Personal Digital Information Masking Preview for Text Anonymization

EPII Personal Digital Information (PDI) Masking Preview Dataset contains 400 text samples demonstrating the identification and masking of sensitive personal information. This preview represents the structure and label taxonomy of a larger commercial dataset containing over 100,000 entries. The dataset was created by ai4privacy and was last updated in April 2026.

Use Cases

• Train a named entity recognition model to identify sensitive PDI entities like names or addresses within text.
• Benchmark the performance of text masking algorithms using the provided label taxonomy for sensitive information categories.
• Develop sequence labeling models for redacting specific personal digital information types demonstrated in the preview samples.

Strengths

• Preview of a larger dataset containing over 100,000 entries.
• Demonstrates a specialized label taxonomy for sensitive personal digital information.

Limitations

• Small preview size of only 400 samples limits model training.
• Preview nature means the full dataset's features and volume are not accessible.

Provenance

Source

ai4privacy

Collection Method

Specialized collection for identifying and masking sensitive personal digital information within text data.

Freshness

Last updated in April 2026.