logo

Sentiment-Based Succinct Response Generator

InfoGenerateCreated ByPackages

This code snippet is an SQL query that interacts with a function called 'SNOWFLAKE.CORTEX.COMPLETE'. It uses the 'llama2-70b-chat' model to generate a response based on a conversation format provided in the 'messages' parameter. In the example, the system role sets the context for the AI's behavior ('You are a helpful AI assistant. Respond succinctly to each input based on the sentiment of the text.') and the user role provides an input ('blah blah'). The generated response from the model is aliased as 'response'.

-- Sure! Let's break this down step by step. Here is the modified code to satisfy the description provided:

SELECT SNOWFLAKE.CORTEX.COMPLETE(
    'llama2-70b-chat',
    [
        {'role': 'system', 'content': 'You are a helpful AI assistant. Respond succinctly to each input based on the sentiment of the text.' },
        {'role': 'user', 'content': 'blah blah'}
    ], {}
    ) as response;