- Acronym:
- HIVE AI
- Full name:
- HIVE AI: Development and pilot implementation of large language models in Polish public administration
- Logotyp:
-
- Duration:
- 20.02.2025 - 31.12.2025
- Aim:
The aim of the Project was to support the digitalization and automation of repetitive official activities in Poland through the pilot implementation of Polish Large Language Models (LLMs) into public administration systems. The task was a direct continuation of the project entitled Polish Large Language Model (PLLuM): it assumed the development of solutions and resources produced, and at the same time a transition from the research phase to the implementation phase.
As part of the work on the data corpus, the ACC Cyfronet AGH team was focus on the process of acquiring, cleaning and evaluating training data for the needs of pre-training and tuning processes. The ACC Cyfronet AGH team provided infrastructure and expert support in the field of data processing and configuration and maintenance of the environment for the needs of the task.
- Contact:
- Łukasz Flis, e-mail: lukasz.flis (at) cyfronet.pl, phone: (+48) 12-632-33-55 ext. 603
CO-FINANCED FROM THE STATE BUDGET
- Type of budget subsidy or name of program or fund:
- Targeted subsidy agreement no. 1/WII/DBI/2025
- Task name:
- HIVE AI: Development and pilot implementation of large language models in Polish public administration
- Funding value:
- 18 076 638,75 zł
- Total value of the task:
- 18 076 638,75 zł
- Brief description of the task:
- The aim of the Project was to support the digitalization and automation of repetitive official activities in Poland through the pilot implementation of Polish Large Language Models (LLMs) into public administration systems. The task was a direct continuation of the project entitled Polish Large Language Model (PLLuM): it assumed the development of solutions and resources produced, and at the same time a transition from the research phase to the implementation phase.
- Date of signing the agreement on financing or co-financing the task:
- 25.03.2025
