2025W

• Experiment Design and Execution / 194.192 / VU

• Understanding Verbatim Memorization in LLMs Through Circuit Discovery / Lasy, I., Knees, P., & Woltran, S. (2025). Understanding Verbatim Memorization in LLMs Through Circuit Discovery. arXiv. https://doi.org/10.48550/ARXIV.2506.21588
• Understanding Verbatim Memorization in LLMs Through Circuit Discovery / Lasy, I., Knees, P., & Woltran, S. (2025). Understanding Verbatim Memorization in LLMs Through Circuit Discovery. In R. Jia, E. Wallace, Y. Huang, T. Pimentel, P. Maini, V. Dankers, J. T.-Z. Wei, & P. Lesci (Eds.), Proceedings of the First Workshop on Large Language Model Memorization (L2M2) (pp. 83–94). Association for Computational Linguistics. https://doi.org/10.18653/v1/2025.l2m2-1.7
  Project: BILAI (2024–2029)
• TU Wien at SemEval-2024 Task 6: Unifying Model-Agnostic and Model-Aware Techniques for Hallucination Detection / Arzt, V., Azarbeik, M. M., Lasy, I., Kerl, T., & Recski, G. (2024). TU Wien at SemEval-2024 Task 6: Unifying Model-Agnostic and Model-Aware Techniques for Hallucination Detection. In A. K. Ojha, A. S. Dogruöz, H. Tayyar Madabushi, G. Da San Martino, S. Rosenthal, & A. Rosá (Eds.), Proceedings of the 18th International Workshop on Semantic Evaluation (SemEval-2024) (pp. 1183–1196). Association for Computational Linguistics. https://doi.org/10.18653/v1/2024.semeval-1.173

• Evaluation of sparse autoencoder-based refusal features in LLMs : a dataset-dependence study / Kerl, T. (2025). Evaluation of sparse autoencoder-based refusal features in LLMs : a dataset-dependence study [Diploma Thesis, Technische Universität Wien]. reposiTUm. https://doi.org/10.34726/hss.2025.130185
  Download: PDF (1.9 MB)