This repository serves as a central directory of all EMNLP'26 publications by the Webis Group.
The code and data associated with the respective papers are found in the repositories linked below:
- Leandra Fichtel, Janek Prange, and Henning Wachsmuth. Attribute-Based Activation Steering of LLMs for Group-Specific Explanation Generation. [paper available soon] [code]
- Klara M. Gutekunst, Yaron Winter, and Martin Potthast. The Impostors Method as a Statistical Test for Authorship Verification With Large Language Models as Sampling Source for Hard Negatives [paper] [code]
- Maximilian Spliethöver, Jannes Lötfering, Dominik Bachmann, Alexander Espig, Willem Zuidema, Katrin Schulz, Henning Wachsmuth. Cultural (In)Competence: Investigating Culture-Specific Stereotype Detection Capabilities of LLMs [paper available soon] [code]
- Maja Stahl, Timon Ziegenbein, and Henning Wachsmuth. Generating Constructive Feedback on Stories via Reinforcement Learning [preprint] [code]