Aleph-Alpha-GermanWeb: Improving German-language LLM pre-training with model-based data curation and synthetic data generation.

Thomas F. Burns, Letitia Parcalabescu, Stephan Wldchen, Michael Barlow, Gregor Ziegltrum, Volker Stampa, Bastian Harren, Bjrn Deiseroth

Browse the full EACL paper archive.