Read the original document by opening this link in a new tab.
Table of Contents
Abstract
Introduction
Related Work
Algorithm of Thoughts
Summary
The document discusses the Algorithm of Thoughts, a novel strategy for enhancing exploration of ideas in Large Language Models (LLMs). It proposes a new mode of in-context learning by employing algorithmic reasoning pathways. The approach aims to reduce query counts in reasoning methods while maintaining performance, leading to a more responsible use of AI resources. The paper outlines the evolution of LLMs, the integration of algorithms for internalizing iterative logic, and the comparison with existing prompting strategies like Chain-of-Thought and Tree of Thoughts.