"RMF-PC Reproducibility Artifacts for Auditable Evaluation of Spam Detection under Bounded Paraphrase Attacks"

Derived reproducibility artifacts supporting the RMF-PC spam-detection study. The archive includes data audit outputs, opaque split and prediction manifests, content hashes, reported metrics, cross-corpus results, confidence intervals, paired statistical tests, and aggregate L2/L3 human-validation outcomes.</p><p>Raw email content, source archives, attack rewrite text, reviewer-level forms, model weights, and original identifiers are excluded. See README and the third-party data notice in the archive.