Benchmark used in our paper, "Towards Reliable Multilingual Judge Models: An Empirical Study".
The full code is available in the GitHub repository [hitz-zentroa/mJudge](https://github.com/hitz-zentroa/mJudge).
The original English partition from which this benchmark was derived can be found in [RECON](https://huggingface.co/datasets/HiTZ/es_eu-recon).

## Dataset Variants

- `recon_es`: All fields translated into Spanish.
- `recon_eu`: All fields translated into Basque.
- `recon_io_es`: Only the model input and output to be evaluated are translated into Spanish; all other fields remain in English.
- `recon_io_eu`: Only the model input and output to be evaluated are translated into Basque; all other fields remain in English.