HomeResearch LibraryLeast-to-Most Prompting Enables Complex Reasoning in La…
system architectureChapter 2ICLR 2023 · 2022

Least-to-Most Prompting Enables Complex Reasoning in Large Language Models

Denny Zhou (Google Brain), Nathanael Schärli (Google Brain)

Abstract

We propose least-to-most prompting, a technique that decomposes complex problems into simpler subproblems and solves them sequentially, with each solution building on previous ones.

Eigenvector Insight — Zone III / PASF-PADE AnalysisNot part of the original paper
Eigenvector Research — Marco van Hurne
How this paper contributes to solving the Zone III problem (PASF-PADE)

Least-to-most prompting formalizes the task decomposition pattern that is central to Zone III workflows. The principle of solving simpler subproblems first and building up to complex solutions is the foundation of reliable long-horizon execution.

Why AI is not sufficient for Zone III without this

Zone III refers to high-complexity, high-risk, long-running agentic workflows — the class of enterprise AI deployments where a single failure can cascade across hundreds of steps. Standard AI models, trained to predict the next token, are not inherently designed for durable, governed, multi-step execution. This paper addresses one or more of the structural gaps that make Zone III deployments unsafe without explicit architectural intervention.

Key Contributions

  • →Least-to-most decomposition
  • →Sequential subproblem solving
  • →Compositional generalization

Topics

task decompositionpromptinghierarchical reasoningsubproblem solving