{"id":"https://openalex.org/W7154135049","doi":"https://doi.org/10.48550/arxiv.2604.08749","title":"A Little Rank Goes a Long Way: Random Scaffolds with LoRA Adapters Are All You Need","display_name":"A Little Rank Goes a Long Way: Random Scaffolds with LoRA Adapters Are All You Need","publication_year":2026,"publication_date":"2026-04-09","ids":{"openalex":"https://openalex.org/W7154135049","doi":"https://doi.org/10.48550/arxiv.2604.08749"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.08749","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08749","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2604.08749","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5086359402","display_name":"Hananel Hazan","orcid":"https://orcid.org/0000-0003-1446-1628"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hazan, Hananel","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5133526620","display_name":"Yanbo Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yanbo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031845068","display_name":"Benedikt Hartl","orcid":"https://orcid.org/0000-0001-7787-4839"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hartl, Benedikt","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5133546303","display_name":"Michael Levin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Levin, Michael","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/T12611","display_name":"Neural Networks and Reservoir Computing","score":0.916700005531311,"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/T12611","display_name":"Neural Networks and Reservoir Computing","score":0.916700005531311,"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/T12808","display_name":"Ferroelectric and Negative Capacitance Devices","score":0.020500000566244125,"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/T12676","display_name":"Machine Learning and ELM","score":0.015799999237060547,"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/curse-of-dimensionality","display_name":"Curse of dimensionality","score":0.7196999788284302},{"id":"https://openalex.org/keywords/initialization","display_name":"Initialization","score":0.6571000218391418},{"id":"https://openalex.org/keywords/subspace-topology","display_name":"Subspace topology","score":0.5577999949455261},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.53329998254776},{"id":"https://openalex.org/keywords/encode","display_name":"ENCODE","score":0.4675999879837036},{"id":"https://openalex.org/keywords/principal-component-analysis","display_name":"Principal component analysis","score":0.4449000060558319},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4002000093460083},{"id":"https://openalex.org/keywords/scheduling","display_name":"Scheduling (production processes)","score":0.361299991607666},{"id":"https://openalex.org/keywords/random-projection","display_name":"Random projection","score":0.34150001406669617}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7397000193595886},{"id":"https://openalex.org/C111030470","wikidata":"https://www.wikidata.org/wiki/Q1430460","display_name":"Curse of dimensionality","level":2,"score":0.7196999788284302},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.6571000218391418},{"id":"https://openalex.org/C32834561","wikidata":"https://www.wikidata.org/wiki/Q660730","display_name":"Subspace topology","level":2,"score":0.5577999949455261},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.53329998254776},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.4675999879837036},{"id":"https://openalex.org/C27438332","wikidata":"https://www.wikidata.org/wiki/Q2873","display_name":"Principal component analysis","level":2,"score":0.4449000060558319},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4438999891281128},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4002000093460083},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.361299991607666},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3582000136375427},{"id":"https://openalex.org/C2777036070","wikidata":"https://www.wikidata.org/wiki/Q18393452","display_name":"Random projection","level":2,"score":0.34150001406669617},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.3402000069618225},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3379000127315521},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.30239999294281006},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.296099990606308},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.28790000081062317},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.28700000047683716},{"id":"https://openalex.org/C122123141","wikidata":"https://www.wikidata.org/wiki/Q176623","display_name":"Random variable","level":2,"score":0.2827000021934509},{"id":"https://openalex.org/C88796919","wikidata":"https://www.wikidata.org/wiki/Q1142907","display_name":"Backbone network","level":2,"score":0.2815000116825104},{"id":"https://openalex.org/C74912251","wikidata":"https://www.wikidata.org/wiki/Q6815727","display_name":"Memory footprint","level":2,"score":0.2752000093460083},{"id":"https://openalex.org/C75553542","wikidata":"https://www.wikidata.org/wiki/Q178161","display_name":"A priori and a posteriori","level":2,"score":0.2741999924182892},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.26759999990463257},{"id":"https://openalex.org/C2984118289","wikidata":"https://www.wikidata.org/wiki/Q29954","display_name":"Power consumption","level":3,"score":0.2653000056743622},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.2597000002861023},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.25429999828338623},{"id":"https://openalex.org/C38858127","wikidata":"https://www.wikidata.org/wiki/Q5441228","display_name":"Feed forward","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.08749","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08749","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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.2604.08749","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.08749","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"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":{"How":[0],"many":[1],"of":[2,59,67,78,162,166,169,190],"a":[3,17,75,191,200,212],"neural":[4],"network's":[5],"parameters":[6],"actually":[7],"encode":[8],"task-specific":[9,71],"information?":[10],"We":[11],"investigate":[12],"this":[13,91],"question":[14],"with":[15,216],"LottaLoRA,":[16],"training":[18,64],"paradigm":[19],"in":[20,172],"which":[21,155],"every":[22],"backbone":[23,95,131,196],"weight":[24],"is":[25,96,114,132,179,197],"drawn":[26],"at":[27,154],"random":[28,55,137,201],"and":[29,120,148,225],"frozen;":[30],"only":[31,65],"low-rank":[32,51],"LoRA":[33,122,152],"adapters":[34,52,209],"are":[35],"trained.":[36],"Across":[37],"nine":[38],"benchmarks":[39],"spanning":[40],"diverse":[41],"architecture":[42],"families":[43],"from":[44],"single-layer":[45],"classifiers":[46],"to":[47,182],"900M":[48],"parameter":[49,84],"Transformers":[50],"over":[53],"frozen":[54,94,130],"backbones":[56],"recover":[57],"96-100%":[58],"fully":[60],"trained":[61],"performance":[62,156],"while":[63],"0.5-40%":[66],"the":[68,82,93,101,112,116,121,129,150,159,163,167,187,195],"parameters.":[69],"The":[70,177],"signal":[72],"therefore":[73],"occupies":[74],"subspace":[76],"orders":[77],"magnitude":[79],"smaller":[80],"than":[81],"full":[83],"count":[85],"suggests.":[86],"Three":[87],"mechanistic":[88],"findings":[89],"underpin":[90],"result:(1)":[92],"actively":[97],"exploited":[98],"when":[99,111],"static":[100],"learned":[102],"scaling~$\u03b2$":[103],"remains":[104,144],"strictly":[105],"positive":[106],"across":[107],"all":[108,125],"architectures":[109,230],"but":[110,134],"scaffold":[113],"destabilized,":[115],"optimizer":[117],"silences":[118],"it":[119,143],"factors":[123],"absorb":[124],"task":[126,217],"information;":[127],"(2)":[128],"preferable":[133],"interchangeable":[135],"any":[136],"initialization":[138],"works":[139],"equally":[140],"well,":[141],"provided":[142],"fixed":[145],"throughout":[146],"training;":[147],"(3)":[149],"minimum":[151],"rank":[153],"saturates":[157],"estimates":[158],"intrinsic":[160],"dimensionality":[161],"task,":[164],"reminiscent":[165],"number":[168],"components":[170],"retained":[171],"Principal":[173],"Component":[174],"Analysis":[175],"(PCA).":[176],"construction":[178],"formally":[180],"analogous":[181],"Reservoir":[183],"Computing":[184],"unfolded":[185],"along":[186],"depth":[188],"axis":[189],"feedforward":[192],"network.":[193],"Because":[194],"determined":[198],"by":[199],"seed":[202,211],"alone,":[203],"models":[204],"can":[205],"be":[206],"distributed":[207],"as":[208,229],"plus":[210],"footprint":[213],"that":[214,223],"grows":[215],"complexity,":[218],"not":[219],"model":[220],"size,":[221],"so":[222],"storage":[224],"memory":[226],"savings":[227],"compound":[228],"scale.":[231]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-04-14T00:00:00"}