annotation-study

This paper explores whether valence features can reflect morality in natural language by analyzing human annotations of moral scenarios, finding significant correlations and achieving a Matthew's correlation coefficient of 0.764 for binary morality classification.