May 2026
One paper on multilingual interpretability is currently under submission.
Yilong Wang
Student Researcher · Multilingual NLP · Mechanistic Interpretation · On-Policy Distillation
I am a master's student at TU Berlin, currently supervised by Qianli Wang. Previously, I was supervised by Qianli Wang and Nils Feldhus on my bachelor's thesis. My research interests include multilingual NLP, mechanistic interpretation, counterfactual, on-policy distillation, reinforcement learning, AI memory, and modality priors.
I am currently seeking internship and PhD position opportunities. Feel free to contact me via email.
News
Jan 2026
One paper on counterfactual generation was submitted earlier this year.
Selected Publications
Macro: Enhancing Multilingual Counterfactual Explanations through Alignment-as-Preference Optimization
Yilong Wang*, Qianli Wang*, Bohao Chu, Yihong Liu, Jing Yang, Simon Ostermann.
In submission, 2026.
* equal contribution.
iFlip: Iterative Feedback-driven Counterfactual Example Refinement
Yilong Wang*, Qianli Wang*, Nils Feldhus.
In submission, 2026.
* equal contribution.
Open Source Projects
Apache SystemDS
Ongoing
Contributor · 2026.05 – Present ·
SYSTEMDS-3942
I am contributing to an open-source project in Apache SystemDS on LLM-assisted generation of data augmentation pipelines. Our work focuses on building a flexible pipeline for comparing published augmentation baselines with LLM-generated strategies across datasets with different modalities.