Tree-of-Thought (ToT)
Also known as: Tree-of-Thought / ToT / 思考の木
A reasoning framework that explores multiple thought branches in a tree structure, evaluating and pruning nodes to find optimal solutions — an extension of Chain-of-Thought for more complex problem solving.
Overview
Proposed in 2023 by Princeton/DeepMind, Tree-of-Thought has the LLM explore multiple candidate thoughts in parallel as a tree, then evaluate and select the most promising branch to continue. It is analogous to minimax search in game trees, applied to language reasoning.
When to apply
ToT suits tasks with wide solution spaces requiring search — creative writing, software architecture decisions, and mathematical proofs. It is more compute-intensive than CoT but improves accuracy on hard reasoning tasks. LangChain and LlamaIndex offer ToT modules.
Related Columns
Related Terms
Feel free to contact us
Contact Us