{"id":"https://openalex.org/W7154134157","doi":"https://doi.org/10.48550/arxiv.2604.08964","title":"Breaking Block Boundaries: Anchor-based History-stable Decoding for Diffusion Large Language Models","display_name":"Breaking Block Boundaries: Anchor-based History-stable Decoding for Diffusion Large Language Models","publication_year":2026,"publication_date":"2026-04-10","ids":{"openalex":"https://openalex.org/W7154134157","doi":"https://doi.org/10.48550/arxiv.2604.08964"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.08964","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08964","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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.08964","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5102789826","display_name":"Shun Zou","orcid":"https://orcid.org/0000-0003-2813-7167"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zou, Shun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100751600","display_name":"Yong Wang","orcid":"https://orcid.org/0000-0002-5383-5736"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100678098","display_name":"Zehui Chen","orcid":"https://orcid.org/0000-0002-1843-4478"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Zehui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133541808","display_name":"Lin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Lin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133490849","display_name":"Chongyang Tao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tao, Chongyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133537188","display_name":"Feng Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Feng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133518005","display_name":"Xiangxiang Chu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chu, Xiangxiang","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":0.23770000040531158,"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":0.23770000040531158,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.1152999997138977,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.10459999740123749,"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/decoding-methods","display_name":"Decoding methods","score":0.9380999803543091},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.7674000263214111},{"id":"https://openalex.org/keywords/sequential-decoding","display_name":"Sequential decoding","score":0.6227999925613403},{"id":"https://openalex.org/keywords/list-decoding","display_name":"List decoding","score":0.5813000202178955},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.5475000143051147},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.5249999761581421},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.4675999879837036},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.46309998631477356}],"concepts":[{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.9380999803543091},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.7674000263214111},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7142000198364258},{"id":"https://openalex.org/C193969084","wikidata":"https://www.wikidata.org/wiki/Q7452500","display_name":"Sequential decoding","level":4,"score":0.6227999925613403},{"id":"https://openalex.org/C204397858","wikidata":"https://www.wikidata.org/wiki/Q4437907","display_name":"List decoding","level":5,"score":0.5813000202178955},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.5475000143051147},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.5249999761581421},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5051000118255615},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.4675999879837036},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.46309998631477356},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.45730000734329224},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.4194999933242798},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.4169999957084656},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.39489999413490295},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3411000072956085},{"id":"https://openalex.org/C2778858076","wikidata":"https://www.wikidata.org/wiki/Q5249539","display_name":"Decodes","level":3,"score":0.3402000069618225},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.32670000195503235},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.32019999623298645},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.3167000114917755},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3043000102043152},{"id":"https://openalex.org/C132727834","wikidata":"https://www.wikidata.org/wiki/Q4892328","display_name":"Berlekamp\u2013Welch algorithm","level":5,"score":0.26989999413490295},{"id":"https://openalex.org/C157125643","wikidata":"https://www.wikidata.org/wiki/Q884707","display_name":"Block code","level":3,"score":0.26260000467300415},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2547000050544739}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.08964","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08964","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.08964","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08964","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":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Diffusion":[0],"Large":[1],"Language":[2],"Models":[3],"(dLLMs)":[4],"have":[5],"recently":[6],"become":[7],"a":[8,107,128],"promising":[9],"alternative":[10],"to":[11,31,56,137],"autoregressive":[12],"large":[13],"language":[14],"models":[15],"(ARMs).":[16],"Semi-autoregressive":[17],"(Semi-AR)":[18],"decoding":[19,28,41,50,80,111,136,172,184],"is":[20,81,95],"widely":[21],"employed":[22],"in":[23,121,169],"base":[24],"dLLMs":[25],"and":[26,72,91,140,147,157],"advanced":[27,171],"strategies":[29],"due":[30],"its":[32],"superior":[33],"performance.":[34,141],"However,":[35],"our":[36,181],"observations":[37],"reveal":[38],"that":[39,151],"Semi-AR":[40],"suffers":[42],"from":[43],"inherent":[44],"block":[45],"constraints,":[46],"which":[47],"cause":[48],"the":[49,67,116,164,178],"of":[51,69,119],"many":[52],"cross-block":[53,135],"stable":[54,70],"tokens":[55,71,120],"be":[57],"unnecessarily":[58],"delayed.":[59],"To":[60],"address":[61],"this":[62],"challenge,":[63],"we":[64,101],"systematically":[65],"investigate":[66],"identification":[68],"present":[73],"three":[74],"key":[75],"findings:":[76],"(1)":[77],"naive":[78],"lookahead":[79],"unreliable,":[82],"(2)":[83],"token":[84,129],"stability":[85,117],"closely":[86],"correlates":[87],"with":[88],"convergence":[89],"trend,":[90],"(3)":[92],"historical":[93],"information":[94],"isolated.":[96],"Building":[97],"on":[98,177],"these":[99],"insights,":[100],"propose":[102],"Anchor-based":[103],"History-stable":[104],"Decoding":[105],"(AHD),":[106],"training-free,":[108],"plug-and-play":[109],"dynamic":[110,125],"strategy.":[112],"Specifically,":[113],"AHD":[114,152,161],"monitors":[115],"trend":[118],"real":[122],"time":[123],"through":[124],"anchors.":[126],"Once":[127],"reaches":[130],"stability,":[131],"it":[132],"initiates":[133],"early":[134],"enhance":[138],"efficiency":[139],"Extensive":[142],"experiments":[143],"across":[144],"language,":[145],"vision-language,":[146],"audio-language":[148],"domains":[149],"demonstrate":[150],"simultaneously":[153],"improves":[154],"both":[155],"performance":[156,165,190],"inference":[158],"efficiency.":[159],"Notably,":[160],"effectively":[162],"reverses":[163],"degradation":[166],"typically":[167],"observed":[168],"existing":[170],"acceleration":[173],"strategies.":[174],"For":[175],"instance,":[176],"BBH":[179],"benchmark,":[180],"approach":[182],"reduces":[183],"steps":[185],"by":[186,191],"80%":[187],"while":[188],"improving":[189],"3.67%.":[192]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-14T00:00:00"}