wehub-resource-sync
111 lines
4.2 KiB
Python
111 lines
4.2 KiB
Python
class COPROTemplate:
|
|
|
|
@staticmethod
|
|
def generate_bootstrap_guidelines(
|
|
original_prompt: str, breadth: int
|
|
) -> str:
|
|
return f"""You are an expert prompt engineer. I need to generate {breadth} distinct, high-quality variations of the following prompt.
|
|
|
|
[ORIGINAL PROMPT]
|
|
{original_prompt}
|
|
|
|
[INSTRUCTIONS]
|
|
Brainstorm exactly {breadth} diverse "Variation Guidelines". Each guideline should be a 1-2 sentence strategy on how to significantly alter or improve the prompt (e.g., changing the tone, adding reasoning steps, enforcing specific output formats, reordering instructions).
|
|
Make sure the guidelines are completely distinct from one another to ensure a wide search space.
|
|
|
|
**
|
|
IMPORTANT: You must only return in JSON format matching the schema.
|
|
Example JSON:
|
|
{{
|
|
"guidelines": [
|
|
"Reframe the prompt to require step-by-step chain of thought before providing the final answer.",
|
|
"Condense the instructions into a highly aggressive, concise format avoiding any pleasantries.",
|
|
"Add strict formatting constraints requiring the output to be bulleted."
|
|
]
|
|
}}
|
|
**
|
|
|
|
JSON:
|
|
"""
|
|
|
|
@staticmethod
|
|
def generate_history_guidelines(
|
|
original_prompt: str, history_text: str, breadth: int
|
|
) -> str:
|
|
return f"""You are an expert prompt engineer and diagnostic system. We are using Coordinate Ascent to optimize a prompt.
|
|
|
|
[ORIGINAL PROMPT]
|
|
{original_prompt}
|
|
|
|
[PAST ATTEMPTS, SCORES, & EVALUATION FEEDBACK]
|
|
{history_text}
|
|
|
|
[INSTRUCTIONS]
|
|
Analyze the [PAST ATTEMPTS, SCORES, & EVALUATION FEEDBACK]. Higher scores are better.
|
|
Crucially, look at the "Evaluation Feedback" for each attempt. This tells you exactly why the prompt lost points (e.g., failed a toxicity metric, missed a formatting constraint).
|
|
|
|
Based on this analysis, brainstorm exactly {breadth} new "Variation Guidelines" to try next.
|
|
These guidelines MUST explicitly address and fix the errors mentioned in the evaluation feedback while maintaining the successful traits of the highest-scoring prompts.
|
|
|
|
**
|
|
IMPORTANT: You must only return in JSON format matching the schema.
|
|
Example JSON:
|
|
{{
|
|
"guidelines": [
|
|
"The highest scoring prompts used step-by-step reasoning, but failed the JSON format metric. Add a strict JSON schema constraint.",
|
|
"Past attempts failed the toxicity metric when being too aggressive. Create a variation that is highly polite but retains the reasoning steps."
|
|
]
|
|
}}
|
|
**
|
|
|
|
JSON:
|
|
"""
|
|
|
|
@staticmethod
|
|
def generate_candidate(
|
|
original_prompt: str, guideline: str, is_list_format: bool = False
|
|
) -> str:
|
|
|
|
# Dynamically instruct the LLM on how to format the revised_prompt field
|
|
if is_list_format:
|
|
format_instruction = (
|
|
"A JSON array of message objects representing the revised conversational prompt "
|
|
'(e.g., [{"role": "system", "content": "..."}, {"role": "user", "content": "..."}]).'
|
|
)
|
|
example_instruction = '[\n {"role": "system", "content": "You are a helpful assistant."},\n {"role": "user", "content": "{{input}}"}\n ]'
|
|
else:
|
|
format_instruction = (
|
|
"The final string representing the optimized revised prompt."
|
|
)
|
|
example_instruction = (
|
|
'"You are a helpful assistant. Please answer: {{input}}"'
|
|
)
|
|
|
|
return f"""You are an expert prompt engineer. Your task is to rewrite a prompt based strictly on a specific optimization guideline.
|
|
|
|
[ORIGINAL PROMPT]
|
|
{original_prompt}
|
|
|
|
[OPTIMIZATION GUIDELINE]
|
|
{guideline}
|
|
|
|
[INSTRUCTIONS]
|
|
Rewrite the [ORIGINAL PROMPT] applying the [OPTIMIZATION GUIDELINE].
|
|
1. The new prompt must fulfill the core task of the original prompt.
|
|
2. DO NOT wrap your revised_prompt in markdown blocks (like ```).
|
|
3. If the original prompt uses variable placeholders (like {{input}}), you MUST retain them.
|
|
|
|
**
|
|
IMPORTANT: You must only return in JSON format matching the schema.
|
|
"revised_prompt" format: {format_instruction}
|
|
|
|
Example JSON:
|
|
{{
|
|
"thought_process": "The guideline asks to make the prompt more concise. I will remove the introductory pleasantries and state the objective directly.",
|
|
"revised_prompt": {example_instruction}
|
|
}}
|
|
**
|
|
|
|
JSON:
|
|
"""
|