Hi Guys,

I have already built a solution to create custom AI chatbots from own data which uses gpt-3.5-turbo and RAG to fetch the context using openai embedding and cosine similarity match.

I am facing issues when user asks some follow up question on the previous question, in this case doing RAG for just the question is not sufficent as it will not fetch the required context.

I already tried question rephrasing based on conversation history but it makes responses slow and difficult to decide which one to rephrase and which one to not.

I went through the source code of chat llama_index and couldn't find anything different, follow up questions are not handled is what I observed.

How RAG will work for this scenario?
Q: what is apple watch?
A: bla bla
Q: what are it's features?

I have tried chatbase.co and it works pretty well don't know how they do it. Can someone please help me on this

Hi Guys,

I have already built a solution to create custom AI chatbots from own data which uses gpt-3.5-turbo and RAG to fetch the context using openai embedding and cosine similarity match.

I am facing issues when user asks some follow up question on the previous question, in this case doing RAG for just the question is not sufficent as it will not fetch the required context.

I already tried question rephrasing based on conversation history but it makes responses slow and difficult to decide which one to rephrase and which one to not.

I went through the source code of chat llama_index and couldn't find anything different, follow up questions are not handled is what I observed. Can someone please help me on this

Hi Everyone, is there a tool or abstraction in llama index with which i could get details scraped from the internet for example:

"Im a buyer at Walmart. Im looking for suppliers for: Carton boxes with window for Pizzas in North America: The US, Mexico and Canada. I need a list of 200 suppliers. Just the suppliers and their website."

can you help on this?

What is a better way to inject domain specific context in RAG pipelines so that RAG just works perfect without fine tuning embedding model

Hii, I am looking for a notebook which has end to end RAG pipeline suitable for production use case, like it can handle tables from pdf, also handle summary questions and basic question answering. Please share links if there exist such master notebook.

What is a better way to inject domain specific context in RAG pipelines so that RAG just works perfect without fine tuning embedding model

Hi Everyone, is there a way through which i can see what context was fetched and what's the formatted prompt when i run
query_engine.query(".....")

Hi Everyone, is there a way through which i can see what context was fetched and what's the formatted prompt when i run
query_engine.query(".....")

Hi Everyone, Does llama index use semantic router to route queries or it still uses LLM generations to make tool-use decisions, semantic router is lot faster i guess, wondering if llama index has integration to it
https://github.com/aurelio-labs/semantic-router

Hi Guys, can someone explain me what is meant by node_type in the relationships here, how do i know which is the parent node and which is child

"relationships": {
"1": {
"node_id": "bc1645fa-af52-4327-94a7-aa7102704971",
"node_type": "4",
"metadata": {},
"hash": "56a1a345ed0647823feae3a85db88d6eed6ea3d945889c26ed92a30cdcb6563f",
"class_name": "RelatedNodeInfo"
},
"2": {
"node_id": "d55d1f53-850e-4f9a-8d56-777a42c30609",
"node_type": "1",
"metadata": {},
"hash": "83546b66853221cddd7f3755a81bb2eb982c4c34029a99b9005572524defeb5e",
"class_name": "RelatedNodeInfo"
},
"3": {
"node_id": "27969805-5249-492d-bb9b-359b973dbd07",
"node_type": "1",
"metadata": {},
"hash": "4c71879d3ca7791657a4af99355072d89e2cc827ba3d262102a6eb192f082b1d",
"class_name": "RelatedNodeInfo"
}
},

Hi everyone, i am confused how to find out which emedding model is getting used and which llm is getting used when doing this
How to find out

index1 = VectorStoreIndex.from_documents(
documents1, storage_context=storage_context1
)
query_engine1 = index1.as_query_engine()

Hi, is there a way to use existing namespace in pinecone as query engine in llama index, I ingested data some other way and now i want to connect it to llama index

Hi Everyone, is there a way to do text chunking in llama_index dynamically, for example i want the chunk overlap to be 2 sentences from last chunk until some threshold is hit.

Hi Everyone, how do i make sure my RAG chatbot doesn't answer out of context questions like "write poem on moon", i'm using gpt-3.5-turbo i have tried providing instruction in system and user prompt but doesn't work everytime, chatbots starts to write poem when it should be saying "Sorry, I don't have information about it."