FREYR: A Framework for Recognizing and Executing Your Requests

2025-01-21

AI Summary

Existing large language models (LLMs) rely on custom prompts or fine-tuning for tool invocation, resulting in poor generalization, high computational overhead, and limited adaptability to novel tools. Method: We propose a lightweight, modular, stepwise tool-calling framework that decouples request parsing, tool selection, and execution. It enables zero-shot tool generalization via instruction decomposition, standardized tool interfaces, and dynamic scheduling—requiring neither model fine-tuning nor prompt engineering. The framework is compatible with mainstream open-source LLMs (e.g., Ollama). Contribution/Results: This work introduces the first fully decoupled, general-purpose tool-calling architecture. Evaluated in real-world video game design tasks, it achieves a 27% higher task completion rate than Ollama’s native tool-calling, with significantly improved response accuracy and robustness over baselines. Our approach establishes a new paradigm for lightweight, scalable tool augmentation of LLMs.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

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.

Problem

• Large Language Models
• Tool Utilization
• Adaptability

Innovation

• FREYR
• modularization
• tool utilization