●researcherThe two annotated datasets with rationale labels fill a gap for training and evaluating explainable multimodal content moderation models.
●policyExplainability-linked detection outputs could support audit requirements for platform content moderation under emerging regulatory frameworks.