LLM이 곧 Human-Level Prompt Engineer: Automatic Prompt Engineer (APE)

# APE 핵심 플로우 — Python 슈도코드
import openai

def generate_prompt_candidates(demos, n=50):
    """
    demos: [(input, output), ...] 형태의 입출력 예시
    """
    demo_str = "\n".join([f"Input: {i}\nOutput: {o}" for i, o in demos])
    meta_prompt = f"""I gave a friend an instruction and five inputs.
 The friend read the instruction and wrote an output for every one of the inputs.
 Here are the input-output pairs:
{demo_str}
The instruction was"""
    
    candidates = []
    for _ in range(n):
        resp = openai.Completion.create(
            model="text-davinci-002",
            prompt=meta_prompt,
            max_tokens=50,
            temperature=0.9
        )
        candidates.append(resp.choices[0].text.strip())
    return candidates

def score_candidate(instruction, val_demos, model="text-davinci-002"):
    """실행 정확도로 프롬프트 품질 평가"""
    correct = 0
    for q, a in val_demos:
        prompt = f"Instruction: {instruction}\nInput: {q}\nOutput:"
        resp = openai.Completion.create(model=model, prompt=prompt, max_tokens=20, temperature=0)
        pred = resp.choices[0].text.strip()
        if pred == a:
            correct += 1
    return correct / len(val_demos)

# 메인 APE 루프
train_demos = [("cat", "c"), ("dog", "d"), ("apple", "a")]  # 예시
val_demos = [("banana", "b"), ("orange", "o")]

candidates = generate_prompt_candidates(train_demos, n=50)
scores = [(c, score_candidate(c, val_demos)) for c in candidates]
best_prompt = max(scores, key=lambda x: x[1])
print(f"Best prompt: {best_prompt[0]} (score: {best_prompt[1]:.2f})")

# Zero-shot CoT 최적화용 템플릿
cot_meta_prompt = """
Instruction: Answer the following question.
Q: {question}
A: Let's <INSERT>. {reasoning}
"""