Stepmothers are mean and academics are pretentious: What do pretrained language models learn about you?
In this paper, we investigate what types of stereotypical information are captured by pretrained language models. We present the first dataset comprising stereotypical attributes of a range of social groups and propose a method to elicit stereotypes encoded by pretrained language models in an unsupervised fashion. Moreover, we link the emergent stereotypes to their manifestation as basic emotions as a means to study their emotional effects in a more generalized manner. To demonstrate how our methods can be used to analyze emotion and stereotype shifts due to linguistic experience, we use fine-tuning on news sources as a case study. Our experiments expose how attitudes towards different social groups vary across models and how quickly emotions and stereotypes can shift at the fine-tuning stage.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
A Bayesian Model of the Litigation Game
Over a century ago, Oliver Wendell Holmes invited scholars to look at the law through the lens of probability theory: "The prophecies of what the courts will do in fact, and nothing more pretentious, are what I mean by t…
modelChoose Your Weapon: Survival Strategies for Depressed AI Academics
Are you an AI researcher at an academic institution? Are you anxious you are not coping with the current pace of AI advancements? Do you feel you have no (or very limited) access to the computational and human resources …
The Proper Use of Google Trends in Forecasting Models
It is widely known that Google Trends have become one of the most popular free tools used by forecasters both in academics and in the private and public sectors. There are many papers, from several different fields, conc…
KeypartX: Graph-based Perception (Text) Representation
The availability of big data has opened up big opportunities for individuals, businesses and academics to view big into what is happening in their world. Previous works of text representation mostly focused on informativ…
InformativenessPriors Persist Through Suppression: A Stroop Paradigm for Lexical Override
Glossaries, technical specifications, and system prompts routinely ask language models to use familiar words in unfamiliar ways. The instruction competes with what the word already means, and even when it wins, the pretr…