Whose Emotions and Moral Sentiments Do Language Models Reflect?

arXiv:2402.11114v1 Announce Type: new
Abstract: Language models (LMs) are known to represent the perspectives of some social groups better than others, which may impact their performance, especially on subjective tasks such as content moderation and hate speech detection. To explore how LMs represent different perspectives, existing research focused on positional alignment, i.e., how closely the models mimic the opinions and stances of different groups, e.g., liberals or conservatives. However, human communication also encompasses emotional and moral dimensions. We define the …

abstract alignment arxiv content moderation cs.cl cs.cy cs.si detection emotions explore hate speech hate speech detection impact language language models lms moderation performance perspectives research social speech tasks type