Techniques6 min read
What Is RAG? Retrieval-Augmented Generation Explained Simply
RAG retrieves relevant documents and adds them to the prompt so the model answers from sources. See how it works and how to prompt for grounded, cited answers.
Grounding
Grounding and RAG: keeping AI answers tied to sources
Grounding means making a model answer from sources you provide instead of from memory. Retrieval-augmented generation (RAG) does this automatically: it searches your documents, puts the relevant passages into the prompt and asks the model to answer from them. Good grounding prompts tell the model to cite the passage it used and to say when the sources do not contain the answer.
Coming soon onApp StoreComing soon onGoogle Play
Structured outputIntermediate
Your app parses the model’s JSON, but sometimes the reply starts with “Sure! Here is the JSON:”. What is the most robust fix?
AAdd “please” to the prompt
BUse structured-output mode with a schema, then validate
CLower the temperature to 0
DStrip the first line of every reply
Sample question · Grounding
A RAG assistant answers confidently even when the retrieved passages don’t cover the question. Which instruction helps most?
- A“Be accurate.”
- B“Answer only from the passages below. If they don’t contain the answer, say so.”
- C“Use your general knowledge as a backup.”
- D“Answer in one sentence.”
Show the answer and why
B. “Answer only from the passages below. If they don’t contain the answer, say so.” Giving the model an explicit, allowed way to say “not in the sources” reduces made-up answers. “Be accurate” gives it nothing to act on.
What the questions test
- Answering only from provided sources
- Citations and quoting the supporting passage
- Handling questions the sources don’t answer
- Placing long documents in the prompt
What you’ll learn
- What RAG is and why it reduces hallucinations
- How to instruct a model to stay inside its sources
- How to ask for citations that can be checked
- How chunking and passage order affect answers
How to practice it in TokIQ
- 1. Open Grounding. Pick it from the topic list on the Quiz tab. It is a Premium topic; the free plan includes Fundamentals, Role & context and Examples.
- 2. Answer one question at a time. You see right away whether you were correct.
- 3. Tap Why? Read the short explanation and answer three follow-up questions on the same idea.
- 4. Watch your accuracy. The Progress tab shows how Grounding compares with your other topics.
Grounding: common questions
What is retrieval-augmented generation (RAG)?
RAG is a pattern where a system retrieves relevant passages from a document store and adds them to the prompt, so the model answers from that material instead of only from its training data. The term comes from Lewis et al. (2020).
Does RAG stop hallucinations completely?
No. It reduces them, but the model can still misread or go beyond the passages. Asking for citations and allowing “I don’t know” makes errors easier to spot.
Further reading
Safety6 min read
Prompt Injection Explained: Direct vs Indirect Attacks and Fixes
Prompt injection tricks an AI into following attacker instructions hidden in input or documents. Learn direct vs indirect attacks and the mitigations that work.
Basics5 min read
11 Common Prompt Engineering Mistakes and How to Fix Them
The prompt engineering mistakes that cause most bad AI output, from missing context to conflicting rules and untested changes, each with a concrete fix.
More topics
Practice grounding a few questions a day.
Short quizzes on real prompting decisions, with an explanation for every answer. Free to start on iPhone and Android.
Coming soon onApp StoreComing soon onGoogle Play