{"id":"https://openalex.org/W4412886660","doi":"https://doi.org/10.18653/v1/2025.acl-long.250","title":"Pandora\u2019s Box or Aladdin\u2019s Lamp: A Comprehensive Analysis Revealing the Role of RAG Noise in Large Language Models","display_name":"Pandora\u2019s Box or Aladdin\u2019s Lamp: A Comprehensive Analysis Revealing the Role of RAG Noise in Large Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412886660","doi":"https://doi.org/10.18653/v1/2025.acl-long.250"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.250","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.250","pdf_url":"https://aclanthology.org/2025.acl-long.250.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.250.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5064434232","display_name":"Jinyang Wu","orcid":"https://orcid.org/0000-0002-3285-7389"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jinyang Wu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100328947","display_name":"Shuai Zhang","orcid":"https://orcid.org/0000-0003-4291-8170"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shuai Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5014276343","display_name":"Feihu Che","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feihu Che","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5023226704","display_name":"Mingkuan Feng","orcid":"https://orcid.org/0000-0001-9035-9014"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mingkuan Feng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018467344","display_name":"Pengpeng Shao","orcid":"https://orcid.org/0000-0003-4229-8768"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pengpeng Shao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5112613657","display_name":"Jianhua Tao","orcid":"https://orcid.org/0000-0002-9344-6428"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jianhua Tao","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":4,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"5019","last_page":"5039"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.7706000208854675,"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":0.7706000208854675,"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/noise","display_name":"Noise (video)","score":0.5497808456420898},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.4506436288356781},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.20248430967330933}],"concepts":[{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.5497808456420898},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4506436288356781},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.20248430967330933},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.250","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.250","pdf_url":"https://aclanthology.org/2025.acl-long.250.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.250","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.250","pdf_url":"https://aclanthology.org/2025.acl-long.250.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.4300000071525574,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412886660.pdf","grobid_xml":"https://content.openalex.org/works/W4412886660.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Retrieval-Augmented":[0],"Generation":[1],"(RAG)":[2],"has":[3,19],"emerged":[4],"as":[5],"a":[6,64,69,74],"key":[7],"method":[8],"to":[9,23,32,43,112,122],"address":[10],"hallucinations":[11,154],"in":[12],"large":[13],"language":[14],"models":[15,22],"(LLMs).While":[16],"recent":[17],"research":[18],"extended":[20],"RAG":[21,71,150],"complex":[24],"noisy":[25],"scenarios,":[26],"these":[27,98],"explorations":[28],"often":[29],"confine":[30],"themselves":[31],"limited":[33],"noise":[34,39,61,108,118,128,133],"types":[35,62],"and":[36,51,67,81,93,117,141,152],"presuppose":[37],"that":[38,97,109,119],"is":[40,110,120,159],"inherently":[41],"detrimental":[42],"LLMs,":[44],"potentially":[45],"deviating":[46],"from":[47,63],"real-world":[48],"retrieval":[49,157],"environments":[50],"restricting":[52],"practical":[53,106],"applicability.In":[54],"this":[55],"paper,":[56],"we":[57,95],"define":[58],"seven":[59],"distinct":[60],"linguistic":[65],"perspective":[66],"establish":[68],"Noise":[70],"Benchmark":[72],"(NoiserBench),":[73],"comprehensive":[75],"evaluation":[76,85],"framework":[77],"encompassing":[78],"multiple":[79],"datasets":[80],"reasoning":[82],"tasks.Through":[83],"empirical":[84],"of":[86,138],"eight":[87],"representative":[88],"LLMs":[89,113,123],"with":[90],"diverse":[91,156],"architectures":[92],"scales,":[94],"reveal":[96],"noises":[99],"can":[100],"be":[101],"further":[102],"categorized":[103],"into":[104],"two":[105],"groups:":[107],"beneficial":[111,115,132],"(aka":[114,124],"noise)":[116],"harmful":[121,125,127],"noise).While":[126],"generally":[129],"impairs":[130],"performance,":[131],"may":[134],"enhance":[135],"several":[136],"aspects":[137],"model":[139],"capabilities":[140],"overall":[142],"performance.Our":[143],"analysis":[144],"offers":[145],"insights":[146],"for":[147],"developing":[148],"robust":[149],"solutions":[151],"mitigating":[153],"across":[155],"scenarios.Code":[158],"available":[160],"at":[161],"https://github.com/jinyangwu/NoiserBench.":[162]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}