{"id":"https://openalex.org/W7214630769","doi":"https://doi.org/10.48550/arxiv.2609.30849","title":"Enhancing Assessment of Self-Consistency in LLM Explanations using Perturbation Strength","display_name":"Enhancing Assessment of Self-Consistency in LLM Explanations using Perturbation Strength","publication_year":2026,"publication_date":"2026-09-25","ids":{"openalex":"https://openalex.org/W7214630769","doi":"https://doi.org/10.48550/arxiv.2609.30849"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.30849","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.30849","pdf_url":"https://arxiv.org/pdf/2609.30849","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.30849","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5045692299","display_name":"Phuong Le","orcid":"https://orcid.org/0000-0003-4724-7118","observed_orcids":["https://orcid.org/0000-0003-4724-7118"]},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Le, Phuong Q.","raw_affiliation_strings":["School of Computing and Information Systems The University of Melbourne , Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computing and Information Systems The University of Melbourne , Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067017155","display_name":"Kemal Kurniawan","orcid":"https://orcid.org/0000-0002-4231-4310","observed_orcids":["https://orcid.org/0000-0002-4231-4310"]},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Kurniawan, Kemal","raw_affiliation_strings":["School of Computer Science and Engineering University of New South Wales , Sydney , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computer Science and Engineering University of New South Wales , Sydney , Australia","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5153212918","display_name":"Jey Han Lau","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I165779595","display_name":"The University of Melbourne","ror":"https://ror.org/01ej9dk98","country_code":"AU","type":"education","lineage":["https://openalex.org/I165779595"]}],"countries":["AU"],"is_corresponding":false,"raw_author_name":"Lau, Jey Han","raw_affiliation_strings":["School of Computing and Information Systems The University of Melbourne , Melbourne , Australia"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Computing and Information Systems The University of Melbourne , Melbourne , Australia","institution_ids":["https://openalex.org/I165779595"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9984999895095825,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9984999895095825,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.0006000000284984708,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.00019999999494757503,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/self-consistency","display_name":"self-consistency","score":0.9599999785423279},{"id":"https://openalex.org/keywords/llm-generated-explanations","display_name":"LLM-generated explanations","score":0.7850000262260437},{"id":"https://openalex.org/keywords/perturbation-strength","display_name":"perturbation strength","score":0.921999990940094},{"id":"https://openalex.org/keywords/llm-as-a-judge","display_name":"LLM-as-a-judge","score":0.906000018119812},{"id":"https://openalex.org/keywords/input-disturbances","display_name":"input perturbation","score":0.984000027179718},{"id":"https://openalex.org/keywords/evaluation-of-explanations","display_name":"explanation evaluation","score":0.5270000100135803},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.5120000243186951}],"concepts":[{"id":"https://openalex.org/C177918212","wikidata":"https://www.wikidata.org/wiki/Q803623","display_name":"Perturbation (astronomy)","level":2,"score":0.8147000074386597},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.4327000081539154},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3801000118255615},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.335999995470047},{"id":"https://openalex.org/C2986219828","wikidata":"https://www.wikidata.org/wiki/Q1193095","display_name":"Coupling strength","level":2,"score":0.3224000036716461},{"id":"https://openalex.org/C165160513","wikidata":"https://www.wikidata.org/wiki/Q7524249","display_name":"Singular perturbation","level":2,"score":0.3093999922275543},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.29600000381469727},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.28949999809265137},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.2727999985218048},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.2587999999523163},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.2572999894618988},{"id":"https://openalex.org/C153359094","wikidata":"https://www.wikidata.org/wiki/Q7207873","display_name":"Poincar\u00e9\u2013Lindstedt method","level":3,"score":0.2508000135421753}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.30849","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.30849","pdf_url":"https://arxiv.org/pdf/2609.30849","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.30849","doi":"https://doi.org/10.48550/arxiv.2609.30849","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.30849","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.30849","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.30849","pdf_url":"https://arxiv.org/pdf/2609.30849","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214630769.pdf","grobid_xml":"https://content.openalex.org/works/W7214630769.grobid-xml"},"referenced_works_count":7,"referenced_works":["https://openalex.org/W2794325560","https://openalex.org/W4385327568","https://openalex.org/W4387561528","https://openalex.org/W4395474395","https://openalex.org/W4401306886","https://openalex.org/W4407806804","https://openalex.org/W4415976144"],"related_works":["https://openalex.org/W4402671681","https://openalex.org/W4415100943","https://openalex.org/W4406844883","https://openalex.org/W7119538405","https://openalex.org/W7166869515","https://openalex.org/W4416047182","https://openalex.org/W4417129603","https://openalex.org/W7166649129","https://openalex.org/W4391272510","https://openalex.org/W4402683840"],"abstract_inverted_index":{"Prior":[0],"work":[1,96],"has":[2],"examined":[3],"the":[4,14,49,108],"self-consistency":[5,50,103],"of":[6,16],"LLM-generated":[7],"explanations":[8,52],"using":[9],"surface-level":[10],"perturbation":[11,35,66,74,110],"methods.":[12],"However,":[13],"strength":[15,36,59,75],"these":[17],"perturbations":[18,86],"is":[19,104],"not":[20],"explicitly":[21],"measured":[22],"and":[23,43,80,83],"controlled.":[24],"In":[25],"this":[26],"work,":[27],"we":[28],"propose":[29],"an":[30],"LLM-as-a-judge":[31],"approach":[32],"to":[33],"measure":[34,76],"in":[37,51],"a":[38,62,101],"unified":[39],"manner":[40],"across":[41,65],"input":[42,85],"CoT":[44,93],"perturbations.":[45,94],"We":[46],"then":[47],"evaluate":[48],"generated":[53],"from":[54],"various":[55],"LLMs":[56,89],"under":[57],"controlled":[58],"conditions,":[60],"ensuring":[61],"fair":[63,105],"comparison":[64],"types.":[67],"Experiments":[68],"show":[69],"that":[70,84,98],"our":[71],"proposed":[72],"LLM-based":[73],"outperforms":[77],"other":[78],"embedding-":[79],"probability-based":[81],"approaches":[82],"generally":[87],"affect":[88],"more":[90],"strongly":[91],"than":[92],"Our":[95],"suggests":[97],"judgments":[99],"about":[100],"model's":[102],"only":[106],"within":[107],"same":[109],"type.":[111]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-29T00:00:00"}