Dataset · ai4privacy
PII masking on 200k examples
ai4privacy/pii-masking-200k
Texts with personally identifiable information in English, French, German, and Italian for privacy masking training.
What's inside
The dataset contains texts in English, French, German, and Italian with PII annotations for masking tasks. Metadata categorizes it under text and token classification, question-answering, summarization, and zero-shot classification. The card calls it an open dataset for privacy masking and references a newer flagship version, openpii-1.5m.
What you can do with it
- Fine-tune an NER model for PII detection
- Set up PII masking in texts
- Test zero-shot PII classification
- Tasks
- Text classification, Token classification, Table question answering, Question answering, Zero shot classification, Summarization
- Languages
- English, French, German, Italian
- Size
- 100K–1M rows
- Downloads this month
- 4.7K
- Likes
- 132
Models trained on it
- OpenMed/privacy-filter-multilingual-v2Token classification 116.2K
- OpenMed/privacy-filter-multilingualToken classification 70.5K
- Isotonic/deberta-v3-base_finetuned_ai4privacy_v2Token classification 25K
- Isotonic/distilbert_finetuned_ai4privacy_v2Token classification 8.5K
- Isotonic/distilbert-base-german-cased_finetuned_ai4privacy_v2Token classification 3K
- rv2307/electra-small-nerToken classification 2.6K
- SoelMgd/bert-pii-detectionToken classification 1.5K
- Spyridon/camembert-pii-fr-v2Token classification 118
How to load it
from datasets import load_dataset
ds = load_dataset("ai4privacy/pii-masking-200k")