prajwalmani/trivia-qa-20k

Dataset Description

This dataset, trivia-qa-20k, is a cleaned and simplified version of the popular mandarjoshi/trivia_qa dataset. It contains 20,000 high-quality, simple question-answer pairs in English.

The data is structured for ease of use in fine-tuning models for straightforward question-answering tasks, where context or evidence is not required.

Note: This is a cleaned and simplified version of the **mandarjoshi/trivia_qa** dataset.

Dataset Structure

The dataset consists of a single train.csv file, containing 20,000 rows.

Data Fields

Each row in the CSV represents a single trivia pair:

• —`question`: (string) A trivia question.
• —`answer`: (string) The corresponding answer.

Example

Here is an example from the dataset:

question,answer
"Si is the symbol for which chemical element?","silicon"
"What is the capital of France?","Paris"