Xinpeng Wang

PhD student, Ludwig-Maximilians-Universität München

Joined

June 2023

Names

Xinpeng Wang (Preferred)

Emails

****@cis.lmu.de (Confirmed)

Personal Links

Career & Education History

PhD student

Ludwig-Maximilians-Universität München (lmu.de)

2022 – 2026

MS student

Technische Universität München (tum.de)

2019 – 2022

Undergrad student

Dalian University of Technology (dlut.edu.cn)

2014 – 2018

Advisors, Relations & Conflicts

PhD Advisor

Hinrich Schuetze

Present

Coworker

Siyao (Logan) Peng

Present

Coworker

Leon Weber

Present

Coworker

Robert Litschko

Present

Coworker

Weißweiler, Leonie

Present

Coworker

Artemova, Ekaterina

Present

PhD Advisor

Barbara Plank

2022 – Present

Expertise

deep learning

Present

natural language processing

Present

computer vision

Present

Publications

Is it Thinking or Cheating? Detecting Implicit Reward Hacking by Measuring Reasoning Effort
Xinpeng Wang, Nitish Joshi, Barbara Plank, Rico Angell, He He
- ICLR 2026 Oral
- Readers: Everyone
Algorithmic Fidelity of Large Language Models in Generating Synthetic German Public Opinions: A Case Study
Bolei Ma, Berk Yoztyurk, Anna-Carolina Haensch, Xinpeng Wang, Markus Herklotz, Frauke Kreuter, Barbara Plank, Matthias Aßenmacher
- Social Sim'25
- Readers: Everyone
Refusal Direction is Universal Across Safety-Aligned Languages
Xinpeng Wang, Mingyang Wang, Yihong Liu, Hinrich Schuetze, Barbara Plank
- NeurIPS 2025 poster
- Readers: Everyone
Think Before Refusal : Triggering Safety Reflection in LLMs to Mitigate False Refusal Behavior
Shengyun Si, Xinpeng Wang, Guangyao Zhai, Nassir Navab, Barbara Plank
- CoRR 2025
- Readers: Everyone
Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
Xinpeng Wang, Chengzhi Hu, Paul Röttger, Barbara Plank
- ICLR 2025
- Readers: Everyone
FinerCut: Finer-grained Interpretable Layer Pruning for Large Language Models
Yang Zhang, Yawei Li, Xinpeng Wang, Qianli Shen, Barbara Plank, Bernd Bischl, Mina Rezaei, Kenji Kawaguchi
- Compression Workshop @ NeurIPS 2024
- Readers: Everyone
Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
Xinpeng Wang, Chengzhi Hu, Paul Röttger, Barbara Plank
- ICLR 2025 Poster
- Readers: Everyone
Look at the Text: Instruction-Tuned Language Models are More Robust Multiple Choice Selectors than You Think
Xinpeng Wang, Chengzhi Hu, Bolei Ma, Paul Rottger, Barbara Plank
- COLM
- Readers: Everyone
Look at the Text: Instruction-Tuned Language Models are More Robust Multiple Choice Selectors than You Think
Xinpeng Wang, Chengzhi Hu, Bolei Ma, Paul Röttger, Barbara Plank
- CoRR 2024
- Readers: Everyone
"Seeing the Big through the Small": Can LLMs Approximate Human Judgment Distributions on NLI from a Few Explanations?
Beiduo Chen, Xinpeng Wang, Siyao Peng, Robert Litschko, Anna Korhonen, Barbara Plank
- CoRR 2024
- Readers: Everyone

View all 26 publications

Co-Authors

View all 39 co-authors