LEAKS detector
detect-secrets PII detection benchmark
Measured on 4 of 41 retained datasets under the same frozen scoring and masking protocol.
16.74%fully hidden annotations
16.99%detected / overlapped
4.03%extra masking
—CPU characters / second
Strengths
Highest complete-masking rates by normalized data type.
Passwords, keys & tokens38.24%468 / 1,224
Bank accounts & cards0.00%0 / 169
Documents & identifiers0.00%0 / 609
Watch list
Lowest observed complete-masking rates; inspect before deployment.
Addresses & locations0.00%0 / 176
Phone numbers & email0.00%0 / 390
People's names0.00%0 / 228
Language coverage
The overall leader can differ from the best choice for one language slice.
| Language | Datasets | Fully hidden ↑ | Detected ↑ | Extra masking ↓ |
|---|---|---|---|---|
| English | 3 | 39.15% | 39.77% | 4.23% |
| Russian | 1 | 1.56% | 1.56% | 0.00% |
Configuration, source and version
- Family
- LEAKS
- Revision
- 1.5.0
- Eligible datasets
- 4 / 41
- Character F1
- 0.366
- Untouched diagnostic
- 83.01% · 2,321 / 2,796
- Benchmark version
- 1.0.3 · 2026-09-09
- CPU seconds / 10k chars
- Not measured
- 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 ↓ |
|---|---|---|---|---|
| hivetrace | 1,667 | 1.56% | 1.56% | 0.00% |
| secrets-issues | 288 | 24.31% | 26.04% | 1.82% |
| leaky-repo | 95 | 32.63% | 34.74% | 0.00% |
| secrets-rules | 746 | 45.71% | 45.71% | 30.78% |