LLMs4EU – Large Language Models for the European Union
The project fine-tunes large language models for five strategic sectors and makes them available in all EU languages, with a focus on SMEs, public institutions and under-resourced languages.
Lead Investigator: Łukasz Kobyliński, Maciej Ogrodniczuk
- Grant Number
- 101198470
- Start Date
- 1 March 2025
- End Date
- 29 February 2028
Project factsheet
| English name: | Large Language Models for the European Union |
| Project type: | A Digital Europe Programme grant (DIGITAL-2024-AI-B-06) |
| Grant agreement number: | 101198470 |
| Duration: | 1 March 2025 – 29 February 2028 |
| Coordinator: | Alliance for Language Technologies (ALT-EDIC) |
| Polish participation: | Linguistic Engineering Group, Institute of Computer Science, Polish Academy of Sciences |
| Polish PI: | Łukasz Kobyliński, Maciej Ogrodniczuk |
| Project website: | https://www.llms4eu.eu/ |
Project summary
The Large Language Models for the European Union (LLMs4EU) project aims to fine-tune LLMs for specific use cases in five strategic sectors: science, public services, tourism, telecommunications, and energy. The goal is to make models available in all EU languages and to encourage their use by European entities, particularly small and medium-sized enterprises (SMEs) and public institutions. The project’s ambition is to make generative AI inclusive and representative, also for under-resourced languages.
Core activities
- Model development: establishing a framework for multi-model fine-tuning and evaluation.
- Resource accessibility: creating and acquiring diverse datasets to ensure comprehensive language coverage.
- Sector-specific integration: fine-tuning models for five domains (science, public services, tourism, telecommunications, energy) to support industry requirements, capture linguistic nuances, and reinforce each sector’s growth potential.
- Compliance & governance: developing guidelines and resources for data compliance (AI Act, GDPR) and Intellectual Property Rights (IPR) to support the long-term adoption of these technologies.
Consortium
Coordinated by the Alliance for Language Technologies (ALT-EDIC), the consortium brings together 69 partners across 20 countries, including national research centres, innovative technology providers, and highly specialised SMEs. The collaboration between established research excellence and agile industry leaders provides the necessary depth to build a competitive AI ecosystem for all European languages.
Polish partners in the consortium:
- Wrocław University of Science and Technology
- Orange Polska
- Institute of Computer Science, Polish Academy of Sciences
- National Information Processing Institute
The full list of partners is available on the project website.