How can I change/modify prompt in chain.call?

[Luisotee]

Hi guys, is there a way for me to add context or to completely change the initial prompt upon every chain.call?

This is my failed attempt at trying to give a new prompt in every chain.call:

Initializing the model:

const OPENROUTER_BASE_URL = "https://openrouter.ai";

const openRouterChat = new ChatOpenAI(
  {
    modelName: LLM_MODEL,
    //streaming: true,
    openAIApiKey: OPENROUTER_API_KEY,
  },
  {
    basePath: `${OPENROUTER_BASE_URL}/api/v1`,
  }
);

const memory = new ConversationSummaryMemory({
  memoryKey: "chat_history",
  inputKey: "input",
  llm: openRouterChat,
});

//${OPEN_ROUTER_SYSTEM_MESSAGE}
const systemMessageOpenRouter = PromptTemplate.fromTemplate(` 
## Current conversation:
    {chat_history}
    Human: {input}
    AI:

    `);

export const chain = new LLMChain({
  llm: openRouterChat,
  prompt: systemMessageOpenRouter,
  memory,
});

Calling the model:

export async function getCompletionWithOpenRouter(
  message: Message,
  context: string,
  streamingReply: Message
) {
  const contextPrompt = PromptTemplate.fromTemplate(` 
  ${OPEN_ROUTER_SYSTEM_MESSAGE}

  ## Current conversation:
      {chat_history}
      Human: {input}
      AI:
  
      `);
  let response = await chain.call(
    { input: message.body, prompt: contextPrompt }
   
  );

  return response.text;
}

[dosubot[bot]]

🤖

Hi @Luisotee,

Nice to meet you! I'm Dosu, a bot here to assist with your questions, help solve bugs, and guide you on becoming a contributor while we wait for a human maintainer. I'm here to help make your coding journey smoother. Let's dive into your issue!

Based on the code you've provided, it seems like you're trying to modify the prompt for each chain.call by passing a new prompt parameter. However, the chain.call method in the LangChainJS framework doesn't handle the prompt parameter in the way you're expecting.

In the LangChainJS framework, the prompt is defined when you initialize the LLMChain object and it's not designed to be changed with each chain.call. The chain.call method only accepts an input parameter, which is inserted into the {input} placeholder in the prompt template.

If you want to change the context or the initial prompt for each chain.call, you might want to consider using the MultiPromptChain.fromLLMAndPrompts method. This method allows you to define different prompt templates for different contexts and use them in your chain.call.

Here's an example of how you can use it:

const physicsTemplate = `You are a very smart physics professor. You are great at answering questions about physics in a concise and easy to understand manner. When you don't know the answer to a question you admit that you don't know.

Here is a question:
{input}
`;

// ...

const multiPromptChain = MultiPromptChain.fromLLMAndPrompts(llm, {
  promptNames,
  promptDescriptions,
  promptTemplates,
});

const testPromise1 = multiPromptChain.call({
  input: "What is the speed of light?",
});

In this example, the input "What is the speed of light?" is inserted into the physicsTemplate, changing the initial prompt to a physics context.

You can find more information about this in the examples/src/chains/multi_prompt.ts file in the LangChainJS repository.

I hope this helps! If you have any other questions, feel free to ask.

-Dosu

Sources

• examples/src/chains/multi_prompt.ts

---

This response is meant to be useful and save you time. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

If you want to continue the conversation, start your reply with @dosu-bot.