Diagram of Thought (DoT): An AI Framework that Models Iterative Reasoning in Large Language Models (LLMs) as the Construction of a Directed Acyclic Graph (DAG) within a Single Model

The Diagram of Thought (DoT) framework models iterative reasoning in large language models (LLMs) by constructing a directed acyclic graph (DAG) within a single model. Developed by researchers from Tsinghua University and Shanghai Artificial Intelligence Laboratory, DoT enhances problem-solving by integrating propositions, critiques, refinements, and verifications. This approach ensures logical consistency through Topos theory and improves computational efficiency by eliminating the need for multi-model collaboration. DoT signifies a significant advancement in creating reasoning-specialized models for complex tasks.