BERT Regard classification model

This model is the result of a project entitled Towards Controllable Biases in Language Generation. It consists of a BERT classifier (no ensemble) trained on 1.7K samples of biased language.

Regard measures language polarity towards and social perceptions of a demographic (compared to sentiment, which only measures overall language polarity).

BibTeX entry and citation info

@article{sheng2019woman,
  title={The woman worked as a babysitter: On biases in language generation},
  author={Sheng, Emily and Chang, Kai-Wei and Natarajan, Premkumar and Peng, Nanyun},
  journal={arXiv preprint arXiv:1909.01326},
  year={2019}
}
Downloads last month
18,946
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Spaces using sasha/regardv3 4