Results updated 19 Sep 2026
MULTI · PII

redact-multi PII detection benchmark results

A frozen evaluation slice with 1,000 rows, 22,864 normalized gold annotations and source-level provenance.

1,000input rows
22,864gold annotations
1,962,299characters
41measured detectors

Dataset provenance

Language group
MULTI
Task
PII
License
other (gated, terms accepted)
Data categories
ACCOUNT, ADDRESS, CONTACT, DATE, ID, NET, ORG, OTHER, 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 fully hidden annotations. Training-source overlaps are excluded.

DetectorFamilyFully hidden ↑Detected / any overlap ↑Extra masking ↓Char F1 ↑
pplxPPLX86.98%89.39%—0.705
gliner-urchadeGLINER82.02%84.54%—0.663
bardsai-euONNX79.98%91.65%—0.755
gliner2-fastinoGLINER278.92%84.54%—0.694
gliner2-hivetrace-omniGLINER278.44%84.03%—0.717
nuner-zeroGLINER73.43%78.46%—0.662
gliner2-hivetrace-uniGLINER272.38%76.73%—0.689
pii-shield-onnxONNX70.23%81.56%—0.329
aparartiOPF68.32%74.10%—0.616
openai-baseOPF68.17%72.01%—0.623
gliner25-fastinoGLINER267.60%71.45%—0.675
opf-kz-ruOPF63.37%69.30%—0.598
mmbert32kHF62.39%89.25%—0.595
gliner2-largeGLINER262.34%67.31%—0.610
gliner-nvidiaGLINER62.17%82.68%—0.623
gliner-multi-v21GLINER60.56%67.84%—0.546
nym-baseHF59.45%81.93%—0.662
gravitee-smallHF59.42%65.43%—0.510
opf-ruOPF58.26%74.80%—0.591
nym-smallONNX57.85%79.33%—0.657
ru-pii-nerRUPII55.93%59.06%—0.607
davlan-xlmrHF55.52%64.35%—0.539
openmed-multilingualHF54.48%77.67%—0.614
davlan-mbertHF53.74%61.69%—0.507
opf-ru-v2OPF52.35%61.83%—0.647
tracioraOPF52.29%60.69%—0.650
gliner2-vladlinvGLINER249.27%50.86%—0.571
openmed-nemotronHF46.51%71.09%—0.484
ner-ru-yqelzHF45.79%55.84%—0.460
kalyan-ettinHF44.22%68.89%—0.543
stanza-ruSTANZA42.40%52.22%—0.277
gliner-pii-edgeGLINER41.73%46.82%—0.483
ru-legal-nerHF39.29%61.07%—0.330
ner-ru-ghermanHF36.48%60.13%—0.478
gliner-stream-piiGLINER35.81%47.61%—0.470
gliner-pii-baseGLINER27.93%33.63%—0.451
fef2-secret-ruHF11.71%15.60%—0.274
rules-ruRULES11.12%11.55%—0.322
spacy-ru-lgSPACY6.82%8.62%—0.096
natashaNATASHA4.14%5.30%—0.076
spacy-alrosaitSPACY0.01%0.10%—0.002