Reasoning, or the ability to assess things rationally by applying logic for complex problem-solving or drawing conclusions, is a central component of human intelligence. Recently, large language models (LLMs) have pushed the state-of-the-art in many natural language processing tasks, and researchers have started investigating the reasoning abilities of these models. Although large language models are very good at processing vast amounts of data and recognizing patterns, they still underperform in complex reasoning tasks, such as argumentative reasoning or defeasible reasoning.

In this project, we aim to research novel approaches to address these challenging reasoning tasks. Moreover, we want to explore these reasoning tasks in a multilingual setting to gain insights in the language dependence of large language models in handling complex reasoning tasks, and in their generalizability and proficiency in multilingual contexts.