ChatGPT Exercises

Fill in the blanks to test your knowledge.

1

The four-part prompt structure for effective ChatGPT prompts is: Role, Context, ___, and Constraints

const promptStructure = {
role: "You are a senior TypeScript engineer...",
context: "I'm building a multi-tenant SaaS app...",
: "Review this code for type safety issues...",
constraints: "Format as [SEVERITY] Line: Description",
};
2

ChatGPT's persistent background context fields — your role, stack, and experience — are configured in the ___ Instructions settings

const persistentContext = {
setting: " Instructions",
field1: "What ChatGPT should know about you",
field2: "How ChatGPT should respond",
scope: "Applies to every conversation automatically",
bestFor: "Stable, long-term preferences",
};
3

The ChatGPT feature that stores notes about you dynamically across sessions — updating as you share information — is called ___

const persistenceFeatures = {
static: "Custom Instructions (you write once)",
dynamic: " (ChatGPT updates over time)",
manage: "Settings → Personalization → Manage ",
maintenance: "Delete stale entries periodically",
};
4

Custom ChatGPT instances with a system prompt, knowledge files, and tool access are called ___

const customInstances = {
name: "s",
layers: [
"System prompt (role, behavior, output format)",
"Knowledge files (PDFs, docs, code)",
"Tool access (web, Code Interpreter, API actions)",
],
access: "GPT store or shared link",
useCase: "Recurring specialized workflows",
};
5

OpenAI's reasoning-optimized models that think through problems step-by-step before responding are called o1 and ___

const modelComparison = {
default: "GPT-4o — fast, multimodal, general purpose",
reasoning: [
"o1 — complex reasoning, most professional use",
" — hardest problems, maximum accuracy",
"o3-mini — reasoning at faster speed",
],
whenToSwitch: "GPT-4o gives confident but wrong answer on hard logic",
};
6

The ChatGPT feature that provides a sandboxed Python execution environment — enabling file analysis, precise calculations, and chart generation — is called Code ___

const codeFeature = {
name: "Code ",
capabilities: [
"Execute Python code",
"Read uploaded CSV, Excel, JSON files",
"Generate matplotlib charts",
"Compute precise calculations",
"Create downloadable files",
],
criticalNote: "Calculations here are exact — unlike plain conversation",
};
7

The code generation approach that shows ChatGPT an existing function and asks for a new one following the same pattern is called ___ application

const generationApproach = {
name: " application",
method: "Show existing function + ask for new one in same ",
benefit: "Output matches codebase style automatically",
vs: "Blank specification (less integrated output)",
example: "Here is getUser — write getProject following exact same ",
};
8

When asking ChatGPT to review its own generated code, the most valuable checklist items are TypeScript violations, unhandled ___ cases, and potential runtime errors

const selfReviewChecklist = {
item1: "TypeScript strict mode violations",
item2: "Unhandled cases (null, empty array, invalid type)",
item3: "Potential runtime errors",
whenToUse: "After every significant code generation task",
prompt: "Review the code you just wrote. Check for these issues...",
};