Dataset Card for the Text Anonymization Benchmark (TAB)
Dataset Summary
This repository contains the v1.0 release of the Text Anonymization Benchmark, a corpus for text anonymization.
The corpus comprises 1,268 English-language court cases from the European Court for Human Rights (ECHR). The documents were manually annotated with information about personal identifiers (including their semantic category and need for masking), confidential attributes and co-reference… See the full description on the dataset page: https://huggingface.co/datasets/doshimit3015/text-anonymization-benchmark.
