Saved in:
Bibliographic Details
Main Authors: Moon, Suhong, Abdulhai, Marwa, Kang, Minwoo, Suh, Joseph, Soedarmadji, Widyadewi, Behar, Eran Kohen, Chan, David M., Canny, John
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2407.06576
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915999093096448
author Moon, Suhong
Abdulhai, Marwa
Kang, Minwoo
Suh, Joseph
Soedarmadji, Widyadewi
Behar, Eran Kohen
Chan, David M.
Canny, John
author_facet Moon, Suhong
Abdulhai, Marwa
Kang, Minwoo
Suh, Joseph
Soedarmadji, Widyadewi
Behar, Eran Kohen
Chan, David M.
Canny, John
contents Large language models (LLMs) are trained from vast repositories of text authored by millions of distinct authors, reflecting an enormous diversity of human traits. While these models bear the potential to be used as approximations of human subjects in behavioral studies, prior efforts have been limited in steering model responses to match individual human users. In this work, we introduce "Anthology", a method for conditioning LLMs to particular virtual personas by harnessing open-ended life narratives, which we refer to as "backstories." We show that our methodology enhances the consistency and reliability of experimental outcomes while ensuring better representation of diverse sub-populations. Across three nationally representative human surveys conducted as part of Pew Research Center's American Trends Panel (ATP), we demonstrate that Anthology achieves up to 18% improvement in matching the response distributions of human respondents and 27% improvement in consistency metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2407_06576
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Virtual Personas for Language Models via an Anthology of Backstories
Moon, Suhong
Abdulhai, Marwa
Kang, Minwoo
Suh, Joseph
Soedarmadji, Widyadewi
Behar, Eran Kohen
Chan, David M.
Canny, John
Computation and Language
Artificial Intelligence
Large language models (LLMs) are trained from vast repositories of text authored by millions of distinct authors, reflecting an enormous diversity of human traits. While these models bear the potential to be used as approximations of human subjects in behavioral studies, prior efforts have been limited in steering model responses to match individual human users. In this work, we introduce "Anthology", a method for conditioning LLMs to particular virtual personas by harnessing open-ended life narratives, which we refer to as "backstories." We show that our methodology enhances the consistency and reliability of experimental outcomes while ensuring better representation of diverse sub-populations. Across three nationally representative human surveys conducted as part of Pew Research Center's American Trends Panel (ATP), we demonstrate that Anthology achieves up to 18% improvement in matching the response distributions of human respondents and 27% improvement in consistency metrics.
title Virtual Personas for Language Models via an Anthology of Backstories
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2407.06576