The tagline of my musing on one of the hot topics around the world an Andaaz-e-Istifsar to Prompt Engineering describes some salient features of the topic of this post. Mirza Ghalib, one the most celebrated personality of the subcontinent says:
देखना तक़रीर कि लज्ज़त कि जो उसने कहा,
मैं ने यह समझा कि गोया ये भी मेरे दिल में है.
To express oneself in the apt and suitable style to convey the thought process is part of the soft skill skills sought in almost every walk of human life. Therefore, I am going to explain what prompt engineering is in simple and straight forward way not making it a puzzle, though due to my poetic temperament, I could not control to refer one couplet from the corpus of Ghalib which optimally explains the very core idea of Prompt Engineering.
When I asked CHATGPT, developed by Open AI which was founded by Sam Altman and Elon Musk for non- profit and when it diverted its way to the for profit, Elon Musk filed a suit against Open AI, which this company won: “Define Prompt Engineering in academic way.”
The response was: “Prompt engineering is the systematic design, structuring, and optimization of inputs (prompts) given to artificial intelligence models—especially large language models (LLMs)—to guide them toward producing accurate, relevant, reliable, and goal-oriented outputs.”
And when I queried from Google “Prompt Engineering”, the reply was:
“Core Framework: The C-O-S-T-A-R Method
To build effective prompts, structure them logically using the COSTAR framework:
• Context (C): Provide background information or the scenario (e.g., “You are an AI consultant…”).
• Objective (O): Clearly state what the AI needs to do (e.g., “Summarize this data into three bullet points.”).
• Style (S): Specify the tone or formatting desired (e.g., “Write in a professional, empathetic tone.”).
• Tactic (T): Outline specific steps the AI should take to break down the problem.
• Audience (A): Identify who the final response is for (e.g., “Explain this for a non-technical audience.”).
• Response Format (R): Define how the output should be delivered (e.g., “Return the results in a JSON object.”).
Advanced Techniques
• Few-Shot Prompting: Providing a few examples of desired input-output pairs to teach the AI the pattern before asking it to do the actual task.
• Chain-of-Thought: Asking the AI to “think step-by-step” before giving its final answer to reduce logic errors.
• Role Prompting: Assigning a specific persona or profession to the AI so it adopts the appropriate knowledge base and style.”
The tagline to my essay – An Andaaz-e-Istifsar means the style of asking something. It is a phrase consisting one word Andaaz from Persian and the second one from Arabic languet meaning prompting.
Prompt Engineering is not a tiny matter but a vast fielding and it has many dimensions which would continue to be explained on this portal for the benefit of human kind. (Hamid Siddiqui)
DATA ANNOTATION AND DATA LABELING
BAKRA EID AND BENGAL
Dasna – A Muslim Majority Town
DAILY ROUTINE OF A MADRASA STUDENT
Leave a Reply