@inproceedings{trevino-etal-2025-benchmarking,
title = "Benchmarking Failures in Tool-Augmented Language Models",
author = "Trevi{\~n}o, Eduardo and
Contant, Hugo and
Ngai, James and
Neubig, Graham and
Wang, Zora Zhiruo",
editor = "Chiruzzo, Luis and
Ritter, Alan and
Wang, Lu",
booktitle = "Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)",
month = apr,
year = "2025",
address = "Albuquerque, New Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.naacl-long.149/",
doi = "10.18653/v1/2025.naacl-long.149",
pages = "2916--2934",
ISBN = "979-8-89176-189-6",
abstract = "The integration of tools has extended the capabilities of language models (LMs) beyond vanilla text generation to versatile scenarios. However, tool-augmented language models (TaLMs) often assume `perfect' information access and tool availability, which may not hold in the real world. To systematically study TaLMs imperfections, we introduce the FAIL-TaLMs benchmark, featuring two major failures: under-specified user queries and non-available tools. FAIL-TaLMS contains 1,749 examples using 906 tools across 21 categories, including single- and multi-tool usage. We evaluate top-performing proprietary and open-source models, and find all current models except for Claude struggle to recognize missing tools or information. Further, to study possible mitigation of the failures, we enable real-time human interaction, named the Ask-and-Help method, to provide missing information or replace non-functional tools. While Ask-and-Help can help models solve tasks more correctly when queries are under-specified, it brings minimal benefit when complex tools are broken."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="trevino-etal-2025-benchmarking">
<titleInfo>
<title>Benchmarking Failures in Tool-Augmented Language Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Eduardo</namePart>
<namePart type="family">Treviño</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hugo</namePart>
<namePart type="family">Contant</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">James</namePart>
<namePart type="family">Ngai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Graham</namePart>
<namePart type="family">Neubig</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zora</namePart>
<namePart type="given">Zhiruo</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Luis</namePart>
<namePart type="family">Chiruzzo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Alan</namePart>
<namePart type="family">Ritter</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Lu</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Albuquerque, New Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-189-6</identifier>
</relatedItem>
<abstract>The integration of tools has extended the capabilities of language models (LMs) beyond vanilla text generation to versatile scenarios. However, tool-augmented language models (TaLMs) often assume ‘perfect’ information access and tool availability, which may not hold in the real world. To systematically study TaLMs imperfections, we introduce the FAIL-TaLMs benchmark, featuring two major failures: under-specified user queries and non-available tools. FAIL-TaLMS contains 1,749 examples using 906 tools across 21 categories, including single- and multi-tool usage. We evaluate top-performing proprietary and open-source models, and find all current models except for Claude struggle to recognize missing tools or information. Further, to study possible mitigation of the failures, we enable real-time human interaction, named the Ask-and-Help method, to provide missing information or replace non-functional tools. While Ask-and-Help can help models solve tasks more correctly when queries are under-specified, it brings minimal benefit when complex tools are broken.</abstract>
<identifier type="citekey">trevino-etal-2025-benchmarking</identifier>
<identifier type="doi">10.18653/v1/2025.naacl-long.149</identifier>
<location>
<url>https://aclanthology.org/2025.naacl-long.149/</url>
</location>
<part>
<date>2025-04</date>
<extent unit="page">
<start>2916</start>
<end>2934</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Benchmarking Failures in Tool-Augmented Language Models
%A Treviño, Eduardo
%A Contant, Hugo
%A Ngai, James
%A Neubig, Graham
%A Wang, Zora Zhiruo
%Y Chiruzzo, Luis
%Y Ritter, Alan
%Y Wang, Lu
%S Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)
%D 2025
%8 April
%I Association for Computational Linguistics
%C Albuquerque, New Mexico
%@ 979-8-89176-189-6
%F trevino-etal-2025-benchmarking
%X The integration of tools has extended the capabilities of language models (LMs) beyond vanilla text generation to versatile scenarios. However, tool-augmented language models (TaLMs) often assume ‘perfect’ information access and tool availability, which may not hold in the real world. To systematically study TaLMs imperfections, we introduce the FAIL-TaLMs benchmark, featuring two major failures: under-specified user queries and non-available tools. FAIL-TaLMS contains 1,749 examples using 906 tools across 21 categories, including single- and multi-tool usage. We evaluate top-performing proprietary and open-source models, and find all current models except for Claude struggle to recognize missing tools or information. Further, to study possible mitigation of the failures, we enable real-time human interaction, named the Ask-and-Help method, to provide missing information or replace non-functional tools. While Ask-and-Help can help models solve tasks more correctly when queries are under-specified, it brings minimal benefit when complex tools are broken.
%R 10.18653/v1/2025.naacl-long.149
%U https://aclanthology.org/2025.naacl-long.149/
%U https://doi.org/10.18653/v1/2025.naacl-long.149
%P 2916-2934
Markdown (Informal)
[Benchmarking Failures in Tool-Augmented Language Models](https://aclanthology.org/2025.naacl-long.149/) (Treviño et al., NAACL 2025)
ACL
- Eduardo Treviño, Hugo Contant, James Ngai, Graham Neubig, and Zora Zhiruo Wang. 2025. Benchmarking Failures in Tool-Augmented Language Models. In Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 2916–2934, Albuquerque, New Mexico. Association for Computational Linguistics.