Generative Language Models Exhibit Social Identity Biases

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Hu, Tiancheng, Kyrychenko, Yara, Rathje, Steve, Collier, Nigel, van der Linden, Sander, Roozenbeek, Jon
Format:	Preprint
Published:	2023
Subjects:	Computation and Language Computers and Society
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866913393073455104
author	Hu, Tiancheng Kyrychenko, Yara Rathje, Steve Collier, Nigel van der Linden, Sander Roozenbeek, Jon
author_facet	Hu, Tiancheng Kyrychenko, Yara Rathje, Steve Collier, Nigel van der Linden, Sander Roozenbeek, Jon
contents	The surge in popularity of large language models has given rise to concerns about biases that these models could learn from humans. We investigate whether ingroup solidarity and outgroup hostility, fundamental social identity biases known from social psychology, are present in 56 large language models. We find that almost all foundational language models and some instruction fine-tuned models exhibit clear ingroup-positive and outgroup-negative associations when prompted to complete sentences (e.g., "We are..."). Our findings suggest that modern language models exhibit fundamental social identity biases to a similar degree as humans, both in the lab and in real-world conversations with LLMs, and that curating training data and instruction fine-tuning can mitigate such biases. Our results have practical implications for creating less biased large-language models and further underscore the need for more research into user interactions with LLMs to prevent potential bias reinforcement in humans.
format	Preprint
id	arxiv_https___arxiv_org_abs_2310_15819
institution	arXiv
publishDate	2023
record_format	arxiv
spellingShingle	Generative Language Models Exhibit Social Identity Biases Hu, Tiancheng Kyrychenko, Yara Rathje, Steve Collier, Nigel van der Linden, Sander Roozenbeek, Jon Computation and Language Computers and Society The surge in popularity of large language models has given rise to concerns about biases that these models could learn from humans. We investigate whether ingroup solidarity and outgroup hostility, fundamental social identity biases known from social psychology, are present in 56 large language models. We find that almost all foundational language models and some instruction fine-tuned models exhibit clear ingroup-positive and outgroup-negative associations when prompted to complete sentences (e.g., "We are..."). Our findings suggest that modern language models exhibit fundamental social identity biases to a similar degree as humans, both in the lab and in real-world conversations with LLMs, and that curating training data and instruction fine-tuning can mitigate such biases. Our results have practical implications for creating less biased large-language models and further underscore the need for more research into user interactions with LLMs to prevent potential bias reinforcement in humans.
title	Generative Language Models Exhibit Social Identity Biases
topic	Computation and Language Computers and Society
url	https://arxiv.org/abs/2310.15819