Dark Mode
Global Commonsense Questions Dataset
Education & Learning Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
This dataset, CommonsenseQA (Multiple-Choice Q&A), is designed for question answering that relies on various types of commonsense knowledge. It features 12,102 questions, each presented with a single correct answer and four distractor answers. The primary purpose of this dataset is to facilitate the training and evaluation of models in their ability to accurately predict answers to multiple-choice questions, particularly those requiring commonsense reasoning. It can also be utilised to uncover new forms of commonsense knowledge essential for question answering.
Columns
The dataset includes the following key columns:
- answerKey: A string representing the correct answer to the question.
- choices: A list of strings, providing the four possible answer options for each question.
- question: A string containing the text of the multiple-choice question.
Distribution
The CommonsenseQA dataset comprises 12,102 questions. It is structured into two main evaluation splits: a "Random split" which serves as the primary evaluation method, and a "Question token split." While the exact file format is not specified in the provided information, data files for marketplace listings are typically in CSV format. Specific numbers for rows or records beyond the total question count are not available.
Usage
This dataset is highly suitable for several applications:
- Training AI models: Develop and train models to predict correct answers to multiple-choice questions requiring commonsense understanding.
- Model evaluation: Assess the performance of various AI and machine learning models on commonsense question-answering tasks.
- Knowledge discovery: Research and identify novel types of commonsense knowledge crucial for accurate question answering.
Coverage
The dataset has a global regional coverage. No specific demographic scope or time range information is provided within the sources.
License
CCO
Who Can Use It
This dataset is ideal for:
- AI and Machine Learning Researchers: For developing and testing algorithms in natural language processing and commonsense reasoning.
- NLP Practitioners: To build and improve systems that require an understanding of everyday knowledge.
- Educational Technology Developers: For creating intelligent tutoring systems or advanced learning analytics tools.
Dataset Name Suggestions
- Commonsense Question Answering Dataset
- Multiple-Choice Commonsense Q&A
- AI Commonsense Challenge Data
- Global Commonsense Questions
Attributes
Original Data Source: CommonsenseQA (Multiple-Choice Q&A)
QUALITY