Dataset · ai4privacy
PII personal data masking
ai4privacy/pii-masking-300k
300k texts across 6 languages with PII annotations for training data removal models.
What's inside
A dataset for training and evaluating models that remove personal and sensitive data from text. This version contains 300k examples; languages: en, fr, de, it, es, nl; size 100K<n<1M. The card references newer openpii-1.5m and openpii-1m versions with 30 and 23 languages and 19 PII classes; the format is not specified in the card.
What you can do with it
- Train a token-classification model for PII detection
- Evaluate text anonymization
- Fine-tune a multilingual personal data NER detector
- Tasks
- Text classification, Token classification, Table question answering, Question answering, Zero shot classification, Summarization
- Languages
- English, French, German, Italian, Spanish, Dutch
- Size
- 100K–1M rows
- Downloads this month
- 4.2K
- Likes
- 121
Models trained on it
- temsa/eurocivic-gliner2-onnx-cpu-v2Token classification 571
- LokaalHub/nl-lokaal-middelToken classification 52
- Aniket2003333333/secguard-decision-model 51
- bengid/pii-redaction-deberta-smallToken classification 29
- gpancardo/beto-piiToken classification 29
- yonigo/distilbert-base-cased-pii-enToken classification 28
- S4QezPbPaV6mQaKPx75R/sentinel-Q8_0-GGUFText generation 28
- bengid/pii-redaction-deberta-xsmallToken classification 26
How to load it
from datasets import load_dataset
ds = load_dataset("ai4privacy/pii-masking-300k")