B

Bert Base Uncased Eurlex

Developed by nlpaueb
A BERT variant optimized for EU legal texts, pre-trained on 12GB of legal corpus, outperforming general BERT models in legal NLP tasks
Downloads 258
Release Time : 3/2/2022

Model Overview

A BERT model series specifically designed for legal natural language processing, this version is optimized for EU legislative texts and suitable for professional scenarios such as legal document analysis and clause classification

Model Features

Legal Domain Specialization
Pre-trained on 116,062 EU legislative documents, significantly improving legal terminology comprehension
Multi-Jurisdiction Coverage
Training data covers EU, UK, European Court, and US legal texts, with cross-jurisdictional adaptability
Efficient Pre-training
Utilized Google Cloud TPU v3-8 for accelerated training, completing 1 million optimization steps

Model Capabilities

Legal text semantic understanding
Legal clause classification
Legal entity recognition
Contract clause analysis

Use Cases

Legislative Analysis
EU Regulation Semantic Search
Quickly locate relevant legal provisions
12% accuracy improvement on EURLEX test set
Judicial Assistance
Case Law-Statute Correlation Analysis
Automatically associate judgments with relevant legal provisions
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase