Vol. 1 No. 1 (2024)
Contributions

Integrating AI Services into Semantic Kernel: A Case Study on Enhancing Functionality with Google PaLM and Large Language Models

Published 2024-06-18 — Updated on 2024-06-21

Versions

How to Cite

Integrating AI Services into Semantic Kernel: A Case Study on Enhancing Functionality with Google PaLM and Large Language Models. (2024). Transactions on Open Source Software Projects, 1(1). https://ojs.codeday.org/tossp/article/view/2 (Original work published 2024)

Abstract

Microsoft's Semantic Kernel is a software development kit (SDK) designed for C# and Python developers to integrate AI large language models (LLM) such as Chat-GPT into their existing applications. The Semantic Kernel leverages AI services like OpenAI, Azure OpenAI, and Hugging Face via their respective APIs. This SDK simplifies the utilization of AI technology for developers by enabling the orchestration of AI plugins.

To become acquainted with the Semantic Kernel project, the team selected a 'good first issue' to address collaboratively. This issue involved modifying the API calls to include a parameter known as logit-bias, which allows developers to influence the probability of certain words appearing in responses from OpenAI models. Logit-bias is a dictionary that maps words to integer values, with -100 completely banning a word and 100 exclusively favoring a word, with varying degrees in between.