{"id":"https://openalex.org/W7154406401","doi":"https://doi.org/10.48550/arxiv.2604.11802","title":"Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?","display_name":"Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?","publication_year":2026,"publication_date":"2026-04-13","ids":{"openalex":"https://openalex.org/W7154406401","doi":"https://doi.org/10.48550/arxiv.2604.11802"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2604.11802","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2604.11802","pdf_url":"https://arxiv.org/pdf/2604.11802","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/2604.11802","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119358191","display_name":"Yuto Harada","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Harada, Yuto","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5002910355","display_name":"Hiro Taiyo Hamada","orcid":"https://orcid.org/0000-0002-7919-9125","observed_orcids":["https://orcid.org/0000-0002-7919-9125"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hamada, Hiro Taiyo","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":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":0.9229000210762024,"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":0.9229000210762024,"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":0.0430000014603138,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.006000000052154064,"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/big-five-personality-traits","display_name":"Big Five personality traits","score":0.8669999837875366},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8019999861717224},{"id":"https://openalex.org/keywords/probing","display_name":"probing","score":0.44200000166893005},{"id":"https://openalex.org/keywords/personality-prediction","display_name":"personality prediction","score":0.36500000953674316}],"concepts":[{"id":"https://openalex.org/C27415008","wikidata":"https://www.wikidata.org/wiki/Q7256382","display_name":"Psychological intervention","level":2,"score":0.6636999845504761},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.574999988079071},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.5665000081062317},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.5210000276565552},{"id":"https://openalex.org/C192209626","wikidata":"https://www.wikidata.org/wiki/Q190909","display_name":"Focus (optics)","level":2,"score":0.47780001163482666},{"id":"https://openalex.org/C187288502","wikidata":"https://www.wikidata.org/wiki/Q641118","display_name":"Personality","level":2,"score":0.427700012922287},{"id":"https://openalex.org/C75684735","wikidata":"https://www.wikidata.org/wiki/Q858810","display_name":"Big data","level":2,"score":0.38830000162124634},{"id":"https://openalex.org/C2865642","wikidata":"https://www.wikidata.org/wiki/Q378132","display_name":"Big Five personality traits","level":3,"score":0.3747999966144562},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3434999883174896},{"id":"https://openalex.org/C2780665704","wikidata":"https://www.wikidata.org/wiki/Q959298","display_name":"Intervention (counseling)","level":2,"score":0.3156000077724457},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3012000024318695},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.29649999737739563},{"id":"https://openalex.org/C77805123","wikidata":"https://www.wikidata.org/wiki/Q161272","display_name":"Social psychology","level":1,"score":0.28790000081062317},{"id":"https://openalex.org/C2776035688","wikidata":"https://www.wikidata.org/wiki/Q1606558","display_name":"Affect (linguistics)","level":2,"score":0.2802000045776367},{"id":"https://openalex.org/C32362775","wikidata":"https://www.wikidata.org/wiki/Q1405524","display_name":"Self-control","level":2,"score":0.2752000093460083}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2604.11802","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2604.11802","pdf_url":"https://arxiv.org/pdf/2604.11802","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.2604.11802","doi":"https://doi.org/10.48550/arxiv.2604.11802","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.11802","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2604.11802","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2604.11802","pdf_url":"https://arxiv.org/pdf/2604.11802","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":[{"id":"https://openalex.org/G5725925787","display_name":null,"funder_award_id":"JPMJMS2295","funder_id":"https://openalex.org/F4320338247","funder_display_name":"Moonshot Research and Development Program"}],"funders":[{"id":"https://openalex.org/F4320338247","display_name":"Moonshot Research and Development Program","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7154406401.pdf","grobid_xml":"https://content.openalex.org/works/W7154406401.grobid-xml"},"referenced_works_count":9,"referenced_works":["https://openalex.org/W2075134121","https://openalex.org/W2786672974","https://openalex.org/W3024936740","https://openalex.org/W4248641702","https://openalex.org/W4387561528","https://openalex.org/W4401306886","https://openalex.org/W4405461402","https://openalex.org/W4405655184","https://openalex.org/W4412945505"],"related_works":["https://openalex.org/W7116864210","https://openalex.org/W4416021888","https://openalex.org/W7163995337","https://openalex.org/W4416026439","https://openalex.org/W7159652104","https://openalex.org/W7167268364","https://openalex.org/W4415193054","https://openalex.org/W4416019581","https://openalex.org/W4404390850","https://openalex.org/W7213623260"],"abstract_inverted_index":{"Using":[0],"psychological":[1],"constructs":[2],"such":[3],"as":[4],"the":[5,41,62,145,185,199,202,211,215],"Big":[6,58,92,108,132,190],"Five,":[7],"large":[8,241],"language":[9],"models":[10],"(LLMs)":[11],"can":[12,24,119,194],"imitate":[13],"specific":[14],"personality":[15,192],"profiles":[16],"and":[17,35,43,64,70,111,123,141,157,221,255],"predict":[18],"a":[19,240,250],"user's":[20],"personality.":[21],"While":[22],"LLMs":[23],"exhibit":[25,158],"behaviors":[26],"consistent":[27],"with":[28,174,237],"these":[29,76,165],"constructs,":[30],"it":[31],"remains":[32,142],"unclear":[33],"where":[34,91],"how":[36,44,75],"they":[37,45],"are":[38,151,217],"represented":[39],"inside":[40],"model":[42,97],"relate":[46,78],"to":[47,73,79,89,106],"behavioral":[48,80,256],"outputs.":[49,81],"To":[50],"address":[51],"this":[52],"gap,":[53],"we":[54,85],"focus":[55],"on":[56,164,239],"questionnaire-operationalized":[57],"Five":[59,93,109,133,191],"concepts,":[60,173,182],"analyze":[61],"formation":[63],"localization":[65],"of":[66,189,243],"their":[67,117],"internal":[68,187],"representations,":[69],"use":[71,87],"interventions":[72,204,238],"examine":[74,90],"representations":[77,122],"In":[82],"our":[83,247],"experiment,":[84],"first":[86],"probing":[88],"information":[94,134],"emerges":[95],"across":[96,161],"depth.":[98],"We":[99,129],"then":[100],"identify":[101],"neurons":[102,150,166],"that":[103,131,184,228],"respond":[104],"selectively":[105],"each":[107],"concept":[110],"test":[112],"whether":[113],"enhancing":[114],"or":[115],"suppressing":[116],"activations":[118],"bias":[120,206],"latent":[121],"label":[124,208],"generation":[125],"in":[126,138,154,210,258],"intended":[127,212],"directions.":[128],"find":[130],"becomes":[135],"rapidly":[136],"decodable":[137],"early":[139],"layers":[140,156],"detectable":[143],"through":[144],"final":[146],"layers,":[147],"while":[148],"concept-selective":[149,244],"most":[152],"prevalent":[153],"mid":[155],"limited":[159],"overlap":[160],"domains.":[162],"Interventions":[163],"consistently":[167],"shift":[168],"probe":[169],"readouts":[170],"toward":[171],"targeted":[172,175],"success":[176],"rates":[177],"exceeding":[178],"0.8":[179],"for":[180],"some":[181],"indicating":[183,227],"model's":[186],"separation":[188],"traits":[193],"be":[195],"causally":[196],"steered.":[197],"At":[198],"label-generation":[200],"level,":[201],"same":[203],"often":[205,222],"generated":[207,232],"distributions":[209],"directions,":[213],"but":[214],"effects":[216],"weaker,":[218],"more":[219],"concept-dependent,":[220],"accompanied":[223],"by":[224],"cross-trait":[225],"spillover,":[226],"comparable":[229],"control":[230,254,257],"over":[231],"labels":[233],"is":[234],"difficult":[235],"even":[236],"fraction":[242],"neurons.":[245],"Overall,":[246],"findings":[248],"reveal":[249],"gap":[251],"between":[252],"representational":[253],"LLMs.":[259]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-04-15T00:00:00"}