[2601.09313] Understanding or Memorizing? A Case Study of German Definite Articles in Language Models

Computer Science > Computation and Language arXiv:2601.09313 (cs) [Submitted on 14 Jan 2026 (v1), last revised 14 Apr 2026 (this version, v2)] Title:Understanding or Memorizing? A Case Study of German Definite Articles in Language Models Authors:Jonathan Drechsel, Erisa Bytyqi, Steffen Herbold View a PDF of the paper titled Understanding or Memorizing? A Case Study of German Definite Articles in Language Models, by Jonathan Drechsel and 2 other authors View PDF HTML (experimental) Abstract:Language models perform well on grammatical agreement, but it is unclear whether this reflects rule-based generalization or memorization. We study this question for German definite singular articles, whose forms depend on gender and case. Using GRADIEND, a gradient-based interpretability method, we learn parameter update directions for gender-case specific article transitions. We find that updates learned for a specific gender-case article transition frequently affect unrelated gender-case settings, with substantial overlap among the most affected neurons across settings. These results argue against a strictly rule-based encoding of German definite articles, indicating that models at least partly rely on memorized associations rather than abstract grammatical rules. Comments: Subjects: Computation and Language (cs.CL); Artificial Intelligence (cs.AI) Cite as: arXiv:2601.09313 [cs.CL]   (or arXiv:2601.09313v2 [cs.CL] for this version)   https://doi.org/10.48550/arXiv.2601.09313 Focus to l...