geckopt
GeckOpt: LLM System Efficiency via Intent-Based Tool Selection
Fore, Michael, Singh, Simranjit, Stamoulis, Dimitrios
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.
2404.15804
Country:
- North America > United States > Florida > Pinellas County > Clearwater (0.06)
- North America > United States > Washington > King County > Redmond (0.05)
- North America > United States > Virginia > Fairfax County > Reston (0.05)
- (2 more...)
Technology: