{"id":"https://openalex.org/W4412945407","doi":"https://doi.org/10.18653/v1/2025.acl-long.529","title":"SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation","display_name":"SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412945407","doi":"https://doi.org/10.18653/v1/2025.acl-long.529"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.529","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.529","pdf_url":"https://aclanthology.org/2025.acl-long.529.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.529.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5113376546","display_name":"Jialong Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jialong Wu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026795459","display_name":"Zhenglin Wang","orcid":"https://orcid.org/0000-0002-5812-2989"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhenglin Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032749334","display_name":"Linhai Zhang","orcid":"https://orcid.org/0000-0003-1009-6870"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Linhai Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yilong Lai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yilong Lai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5066124369","display_name":"Yuanyuan He","orcid":"https://orcid.org/0000-0003-1804-0302"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yulan He","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5074341211","display_name":"Deyu Zhou","orcid":"https://orcid.org/0000-0003-4194-6356"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deyu Zhou","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":"10775","last_page":"10790"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9883000254631042,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.9883000254631042,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T11181","display_name":"Advanced Data Storage Technologies","score":0.970300018787384,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10054","display_name":"Parallel Computing and Optimization Techniques","score":0.9692000150680542,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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/computer-science","display_name":"Computer science","score":0.7005138397216797},{"id":"https://openalex.org/keywords/cache","display_name":"Cache","score":0.6774787902832031},{"id":"https://openalex.org/keywords/scope","display_name":"Scope (computer science)","score":0.6178957223892212},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.5674720406532288},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5206286311149597},{"id":"https://openalex.org/keywords/compression","display_name":"Compression (physics)","score":0.41500234603881836},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.16793662309646606},{"id":"https://openalex.org/keywords/operating-system","display_name":"Operating system","score":0.13372865319252014},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.07197117805480957},{"id":"https://openalex.org/keywords/geology","display_name":"Geology","score":0.06383255124092102},{"id":"https://openalex.org/keywords/materials-science","display_name":"Materials science","score":0.05742835998535156}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7005138397216797},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.6774787902832031},{"id":"https://openalex.org/C2778012447","wikidata":"https://www.wikidata.org/wiki/Q1034415","display_name":"Scope (computer science)","level":2,"score":0.6178957223892212},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5674720406532288},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5206286311149597},{"id":"https://openalex.org/C180016635","wikidata":"https://www.wikidata.org/wiki/Q2712821","display_name":"Compression (physics)","level":2,"score":0.41500234603881836},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.16793662309646606},{"id":"https://openalex.org/C111919701","wikidata":"https://www.wikidata.org/wiki/Q9135","display_name":"Operating system","level":1,"score":0.13372865319252014},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.07197117805480957},{"id":"https://openalex.org/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"score":0.06383255124092102},{"id":"https://openalex.org/C192562407","wikidata":"https://www.wikidata.org/wiki/Q228736","display_name":"Materials science","level":0,"score":0.05742835998535156},{"id":"https://openalex.org/C151730666","wikidata":"https://www.wikidata.org/wiki/Q7205","display_name":"Paleontology","level":1,"score":0.0},{"id":"https://openalex.org/C159985019","wikidata":"https://www.wikidata.org/wiki/Q181790","display_name":"Composite material","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.529","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.529","pdf_url":"https://aclanthology.org/2025.acl-long.529.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.529","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.529","pdf_url":"https://aclanthology.org/2025.acl-long.529.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":[],"awards":[{"id":"https://openalex.org/G3537818187","display_name":null,"funder_award_id":"62176053","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"},{"id":"https://openalex.org/F4320324856","display_name":"Southeast University","ror":"https://ror.org/04ct4d772"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412945407.pdf","grobid_xml":"https://content.openalex.org/works/W4412945407.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4241523039","https://openalex.org/W2360028903","https://openalex.org/W4280543773","https://openalex.org/W178231042","https://openalex.org/W2366083136","https://openalex.org/W2387622493","https://openalex.org/W1932132538","https://openalex.org/W2357832196","https://openalex.org/W1566252468","https://openalex.org/W4200375594"],"abstract_inverted_index":{"Key-Value":[0],"(KV)":[1],"cache":[2,87,99],"has":[3],"become":[4],"a":[5,78,112,155],"bottleneck":[6],"of":[7,59,65,149],"LLMs":[8],"for":[9,20,34,125],"long-context":[10],"generation.Despite":[11],"the":[12,18,21,40,48,57,60,71,90,97,101,108,126,145],"numerous":[13],"efforts":[14],"in":[15,70],"this":[16],"area,":[17],"optimization":[19,30,88],"decoding":[22,93,127],"phase":[23,50,103],"is":[24,31,95,104,118],"generally":[25],"ignored.However,":[26],"we":[27],"believe":[28],"such":[29],"crucial,":[32],"especially":[33],"long-output":[35],"generation":[36],"tasks":[37,73],"based":[38,115],"on":[39,116,142],"following":[41],"two":[42],"observations:":[43],"(i)":[44],"Excessive":[45],"compression":[46,161],"during":[47,89,100],"prefill":[49,91,102],"which":[51],"requires":[52],"specific":[53],"full":[54],"context,":[55],"impairs":[56],"comprehension":[58],"reasoning":[61,72],"task;":[62],"(ii)":[63],"Deviation":[64],"heavy":[66,123],"hitters":[67,124],"1":[68],"occurs":[69],"with":[74],"long":[75],"outputs.Therefore,":[76],"SCOPE,":[77],"simple":[79],"yet":[80],"efficient":[81],"framework":[82],"that":[83],"separately":[84],"performs":[85],"KV":[86,98,160],"and":[92,130,138,147,151],"phases,":[94],"introduced.Specifically,":[96],"preserved":[105],"to":[106,120,157],"maintain":[107],"essential":[109,122],"information,":[110],"while":[111],"novel":[113],"strategy":[114],"sliding":[117],"proposed":[119],"select":[121],"phase.Memory":[128],"usage":[129],"memory":[131],"transfer":[132],"are":[133],"further":[134],"optimized":[135],"using":[136],"adaptive":[137],"discontinuous":[139],"strategies.Extensive":[140],"experiments":[141],"LONGGENBENCH":[143],"show":[144],"effectiveness":[146],"generalization":[148],"SCOPE":[150],"its":[152],"compatibility":[153],"as":[154],"plug-in":[156],"other":[158],"prefill-only":[159],"methods.":[162],"2":[163]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}
