Geckopt: LLM System Efficiency Via Intent-based Tool Selection

Fore Michael, Singh Simranjit, Stamoulis Dimitrios. Arxiv 2024

[Paper]    
Efficiency And Optimization GPT Model Architecture Prompting Reinforcement Learning Tools

In this preliminary study, we investigate a GPT-driven intent-based reasoning approach to streamline tool selection for large language models (LLMs) aimed at system efficiency. By identifying the intent behind user prompts at runtime, we narrow down the API toolset required for task execution, reducing token consumption by up to 24.6%. Early results on a real-world, massively parallel Copilot platform with over 100 GPT-4-Turbo nodes show cost reductions and potential towards improving LLM-based system efficiency.

Similar Work