Recent advancements have enabled Large Language Models (LLMs) to function as agents that can perform actions using external tools. This requires registering, i.e., integrating tool information into the LLM context prior to taking actions. Current methods indiscriminately incorporate all candidate tools into the agent's context and retain them across multiple reasoning steps. This process remains opaque to LLM agents and is not integrated into their reasoning procedures, leading to inefficiencies due to increased context length from irrelevant tools. To address this, we introduce EcoAct, a tool using algorithm that allows LLMs to selectively register tools as needed, optimizing context use. By integrating the tool registration process into the reasoning procedure, EcoAct reduces computational costs by over 50% in multiple steps reasoning tasks while maintaining performance, as demonstrated through extensive experiments. Moreover, it can be plugged into any reasoning pipeline with only minor modifications to the prompt, making it applicable to LLM agents now and future.
@article{arxiv.2411.01643,
title = {EcoAct: Economic Agent Determines When to Register What Action},
author = {Shaokun Zhang and Jieyu Zhang and Dujian Ding and Mirian Hipolito Garcia and Ankur Mallick and Daniel Madrigal and Menglin Xia and Victor Rühle and Qingyun Wu and Chi Wang},
journal= {arXiv preprint arXiv:2411.01643},
year = {2024}
}