Toward Generalized Cross-Lingual Hateful Language Detection with Web-Scale Data and Ensemble LLM Annotations

Dang Hai Dang, Jelena Mitrović, Michael Granitzer


Abstract
We study whether large-scale unlabelled web data and LLM-based synthetic annotations can improve multilingual hate speech detection. Starting from texts crawled via OpenWebSearch (OWS) in four languages (English, German, Spanish, Vietnamese), we pursue two complementary strategies. First, we apply continued pre-training to BERT models by continuing masked language modelling on unlabelled OWS texts before supervised fine-tuning, and show that this yields an average macro-F1 gain of approximately 3% over standard baselines across sixteen benchmarks, with stronger gains in low-resource settings. Second, we use four open-source LLMs (Mistral-7B, Llama3.1-8B, Gemma2-9B, Qwen2.5-14B) to produce synthetic annotations through three ensemble strategies: mean averaging, majority voting, and a LightGBM meta-learner. The LightGBM ensemble consistently outperforms the other strategies. Fine-tuning on these synthetic labels substantially benefits a small model Llama3.2-1B: +11% pooled F1), but provides only a modest gain for the larger Qwen2.5-14B (+0.6%). Our results indicate that the combination of web-scale unlabelled data and LLM-ensemble annotations is most valuable for smaller models and low-data languages.
Anthology ID:
2026.lrec-1.38
Volume:
Proceedings of the Fifteenth Language Resources and Evaluation Conference
Month:
May
Year:
2026
Address:
Palma de Mallorca, Spain
Editors:
Stelios Piperidis, Núria Bel, Henk van den Heuvel, Nancy Ide, Simon Krek, Antonio Toral
Venue:
LREC
SIG:
Publisher:
ELRA Language Resource Association
Note:
Pages:
549–559
Language:
External URL:
https://lrec.elra.info/lrec2026-main-038
DOI:
10.63317/22xjdmqtv6mk
Bibkey:
Cite (ACL):
Dang Hai Dang, Jelena Mitrović, and Michael Granitzer. 2026. Toward Generalized Cross-Lingual Hateful Language Detection with Web-Scale Data and Ensemble LLM Annotations. In Proceedings of the Fifteenth Language Resources and Evaluation Conference, pages 549–559, Palma de Mallorca, Spain. ELRA Language Resource Association.
Cite (Informal):
Toward Generalized Cross-Lingual Hateful Language Detection with Web-Scale Data and Ensemble LLM Annotations (Dang et al., LREC 2026)
Copy Citation: