{"id":"https://openalex.org/W7212029997","doi":"https://doi.org/10.48550/arxiv.2609.08615","title":"Why shared attention vectors fail: a case for outcome-indexed tuning","display_name":"Why shared attention vectors fail: a case for outcome-indexed tuning","publication_year":2026,"publication_date":"2026-09-08","ids":{"openalex":"https://openalex.org/W7212029997","doi":"https://doi.org/10.48550/arxiv.2609.08615"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.08615","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.08615","pdf_url":"https://arxiv.org/pdf/2609.08615","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/2609.08615","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5145960667","display_name":"Lenard Dome","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4405258522","display_name":"Deutsches Zentrum f\u00fcr Psychische Gesundheit","ror":"https://ror.org/00tkfw097","country_code":"DE","type":"nonprofit","lineage":["https://openalex.org/I4405256427","https://openalex.org/I4405258522"]},{"id":"https://openalex.org/I8087733","display_name":"University of T\u00fcbingen","ror":"https://ror.org/03a1kwz48","country_code":"DE","type":"education","lineage":["https://openalex.org/I8087733"]}],"countries":["DE"],"is_corresponding":true,"raw_author_name":"Dome, Lenard","raw_affiliation_strings":["Department of Psychiatry and Psychotherapy , Faculty of Medicine University of T\u00fcbingen , T\u00fcbingen , Germany","German Center for Mental Health (DZPG) , T\u00fcbingen , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Psychiatry and Psychotherapy , Faculty of Medicine University of T\u00fcbingen , T\u00fcbingen , Germany","institution_ids":["https://openalex.org/I8087733"]},{"raw_affiliation_string":"German Center for Mental Health (DZPG) , T\u00fcbingen , Germany","institution_ids":["https://openalex.org/I4405258522"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5145960667"],"corresponding_institution_ids":["https://openalex.org/I4405258522","https://openalex.org/I8087733"],"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.3131999969482422,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.3131999969482422,"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/T10320","display_name":"Neural Networks and Applications","score":0.179299995303154,"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":0.16249999403953552,"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/gradient-based-learning","display_name":"gradient-based learning","score":0.5659999847412109},{"id":"https://openalex.org/keywords/attentional-stability","display_name":"attention stability","score":0.42399999499320984},{"id":"https://openalex.org/keywords/generalization","display_name":"generalization","score":0.5099999904632568},{"id":"https://openalex.org/keywords/predictive-features","display_name":"predictive features","score":0.3009999990463257}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6383000016212463},{"id":"https://openalex.org/C33676613","wikidata":"https://www.wikidata.org/wiki/Q13415176","display_name":"Dimension (graph theory)","level":2,"score":0.5097000002861023},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.503600001335144},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.41510000824928284},{"id":"https://openalex.org/C28006648","wikidata":"https://www.wikidata.org/wiki/Q6934509","display_name":"Multi-task learning","level":3,"score":0.39089998602867126},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.38830000162124634},{"id":"https://openalex.org/C40506919","wikidata":"https://www.wikidata.org/wiki/Q7452469","display_name":"Sequence learning","level":2,"score":0.3831999897956848},{"id":"https://openalex.org/C2993807640","wikidata":"https://www.wikidata.org/wiki/Q103709453","display_name":"Attention network","level":2,"score":0.3765000104904175},{"id":"https://openalex.org/C2779918689","wikidata":"https://www.wikidata.org/wiki/Q3771842","display_name":"Stimulus (psychology)","level":2,"score":0.3709000051021576},{"id":"https://openalex.org/C106487976","wikidata":"https://www.wikidata.org/wiki/Q685816","display_name":"Matrix (chemical analysis)","level":2,"score":0.33660000562667847},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3192000091075897},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.31450000405311584},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.2996000051498413},{"id":"https://openalex.org/C2989146674","wikidata":"https://www.wikidata.org/wiki/Q4818236","display_name":"Selective attention","level":3,"score":0.2822999954223633},{"id":"https://openalex.org/C2777379011","wikidata":"https://www.wikidata.org/wiki/Q938545","display_name":"Implicit learning","level":3,"score":0.26350000500679016},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.2533000111579895}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.08615","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.08615","pdf_url":"https://arxiv.org/pdf/2609.08615","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.2609.08615","doi":"https://doi.org/10.48550/arxiv.2609.08615","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.08615","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:2609.08615","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.08615","pdf_url":"https://arxiv.org/pdf/2609.08615","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/W7212029997.pdf","grobid_xml":"https://content.openalex.org/works/W7212029997.grobid-xml"},"referenced_works_count":13,"referenced_works":["https://openalex.org/W134309601","https://openalex.org/W1967011375","https://openalex.org/W2060535027","https://openalex.org/W2110278466","https://openalex.org/W2494089081","https://openalex.org/W2951896810","https://openalex.org/W2964608093","https://openalex.org/W3139392167","https://openalex.org/W4239618397","https://openalex.org/W4300402905","https://openalex.org/W4310875024","https://openalex.org/W4409332354","https://openalex.org/W4414608904"],"related_works":["https://openalex.org/W7208771770","https://openalex.org/W6962848750","https://openalex.org/W4414897147","https://openalex.org/W3199107222","https://openalex.org/W4387171255","https://openalex.org/W7117651376","https://openalex.org/W2859017189","https://openalex.org/W1587995223","https://openalex.org/W4399596531","https://openalex.org/W4406890777"],"abstract_inverted_index":{"Dimensional":[0],"attention":[1,11,117,129],"in":[2],"learning":[3,66,71,153],"is":[4],"often":[5],"implemented":[6],"as":[7],"a":[8,19,142],"globally":[9,85],"shared":[10,52,86,100,128],"vector,":[12],"where":[13,32,44],"each":[14],"stimulus":[15],"dimension":[16],"corresponds":[17],"to":[18,58,124,132],"single":[20],"scalar.":[21],"These":[22,134],"scalars":[23],"are":[24,141],"learned":[25],"by":[26,77],"models":[27,45,64,151],"through":[28],"gradient-descent":[29],"on":[30],"error,":[31],"predictive":[33],"features":[34],"acquire":[35],"more":[36,47],"salience.":[37],"We":[38,74,93],"show":[39,120],"that":[40,83,121,137],"under":[41,106,154],"multi-outcome":[42,155],"learning,":[43],"predict":[46],"than":[48],"one":[49],"outcome,":[50],"this":[51,76],"vector":[53],"becomes":[54],"unstable;":[55],"it":[56,108],"collapses":[57],"its":[59],"bounds":[60],"and":[61,72,102,119],"prevents":[62],"the":[63,98,104,115],"from":[65],"meaningful":[67,125],"attentional":[68,81,87,139,147],"tunings":[69],"for":[70,145],"generalization.":[73],"address":[75],"introducing":[78],"an":[79,90,95],"outcome-indexed":[80,91,138],"matrix":[82],"converts":[84],"tuning":[88],"into":[89],"representation.":[92],"present":[94],"analysis":[96],"of":[97,152],"unstable":[99],"vectors":[101,130],"derive":[103],"conditions":[105],"which":[107,149],"holds.":[109],"Empirically,":[110],"three":[111],"synthetic":[112],"experiments":[113],"benchmark":[114],"proposed":[116],"matrices":[118,140],"they":[122],"converge":[123],"representations,":[126],"something":[127],"fail":[131],"do.":[133],"results":[135],"suggest":[136],"general":[143],"fix":[144],"gradient-based":[146],"processes,":[148],"improves":[150],"conditions.":[156]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-10T00:00:00"}