{"id":"https://openalex.org/W7136293840","doi":"https://doi.org/10.48550/arxiv.2603.12396","title":"Test-Time Strategies for More Efficient and Accurate Agentic RAG","display_name":"Test-Time Strategies for More Efficient and Accurate Agentic RAG","publication_year":2026,"publication_date":"2026-03-12","ids":{"openalex":"https://openalex.org/W7136293840","doi":"https://doi.org/10.48550/arxiv.2603.12396"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.12396","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12396","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.12396","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5040603227","display_name":"Brian Zhang","orcid":"https://orcid.org/0009-0003-1047-2247"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Zhang, Brian","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129507454","display_name":"Deepti Guntur","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guntur, Deepti","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5072404474","display_name":"Zhiyang Zuo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zuo, Zhiyang","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129417980","display_name":"Abhinav Sharma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sharma, Abhinav","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129601735","display_name":"Shreyas Chaudhari","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chaudhari, Shreyas","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129459235","display_name":"Wenlong Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Wenlong","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028863551","display_name":"Franck Dernoncourt","orcid":"https://orcid.org/0000-0002-1119-1346"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dernoncourt, Franck","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122953241","display_name":"Puneet Mathur","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mathur, Puneet","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129553800","display_name":"Ryan Rossi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rossi, Ryan","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5016933602","display_name":"Nedim Lipka","orcid":"https://orcid.org/0000-0002-3779-7784"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lipka, Nedim","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5040603227"],"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":0.4997999966144562,"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":0.4997999966144562,"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/T13274","display_name":"Expert finding and Q&A systems","score":0.2248000055551529,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.18320000171661377,"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/pipeline","display_name":"Pipeline (software)","score":0.6704000234603882},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5388000011444092},{"id":"https://openalex.org/keywords/contextualization","display_name":"Contextualization","score":0.4106999933719635},{"id":"https://openalex.org/keywords/face","display_name":"Face (sociological concept)","score":0.33320000767707825},{"id":"https://openalex.org/keywords/natural-language-generation","display_name":"Natural language generation","score":0.3093999922275543},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.2678999900817871}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7508000135421753},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6704000234603882},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5388000011444092},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.42980000376701355},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4178999960422516},{"id":"https://openalex.org/C2780712339","wikidata":"https://www.wikidata.org/wiki/Q5165204","display_name":"Contextualization","level":3,"score":0.4106999933719635},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3431999981403351},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.33320000767707825},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3257000148296356},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.3093999922275543},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.2678999900817871},{"id":"https://openalex.org/C161156560","wikidata":"https://www.wikidata.org/wiki/Q1638872","display_name":"Document retrieval","level":2,"score":0.26409998536109924},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.263700008392334},{"id":"https://openalex.org/C33326189","wikidata":"https://www.wikidata.org/wiki/Q17092450","display_name":"Information integration","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.12396","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12396","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.12396","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.12396","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Retrieval-Augmented":[0],"Generation":[1],"(RAG)":[2],"systems":[3],"face":[4],"challenges":[5,44],"with":[6,120],"complex,":[7],"multihop":[8],"questions,":[9],"and":[10,43,67,95,111,137,158,178,195],"agentic":[11],"frameworks":[12],"such":[13,31],"as":[14],"Search-R1":[15,80,189],"(Jin":[16],"et":[17,134,142],"al.,":[18,135,143],"2025),":[19],"which":[20],"operates":[21],"iteratively,":[22],"have":[23],"been":[24],"proposed":[25],"to":[26,59,78,82,101,187],"address":[27],"these":[28,84],"complexities.":[29],"However,":[30],"approaches":[32,129],"can":[33,57],"introduce":[34],"inefficiencies,":[35],"including":[36],"repetitive":[37],"retrieval":[38,61,196],"of":[39,92,155,162,182],"previously":[40,117],"processed":[41],"information":[42,105],"in":[45,175],"contextualizing":[46],"retrieved":[47,107,118],"results":[48],"effectively":[49],"within":[50],"the":[51,79,90,121,131,138,147,159,180,188],"current":[52],"generation":[53],"prompt.":[54],"Such":[55],"issues":[56],"lead":[58],"unnecessary":[60],"turns,":[62],"suboptimal":[63],"reasoning,":[64,110],"inaccurate":[65],"answers,":[66],"increased":[68],"token":[69],"consumption.":[70],"In":[71],"this":[72],"paper,":[73],"we":[74,88],"investigate":[75],"test-time":[76],"modifications":[77],"pipeline":[81],"mitigate":[83],"identified":[85],"shortcomings.":[86],"Specifically,":[87],"explore":[89],"integration":[91],"two":[93],"components":[94],"their":[96],"combination:":[97],"a":[98,112,172],"contextualization":[99],"module":[100,114],"better":[102],"integrate":[103],"relevant":[104,124],"from":[106],"documents":[108,119],"into":[109],"de-duplication":[113],"that":[115],"replaces":[116],"next":[122],"most":[123],"ones.":[125],"We":[126],"evaluate":[127],"our":[128],"using":[130],"HotpotQA":[132],"(Yang":[133],"2018)":[136],"Natural":[139],"Questions":[140],"(Kwiatkowski":[141],"2019)":[144],"datasets,":[145],"reporting":[146],"exact":[148],"match":[149],"(EM)":[150],"score,":[151],"an":[152],"LLM-as-a-Judge":[153],"assessment":[154],"answer":[156,193],"correctness,":[157],"average":[160],"number":[161,181],"turns.":[163],"Our":[164],"best-performing":[165],"variant,":[166],"utilizing":[167],"GPT-4.1-mini":[168],"for":[169],"contextualization,":[170],"achieves":[171],"5.6%":[173],"increase":[174],"EM":[176],"score":[177],"reduces":[179],"turns":[183],"by":[184],"10.5%":[185],"compared":[186],"baseline,":[190],"demonstrating":[191],"improved":[192],"accuracy":[194],"efficiency.":[197]},"counts_by_year":[],"updated_date":"2026-03-17T07:05:13.627479","created_date":"2026-03-17T00:00:00"}