Skip to content

Dataset · mteb

Tweet sentiment extraction for MTEB

mteb/tweet_sentiment_extraction

A set of English tweets from a Kaggle competition for evaluating and training text embeddings in the sentiment analysis task.

What's inside

The dataset collects English social media texts for the text-classification sentiment task. The data volume ranges from 10 to 100 thousand examples. Originally taken from a Kaggle competition, the current version is prepared for evaluating embedding models within the MTEB benchmark.

What you can do with it

  • Evaluating quality of text embedding models via MTEB library
  • Training sentiment classifiers for short social media posts
Tasks
Text classification
Languages
English
Size
10K–100K rows
Downloads this month
5.1K
Likes
38

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("mteb/tweet_sentiment_extraction")