ChainOfThoughtStrategy<T>
Models & Types · AiDotNet.Reasoning.Strategies
Implements Chain-of-Thought (CoT) reasoning that solves problems through explicit step-by-step thinking.
Inheritance: ReasoningStrategyBase<T> → ChainOfThoughtStrategy<T>
For Beginners
Chain-of-Thought (CoT) is a reasoning approach where the AI explicitly shows its work, step by step, similar to how you would solve a math problem by writing down each step.
How it works: Given: “What is 15% of 240?”
Step 1: Convert percentage to decimal
- 15% = 15/100 = 0.15
Step 2: Multiply by the number
- 0.15 × 240 = 36
Step 3: State the answer
- The answer is 36
Why it’s effective:
- Makes reasoning transparent and verifiable
- Catches logical errors early
- Improves accuracy on complex problems
- Allows debugging when answers are wrong
Based on research: “Chain-of-Thought Prompting Elicits Reasoning in Large Language Models” (Wei et al., 2022) showed 3-5x improvements on reasoning tasks when models show their work.
How It Works
Example Usage:
Constructors
| Constructor | Summary |
|---|---|
ChainOfThoughtStrategy(IChatClient<T> chatModel, IEnumerable<IAgentTool> tools, bool useJsonFormat) |
Initializes a new instance of the ChainOfThoughtStrategy class. |
Properties
| Property | Summary |
|---|---|
string Description { get; } |
|
string StrategyName { get; } |