MaiNLP

MaiNLP research lab

"Standard-to-dialect transfer trends differ across text and speech: A case study on intent and topic classification in German dialects" (Blaschke, Winkler & Plank, ACL 2026)

CLIMATELI: Evaluating Entity Linking on Climate Change Data

We introduce TruthQuest, a benchmark designed to evaluate the suppositional reasoning capabilities of large language models through knights and knaves puzzles.

In this project, we evaluate inferential strategies employed by large language models in propositional logic problems and compare them to strategies observed in humans.

A unified framework for attributing model behavior to model components, training data, and training dynamics.

Implementation and Outputs of the ACL 2026 Findings paper - Decoupling the Effect of Chain-of-Thought Reasoning: A Human Label Variation Perspective

Repository for paper MAKIEval: A Multilingual Automatic WiKIdata-based Framework for Cultural Awareness Evaluation for LLMs

Dataset and Implementation of the ACL 2026 Findings paper "Agree, Disagree, Explain: Decomposing Human Label Variation in NLI through the Lens of Explanations"