Guardado en:
Detalles Bibliográficos
Autores principales: Zhao, Xiutian, Wang, Ke, Peng, Wei
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2410.08851
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910645443624960
author Zhao, Xiutian
Wang, Ke
Peng, Wei
author_facet Zhao, Xiutian
Wang, Ke
Peng, Wei
contents Despite large language models' (LLMs) recent advancements, their bias and hallucination issues persist, and their ability to offer consistent preferential rankings remains underexplored. This study investigates the capacity of LLMs to provide consistent ordinal preferences, a crucial aspect in scenarios with dense decision space or lacking absolute answers. We introduce a formalization of consistency based on order theory, outlining criteria such as transitivity, asymmetry, reversibility, and independence from irrelevant alternatives. Our diagnostic experiments on selected state-of-the-art LLMs reveal their inability to meet these criteria, indicating a strong positional bias and poor transitivity, with preferences easily swayed by irrelevant alternatives. These findings highlight a significant inconsistency in LLM-generated preferential rankings, underscoring the need for further research to address these limitations.
format Preprint
id arxiv_https___arxiv_org_abs_2410_08851
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Measuring the Inconsistency of Large Language Models in Preferential Ranking
Zhao, Xiutian
Wang, Ke
Peng, Wei
Computation and Language
Despite large language models' (LLMs) recent advancements, their bias and hallucination issues persist, and their ability to offer consistent preferential rankings remains underexplored. This study investigates the capacity of LLMs to provide consistent ordinal preferences, a crucial aspect in scenarios with dense decision space or lacking absolute answers. We introduce a formalization of consistency based on order theory, outlining criteria such as transitivity, asymmetry, reversibility, and independence from irrelevant alternatives. Our diagnostic experiments on selected state-of-the-art LLMs reveal their inability to meet these criteria, indicating a strong positional bias and poor transitivity, with preferences easily swayed by irrelevant alternatives. These findings highlight a significant inconsistency in LLM-generated preferential rankings, underscoring the need for further research to address these limitations.
title Measuring the Inconsistency of Large Language Models in Preferential Ranking
topic Computation and Language
url https://arxiv.org/abs/2410.08851