How do I configure the OpenAi chat model node?#
A chat model is a type of Large Language Model (LLM) that returns text that is statistically likely to meet the requirements of the user prompt.
The OpenAi chat model uses tokenisation, embedding and transformation functions to determine responses to prompts.
Before you begin#
- Learn about Tokenization, Embedding and Transformation functions in LLMs
- Learn about OpenAi
- Learn about OpenAi models which are available depending on your OpenAi account settings
- Learn how to authenticate with OpenAi
- Learn about pathways
- Learn about pathway nodes
- Learn how to add an Agent nodes to the pathway
Step 1 - Link the OpenAi chat model node to the Agent node#
- Click the Models button on the Agent.
- Search for OpenAi chat model in the node search field.
- Click the model to automatically link to the agent.
Step 2 - set OpenAi parameters#
- Choose an OpenAi credential
- Choose an OpenAi model from those available in your account
Step 3 - Decide how parameters are to be added#
There are three ways to add parameters to Fixed or Expression fields in a given node:
Method | Description | Additional information |
---|---|---|
Load from previous nodes | Click Execute previous nodes in the Input panel to load parameters from connected and configured nodes | Not available for Triggers |
Add a fixed value | Click Fixed on available fields to enter plain-text or JSON values | Learn how to edit fixed responses |
Add a JSON expression | Click Expression on available fields to enter a JSON expression | Learn how to edit expressions |
Note
The Expression editor loads all possible parameters from connected nodes. These can then be added to fields as required
Step 3 - define optional LLM Token parameters#
Click Add option to define any of the following fields:
Field | Parameter | Description | Additional information |
---|---|---|---|
Frequency penalty | frequency_penalty |
Relates to repeated tokens that exist in the prompt sent by the Agent node which may result in repetition. | Positive decimal values decrease repetition |
Maximum number of tokens | max_tokens |
The maximum number of tokens to generate | OpenAi Max Tokens parameter |
Presence penalty | presence_penalty |
Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics | OpenAi presence pentalty parameter |
Step 4 - define optional LLM response parameters#
Click Add option to define any of the following fields:
Field | Parameter | Description | Additional information |
---|---|---|---|
Response format | response_format |
Choose plain text or valid JSON | |
Timeout | timeout |
Integer value representing milliseconds before a LLM response should timeout | |
Max retries | max_retries |
Integer value representing the number of times the LLM should retry processing the prompt to create a response |
Step 5 - define optional sampling methods#
Choose one of the following to define how repetitive or unique responses will be:
Field | Parameter | Description | Additional information |
---|---|---|---|
Sampling temperature | temperature |
A scale of values which increase repetitive responses from the LLM the closer the temperature is to zero | OpenAi Sampling temperature parameter |
Top P | top_p |
Scale of decimal values between 0 and 1 that determines the diversity of responses using nucleus sampling | OpenAi Top_P Nucleus Sampling parameter |