{"id":"https://openalex.org/W7171708250","doi":"https://doi.org/10.48550/arxiv.2607.25600","title":"Beyond Self-Knowledge: Propagating Uncertainty Across Reasoning and Retrieval in LLMs","display_name":"Beyond Self-Knowledge: Propagating Uncertainty Across Reasoning and Retrieval in LLMs","publication_year":2026,"publication_date":"2026-07-28","ids":{"openalex":"https://openalex.org/W7171708250","doi":"https://doi.org/10.48550/arxiv.2607.25600"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.25600","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.25600","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2607.25600","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5143930576","display_name":"Chandan Kumar Sah","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sah, Chandan Kumar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5144004891","display_name":"Li Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5111296074","display_name":"Xiaoli Lian","orcid":"https://orcid.org/0000-0002-6100-7068"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lian, Xiaoli","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T10028","display_name":"Topic Modeling","score":0.857200026512146,"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/T10028","display_name":"Topic Modeling","score":0.857200026512146,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.053199999034404755,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T13274","display_name":"Expert finding and Q&A systems","score":0.03290000185370445,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/security-token","display_name":"Security token","score":0.5848000049591064},{"id":"https://openalex.org/keywords/question-answering","display_name":"Question answering","score":0.52920001745224},{"id":"https://openalex.org/keywords/test","display_name":"Test (biology)","score":0.4684999883174896},{"id":"https://openalex.org/keywords/relevance","display_name":"Relevance (law)","score":0.4185999929904938},{"id":"https://openalex.org/keywords/calibration","display_name":"Calibration","score":0.40720000863075256},{"id":"https://openalex.org/keywords/document-retrieval","display_name":"Document retrieval","score":0.3968999981880188},{"id":"https://openalex.org/keywords/confidence-interval","display_name":"Confidence interval","score":0.3774000108242035},{"id":"https://openalex.org/keywords/low-confidence","display_name":"Low Confidence","score":0.3425999879837036}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6126999855041504},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5848000049591064},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.52920001745224},{"id":"https://openalex.org/C2777267654","wikidata":"https://www.wikidata.org/wiki/Q3519023","display_name":"Test (biology)","level":2,"score":0.4684999883174896},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4320000112056732},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.4262000024318695},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.4185999929904938},{"id":"https://openalex.org/C165838908","wikidata":"https://www.wikidata.org/wiki/Q736777","display_name":"Calibration","level":2,"score":0.40720000863075256},{"id":"https://openalex.org/C161156560","wikidata":"https://www.wikidata.org/wiki/Q1638872","display_name":"Document retrieval","level":2,"score":0.3968999981880188},{"id":"https://openalex.org/C44249647","wikidata":"https://www.wikidata.org/wiki/Q208498","display_name":"Confidence interval","level":2,"score":0.3774000108242035},{"id":"https://openalex.org/C2909755999","wikidata":"https://www.wikidata.org/wiki/Q4751126","display_name":"Low Confidence","level":2,"score":0.3425999879837036},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3393000066280365},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3271999955177307},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.32010000944137573},{"id":"https://openalex.org/C2778067643","wikidata":"https://www.wikidata.org/wiki/Q166507","display_name":"Interval (graph theory)","level":2,"score":0.31439998745918274},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.29339998960494995},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.2881999909877777},{"id":"https://openalex.org/C551230270","wikidata":"https://www.wikidata.org/wiki/Q4368942","display_name":"Data retrieval","level":2,"score":0.27549999952316284},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.25929999351501465},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.25589999556541443},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.25600","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.25600","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2607.25600","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.25600","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,"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"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/10","display_name":"Reduced inequalities","score":0.5546808838844299}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Retrieval-augmented":[0],"generation":[1],"improves":[2],"knowledge-intensive":[3],"question":[4],"answering,":[5],"but":[6],"indiscriminate":[7],"retrieval":[8,32,66,95,108,133,168],"can":[9,25],"introduce":[10],"irrelevant":[11],"evidence":[12,189],"and":[13,43,56,93,109,191],"unnecessary":[14],"computation.":[15],"We":[16,81],"investigate":[17],"whether":[18],"verbalized":[19],"confidence":[20,44],"from":[21],"black-box":[22],"language":[23],"models":[24],"serve":[26],"as":[27,159],"an":[28,150,160],"actionable":[29],"signal":[30],"for":[31,106,111],"routing.":[33],"Our":[34],"method,":[35],"BeyondUncertainty,":[36],"first":[37],"elicits":[38],"a":[39,48,69,140,184],"structured":[40],"provisional":[41,78],"answer":[42,71,79],"estimate,":[45],"then":[46],"applies":[47],"model-specific":[49],"threshold":[50],"selected":[51],"on":[52,126],"held-out":[53],"validation":[54],"data":[55],"frozen":[57],"before":[58],"test":[59],"evaluation.":[60],"Low-confidence":[61],"questions":[62,75,130],"receive":[63],"top-5":[64],"TF-IDF":[65],"followed":[67],"by":[68,118,181],"second":[70],"call,":[72],"whereas":[73],"high-confidence":[74],"return":[76],"the":[77,127,174],"directly.":[80],"evaluate":[82],"27,000":[83],"policy":[84],"instances":[85],"across":[86],"six":[87],"QA":[88],"benchmarks,":[89],"three":[90,94],"model":[91],"families,":[92],"policies.":[96],"BeyondUncertainty":[97],"achieves":[98],"0.483":[99],"mean":[100],"token-level":[101],"F1,":[102],"compared":[103],"with":[104,149],"0.467":[105],"always":[107,122],"0.401":[110],"no":[112],"retrieval,":[113],"while":[114],"reducing":[115],"retrieved":[116],"passages":[117],"20.4\\%":[119],"relative":[120],"to":[121,132],"retrieval.":[123],"When":[124],"matched":[125],"number":[128],"of":[129,146,153],"routed":[131],"within":[134],"each":[135],"dataset-model":[136],"cell,":[137],"it":[138],"outperforms":[139],"post-hoc":[141],"random":[142],"allocation":[143],"in":[144],"17":[145],"18":[147],"settings,":[148],"average":[151],"gain":[152],"0.024":[154],"F1.":[155],"Although":[156],"poorly":[157],"calibrated":[158],"absolute":[161],"probability,":[162],"probe":[163,176],"uncertainty":[164],"modestly":[165],"predicts":[166],"question-level":[167],"benefit":[169],"(AUROC":[170],"=":[171],"0.628).":[172],"However,":[173],"additional":[175],"increases":[177],"total":[178],"token":[179,193],"usage":[180],"28.2\\%,":[182],"revealing":[183],"trade-off":[185],"between":[186],"more":[187],"selective":[188],"acquisition":[190],"end-to-end":[192],"efficiency.":[194]},"counts_by_year":[],"updated_date":"2026-08-08T07:41:36.138363","created_date":"2026-07-30T00:00:00"}