{"id":"https://openalex.org/W7131412136","doi":"https://doi.org/10.48550/arxiv.2602.19058","title":"Do LLMs and VLMs Share Neurons for Inference? Evidence and Mechanisms of Cross-Modal Transfer","display_name":"Do LLMs and VLMs Share Neurons for Inference? Evidence and Mechanisms of Cross-Modal Transfer","publication_year":2026,"publication_date":"2026-02-22","ids":{"openalex":"https://openalex.org/W7131412136","doi":"https://doi.org/10.48550/arxiv.2602.19058"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.19058","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19058","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.2602.19058","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126843180","display_name":"Chenhang Cui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cui, Chenhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126798276","display_name":"An Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, An","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126837969","display_name":"Yuxin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Yuxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124927398","display_name":"Gelei Deng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Deng, Gelei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126688538","display_name":"Jingnan Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Jingnan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5084611756","display_name":"Zhenkai Liang","orcid":"https://orcid.org/0000-0001-7138-5030"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liang, Zhenkai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126800466","display_name":"Xiang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Xiang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126788620","display_name":"Tat-Seng Chua","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chua, Tat-Seng","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":0.4821999967098236,"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":0.4821999967098236,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.08399999886751175,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.0803999975323677,"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/inference","display_name":"Inference","score":0.8883000016212463},{"id":"https://openalex.org/keywords/perception","display_name":"Perception","score":0.5626000165939331},{"id":"https://openalex.org/keywords/causal-inference","display_name":"Causal inference","score":0.5367000102996826},{"id":"https://openalex.org/keywords/encode","display_name":"ENCODE","score":0.47540000081062317},{"id":"https://openalex.org/keywords/reciprocal","display_name":"Reciprocal","score":0.38580000400543213},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.37450000643730164},{"id":"https://openalex.org/keywords/unification","display_name":"Unification","score":0.3434000015258789},{"id":"https://openalex.org/keywords/mechanism","display_name":"Mechanism (biology)","score":0.33709999918937683}],"concepts":[{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.8883000016212463},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.5626000165939331},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5429999828338623},{"id":"https://openalex.org/C158600405","wikidata":"https://www.wikidata.org/wiki/Q5054566","display_name":"Causal inference","level":2,"score":0.5367000102996826},{"id":"https://openalex.org/C66746571","wikidata":"https://www.wikidata.org/wiki/Q1134833","display_name":"ENCODE","level":3,"score":0.47540000081062317},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4052000045776367},{"id":"https://openalex.org/C2777742833","wikidata":"https://www.wikidata.org/wiki/Q1964083","display_name":"Reciprocal","level":2,"score":0.38580000400543213},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.37450000643730164},{"id":"https://openalex.org/C96146094","wikidata":"https://www.wikidata.org/wiki/Q609057","display_name":"Unification","level":2,"score":0.3434000015258789},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.33709999918937683},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.32580000162124634},{"id":"https://openalex.org/C3746660","wikidata":"https://www.wikidata.org/wiki/Q1068763","display_name":"Rule of inference","level":2,"score":0.3181000053882599},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.30709999799728394},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.30410000681877136},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.29409998655319214},{"id":"https://openalex.org/C39743133","wikidata":"https://www.wikidata.org/wiki/Q3262595","display_name":"Signalling","level":2,"score":0.27730000019073486},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.26840001344680786},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.26499998569488525},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.2542000114917755},{"id":"https://openalex.org/C2776175482","wikidata":"https://www.wikidata.org/wiki/Q1195816","display_name":"Transfer (computing)","level":2,"score":0.2524000108242035},{"id":"https://openalex.org/C169760540","wikidata":"https://www.wikidata.org/wiki/Q207011","display_name":"Neuroscience","level":1,"score":0.25130000710487366},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.2508000135421753}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.19058","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19058","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.2602.19058","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19058","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":[{"id":"https://metadata.un.org/sdg/16","display_name":"Peace, Justice and strong institutions","score":0.7402652502059937}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Large":[0],"vision-language":[1],"models":[2,19],"(LVLMs)":[3],"have":[4],"rapidly":[5],"advanced":[6],"across":[7],"various":[8],"domains,":[9],"yet":[10],"they":[11],"still":[12],"lag":[13],"behind":[14],"strong":[15],"text-only":[16],"large":[17,59],"language":[18],"(LLMs)":[20],"on":[21,44,109],"tasks":[22],"that":[23,92,122,191],"require":[24],"multi-step":[25,69],"inference":[26,70,81,125,160,182,206],"and":[27,76,98,147,166,175,200],"compositional":[28],"decision-making.":[29],"Motivated":[30],"by":[31],"their":[32,103],"shared":[33,72,94,137,192],"transformer":[34],"architectures,":[35],"we":[36,55,89,112],"investigate":[37],"whether":[38],"the":[39,52,65,153],"two":[40],"model":[41],"families":[42],"rely":[43],"common":[45],"internal":[46],"computation":[47],"for":[48],"such":[49],"inference.":[50,107],"At":[51],"neuron":[53],"level,":[54],"uncover":[56],"a":[57,79,119,140],"surprisingly":[58],"overlap:":[60],"more":[61],"than":[62],"half":[63],"of":[64,143,205],"top-activated":[66],"units":[67],"during":[68],"are":[71],"between":[73,198],"representative":[74],"LLMs":[75,128,199],"LVLMs,":[77,201],"revealing":[78],"modality-invariant":[80],"subspace.":[82,155],"Through":[83],"causal":[84],"probing":[85],"via":[86],"activation":[87],"amplification,":[88],"further":[90],"show":[91],"these":[93,149],"neurons":[95,193],"encode":[96],"consistent":[97],"interpretable":[99,196],"concept-level":[100],"effects,":[101],"demonstrating":[102],"functional":[104],"contribution":[105],"to":[106,129,135],"Building":[108],"this":[110],"insight,":[111],"propose":[113],"Shared":[114],"Neuron":[115],"Low-Rank":[116],"Fusion":[117],"(SNRF),":[118],"parameter-efficient":[120],"framework":[121],"transfers":[123],"mature":[124],"circuitry":[126],"from":[127],"LVLMs.":[130],"SNRF":[131,178],"profiles":[132],"cross-model":[133],"activations":[134],"identify":[136],"neurons,":[138],"computes":[139],"low-rank":[141],"approximation":[142],"inter-model":[144],"weight":[145],"differences,":[146],"injects":[148],"updates":[150],"selectively":[151],"within":[152],"shared-neuron":[154],"This":[156],"mechanism":[157],"strengthens":[158],"multimodal":[159,170,209],"performance":[161,183],"with":[162],"minimal":[163],"parameter":[164],"changes":[165],"requires":[167],"no":[168],"large-scale":[169],"fine-tuning.":[171],"Across":[172],"diverse":[173],"mathematics":[174],"perception":[176],"benchmarks,":[177],"consistently":[179],"enhances":[180],"LVLM":[181],"while":[184],"preserving":[185],"perceptual":[186],"capabilities.":[187],"Our":[188,211],"results":[189],"demonstrate":[190],"form":[194],"an":[195],"bridge":[197],"enabling":[202],"low-cost":[203],"transfer":[204],"ability":[207],"into":[208],"models.":[210],"code":[212],"is":[213],"available":[214],"at":[215],"[https://github.com/chenhangcuisg-code/Do-LLMs-VLMs-Share-Neurons](https://github.com/chenhangcuisg-code/Do-LLMs-VLMs-Share-Neurons).":[216]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-02-26T00:00:00"}