ZurichNLP/romansh-mt-evaluation

---
configs:
- config_name: document_accuracy
data_files:
- split: train
path: document_accuracy/**/*.jsonl
- config_name: segment_accuracy
data_files:
- split: train
path: segment_accuracy/**/*.jsonl
- config_name: segment_fluency
data_files:
- split: train
path: segment_fluency/**/*.jsonl
license: apache-2.0
language:
- de
- rm
---

Human quality ratings of machine translations from German into 6 varieties of Romansh.

## Citation
```bibtex
@misc{vamvas-et-al-2026-translation,
title={Translation Asymmetry in LLMs as a Data Augmentation Factor: A Case Study for 6 Romansh Language Varieties},
author={Jannis Vamvas and Ignacio Pérez Prat and Angela Heldstab and Dominic P. Fischer and Sina Ahmadi and Rico Sennrich},
year={2026},
eprint={2603.25489},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2603.25489},
}
```

---
配置项：
- 配置名称：document_accuracy
数据文件：
- 拆分集：训练集（train）
路径：document_accuracy/**/*.jsonl
- 配置名称：segment_accuracy
数据文件：
- 拆分集：训练集（train）
路径：segment_accuracy/**/*.jsonl
- 配置名称：segment_fluency
数据文件：
- 拆分集：训练集（train）
路径：segment_fluency/**/*.jsonl
许可证：Apache 2.0
语言：
- 德语（de）
- 罗曼什语（rm）
---

本数据集收录了德语译为6种罗曼什语（Romansh）变体的机器翻译的人工质量评级。

## 引用
bibtex
@misc{vamvas-et-al-2026-translation,
title={作为数据增强因素的大语言模型（Large Language Model，LLM）翻译不对称性：针对6种罗曼什语（Romansh）变体的案例研究},
author={Jannis Vamvas and Ignacio Pérez Prat and Angela Heldstab and Dominic P. Fischer and Sina Ahmadi and Rico Sennrich},
year={2026},
eprint={2603.25489},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2603.25489},
}