Ontology-Guided Prompting for Zero-Shot Information Extraction: Data, Code, Annotations, and Evaluation Artifacts

Abstract

This record contains the data, code, annotations, and evaluation artifacts for a controlled study of ontology-guided prompting in zero-shot information extraction. The experiment compares a generic extraction prompt with an ontology-guided variant using 20 synthetic single-entity texts constructed for four selected ontology classes, two language models (Qwen3 0.6B and Mistral Small 3.1 24B), and two prompting conditions, resulting in 80 model outputs. The archive includes the synthetic dataset, baseline and ontology-guided prompts, reference extractions, model-generation wrappers and configurations, released raw outputs, attribute-level manual scores, the scoring rubric and annotation interfaces, the class-specific schema inventory, aggregation code, and derived result tables. The package supports inspection of the experimental materials and deterministic recomputation of the aggregate results from the released annotations. The target classes, their attribute inventories, and the natural-language class and attribute descriptions used in the ontology-guided prompts are derived from: Hartley III, D. S. (2021). An Ontology of Modern Conflict: Including Conventional Combat and Unconventional Conflict. Springer International Publishing. https://doi.org/10.1007/978-3-030-53214-7