Tree-of-thought (ToT) is a prompting technique that asks a language model to explore several intermediate reasoning paths instead of following one sequence from start to finish. It then compares the options and selects the most promising direction. section 7.4
A ToT prompt typically does the following:
- Generates several candidate strategies or hypotheses in parallel. section 7.4
- Evaluates the benefits, trade-offs, or promise of each candidate. section 7.4
- Selects a direction rather than committing immediately to the first reasoning chain. section 7.4
Customer-retention strategies
A prompt might ask for three ways to improve customer retention, request the potential benefits of each, and then require a recommendation. Possible branches include subscription models, loyalty programs, and personalized offers. The model compares these options before converging on one approach. section 7.4
Unlike chain-of-thought prompting, which produces one step-by-step sequence, ToT maintains multiple possible paths at each stage. This branching can help with logic puzzles, scientific problem-solving, and search-heavy tasks where divergent thinking matters. Its benefits are not automatic: prompting approaches still require evaluation, and post-training or hybrid methods may be more suitable in some cases. section 7.4 section 7.13
Table 7.3 Common advanced prompting techniques
| Technique |
Description |
Example prompt |
|---|---|---|
| Chain-of-thought |
Generate step-by-step reasoning before answering |
“Let’s think step by step. John has 3 apples . . .” |
| Tree-of-thought |
Explore multiple reasoning paths in parallel |
“List three strategies, evaluate each, then choose the best.” |
| Self-consistency |
Generate multiple reasoning chains and aggregate |
“Generate five solutions, then select the most common answer.” |
| Reflexion |
Critique and revise output iteratively |
“Review your answer for errors and revise if needed.” |
| Verification |
Add a final self-check before finalizing the response |
“Double-check your answer for accuracy before submitting.” |
| Decomposition |
Break the problem into subtasks executed sequentially |
“Step 1: Extract metrics. Step 2: Summarize trends. Step 3: Write summary.” |
| Least-to-most |
Solve easier subproblems first, then build up |
“First, list known variables. Then calculate step by step.” |
| Pyramid priming |
Build knowledge gradually in layers |
“Step 1: Define carbon neutrality. Step 2: List targets. Step 3: Analyze impact.” |
| Generated knowledge |
Surface key facts before answering |
“List key concepts related to quantum entanglement before answering.” |
| Role-based prompting |
Assign a role or persona to guide behavior |
“You are an experienced employment lawyer advising a client.” |
| Prompt chaining |
Connect multiple prompts into a pipeline |
“Extract relevant quotes → Summarize → Generate final answer.” |
| Conversational context |
Summarize prior dialogue to preserve coherence |
“Summarize the student’s understanding before proceeding.” |
Compare tree-of-thought with chain-of-thought, self-consistency, reflexion, verification, decomposition, and other advanced prompting techniques. section 7.4
