News

Three papers from our department have been accepted for the  Conference on Language Models (COLM) 2026. Congratulations to all authors!

“When Contextual Inference Fails: Cancelability in Interactive Instruction Following” by
Natalia Bila, Kata Naszadi, Alexandra Mayn, Christof Monz. 
 “Barriers to Universal Reasoning With Transformers (And How to Overcome Them)” by
Oliver Kraus, Yash Sarrof, Yuekun Yao, Alexander Koller, Michael Hahn. 
 “Algebraic Decomposition Theory for Transformer Length… [...]

On May, 15 2026 a new paper from Annemarie Verkerk and Luigi Talamo appeared in "Linguistics": 
"They saw it, onu, 它, coming: an information-theoretic study of cross-linguistic variation in personal pronouns". 

This paper is a SFB 1102 collaboration with  Julius Steuer , now at the Heidelberg Institute for Theoretical Studies, where the surprisal of personal pronouns (Engl: she, you, they, ...; Germ: sie, wir, du, ...) was assessed in seventeen languages from around the world. The authors find…

[...]

We are proud to announce that four papers from our department have won awards in various categories.
Congratulations to all the authors!

Best Paper Award bei CMCL  2026
“Correlating Language Model Surprisal With Cloze and Plausibility: Getting the Best of Both Measures”. By
Kate Rebecca Belcher, Matthew W. Crocker. 
 Honorable Mention Award at  CMCL 2026
“Is Cross-Lingual Transfer in Bilingual Models Human-Like? A Study with Overlapping Word Forms in Dutch and English” By
Iza Škrjanec, Irene… [...]

Two scientific publications from our department were accepted at   Interspeech 2026. Congratulations to both authors!

“Multimodal Jailbreaks for Multimodal LLMs” by
Aravind Krishnan, Karolina Stańcza, Dietrich Klakow"How do word frequency and syllable surprisal affect response time and acoustic duration in sentence formulation?" by
Ivan Yuen, Bernd Möbius, Bistra Andreeva, Mitko Sabev [...]

Numerous scientific publications from our department were accepted at the the International Conference on Machine Learning (ICML)
Congratulations to all authors!

“Provably Learning Attention with Queries” by
Satwik Bhattamishra, Kulin Shah, Michael Hahn, Varun Kanade.“Discovering Interpretable Algorithms by Decompiling Transformers to RASP” by
Xinting Huang, Aleksandra Bakalova, Satwik Bhattamishra, William Merrill, Michael Hahn.“A Framework for Understanding Learnability in Transformers” by
Blanka… [...]

Dr. Annemarie has been appointed to a professorship in ‘Language Science’ in the department Language Science and Technology, effective on 31 May 2026. Since 2019, she has been supported by a junior professorship at Saarland University through the WISNA programme. Prof. Verkerk’s areas of specialisation are linguistic typology and language evolution. Using statistical methods, the Dutch-born researcher examines differences and similarities between languages, taking into account both historical…

[...]

Numerous scientific publications from our department were accepted at the  Cognitive Science Society (CogSci) 2026.
Congratulations to all authors!

"Reflecting the Readers of Today: An Update of the American English Author Recognition Task" by Ana Celmare, Ted Sanders, Merel Scholman"Referring Effectively to a Group of Objects" by Emilia Ellsiepen, Alexandra Mayn, Natalia Bila, Vera Demberg"Expectation-based Linearization: Evidence from German Word-Order and Information Structure" by Torsten Kai… [...]

Numerous scientific publications from our department were accepted at the Association for Computational Linguistics (ACL) 2026
Congratulations to all authors!

Main conference:

“Disco-RAG: Discourse-Aware Retrieval-Augmented Generation” 
by Dongqi Liu, Hang Ding, Qiming Feng, Jian Li, Xurong Xie, Zhucun Xue, Chengjie Wang, Jiangning Zhang, Yabiao Wang."From Weights to Activations: Is Steering the Next Frontier of Adaptation?"
by Simon Ostermann, Daniil Gurgurov, Tanja Baeumel, Michael A.… [...]

The Heinz Maier-Leibnitz Prize is Germany’s most prestigious award for early-career researchers. One of the 10 coveted awards in 2026 goes to Dr. Michael Hahn, a professor of computational linguistics at Saarland University. [...]

Artificial intelligence has become an integral part of everyday (working) life. However, AI-based systems – for example, those used to improve medical diagnoses – have one major shortcoming: they are based on statistical patterns and cannot ‘understand’ how cause and effect are related, unlike humans. Newly appointed professor Sara Magliacane aims to fill this gap with her research: ‘We often don't know exactly what the neural networks in the background of an AI-supported system are calculating,…

[...]