Studying and improving reasoning in humans and machines

TL;DR

This study compares reasoning in humans and large language models using cognitive psychology tools, revealing similarities and differences, and highlighting the evolution of LLM reasoning capabilities and their response to prompting strategies.

Contribution

It introduces a novel comparative framework for analyzing human and LLM reasoning using classical cognitive experiments, revealing model limitations and improvements over recent releases.

Findings

Most models exhibit reasoning errors similar to humans.

Recent LLMs show significant reduction in reasoning limitations.

Humans and LLMs respond differently to prompting strategies.

Abstract

In the present study, we investigate and compare reasoning in large language models (LLM) and humans using a selection of cognitive psychology tools traditionally dedicated to the study of (bounded) rationality. To do so, we presented to human participants and an array of pretrained LLMs new variants of classical cognitive experiments, and cross-compared their performances. Our results showed that most of the included models presented reasoning errors akin to those frequently ascribed to error-prone, heuristic-based human reasoning. Notwithstanding this superficial similarity, an in-depth comparison between humans and LLMs indicated important differences with human-like reasoning, with models limitations disappearing almost entirely in more recent LLMs releases. Moreover, we show that while it is possible to devise strategies to induce better performance, humans and machines are not equally-responsive to the same prompting schemes. We conclude by discussing the epistemological implications and challenges of comparing human and machine behavior for both artificial intelligence and cognitive psychology.