Module 1: Introduction to LLM and Prompting
--- Simplified Notes ---
Introduction to Large Language Models
- Text Generation Models: AI that writes human-like text.
- Large Language Models (LLMs): Big models like ChatGPT trained on a lot of data.
- Why LLMs are Special: They understand and generate language well.
- History: From simple models to GPT, BERT, etc.
- LLMs in the Market: ChatGPT, Gemini, LLaMA, etc.
Understanding Prompting and Prompt Techniques
- Five Principles:
1. Be clear
2. Be specific
3. Use examples
4. Set a goal
5. Use step-by-step instructions
- What is a Prompt: Instructions given to LLMs.
- How Prompts Work: LLMs read and respond based on intent.
- Types of Prompts:
- Instruction-based
- Question-based
- Role-based
- Prompt Components:
- Task
- Input
- Output
- Defining Personality in Prompts: Ask model to act like a teacher, poet, etc.
- Mix and Match Strategy: Combine methods for better responses.
- Challenges:
- Misunderstood prompts
- Wrong or biased answers
- Not always updated
--- 5-Mark Answers ---
Q: What are Large Language Models?
> LLMs are AI trained on large text data to generate human-like text. Examples: ChatGPT, Bard.
Q: Components of a Prompt?
> Task, Input, Output.
Q: Challenges of Prompts?
> Misunderstanding, wrong data, outdated info, bias.
--- Case Study ---
Case: Customer Support Bot
- Old Prompt: Tell me about refunds.
- New Prompt: Act like a customer service agent. Explain the refund policy in simple words. Be
polite.
- Result: 35% better customer satisfaction.
--- MCQs ---
1. Main function of LLM?
A. Store data
B. Generate human-like text
C. Translate images
D. Count numbers
Answer: B
2. What is a prompt?
A. Error
B. Voice command
C. Instruction
D. Software
Answer: C
3. NOT a prompt component?
A. Task
B. Input
C. Speaker
D. Output
Answer: C
4. Improves prompt performance?
A. Ignoring output
B. Being vague
C. Mix and Match
D. General questions
Answer: C
5. Limitation of LLMs?
A. Never wrong
B. Biased
C. Always updated
D. No prompts needed
Answer: B