зеркало из
https://github.com/iharh/notes.git
synced 2025-10-29 20:56:06 +02:00
32 строки
1.2 KiB
Plaintext
32 строки
1.2 KiB
Plaintext
8 components
|
|
|
|
1. role
|
|
"Act as business analyst," or "You are a creative and inspirational technical expert at EPAM, with an extensive technical vocabulary."
|
|
|
|
2. context
|
|
?
|
|
|
|
3. instruction
|
|
The Instruction could be anything from summarizing text to generating a response to a specific question,
|
|
but the more precise and succinct you define the ask,
|
|
the better the chances are that you will receive the desired result.
|
|
|
|
4. data
|
|
The Data component includes any relevant data or information that the LLM should consider when generating a response.
|
|
This could include specific keywords or phrases, documents, or other relevant information.
|
|
|
|
5. output
|
|
defines the format and structure of the desired output
|
|
|
|
6. example
|
|
provides an example of the type of response the LLM should generate
|
|
|
|
7. constraints
|
|
defines any limitations that the LLM should work within when generating a response.
|
|
This could include time constraints, computational resources, or any other relevant limitations.
|
|
This is where you fine-tune the model to get your desired response.
|
|
|
|
8. evaluation criteria
|
|
defines how the output generated by the LLM will be evaluated.
|
|
this may include criteria for accuracy, specificity, or any other relevant factors.
|