Flashcards · Prompt Engineering · Free
Prompt Engineering flashcards, generated for you.
Example Prompt Engineering study cards to learn from right now — then generate a full set from your own notes (plus a practice quiz) and export to Quizlet or Anki. Free, no account needed.
Example Prompt Engineering flashcards
Why does telling an AI to 'think step-by-step' improve its answers?
It forces the model to externalize intermediate reasoning instead of jumping to conclusions. This mimics how humans solve complex problems and reduces errors by making each reasoning step verifiable.
What is the core intuition behind 'prompt engineering'?
The quality of a model's output depends heavily on how you frame your request. Better prompts = clearer communication of what you want, leading to better results.
Why does providing examples in a prompt (few-shot learning) work?
Examples show the model the pattern, format, and style you expect. It uses these as templates to understand context better than abstract instructions alone.
What does 'temperature' control in a model, intuitively?
Temperature controls randomness in responses. Low temperature (e.g., 0.2) makes outputs predictable and focused; high temperature (e.g., 1.0+) makes them creative and varied.
What is 'chain-of-thought prompting' and why does it work?
Asking the model to explain its reasoning before answering. It works because verbalizing intermediate steps reduces logical errors—the model catches its own mistakes while reasoning aloud.
How does 'prompt injection' happen, and what's the underlying vulnerability?
A user embeds malicious instructions within input data, treating user input as trusted code. The model can't distinguish between the original prompt and user-supplied text, executing unintended commands.
What is 'in-context learning' at a mathematical level?
The model uses attention mechanisms to weight tokens from examples and instructions. Contextual embeddings shift the model's prediction distribution toward patterns seen in the prompt, without updating weights.
Why does 'role-playing' (e.g., 'You are an expert historian') improve outputs?
It biases the model's token predictions toward high-probability responses associated with that role. The prompt conditions the model's attention and output distribution to match relevant knowledge patterns.
What is the relationship between prompt length and model performance?
Longer prompts with clear structure improve reasoning but increase computational cost (more tokens = slower, more expensive). Beyond a point, irrelevant detail adds noise and reduces precision.
What mathematical principle explains why 'explicit constraints' in prompts reduce errors?
Constraints reduce the output space and increase the prior probability of valid answers. Mathematically, they shift the model's conditional probability distribution P(output | prompt) toward desired outcomes by narrowing possibilities.
Make your own Prompt Engineering study set
Flashcards for related topics
Studying Prompt Engineering to build with AI? MindloomHQ turns it into real skills — structured courses, agent projects, and certificates.
Explore MindloomHQ →