TaatiTeam/OCW

Only Connect Wall (OCW)数据集包含618个来自Only Connect quiz show的Connecting Walls，涵盖了15个赛季的节目内容。每个wall包含真实的分组和连接信息，以及记录的人类表现。数据集可用于文本分类任务，并提供了两个不同版本的数据集，分别减少了或完全移除了每个wall中的干扰项。The Only Connect Wall (OCW) dataset contains 618 Connecting Walls sourced from the Only Connect quiz show, spanning 15 seasons of the program. Each wall includes authentic grouping and connection information, along with recorded human performance data. This dataset is applicable to text classification tasks, and two distinct variants are provided, which respectively reduce or completely remove the distractor items within each wall.

原始信息汇总

Only Connect Wall Dataset

概述

• 名称: Only Connect Wall Dataset
• 许可证: MIT
• 任务类别: 文本分类
• 语言: 英语
• 标签:
  • creative problem solving
  • puzzles
  • fixation effect
  • large language models
  • only connect
  • quiz show
  • connecting walls
• 大小类别: n<1K

数据集详情

• 来源: Only Connect quiz show 的 Round 3: Connecting Wall 环节
• 内容: 包含618个“Connecting Walls”，来自15季的节目
• 数据: 每个墙包含实际的组和连接，以及记录的人类表现

使用方法

python

pip install datasets

from datasets import load_dataset

dataset = load_dataset("TaatiTeam/OCW")

示例：获取训练集中第一个示例的wall_id

dataset["train"]["wall_id"][0]

或获取测试集中前10个示例的words

dataset["test"]["words"][0:10]

额外版本

• ocw_randomized: 显著减少每个墙中的红鲱鱼
• ocw_wordnet: 完全移除每个墙中的红鲱鱼

python

pip install datasets

from datasets import load_dataset

ocw_randomized = load_dataset("TaatiTeam/OCW", "ocw_randomized") ocw_wordnet = load_dataset("TaatiTeam/OCW", "ocw_wordnet")

引用

@article{alavi2024large, title={Large Language Models are Fixated by Red Herrings: Exploring Creative Problem Solving and Einstellung Effect using the Only Connect Wall Dataset}, author={Alavi Naeini, Saeid and Saqur, Raeid and Saeidi, Mozhgan and Giorgi, John and Taati, Babak}, journal={Advances in Neural Information Processing Systems}, volume={36}, year={2024} }