Published in the Proceedings of the First Workshop on Language Models for Low-Resource Languages (LoResLM 2025), at COLING 2025, Abu Dhabi, UAE.
The paper releases PersianMCQ-Instruct, a dataset of over 12,000 multiple-choice questions generated from more than 4,000 Persian Wikipedia pages, together with three fine-tuned models (Gemma2-9B, Llama3.1-8B and Mistral-7B) for Persian question generation. Human evaluation and fine-tuning experiments showed clear improvements in Persian MCQ generation.
Authors: Kamyar Zeinalipour, Neda Jamshidi, Fahimeh Akbari, Marco Maggini, Monica Bianchini, Marco Gori.
Read the paper | Code | Models and data
All the lab's language releases are listed in Language models and datasets.