@inproceedings{chen-etal-2026-rfcllm,
title = "{RFCLLM}: Evaluating {LLM}s' Reasoning Ability of Network Protocol State Machines",
author = "Chen, Anqi and
Goldwasser, Dan and
Nita-Rotaru, Cristina",
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.2161/",
doi = "10.18653/2026.findings-emnlp.2161",
pages = "46227--46246",
ISBN = "979-8-89176-462-0",
abstract = "Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations. LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in practice. The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly. We examine the degree to which an LLM{'}s implicit representation of a finite-state transition system{--}defined via natural language descriptions{--}aligns with a manually generated ground-truth model.\par We designed 4 tasks and 1482 task queries for 16 protocols. We evaluated different judge biases, observed the inherent difficulty gaps between tasks, looked into the effect of 4 context types, and the influence of protocol characteristics. Our work contributes to a step toward verifying whether LLMs can really be trusted in FSM (Finite State Machine) reasoning of protocol specifications."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-rfcllm">
<titleInfo>
<title>RFCLLM: Evaluating LLMs’ Reasoning Ability of Network Protocol State Machines</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anqi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dan</namePart>
<namePart type="family">Goldwasser</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Cristina</namePart>
<namePart type="family">Nita-Rotaru</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>Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations. LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in practice. The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly. We examine the degree to which an LLM’s implicit representation of a finite-state transition system–defined via natural language descriptions–aligns with a manually generated ground-truth model.\par We designed 4 tasks and 1482 task queries for 16 protocols. We evaluated different judge biases, observed the inherent difficulty gaps between tasks, looked into the effect of 4 context types, and the influence of protocol characteristics. Our work contributes to a step toward verifying whether LLMs can really be trusted in FSM (Finite State Machine) reasoning of protocol specifications.</abstract>
<identifier type="citekey">chen-etal-2026-rfcllm</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.2161</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2161/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>46227</start>
<end>46246</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T RFCLLM: Evaluating LLMs’ Reasoning Ability of Network Protocol State Machines
%A Chen, Anqi
%A Goldwasser, Dan
%A Nita-Rotaru, Cristina
%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 chen-etal-2026-rfcllm
%X Mapping textual specifications into formal representations is essential for ensuring the correctness of protocol designs and implementations. LLM-generated mappings, used for networking security or testing, are assumed to capture a perfect understanding of the specification, which may not hold in practice. The goal of this paper is to assess the extent to which LLMs can interpret the specification correctly. We examine the degree to which an LLM’s implicit representation of a finite-state transition system–defined via natural language descriptions–aligns with a manually generated ground-truth model.\par We designed 4 tasks and 1482 task queries for 16 protocols. We evaluated different judge biases, observed the inherent difficulty gaps between tasks, looked into the effect of 4 context types, and the influence of protocol characteristics. Our work contributes to a step toward verifying whether LLMs can really be trusted in FSM (Finite State Machine) reasoning of protocol specifications.
%R 10.18653/2026.findings-emnlp.2161
%U https://aclanthology.org/2026.findings-emnlp.2161/
%U https://doi.org/10.18653/2026.findings-emnlp.2161
%P 46227-46246
Markdown (Informal)
[RFCLLM: Evaluating LLMs’ Reasoning Ability of Network Protocol State Machines](https://aclanthology.org/2026.findings-emnlp.2161/) (Chen et al., Findings 2026)
ACL