llama-ai4privacy-multilingual-anonymiser-openpii Open Source Model - Supports PII Information Identification and Anonymization in 8 Languages

Llama Ai4privacy Multilingual Anonymiser Openpii

Developed by ai4privacy

Multilingual PII anonymization model based on ModernBERT architecture, supporting personal identity information recognition and masking in 8 languages

Sequence Labeling

Transformers

Supports Multiple LanguagesOpen Source License:MIT #Multilingual PII Recognition #High-precision Anonymization #Personal Data Masking

Downloads 127

Release Time : 3/23/2025

Model Overview

This model is specifically designed for identifying and anonymizing Personally Identifiable Information (PII) in text, supporting multiple PII types and languages, suitable for privacy protection scenarios

Model Features

Multilingual Support

Supports PII recognition in 8 languages, including major European and Asian languages

High-precision Recognition

Achieves an overall F1 score of 98.59% on test sets, with most PII types exceeding 99% accuracy

Comprehensive PII Coverage

Supports 21 PII label recognitions, including various sensitive information such as names, ID numbers, and contact details

Zero False Positives

Most PII types have 0 false positive rates, ensuring non-PII content won't be incorrectly flagged

Model Capabilities

PII recognition in text

Multilingual PII detection

Sensitive information anonymization

Privacy data protection

Use Cases

Data Privacy Protection

User Data Anonymization

Automatically identifies and anonymizes PII information before storing or sharing user data

Effectively reduces data breach risks and complies with privacy regulations like GDPR

Log Masking

Automatically masks sensitive information in system logs

Protects user privacy in logs while preserving log analysis value

Compliance Checking

Privacy Compliance Audit

Automatically detects potential PII information in documents and databases

Helps organizations identify privacy compliance risk points

🚀 Multilingual PII Anonymiser Evaluation

This project focuses on evaluating a multilingual anonymiser. It provides detailed evaluation metrics for different PII (Personally Identifiable Information) labels and also discusses the model's behavior and limitations.

🚀 Quick Start

The model is based on answerdotai/ModernBERT-base and uses the transformers library. It supports multiple languages including French, English, German, Telugu, Hindi, Italian, Spanish, and Dutch. The dataset used for evaluation is ai4privacy/open-pii-masking-500k-ai4privacy.

📦 Model Information

Property	Details
Model Type	Based on `answerdotai/ModernBERT-base`
Training Data	ai4privacy/open-pii-masking-500k-ai4privacy
Library Name	`transformers`
Tags	PII
License	MIT

💻 Evaluation Metrics

The table below summarizes the detailed evaluation results per PII label:

Label	TP	FP	FN	Accuracy	Precision	Recall	F1 Score
SURNAME	3722	0	28	99.25%	100.0%	99.25%	99.63%
O (Non-PII)	0	400	0	99.30%	n/a	n/a	n/a
TIME	1936	0	0	100.0%	100.0%	100.0%	100.0%
DRIVERLICENSENUM	505	0	2	99.61%	100.0%	99.61%	99.80%
PASSPORTNUM	564	0	2	99.65%	100.0%	99.65%	99.82%
GIVENNAME	7548	0	172	97.77%	100.0%	97.77%	98.87%
TELEPHONENUM	3641	0	0	100.0%	100.0%	100.0%	100.0%
BUILDINGNUM	407	0	19	95.54%	100.0%	95.54%	97.72%
AGE	168	0	1	99.41%	100.0%	99.41%	99.70%
DATE	2335	0	0	100.0%	100.0%	100.0%	100.0%
CITY	1672	0	130	92.79%	100.0%	92.79%	96.26%
TITLE	349	0	35	90.89%	100.0%	90.89%	95.23%
IDCARDNUM	1998	0	22	98.91%	100.0%	98.91%	99.45%
GENDER	121	0	0	100.0%	100.0%	100.0%	100.0%
CREDITCARDNUMBER	557	0	1	99.82%	100.0%	99.82%	99.91%
SEX	78	0	1	98.73%	100.0%	98.73%	99.36%
STREET	1368	0	19	98.63%	100.0%	98.63%	99.31%
TAXNUM	345	0	12	96.64%	100.0%	96.64%	98.29%
EMAIL	2606	0	2	99.92%	100.0%	99.92%	99.96%
SOCIALNUM	411	0	11	97.39%	100.0%	97.39%	98.68%
ZIPCODE	406	0	20	95.31%	100.0%	95.31%	97.60%

Overall Evaluation

Accuracy: 99.01%
Precision: 98.72%
Recall: 98.47%
F1 Score: 98.59%
Total True Positives (TP): 30,737
Total False Positives (FP): 400
Total False Negatives (FN): 477

Macro-Averaged Metrics

Accuracy: 98.35%
Precision: 95.24%
Recall: 93.35%
F1 Score: 94.29%

📚 Model Behavior & Limitations

Evaluation Focus:
The metrics shown above reflect performance on the test split of the open-pii-masking-500k-ai4privacy dataset. Real-world performance may vary and requires additional measures. Feel free to contact support@ai4privacy.com for assistance.

📄 Disclaimer

This model card details the evaluation metrics and fine-tuning parameters for the multilingual anonymiser. Please note:

The model is provided as-is under the MIT License.
It is intended solely for redaction purposes and does not perform full PII classification.
Users should carefully test and evaluate its performance on their own data before deploying in production environments.

Ai4Privacy – Committed to protecting personal data in the age of AI.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご