Blind datasets to test CAMS

A blinded validation corpus consisting of 24 anonymized societal datasets developed for rigorous, out-of-sample testing of the Complex Adaptive Model of Societies (CAMS) v3.2-R framework. The corpus is divided into two balanced sets of 12 datasets each: - Calibration Set (12 datasets): Used for model tuning, threshold setting, and internal consistency checks. - Blind Test Set (12 datasets): Fully withheld during development. Each dataset is stripped of identifying information (society name, year, and geographic markers) so that scorers evaluate the material without prior knowledge of the case. Each dataset includes: - Redacted primary source material spanning multiple decades - Ensemble-scored values for the eight CAMS nodes across the four core metrics (Coherence, Capacity, Stress, Abstraction) All datasets are fully anonymized to support unbiased evaluation. The blind test set remains locked until calibration is complete. Purpose: Independent, reproducible validation of the full CAMS JUNO diagnostic framework under blinded conditions.