@inproceedings{yu-etal-2026-prfc,
title = "{PRFC}: Learning Progressive Recovery from Function Calling Errors for Tool-Using {LLM}s",
author = "Yu, Beibei and
Shen, Tao and
Chen, Ling",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.1762/",
doi = "10.18653/2026.findings-emnlp.1762",
pages = "37685--37706",
ISBN = "979-8-89176-462-0",
abstract = "Tool-using large language models (LLMs) can make multiple attempts to complete a task, using execution feedback from failed tool calls to identify errors and revise subsequent actions. However, their recovery performance often plateaus in later attempts, as models may repeat similar tool-call errors instead of making effective corrections. To address this limitation, we introduce PRFC, a reinforcement learning method that models error recovery as a sequential decision-making process. PRFC uses a progress-aware reward that assigns credit only when a later attempt improves upon previous failures, encouraging steady improvement across the recovery trajectory. Across four benchmarks with a Qwen3-4B backbone, PRFC improves multi-turn recovery on BFCL-v3 by an absolute 6.4{\%} and achieves the highest mean score across three single-step tasks (API-Bank, T-Eval, and Nestful). These results show that modeling progress across attempts provides an effective learning signal for multi-attempt tool-use recovery. These results suggest that explicitly modeling progress across attempts provides a useful learning signal for tool-use recovery."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yu-etal-2026-prfc">
<titleInfo>
<title>PRFC: Learning Progressive Recovery from Function Calling Errors for Tool-Using LLMs</title>
</titleInfo>
<name type="personal">
<namePart type="given">Beibei</namePart>
<namePart type="family">Yu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tao</namePart>
<namePart type="family">Shen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ling</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Tool-using large language models (LLMs) can make multiple attempts to complete a task, using execution feedback from failed tool calls to identify errors and revise subsequent actions. However, their recovery performance often plateaus in later attempts, as models may repeat similar tool-call errors instead of making effective corrections. To address this limitation, we introduce PRFC, a reinforcement learning method that models error recovery as a sequential decision-making process. PRFC uses a progress-aware reward that assigns credit only when a later attempt improves upon previous failures, encouraging steady improvement across the recovery trajectory. Across four benchmarks with a Qwen3-4B backbone, PRFC improves multi-turn recovery on BFCL-v3 by an absolute 6.4% and achieves the highest mean score across three single-step tasks (API-Bank, T-Eval, and Nestful). These results show that modeling progress across attempts provides an effective learning signal for multi-attempt tool-use recovery. These results suggest that explicitly modeling progress across attempts provides a useful learning signal for tool-use recovery.</abstract>
<identifier type="citekey">yu-etal-2026-prfc</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.1762</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.1762/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>37685</start>
<end>37706</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PRFC: Learning Progressive Recovery from Function Calling Errors for Tool-Using LLMs
%A Yu, Beibei
%A Shen, Tao
%A Chen, Ling
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F yu-etal-2026-prfc
%X Tool-using large language models (LLMs) can make multiple attempts to complete a task, using execution feedback from failed tool calls to identify errors and revise subsequent actions. However, their recovery performance often plateaus in later attempts, as models may repeat similar tool-call errors instead of making effective corrections. To address this limitation, we introduce PRFC, a reinforcement learning method that models error recovery as a sequential decision-making process. PRFC uses a progress-aware reward that assigns credit only when a later attempt improves upon previous failures, encouraging steady improvement across the recovery trajectory. Across four benchmarks with a Qwen3-4B backbone, PRFC improves multi-turn recovery on BFCL-v3 by an absolute 6.4% and achieves the highest mean score across three single-step tasks (API-Bank, T-Eval, and Nestful). These results show that modeling progress across attempts provides an effective learning signal for multi-attempt tool-use recovery. These results suggest that explicitly modeling progress across attempts provides a useful learning signal for tool-use recovery.
%R 10.18653/2026.findings-emnlp.1762
%U https://aclanthology.org/2026.findings-emnlp.1762/
%U https://doi.org/10.18653/2026.findings-emnlp.1762
%P 37685-37706
Markdown (Informal)
[PRFC: Learning Progressive Recovery from Function Calling Errors for Tool-Using LLMs](https://aclanthology.org/2026.findings-emnlp.1762/) (Yu et al., Findings 2026)
ACL