UNH at CheckThat! 2025: Fine-tuning Vs Prompting in Claim Extraction
This work addresses claim extraction for fact-checking, but it is incremental as it compares existing techniques without introducing new methods.
The paper tackled claim extraction from social media by comparing fine-tuning and prompting methods, achieving a best METEOR score with fine-tuned FLAN-T5 but noting that other methods sometimes yield higher-quality claims despite lower scores.
We participate in CheckThat! Task 2 English and explore various methods of prompting and in-context learning, including few-shot prompting and fine-tuning with different LLM families, with the goal of extracting check-worthy claims from social media passages. Our best METEOR score is achieved by fine-tuning a FLAN-T5 model. However, we observe that higher-quality claims can sometimes be extracted using other methods, even when their METEOR scores are lower.