A Dataset of Cultural Heritage Manipulation on English Wikipedia in the Russo–Ukrainian Context

This dataset contains revision-level data collected from culturally and politically salient English Wikipedia articles related to the Russo–Ukrainian context. The resource is designed to support the study of cultural heritage manipulation, narrative reframing, and subtle discursive interventions in collaborative knowledge systems. The Zenodo record contains two CSV files: silver_dataset.csv — a large-scale dataset containing 69,908 Wikipedia revision pairs annotated with LLM-generated silver labels (Weaponized / Not Weaponized). Each row contains a revision and its preceding version, enabling fine-grained analysis of textual changes and narrative shifts. gold_dataset.csv — an expert-validated subset containing 6,922 revision pairs manually reviewed for annotation quality and consistency. This subset is intended for evaluation and benchmarking purposes. Both datasets include revision metadata such as timestamps, editor type (registered user, anonymous IP editor, or bot), edit comments, size deltas, and textual diffs between revisions. User identifiers have been anonymized prior to release. The dataset supports research in information retrieval, natural language processing, computational social science, digital humanities, and the analysis of narrative manipulation in large-scale collaborative platforms.