Saved in:
Bibliographic Details
Main Authors: Chance, Christina, Yin, Da, Wang, Dakuo, Chang, Kai-Wei
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2310.10865
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910149402165248
author Chance, Christina
Yin, Da
Wang, Dakuo
Chang, Kai-Wei
author_facet Chance, Christina
Yin, Da
Wang, Dakuo
Chang, Kai-Wei
contents In this paper, we study whether language models are affected by learned gender stereotypes during the comprehension of stories. Specifically, we investigate how models respond to gender stereotype perturbations through counterfactual data augmentation. Focusing on Question Answering (QA) tasks in fairytales, we modify the FairytaleQA dataset by swapping gendered character information and introducing counterfactual gender stereotypes during training. This allows us to assess model robustness and examine whether learned biases influence story comprehension. Our results show that models exhibit slight performance drops when faced with gender perturbations in the test set, indicating sensitivity to learned stereotypes. However, when fine-tuned on counterfactual training data, models become more robust to anti-stereotypical narratives. Additionally, we conduct a case study demonstrating how incorporating counterfactual anti-stereotype examples can improve inclusivity in downstream applications.
format Preprint
id arxiv_https___arxiv_org_abs_2310_10865
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Will the Prince Get True Love's Kiss? On the Model Sensitivity to Gender Perturbation over Fairytale Texts
Chance, Christina
Yin, Da
Wang, Dakuo
Chang, Kai-Wei
Computation and Language
In this paper, we study whether language models are affected by learned gender stereotypes during the comprehension of stories. Specifically, we investigate how models respond to gender stereotype perturbations through counterfactual data augmentation. Focusing on Question Answering (QA) tasks in fairytales, we modify the FairytaleQA dataset by swapping gendered character information and introducing counterfactual gender stereotypes during training. This allows us to assess model robustness and examine whether learned biases influence story comprehension. Our results show that models exhibit slight performance drops when faced with gender perturbations in the test set, indicating sensitivity to learned stereotypes. However, when fine-tuned on counterfactual training data, models become more robust to anti-stereotypical narratives. Additionally, we conduct a case study demonstrating how incorporating counterfactual anti-stereotype examples can improve inclusivity in downstream applications.
title Will the Prince Get True Love's Kiss? On the Model Sensitivity to Gender Perturbation over Fairytale Texts
topic Computation and Language
url https://arxiv.org/abs/2310.10865