New preprint: BLiMPPI, a multilingual benchmark for polarity items

BLiMPPI: Benchmark of Linguistic Minimal Pairs for Polarity Items by Arzu Burcu Güven, Anna Rogers, Leonie Weissweiler, Joakim Nivre, and Rob van der Goot is now on arXiv. BLiMPPI contains over 12,000 negative polarity item minimal pairs across 33 languages, built by a pipeline that discovers NPIs in Universal Dependencies treebanks, labels their licensing environments, and perturbs them into minimal pairs.

Jacob Lee Suchardt presented a poster at the 4th UniDive General Meeting

Jacob Lee Suchardt presented a poster on Towards Transcription of (Un-)seen Languages with Speech-to-IPA Models, based on his EMNLP 2025 paper Towards Language-Agnostic STIPA: Universal Phonetic Transcription to Support Language Documentation at Scale, at the 4th UniDive General Meeting, hosted by the Research Institute for Artificial Intelligence of the Romanian Academy (RACAI) in Bucharest.

Leonie Weissweiler took part in the ML4Chem 2026 career round table

Leonie Weissweiler joined the career round table at ML4Chem 2026 — Modern ML in Chemistry, a discussion of research careers inside and outside academia, hosted by the Faculty of Chemistry at Leipzig University.

Leonie Weissweiler presented at the Digital Humanities Open Garden 2026

Leonie Weissweiler gave a talk on Hybrid Human-LLM Corpus Construction and LLM Evaluation for the Caused-Motion Construction at the Digital Humanities Open Garden 2026, hosted by the FDHL at the Sächsische Akademie der Wissenschaften zu Leipzig.

Leonie Weissweiler gave invited talks at Heidelberg University and HITS

Leonie Weissweiler gave an invited talk on "Expanding the Linguistic Evaluation of Language Models, or: What to do now that LMs are good" at the Computational Linguistics Colloquium at Heidelberg University, and at the Heidelberg Institute for Theoretical Studies (HITS).