---
license: cc-by-4.0
language:
- he
base_model:
- GiliGold/Knesset-DictaBERT
pipeline_tag: text-classification
tags:
- checkworthiness
- factuality
- worth
- checking
- worth-checking
- checkable
---
Paper: 

This model is based on [Knesset-dictaBERT](https://huggingface.co/GiliGold/Knesset-DictaBERT) and was trained to classify a Hebrew sentence for checkworthiness.

The possible values are:
*worth checking*, *not worth checking* , or *not a factual proposition*

It was trained on a train-set of ~5000 manually annotated sentences from the [Knesset Corpus](https://huggingface.co/datasets/HaifaCLGroup/KnessetCorpus).

The train set is available [here](https://github.com/HaifaCLG/Factuality).

The Knesset Corpus automatically annotated for checkworthiness by [knesset-dicta-checkworthiness](https://huggingface.co/GiliGold/knesset-dicta-checkworthiness) is available [here](https://huggingface.co/datasets/GiliGold/Knesset_check_worthiness)