# Sentiment Analysis outputs

**URL:** <https://discuss.huggingface.co/t/sentiment-analysis-outputs/27660>\
**Category:** Beginners\
**Created:** [December 11, 2022, 8:12pm UTC](https://discuss.huggingface.co/t/sentiment-analysis-outputs/27660 "2022-12-11T20:12:06Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![JohnnyOmega](https://avatars.discourse-cdn.com/v4/letter/j/3ec8ea/32.png) [@JohnnyOmega](https://discuss.huggingface.co/u/JohnnyOmega)\
**Post date:** [December 11, 2022, 8:12pm UTC](https://discuss.huggingface.co/t/sentiment-analysis-outputs/27660/1 "2022-12-11T20:12:06Z")

</div>

Hello. Sentiment analysis models such as [bertweet](https://huggingface.co/finiteautomata/bertweet-base-sentiment-analysis) give two outputs. A classification and a number. Is that number a measure of confidence in the classification? Or is it a probability?

Also, do I need to do any sort of tokenizing or preprocessing before feeding these models text? Or does transformers.pipeline do some of this preprocessing for me?

Thanks.