| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Crash narratives dataset | Precision100 | 29 | 1mo ago | ||
| MathEd-PII (test) | Claude 4.5 Opus | Precision93.4 | 14 | 3mo ago | |
| PIIBench corrected (test 5k) | Direct DeBERTa | F1 Score64.76 | 11 | 7d ago | |
| PII-Bench | GLiNER2 Multi | Name F185.2 | 10 | 26d ago | |
| PIIBench 1,398-record (test) | F1 Score13.85 | 8 | 7d ago | ||
| PUPA | F1 Score68.48 | 8 | 1mo ago | ||
| PII-Distract v1 (test) | Entity F1 Score94.8 | 8 | 3mo ago | ||
| PII-Hard v1 (test) | DeepSeekV3 | Entity F1 Score0.893 | 8 | 3mo ago | |
| PII-Multi v1 (test) | Llama3.1 | Entity F1 Score0.942 | 8 | 3mo ago | |
| PII-Single v1 (test) | DeepSeekV3 | Entity F1 Score92.1 | 8 | 3mo ago | |
| PII-Real | DeepSeekV3 | Strict F192.3 | 7 | 3mo ago | |
| Reddit 150 samples (test) | Llama-3.1-8B (FT) | Span Precision86.18 | 7 | 3mo ago | |
| CAPID (test) | Llama-3.1-8B (FT) | Span Precision96.5 | 7 | 3mo ago | |
| SPY (Synthetic PII Yesterday) | urchade/gliner_multi_pii-v1 | Legal Precision52.2 | 5 | 21d ago | |
| CPPB | BodhiPromptShield | Span F192 | 3 | 1mo ago | |
| PIIBench (test) | Direct DeBERTa | F1 Score64.55 | 2 | 7d ago | |
| Visual Redactions (test) | Mask R-CNN | Dice75.83 | 2 | 2mo ago |
