{"id":"https://openalex.org/W6982336101","doi":null,"title":"Impact of chunking granularity on accuracy and token consumption in retrieval-augmented generation for question-answering","display_name":"Impact of chunking granularity on accuracy and token consumption in retrieval-augmented generation for question-answering","publication_year":2025,"publication_date":"2025-07-17","ids":{"openalex":"https://openalex.org/W6982336101"},"language":"en","primary_location":{"id":"pmh:oai:aaltodoc.aalto.fi:123456789/137892","doi":null,"is_oa":true,"landing_page_url":"https://aaltodoc.aalto.fi/handle/123456789/137892","pdf_url":"https://aaltodoc.aalto.fi/bitstreams/ed0e3647-2203-4a63-aea7-fa676720906c/download","source":{"id":"https://openalex.org/S4306401662","display_name":"Aaltodoc (Aalto University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I9927081","host_organization_name":"Aalto University","host_organization_lineage":["https://openalex.org/I9927081"],"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":"G2 Pro gradu, diplomity\u00f6"},"volume":null,"type":"dissertation","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://aaltodoc.aalto.fi/bitstreams/ed0e3647-2203-4a63-aea7-fa676720906c/download","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Lensu, Urho","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Lensu, Urho","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":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":true,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9726999998092651,"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":0.9726999998092651,"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":0.022600000724196434,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.0010999999940395355,"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/retrieval-augmented-generation","display_name":"retrieval-augmented generation","score":0.9169999957084656},{"id":"https://openalex.org/keywords/document-chunking","display_name":"document chunking","score":0.9769999980926514},{"id":"https://openalex.org/keywords/token-efficiency","display_name":"token efficiency","score":0.6869999766349792},{"id":"https://openalex.org/keywords/question-answering","display_name":"question answering","score":0.9399999976158142},{"id":"https://openalex.org/keywords/dense-retrieval","display_name":"dense retrieval","score":0.7990000247955322},{"id":"https://openalex.org/keywords/token-consumption","display_name":"token consumption","score":0.4790000021457672},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.5709999799728394}],"concepts":[{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.8648999929428101},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8384000062942505},{"id":"https://openalex.org/C203357204","wikidata":"https://www.wikidata.org/wiki/Q1089605","display_name":"Chunking (psychology)","level":2,"score":0.8276000022888184},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.6190999746322632},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5605999827384949},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.42239999771118164},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.4092999994754791},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.39469999074935913},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3625999987125397},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.329800009727478},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.32199999690055847},{"id":"https://openalex.org/C48105269","wikidata":"https://www.wikidata.org/wiki/Q1141160","display_name":"Header","level":2,"score":0.3215000033378601},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.32019999623298645},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.3199000060558319},{"id":"https://openalex.org/C81669768","wikidata":"https://www.wikidata.org/wiki/Q2359161","display_name":"Precision and recall","level":2,"score":0.3107999861240387},{"id":"https://openalex.org/C203005215","wikidata":"https://www.wikidata.org/wiki/Q79798","display_name":"Machine translation","level":2,"score":0.2888000011444092},{"id":"https://openalex.org/C148524875","wikidata":"https://www.wikidata.org/wiki/Q6975395","display_name":"F1 score","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.27950000762939453},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.2736000120639801},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.2705000042915344},{"id":"https://openalex.org/C2776187449","wikidata":"https://www.wikidata.org/wiki/Q1513879","display_name":"Natural language generation","level":3,"score":0.25209999084472656},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.25029999017715454}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:aaltodoc.aalto.fi:123456789/137892","doi":null,"is_oa":true,"landing_page_url":"https://aaltodoc.aalto.fi/handle/123456789/137892","pdf_url":"https://aaltodoc.aalto.fi/bitstreams/ed0e3647-2203-4a63-aea7-fa676720906c/download","source":{"id":"https://openalex.org/S4306401662","display_name":"Aaltodoc (Aalto University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I9927081","host_organization_name":"Aalto University","host_organization_lineage":["https://openalex.org/I9927081"],"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":"G2 Pro gradu, diplomity\u00f6"}],"best_oa_location":{"id":"pmh:oai:aaltodoc.aalto.fi:123456789/137892","doi":null,"is_oa":true,"landing_page_url":"https://aaltodoc.aalto.fi/handle/123456789/137892","pdf_url":"https://aaltodoc.aalto.fi/bitstreams/ed0e3647-2203-4a63-aea7-fa676720906c/download","source":{"id":"https://openalex.org/S4306401662","display_name":"Aaltodoc (Aalto University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I9927081","host_organization_name":"Aalto University","host_organization_lineage":["https://openalex.org/I9927081"],"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":"G2 Pro gradu, diplomity\u00f6"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W6982336101.pdf","grobid_xml":"https://content.openalex.org/works/W6982336101.grobid-xml"},"referenced_works_count":97,"referenced_works":["https://openalex.org/W2017561954","https://openalex.org/W2116148009","https://openalex.org/W2259472270","https://openalex.org/W2626778328","https://openalex.org/W2896457183","https://openalex.org/W2981852735","https://openalex.org/W2983088655","https://openalex.org/W3027879771","https://openalex.org/W3030163527","https://openalex.org/W3122916846","https://openalex.org/W3132607382","https://openalex.org/W3161970973","https://openalex.org/W3185341429","https://openalex.org/W4225141805","https://openalex.org/W4295857769","https://openalex.org/W4296713955","https://openalex.org/W4307079201","https://openalex.org/W4309674289","https://openalex.org/W4321367331","https://openalex.org/W4362515116","https://openalex.org/W4367628377","https://openalex.org/W4381332452","https://openalex.org/W4382246105","https://openalex.org/W4384388099","https://openalex.org/W4386435778","https://openalex.org/W4386755313","https://openalex.org/W4387390366","https://openalex.org/W4387800173","https://openalex.org/W4388184452","https://openalex.org/W4389984066","https://openalex.org/W4390607214","https://openalex.org/W4390647650","https://openalex.org/W4390790246","https://openalex.org/W4391136507","https://openalex.org/W4391158659","https://openalex.org/W4391709255","https://openalex.org/W4391876619","https://openalex.org/W4391971084","https://openalex.org/W4393435697","https://openalex.org/W4393487710","https://openalex.org/W4394780591","https://openalex.org/W4394838906","https://openalex.org/W4394929444","https://openalex.org/W4394947112","https://openalex.org/W4395686609","https://openalex.org/W4396242252","https://openalex.org/W4399356589","https://openalex.org/W4399530612","https://openalex.org/W4399794929","https://openalex.org/W4401042753","https://openalex.org/W4401306278","https://openalex.org/W4401307379","https://openalex.org/W4401857375","https://openalex.org/W4402482792","https://openalex.org/W4402854593","https://openalex.org/W4403363055","https://openalex.org/W4403578737","https://openalex.org/W4403583523","https://openalex.org/W4403885195","https://openalex.org/W4404129812","https://openalex.org/W4404342118","https://openalex.org/W4404345592","https://openalex.org/W4404534210","https://openalex.org/W4404781852","https://openalex.org/W4404782267","https://openalex.org/W4404783118","https://openalex.org/W4404988050","https://openalex.org/W4405468714","https://openalex.org/W4405518282","https://openalex.org/W4406031095","https://openalex.org/W4406085118","https://openalex.org/W4409248734","https://openalex.org/W4409657237","https://openalex.org/W4410202262","https://openalex.org/W4411113004","https://openalex.org/W4411119855","https://openalex.org/W4411120199","https://openalex.org/W4411120544","https://openalex.org/W4411368489","https://openalex.org/W4411403346","https://openalex.org/W4412158322","https://openalex.org/W4412396177","https://openalex.org/W4412945697","https://openalex.org/W4416034797","https://openalex.org/W4416035209","https://openalex.org/W4416379250","https://openalex.org/W4416905789","https://openalex.org/W6910741274","https://openalex.org/W6922346929","https://openalex.org/W6928994809","https://openalex.org/W6929646337","https://openalex.org/W6941243218","https://openalex.org/W6948434932","https://openalex.org/W7118085507","https://openalex.org/W7133185274","https://openalex.org/W7133205974","https://openalex.org/W7134117706"],"related_works":["https://openalex.org/W7154900482","https://openalex.org/W7125352456","https://openalex.org/W7162122905","https://openalex.org/W7214360078","https://openalex.org/W7171663821","https://openalex.org/W4416574656","https://openalex.org/W7167231154","https://openalex.org/W7125362215","https://openalex.org/W4409183391","https://openalex.org/W7162291267"],"abstract_inverted_index":{"This":[0,68,261],"research":[1,69,101],"investigates":[2,71],"the":[3,28,60,87,162,239],"role":[4],"of":[5,89,121,164,180],"document":[6],"chunking":[7,47,73],"in":[8,82,263],"retrieval-augmented":[9],"generation":[10],"systems.":[11],"Even":[12],"though":[13],"RAG":[14,76,92,142,199],"addresses":[15],"LLM":[16],"drawbacks":[17],"such":[18],"as":[19],"knowledge":[20],"cutoffs":[21],"and":[22,37,65,79,97,111,124,133,147,155,182,195,225,255,285],"errors,":[23],"performance":[24],"is":[25],"impacted":[26],"by":[27,252,272],"techniques":[29,48],"employed":[30],"to":[31,85,107,244,267],"divide":[32],"documents":[33],"into":[34],"searchable":[35],"sections,":[36],"these,":[38],"particularly":[39],"concerning":[40],"chunk":[41,119,240],"granularity,":[42],"are":[43,145],"often":[44,204,247],"underexplored.":[45],"Conventional":[46],"can":[49],"fragment":[50],"context":[51,196],"or":[52,276],"overlook":[53],"complex":[54],"details":[55],"within":[56],"documents,":[57],"which":[58],"challenges":[59],"balance":[61,192],"between":[62,193],"retrieval":[63,153],"accuracy":[64,78,194,234,269],"token":[66,80,250,289],"consumption.":[67],"systematically":[70],"how":[72],"choices":[74],"affect":[75],"system":[77,154,172,212],"efficiency":[81],"question-answering,":[83],"aiming":[84],"inform":[86],"design":[88,106],"more":[90,257],"resource-conscious":[91],"systems":[93],"that":[94,188],"consider":[95],"operational":[96],"environmental":[98],"costs.":[99],"The":[100,209],"employs":[102,128],"a":[103,149,169,177],"quantitative,":[104],"experimental":[105],"compare":[108],"fixed-size,":[109],"sentence-aware,":[110],"hybrid":[112],"chunking,":[113],"each":[114],"applied":[115],"at":[116],"target":[117],"maximum":[118],"granularities":[120],"256,":[122],"512,":[123],"1024":[125],"tokens.":[126],"It":[127],"twenty":[129],"10-Q":[130],"financial":[131],"reports":[132],"an":[134,156],"existing":[135],"195-pair":[136],"question-answer":[137],"dataset":[138],"from":[139,242],"GitHub.":[140],"Two":[141],"pipeline":[143],"architectures":[144],"implemented":[146],"evaluated,":[148],"baseline":[150],"NaiveRAG":[151],"dense":[152],"advanced":[157,210],"graph-based":[158],"PathRAG":[159,211],"system.":[160],"While":[161,278],"number":[163],"retrieved":[165,249],"tokens":[166,246,265],"measures":[167],"efficiency,":[168],"multifaceted":[170],"evaluation":[171],"evaluates":[173],"answer":[174],"quality":[175],"using":[176],"wide":[178],"range":[179],"lexical":[181],"semantic":[183],"similarity":[184],"indicators.":[185],"Findings":[186],"indicate":[187],"512-token":[189,202],"chunks":[190,231,293],"generally":[191],"for":[197,207,222],"both":[198],"systems,":[200],"with":[201,283,291],"methods":[203],"proving":[205],"effective":[206],"NaiveRAG.":[208],"also":[213,281],"effectively":[214],"utilizes":[215],"concise":[216],"256-token":[217],"chunks,":[218],"especially":[219],"fixed-size":[220],"ones":[221],"specific":[223],"metrics":[224],"tabular":[226],"data.":[227],"Conversely,":[228],"large":[229],"1024-token":[230,292],"consistently":[232,294],"diminish":[233],"across":[235],"all":[236],"configurations.":[237],"Increasing":[238],"size":[241],"256":[243],"512":[245],"inflates":[248],"counts":[251],"over":[253],"half":[254],"sometimes":[256],"than":[258],"doubles":[259],"them.":[260],"increase":[262],"consumed":[264],"leads":[266],"modest":[268],"gains,":[270],"occasionally":[271],"two":[273],"percentage":[274],"points":[275],"less.":[277],"preferred":[279],"granularity":[280],"shifts":[282],"question":[284],"source":[286],"type,":[287],"further":[288],"escalation":[290],"degrades":[295],"overall":[296],"performance.":[297]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-10T00:00:00"}