The source code for the paper of "PerQ: Efficient Evaluation of Multilingual Text Platform-Adaptation Quality", accepted to the AACL-IJCNLP 2026 conference, to replicate the results.
The anonymized and processed version of the generated dataset is available on Zenodo.
If you use the data, code, or the information in this repository, cite the following paper, please (also available on arXiv).
@misc{macko2025perqefficientevaluationmultilingual,
title={PerQ: Efficient Evaluation of Multilingual Text Personalization Quality},
author={Dominik Macko and Andrew Pulver},
year={2025},
eprint={2509.25903},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2509.25903},
}