RU · PII
rubai-ru PII detection benchmark results
A frozen evaluation slice with 1,500 rows, 3,458 normalized gold annotations and source-level provenance.
1,500input rows
3,458gold annotations
362,298characters
49measured detectors
Dataset provenance
- Language group
- RU
- Task
- PII
- License
- apache-2.0
- Data categories
- ADDRESS, CONTACT, ID, PERSON
- 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 fully hidden annotations. Training-source overlaps are excluded.
| Detector | Family | Fully hidden ↑ | Detected / any overlap ↑ | Extra masking ↓ | Char F1 ↑ |
|---|---|---|---|---|---|
| openai-base | OPF | 26.17% | 93.18% | — | 0.924 |
| bardsai-eu | ONNX | 18.16% | 94.82% | — | 0.783 |
| pplx | PPLX | 10.70% | 99.54% | — | 0.941 |
| ru-pii-ner | RUPII | 8.01% | 94.88% | — | 0.923 |
| gliner25-fastino | GLINER2 | 7.92% | 97.95% | — | 0.957 |
| gliner25-fastino-ru | GLINER2 | 7.87% | 97.83% | — | 0.961 |
| nuner-zero | GLINER | 7.55% | 95.37% | — | 0.894 |
| gliner2-fastino | GLINER2 | 7.29% | 69.38% | — | 0.633 |
| gliner2-hivetrace-omni | GLINER2 | 7.26% | 74.58% | — | 0.653 |
| gliner2-hivetrace-omni-ru | GLINER2 | 7.23% | 72.04% | — | 0.637 |
| gliner2-fastino-ru | GLINER2 | 7.17% | 68.57% | — | 0.635 |
| gliner-urchade | GLINER | 7.17% | 78.48% | — | 0.748 |
| gliner-nvidia | GLINER | 7.17% | 80.36% | — | 0.764 |
| traciora | OPF | 7.11% | 95.89% | — | 0.924 |
| gliner2-vladlinv | GLINER2 | 6.82% | 80.19% | — | 0.861 |
| ru-legal-ner | HF | 6.74% | 89.97% | — | 0.697 |
| gliner2-vladlinv-ru | GLINER2 | 6.62% | 77.01% | — | 0.850 |
| gliner2-large | GLINER2 | 6.45% | 89.59% | — | 0.750 |
| gliner-pii-edge | GLINER | 6.33% | 75.53% | — | 0.661 |
| gliner-pii-base | GLINER | 6.13% | 63.79% | — | 0.534 |
| gliner-nvidia-ru | GLINER | 6.04% | 76.08% | — | 0.743 |
| gliner-multi-v21 | GLINER | 6.04% | 86.41% | — | 0.760 |
| apararti | OPF | 5.81% | 94.94% | — | 0.928 |
| fef2-secret-ru | HF | 5.81% | 73.11% | — | 0.430 |
| mmbert32k | HF | 5.67% | 98.38% | — | 0.651 |
| opf-ru | OPF | 5.64% | 97.60% | — | 0.639 |
| gliner-multi-v21-ru | GLINER | 5.64% | 83.05% | — | 0.781 |
| opf-ru-v2 | OPF | 5.44% | 94.53% | — | 0.919 |
| stanza-ru | STANZA | 5.44% | 62.17% | — | 0.348 |
| nym-base | HF | 5.29% | 98.38% | — | 0.730 |
| pii-shield-onnx | ONNX | 5.12% | 88.29% | — | 0.845 |
| nym-small | ONNX | 5.00% | 98.26% | — | 0.716 |
| gliner2-hivetrace-uni | GLINER2 | 4.80% | 49.02% | — | 0.342 |
| ner-ru-yqelz | HF | 4.51% | 56.28% | — | 0.457 |
| davlan-xlmr | HF | 4.45% | 55.21% | — | 0.417 |
| natasha | NATASHA | 4.34% | 53.70% | — | 0.300 |
| opf-kz-ru | OPF | 4.22% | 93.00% | — | 0.915 |
| gliner-stream-pii | GLINER | 4.19% | 85.83% | — | 0.777 |
| spacy-ru-lg | SPACY | 4.05% | 52.37% | — | 0.300 |
| davlan-mbert | HF | 4.05% | 55.12% | — | 0.408 |
| openmed-multilingual | HF | 3.96% | 92.74% | — | 0.633 |
| gravitee-small | HF | 3.59% | 61.57% | — | 0.595 |
| gliner2-hivetrace-uni-ru | GLINER2 | 3.59% | 32.07% | — | 0.225 |
| gliner-urchade-ru | GLINER | 3.09% | 67.84% | — | 0.732 |
| openmed-nemotron | HF | 2.89% | 84.47% | — | 0.664 |
| kalyan-ettin | HF | 2.69% | 65.12% | — | 0.446 |
| spacy-alrosait | SPACY | 2.66% | 50.43% | — | 0.628 |
| ner-ru-gherman | HF | 1.76% | 55.32% | — | 0.655 |
| rules-ru | RULES | 0.17% | 3.44% | — | 0.051 |