Why RAG gives agents knowledge in Agentic AI - Test Your Understanding

Practice - 5 Tasks

1fill in blank

easy

Complete the code to load documents for retrieval in RAG.

Agentic AI

from langchain.document_loaders import TextLoader
loader = TextLoader('[1]')
docs = loader.load()

Drag options to blanks, or click blank then click option'

Aagent.py

Bdata.txt

Cconfig.json

Dmodel.py

Attempts:

3 left

💡 Hint

Common Mistakes

Choosing a code or config file instead of a text file.

The documents to retrieve knowledge from must be loaded from a text file, so 'data.txt' is correct.

2fill in blank

medium

Complete the code to create a retriever from documents for RAG.

Agentic AI

from langchain.vectorstores import FAISS
retriever = FAISS.from_documents(docs, [1]).as_retriever()

Drag options to blanks, or click blank then click option'

Aloss_function

Btokenizer

Coptimizer

Dembedding_model

Attempts:

3 left

💡 Hint

Common Mistakes

Using tokenizer or optimizer which are unrelated to vector search.

The retriever needs an embedding model to convert documents into vectors for similarity search.

3fill in blank

hard

Fix the error in the RAG agent creation by completing the missing argument.

Agentic AI

from langchain.chains import RetrievalQA
rag_agent = RetrievalQA.from_chain_type(llm=llm, retriever=[1], return_source_documents=True)

Drag options to blanks, or click blank then click option'

Aretriever

Bembedding_model

Ctokenizer

Doptimizer

Attempts:

3 left

💡 Hint

Common Mistakes

Passing tokenizer or embedding model instead of retriever.

The RAG agent requires a retriever to fetch relevant documents for answering queries.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that filters documents by length and stores their text.

Agentic AI

filtered_docs = {doc.metadata['title']: doc.page_content for doc in docs if len(doc.page_content) [1] 100 and doc.metadata['source'] [2] 'trusted_source'}

Drag options to blanks, or click blank then click option'

A<

B>

C==

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using '>' for length or '!=' for source causes wrong filtering.

We filter documents shorter than 100 characters and from the trusted source.

5fill in blank

hard

Fill all three blanks to build a RAG pipeline that embeds, retrieves, and answers questions.

Agentic AI

embedding = [1]()
retriever = FAISS.from_documents(docs, embedding).as_retriever()
rag_agent = RetrievalQA.from_chain_type(llm=[2](), retriever=retriever, return_source_documents=True)
answer = rag_agent.run([3])

Drag options to blanks, or click blank then click option'

AOpenAIEmbeddings

BOpenAI

C'What is RAG?'

DGPT2

Attempts:

3 left

💡 Hint

Common Mistakes

Using GPT2 as LLM or passing a non-string question.

The embedding model is OpenAIEmbeddings, the language model is OpenAI, and the question is a string.