D

Deberta V3 Base Finetuned Ai4privacy V2

Developed by Isotonic
This model is a privacy protection model fine-tuned based on the Microsoft DeBERTa-v3-base architecture, specifically designed to identify and mask Personally Identifiable Information (PII) in text.
Downloads 33.35k
Release Time : 12/1/2023

Model Overview

Fine-tuned on the world's largest open-source privacy dataset, this model can identify 54 types of sensitive information and is suitable for privacy protection in AI assistants and LLM scenarios.

Model Features

Comprehensive PII Identification Capability
Supports identification of 54 types of Personally Identifiable Information, including financial information, contact details, biometric features, etc.
High-precision Identification
Achieves an overall F1 score of 97.57% on the evaluation set, with recognition accuracy exceeding 99% for most categories.
Multi-domain Applicability
Training data covers multiple domains including business, education, psychology, and law.

Model Capabilities

PII Identification in Text
Sensitive Information Classification
Privacy Data Masking
Multi-type Financial Information Detection
Contact Information Recognition

Use Cases

Privacy Protection
AI Chat Log Anonymization
Automatically identifies and masks sensitive information in chat logs
Effectively protects user privacy while preserving conversation context
Document Privacy Review
Scans documents for Personally Identifiable Information
Helps organizations comply with privacy regulations such as GDPR
Financial Security
Payment Information Detection
Identifies credit card numbers, bank accounts and other financial information
F1 score exceeds 99%
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase