Overview of the Named Entity Recognition Task at EvaLatin 2026

Abstract

This paper describes the organisation and results of the Named Entity Recognition and Classification (NERC) shared task, conducted as part of EvaLatin 2026. The fourth edition of this evaluation campaign for Natural Language Processing on Latin features two shared tasks, i.e. Dependency Parsing and NERC. After introducing the objective of the task and presenting the Ancient Named Entities Special Interest Group, which aims to address the specific challenges that this task presents, this overview details the annotation tagset, the data provided to the participants and their format. The evaluation metrics and the scorer are also described. Finally, the methodology used by each participating team and their results are presented and discussed.

Resources

Details