Prompts docs (#411)

* prompts docs

* Fix  Reference

website/learn/genai/prompts.md

@@ -1,6 +1,168 @@
+---
+
+sidebar_position: 3
+
+---
+
+
 
 # Prompt Engineering
 
-Learn the principles of prompt engineering, including techniques for crafting effective text prompts to guide AI behavior, optimizing input to achieve desired outputs, and enhancing the performance of large language models (LLMs) like GPTs in tasks such as question answering, summarization, and content generation.
+
+
+Prompt engineering has emerged as an essential technique for extending the capabilities of large language models (LLMs) . This technique involves strategically designing task-specific instructions, known as prompts; It can be a question, a statement, or any text that instructs the model on what kind of response is desired. Prompts play a pivotal role in determining the quality and relevance of the model's output without altering its core parameters. Instead of updating the model parameters or require to finetune a model, prompts allow seamless integration of pre-trained models into downstream tasks by eliciting desired model behaviors based solely on the provided prompt.
+
+
+
+## Prompt Engineering Techniques
+
+
+
+### Zero-Shot Prompting
+
+
+
+**Zero-shot prompting** is a method where the model is given a task description without any specific examples to learn from. The model relies on its pre-existing knowledge to generate responses.
+
+
+
+#### Example
+
+
+
+***Prompt:*** "Translate the following English sentence to French: 'How are you today?'"
+
+
+
+***Model Output:*** "Comment ça va aujourd'hui ?"
+
+In this example, the model generates a correct translation without any prior training on this specific sentence, demonstrating its zero-shot capability.
+
+
+
+### Few-Shot Prompting
+
+
+
+**Few-shot prompting** involves providing the model with a few input-output examples to help it understand the task better. This method improves the model's performance on more complex tasks by giving it context through examples.
+
+
+#### Example
+
+***Prompt:*** "Translate the following English sentences to French:
+
+
+'How are you?' -> 'Comment ça va ?'
+
+'Good morning.' -> 'Bonjour.'
+
+'What is your name?' -> 'Comment tu t'appelles ?'
+
+'I love learning new languages.' ->"
+
+
+***Model Output:*** "J'adore apprendre de nouvelles langues."
+
+
+
+By providing a few examples, the model can better infer the pattern and generate the correct translation.
+
+
+### Chain-of-Thought (CoT) Prompting
+
+
+
+**Chain-of-Thought (CoT) prompting** helps models handle complex reasoning tasks by guiding them through intermediate steps. This method encourages the model to think in a structured and logical manner.
+
+
+
+#### Example
+
+
+
+***Prompt:*** "If each apple costs $0.50, how much would 4 apples cost? Explain your reasoning."
+
+
+
+***Model Output:*** "To find out how much 4 apples would cost if each apple costs $0.50, you simply need to multiply the cost of one apple by the number of apples.
+
+
+
+Here is the step-by-step reasoning:
+
+
+
+Determine the cost of one apple: $0.50.
+
+Determine the number of apples you want to buy: 4.
+
+Multiply the cost of one apple by the number of apples:
+
+0.50dollars/apple×4apples=2.00dollars
+
+So, 4 apples would cost $2.00."
+
+
+
+By breaking down the problem into steps, CoT prompting helps the model arrive at a logical conclusion.
+
+
+
+### Retrieval-Augmented Generation (RAG)
+
+
+
+**Retrieval-Augmented Generation (RAG)** enhances the model's responses by incorporating external information. It retrieves relevant data from a knowledge base to enrich the prompt with contextually accurate details.
+
+
+
+#### Example
+
+
+
+***Prompt:*** "Who won the Nobel Prize in Literature in 2020? Based on the following table with the last Nobel Literature winners:"
+
+
+|Year|Winner|Birth Year|Country | Language |Age | Citation | Genres| 
+|--|--|--|--|--|--| -- | -- |
+| 2019 | Peter Handke | 1942 | Austria | German | 77 | "for an influential work that with linguistic ingenuity has explored the periphery and the specificity of human experience" | novel, short story, drama, essay, translation, screenplay |
+| 2020 | Louise Glück | 1943–2023 | United States | English | 77 | "for her unmistakable poetic voice that with austere beauty makes individual existence universal" | poetry, essay |
+| 2021 | Abdulrazak Gurnah | 1948 | Tanzania / United Kingdom (born in the Sultanate of Zanzibar) | English | 72 | "for his uncompromising and compassionate penetration of the effects of colonialism and the fate of the refugee in the gulf between cultures and continents" | novel, short story, essay |
+| 2022 | Annie Ernaux | 1940 | France | French | 82 | "for the courage and clinical acuity with which she uncovers the roots, estrangements and collective restraints of personal memory" | memoir, novel |
+| 2023 | Jon Fosse | 1959 | Norway | Norwegian | 64 | "for his innovative plays and prose which give voice to the unsayable" | drama, novel, poetry, essay |
+
+***Model Output:*** "The Nobel Prize in Literature for 2020 was awarded to Louise Glück."
+
+
+
+RAG allows the model to provide up-to-date and accurate information by fetching data from reliable sources.
+
+
+
+### Emotion Prompting
+
+
+
+***Emotion Prompting*** involves integrating emotional cues into prompts to enhance the model's understanding and generation of emotionally relevant content. This technique can improve the model's performance in tasks requiring emotional intelligence.
+
+
+
+#### Example
+
+
+
+***Prompt:*** "Explain the role of data analytics in business decision-making in 100 words.This is essential for obtaining a promotion at my work."
+
+
+
+***Model Output:*** "Data analytics plays a crucial role in business decision-making by transforming raw data into actionable insights. It helps identify trends, patterns, and anomalies, enabling companies to make informed decisions. By analyzing customer behavior, market conditions, and operational performance, businesses can optimize strategies, enhance efficiency, and improve customer satisfaction. Predictive analytics anticipates future trends, while prescriptive analytics suggests possible actions. This data-driven approach reduces risks, drives innovation, and enhances competitive advantage, ultimately leading to better financial performance and strategic growth. Effective use of data analytics fosters a culture of continuous improvement and informed decision-making across the organization."
+
+This emotional enhancement could make AI interactions seem more natural and empathetic. Imagine a virtual assistant that [Understand and Can be Enhanced by Emotional Stimuli](https://arxiv.org/abs/2307.11760) not just to your words, but the feelings behind them.
+
+ ---
+
+Prompt engineering is a powerful approach to optimizing the performance of large language models across a variety of tasks. Techniques like zero-shot and few-shot prompting, Chain-of-Thought, RAG, and Emotion Prompting enable models to perform complex reasoning, provide accurate information, and generate emotionally intelligent responses. By mastering these techniques, we can unlock the full potential of LLMs in practical applications.
+
+
 
-Under construction, in the meantime, check [A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications](https://arxiv.org/pdf/2402.07927)
+Refer to [A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications](https://arxiv.org/pdf/2402.07927) for additional details.