anonymizer

To redact PII in CSV datasets, you can use automated detection to identify sensitive entities like names and emails, then apply replacement strategies such as redaction or hashing to sanitize the data before sharing.

Anonymizing Parquet files involves detecting personally identifiable information and applying privacy strategies like redaction, hashing, or LLM-based rewriting to replace sensitive entities while preserving dataset context.

Yes, you can preview PII detection accuracy by running the anonymization process on a small sample of your dataset first, allowing you to verify entity detection and validate results before executing a full-scale job.

Yes, the NeMo Anonymizer supports LLM-based substitution and rewriting, using model-backed inference to replace sensitive personal information with contextually appropriate alternatives for compliance.

Available privacy strategies for tabular data sanitization include redaction, hashing, annotation, and LLM-backed substitution, providing flexible options to securely handle sensitive entities like locations and phone numbers.

Model-backed entity detection works by integrating with inference gateways to identify sensitive personal information in tabular datasets, providing scalable and accurate validation for production-grade data sanitization.