Max Pellert

Andrea Cera Palatsi, Samuel Martin-Gutierrez, Ana S. Cardenal et al.

Large language models (LLMs) are increasingly used both to make decisions in domains such as health, education and law, and to simulate human behavior. Yet how closely LLMs mirror actual human decision-making remains poorly understood. This gap is critical: misalignment could produce harmful outcomes in practical applications, while failure to replicate human behavior renders LLMs ineffective for social simulations. Here, we address this gap by developing a digital twin of game-theoretic experiments and introducing a systematic prompting and probing framework for machine-behavioral evaluation. Testing three open-source models (Llama, Mistral and Qwen), we find that Llama reproduces human cooperation patterns with high fidelity, capturing human deviations from rational choice theory, while Qwen aligns closely with Nash equilibrium predictions. Notably, we achieved population-level behavioral replication without persona-based prompting, simplifying the simulation process. Extending beyond the original human-tested games, we generate and preregister testable hypotheses for novel game configurations outside the original parameter grid. Our findings demonstrate that appropriately calibrated LLMs can replicate aggregate human behavioral patterns and enable systematic exploration of unexplored experimental spaces, offering a complementary approach to traditional research in the social and behavioral sciences that generates new empirical predictions about human social decision-making.

Georg Ahnert, Max Pellert, David Garcia et al.

This paper proposes temporally aligned Large Language Models (LLMs) as a tool for longitudinal analysis of social media data. We fine-tune Temporal Adapters for Llama 3 8B on full timelines from a panel of British Twitter users, and extract longitudinal aggregates of emotions and attitudes with established questionnaires. We focus our analysis on the beginning of the COVID-19 pandemic that had a strong impact on public opinion and collective emotions. We validate our estimates against representative British survey data and find strong positive, significant correlations for several collective emotions. The obtained estimates are robust across multiple training seeds and prompt formulations, and in line with collective emotions extracted using a traditional classification model trained on labeled data. We demonstrate the flexibility of our method on questions of public opinion for which no pre-trained classifier is available. Our work extends the analysis of affect in LLMs to a longitudinal setting through Temporal Adapters. It enables flexible, new approaches towards the longitudinal analysis of social media data.

Mats Faulborn, Indira Sen, Max Pellert et al.

Prompt-based language models like GPT4 and LLaMa have been used for a wide variety of use cases such as simulating agents, searching for information, or for content analysis. For all of these applications and others, political biases in these models can affect their performance. Several researchers have attempted to study political bias in language models using evaluation suites based on surveys, such as the Political Compass Test (PCT), often finding a particular leaning favored by these models. However, there is some variation in the exact prompting techniques, leading to diverging findings, and most research relies on constrained-answer settings to extract model responses. Moreover, the Political Compass Test is not a scientifically valid survey instrument. In this work, we contribute a political bias measured informed by political science theory, building on survey design principles to test a wide variety of input prompts, while taking into account prompt sensitivity. We then prompt 11 different open and commercial models, differentiating between instruction-tuned and non-instruction-tuned models, and automatically classify their political stances from 88,110 responses. Leveraging this dataset, we compute political bias profiles across different prompt variations and find that while PCT exaggerates bias in certain models like GPT3.5, measures of political bias are often unstable, but generally more left-leaning for instruction-tuned models. Code and data are available on: https://github.com/MaFa211/theory_grounded_pol_bias

Max Pellert, Jana Lasser, Hannah Metzler et al.

To track online emotional expressions of the Austrian population close to real-time during the COVID-19 pandemic, we build a self-updating monitor of emotion dynamics using digital traces from three different data sources. This enables decision makers and the interested public to assess issues such as the attitude towards counter-measures taken during the pandemic and the possible emergence of a (mental) health crisis early on. We use web scraping and API access to retrieve data from the news platform derstandard.at, Twitter and a chat platform for students. We document the technical details of our workflow in order to provide materials for other researchers interested in building a similar tool for different contexts. Automated text analysis allows us to highlight changes of language use during COVID-19 in comparison to a neutral baseline. We use special word clouds to visualize that overall difference. Longitudinally, our time series show spikes in anxiety that can be linked to several events and media reporting. Additionally, we find a marked decrease in anger. The changes last for remarkably long periods of time (up to 12 weeks). We discuss these and more patterns and connect them to the emergence of collective emotions. The interactive dashboard showcasing our data is available online under http://www.mpellert.at/covid19_monitor_austria/. Our work has attracted media attention and is part of an web archive of resources on COVID-19 collected by the Austrian National Library.