Evolutionary System 2 Reasoning: An Empirical Proof

TL;DR

This paper introduces an evolutionary framework to enhance reasoning abilities in large language models, demonstrating that even weaker models can develop strong reasoning skills through evolutionary optimization.

Contribution

The paper proposes the ERO framework that evolves LLMs to improve their reasoning ability, showing that simple evolutionary strategies can significantly enhance weaker models.

Findings

GPT-5 shows limited reasoning ability

Weak models can be evolved to strong reasoners

Evolutionary optimization improves reasoning performance

Abstract

Machine intelligence marks the ultimate dream of making machines' intelligence comparable to human beings. While recent progress in Large Language Models (LLMs) show substantial specific skills for a wide array of downstream tasks, they more or less fall shorts in general intelligence. Following correlation between intelligence and system 2 reasoning (slow thinking), in this paper, we aim to answering a worthwhile research question: could machine intelligence such as LLMs be evolved to acquire reasoning ability (not specific skill) just like our human beings? To this end, we propose evolutionary reasoning optimization (ERO) framework which performs survival of the fittest over a population of LLMs to search for individual with strong reasoning ability. Given a reasoning task, ERO first initializes multiple LLMs as a population, after which an evolutionary strategy evolves the population to maximize quantified reasoning score of the best individual. Based on experiments on representative testsuites, we claim two surprising empirical discoveries: i) the latest LLMs such as GPT-5 still show limited system 2 reasoning ability; ii) with simple evolution-loop of ERO, a relatively weak model (Qwen-7B) could be enhanced to emerge powerful reasoning ability. Our project can be accessed at https://github.com/MetaEvo/ERO for reproduction needs.