This dataset contains LLM-based translations (via Aya-Expanse) of the full fact-space for SemEval Task 7 - Multilingual and Crosslingual Fact-Checked Claim Retrieval. However, it was not used in the final pipeline due to a lack of performance gains. Further improvements via translation refinement of the fact-space were not pursued due to high computational costs, and no definitive conclusions were drawn about the feasibility of this direction.
