매일 아침에 프롬프트 연습문제 받아보기 (feat. ChatGPT Task)
T
Two_Jay
Tell me a practice problem to master prompt engineering.
### Contents
When creating a problem, the following elements should be included:
- Prompt problem title
- Background situation of the problem
- Requirements in the problem
- Minimum criteria that the completed prompt should satisfy
- Challenging tasks that would be good to achieve as a bonus in the completed prompt
### Guidelines
When creating a problem, you must follow these guidelines:
- Create a practice problem for LLM prompt engineering.
- Check each element and if there are any elements that are impossible to solve or
conflicting requirements, integrate those elements into one direction.
- The result can be a prompt, code, or document.
- Create a problem that a real prompt engineer would be required to solve at a
professional level.
- The minimum requirements should be very very high reliability.
- 'Bonus tasks' should include higher-level requirements that are not covered
in the 'minimum criteria'. These requirements should be sufficiently challenging
and should be achievable on the premise that the minimum criteria have been met.
- 'Minimum criteria' and 'bonus tasks' should remain at a level that provides
appropriate clues to start with, and should not be clear enough to include or imply
the correct answer. The content should be structured so that the user can
creatively solve the problem.
#### Deliverables
- The result of the problem you generate should be a prompt design or a simple agent
design.
- The result should be achievable using LLM.
#### Error case
- The prompt task should include cases that may be considered errors.
- In this case, a simple example should be included.
- The content of the error case should be a case that can occur in reality.
### Output Format
- Insert a divider between each element.
- Do not use markdown format, instead use plain text.
- Respond in Korean.
