Georg Ahnert

PhD Student, University of Mannheim

ahnert [at] uni-mannheim.de

github.com/wanlo

@wanlo.bsky.social

Hi, I'm Georg!

I‘m a PhD student at the Chair for Data Science in the Economic and Social Sciences at University of Mannheim under the supervision of Markus Strohmaier and Florian Keusch. I previously obtained a M.Sc. in Computational Social Systems from RWTH Aachen University. I am passionate about social data science, natural language processing, and fairness in machine learning. Currently, I‘m thinking about how to use large language models in survey research.

Selected Publications

  • QSTN: A Modular Framework for Robust Questionnaire Inference with LLMs 
    Maximilian Kreutner, Jens Rupprecht, Georg Ahnert, Ahmed Salem, & Markus Strohmaier (2025). 
    » preprint » code » live demo
  • Survey Response Generation: Generating Closed-Ended Survey Responses In-Silico with Large Language Models 
    Georg Ahnert, Anna-Carolina Haensch, Barbara Plank, & Markus Strohmaier (2025). 
    » preprint
  • Connecting Natural Language Processing and Survey Methodology: Potentials, Challenges, and Open Questions 
    Indira Sen, Bolei Ma, Georg Ahnert, Anna-Carolina Haensch, Tobias Holtdirk, Frauke Kreuter, & Markus Strohmaier (2025). 
    » preprint
  • The Prompt Makes the Person(a): A Systematic Evaluation of Sociodemographic Persona Prompting for Large Language Models 
    Marlene Lutz, Indira Sen, Georg Ahnert, Elisa Rogers, & Markus Strohmaier (2025). EMNLP Findings 2025 » paper
  • Simulating Persuasive Dialogues on Meat Reduction with Generative Agents 
    Georg Ahnert, Elena Wurth, Markus Strohmaier, & Jutta Mata (2025). NLPSI workshop @ ICWSM'25 
    » paper » code
  • Extracting Affect Aggregates from Longitudinal Social Media Data with Temporal Adapters for Large Language Models 
    Georg Ahnert, Max Pellert, David Garcia, & Markus Strohmaier (2025). ICWSM'25 
    » paper » code » poster at IC2S2‘24
  • Identifying Themes of Right-Wing Extremism in Hindutva Discourse on Twitter 
    Piyush Ghasiya, Georg Ahnert, & Kazutoshi Sasahara (2023). Social Media + Society, 9(3) 
    » paper » data » poster at IC2S2‘23 » covered on Newsweek Japan

For more publications, see Semantic Scholar or Google Scholar.

Education

  • Research Stay at the SODA lab —
    With Frauke Kreuter & Anna-Carolina Haensch —
    LMU Munich, 2025
  • Computations Social Systems, M.Sc. (with distinction) —
    RWTH Aachen University, 2020–2023 —
    Master's thesis on Fairness-Aware Ranking Recovery from Pairwise Comparisons —
    Winner of the GOR'24 Thesis Award
  • Student Assistant at the Department for Computational Social Science —
    With Claudia Wagner & David Schoch —
    GESIS Leibniz Institute for the Social Sciences, 2023
  • Research Internship at the Computational Social Science Lab —
    With Kazutoshi Sasahara & Piyush Ghasiya —
    Tokyo Institute of Technology, 2022
  • Computer Science, B.Sc. (with distinction) —
    RWTH Aachen University, 2017–2020 —
    Bachelor's thesis on Human Perceptions of Algorithmic Fairness

Teaching

  • Social Media Data Analysis
    Summer 2026, lead by Indira Sen, and together with Maximilian Kreutner 
  • Large Language Models for the Economic and Social Sciences
    Winter 2025, lead by Indira Sen, and together with Abigail Hayes 
    » course webpage
  • Scientific Programming with Python
    2023–2025, weekly lectures and exercises 
    » course webpage
  • Team Project Supervision
    Topic for winter 2025: Building and Benchmarking Historical LLMs
    Topic for summer 2025: Self-Organization Among Large Language Model Agents
  • Master's Theses Supervision
    Selected topics supervised so far:
    » Comparing Life Satisfaction Scales for Synthetic Respondents: An LLM Approach
    » Feature Selection by LLMs for the Generation of Synthetic Survey Responses
    » Human Survey Response and Prompt Perturbation Biases in Large Language Models
    » GuardBench: Assessing Guardrails of Large Language Models
    » A User Interface for Interaction Studies with Large Language Models

Projects

  • Values to Country, a webapp that lets users find the country closest to their personal values built upon data from the World Value Survey 
    » try it yourself » code
  • GESIS WebBot, a browser plugin that automates data retrieval from various search engines for the department for Computational Social Science at GESIS 
    » code and documentation
  • Vogelnest, a user management webapp with LDAP integration for the student's initiative Studies Without Borders 
    » code