A4 Vertaisarvioitu artikkeli konferenssijulkaisussa
FinerWeb-10BT: Refining Web Data with LLM-Based Line-Level Filtering
Tekijät: Henriksson, Erik; Tarkka, Otto; Ginter, Filip
Toimittaja: Johansson, Richard; Stymne, Sara
Konferenssin vakiintunut nimi: Nordic Conference on Computational Linguistics and Baltic Conference on Human Language Technologies
Julkaisuvuosi: 2025
Lehti: NEALT proceedings series
Kokoomateoksen nimi: Proceedings of the Joint 25th Nordic Conference on Computational Linguistics and 11th Baltic Conference on Human Language Technologies (NoDaLiDa/Baltic-HLT 2025)
Vuosikerta: 57
Aloitussivu: 258
Lopetussivu: 268
ISBN: 978-9908-53-109-0
ISSN: 1736-8197
eISSN: 1736-6305
Julkaisun avoimuus kirjaamishetkellä: Avoimesti saatavilla
Julkaisukanavan avoimuus : Kokonaan avoin julkaisukanava
Verkko-osoite: https://aclanthology.org/2025.nodalida-1.27/
Rinnakkaistallenteen osoite: https://research.utu.fi/converis/portal/detail/Publication/506553763
Data quality is crucial for training Large Language Models (LLMs). Traditional heuristic filters often miss low-quality text or mistakenly remove valuable content. In this paper, we introduce an LLM-based line-level filtering method to enhance training data quality. We use GPT-4o mini to label a 20,000-document sample from FineWeb at the line level, allowing the model to create descriptive labels for low-quality lines. These labels are grouped into nine main categories, and we train a DeBERTa-v3 classifier to scale the filtering to a 10B-token subset of FineWeb. To test the impact of our filtering, we train GPT-2 models on both the original and the filtered datasets. The results show that models trained on the filtered data achieve higher accuracy on the HellaSwag benchmark and reach their performance targets faster, even with up to 25% less data. This demonstrates that LLM-based line-level filtering can significantly improve data quality and training efficiency for LLMs. We release our quality-annotated dataset, FinerWeb-10BT, and the codebase to support further work in this area.
Ladattava julkaisu This is an electronic reprint of the original article. |  |
Julkaisussa olevat rahoitustiedot:
This project has received funding from the European Union’s Horizon Europe research and innovation programme under Grant agreement No 101070350 and from UK Research and Innovation (UKRI) under the UK government’s Horizon Europe funding guarantee [grant number 10052546]. This work was supported by the Research Council of Finland.
