I am a Ph.D. student at Informatics department at LMU Munich in Germany. Previously, I obtained my master at LMU Munich in Physics. My current research interests focus on LLM training and alignment.
I’m currently looking for motivated master’s and bachelor’s students in Munich for thesis projects or guided research who want to work on something actually cool in AI agents! Let’s have a chat.
Contact me (for any reason): Liu@dbs.ifi.lmu.de
Room G U104, Oettingenstraße 67, Munich
2. [Multimodal] [Safety] Multimodal Pragmatic Jailbreak on Text-to-image Models
Tong Liu, Zhixin Lai, Jiawen Wang, Gengyuan Zhang, Shuo Chen, Philip Torr, Vera Demberg, Volker Tresp, Jindong Gu
Association for Computational Linguistics (ACL) main, 2025.
arXiv code
We first propose that AI safety can be 𝐦𝐮𝐥𝐭𝐢𝐦𝐨𝐝𝐚𝐥!
We get best paper award 🏆 at ReGenAI@CVPR 2025.
1. [LLMs] [Psycholinguistics] Temperature-scaling Surprisal Estimates Improve Fit to Human Reading Times – But Does It Do So for the "Right Reasons"?
Tong Liu, Iza Škrjanec, Vera Demberg
Association for Computational Linguistics (ACL) main, 2024.
arXiv code
We show that temperature itself can substantially improve LLM human-likeness.
Tokenization is the gap.
6. [Multimodal] [RL] Rethinking Multi-Image Re-Representation in Multi-Image Understanding
Gengyuan Zhang, Xiao Han, Xinyu Xie, Tong Liu, Volker Tresp
arXiv5. [Agent] [Safety] Deep Research Brings Deeper Harm
Shuo Chen, Zonggen Li, Zhen Han, Bailan He, Tong Liu, Haokun Chen, Georg Groh, Philip Torr, Volker Tresp, Jindong Gu
Conference on Language Modeling (COLM), 2026.
arXiv code4. [Agent] [RL] On Effectiveness and Efficiency of Agentic Tool-calling and RL Training
Tong Liu, Cheng Qian, Matej Cief, Yuan He, Daniele Dan, Nikolaos Aletras, Gabriella Kazai
International Conference on Machine Learning (ICML), 2026.
arXiv3. [LLMs] [Alignment] FocalPO: Enhancing Preference Optimizing by Focusing on Correct Preference Rankings
Tong Liu, Xiao Yu, Wenxuan Zhou, Jindong Gu, Volker Tresp
Association for Computational Linguistics (ACL) main, 2025.
arXiv code2. [LLMs] [PEFT] HiFT: A Hierarchical Full Parameter Fine-Tuning Strategy
Yongkang Liu, Yiqun Zhang, Qian Li, Tong Liu, Shi Feng, Daling Wang, Yifei Zhang, Hinrich Schütze
Empirical Methods in Natural Language Processing (EMNLP) main, 2024.
arXiv code1. [Graph] [Calibration] On Calibration of Graph Neural Networks for Node Classification
Tong Liu*, Yushan Liu*, Marcel Hildebrandt, Mitchell Joblin, Hang Li, Volker Tresp
International Joint Conference on Neural Networks (IJCNN) (Oral), 2022. arXiv code