A Systematic Comparison of Syllogistic Reasoning in Humans and Language Models

TL;DR

This study compares human and language model syllogistic reasoning, revealing that larger models are more logical but still exhibit human-like biases and errors, with some models surpassing human logical performance.

Contribution

It systematically analyzes how transformer language models replicate or overcome human syllogistic reasoning biases, highlighting size-dependent improvements and persistent errors.

Findings

Larger models are more logical than smaller ones and humans.

Models show sensitivity to irrelevant variable ordering.

Models sometimes make confident, incorrect inferences.

Abstract

A central component of rational behavior is logical inference: the process of determining which conclusions follow from a set of premises. Psychologists have documented several ways in which humans' inferences deviate from the rules of logic. Do language models, which are trained on text generated by humans, replicate such human biases, or are they able to overcome them? Focusing on the case of syllogisms -- inferences from two simple premises -- we show that, within the PaLM2 family of transformer language models, larger models are more logical than smaller ones, and also more logical than humans. At the same time, even the largest models make systematic errors, some of which mirror human reasoning biases: they show sensitivity to the (irrelevant) ordering of the variables in the syllogism, and draw confident but incorrect inferences from particular syllogisms (syllogistic fallacies). Overall, we find that language models often mimic the human biases included in their training data, but are able to overcome them in some cases.