FREYR: A Framework for Recognizing and Executing Your Requests
This addresses the adaptability and resource limitations of current tool usage approaches for conversational agents, though it appears incremental as a modularization of existing concepts.
The paper tackles the problem of extending large language models' capabilities through tool usage by introducing FREYR, a framework that modularizes the process into separate steps, achieving superior performance compared to conventional methods as evaluated on real-world video game design test cases against the Ollama API.
Large language models excel as conversational agents, but their capabilities can be further extended through tool usage, i.e.: executable code, to enhance response accuracy or address specialized domains. Current approaches to enable tool usage often rely on model-specific prompting or fine-tuning a model for function-calling instructions. Both approaches have notable limitations, including reduced adaptability to unseen tools and high resource requirements. This paper introduces FREYR, a streamlined framework that modularizes the tool usage process into separate steps. Through this decomposition, we show that FREYR achieves superior performance compared to conventional tool usage methods. We evaluate FREYR on a set of real-world test cases specific for video game design and compare it against traditional tool usage as provided by the Ollama API.