Anthropic/persuasion

---
license: cc-by-nc-sa-4.0
language:
- en
size_categories:
- 1K<n<10K
---
# Dataset Card for Persuasion Dataset

## Dataset Summary
The Persuasion Dataset contains claims and corresponding human-written and model-generated arguments, along with persuasiveness scores.
This dataset was created for research on measuring the persuasiveness of language models, as described in this blog post: [Measuring the Persuasiveness of Language Models](https://www.anthropic.com/news/measuring-model-persuasiveness).

## Dataset Description

The dataset consists of a CSV file with the following columns:
- **worker\_id**: Id of the participant who annotated their initial and final stance on the claim.
- **claim**: The claim for which the argument was generated.
- **argument**: The generated argument, either by a human or a language model.
- **source**: The source of the argument (model name or "Human").
- **prompt\_type**: The prompt type used to generate the argument.
- **rating\_initial**: The participant's initial rating of the claim.
- **rating\_final**: The participant's final rating of the claim after reading the argument.

## Usage
```python
from datasets import load_dataset
# Loading the data
dataset = load_dataset("Anthropic/persuasion")
```
## Contact
For questions, you can email esin at anthropic dot com

## Citation
If you would like to cite our work or data, you may use the following bibtex citation:

```
@online{durmus2024persuasion,
author = {Esin Durmus and Liane Lovitt and Alex Tamkin and Stuart Ritchie and Jack Clark and Deep Ganguli},
title = {Measuring the Persuasiveness of Language Models},
date = {2024-04-09},
year = {2024},
url = {https://www.anthropic.com/news/measuring-model-persuasiveness},
}

```

license: 知识共享署名-非商业性使用-相同方式共享 4.0 国际协议（CC BY-NC-SA 4.0）
language:
- en
size_categories:
- 1K<n<10K
---
# 《说服性数据集》数据集卡片

## 数据集概述
本说服性数据集包含主张、对应人类撰写及大语言模型（Large Language Model）生成的论证文本，以及说服力评分。本数据集专为测量大语言模型说服力的研究而构建，相关细节可参阅博文《Measuring the Persuasiveness of Language Models》：https://www.anthropic.com/news/measuring-model-persuasiveness。

## 数据集说明
本数据集采用CSV格式文件存储，包含以下字段：
- **worker_id**：对该主张标注初始与最终立场的参与者编号
- **claim**：对应生成论证文本的主张内容
- **argument**：生成的论证文本，来源可为人类或大语言模型
- **source**：论证文本的来源（模型名称或"Human"）
- **prompt_type**：用于生成该论证文本的提示词类型
- **rating_initial**：参与者对该主张的初始评分
- **rating_final**：参与者阅读该论证文本后对该主张的最终评分

## 使用示例
python
from datasets import load_dataset
# 加载数据集
dataset = load_dataset("Anthropic/persuasion")

## 联系方式
如有疑问，请发送邮件至[e-mail]（原格式为esin at anthropic dot com）。

## 引用格式
若需引用本研究或数据集，请使用以下BibTeX格式引用：

@online{durmus2024persuasion,
author = {Esin Durmus and Liane Lovitt and Alex Tamkin and Stuart Ritchie and Jack Clark and Deep Ganguli},
title = {Measuring the Persuasiveness of Language Models},
date = {2024-04-09},
year = {2024},
url = {https://www.anthropic.com/news/measuring-model-persuasiveness},
}

原始信息汇总

数据集概述

数据集名称

Persuasion Dataset

数据集摘要

该数据集包含声明以及相应的人工编写和模型生成的论点，以及说服力评分。此数据集用于研究测量语言模型的说服力。

数据集描述

数据集由一个CSV文件组成，包含以下列：

• worker_id: 参与者的ID，标注其对声明的初始和最终立场。
• claim: 生成论点所针对的声明。
• argument: 生成的论点，由人或语言模型产生。
• source: 论点的来源（模型名称或“Human”）。
• prompt_type: 用于生成论点的提示类型。
• rating_initial: 参与者对声明的初始评分。
• rating_final: 参与者阅读论点后对声明的最终评分。

许可证

cc-by-nc-sa-4.0

语言

英语

大小分类

1K<n<10K

联系方式

邮箱：[e-mail]

引用信息

@online{durmus2024persuasion, author = {Esin Durmus and Liane Lovitt and Alex Tamkin and Stuart Ritchie and Jack Clark and Deep Ganguli}, title = {Measuring the Persuasiveness of Language Models}, date = {2024-04-09}, year = {2024}, url = {https://www.anthropic.com/news/measuring-model-persuasiveness}, }

搜集汇总

背景与挑战

背景概述

The Persuasion Dataset by Anthropic is a research-focused collection featuring claims paired with arguments from both humans and language models, annotated with persuasiveness scores. It includes 3,939 rows of data, aimed at analyzing the effectiveness of model-generated persuasive text compared to human efforts.