{"id":"https://openalex.org/W7165155381","doi":"https://doi.org/10.48550/arxiv.2606.19268","title":"Patnaik-Pearson intrinsic dimension for internal representations of neural networks","display_name":"Patnaik-Pearson intrinsic dimension for internal representations of neural networks","publication_year":2026,"publication_date":"2026-06-17","ids":{"openalex":"https://openalex.org/W7165155381","doi":"https://doi.org/10.48550/arxiv.2606.19268"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2606.19268","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2606.19268","pdf_url":"https://arxiv.org/pdf/2606.19268","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2606.19268","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5019713454","display_name":"Tom Hadfield","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I47508984","display_name":"Imperial College London","ror":"https://ror.org/041kmwe10","country_code":"GB","type":"education","lineage":["https://openalex.org/I47508984"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Hadfield, Tom","raw_affiliation_strings":["Department of Mathematics , Imperial College London 180 Queen's Gate , London SW7 5HF , United Kingdom"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Mathematics , Imperial College London 180 Queen's Gate , London SW7 5HF , United Kingdom","institution_ids":["https://openalex.org/I47508984"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":["https://openalex.org/A5019713454"],"corresponding_institution_ids":["https://openalex.org/I47508984"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"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/T10320","display_name":"Neural Networks and Applications","score":0.4043000042438507,"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":0.4043000042438507,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.13899999856948853,"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/T12536","display_name":"Topological and Geometric Data Analysis","score":0.07989999651908875,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/intrinsic-dimension","display_name":"intrinsic dimension","score":0.9300000071525574},{"id":"https://openalex.org/keywords/data-manifold","display_name":"data manifold","score":0.8880000114440918},{"id":"https://openalex.org/keywords/transformer","display_name":"transformer models","score":0.8960000276565552},{"id":"https://openalex.org/keywords/weighting-matrix","display_name":"weight matrix","score":0.5870000123977661},{"id":"https://openalex.org/keywords/pareto-distribution","display_name":"Pareto distribution","score":0.8870000243186951},{"id":"https://openalex.org/keywords/tail-exponent","display_name":"tail exponent","score":0.8199999928474426},{"id":"https://openalex.org/keywords/free-probability","display_name":"free probability","score":0.7720000147819519},{"id":"https://openalex.org/keywords/token-embeddings","display_name":"token embeddings","score":0.8759999871253967},{"id":"https://openalex.org/keywords/bert","display_name":"BERT","score":0.8259999752044678}],"concepts":[{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.8019000291824341},{"id":"https://openalex.org/C30732413","wikidata":"https://www.wikidata.org/wiki/Q17092636","display_name":"Intrinsic dimension","level":3,"score":0.765500009059906},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6449000239372253},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.5947999954223633},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.49050000309944153},{"id":"https://openalex.org/C119322782","wikidata":"https://www.wikidata.org/wiki/Q2662236","display_name":"VC dimension","level":2,"score":0.4546999931335449},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.4478999972343445},{"id":"https://openalex.org/C115311070","wikidata":"https://www.wikidata.org/wiki/Q5347255","display_name":"Effective dimension","level":3,"score":0.429500013589859},{"id":"https://openalex.org/C529865628","wikidata":"https://www.wikidata.org/wiki/Q1790740","display_name":"Manifold (fluid mechanics)","level":2,"score":0.3977000117301941},{"id":"https://openalex.org/C2780388253","wikidata":"https://www.wikidata.org/wiki/Q5421508","display_name":"Exponent","level":2,"score":0.37560001015663147},{"id":"https://openalex.org/C168185519","wikidata":"https://www.wikidata.org/wiki/Q5156577","display_name":"Complex dimension","level":2,"score":0.36160001158714294},{"id":"https://openalex.org/C184720557","wikidata":"https://www.wikidata.org/wiki/Q7825049","display_name":"Topology (electrical circuits)","level":2,"score":0.35760000348091125},{"id":"https://openalex.org/C110601934","wikidata":"https://www.wikidata.org/wiki/Q5172849","display_name":"Correlation dimension","level":4,"score":0.3407000005245209},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.3353999853134155},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.33390000462532043},{"id":"https://openalex.org/C26546657","wikidata":"https://www.wikidata.org/wiki/Q1412452","display_name":"Fractal dimension","level":3,"score":0.3206000030040741},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3188999891281128},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.30320000648498535},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.28859999775886536},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28769999742507935},{"id":"https://openalex.org/C202444582","wikidata":"https://www.wikidata.org/wiki/Q837863","display_name":"Pure mathematics","level":1,"score":0.2833999991416931},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.27889999747276306},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.26269999146461487},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C121864883","wikidata":"https://www.wikidata.org/wiki/Q677916","display_name":"Statistical physics","level":1,"score":0.26089999079704285},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.25760000944137573},{"id":"https://openalex.org/C95341145","wikidata":"https://www.wikidata.org/wiki/Q5277265","display_name":"Dimension theory (algebra)","level":3,"score":0.2551000118255615},{"id":"https://openalex.org/C168110828","wikidata":"https://www.wikidata.org/wiki/Q1331626","display_name":"Spectral density","level":2,"score":0.2531000077724457}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2606.19268","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2606.19268","pdf_url":"https://arxiv.org/pdf/2606.19268","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2606.19268","doi":"https://doi.org/10.48550/arxiv.2606.19268","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.19268","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,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2606.19268","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2606.19268","pdf_url":"https://arxiv.org/pdf/2606.19268","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7165155381.pdf","grobid_xml":"https://content.openalex.org/works/W7165155381.grobid-xml"},"referenced_works_count":27,"referenced_works":["https://openalex.org/W1593775104","https://openalex.org/W2016646427","https://openalex.org/W2047627004","https://openalex.org/W2316416340","https://openalex.org/W2626778328","https://openalex.org/W2754478492","https://openalex.org/W2896457183","https://openalex.org/W2907762094","https://openalex.org/W2946857231","https://openalex.org/W3039502308","https://openalex.org/W4224913930","https://openalex.org/W4226434736","https://openalex.org/W4293138820","https://openalex.org/W4398187977","https://openalex.org/W4399555118","https://openalex.org/W4403573173","https://openalex.org/W4404976256","https://openalex.org/W4407093434","https://openalex.org/W4411881822","https://openalex.org/W4415035773","https://openalex.org/W4415087802","https://openalex.org/W4416075919","https://openalex.org/W4416968619","https://openalex.org/W6892060006","https://openalex.org/W7101386660","https://openalex.org/W7104280109","https://openalex.org/W7135050644"],"related_works":["https://openalex.org/W2946857231","https://openalex.org/W7160329045","https://openalex.org/W7155410031","https://openalex.org/W2007023969","https://openalex.org/W7103752933","https://openalex.org/W2947024452","https://openalex.org/W4400335910","https://openalex.org/W3019142305","https://openalex.org/W1489992831","https://openalex.org/W7133237026"],"abstract_inverted_index":{"We":[0,57,144],"define":[1],"a":[2,9,83,114,134],"new":[3],"measure":[4],"of":[5,8,24,41,53,61,69,104,116,129,133,160,165,172,183],"intrinsic":[6,50,63],"dimension":[7,51,64,92,132,159,175],"data":[10,73,135,163],"manifold,":[11],"which":[12],"we":[13,88,125],"call":[14],"the":[15,36,48,90,94,105,110,127,130,138,149,157,161,170,173,181,184,191],"Patnaik-Pearson":[16,91,131,158,174],"dimension,":[17],"and":[18,38,44,96,99,118,151,168,187],"apply":[19,145],"this":[20,33,62,146],"to":[21,93,141,148,154],"internal":[22],"representations":[23],"neural":[25,70,142],"networks,":[26],"in":[27],"particular":[28],"transformers.":[29],"The":[30],"inspiration":[31],"for":[32,75,109,190],"comes":[34],"from":[35],"HTSR":[37,95],"SETOL":[39,97],"work":[40],"Martin,":[42],"Mahoney":[43],"Hinrichs,":[45],"combined":[46],"with":[47],"TwoNN":[49],"estimator":[52],"Facco":[54],"et":[55],"al.":[56],"prove":[58],"various":[59],"properties":[60],"estimator.":[65],"Treating":[66],"weight":[67,76],"matrices":[68,77],"networks":[71],"as":[72,176],"manifolds,":[74],"whose":[78],"empirical":[79],"spectral":[80],"density":[81],"follows":[82],"Pareto":[84],"(power":[85],"law)":[86],"distribution,":[87],"relate":[89],"analysis,":[98],"show":[100],"that":[101],"critical":[102],"values":[103],"tail":[106],"exponent":[107],"coincide":[108],"two":[111],"approaches.":[112],"Using":[113],"combination":[115],"theoretical":[117],"numerical":[119,192],"techniques,":[120],"including":[121],"Voiculescu's":[122],"free":[123],"probability,":[124],"study":[126],"behaviour":[128],"manifold":[136,164],"under":[137],"transformations":[139],"typical":[140],"networks.":[143],"machinery":[147],"BERT-base":[150],"DeepSeek-R1-Distill-Qwen-1":[152],"models,":[153],"investigate":[155],"first":[156],"initial":[162],"token":[166,177],"embeddings,":[167],"second":[169],"evolution":[171],"embeddings":[178],"pass":[179],"through":[180],"layers":[182],"model.":[185],"Code":[186],"notebooks":[188],"used":[189],"results":[193],"presented":[194],"here":[195],"is":[196],"available":[197],"at":[198],"https://github.com/tdhadfield/PatnaikPearson":[199]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-06-19T00:00:00"}