EN · PII
kiji-en PII detection benchmark results
A frozen evaluation slice with 1,000 rows, 7,623 normalized gold annotations and source-level provenance.
1,000input rows
7,623gold annotations
385,103characters
41measured detectors
Dataset provenance
- Language group
- EN
- Task
- PII
- License
- apache-2.0
- Data categories
- ACCOUNT, ADDRESS, CONTACT, DATE, ID, NET, ORG, PERSON, SECRET
- Project generated
- No
- Corrupted copy
- No
Rows without annotations are not verified clean. Reported extra masking is not a human-confirmed false-positive rate.
Detector outcomes
Sorted by untouched annotations. Training-source overlaps are excluded.
| Detector | Family | Untouched ↓ | Fully hidden ↑ | Extra masking ↓ | Char F1 ↑ |
|---|---|---|---|---|---|
| nym-small | ONNX | 2.26% | 96.09% | — | 0.907 |
| nym-base | HF | 2.74% | 95.36% | — | 0.901 |
| gliner2-fastino | GLINER2 | 3.21% | 96.05% | — | 0.878 |
| gliner-nvidia | GLINER | 3.63% | 93.11% | — | 0.874 |
| pplx | PPLX | 4.07% | 95.74% | — | 0.863 |
| bardsai-eu | ONNX | 4.18% | 89.69% | — | 0.842 |
| gravitee-small | HF | 4.25% | 95.26% | — | 0.923 |
| gliner2-hivetrace-omni | GLINER2 | 4.36% | 95.04% | — | 0.883 |
| openmed-multilingual | HF | 4.83% | 93.31% | — | 0.894 |
| openmed-nemotron | HF | 8.23% | 89.09% | — | 0.856 |
| kalyan-ettin | HF | 8.93% | 86.00% | — | 0.845 |
| gliner-stream-pii | GLINER | 9.94% | 85.82% | — | 0.849 |
| mmbert32k | HF | 12.84% | 77.45% | — | 0.845 |
| gliner2-large | GLINER2 | 13.28% | 85.92% | — | 0.850 |
| apararti | OPF | 14.61% | 84.77% | — | 0.825 |
| openai-base | OPF | 15.19% | 84.36% | — | 0.810 |
| opf-kz-ru | OPF | 16.25% | 82.82% | — | 0.824 |
| traciora | OPF | 20.08% | 76.64% | — | 0.794 |
| opf-ru-v2 | OPF | 23.31% | 73.61% | — | 0.788 |
| nuner-zero | GLINER | 23.84% | 73.68% | — | 0.804 |
| opf-ru | OPF | 24.70% | 69.15% | — | 0.783 |
| gliner25-fastino | GLINER2 | 25.00% | 74.09% | — | 0.802 |
| stanza-ru | STANZA | 27.36% | 70.38% | — | 0.458 |
| gliner2-hivetrace-uni | GLINER2 | 27.56% | 71.43% | — | 0.788 |
| gliner-multi-v21 | GLINER | 30.30% | 67.32% | — | 0.668 |
| ru-pii-ner | RUPII | 32.40% | 66.84% | — | 0.715 |
| gliner-pii-base | GLINER | 36.36% | 62.06% | — | 0.750 |
| davlan-xlmr | HF | 37.56% | 60.04% | — | 0.646 |
| gliner-urchade | GLINER | 38.41% | 60.13% | — | 0.746 |
| ner-ru-gherman | HF | 38.44% | 56.25% | — | 0.591 |
| gliner-pii-edge | GLINER | 39.41% | 59.11% | — | 0.718 |
| davlan-mbert | HF | 42.04% | 56.33% | — | 0.630 |
| gliner2-vladlinv | GLINER2 | 44.79% | 54.87% | — | 0.663 |
| ner-ru-yqelz | HF | 44.94% | 51.65% | — | 0.575 |
| pii-shield-onnx | ONNX | 56.54% | 37.58% | — | 0.637 |
| ru-legal-ner | HF | 63.24% | 28.02% | — | 0.554 |
| spacy-ru-lg | SPACY | 79.34% | 20.54% | — | 0.243 |
| fef2-secret-ru | HF | 84.82% | 13.71% | — | 0.310 |
| natasha | NATASHA | 85.81% | 14.04% | — | 0.203 |
| rules-ru | RULES | 91.87% | 7.84% | — | 0.317 |
| spacy-alrosait | SPACY | 100.00% | 0.00% | — | 0.000 |