

• Enhancing and Evaluating Probabilistic Circuits for High-Resolution Lossless Image Compression.
  Daniel Severo, Jingtong Su, Anji Liu, Jeff Johnson, Brian Karrer, Guy Van den Broeck, Matthew J. Muckley, and Karen Ullrich.
  Data Compression Conference 2025.
  

• Mission Impossible: A Statistical Perspective on Jailbreaking LLMs.
  Jingtong Su, Julia Kempe, and Karen Ullrich.
  NeurIPS 2024. [slides]
  

• On the Robustness of Neural Collapse and the Neural Collapse of Robustness.
  Jingtong Su, Ya Shi Zhang, Nikolaos Tsilivis, and Julia Kempe.
  TMLR.
  

• On the Connection between Invariant Learning and Adversarial Training for Out-of-Distribution Generalization.
  Shiji Xin, Yifei Wang, Jingtong Su, and Yisen Wang.
  AAAI 2023 (Oral).

• Wavelets beat Monkeys at Adversarial Robustness.
  Jingtong Su and Julia Kempe.
  NeurIPS Workshop on Machine Learning and the Physical Sciences, 2022.
  

• Adversarial Noise Injection for Learned Turbulence Simulations.
  Jingtong Su, Julia Kempe, Drummond Fielding, Nikolaos Tsilivis, Miles Cranmer, and Shirley Ho.
  NeurIPS Workshop on Machine Learning and the Physical Sciences, 2022.
  

• Can We achieve Robustness from Data Alone?
  Nikolaos Tsilivis, Jingtong Su, and Julia Kempe.
  ICML Workshop on New Frontiers in Adversarial Machine Learning, 2022.
  

• Sanity-Checking Pruning Methods: Random Tickets can Win the Jackpot.
  Jingtong Su*, Yihang Chen*, Tianle Cai*, Tianhao Wu, Ruiqi Gao, Liwei Wang, and Jason D. Lee.
  NeurIPS 2020. [slides] [code]
  (* denotes equal contribution)

Selected Awards

• Outstanding Graduate in Beijing (2021)

• Merit Student Pacesetter at PKU (2019)

• Gold Medal at Chinese Physics Olympiad (2016)

• Bronze Medal at Chinese Physics Olympiad (2015)