{"id":"https://openalex.org/W7126395107","doi":"https://doi.org/10.18653/v1/2024.findings-eacl.58","title":"The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Models","display_name":"The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Models","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W7126395107","doi":"https://doi.org/10.18653/v1/2024.findings-eacl.58"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2024.findings-eacl.58","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-eacl.58","pdf_url":"https://aclanthology.org/2024.findings-eacl.58.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EACL 2024","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2024.findings-eacl.58.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5124622240","display_name":"Anton Razzhigaev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Anton Razzhigaev","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093257146","display_name":"Matvey Mikhalchuk","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Matvey Mikhalchuk","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124544916","display_name":"Elizaveta Goncharova","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Elizaveta Goncharova","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124679531","display_name":"Ivan Oseledets","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ivan Oseledets","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5124552210","display_name":"Denis Dimitrov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Denis Dimitrov","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5124602109","display_name":"Andrey Kuznetsov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andrey Kuznetsov","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.3055,"has_fulltext":true,"cited_by_count":1,"citation_normalized_percentile":{"value":0.70252033,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":95,"max":98},"biblio":{"volume":null,"issue":null,"first_page":"868","last_page":"874"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.033399999141693115,"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/T10320","display_name":"Neural Networks and Applications","score":0.033399999141693115,"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/T11883","display_name":"Embodied and Extended Cognition","score":0.027300000190734863,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}},{"id":"https://openalex.org/T10427","display_name":"Visual perception and processing mechanisms","score":0.02329999953508377,"subfield":{"id":"https://openalex.org/subfields/2805","display_name":"Cognitive Neuroscience"},"field":{"id":"https://openalex.org/fields/28","display_name":"Neuroscience"},"domain":{"id":"https://openalex.org/domains/1","display_name":"Life Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/anisotropy","display_name":"Anisotropy","score":0.4713999927043915},{"id":"https://openalex.org/keywords/field","display_name":"Field (mathematics)","score":0.29989999532699585},{"id":"https://openalex.org/keywords/dimension","display_name":"Dimension (graph theory)","score":0.2992999851703644},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.2888000011444092},{"id":"https://openalex.org/keywords/mathematical-model","display_name":"Mathematical model","score":0.2272000014781952}],"concepts":[{"id":"https://openalex.org/C85725439","wikidata":"https://www.wikidata.org/wiki/Q466686","display_name":"Anisotropy","level":2,"score":0.4713999927043915},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.46619999408721924},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.29989999532699585},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.2992999851703644},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2888000011444092},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26930001378059387},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.24529999494552612},{"id":"https://openalex.org/C76969082","wikidata":"https://www.wikidata.org/wiki/Q486902","display_name":"Mathematical model","level":2,"score":0.2272000014781952},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.2222999930381775},{"id":"https://openalex.org/C198352243","wikidata":"https://www.wikidata.org/wiki/Q37105","display_name":"Line (geometry)","level":2,"score":0.22110000252723694}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2024.findings-eacl.58","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-eacl.58","pdf_url":"https://aclanthology.org/2024.findings-eacl.58.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EACL 2024","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2024.findings-eacl.58","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-eacl.58","pdf_url":"https://aclanthology.org/2024.findings-eacl.58.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EACL 2024","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7126395107.pdf","grobid_xml":"https://content.openalex.org/works/W7126395107.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"In":[0],"this":[1],"study,":[2],"we":[3,61],"present":[4],"an":[5,77],"investigation":[6],"into":[7,79,102],"the":[8,21,41,46,52,64,71,80,91,94,111],"anisotropy":[9,30,43,56],"dynamics":[10],"and":[11,25,115],"intrinsic":[12,65],"dimension":[13,66],"of":[14,67,74,93,113],"embeddings":[15,68],"in":[16,32,45,58],"transformer":[17,33],"architectures,":[18],"focusing":[19],"on":[20,110],"dichotomy":[22],"between":[23],"encoders":[24,114],"decoders.Our":[26],"findings":[27],"reveal":[28],"that":[29,63],"profile":[31],"decoders":[34,116],"exhibits":[35],"a":[36,87,100],"distinct":[37],"bell-shaped":[38],"curve,":[39],"with":[40,96],"highest":[42],"concentrations":[44],"middle":[47],"layers.This":[48],"pattern":[49],"diverges":[50],"from":[51],"more":[53,103],"uniformly":[54],"distributed":[55],"observed":[57],"encoders.In":[59],"addition,":[60],"found":[62],"increases":[69],"during":[70],"initial":[72],"phases":[73],"training,":[75],"indicating":[76],"expansion":[78],"higher-dimensional":[81],"space.Which":[82],"is":[83],"then":[84],"followed":[85],"by":[86],"compression":[88],"phase":[89],"towards":[90],"end":[92],"training":[95],"dimensionality":[97],"decrease,":[98],"suggesting":[99],"refinement":[101],"compact":[104],"representations.Our":[105],"results":[106],"provide":[107],"fresh":[108],"insights":[109],"understanding":[112],"embedding":[117],"properties.":[118]},"counts_by_year":[{"year":2026,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2026-02-02T00:00:00"}