Chain-of-Tools: Utilizing Massive Unseen Tools in the CoT Reasoning of Frozen Language Models

Mengsong Wu Tong Zhu Han Han Xiang Zhang · 2025

Read Chain-of-Tools: Utilizing Massive Unseen Tools in the CoT Reasoning of Frozen Language Models by Mengsong Wu Tong Zhu Han Han Xiang Zhang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-16779.txt.

Description

Mengsong Wu and five co-authors taught frozen language models to call vast sets of tools. Their paper, from twenty twenty-five, deals with tool learning. Past ways tune the model so it knows just the tools seen in training. Or they stuff tool demos in the prompt, which wastes space. Their way, Chain-of-Tools, leans on the strong sense vectors of a frozen model. The model calls tools step by step inside its chain of thought.

The tool pool can be huge and can hold tools the model has never seen. The team built a new set of plain tool questions to test that case. Trials on math and fact benchmarks back the claim. Readers get tool use with no tuning bill.

Catalog record

CREATORMengsong Wu Tong Zhu Han Han Xiang Zhang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE33 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-16779

Browse the full library · About the Tower of Babel