{"id":"https://openalex.org/W7160900096","doi":"https://doi.org/10.48550/arxiv.2605.08177","title":"Echo-LoRA: Parameter-Efficient Fine-Tuning via Cross-Layer Representation Injection","display_name":"Echo-LoRA: Parameter-Efficient Fine-Tuning via Cross-Layer Representation Injection","publication_year":2026,"publication_date":"2026-05-05","ids":{"openalex":"https://openalex.org/W7160900096","doi":"https://doi.org/10.48550/arxiv.2605.08177"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.08177","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08177","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.2605.08177","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135992315","display_name":"Yihang Peng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Peng, Yihang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135934409","display_name":"Peng Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Peng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135917572","display_name":"Jie Gong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Jie","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135937442","display_name":"Xingyuan Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Xingyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5043925436","display_name":"Lingjiao Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Lingjiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135971104","display_name":"Ning Su","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Su, Ning","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135999194","display_name":"Yan Ran","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ran, Yan","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":0.15039999783039093,"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":0.15039999783039093,"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/T11948","display_name":"Machine Learning in Materials Science","score":0.11169999837875366,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials 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":0.09080000221729279,"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/representation","display_name":"Representation (politics)","score":0.6593999862670898},{"id":"https://openalex.org/keywords/path","display_name":"Path (computing)","score":0.6218000054359436},{"id":"https://openalex.org/keywords/projection","display_name":"Projection (relational algebra)","score":0.5429999828338623},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.4724000096321106},{"id":"https://openalex.org/keywords/boundary","display_name":"Boundary (topology)","score":0.4551999866962433},{"id":"https://openalex.org/keywords/intersection","display_name":"Intersection (aeronautics)","score":0.4507000148296356},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.43639999628067017},{"id":"https://openalex.org/keywords/signal","display_name":"SIGNAL (programming language)","score":0.4124000072479248}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6758000254631042},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6593999862670898},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.6218000054359436},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5440999865531921},{"id":"https://openalex.org/C57493831","wikidata":"https://www.wikidata.org/wiki/Q3134666","display_name":"Projection (relational algebra)","level":2,"score":0.5429999828338623},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.4724000096321106},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.4551999866962433},{"id":"https://openalex.org/C64543145","wikidata":"https://www.wikidata.org/wiki/Q162942","display_name":"Intersection (aeronautics)","level":2,"score":0.4507000148296356},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.43639999628067017},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4309999942779541},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.4124000072479248},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.4002000093460083},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.39959999918937683},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.3684000074863434},{"id":"https://openalex.org/C194544171","wikidata":"https://www.wikidata.org/wiki/Q21105679","display_name":"Gating","level":2,"score":0.3465000092983246},{"id":"https://openalex.org/C2779426996","wikidata":"https://www.wikidata.org/wiki/Q18389128","display_name":"Echo (communications protocol)","level":2,"score":0.32330000400543213},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.3028999865055084},{"id":"https://openalex.org/C198352243","wikidata":"https://www.wikidata.org/wiki/Q37105","display_name":"Line (geometry)","level":2,"score":0.29660001397132874},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.2906999886035919},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.28700000047683716},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2685000002384186},{"id":"https://openalex.org/C13355873","wikidata":"https://www.wikidata.org/wiki/Q2920850","display_name":"Connection (principal bundle)","level":2,"score":0.2612999975681305}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.08177","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08177","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.2605.08177","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.08177","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":{"Parameter-efficient":[0],"fine-tuning":[1],"(PEFT)":[2],"has":[3],"become":[4],"a":[5,61,83],"practical":[6],"route":[7],"for":[8,66],"adapting":[9],"large":[10],"language":[11],"models":[12],"to":[13,26,30,94,114,121],"downstream":[14],"tasks,":[15],"with":[16,167],"LoRA-style":[17],"methods":[18],"being":[19],"particularly":[20],"attractive":[21],"because":[22],"they":[23],"are":[24,112],"inexpensive":[25],"train":[27],"and":[28,47,87,91,109,120,127,149,191],"easy":[29],"deploy.":[31],"Most":[32],"LoRA":[33,101,138,153],"variants,":[34],"however,":[35],"revise":[36],"the":[37,41,49,96,123,136,159,169,182,186],"update":[38],"rule":[39],"within":[40],"weight":[42],"space":[43],"of":[44],"each":[45],"layer":[46],"leave":[48],"intermediate":[50],"representations":[51],"formed":[52],"by":[53,140],"deeper":[54,77],"layers":[55],"largely":[56],"unused.":[57],"We":[58],"propose":[59],"Echo-LoRA,":[60],"cross-layer":[62],"representation":[63],"injection":[64],"method":[65],"parameter-efficient":[67],"fine-tuning.":[68],"During":[69],"training,":[70,180],"Echo-LoRA":[71,134],"collects":[72],"boundary":[73],"hidden":[74],"states":[75],"from":[76],"source":[78],"layers,":[79],"aggregates":[80],"them":[81],"into":[82,99],"sample-level":[84],"echo":[85],"representation,":[86],"uses":[88],"lightweight":[89],"projection":[90],"gating":[92],"networks":[93],"inject":[95],"resulting":[97],"signal":[98],"shallow":[100],"or":[102],"DoRA":[103],"modules.":[104],"Answer-only":[105],"masking,":[106],"masked":[107],"distillation,":[108],"stochastic":[110],"routing":[111],"used":[113],"keep":[115],"this":[116],"auxiliary":[117],"path":[118,176],"stable":[119],"reduce":[122],"gap":[124],"between":[125],"training":[126],"inference.":[128],"On":[129],"eight":[130],"commonsense":[131],"reasoning":[132],"benchmarks,":[133],"exceeds":[135],"reported":[137],"baselines":[139,154],"5.7":[141],"percentage":[142],"points":[143],"on":[144],"average":[145,160],"across":[146],"LLaMA-7B,":[147],"LLaMA2-7B,":[148],"LLaMA3-8B.":[150],"Under":[151],"reproduced":[152],"in":[155],"our":[156],"unified":[157],"implementation,":[158],"gain":[161,170],"is":[162,171,177],"3.0":[163],"points;":[164],"when":[165],"combined":[166],"DoRA,":[168],"2.7":[172],"points.":[173],"The":[174],"Echo":[175],"discarded":[178],"after":[179],"so":[181],"deployed":[183],"model":[184],"keeps":[185],"original":[187],"low-rank":[188],"LoRA/DoRA":[189],"form":[190],"adds":[192],"neither":[193],"inference-time":[194],"parameters":[195],"nor":[196],"inference":[197],"computation.":[198]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-13T00:00:00"}