Inspired by our recent LLäMmlein series we created a new German state of the art encoder model family in cooperation with the Computerphilologie!
Mehr
Our LLäMmlein paper got accepted for ACL 25! See you in Vienna!
Mehr
Im Interview mit heise.de spricht Prof. Dr. Andreas Hotho über die Herausforderungen bei der Entwicklung deutscher Sprachmodelle und stellt das komplett auf deutschen Daten trainierte Sprachmodell LLämMlein vor.
Mehr
We’re excited to share that our paper, “Exploring Design Choices for Autoregressive Deep Learning Climate Models,” has been accepted to the Tackling Climate Change with Machine Learning (TCCML) Workshop at ICLR 2025!
Mehr
We participated in two SemEval tasks and placed first in SemEval'25 Task 2!! OTTERly fantastic!
Mehr
Our paper "Assessing the State of the Art in Scene Segmentation" has been accepted at the 2025 Annual Conference of the Nations of the Americas Chapter of the ACL. In our paper, we assess the state of the art for scene segmentation for german novels and compare BERT-based to current Llama models.
Mehr
Our paper "Modeling and Analyzing the Influence of Non-Item Pages on Sequential Next-Item Prediction" just got accepted at the ACM Journal "Transactions on Recommender System". In our work, we introduce and analyze the utility of non-item interactions for sequential recommender systems.
Mehr
We created two German-only decoder models, LLäMmlein 120M and 1B, from scratch!
Mehr
Janna Omeliyanenko just gave a keynote on "Efficient Language Models as Continual Knowledge Bases" at the Workshop on Knowledge Base Construction from Pre-Trained Language Models at the ISWC 2024.
Mehr