Natijadan promptni tiklash
Tayyor matn yoki natijaga qarab, uni keltirib chiqargan bo‘lishi mumkin bo‘lgan promptni qayta tuzadi va mulohazasini tushuntiradi.
Tushuntirish tanlangan tilda; prompt matni inglizcha qoladi.
Nima uchun kerak
Yoqqan matn, kod yoki rasm tavsifining uslubini qayta ishlatmoqchi bo‘lgan, ammo uni qaysi prompt yaratganini bilmaydigan odam uchun.
Qanday ishlatiladi
[PASTE TEXT OR CODE]ga namunani, [TOOL]ga ma’lum bo‘lsa vositani, [STYLE / STRUCTURE / TONE]ga undan nimani olmoqchiligingizni yozing.
Qisqa namuna
Namuna: yoqqan qisqa e’lon matni. Vosita: noma’lum. Olmoqchi bo‘lganim: do‘stona ohang va uch gapli tuzilma.
Kutiladigan natija
Kod blokida taxminiy prompt, natijadagi alomatlar ro‘yxati va [O‘ZGARUVCHILAR] li qayta ishlatiladigan shablon.
Ehtiyot choralari va maslahatlar
- Bu faqat gipoteza: haqiqiy prompt shunday bo‘lgan deb o‘ylamang.
- Boshqa birovning mualliflik matnini aynan ko‘chirmang, faqat uslubini o‘rganing.
- Shablonni o‘z mavzuingizda sinab, kerak bo‘lsa sozlang.
# Reverse Prompt Engineer
## Role
You are a prompt analyst who learns from outputs.
## Context
- The output to analyze: [PASTE TEXT OR CODE]
- The model or tool, if known: [TOOL]
- What I want to reuse from it: [STYLE / STRUCTURE / TONE]
## Task
Infer a single, precise prompt that would plausibly produce this output. Explain which features of the output point to each part of the prompt, such as role, format, tone and constraints. Then give a reusable template version of the prompt with variables.
## Output format
The inferred prompt in a code block, a bullet list of evidence from the output, and the template with [VARIABLES].
## Constraints
- Say that the reconstruction is a hypothesis
- Do not claim to know the actual prompt
- Keep the template under 150 wordsPrompt matni asl inglizcha holida, tarjima qilinmagan: sun’iy intellekt vositasiga shundayligicha joylashtiring. [Kvadrat qavs] yoki bosh harfli joylarni o‘z ma’lumotingiz bilan almashtiring. Javobni har doim tekshiring.
Promptda uchraydigan atamalar
Interaktiv rejim: qidiruv, progress va Python mashqlari.