A catalogue of training datasets per language along with automatic analytics reports.

Full communication and dissemination and exploitation strategy and schedule of activities to implement during the project duration.

Texts (with metadata) used to train the OpenEuroLLM model available at mid-project.

Initial release of LLM models (tokenizers and model weights).

Code package in Python containing model evaluation procedures. The package will be open-sourced on the proposed due date after having iterated on feedback provided by other WPs.

Texts (with metadata) used to train the final OpenEuroLLM model(s).

Written report on strategic advice of the OSPB and community feedback on the development of OpenEuroLLM.

Final release of LLM models (tokenizers and model weights).

Final report on model training, including all necessary details for open publishing and regulatory compliance.

Technical report on the work made in the Evaluation workpackage. It will include our findings in evaluating LLMs on multilingual and regulatory aspects.

Evaluation of the impact of the overall strategy in aspects of dissemination, exploitation and communication.