SeaEval/SeaEval_datasets

---
license: cc-by-nc-4.0
configs:

- config_name: cross_xquad
data_files:
- split: test
path: "cross_xquad.json"

- config_name: cross_mmlu
data_files:
- split: test
path: "cross_mmlu.json"

- config_name: cross_logiqa
data_files:
- split: test
path: "cross_logiqa.json"

- config_name: us_eval
data_files:
- split: test
path: "us_eval.json"

- config_name: sg_eval
data_files:
- split: test
path: "sg_eval.json"

- config_name: cn_eval
data_files:
- split: test
path: "cn_eval.json"

- config_name: ph_eval
data_files:
- split: test
path: "ph_eval.json"

- config_name: flores_ind2eng
data_files:
- split: test
path: "flores_ind2eng.json"

- config_name: flores_vie2eng
data_files:
- split: test
path: "flores_vie2eng.json"

- config_name: flores_zho2eng
data_files:
- split: test
path: "flores_zho2eng.json"

- config_name: flores_zsm2eng
data_files:
- split: test
path: "flores_zsm2eng.json"

- config_name: mmlu
data_files:
- split: test
path: "mmlu.json"

- config_name: mmlu_full
data_files:
- split: test
path: "mmlu_full.json"

- config_name: c_eval
data_files:
- split: test
path: "c_eval.json"

- config_name: c_eval_full
data_files:
- split: test
path: "c_eval_full.json"

- config_name: cmmlu
data_files:
- split: test
path: "cmmlu.json"

- config_name: cmmlu_full
data_files:
- split: test
path: "cmmlu_full.json"

- config_name: zbench
data_files:
- split: test
path: "zbench.json"

- config_name: ind_emotion
data_files:
- split: test
path: "ind_emotion.json"

- config_name: ocnli
data_files:
- split: test
path: "ocnli.json"

- config_name: c3
data_files:
- split: test
path: "c3.json"

- config_name: dream
data_files:
- split: test
path: "dream.json"

- config_name: samsum
data_files:
- split: test
path: "samsum.json"

- config_name: dialogsum
data_files:
- split: test
path: "dialogsum.json"

- config_name: sst2
data_files:
- split: test
path: "sst2.json"

- config_name: cola
data_files:
- split: test
path: "cola.json"

- config_name: qqp
data_files:
- split: test
path: "qqp.json"

- config_name: mnli
data_files:
- split: test
path: "mnli.json"

- config_name: qnli
data_files:
- split: test
path: "qnli.json"

- config_name: wnli
data_files:
- split: test
path: "wnli.json"

- config_name: rte
data_files:
- split: test
path: "rte.json"

- config_name: mrpc
data_files:
- split: test
path: "mrpc.json"

- config_name: indommlu
data_files:
- split: test
path: "indommlu.json"

---

\[GitHub\]: https://github.com/SeaEval/SeaEval \[Website\]: https://seaeval.github.io/

```
@article{SeaEval,
title={SeaEval for Multilingual Foundation Models: From Cross-Lingual Alignment to Cultural Reasoning},
author={Wang, Bin and Liu, Zhengyuan and Huang, Xin and Jiao, Fangkai and Ding, Yang and Aw, Ai Ti and Chen, Nancy F.},
journal={NAACL},
year={2024}
}
```

原始信息汇总

数据集概述

该数据集包含多个配置文件，每个配置文件对应不同的测试数据集。以下是各配置文件及其对应的测试数据文件路径：

• cross_xquad

  • 测试数据文件路径: cross_xquad.json

• cross_mmlu

  • 测试数据文件路径: cross_mmlu.json

• cross_logiqa

  • 测试数据文件路径: cross_logiqa.json

• us_eval

  • 测试数据文件路径: us_eval.json

• sg_eval

  • 测试数据文件路径: sg_eval.json

• cn_eval

  • 测试数据文件路径: cn_eval.json

• ph_eval

  • 测试数据文件路径: ph_eval.json

• flores_ind2eng

  • 测试数据文件路径: flores_ind2eng.json

• flores_vie2eng

  • 测试数据文件路径: flores_vie2eng.json

• flores_zho2eng

  • 测试数据文件路径: flores_zho2eng.json

• flores_zsm2eng

  • 测试数据文件路径: flores_zsm2eng.json

• mmlu

  • 测试数据文件路径: mmlu.json

• mmlu_full

  • 测试数据文件路径: mmlu_full.json

• c_eval

  • 测试数据文件路径: c_eval.json

• c_eval_full

  • 测试数据文件路径: c_eval_full.json

• cmmlu

  • 测试数据文件路径: cmmlu.json

• cmmlu_full

  • 测试数据文件路径: cmmlu_full.json

• zbench

  • 测试数据文件路径: zbench.json

• ind_emotion

  • 测试数据文件路径: ind_emotion.json

• ocnli

  • 测试数据文件路径: ocnli.json

• c3

  • 测试数据文件路径: c3.json

• dream

  • 测试数据文件路径: dream.json

• samsum

  • 测试数据文件路径: samsum.json

• dialogsum

  • 测试数据文件路径: dialogsum.json

• sst2

  • 测试数据文件路径: sst2.json

• cola

  • 测试数据文件路径: cola.json

• qqp

  • 测试数据文件路径: qqp.json

• mnli

  • 测试数据文件路径: mnli.json

• qnli

  • 测试数据文件路径: qnli.json

• wnli

  • 测试数据文件路径: wnli.json

• rte

  • 测试数据文件路径: rte.json

• mrpc

  • 测试数据文件路径: mrpc.json

• indommlu

  • 测试数据文件路径: indommlu.json

许可证

该数据集的许可证为 cc-by-nc-4.0。