Tree of Thoughts

Introduction

Tree of Thoughts (ToT) generalises CoT into a search — at each step the model proposes multiple 'thoughts', evaluates them, and explores the most promising branch. It tackles problems where CoT alone gets stuck.

Beginner analogy: Solving a maze by trying multiple paths and backtracking — not just charging forward.

Understanding the topic

Core concepts to understand:

• Generate K thoughts at each step.
• Evaluate each branch (model self-rates or external scorer).
• Expand the best branch; prune the rest.
• Used for: planning, code synthesis, complex puzzles.
• Much more expensive than CoT; reserve for hard problems.

Syntax reference

Visual workflow / architecture:

       Root
      /  |  \
    A    B    C   ← generate
   /|   /|   /|
  ...evaluate each...
   ▼
   pick best, expand again

Real-world use

Used in research-grade prompting (Google's ToT paper) and high-stakes agent planning. Tools like LangGraph implement ToT-style flows.

Best practices

• Use only when CoT and self-consistency aren't enough.
• Cap depth and branching to control cost.
• Score branches with a fast cheap model.

Hands-on exercise

Interview preparation — practice these questions:

• Q1. ToT vs CoT?
• Q2. How do you evaluate branches?
• Q3. Cost concerns?
• Q4. When is ToT overkill?