Here is how to augment LLMs with tools

• a list of the possible and description of what they are and how to use them

• the template of the Reasoning-Act (ReAct) prompt technique

• the scratch book showing the results of the previous steps

• the output indicator to guide the LLM in formatting its output correctly

The ReAct technique forces the LLM to think about the next step to solve the question and choose a tool and a tool input to get more information based on that thought. We then extract the tool name and input with Regex and programmatically call the tool with the input and get the response. For example, one tool could be the Python package of the Wikipedia search engine.

We use the tool response to help further the LLM investigation to find the right answer. An agent is a wrapper around an LLM that is augmented with a bunch of tools. The agent iterates until the answer is found:

Author

AiUTOMATING PEOPLE, ABN ASIA was founded by people with deep roots in academia, with work experience in the US, Holland, Hungary, Japan, South Korea, Singapore, and Vietnam. ABN Asia is where academia and technology meet opportunity. With our cutting-edge solutions and competent software development services, we're helping businesses level up and take on the global scene. Our commitment: Faster. Better. More reliable. In most cases: Cheaper as well.

Feel free to reach out to us whenever you require IT services, digital consulting, off-the-shelf software solutions, or if you'd like to send us requests for proposals (RFPs). You can contact us at [email protected]. We're ready to assist you with all your technology needs.

© ABN ASIA