Planning and Editing What You Retrieve for Enhanced Tool Learning

Tenghao Huang, Dongwon Jung, Muhao Chen

Abstract:

Recent advancements in integrating external
tools with Large Language Models (LLMs)
have opened new frontiers, with applications
in mathematical reasoning, code generators,
and smart assistants. However, existing meth-
ods, relying on simple one-time retrieval strate-
gies, fall short on effectively and accurately
shortlisting relevant tools. This paper intro-
duces a novel PLUTO ( P lanning, L earning,
and U nderstanding for TOols) approach, en-
compassing “Plan-and-Retrieve (P&R)” and
“Edit-and-Ground (E&G)” paradigms. The
P&R paradigm consists of a neural retrieval
module for shortlisting relevant tools and an
LLM-based query planner that decomposes
complex queries into actionable tasks, enhanc-
ing the effectiveness of tool utilization. The
E&G paradigm utilizes LLMs to enrich tool de-
scriptions based on user scenarios, bridging the
gap between user queries and tool functionali-
ties. Experiment results demonstrate that these
paradigms significantly improve the recall and
NDCG in tool retrieval tasks, significantly sur-
passing current state-of-the-art models.