Get the latest news
Subscribe to receive updates from the School of Data Science.
Postdoctoral Researcher
Hannah Cyberey
Email: yc4dx@virginia.edu
Additional Contact Information
Hannah Cyberey is a postdoctoral research associate in the School of Data Science at the University of Virginia. She currently works in Professor Alex Gates’ Connected Data Hub, studying the global competitiveness of national research infrastructures for emerging AI technologies as part of UVA’s National Security Data and Policy Institute. Her PhD research focuses on trustworthy natural language processing, addressing evaluation reliability, robustness, fairness, and censorship in large language models. She received her PhD in Computer Science from the University of Virginia and her BS in Information Management from Chang Gung University.
Education
- Ph.D. in Computer Science, University of Virginia
- B.S. in Information Management, Chang Gung University
Selected Publications
Hannah Cyberey, Yangfeng Ji, David Evans. Unsupervised Concept Vector Extraction for Bias Control in LLMs. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing (EMNLP), Nov 2025.
Hannah Cyberey, David Evans. Steering the CensorShip: Uncovering Representation Vectors for LLM “Thought” Control. In Conference on Language Modeling (COLM). Oct 2025.
Hannah Cyberey, Yangfeng Ji, David Evans. Do Prevalent Bias Metrics Capture Allocational Harms from LLMs? In Proceedings of the Fifth Workshop on Insights from Negative Results in NLP. May 2025.
Hannah Cyberey, Yangfeng Ji, David Evans. Addressing Both Statistical and Causal Gender Fairness in NLP Models. In Findings of the Association for Computational Linguistics: NAACL 2024. Jun 2024.
Hannah Cyberey, Yangfeng Ji, David Evans. Balanced Adversarial Training: Balancing Tradeoffs Between Oversensitivity and Undersensitivity in NLP Models. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing (EMNLP). Oct 2022.
Hannah Cyberey, Yangfeng Ji, David Evans. Finding Friends and Flipping Frenemies: Automatic Paraphrase Dataset Augmentation Using Graph Theory. In Findings of the Association for Computational Linguistics: EMNLP 2020. Nov 2020.
Hannah Cyberey, Yangfeng Ji, David Evans. Pointwise Paraphrase Appraisal Is Potentially Problematic. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop. Jul 2020.