How to remove the context and prompts in the answer generated by LLM and langchain?

[ztl58553116]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding)
retriever = vectordb.as_retriever() 
memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)

qa = ConversationalRetrievalChain.from_llm(
        return_model,
        retriever,
        memory=memory,
    )
result = qa({"question": "现在你是一名专业友好的中文心理咨询师,你的任务是用中文与用户对话,让他们开心快乐。我们鼓励你使用合适的表情符号,让用户感到精力充沛和满足。请注意,你的答案应该在100到200个单词之间。"})
question = ""
while question != "退出":
    question = input("请输入您的问题：")
    if question == "退出":
        break
    result = qa({"question": question})
    print(result['answer'])

Description

I'm a beginner in langchain, and I'm trying to use langchain to build a chinese chatbot that could comfort consultants. When I created a instance of "ConversationalRetrievalChain" and tried to use it to generate answer, I found that it always contained context and instructions information. I tried many ways but failed to remove them. Did anyone know how to solve the problem? I sincerely appreciate your replies.

System Info

langchain==0.2.0
langchain-community==0.2.0
langchain-core==0.2.0
langchain-text-splitters==0.2.0
langchainplus-sdk==0.0.20
requirements.txt

[feijoes]

Some of the context is derived from the condense_question_prompt: BasePromptTemplate = CONDENSE_QUESTION_PROMPT. If you do not provide a condense_question_prompt, the default will be CONDENSE_QUESTION_PROMPT.

[ztl58553116]

@feijoes Thank you so much! I will change the condense_question_prompt immediately.

[feijoes]

Hey @ztl58553116, did that solve the issue? If yes, please add it as an answer to this issue.