News

I took part in 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.

Invited 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 Saxon Academy of Sciences and Humanities in Leipzig.

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

I'm going to be a tutorial chair at COLING 2027 in Macau, together with Michael Roth. We can be reached at coling27tutorialchairs@googlegroups.com.

Invited talk on Computational Typology? in the Department of Linguistic and Cultural Evolution at the Max Planck Institute for Evolutionary Anthropology in Leipzig.

Our paper Language Models Learn Constructional Semantics, Not To Mention Syntax was accepted to CoNLL 2026.

I've started a new position as Assistant Professor for Natural Language Processing at the University of Leipzig Institute for Computer Science and started my own research group, the LION Lab (Linguistically-Oriented NLP).

Invited talk at Stockholm University on "Comprehensively Evaluating Language in Language Models".

Invited talk at Gothenburg University on "Comprehensively Evaluating Language in Language Models".

Our paper MultiBLiMP 1.0: A Massively Multilingual Benchmark of Linguistic Minimal Pairs has been published at TACL.

Online invited talk at the University of Vienna on "Comprehensively Evaluating Language in Language Models".

Invited talk at ITU Copenhagen on "Comprehensively Evaluating Language in Language Models".

Our paper Constructions are Revealed in Word Distributions was featured as a Senior Area Chair Highlight at EMNLP!

Our paper Hybrid Human-LLM Corpus Construction and LLM Evaluation for the Caused-Motion Construction has been published in NEJLT.

I joined the European Laboratory for Learning and Intelligent Systems (ELLIS) e.V. as a member.

Invited talk at Bochum University on "Constructions all the way down: rethinking linguistic generalisation in LLMs".

I was a co-organiser of the Second International Workshop on Construction Grammars and NLP, co-located with the 16th International Conference on Computational Semantics (IWCS) in Düsseldorf.

I started a postdoc at Uppsala University Computational Linguistics, working with Joakim Nivre on topics related to constructions and multilinguality, with continued funding from the Walter Benjamin Fellowship of the German Research Foundation (DFG).

Invited talk at JKU Linz on "Comprehensively Evaluating Language in Language Models".

Our paper MultiBLiMP 1.0: A Massively Multilingual Benchmark of Linguistic Minimal Pairs was accepted to TACL and will be presented at EMNLP 2025.

Two papers accepted to EMNLP 2025! Constructions are Revealed in Word Distributions and BabyLM's First Constructions: Causal probing provides a signal of learning.

Our paper Both Direct and Indirect Evidence Contribute to Dative Alternation Preferences in Language Models was accepted to COLM 2025.

Invited talk at TUM Heilbronn on "Comprehensively Evaluating Language in Language Models".

Our paper Linguistic generalizations are not rules: Impacts on evaluation of LMs was accepted to the Second International Workshop on Construction Grammars and NLP.

Invited talk at Copenhagen University on "Constructions all the way down: rethinking linguistic generalisation in LLMs".

Invited talks at LMU Munich and Leipzig University on "Understanding Language in Language Models: The Role of Constructions in LM Interpretability".

Invited talks at Stanford University and the University of Nevada at Reno on the topic of "Constructions all the way down: rethinking linguistic generalisation in LLMs".

I received a Short Term Scientific Mission (STSM) grant through the UniDive COST Action to visit Uppsala University for one month from May to June 2025.

Our paper Derivational Morphology Reveals Analogical Generalization in Large Language Models was accepted to PNAS.

My DFG Walter Benjamin Fellowship proposal A Computational Exploration of Construction Learning was approved!

Our paper Models Can and Should Embrace the Communicative Nature of Human-Generated Math was accepted at the NEURIPS 2024 workshop on Mathematical Reasoning and AI!

Our paper SYNTHEVAL: Hybrid Behavioral Testing of NLP Models with Synthetic CheckLists was accepted to Findings of EMNLP 2024.

I started my postdoc at UT Austin Linguistics with Kyle Mahowald!

I gave a keynote at KONVENS 2024 on "Constructions all the way down: Rethinking compositionality in LLMs". The slides are available here.

I was a co-organiser of the Sixth Workshop on Teaching NLP, co-located with ACL 2024 in Bangkok.

On July 3rd, I successfully defended my PhD thesis "Computational Approaches to Construction Grammar and Morphology" and graduated summa cum laude with a Dr.Phil. in Computational Linguistics! A recording of the 15-minute defence talk is available here.

Invited talks at NYU, Boston University, MIT, UPenn, CMU, and Cornell on the topic of "Finding the Limits of LLMs with Constructions".

I will be visiting Adele Goldberg at Princeton University for 3 months, starting in April 2024, to work on Construction Grammar and LLMs!

Three papers accepted to LREC-COLING 2024! UCxn: Typologically Informed Annotation of Constructions Atop Universal Dependencies, Constructions Are So Difficult That Even Large Language Models Get Them Right for the Wrong Reasons and Verbing Weirds Language (Models): Evaluation of English Zero-Derivation in Five LLMs.

Invited talk at the SAIL workshop on Fundamental Limits of Large Language Models at the University of Bielefeld. Check out the video here!

Two papers accepted to EMNLP 2023! Counting the Bugs in ChatGPT's Wugs: A Multilingual Investigation into the Morphological Capabilities of a Large Language Model and Crosslingual Transfer Learning for Low-Resource Languages Based on Multilingual Colexification Graphs.

Invited talk at the "Poetik, Praxis und Hermeneutik Künstlicher Intelligenz" workshop at the Fritz Thyssen Foundation in Cologne, on LLMs as the newest battlefield of the Linguistics Wars, featuring opinions by Noam Chomsky, Adele Goldberg, Steve Piantadosi and myself. Slides (in German) are available here.

Invited talk at Bar Ilan University NLP.

Two papers accepted to ACL 2023! A Crosslingual Investigation of Conceptualization in 1335 Languages and How to Distill your BERT: An Empirical Study on the Impact of Weight Initialisation and Distillation Objectives.

I presented The Better Your Syntax, the Better Your Semantics? Probing Pretrained Language Models for the English Comparative Correlative at the 19th Workshop on Multiword Expressions at EACL. Check out the slides!

Invited talk at Ben Roth's Lab at the University of Vienna.

Invited talk at Ryan Cotterell's Lab at ETH Zürich on construction grammar and how we can use it to formulate new goals for syntactic and semantic probing.

Invited talk at MunichNLP on the joint history of NLP and Linguistics and how we can learn from each other going forward.

Our paper Construction Grammar Provides Unique Insight into Neural Language Models was accepted to the CxGs+NLP Workshop at the Georgetown University Round Table 2023! Check it out here.

Our paper The Better Your Syntax, the Better Your Semantics? Probing Pretrained Language Models for the English Comparative Correlative was accepted to EMNLP 2022! Check it out here.

Invited at the 8th International ScaDS Summer School 2022 in Leipzig on the joint history of NLP and Linguistics and how we can learn from each other going forward! Here are the abstract and slides.

I will be visiting David Mortensen and Lori Levin's Linguistics Lab at the Language Technologies Institute at Carnegie Mellon University for 3 months, starting in mid-July 2022, to work on Construction Grammar and NLP!

I presented our paper CaMEL: Case Marker Extraction without Labels 🐫 at ACL 2022 in Dublin! Click here for the slides.

Our paper CaMEL: Case Marker Extraction without Labels 🐫 was accepted to ACL 2022! Check it out here.