PII Masking Financial Text Preview

50 sample entries from the PII-Masking-2M European release by AI4Privacy. The preview contains redacted source text and PII values, with the full dataset containing 200k entries. The dataset was last updated in April 2026.

Use Cases

• Training models to identify and mask PII labels in financial text snippets.
• Benchmarking named entity recognition (NER) systems on redacted financial PII entities.
• Analyzing label and language distributions across European financial documents.
• Developing synthetic data generation techniques for privacy-preserving financial AI.

Strengths

• 50 sample entries provided for preview
• Part of a larger 200k entry dataset
• Focuses on European financial text coverage

Limitations

• Preview size is very small (50 rows)
• Source text and PII values are redacted in the preview
• Full dataset access requires contacting the author

Provenance

Source

AI4Privacy

Collection Method

Part of the PII-Masking-2M European release; method not specified.

Freshness

Last updated April 2026.

Geography

European coverage