Using LLMs and Preference Optimization for Agreement-Aware HateWiC Classification
(2025) Proceedings of the The 9th Workshop on Online Abuse and Harms (WOAH) — Vol. 9, p. 538-547 (2025)
Files
2025.woah-1.47.pdf
Open Access - Adobe PDF
- 501.9 KB
Details
- Authors
- Abstract
- Annotator disagreement poses a significant challenge in subjective tasks like hate speech detection. In this paper, we introduce a novel variant of the HateWiC task that explicitly models annotator agreement by estimating the proportion of annotators who classify the meaning of a term as hateful. To tackle this challenge , we explore the use of Llama 3 models fine-tuned through Direct Preference Optimization (DPO). Our experiments show that while LLMs perform well for majority-based hate classification, they struggle with the more complex agreement-aware task. DPO fine-tuning offers improvements, particularly when applied to instruction-tuned models. Our results emphasize the need for improved modeling of subjec-tivity in hate classification and this study can serve as foundation for future advancements.
- Affiliations
Citations
Loftus, S., Mülthaler, A., Hoeken, S., Zarrieß, S., & Alaçam, Ö. (2025). Using LLMs and Preference Optimization for Agreement-Aware HateWiC Classification. Proceedings of the The 9th Workshop on Online Abuse and Harms (WOAH), 9, 538-547. https://hdl.handle.net/2078.5/278595 (Original work published 2025)