AKCIT-FN at CheckThat! 2025: Switching Fine-Tuned SLMs and LLM Prompting for Multilingual Claim Normalization
This work addresses claim normalization for automated fact-checking pipelines across multiple languages, with incremental improvements in performance.
The paper tackled multilingual claim normalization across 20 languages by using fine-tuned small language models for supervised languages and LLM prompting for zero-shot scenarios, achieving top-three rankings in 15 languages, including second place in 8 languages and an average METEOR score of 0.5290 for Portuguese.
Claim normalization, the transformation of informal social media posts into concise, self-contained statements, is a crucial step in automated fact-checking pipelines. This paper details our submission to the CLEF-2025 CheckThat! Task~2, which challenges systems to perform claim normalization across twenty languages, divided into thirteen supervised (high-resource) and seven zero-shot (no training data) tracks. Our approach, leveraging fine-tuned Small Language Models (SLMs) for supervised languages and Large Language Model (LLM) prompting for zero-shot scenarios, achieved podium positions (top three) in fifteen of the twenty languages. Notably, this included second-place rankings in eight languages, five of which were among the seven designated zero-shot languages, underscoring the effectiveness of our LLM-based zero-shot strategy. For Portuguese, our initial development language, our system achieved an average METEOR score of 0.5290, ranking third. All implementation artifacts, including inference, training, evaluation scripts, and prompt configurations, are publicly available at https://github.com/ju-resplande/checkthat2025_normalization.