AILS-NTUA at SemEval-2025 Task 3: Leveraging Large Language Models and Translation Strategies for Multilingual Hallucination Detection
This addresses an underexplored problem in NLP for multilingual applications, but it is incremental as it builds on existing LLM and translation methods.
The paper tackled multilingual hallucination detection by proposing a training-free LLM prompting strategy that translates text into English, achieving competitive rankings including two first positions in low-resource languages.
Multilingual hallucination detection stands as an underexplored challenge, which the Mu-SHROOM shared task seeks to address. In this work, we propose an efficient, training-free LLM prompting strategy that enhances detection by translating multilingual text spans into English. Our approach achieves competitive rankings across multiple languages, securing two first positions in low-resource languages. The consistency of our results highlights the effectiveness of our translation strategy for hallucination detection, demonstrating its applicability regardless of the source language.