Scribd
Upload
4 views2 pages

Prompt Design Experiment

Description

Uploaded by

nurbashanagababu4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
4 views2 pages

Prompt Design Experiment

Description

Uploaded by

nurbashanagababu4
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Prompt Design Experiment

Experimenting with Zero-Shot, Few-Shot, and Chain-of-Thought Prompting

Experiment Setup:

Task/Topic: Solving a math problem (Finding the area under a curve).

Goal: Use each prompt technique (zero-shot, few-shot, and chain-of-thought) to solve the problem.

1. Zero-Shot Prompt:

Task: Find the area under the curve f(x) = x^2 from x = 1 to x = 3.

Response: The area under the curve can be found by integrating f(x) = x^2 over the interval from x =

1 to x = 3. The integral of x^2 is x^3/3. Evaluating from 1 to 3: 3^3/3 - 1^3/3 = 9 - 1/3 = 26/3. The

area is 26/3.

Effectiveness: Correct answer, but lacks explanation.

2. Few-Shot Prompt:

Examples:

Example 1: Find the area under the curve f(x) = 2x from x = 1 to x = 4. Answer: 15.

Example 2: Find the area under the curve f(x) = 3x^2 from x = 0 to x = 2. Answer: 8.

Task: Find the area under the curve f(x) = x^2 from x = 1 to x = 3.

Response: The model correctly follows the examples to solve the task, outputting the correct result.

Effectiveness: The examples improved the response, but there's still a lack of explanation.

3. Chain-of-Thought Prompt:

Task: Find the area under the curve f(x) = x^2 from x = 1 to x = 3.

Response:

1. To find the area, we compute the definite integral of f(x) = x^2 over [1, 3].

2. The integral of f(x) is x^3/3.


3. Evaluate this at the bounds: at x = 3, 3^3/3 = 9; at x = 1, 1^3/3 = 1/3.

4. The area is 9 - 1/3 = 26/3.

Effectiveness: Step-by-step reasoning helped the model clearly explain each part, improving clarity

and accuracy.

Analysis of Effectiveness:

Zero-Shot: Quick and efficient for simple tasks, but lacks reasoning or explanation for complex

problems.

Few-Shot: Useful for tasks that benefit from pattern recognition through examples, but might still

lack clarity.

Chain-of-Thought: Ideal for tasks requiring deep reasoning, where step-by-step explanation leads to

better understanding and accuracy.

Conclusion:

Choosing the right prompting technique depends on the task. For simple queries, zero-shot may

suffice, while few-shot works well for tasks needing examples. Chain-of-thought is most effective for

complex reasoning tasks, though it requires more time and explanation.

You might also like