{"id":"https://openalex.org/W7161842310","doi":"https://doi.org/10.48550/arxiv.2605.18822","title":"Hybrid-LoRA: Bridging Full Fine-Tuning and Low-Rank Adaptation for Post-Training","display_name":"Hybrid-LoRA: Bridging Full Fine-Tuning and Low-Rank Adaptation for Post-Training","publication_year":2026,"publication_date":"2026-05-12","ids":{"openalex":"https://openalex.org/W7161842310","doi":"https://doi.org/10.48550/arxiv.2605.18822"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.18822","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18822","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2605.18822","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5047129925","display_name":"Chengqian Zhang","orcid":"https://orcid.org/0000-0001-7829-0587"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Chengqian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136571396","display_name":"Wei Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5103224637","display_name":"Kyumin Lee","orcid":"https://orcid.org/0000-0002-9004-1740"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lee, Kyumin","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.2630000114440918,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.2630000114440918,"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/T10028","display_name":"Topic Modeling","score":0.19869999587535858,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.12099999934434891,"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/bridging","display_name":"Bridging (networking)","score":0.8062999844551086},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6715999841690063},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.6360999941825867},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.6244000196456909},{"id":"https://openalex.org/keywords/verifiable-secret-sharing","display_name":"Verifiable secret sharing","score":0.4122999906539917},{"id":"https://openalex.org/keywords/sensitivity","display_name":"Sensitivity (control systems)","score":0.31299999356269836}],"concepts":[{"id":"https://openalex.org/C174348530","wikidata":"https://www.wikidata.org/wiki/Q188635","display_name":"Bridging (networking)","level":2,"score":0.8062999844551086},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7688000202178955},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6715999841690063},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.6360999941825867},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.6244000196456909},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4871000051498413},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.44200000166893005},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.4122999906539917},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.31299999356269836},{"id":"https://openalex.org/C2778915421","wikidata":"https://www.wikidata.org/wiki/Q3643177","display_name":"Performance improvement","level":2,"score":0.30559998750686646},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.30140000581741333},{"id":"https://openalex.org/C164226766","wikidata":"https://www.wikidata.org/wiki/Q7293202","display_name":"Rank (graph theory)","level":2,"score":0.3005000054836273},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.27570000290870667},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2524999976158142}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.18822","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18822","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.18822","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.18822","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Post-training":[0],"has":[1,28],"become":[2],"essential":[3],"for":[4,37,96],"adapting":[5,128],"large":[6],"language":[7],"models":[8],"(LLMs)":[9],"to":[10,91,116,124,140,145,148,191],"complex":[11,97],"downstream":[12],"behaviors,":[13],"including":[14],"instruction":[15],"following,":[16],"preference":[17],"alignment,":[18],"and":[19,47,63,193],"multi-step":[20],"reasoning.":[21],"Reinforcement":[22],"learning":[23],"with":[24,41,55,132,172],"verifiable":[25],"rewards":[26],"(RLVR)":[27],"recently":[29],"emerged":[30],"as":[31,45,74],"a":[32,86,117,136,152,166],"particularly":[33],"effective":[34],"post-training":[35,54,95,109,185],"paradigm":[36],"improving":[38],"reasoning":[39,98],"capabilities,":[40],"critic-free":[42],"algorithms":[43],"such":[44,73],"GRPO":[46],"GSPO":[48],"enabling":[49],"scalable":[50],"optimization.":[51],"However,":[52],"RLVR":[53],"full":[56,92,114,162,168],"fine-tuning":[57,70,93,115,163,169],"(FFT)":[58],"requires":[59],"substantial":[60],"GPU":[61],"memory":[62],"incurs":[64],"high":[65],"training":[66],"costs.":[67],"Although":[68],"parameter-efficient":[69],"(PEFT)":[71],"methods,":[72],"Low-Rank":[75],"Adaptation":[76],"(LoRA),":[77],"effectively":[78],"reduce":[79],"computational":[80],"costs,":[81],"they":[82],"often":[83],"suffer":[84],"from":[85],"noticeable":[87],"performance":[88,164],"gap":[89],"compared":[90],"in":[94],"tasks.":[99],"In":[100],"this":[101],"paper,":[102],"we":[103],"propose":[104],"Hybrid-LoRA,":[105],"an":[106],"efficient":[107],"hybrid":[108],"framework":[110],"that":[111,158],"selectively":[112],"applies":[113],"small":[118],"subset":[119],"of":[120,189],"modules":[121,143,176],"less":[122],"suited":[123],"low-rank":[125,149],"adaptation,":[126],"while":[127],"the":[129,173,198],"remaining":[130,174],"components":[131],"LoRA.":[133],"We":[134],"introduce":[135],"novel":[137],"Hybrid-LoRA":[138,159],"Score":[139],"rank":[141],"candidate":[142,175],"according":[144],"their":[146],"sensitivity":[147],"adaptation":[150],"under":[151,165],"fixed":[153],"parameter":[154],"budget.":[155],"Experiments":[156],"show":[157],"closely":[160],"matches":[161],"10%":[167],"module":[170],"budget,":[171],"adapted":[177],"by":[178],"LoRA,":[179],"consistently":[180],"outperforming":[181],"four":[182],"state-of-the-art":[183],"PEFT":[184],"baselines,":[186],"achieving":[187],"improvements":[188],"up":[190],"5.65%":[192],"on":[194],"average":[195],"4.36%":[196],"over":[197],"best":[199],"baseline.":[200]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-21T00:00:00"}