2npc/KnowledgeBerg
收藏Hugging Face2026-04-18 更新2026-04-26 收录
下载链接：
https://hf-mirror.com/datasets/2npc/KnowledgeBerg
下载链接
链接失效反馈官方服务：
资源简介：
---
configs:
- config_name: default
data_files:
- split: train
path: data/English/tip_of_iceberg_en.csv
- config_name: English
data_files:
- split: train
path: data/English/tip_of_iceberg_en.csv
- config_name: Arabic
data_files:
- split: train
path: data/Arabic/tip_of_iceberg_ar.csv
- config_name: Bengali
data_files:
- split: train
path: data/Bengali/tip_of_iceberg_bn.csv
- config_name: Chinese
data_files:
- split: train
path: data/Chinese/tip_of_iceberg_zh.csv
- config_name: Dutch
data_files:
- split: train
path: data/Dutch/tip_of_iceberg_nl.csv
- config_name: French
data_files:
- split: train
path: data/French/tip_of_iceberg_fr.csv
- config_name: German
data_files:
- split: train
path: data/German/tip_of_iceberg_de.csv
- config_name: Hindi
data_files:
- split: train
path: data/Hindi/tip_of_iceberg_hi.csv
- config_name: Italian
data_files:
- split: train
path: data/Italian/tip_of_iceberg_it.csv
- config_name: Japanese
data_files:
- split: train
path: data/Japanese/tip_of_iceberg_ja.csv
- config_name: Korean
data_files:
- split: train
path: data/Korean/tip_of_iceberg_ko.csv
- config_name: Russian
data_files:
- split: train
path: data/Russian/tip_of_iceberg_ru.csv
- config_name: Spanish
data_files:
- split: train
path: data/Spanish/tip_of_iceberg_es.csv
- config_name: Swahili
data_files:
- split: train
path: data/Swahili/tip_of_iceberg_sw.csv
- config_name: Telugu
data_files:
- split: train
path: data/Telugu/tip_of_iceberg_te.csv
- config_name: Thai
data_files:
- split: train
path: data/Thai/tip_of_iceberg_th.csv
- config_name: Turkish
data_files:
- split: train
path: data/Turkish/tip_of_iceberg_tr.csv
---
# KnowledgeBerg
KnowledgeBerg is a multilingual benchmark for knowledge-grounded reasoning over enumerated answer sets.
The default configuration is English. Use the dataset viewer's configuration selector to open each language separately.
## Data
Each language is stored as one CSV file under `data/<Language>/`.
Columns:
- `id`
- `original_id`
- `domain`
- `original_question`
- `original_answer`
- `question_type`
- `implicit_question`
- `options`
- `correct_option`
- `reasoning_depth`
- `knowledge_width`
- `validation_cot`
提供机构：
2npc
