The development of Large Language
Models (LLMs) such as GPT has led to
significant advances in automatic content
generation. However, their application in
specific contexts, such as the creation of an
Algerian encyclopedic dictionary, requires
finetuning the models to ensure the
relevance, accuracy, and coherence of the
generated content.
This project is part of an explainable AI
approach, where mechanisms are
developed to structure the generated
information and validate it in collaboration
with experts. The system aims to enrich the
Algerian cultural heritage database while
integrating modern natural language
processing techniques to ensure
consistency and reliability in content
generation.