Evaluating Named Entity Recognition Models for Russian Cultural News Texts: From BERT to LLM

Maria Levchenko · 2025

Read Evaluating Named Entity Recognition Models for Russian Cultural News Texts: From BERT to LLM by Maria Levchenko (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-02589.txt.

Description

Maria Levchenko compares models that find person names in Russian cultural news texts, in a paper from 2025. The texts come from a St. Petersburg events guide spanning 1999 to 2019. Russian word forms shift a lot, which makes name finding harder than in English.

Established transformer tools were tested next to recent large language models. The best score came from the newest large model with a prompt that asked for structured output. An earlier large model gave the cleanest calls, with the fewest false hits. A follow-up test with a still newer model scored slightly higher with simple prompts. Levchenko draws lessons for heritage workers who mine such archives.

Catalog record

CREATORMaria Levchenko
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2506-02589

Browse the full library · About the Tower of Babel