MULTI · PII
dialogpii-multi PII detection benchmark results
A frozen evaluation slice with 500 rows, 10,345 normalized gold annotations and source-level provenance.
500input rows
10,345gold annotations
1,964,464characters
41measured detectors
Dataset provenance
- Language group
- MULTI
- Task
- PII
- License
- cc-by-4.0
- Data categories
- ADDRESS, CONTACT, ID, NET, ORG, 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 untouched annotations. Training-source overlaps are excluded.
| Detector | Family | Untouched ↓ | Fully hidden ↑ | Extra masking ↓ | Char F1 ↑ |
|---|---|---|---|---|---|
| pii-shield-onnx | ONNX | 17.76% | 74.31% | — | 0.119 |
| gliner2-fastino | GLINER2 | 22.10% | 71.70% | — | 0.538 |
| bardsai-eu | ONNX | 22.94% | 64.74% | — | 0.718 |
| gliner-urchade | GLINER | 24.93% | 70.47% | — | 0.459 |
| gliner2-hivetrace-omni | GLINER2 | 26.04% | 66.34% | — | 0.604 |
| nym-base | HF | 27.64% | 55.49% | — | 0.696 |
| pplx | PPLX | 29.16% | 65.15% | — | 0.679 |
| gliner-multi-v21 | GLINER | 29.91% | 64.06% | — | 0.458 |
| davlan-xlmr | HF | 30.58% | 60.63% | — | 0.680 |
| mmbert32k | HF | 31.58% | 49.71% | — | 0.485 |
| nuner-zero | GLINER | 31.80% | 62.71% | — | 0.443 |
| nym-small | ONNX | 32.04% | 51.45% | — | 0.667 |
| gliner-nvidia | GLINER | 32.45% | 57.27% | — | 0.479 |
| ner-ru-yqelz | HF | 34.72% | 56.33% | — | 0.644 |
| davlan-mbert | HF | 35.40% | 54.99% | — | 0.635 |
| gliner25-fastino | GLINER2 | 35.98% | 59.15% | — | 0.560 |
| gliner2-large | GLINER2 | 39.32% | 54.87% | — | 0.433 |
| opf-ru | OPF | 40.42% | 44.62% | — | 0.359 |
| gliner-pii-base | GLINER | 41.62% | 53.31% | — | 0.532 |
| gliner2-hivetrace-uni | GLINER2 | 42.81% | 49.21% | — | 0.511 |
| openmed-multilingual | HF | 43.08% | 40.57% | — | 0.527 |
| gliner-pii-edge | GLINER | 43.11% | 51.06% | — | 0.537 |
| ner-ru-gherman | HF | 43.75% | 40.02% | — | 0.520 |
| ru-legal-ner | HF | 44.20% | 45.39% | — | 0.111 |
| stanza-ru | STANZA | 45.04% | 49.77% | — | 0.164 |
| gliner2-vladlinv | GLINER2 | 47.31% | 47.13% | — | 0.550 |
| openmed-nemotron | HF | 48.70% | 34.19% | — | 0.308 |
| gliner-stream-pii | GLINER | 52.78% | 39.76% | — | 0.518 |
| gravitee-small | HF | 53.99% | 39.35% | — | 0.249 |
| traciora | OPF | 54.21% | 36.49% | — | 0.464 |
| kalyan-ettin | HF | 55.26% | 32.18% | — | 0.459 |
| ru-pii-ner | RUPII | 60.45% | 35.68% | — | 0.392 |
| apararti | OPF | 61.77% | 30.81% | — | 0.414 |
| openai-base | OPF | 64.13% | 31.43% | — | 0.363 |
| opf-kz-ru | OPF | 70.45% | 24.65% | — | 0.367 |
| opf-ru-v2 | OPF | 71.06% | 22.64% | — | 0.409 |
| spacy-ru-lg | SPACY | 79.94% | 18.37% | — | 0.123 |
| natasha | NATASHA | 91.02% | 8.43% | — | 0.096 |
| rules-ru | RULES | 97.19% | 2.14% | — | 0.100 |
| fef2-secret-ru | HF | 99.01% | 0.75% | — | 0.018 |
| spacy-alrosait | SPACY | 99.86% | 0.06% | — | 0.002 |