{"id":"https://openalex.org/W7128625403","doi":"https://doi.org/10.48550/arxiv.2602.09413","title":"LARV: Data-Free Layer-wise Adaptive Rescaling Veneer for Model Merging","display_name":"LARV: Data-Free Layer-wise Adaptive Rescaling Veneer for Model Merging","publication_year":2026,"publication_date":"2026-02-10","ids":{"openalex":"https://openalex.org/W7128625403","doi":"https://doi.org/10.48550/arxiv.2602.09413"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.09413","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","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":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5125679610","display_name":"Xinyu Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Xinyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125661334","display_name":"Ke Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Ke","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125664616","display_name":"Fei Dou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Dou, Fei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125631508","display_name":"Jinbo Bi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bi, Jinbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125630937","display_name":"Jin Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Jin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.49480000138282776,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10036","display_name":"Advanced Neural Network Applications","score":0.49480000138282776,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.1298999935388565,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"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":0.06400000303983688,"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/scaling","display_name":"Scaling","score":0.6470000147819519},{"id":"https://openalex.org/keywords/simple","display_name":"Simple (philosophy)","score":0.611299991607666},{"id":"https://openalex.org/keywords/interference","display_name":"Interference (communication)","score":0.5760999917984009},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4869000017642975},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.44600000977516174},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.41679999232292175},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.3725999891757965}],"concepts":[{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.6470000147819519},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.611299991607666},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6090999841690063},{"id":"https://openalex.org/C32022120","wikidata":"https://www.wikidata.org/wiki/Q797225","display_name":"Interference (communication)","level":3,"score":0.5760999917984009},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4869000017642975},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.45329999923706055},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.44600000977516174},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.41679999232292175},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.3725999891757965},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.36570000648498535},{"id":"https://openalex.org/C2779225514","wikidata":"https://www.wikidata.org/wiki/Q1686140","display_name":"Veneer","level":2,"score":0.33219999074935913},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.33059999346733093},{"id":"https://openalex.org/C42058472","wikidata":"https://www.wikidata.org/wiki/Q810214","display_name":"Base (topology)","level":2,"score":0.3305000066757202},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2921000123023987},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.28459998965263367},{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.2549999952316284},{"id":"https://openalex.org/C2988430800","wikidata":"https://www.wikidata.org/wiki/Q428971","display_name":"Scaling law","level":3,"score":0.25369998812675476}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.09413","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.09413","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.09413","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":"pmh:doi:10.48550/arxiv.2602.09413","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"sustainable_development_goals":[{"score":0.8178715705871582,"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Model":[0],"merging":[1,20,98,245],"aims":[2],"to":[3,15,55,86,119,130,187],"combine":[4],"multiple":[5],"fine-tuned":[6],"models":[7],"into":[8,77,146,246],"a":[9,67,83,110,149,247,253],"single":[10],"multi-task":[11],"model":[12,244],"without":[13],"access":[14],"training":[16],"data.":[17],"Existing":[18],"task-vector":[19,79,135,204],"methods":[21],"such":[22],"as":[23],"TIES,":[24],"TSV-M,":[25],"and":[26,81,92,105,143,169,191,221,227],"Iso-C/CTS":[27],"differ":[28],"in":[29,46],"their":[30],"aggregation":[31],"rules":[32],"but":[33],"treat":[34],"all":[35,203],"layers":[36,52,59],"nearly":[37],"uniformly.":[38],"This":[39],"assumption":[40],"overlooks":[41],"the":[42,127,175,188],"strong":[43],"layer-wise":[44],"heterogeneity":[45],"large":[47],"vision":[48],"transformers,":[49],"where":[50],"shallow":[51],"are":[53],"sensitive":[54],"interference":[56,104,236],"while":[57,178,237],"deeper":[58],"encode":[60],"stable":[61],"task-specific":[62],"features.":[63],"We":[64],"introduce":[65],"LARV,":[66],"training-free,":[68],"data-free,":[69],"merger-agnostic":[70],"Layer-wise":[71],"Adaptive":[72],"Rescaling":[73],"Veneer":[74],"that":[75,171,232],"plugs":[76],"any":[78],"merger":[80,190],"assigns":[82],"per-layer":[84],"scale":[85],"each":[87],"task":[88],"vector":[89],"before":[90],"aggregation,":[91],"show":[93,170],"it":[94],"consistently":[95,201],"boosts":[96],"diverse":[97],"rules.":[99],"LARV":[100,137,184,200,213,233],"adaptively":[101],"suppresses":[102,234],"shallow-layer":[103,235],"amplifies":[106],"deeper-layer":[107],"alignment":[108],"using":[109],"simple":[111,139],"deterministic":[112],"schedule,":[113],"requiring":[114],"no":[115],"retraining":[116],"or":[117,166],"modification":[118],"existing":[120],"mergers.":[121],"To":[122],"our":[123],"knowledge,":[124],"this":[125],"is":[126,185],"first":[128],"work":[129],"perform":[131],"layer-aware":[132,249],"scaling":[133,162],"for":[134,209],"merging.":[136],"computes":[138],"data-free":[140],"layer":[141],"proxies":[142],"turns":[144],"them":[145],"scales":[147],"through":[148],"lightweight":[150],"rule;":[151],"we":[152],"study":[153],"several":[154],"instantiations":[155],"within":[156],"one":[157],"framework":[158],"(e.g.,":[159],"tiered":[160,172],"two/three-level":[161],"with":[163,197],"fixed":[164],"values,":[165],"continuous":[167,179],"mappings)":[168],"choices":[173],"offer":[174],"best":[176],"robustness,":[177],"mappings":[180],"remain":[181],"an":[182],"ablation.":[183],"orthogonal":[186],"base":[189],"adds":[192],"negligible":[193],"cost.":[194],"On":[195],"FusionBench":[196],"Vision":[198],"Transformers,":[199],"improves":[202],"baselines":[205],"across":[206],"8/14/20-task":[207],"settings;":[208],"example,":[210],"Iso-C":[211],"+":[212],"reaches":[214],"85.9%":[215],"on":[216,219,223],"ViT-B/32,":[217],"89.2%":[218],"ViT-B/16,":[220],"92.6%":[222],"ViT-L/14.":[224],"Layerwise":[225],"analysis":[226],"corruption":[228],"tests":[229],"further":[230],"indicate":[231],"modestly":[238],"amplifying":[239],"deeper,":[240],"task-stable":[241],"features,":[242],"turning":[243],"robust,":[248],"procedure":[250],"rather":[251],"than":[252],"uniform":[254],"one.":[255]},"counts_by_year":[],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-02-12T00:00:00"}