---
language:
- ar
- bn
- ca
- da
- de
- es
- eu
- fr
- gu
- hi
- hr
- hu
- hy
- id
- it
- kn
- ml
- mr
- ne
- nl
- pt
- ro
- ru
- sk
- sr
- sv
- ta
- te
- uk
- vi
license: cc-by-nc-4.0
---
# okapi_mmlu
<!-- Provide a quick summary of the dataset. -->
Multilingual translation of [Measuring Massive Multitask Language Understanding (MMLU)](https://arxiv.org/abs/2009.03300).
## Dataset Details
### Dataset Description
<!-- Provide a longer summary of what this dataset is. -->
MMLU is a benchmark that measured a text model’s multitask accuracy.
The test covers 57 tasks including elementary mathematics, US history, computer
science, law, and more. To attain high accuracy on this test, models must possess
extensive world knowledge and problem solving ability. By comprehensively evaluating the breadth and depth of a model’s academic and professional understanding, MMLU can be used to analyze models across many tasks and to identify important shortcomings.
- **Curated by:** Dac Lai, Viet and Van Nguyen, Chien and Ngo, Nghia Trung and Nguyen, Thuat and Dernoncourt, Franck and Rossi, Ryan A and Nguyen, Thien Huu
- **License:** The datasets are CC BY NC 4.0 (allowing only non-commercial use).
### Dataset Sources
<!-- Provide the basic links for the dataset. -->
- **Repository:** http://nlp.uoregon.edu/download/okapi-eval/datasets/
- **Paper:** Okapi ([Lai et al., 2023](https://arxiv.org/abs/2307.16039))
## Citation
<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->
```bibtex
@article{dac2023okapi,
title={Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback},
author={Dac Lai, Viet and Van Nguyen, Chien and Ngo, Nghia Trung and Nguyen, Thuat and Dernoncourt, Franck and Rossi, Ryan A and Nguyen, Thien Huu},
journal={arXiv e-prints},
pages={arXiv--2307},
year={2023}
}
```
```bibtex
@article{hendryckstest2021,
title={Measuring Massive Multitask Language Understanding},
author={Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt},
journal={Proceedings of the International Conference on Learning Representations (ICLR)},
year={2021}
}
```
language:
- 阿拉伯语(ar)
- 孟加拉语(bn)
- 加泰罗尼亚语(ca)
- 丹麦语(da)
- 德语(de)
- 西班牙语(es)
- 巴斯克语(eu)
- 法语(fr)
- 古吉拉特语(gu)
- 印地语(hi)
- 克罗地亚语(hr)
- 匈牙利语(hu)
- 亚美尼亚语(hy)
- 印尼语(id)
- 意大利语(it)
- 卡纳达语(kn)
- 马拉雅拉姆语(ml)
- 马拉地语(mr)
- 尼泊尔语(ne)
- 荷兰语(nl)
- 葡萄牙语(pt)
- 罗马尼亚语(ro)
- 俄语(ru)
- 斯洛伐克语(sk)
- 塞尔维亚语(sr)
- 瑞典语(sv)
- 泰米尔语(ta)
- 泰卢固语(te)
- 乌克兰语(uk)
- 越南语(vi)
license: cc-by-nc-4.0
# okapi_mmlu
<!-- 数据集简要说明 -->
本数据集为《大规模多任务语言理解测评(Measuring Massive Multitask Language Understanding,MMLU)》的多语言翻译版本。
## 数据集详情
### 数据集说明
<!-- 本数据集的详细描述 -->
MMLU是一项用于测评文本模型多任务准确率的基准测试。该测试涵盖57项任务,包括初等数学、美国历史、计算机科学、法学等。要在该测试中取得高准确率,模型需具备广博的世界知识与问题求解能力。通过全面评估模型在学术与专业理解层面的广度与深度,MMLU可用于跨多任务分析模型,并识别其关键短板。
- **数据整理者:** Dac Lai, Viet、Van Nguyen, Chien、Ngo, Nghia Trung、Nguyen, Thuat、Dernoncourt, Franck、Rossi, Ryan A、Nguyen, Thien Huu
- **许可证:** 本数据集采用知识共享署名-非商业性使用4.0国际许可协议(CC BY-NC 4.0),仅允许非商业性使用。
### 数据集来源
<!-- 数据集的基础链接信息 -->
- **仓库地址:** http://nlp.uoregon.edu/download/okapi-eval/datasets/
- **相关论文:** 《Okapi》(Lai等人,2023,https://arxiv.org/abs/2307.16039)
## 引用格式
<!-- 若有介绍该数据集的论文或博客文章,需在此处添加APA及Bibtex引用格式 -->
bibtex
@article{dac2023okapi,
title={Okapi: Instruction-tuned Large Language Models in Multiple Languages with Reinforcement Learning from Human Feedback},
author={Dac Lai, Viet and Van Nguyen, Chien and Ngo, Nghia Trung and Nguyen, Thuat and Dernoncourt, Franck and Rossi, Ryan A and Nguyen, Thien Huu},
journal={arXiv e-prints},
pages={arXiv--2307},
year={2023}
}
bibtex
@article{hendryckstest2021,
title={Measuring Massive Multitask Language Understanding},
author={Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt},
journal={Proceedings of the International Conference on Learning Representations (ICLR)},
year={2021}
}