SPACY detector
spacy-ru-lg PII detection benchmark
Measured on 41 of 41 retained datasets under the same frozen scoring and masking protocol.
26.75%fully hidden annotations
31.36%detected / overlapped
3.25%extra masking
52,320CPU characters / second
Strengths
Highest complete-masking rates by normalized data type.
People's names52.59%34,571 / 65,736
Organizations46.18%8,901 / 19,274
Addresses & locations36.75%14,554 / 39,602
Watch list
Lowest observed complete-masking rates; inspect before deployment.
Network identifiers1.31%337 / 25,631
Dates & times1.33%73 / 5,502
Logins & usernames1.66%189 / 11,382
Language coverage
The overall leader can differ from the best choice for one language slice.
| Language | Datasets | Fully hidden ↑ | Detected ↑ | Extra masking ↓ |
|---|---|---|---|---|
| English | 17 | 13.37% | 13.85% | 3.09% |
| Russian | 19 | 36.56% | 43.41% | 3.69% |
| Multilingual | 5 | 11.13% | 12.90% | 1.54% |
Configuration, source and version
- Family
- SPACY
- Revision
- fa0b3a0086b7b0d6a424d999ce25480bd67dd455
- Eligible datasets
- 41 / 41
- Character F1
- 0.326
- Untouched diagnostic
- 68.64% · 156,139 / 227,466
- Benchmark version
- 1.0.3 · 2026-09-09
- CPU seconds / 10k chars
- 0.191
- GPU seconds / 10k chars
- Not measured
Results by dataset
Complete masking is primary; detection shows whether any character overlapped.
| Dataset | Gold annotations | Fully hidden ↑ | Detected ↑ | Extra masking ↓ |
|---|---|---|---|---|
| synth-secrets-en | 581 | 0.00% | 0.00% | 0.78% |
| synth-secrets-ru | 581 | 0.69% | 0.69% | 0.27% |
| arthur-passwords | 280 | 1.07% | 1.07% | 4.96% |
| synth-env-configs | 392 | 1.28% | 1.28% | 0.41% |
| corrupt-secrets-issues | 288 | 2.78% | 5.56% | 3.18% |
| privy | 1,899 | 3.21% | 4.27% | 3.38% |
| synth-wiki-tables | 13,094 | 3.41% | 8.27% | — |
| secrets-issues | 288 | 3.47% | 4.86% | 2.51% |
| rubai-ru | 3,458 | 4.05% | 52.37% | — |
| leaky-repo | 95 | 4.21% | 4.21% | 1.03% |
| leak-museum | 101 | 4.95% | 4.95% | 3.43% |
| ameau01 | 2,242 | 5.49% | 5.49% | 3.74% |
| synth-jira-comments | 2,705 | 5.84% | 10.02% | — |
| secrets-rules | 746 | 6.70% | 7.24% | 12.49% |
| redact-multi | 22,864 | 6.82% | 8.62% | — |
| gretel-multi | 4,484 | 7.83% | 13.80% | 1.56% |
| nym-multi | 8,169 | 8.67% | 9.65% | 0.00% |
| nym-en | 3,584 | 9.88% | 10.99% | 9.02% |
| synth-ru-tickets | 32,099 | 10.01% | 12.63% | — |
| creddata | 776 | 10.31% | 10.82% | 4.22% |
| nemotron-pii | 9,391 | 10.89% | 11.18% | 0.43% |
| corrupt-hivetrace | 1,667 | 12.42% | 21.12% | 8.16% |
| dialogpii-en | 3,087 | 13.74% | 14.35% | 1.58% |
| tonicai | 2,417 | 14.11% | 14.11% | 0.64% |
| hivetrace | 1,667 | 15.72% | 25.19% | 0.64% |
| dialogpii-multi | 10,345 | 18.37% | 20.06% | — |
| kiji-multi | 7,565 | 18.89% | 19.02% | — |
| kiji-en | 7,623 | 20.54% | 20.66% | — |
| corrupt-redmadrobot | 5,531 | 21.73% | 25.33% | 6.20% |
| alexen2 | 1,261 | 22.52% | 46.71% | 10.50% |
| tab-echr | 3,830 | 25.35% | 26.58% | — |
| russian-pii-66k | 4,805 | 29.30% | 37.84% | — |
| nym-ru | 9,497 | 30.50% | 36.66% | — |
| redmadrobot | 5,516 | 30.89% | 34.23% | 2.98% |
| redact-ru | 9,230 | 42.12% | 49.60% | 7.07% |
| alrosait | 1,862 | 48.07% | 85.66% | 2.74% |
| multiconer-ru | 1,208 | 50.17% | 67.88% | 6.99% |
| scanpatch | 8,708 | 50.93% | 57.44% | 1.90% |
| jayguard | 1,195 | 52.55% | 79.25% | 3.26% |
| nerel | 24,369 | 82.10% | 88.27% | — |
| factrueval | 7,966 | 93.95% | 95.39% | — |