Measuring the Inconsistency of Large Language Models in Preferential Ranking

TL;DR

This paper evaluates the consistency of large language models in generating preferential rankings, revealing significant biases and violations of logical criteria, which impact their reliability in decision-making tasks.

Contribution

It introduces a formal framework for measuring ranking consistency and empirically demonstrates the limitations of current LLMs in this regard.

Findings

LLMs show strong positional bias in rankings

Preferences are easily influenced by irrelevant alternatives

LLMs exhibit poor transitivity in preferences

Abstract

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.