Authors: Maciej Besta, Nils Blach, Ales Kubicek, Robert Gerstenberger, Michal Podstawski, Lukas Gianinazzi, Joanna Gajda, Tomasz Lehmann, Hubert Niewiadomski, Piotr Nyczyk, Torsten Hoefler

Published on: August 18, 2023

Impact Score: 8.22

Arxiv code: Arxiv:2308.09687

Summary

• What is new: Introduces Graph of Thoughts (GoT), a new framework for enhancing prompting capabilities in large language models (LLMs) beyond the current Chain-of-Thought and Tree of Thoughts paradigms.
• Why this is important: Existing prompting methods in LLMs are limited in how they can manage and leverage the relationships between different units of information.
• What the research proposes: GoT models the information generated by LLMs as an arbitrary graph, allowing for more sophisticated handling and combination of thoughts.
• Results: Demonstrated a 62% improvement in sorting task quality and a 31% reduction in costs compared to the Tree of Thoughts approach.

Technical Details

Technological frameworks used: Graph of Thoughts (GoT)

Models used: Large Language Models (LLMs)

Data used: nan

Potential Impact

This work could disrupt markets that rely on sophisticated data analysis and decision-making processes, including tech companies specializing in AI and machine learning, as well as industries leveraging these technologies for innovation.

Want to implement this idea in a business?

We have generated a startup concept here: MindGraph.