{"id":"https://openalex.org/W7214896557","doi":"https://doi.org/10.48550/arxiv.2609.37991","title":"Which Attention Heads are like the Human Head? Not the Ones that Compute","display_name":"Which Attention Heads are like the Human Head? Not the Ones that Compute","publication_year":2026,"publication_date":"2026-09-29","ids":{"openalex":"https://openalex.org/W7214896557","doi":"https://doi.org/10.48550/arxiv.2609.37991"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.37991","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.37991","pdf_url":"https://arxiv.org/pdf/2609.37991","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":null,"license_id":null,"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/2609.37991","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5033648655","display_name":"Christopher Pinier","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I887064364","display_name":"University of Amsterdam","ror":"https://ror.org/04dkp9463","country_code":"NL","type":"education","lineage":["https://openalex.org/I887064364"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Pinier, Christopher","raw_affiliation_strings":["Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands","institution_ids":["https://openalex.org/I887064364"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127131058","display_name":"Gustaw Opie\u0142ka","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I887064364","display_name":"University of Amsterdam","ror":"https://ror.org/04dkp9463","country_code":"NL","type":"education","lineage":["https://openalex.org/I887064364"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Opie\u0142ka, Gustaw","raw_affiliation_strings":["Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands","institution_ids":["https://openalex.org/I887064364"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5071071057","display_name":"Hannes Rosenbusch","orcid":"https://orcid.org/0000-0002-4983-3615","observed_orcids":["https://orcid.org/0000-0002-4983-3615"]},"institutions":[{"id":"https://openalex.org/I887064364","display_name":"University of Amsterdam","ror":"https://ror.org/04dkp9463","country_code":"NL","type":"education","lineage":["https://openalex.org/I887064364"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Rosenbusch, Hannes","raw_affiliation_strings":["Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands","institution_ids":["https://openalex.org/I887064364"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5153583374","display_name":"Taylor Webb","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I20089843","display_name":"Princeton University","ror":"https://ror.org/00hx57361","country_code":"US","type":"education","lineage":["https://openalex.org/I20089843"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Webb, Taylor","raw_affiliation_strings":["Princeton Neuroscience Institute , Princeton University , Princeton , NJ , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Princeton Neuroscience Institute , Princeton University , Princeton , NJ , USA","institution_ids":["https://openalex.org/I20089843"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102805257","display_name":"Michael D. Nunez","orcid":"https://orcid.org/0000-0002-9965-6282","observed_orcids":["https://orcid.org/0000-0002-9965-6282"]},"institutions":[{"id":"https://openalex.org/I887064364","display_name":"University of Amsterdam","ror":"https://ror.org/04dkp9463","country_code":"NL","type":"education","lineage":["https://openalex.org/I887064364"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Nunez, Michael D.","raw_affiliation_strings":["Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands","institution_ids":["https://openalex.org/I887064364"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5021896449","display_name":"Claire E. Stevenson","orcid":"https://orcid.org/0000-0003-1797-9836","observed_orcids":["https://orcid.org/0000-0003-1797-9836"]},"institutions":[{"id":"https://openalex.org/I887064364","display_name":"University of Amsterdam","ror":"https://ror.org/04dkp9463","country_code":"NL","type":"education","lineage":["https://openalex.org/I887064364"]}],"countries":["NL"],"is_corresponding":false,"raw_author_name":"Stevenson, Claire E.","raw_affiliation_strings":["Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Psychological Methods , University of Amsterdam , Amsterdam , The Netherlands","institution_ids":["https://openalex.org/I887064364"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9990000128746033,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.9990000128746033,"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/T10028","display_name":"Topic Modeling","score":9.999999747378752e-05,"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/T11605","display_name":"Visual Attention and Saliency Detection","score":0.0,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/electroencephalography","display_name":"EEG","score":0.36399999260902405},{"id":"https://openalex.org/keywords/pattern-completion","display_name":"pattern completion","score":0.6660000085830688},{"id":"https://openalex.org/keywords/concept-vectors","display_name":"concept vectors","score":0.7770000100135803},{"id":"https://openalex.org/keywords/interpretability","display_name":"interpretability","score":0.5559999942779541},{"id":"https://openalex.org/keywords/salience","display_name":"salience","score":0.49799999594688416}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.8342999815940857},{"id":"https://openalex.org/C2778738651","wikidata":"https://www.wikidata.org/wiki/Q16546687","display_name":"Novelty","level":2,"score":0.6194999814033508},{"id":"https://openalex.org/C108154423","wikidata":"https://www.wikidata.org/wiki/Q1469792","display_name":"Salience (neuroscience)","level":2,"score":0.5834000110626221},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.531499981880188},{"id":"https://openalex.org/C142853389","wikidata":"https://www.wikidata.org/wiki/Q744778","display_name":"Association (psychology)","level":2,"score":0.5281999707221985},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5101000070571899},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.489300012588501},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.48510000109672546},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.3903999924659729},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.37880000472068787},{"id":"https://openalex.org/C166151441","wikidata":"https://www.wikidata.org/wiki/Q4923601","display_name":"Causation","level":2,"score":0.37369999289512634},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.3702999949455261},{"id":"https://openalex.org/C139676723","wikidata":"https://www.wikidata.org/wiki/Q1193832","display_name":"Sign (mathematics)","level":2,"score":0.35179999470710754},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.33559998869895935},{"id":"https://openalex.org/C66024118","wikidata":"https://www.wikidata.org/wiki/Q1122506","display_name":"Computational model","level":2,"score":0.3197999894618988},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.3075000047683716},{"id":"https://openalex.org/C556910895","wikidata":"https://www.wikidata.org/wiki/Q384188","display_name":"Magnetoencephalography","level":3,"score":0.30149999260902405},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.2955000102519989},{"id":"https://openalex.org/C2780012671","wikidata":"https://www.wikidata.org/wiki/Q1343870","display_name":"Corollary","level":2,"score":0.28859999775886536},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.2883000075817108},{"id":"https://openalex.org/C169900460","wikidata":"https://www.wikidata.org/wiki/Q2200417","display_name":"Cognition","level":2,"score":0.2851000130176544},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2732999920845032},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2703000009059906},{"id":"https://openalex.org/C46312422","wikidata":"https://www.wikidata.org/wiki/Q11024","display_name":"Communication","level":1,"score":0.25940001010894775}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.37991","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.37991","pdf_url":"https://arxiv.org/pdf/2609.37991","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.37991","doi":"https://doi.org/10.48550/arxiv.2609.37991","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.37991","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:2609.37991","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.37991","pdf_url":"https://arxiv.org/pdf/2609.37991","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214896557.pdf","grobid_xml":"https://content.openalex.org/works/W7214896557.grobid-xml"},"referenced_works_count":10,"referenced_works":["https://openalex.org/W2058616551","https://openalex.org/W2066261007","https://openalex.org/W2892147425","https://openalex.org/W3210923133","https://openalex.org/W4220949944","https://openalex.org/W4308341600","https://openalex.org/W4310572212","https://openalex.org/W4401306886","https://openalex.org/W4411119427","https://openalex.org/W6910721441"],"related_works":["https://openalex.org/W7204916316","https://openalex.org/W7216323833","https://openalex.org/W7155485083","https://openalex.org/W6959727613","https://openalex.org/W7214934955","https://openalex.org/W7161834385","https://openalex.org/W4386469863","https://openalex.org/W7102397576","https://openalex.org/W7161477493","https://openalex.org/W7161980384"],"abstract_inverted_index":{"Brain-AI":[0],"alignment":[1,113,204],"is":[2,25,65,167,196],"often":[3],"interpreted":[4],"as":[5],"a":[6],"sign":[7],"that":[8,82,160],"model":[9,23,209],"and":[10,44,54,101,154,181,213,222],"brain":[11],"perform":[12],"similar":[13],"computations.":[14],"Whether":[15],"the":[16,90,143,157,208,211,220,224],"aligned":[17],"units":[18],"are":[19,182],"causally":[20],"involved":[21],"in":[22],"computation":[24],"rarely":[26],"checked.":[27],"On":[28],"an":[29],"abstract":[30,97],"pattern-completion":[31],"task":[32,51],"(AAABAAA":[33],"$\\rightarrow$":[34],"B),":[35],"we":[36,132],"compare":[37,78],"LLM":[38],"attention-head":[39],"representations":[40],"with":[41,117,123,184],"human":[42],"EEG":[43],"test":[45],"how":[46,207,217],"ablating":[47],"those":[48],"heads":[49,58,72,176],"affects":[50],"performance.":[52],"Alignment":[53],"causation":[55],"dissociate:":[56],"brain-aligned":[57,130],"contribute":[59,177],"to":[60,89,109,156,179],"performance,":[61],"but":[62],"their":[63,107],"removal":[64,70,166,195],"substantially":[66,197],"less":[67,168],"disruptive":[68,199],"than":[69,170,200],"of":[71],"selected":[73,105],"via":[74],"attribution":[75],"patching.":[76],"We":[77],"two":[79],"head":[80],"sets":[81],"prior":[83],"interpretability":[84],"work":[85],"defines":[86],"without":[87],"reference":[88],"brain:":[91],"concept":[92],"vectors":[93,103],"(CVs),":[94],"which":[95],"represent":[96],"patterns":[98],"across":[99,127],"formats,":[100],"function":[102],"(FVs),":[104],"for":[106],"contribution":[108],"correct-answer":[110],"prediction.":[111],"Brain":[112,203],"shows":[114],"little":[115],"association":[116,122],"FV":[118],"scores,":[119],"while":[120],"its":[121,165],"CV":[124],"scores":[125],"varies":[126],"models.":[128],"Among":[129],"heads,":[131],"find":[133],"recurring":[134],"attention":[135],"profiles:":[136],"one":[137],"emphasizes":[138],"distinctive":[139],"elements":[140,146,159],"(novelty":[141],"heads),":[142],"other":[144],"repeating":[145],"(repetition":[147],"heads).":[148],"The":[149],"novelty":[150],"family":[151],"tracks":[152],"salience":[153],"attends":[155],"same":[158],"humans":[161],"look":[162],"at,":[163],"yet":[164],"damaging":[169],"random":[171],"ablation":[172],"on":[173],"average.":[174],"Repetition":[175],"modestly":[178],"performance":[180],"associated":[183],"abstract-pattern":[185],"representation":[186],"(CVs).":[187],"Across":[188],"17":[189],"models":[190],"spanning":[191],"3B-72B":[192],"parameters,":[193],"FV-ranked":[194],"more":[198],"brain-ranked":[201],"removal.":[202],"thus":[205],"captures":[206,216],"reads":[210],"stimulus,":[212],"only":[214],"faintly":[215],"it":[218],"represents":[219],"pattern":[221],"solves":[223],"task.":[225]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-01T00:00:00"}
