wikipedia-sections

# Dataset Card for Wikipedia Sections

This dataset contains pairs and triplets that can be used to train and finetune Sentence Transformer embedding models. The dataset originates from [Dor et al.](https://aclanthology.org/P18-2009.pdf), and was downloaded from [this download link](https://sbert.net/datasets/wikipedia-sections-triplets.zip).
Notably, the "anchor" column contains sentences from Wikipedia, wheras the "positive" column contains other sentences from the same section. The "negative" column contains sentences from other sections.

## Dataset Subsets

### `pair` subset

* Columns: "anchor", "positive"
* Column types: `str`, `str`
* Examples:
```python

```
* Collection strategy: Reading the Wikipedia Sections dataset from https://sbert.net.
* Deduplified: Yes

### `triplet` subset

* Columns: "anchor", "positive", "negative"
* Column types: `str`, `str`, `str`
* Examples:
```python

```
* Collection strategy: Reading the Wikipedia Sections dataset from https://sbert.net.
* Deduplified: Yes

# 维基百科分段数据集卡片
本数据集包含可用于训练与微调句子转换器（Sentence Transformer）嵌入模型的样本对与样本三元组。本数据集源自Dor等人的研究（[Dor et al.](https://aclanthology.org/P18-2009.pdf)），并从[该下载链接](https://sbert.net/datasets/wikipedia-sections-triplets.zip)获取。
值得注意的是，「锚点（anchor）」列包含来自维基百科的句子，「正样本（positive）」列包含同一段落下的其他句子，而「负样本（negative）」列则包含来自其他段落的句子。

## 数据集子集
### `pair` 子集
* 列名："anchor", "positive"
* 列类型：`str`, `str`
* 示例：
python

* 采集策略：从https://sbert.net获取维基百科分段数据集
* 去重情况：是

### `triplet` 子集
* 列名："anchor", "positive", "negative"
* 列类型：`str`, `str`, `str`
* 示例：
python

* 采集策略：从https://sbert.net获取维基百科分段数据集
* 去重情况：是

搜集汇总

背景与挑战

背景概述

该数据集包含用于训练Sentence Transformer嵌入模型的句子对和三元组，数据源自维基百科，其中'anchor'列为维基百科句子，'positive'列为同一章节的其他句子，'negative'列为其他章节的句子。数据集提供pair和triplet两个子集，并采用Apache 2.0许可证。