Benchmarking the rationality of AI decision making using the transitivity axiom

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Song, Kiwon, Jennings III, James M., Davis-Stober, Clintin P.
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866929716004388864
author Song, Kiwon
Jennings III, James M.
Davis-Stober, Clintin P.
author_facet Song, Kiwon
Jennings III, James M.
Davis-Stober, Clintin P.
contents Fundamental choice axioms, such as transitivity of preference, provide testable conditions for determining whether human decision making is rational, i.e., consistent with a utility representation. Recent work has demonstrated that AI systems trained on human data can exhibit similar reasoning biases as humans and that AI can, in turn, bias human judgments through AI recommendation systems. We evaluate the rationality of AI responses via a series of choice experiments designed to evaluate transitivity of preference in humans. We considered ten versions of Meta's Llama 2 and 3 LLM models. We applied Bayesian model selection to evaluate whether these AI-generated choices violated two prominent models of transitivity. We found that the Llama 2 and 3 models generally satisfied transitivity, but when violations did occur, occurred only in the Chat/Instruct versions of the LLMs. We argue that rationality axioms, such as transitivity of preference, can be useful for evaluating and benchmarking the quality of AI-generated responses and provide a foundation for understanding computational rationality in AI systems more generally.
format Preprint
id arxiv_https___arxiv_org_abs_2502_10554
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Benchmarking the rationality of AI decision making using the transitivity axiom
Song, Kiwon
Jennings III, James M.
Davis-Stober, Clintin P.
Artificial Intelligence
Fundamental choice axioms, such as transitivity of preference, provide testable conditions for determining whether human decision making is rational, i.e., consistent with a utility representation. Recent work has demonstrated that AI systems trained on human data can exhibit similar reasoning biases as humans and that AI can, in turn, bias human judgments through AI recommendation systems. We evaluate the rationality of AI responses via a series of choice experiments designed to evaluate transitivity of preference in humans. We considered ten versions of Meta's Llama 2 and 3 LLM models. We applied Bayesian model selection to evaluate whether these AI-generated choices violated two prominent models of transitivity. We found that the Llama 2 and 3 models generally satisfied transitivity, but when violations did occur, occurred only in the Chat/Instruct versions of the LLMs. We argue that rationality axioms, such as transitivity of preference, can be useful for evaluating and benchmarking the quality of AI-generated responses and provide a foundation for understanding computational rationality in AI systems more generally.
title Benchmarking the rationality of AI decision making using the transitivity axiom
topic Artificial Intelligence
url https://arxiv.org/abs/2502.10554