Model Parameter Values for LLM Integration in SuiteScript 2.1

TL;DR: This article discusses how to utilize model parameter values within the N/llm module in SuiteScript 2.1. This enables developers to customize responses from large language models (LLMs) such as Cohere Command A and OpenAI GPT-OSS for applications in Oracle Cloud Infrastructure (OCI).

What is the N/llm Module?

The N/llm module is a critical component of SuiteScript 2.1 that allows developers to interact with large language models to generate text-based responses. With support for models hosted in Oracle Cloud Infrastructure, this module offers advanced capabilities for integrating generative AI into NetSuite applications.

Customizing Model Responses with Parameters

When calling certain methods within the N/llm module, you can provide model parameter values to tailor how the LLM generates its output. Here are some key parameters you can adjust:

Key Model Parameters

You can leverage these parameters to enhance the model's responses. For example, adjusting the temperature can increase or decrease the creativity of the response — higher values (up to 1 for Cohere and 2 for OpenAI) lead to more diverse outputs.

Supported Methods

The following methods within the N/llm module utilize these model parameters:

• llm.generateText(options)
• llm.generateText.promise(options)
• llm.generateTextStreamed(options)
• llm.generateTextStreamed.promise(options)

When using these methods, ensure you provide the appropriate options for your use case to optimize the AI output. For example, in the llm.generateText(options) method, the prompt parameter must be specified unless options.toolResults is provided.

Conclusion

The N/llm module offers powerful capabilities for integrating generative AI into NetSuite applications. By customizing model parameter values, developers can ensure that the LLM produces responses suited to specific business needs.

Who This Affects

• Developers: Those implementing AI features in NetSuite.
• Administrators: Those configuring SuiteScript for generative AI applications.

Key Takeaways

• Customize LLM responses with model parameters like temperature and maxTokens.
• Methods like generateText support these configurations for tailored outputs.
• Understanding default values and accepted ranges is crucial for effective integration.