There is a newer version of the record available.

Published August 17, 2026 | Version 1.0.0
Other Restricted

NLP Framework for Bias Indicator Analysis in Legal Text: A Law-Sensitive Severity Ladder and Benchmark Evaluation across Eight English Legal Datasets

  • 1. COMSATS University Islamabad Pakistan, Lahore Campus

Description

This repository contains the data and computational code associated with the study “NLP Framework for Bias Indicator Analysis in Legal Text: A Law-Sensitive Severity Ladder and Benchmark Evaluation Across Eight English Legal Datasets.” The materials support the implementation, analysis, severity scoring, statistical evaluation, and benchmark experiments reported in the manuscript. The study examines textual bias-indicator signals across eight English legal datasets and develops a law-sensitive severity ladder for interpretable analysis. The repository includes the research data package and computational code used for the reported experiments. The benchmark evaluation includes lexical, embedding-based, and transformer-based methods. Original third-party datasets should be obtained from their respective official sources and used according to their applicable licenses and terms. The data package includes materials associated with the 276,240-document silver-label benchmark corpus described in the manuscript.

Files

Restricted

The record is publicly accessible, but files are restricted. Log in to check if you have access.

Additional details

Software

Programming language
Python
Development Status
Active