Andreas Waldis

Andreas Waldis

PostDoc · Junior Group Lead · NLP Researcher

CSP Lab, University of Tübingen

About

I’m a Postdoc and Junior Group Lead at the University of Tübingen, where I work with Michael Franke. in the Cognitive Science & Pragmatics group (CSP Lab) . Before moving to Tübingen, I completed my PhD at the UKP Lab (TU Darmstadt), supervised by Prof. Dr. Iryna Gurevych and worked as a Senior Research Associate at the Lucerne University of Applied Sciences and Arts.

My research in natural language processing focuses on interpretability methods, their robustness, and the generalizability of language models. I am particularly interested on how these models understand language and their alignment with societal norms and values for a more reliable model behavior. I enjoy designing extensive experiments that connect multiple perspectives of language models—including model behavior, causal mechanisms, and internal representations—to connect empirical findings with linguistic and cognitive theories to compare and contrast human and model behavior.

During my PhD, I worked on the reliability of language models within computational argumentation, exploring the integration of external knowledge, out-of-distribution generalization, linguistic competence, and societal and safety aspects.

Selected Work

Full list available on Google Scholar.

Events that I'm involved in

  • 2025
    First Workshop on the Interplay of Model Behavior and Model Internals
  • 2025
    The First European CL Research Summit for Junior Researchers · Topic: Bias in NLP and its societal impacts
  • 2024
    First shared task on perspective argument retrieval · ArgMining Workshop @ ACL 2024

Talks

  • 2025
    Assessing the Linguistic Competence of Language Models using Classifier-Based Probing
    Oberseminar of General Linguistics, University of Tübingen
  • 2025
    Why and How We Should Study Internal Representations of LMs
    Data Science Group, Hamburg University
  • 2024
    Linguistic Competence of Language Models and the Role of their Internal Representations
    Department of Computational Linguistics, University of Zurich
  • 2024
    Benchmark the Linguistic Competence of Language Models
    Institute for Natural Language Processing, University of Stuttgart
    Seminar, IBM Research