How to properly use ConversationalRetrievalChain.from_llm

[uzumakinaruto19]

Hi, So I have a rag application/chatbot, uses conversationalretrivalqa chain from Langchain, say if for questions like 'Hi' and all retrieval is happening, and its returning random documents How do I make the llm answer directly without retrieval for questions like this.? And one more thing how do I implement a memory(longterm will be better) with conversationalretrivalqa.from_llm chain..whatever I tried is not working, I tried with the Runnablehistory but that screws up the retrieval Does anyone have any workaround on that.? Any help will be appreciated ,thanks

`
chain = ConversationalRetrievalChain.from_llm(

    llm=llm,
    
    chain_type="stuff",
    retriever=ContextualCompressionRetriever(
        base_compressor=compressor,base_retriever=retriever),
    memory=memory,
    combine_docs_chain_kwargs={"prompt": qa_prompt},
    output_key="result",
    verbose=True,
    )`