Benchmarking progress in Natural Language Processing applications to Moral Foundations Theory.


Moral Foundations Theory (MFT) is a theory of moral psychology that proposes that several univeral foundations underlie human morality. Since its inception, various methods have been proposed to operationalize MFT on natural language data, including dictionary-based approaches and deep neural networks.

Natural Language Processing (NLP) tasks benefit from standardized dataset, model, and performance repositories. See HuggingFace Dataset Hub, HuggingFace Model Hub , and PaperswithCode.

NLP applications of MFT have some commonalities with mainstream NLP tasks, but also have some differences. This site aims to cater to the unique needs of MFT NLP approaches, and synchronize with standard ecosystem tools wherever possible.

This site aims to provide an inventory of tasks, datasets, NLP methods for MFT and their performance results, to facilitate the development of improved methods. It is mainly intended as a resource for researchers in NLP, MFT, and related fields.

Moral Sentiment Analysis

Moral Sentiment Analysis is the task of classifying the moral sentiment of a text.