|
Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
Xinpeng Wang, Chengzhi Hu, Paul Röttger, Barbara Plank
Preprint, 2024
paper /
We propose a surgical and flexible approach to mitigate the false refusal in LLMs with minimal effect on performance and inference cost.
|
|
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
EMNLP Findings, 2024
arxiv /
This study proposes to exploit LLMs to approximate human judgment distributions using a small number of expert labels and explanations on NLI.
|
|
The Potential and Challenges of Evaluating Attitudes, Opinions, and Values in Large Language Models
Bolei Ma*, Xinpeng Wang*, Tiancheng Hu, Anna-Carolina Haensch, Michael A. Hedderich, Barbara Plank, Frauke Kreuter
EMNLP Findings, 2024
arxiv /
We review recent research on evaluating Attitudes, Opinions, and Values in LLMs, highlighting the potential and challenges and offering suggestions for future research.
|
|
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
COLM, 2024
arxiv /
We showed that text answers are more robust than first token answer in instruction-tuned language models, even debiased with SOTA first-token debiasing method.
|
|
"My Answer is C": First-Token Probabilities Do Not Match Text Answers in Instruction-Tuned Language Models
Xinpeng Wang, Bolei Ma, Chengzhi Hu, Leon Weber-Genzel, Paul Röttger, Frauke Kreuter, Dirk Hovy, Barbara Plank
ACL Findings, 2024
arxiv /
We showed that the first-token probability evaluation does not match text answers in instruction-tuned language models.
|
|
ACTOR: Active Learning with Annotator-specific Classification Heads to Embrace Human Label Variation
Xinpeng Wang, Barbara Plank
EMNLP, 2023
arxiv /
We proposed an active learning framework that utilizes a muli-head model to model individual annotators. We designed different acquisition functions and showed our active learning setup achieved performance comparable to full-scale training while saving up to 70% of the annotation budget.
|
|
How to Distill your BERT: An Empirical Study on the Impact of Weight Initialisation and Distillation Objectives
Xinpeng Wang, Leonie Weissweiler, Hinrich Schütze, Barbara Plank
ACL, 2023
arxiv /
code /
We showed that using lower teacher layers for pre-loading student model gives significant performance improvement compared to higher layers.
We also studied the robustness of different distillation objectives under various initialisation choices.
|
|
Xinpeng Wang, Chandan Yeshwanth, Matthias Nießner
3DV, 2021
oral
arxiv /
video /
code /
We proposed a transformer model for scene generation conditioned on room layout and text description.
|
Projects
These include coursework and practical course projects.
|
|
Domain Specific Multi-Lingually Aligned Word Embeddings
Machine Learning for Natural Language Processing Applications
2021-07
report /
|
|
Curiosity Driven Reinforcement Learning
Advanced Deep Learning in Robotics
2021-03
report /
Evaluated and compared the count-based and prediction-based curiosity driven learning on different Atari game environments.
|
|
Introduction to Deep Learning (IN2346)
SS 2020, WS2020/2021
Teaching Assistant
website /
|
|