About me

I am currently a third-year PhD student at Pennsylvania State University. I am fortunate to be advised by Dr. Jinyuan Jia. My research focuses on AI Security & Trustworthy ML, recently focusing on LLM robustness and transparency. Previously, I earned double bachelor’s degrees in Mechanical Engineering and Computer Science from RPI, followed by a master’s in Computer Science from Duke University.

Current research focuses

• Transparency of LLM-empowered AI systems for security purposes
• Automatic Red-Teaming for LLM Agents
• Provably secure/robust machine learning systems

Selected Publications

* Equal contribution

• Runpeng Geng, Yanting Wang, Chenlong Yin, Minhao Cheng, Ying Chen, Jinyuan Jia. PISanitizer: Preventing Prompt Injection to Long-Context LLMs via Prompt Sanitization, In EMNLP, 2026. 🔗 Code
• Chenlong Yin, Runpeng Geng, Yanting Wang, and Jinyuan Jia. PISmith: Reinforcement Learning-based Red Teaming for Prompt Injection Defenses, In COLM, 2026. 🔗 Code
• *Runpeng Geng, *Chenlong Yin, Yanting Wang, Ying Chen, and Jinyuan Jia. PIArena: A Platform for Prompt Injection Evaluation, In ACL, 2026. 🔗 Code
• Yanting Wang, Runpeng Geng, Ying Chen, and Jinyuan Jia. AttnTrace: Attention-based Context Traceback for Long-Context LLMs, In IEEE S&P, 2026. 🔗 Code Demo (HF Spaces)
• Yanting Wang and Jinyuan Jia. EnsembleSHAP: Faithful and Certifiably Robust Attribution for Random Subspace Method, In ICLR, 2026.
• *Yanting Wang, *Wei Zou, Runpeng Geng, and Jinyuan Jia. TracLLM: A Generic Framework for Attributing Long Context LLMs, In USENIX Security, 2025. 🔗 Code
• Yupei Liu, Yanting Wang, and Jinyuan Jia. TrojanDec: Data-free Detection of Trojan Inputs in Self-supervised Learning, In AAAI, 2025.
• Yanting Wang, Wei Zou, Jinyuan Jia. FCert: Certifiably Robust Few-Shot Classification in the Era of Foundation Models, In IEEE S&P, 2024. 🔗 Code
• Yanting Wang, Hongye Fu, Wei Zou, and Jinyuan Jia. MMCert: Provable Defense against Adversarial Attacks to Multi-modal Models, In CVPR, 2024. 🔗 Code

Preprints

• Yanting Wang, Runpeng Geng, Jinghui Chen, Minhao Cheng, Jinyuan Jia.
  TASO: Jailbreak LLMs via Alternative Template and Suffix Optimization, In arXiv, 2025.
• Wei Zou, Yupei Liu, Yanting Wang, Ying Chen, Neil Gong, and Jinyuan Jia.
  PIShield: Detecting Prompt Injection Attacks via Intrinsic LLM Features, In arXiv, 2025. 🔗 Code
• Runpeng Geng, Yanting Wang, Ying Chen, and Jinyuan Jia.
  UniC-RAG: Universal Knowledge Corruption Attacks to Retrieval-Augmented Generation, In arXiv, 2025.
• Yupei Liu, Yanting Wang, Yuqi Jia, Jinyuan Jia, and Neil Zhenqiang Gong.
  SecInfer: Preventing Prompt Injection via Inference-Time Scaling, In arXiv, 2025.
• Yuzhou Nie, Yanting Wang, Jinyuan Jia, Michael J. De Lucia, Nathaniel D. Bastian, Wenbo Guo, and Dawn Song. TrojFM: Resource-efficient Backdoor Attacks against Very Large Foundation Models, In arxiv, 2024.

Experience

Research Scientist Intern (AI Security), ByteDance, San Jose, CA | May 2026 – Present

• Research on automated red-teaming for LLM agents.

Professional Service

Program Committee member:

Asia CCS, 2027; AAAI, 2026

Reviewer:

COLM, 2026; NeurIPS, 2026; Pattern Recognition, 2025; TNNLS, 2024

Personal

I’m always interested in trying new things. Outside of work, I enjoy playing all kinds of ball sports, hiking, and snowboarding.