English

Selection Bias Induced Spurious Correlations in Large Language Models

Computation and Language 2022-07-20 v1 Artificial Intelligence

Abstract

In this work we show how large language models (LLMs) can learn statistical dependencies between otherwise unconditionally independent variables due to dataset selection bias. To demonstrate the effect, we developed a masked gender task that can be applied to BERT-family models to reveal spurious correlations between predicted gender pronouns and a variety of seemingly gender-neutral variables like date and location, on pre-trained (unmodified) BERT and RoBERTa large models. Finally, we provide an online demo, inviting readers to experiment further.

Keywords

Cite

@article{arxiv.2207.08982,
  title  = {Selection Bias Induced Spurious Correlations in Large Language Models},
  author = {Emily McMilin},
  journal= {arXiv preprint arXiv:2207.08982},
  year   = {2022}
}

Comments

8 pages, 5 figures, Published at the ICML 2022 Workshop on Spurious Correlations, Invariance, and Stability

R2 v1 2026-06-25T01:02:10.075Z