{"id":"https://openalex.org/W7134072488","doi":"https://doi.org/10.48550/arxiv.2603.05280","title":"Layer by layer, module by module: Choose both for optimal OOD probing of ViT","display_name":"Layer by layer, module by module: Choose both for optimal OOD probing of ViT","publication_year":2026,"publication_date":"2026-03-05","ids":{"openalex":"https://openalex.org/W7134072488","doi":"https://doi.org/10.48550/arxiv.2603.05280"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.05280","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05280","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.2603.05280","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5093123204","display_name":"Ambroise Odonnat","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Odonnat, Ambroise","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5036354530","display_name":"Vasilii Feofanov","orcid":"https://orcid.org/0000-0002-5777-4205"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feofanov, Vasilii","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5125412812","display_name":"Laetitia Chapel","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chapel, Laetitia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5067296471","display_name":"Romain Tavenard","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tavenard, Romain","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5125390829","display_name":"Ievgen Redko","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Redko, Ievgen","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/T11128","display_name":"Transition Metal Oxide Nanomaterials","score":0.20080000162124634,"subfield":{"id":"https://openalex.org/subfields/2507","display_name":"Polymers and Plastics"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11128","display_name":"Transition Metal Oxide Nanomaterials","score":0.20080000162124634,"subfield":{"id":"https://openalex.org/subfields/2507","display_name":"Polymers and Plastics"},"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/T10623","display_name":"Thin-Film Transistor Technologies","score":0.1590999960899353,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.1500999927520752,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/discriminative-model","display_name":"Discriminative model","score":0.8511000275611877},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.5564000010490417},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.5149000287055969},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.5076000094413757},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.475600004196167},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.37310001254081726},{"id":"https://openalex.org/keywords/feed-forward","display_name":"Feed forward","score":0.359499990940094}],"concepts":[{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.8511000275611877},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5564000010490417},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.54830002784729},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5149000287055969},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5076000094413757},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.475600004196167},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4697999954223633},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.37310001254081726},{"id":"https://openalex.org/C38858127","wikidata":"https://www.wikidata.org/wiki/Q5441228","display_name":"Feed forward","level":2,"score":0.359499990940094},{"id":"https://openalex.org/C2779227376","wikidata":"https://www.wikidata.org/wiki/Q6505497","display_name":"Layer (electronics)","level":2,"score":0.35850000381469727},{"id":"https://openalex.org/C689293","wikidata":"https://www.wikidata.org/wiki/Q751484","display_name":"Block design","level":2,"score":0.3287999927997589},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3050000071525574},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.28349998593330383},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.27799999713897705},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.2669999897480011},{"id":"https://openalex.org/C136389625","wikidata":"https://www.wikidata.org/wiki/Q334384","display_name":"Supervised learning","level":3,"score":0.2599000036716461},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.25839999318122864},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2574999928474426},{"id":"https://openalex.org/C24326235","wikidata":"https://www.wikidata.org/wiki/Q126095","display_name":"Electronic engineering","level":1,"score":0.25679999589920044},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2567000091075897},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.05280","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05280","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.2603.05280","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.05280","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":[{"score":0.7602128386497498,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0],"studies":[1],"have":[2],"observed":[3],"that":[4,74,105],"intermediate":[5,53],"layers":[6,54],"of":[7,52,68,86,108,134],"foundation":[8],"models":[9,32],"often":[10],"yield":[11],"more":[12],"discriminative":[13,37],"representations":[14],"than":[15],"the":[16,50,83,99,116,119,123,131,135,142],"final":[17],"layer.":[18],"While":[19],"initially":[20],"attributed":[21],"to":[22,48],"autoregressive":[23],"pretraining,":[24],"this":[25,41],"phenomenon":[26],"has":[27],"also":[28],"been":[29],"identified":[30],"in":[31,55,89],"trained":[33],"via":[34],"supervised":[35],"and":[36,79],"self-supervised":[38],"objectives.":[39],"In":[40],"paper,":[42],"we":[43,72,93],"conduct":[44],"a":[45,65,95],"comprehensive":[46],"study":[47],"analyze":[49],"behavior":[51],"pretrained":[56],"vision":[57],"transformers.":[58],"Through":[59],"extensive":[60],"linear":[61],"probing":[62,107,115],"experiments":[63],"across":[64],"diverse":[66],"set":[67],"image":[69],"classification":[70],"benchmarks,":[71],"find":[73],"distribution":[75,128],"shift":[76,143],"between":[77],"pretraining":[78],"downstream":[80],"data":[81],"is":[82,112,139,144],"primary":[84],"cause":[85],"performance":[87,125],"degradation":[88],"deeper":[90],"layers.":[91],"Furthermore,":[92],"perform":[94],"fine-grained":[96],"analysis":[97],"at":[98],"module":[100,138],"level.":[101],"Our":[102],"findings":[103],"reveal":[104],"standard":[106],"transformer":[109],"block":[110],"outputs":[111],"suboptimal;":[113],"instead,":[114],"activation":[117],"within":[118],"feedforward":[120],"network":[121],"yields":[122],"best":[124],"under":[126],"significant":[127],"shift,":[129],"whereas":[130],"normalized":[132],"output":[133],"multi-head":[136],"self-attention":[137],"optimal":[140],"when":[141],"weak.":[145]},"counts_by_year":[],"updated_date":"2026-08-16T07:02:28.622633","created_date":"2026-03-07T00:00:00"}