Which technique asks a model to reason through intermediate steps?
Why D is the answer
Chain-of-thought prompting
Chain of thought spells out reasoning. The others do not change how the task is framed.
In detail
- Chain of thought asks for intermediate steps.
- Steps can be wrong too, so check them.
Use case
An illustrative scenario, not a real incident.
A model works through a billing problem step by step, so the team can see where it went wrong.
See it in different ways
flowchart TD
A{"What does the model need?"} -->|"Clear task only"| B["Zero-shot"]
A -->|"An example format"| C["Few-shot"]
A -->|"Reasoning"| D["Chain of thought"]stateDiagram-v2 [*] --> Tokenize Tokenize --> Pretraining Pretraining --> FineTuning FineTuning --> RLHF RLHF --> Evaluation Evaluation --> [*]
flowchart LR S0["Question"] --> S1["Intermediate steps"] --> S2["Answer"]
Key point: Chain of thought spells out reasoning.
Why the other options are wrong
- A. Token streaming
- Streaming changes delivery.
- B. Top-k sampling
- Top-k changes sampling.
- C. Weight quantization
- Quantization changes weights.