Best prompt formulas to get highly specific code code explanations for students

Students often get frustrated with AI code explanations because the default response is a generic description of the concept ("recursion is when a function calls itself...") rather than an explanation of their specific code. The fix isn't a magic phrase — it's giving the assistant the actual code and asking it to walk through that code specifically, the same discipline programmers already use in "rubber duck debugging": explaining a program line by line, out loud, to a listener with zero context, which is what actually surfaces the parts you don't understand yet.

A prompt formula that produces specific explanations

1. Paste the real code, not a description of it. "Explain how a for-loop works" gets a textbook answer. Pasting your actual homework function and asking "explain this" gets an explanation tied to your specific variable names and logic.

2. Ask explicitly for line-by-line or block-by-block. Without this, the assistant will often default to a high-level summary of what the whole function does, skipping the individual steps that are usually where the confusion actually is.

3. Ask it to explain the "why," not just the "what." "What does this line do" gets you a restatement of the syntax. "Why is this line here" gets you the actual reasoning — what problem it solves in the context of the rest of the function.

# Prompt that produces a specific, line-by-line explanation

Here is my code: [paste your actual function/homework code]

Walk through it line by line. For each line or logical block:
1. State what it does in plain language.
2. State why it's needed - what would break or behave
   differently if this line were removed.
Do not summarize the whole function first - start directly
with line 1.

This is essentially the same principle as rubber duck debugging, just run in reverse: instead of you explaining the code to a silent listener to find your own gaps, you're asking the assistant to do the line-by-line explaining for you. The core technique — walking through code step by step rather than describing it in the abstract — is the same one described in The Pragmatic Programmer, the book that popularized rubber duck debugging as a practice: explaining code in that much detail is what makes gaps in understanding "leap off the screen." Used this way with an AI assistant, it turns a vague "I don't get it" into a specific list of lines you can ask follow-up questions about.