Understanding the Uncertainty of LLM Explanations: A Perspective Based on Reasoning Topology Article Swipe

PDF

Longchao Da , Xiaoou Liu , Jia-Xiao Dai , Lu Cheng , Yaqing Wang , Wei Hua ·

YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2502.17026

Understanding the uncertainty in large language model (LLM) explanations is important for evaluating their faithfulness and reasoning consistency, and thus provides insights into the reliability of LLM's output regarding a question. In this work, we propose a novel framework that quantifies uncertainty in LLM explanations through a reasoning topology perspective. By designing a structural elicitation strategy, we guide the LLMs to frame the explanations of an answer into a graph topology. This process decomposes the explanations into the knowledge related sub-questions and topology-based reasoning structures, which allows us to quantify uncertainty not only at the semantic level but also from the reasoning path. It further brings convenience to assess knowledge redundancy and provide interpretable insights into the reasoning process. Our method offers a systematic way to interpret the LLM reasoning, analyze limitations, and provide guidance for enhancing robustness and faithfulness. This work pioneers the use of graph-structured uncertainty measurement in LLM explanations and demonstrates the potential of topology-based quantification.

Related Topics

Concepts

No concepts available.

Metadata

Type: preprint
Language: en
Landing Page: http://arxiv.org/abs/2502.17026
PDF: https://arxiv.org/pdf/2502.17026
OA Status: green
Cited By: 1
OpenAlex ID: https://openalex.org/W4414847465

All OpenAlex metadata

Raw OpenAlex JSON

OpenAlex ID: https://openalex.org/W4414847465

Canonical identifier for this work in OpenAlex
DOI: https://doi.org/10.48550/arxiv.2502.17026

Digital Object Identifier
Title: Understanding the Uncertainty of LLM Explanations: A Perspective Based on Reasoning Topology

Work title
Type: preprint

OpenAlex work type
Language: en

Primary language
Publication year: 2025

Year of publication
Publication date: 2025-02-24

Full publication date if available
Authors: Longchao Da, Xiaoou Liu, Jia-Xiao Dai, Lu Cheng, Yaqing Wang, Wei Hua

List of authors in order
Landing page: https://arxiv.org/abs/2502.17026

Publisher landing page
PDF URL: https://arxiv.org/pdf/2502.17026

Direct link to full text PDF
Open access: Yes

Whether a free full text is available
OA status: green

Open access status per OpenAlex
OA URL: https://arxiv.org/pdf/2502.17026

Direct OA link when available
Cited by: 1

Total citation count in OpenAlex
Citations by year (recent): 2025: 1

Per-year citation counts (last 5 years)

Full payload

id	https://openalex.org/W4414847465
doi	https://doi.org/10.48550/arxiv.2502.17026
ids.doi	https://doi.org/10.48550/arxiv.2502.17026
ids.openalex	https://openalex.org/W4414847465
fwci
type	preprint
title	Understanding the Uncertainty of LLM Explanations: A Perspective Based on Reasoning Topology
biblio.issue
biblio.volume
biblio.last_page
biblio.first_page
topics[0].id	https://openalex.org/T13643
topics[0].field.id	https://openalex.org/fields/33
topics[0].field.display_name	Social Sciences
topics[0].score	0.9574999809265137
topics[0].domain.id	https://openalex.org/domains/2
topics[0].domain.display_name	Social Sciences
topics[0].subfield.id	https://openalex.org/subfields/3320
topics[0].subfield.display_name	Political Science and International Relations
topics[0].display_name	Artificial Intelligence in Law
topics[1].id	https://openalex.org/T13851
topics[1].field.id	https://openalex.org/fields/17
topics[1].field.display_name	Computer Science
topics[1].score	0.930400013923645
topics[1].domain.id	https://openalex.org/domains/3
topics[1].domain.display_name	Physical Sciences
topics[1].subfield.id	https://openalex.org/subfields/1702
topics[1].subfield.display_name	Artificial Intelligence
topics[1].display_name	Law, AI, and Intellectual Property
is_xpac	False
apc_list
apc_paid
language	en
locations[0].id	pmh:oai:arXiv.org:2502.17026
locations[0].is_oa	True
locations[0].source.id	https://openalex.org/S4306400194
locations[0].source.issn
locations[0].source.type	repository
locations[0].source.is_oa	True
locations[0].source.issn_l
locations[0].source.is_core	False
locations[0].source.is_in_doaj	False
locations[0].source.display_name	arXiv (Cornell University)
locations[0].source.host_organization	https://openalex.org/I205783295
locations[0].source.host_organization_name	Cornell University
locations[0].source.host_organization_lineage	https://openalex.org/I205783295
locations[0].license
locations[0].pdf_url	https://arxiv.org/pdf/2502.17026
locations[0].version	submittedVersion
locations[0].raw_type	text
locations[0].license_id
locations[0].is_accepted	False
locations[0].is_published	False
locations[0].raw_source_name
locations[0].landing_page_url	http://arxiv.org/abs/2502.17026
locations[1].id	doi:10.48550/arxiv.2502.17026
locations[1].is_oa	True
locations[1].source.id	https://openalex.org/S4306400194
locations[1].source.issn
locations[1].source.type	repository
locations[1].source.is_oa	True
locations[1].source.issn_l
locations[1].source.is_core	False
locations[1].source.is_in_doaj	False
locations[1].source.display_name	arXiv (Cornell University)
locations[1].source.host_organization	https://openalex.org/I205783295
locations[1].source.host_organization_name	Cornell University
locations[1].source.host_organization_lineage	https://openalex.org/I205783295
locations[1].license
locations[1].pdf_url
locations[1].version
locations[1].raw_type	article
locations[1].license_id
locations[1].is_accepted	False
locations[1].is_published
locations[1].raw_source_name
locations[1].landing_page_url	https://doi.org/10.48550/arxiv.2502.17026
indexed_in	arxiv, datacite
authorships[0].author.id	https://openalex.org/A5074815612
authorships[0].author.orcid	https://orcid.org/0009-0000-8631-9634
authorships[0].author.display_name	Longchao Da
authorships[0].author_position	first
authorships[0].raw_author_name	Da, Longchao
authorships[0].is_corresponding	False
authorships[1].author.id	https://openalex.org/A5100683738
authorships[1].author.orcid	https://orcid.org/0000-0002-0958-8758
authorships[1].author.display_name	Xiaoou Liu
authorships[1].author_position	middle
authorships[1].raw_author_name	Liu, Xiaoou
authorships[1].is_corresponding	False
authorships[2].author.id	https://openalex.org/A5081909977
authorships[2].author.orcid	https://orcid.org/0000-0002-0203-1279
authorships[2].author.display_name	Jia-Xiao Dai
authorships[2].author_position	middle
authorships[2].raw_author_name	Dai, Jiaxin
authorships[2].is_corresponding	False
authorships[3].author.id	https://openalex.org/A5091262319
authorships[3].author.orcid	https://orcid.org/0000-0003-3712-4165
authorships[3].author.display_name	Lu Cheng
authorships[3].author_position	middle
authorships[3].raw_author_name	Cheng, Lu
authorships[3].is_corresponding	False
authorships[4].author.id	https://openalex.org/A5051741610
authorships[4].author.orcid	https://orcid.org/0000-0003-1457-1114
authorships[4].author.display_name	Yaqing Wang
authorships[4].author_position	middle
authorships[4].raw_author_name	Wang, Yaqing
authorships[4].is_corresponding	False
authorships[5].author.id	https://openalex.org/A5100403937
authorships[5].author.orcid	https://orcid.org/0000-0002-2047-9712
authorships[5].author.display_name	Wei Hua
authorships[5].author_position	last
authorships[5].raw_author_name	Wei, Hua
authorships[5].is_corresponding	False
has_content.pdf	False
has_content.grobid_xml	False
is_paratext	False
open_access.is_oa	True
open_access.oa_url	https://arxiv.org/pdf/2502.17026
open_access.oa_status	green
open_access.any_repository_has_fulltext	False
created_date	2025-10-10T00:00:00
display_name	Understanding the Uncertainty of LLM Explanations: A Perspective Based on Reasoning Topology
has_fulltext	False
is_retracted	False
updated_date	2025-11-06T06:51:31.235846
primary_topic.id	https://openalex.org/T13643
primary_topic.field.id	https://openalex.org/fields/33
primary_topic.field.display_name	Social Sciences
primary_topic.score	0.9574999809265137
primary_topic.domain.id	https://openalex.org/domains/2
primary_topic.domain.display_name	Social Sciences
primary_topic.subfield.id	https://openalex.org/subfields/3320
primary_topic.subfield.display_name	Political Science and International Relations
primary_topic.display_name	Artificial Intelligence in Law
cited_by_count	1
counts_by_year[0].year	2025
counts_by_year[0].cited_by_count	1
locations_count	2
best_oa_location.id	pmh:oai:arXiv.org:2502.17026
best_oa_location.is_oa	True
best_oa_location.source.id	https://openalex.org/S4306400194
best_oa_location.source.issn
best_oa_location.source.type	repository
best_oa_location.source.is_oa	True
best_oa_location.source.issn_l
best_oa_location.source.is_core	False
best_oa_location.source.is_in_doaj	False
best_oa_location.source.display_name	arXiv (Cornell University)
best_oa_location.source.host_organization	https://openalex.org/I205783295
best_oa_location.source.host_organization_name	Cornell University
best_oa_location.source.host_organization_lineage	https://openalex.org/I205783295
best_oa_location.license
best_oa_location.pdf_url	https://arxiv.org/pdf/2502.17026
best_oa_location.version	submittedVersion
best_oa_location.raw_type	text
best_oa_location.license_id
best_oa_location.is_accepted	False
best_oa_location.is_published	False
best_oa_location.raw_source_name
best_oa_location.landing_page_url	http://arxiv.org/abs/2502.17026
primary_location.id	pmh:oai:arXiv.org:2502.17026
primary_location.is_oa	True
primary_location.source.id	https://openalex.org/S4306400194
primary_location.source.issn
primary_location.source.type	repository
primary_location.source.is_oa	True
primary_location.source.issn_l
primary_location.source.is_core	False
primary_location.source.is_in_doaj	False
primary_location.source.display_name	arXiv (Cornell University)
primary_location.source.host_organization	https://openalex.org/I205783295
primary_location.source.host_organization_name	Cornell University
primary_location.source.host_organization_lineage	https://openalex.org/I205783295
primary_location.license
primary_location.pdf_url	https://arxiv.org/pdf/2502.17026
primary_location.version	submittedVersion
primary_location.raw_type	text
primary_location.license_id
primary_location.is_accepted	False
primary_location.is_published	False
primary_location.raw_source_name
primary_location.landing_page_url	http://arxiv.org/abs/2502.17026
publication_date	2025-02-24
publication_year	2025
referenced_works_count	0
abstract_inverted_index.a	29, 36, 46, 52, 68, 122
abstract_inverted_index.By	50
abstract_inverted_index.In	31
abstract_inverted_index.It	103
abstract_inverted_index.an	65
abstract_inverted_index.at	93
abstract_inverted_index.in	3, 42, 149
abstract_inverted_index.is	9
abstract_inverted_index.of	25, 64, 145, 156
abstract_inverted_index.to	60, 88, 107, 125
abstract_inverted_index.us	87
abstract_inverted_index.we	34, 56
abstract_inverted_index.LLM	43, 128, 150
abstract_inverted_index.Our	119
abstract_inverted_index.and	15, 18, 81, 111, 132, 138, 152
abstract_inverted_index.but	97
abstract_inverted_index.for	11, 135
abstract_inverted_index.not	91
abstract_inverted_index.the	1, 23, 58, 62, 74, 77, 94, 100, 116, 127, 143, 154
abstract_inverted_index.use	144
abstract_inverted_index.way	124
abstract_inverted_index.LLMs	59
abstract_inverted_index.This	71, 140
abstract_inverted_index.also	98
abstract_inverted_index.from	99
abstract_inverted_index.into	22, 67, 76, 115
abstract_inverted_index.only	92
abstract_inverted_index.that	39
abstract_inverted_index.this	32
abstract_inverted_index.thus	19
abstract_inverted_index.work	141
abstract_inverted_index.(LLM)	7
abstract_inverted_index.LLM's	26
abstract_inverted_index.frame	61
abstract_inverted_index.graph	69
abstract_inverted_index.guide	57
abstract_inverted_index.large	4
abstract_inverted_index.level	96
abstract_inverted_index.model	6
abstract_inverted_index.novel	37
abstract_inverted_index.path.	102
abstract_inverted_index.their	13
abstract_inverted_index.which	85
abstract_inverted_index.work,	33
abstract_inverted_index.allows	86
abstract_inverted_index.answer	66
abstract_inverted_index.assess	108
abstract_inverted_index.brings	105
abstract_inverted_index.method	120
abstract_inverted_index.offers	121
abstract_inverted_index.output	27
abstract_inverted_index.analyze	130
abstract_inverted_index.further	104
abstract_inverted_index.process	72
abstract_inverted_index.propose	35
abstract_inverted_index.provide	112, 133
abstract_inverted_index.related	79
abstract_inverted_index.through	45
abstract_inverted_index.guidance	134
abstract_inverted_index.insights	21, 114
abstract_inverted_index.language	5
abstract_inverted_index.pioneers	142
abstract_inverted_index.process.	118
abstract_inverted_index.provides	20
abstract_inverted_index.quantify	89
abstract_inverted_index.semantic	95
abstract_inverted_index.topology	48
abstract_inverted_index.designing	51
abstract_inverted_index.enhancing	136
abstract_inverted_index.framework	38
abstract_inverted_index.important	10
abstract_inverted_index.interpret	126
abstract_inverted_index.knowledge	78, 109
abstract_inverted_index.potential	155
abstract_inverted_index.question.	30
abstract_inverted_index.reasoning	16, 47, 83, 101, 117
abstract_inverted_index.regarding	28
abstract_inverted_index.strategy,	55
abstract_inverted_index.topology.	70
abstract_inverted_index.decomposes	73
abstract_inverted_index.evaluating	12
abstract_inverted_index.quantifies	40
abstract_inverted_index.reasoning,	129
abstract_inverted_index.redundancy	110
abstract_inverted_index.robustness	137
abstract_inverted_index.structural	53
abstract_inverted_index.systematic	123
abstract_inverted_index.convenience	106
abstract_inverted_index.elicitation	54
abstract_inverted_index.measurement	148
abstract_inverted_index.reliability	24
abstract_inverted_index.structures,	84
abstract_inverted_index.uncertainty	2, 41, 90, 147
abstract_inverted_index.consistency,	17
abstract_inverted_index.demonstrates	153
abstract_inverted_index.explanations	8, 44, 63, 75, 151
abstract_inverted_index.faithfulness	14
abstract_inverted_index.limitations,	131
abstract_inverted_index.perspective.	49
abstract_inverted_index.Understanding	0
abstract_inverted_index.faithfulness.	139
abstract_inverted_index.interpretable	113
abstract_inverted_index.sub-questions	80
abstract_inverted_index.topology-based	82, 157
abstract_inverted_index.quantification.	158
abstract_inverted_index.graph-structured	146
cited_by_percentile_year
countries_distinct_count	0
institutions_distinct_count	6
citation_normalized_percentile