Can Large Language Models (or Humans) Disentangle Text?

Nicolas Audinet de Pieuchon, Adel Daoud, Connor Jerzak, Moa Johansson, Richard Johansson


Abstract
We investigate the potential of large language models (LLMs) to disentangle text variables—to remove the textual traces of an undesired forbidden variable in a task sometimes known as text distillation and closely related to the fairness in AI and causal inference literature. We employ a range of various LLM approaches in an attempt to disentangle text by identifying and removing information about a target variable while preserving other relevant signals. We show that in the strong test of removing sentiment, the statistical association between the processed text and sentiment is still detectable to machine learning classifiers post-LLM-disentanglement. Furthermore, we find that human annotators also struggle to disentangle sentiment while preserving other semantic content. This suggests there may be limited separability between concept variables in some text contexts, highlighting limitations of methods relying on text-level transformations and also raising questions about the robustness of disentanglement methods that achieve statistical independence in representation space.
Anthology ID:
2024.nlpcss-1.5
Volume:
Proceedings of the Sixth Workshop on Natural Language Processing and Computational Social Science (NLP+CSS 2024)
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Dallas Card, Anjalie Field, Dirk Hovy, Katherine Keith
Venues:
NLP+CSS | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
57–67
Language:
URL:
https://aclanthology.org/2024.nlpcss-1.5
DOI:
10.18653/v1/2024.nlpcss-1.5
Bibkey:
Cite (ACL):
Nicolas Audinet de Pieuchon, Adel Daoud, Connor Jerzak, Moa Johansson, and Richard Johansson. 2024. Can Large Language Models (or Humans) Disentangle Text?. In Proceedings of the Sixth Workshop on Natural Language Processing and Computational Social Science (NLP+CSS 2024), pages 57–67, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
Can Large Language Models (or Humans) Disentangle Text? (Audinet de Pieuchon et al., NLP+CSS-WS 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.nlpcss-1.5.pdf