Pii Benchmark: Russian PII NER Evaluation Dataset

Russian-language text data designed for evaluating Personally Identifiable Information detection and Named Entity Recognition systems. The dataset was created by redmadrobot-rnd and last updated on 2026-06-09. It targets guardrail and anonymization pipelines that must find personal data and Russian identity-document numbers.

Use Cases

• Benchmarking PII detection systems based on Russian-language text examples.
• Evaluating Named Entity Recognition models on Russian personal data categories.
• Testing guardrail and anonymization pipelines for reliability against real examples.

Strengths

• Focuses on Russian-language PII, a specific and important domain for privacy.
• Designed for evaluating real-world anonymization pipelines, suggesting practical relevance.
• Includes manually annotated examples, indicating a level of curation.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

redmadrobot-rnd

Collection Method

Likely combines real, manually annotated examples.

Freshness

Last updated 2026-06-09 12:53:01; freshness should be verified.

Geography

Russian-language text.