Wenyang Tang

• yang.t.skywalker [at] gmail.com
• notoookay
• @wenyang_t

Hi 👋, I'm a graduate student specializing in Natural Language Processing (NLP), with a particular focus on improving the factuality and reliability of Large Language Models. My research aims to advance how AI systems can generate truthful and verifiable information, making them more reliable and trustworthy for real-world applications.

My research interests include:

1. Hallucination Detection and LLM Factuality: My work focuses on developing methods to detect and prevent language models from generating false or unsupported information, while improving their ability to produce factual and verifiable outputs. This includes designing novel evaluation frameworks and techniques to enhance the truthfulness and reliability of large language models.
2. Retrieval-Augmented Language Models: I'm exploring ways to enhance the performance of large language models by efficiently retrieving and incorporating relevant information from external knowledge sources.
3. Reinforcement Learning: Drawing parallels to human learning processes, I'm also interested in applying reinforcement learning techniques to NLP tasks, particularly in scenarios where continuous adaptation and learning from feedback are crucial.

Feel free to reach out if you're interested in discussing NLP, reinforcement learning, or potential collaborations.

Publications

• Internal Activations Reveal LLMs’ Context Faith: A Linear Direction for Context Distrust Control

  Fengwen Zhai, Wenyang Tang, Jing Jin, Xiangde Jiang, Zhicheng Han, Hu Han

  KnowFM @ ACL 2026 Workshop (non-archival)

  [paper coming soon] [code coming soon]

  [citation]

  @inproceedings{
    zhai2026internal,
    title={Internal Activations Reveal {LLM}s{\textquoteright} Context Faith: A Linear Direction for Context Distrust Control},
    author={Fengwen ZHAI and Wenyang Tang and Jin Jing and Xiangde Jiang and Zhicheng Han},
    booktitle={4th Workshop on Towards Knowledgeable Foundation Models at ACL 2026},
    year={2026},
    url={https://openreview.net/forum?id=OarPb3fqy0}
  }

• RAG-LER: Ranking Adapted Generation with Language-model Enabled Regulation

  Fengwen Zhai, Wenyang Tang, Jing Jin

  Neurocomputing, 2025

  [paper] [code]

  [citation]

  @article{ZHAI2025131514,
    title = {RAG-LER: Ranking adapted generation with language-model enabled regulation},
    journal = {Neurocomputing},
    volume = {656},
    pages = {131514},
    year = {2025},
    issn = {0925-2312},
    doi = {https://doi.org/10.1016/j.neucom.2025.131514},
    url = {https://www.sciencedirect.com/science/article/pii/S0925231225021861},
    author = {Fengwen Zhai and Wenyang Tang and Jing Jin},
    keywords = {Language modeling, Retrieval augmented generation, Information retrieval, Re-ranking}
  }