AI RESEARCH
Subject-level Inference for Realistic Text Anonymization Evaluation
arXiv CS.CL
•
ArXi:2604.21211v1 Announce Type: new Current text anonymization evaluation relies on span-based metrics that fail to capture what an adversary could actually infer, and assumes a single data subject, ignoring multi-subject scenarios. To address these limitations, we present SPIA (Subject-level PII Inference Assessment), the first benchmark that shifts the unit of evaluation from text spans to individuals, comprising 675 documents across legal and online domains with novel subject-level protection metrics.