ALE: Context fragments supplied to the generative AI assistant of the ALE intelligent environment (D6)

Samples of the context supplied to the generative AI assistant of the ALE intelligent environment for agricultural advisory services under condition C2 of the evaluation protocol D9 (10.5281/zenodo.23879795): 128 contexts (32 queries × budgets of 200, 400, 600 and 800 tokens) built by the context module of the ALE platform 0.3.0 from the open ALE knowledge base - the ALE ontology 1.0.0 with the AKIS and AGROVOC modules (10.5281/zenodo.23643625) and synthetic test data. For each context the dataset gives the query category (K1-K4), entry nodes, selected facts (also as RDF N-Quads, one named graph per context), constraints (allowed registered uses, expired registrations, insufficient data), removed personal data, sources, the context text passed to the model, its size in tokens and the building time. Queries come from the published acceptance cases, competency questions D8 and the D9 corpus example. The knowledge base is synthetic; the dataset shows the format and behaviour of context building, not its quality on real data. Terminology: ALE stands for Advisory Learning Ecosystem (the ontology); the software system is called the ALE intelligent environment. Funding: Ministry of Education and Science of Ukraine, project "Development of an Intelligent Support Environment for Agricultural Advisory Services Based on the Integration of Semantic Wiki Technologies and Artificial Intelligence" (state registration No. 0126U000765).