{"id":"https://openalex.org/W7204190651","doi":"https://doi.org/10.48550/arxiv.2608.22117","title":"TANGO: Treating Tokens as Operators","display_name":"TANGO: Treating Tokens as Operators","publication_year":2026,"publication_date":"2026-08-22","ids":{"openalex":"https://openalex.org/W7204190651","doi":"https://doi.org/10.48550/arxiv.2608.22117"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.22117","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22117","pdf_url":"https://arxiv.org/pdf/2608.22117","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","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.22117","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5092911518","display_name":"Joshua Nunley","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210119109","display_name":"Indiana University Bloomington","ror":"https://ror.org/02k40bc56","country_code":"US","type":"education","lineage":["https://openalex.org/I4210119109","https://openalex.org/I592451"]},{"id":"https://openalex.org/I592451","display_name":"Indiana University","ror":"https://ror.org/01kg8sb98","country_code":"US","type":"education","lineage":["https://openalex.org/I592451"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Nunley, Joshua","raw_affiliation_strings":["Department of Informatics Luddy School of Informatics, Computing, and Engineering Cognitive Science Program Indiana University Bloomington"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Informatics Luddy School of Informatics, Computing, and Engineering Cognitive Science Program Indiana University Bloomington","institution_ids":["https://openalex.org/I4210119109","https://openalex.org/I592451"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5092911518"],"corresponding_institution_ids":["https://openalex.org/I4210119109","https://openalex.org/I592451"],"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/T10028","display_name":"Topic Modeling","score":0.885699987411499,"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/T10028","display_name":"Topic Modeling","score":0.885699987411499,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.07760000228881836,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.00839999970048666,"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/tango","display_name":"tango","score":0.7360000014305115},{"id":"https://openalex.org/keywords/transformer","display_name":"transformer models","score":0.4359999895095825},{"id":"https://openalex.org/keywords/natural-language-modeling","display_name":"natural language modeling","score":0.6489999890327454},{"id":"https://openalex.org/keywords/mathematical-reasoning","display_name":"mathematical reasoning","score":0.30300000309944153},{"id":"https://openalex.org/keywords/negative-log-likelihood","display_name":"negative log-likelihood","score":0.5770000219345093},{"id":"https://openalex.org/keywords/computational-efficiency","display_name":"computational efficiency","score":0.37700000405311584}],"concepts":[{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.6132000088691711},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.593999981880188},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5432999730110168},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.492900013923645},{"id":"https://openalex.org/C129844170","wikidata":"https://www.wikidata.org/wiki/Q41299","display_name":"Quadratic equation","level":2,"score":0.4616999924182892},{"id":"https://openalex.org/C194544171","wikidata":"https://www.wikidata.org/wiki/Q21105679","display_name":"Gating","level":2,"score":0.44269999861717224},{"id":"https://openalex.org/C141603448","wikidata":"https://www.wikidata.org/wiki/Q134830","display_name":"Prefix","level":2,"score":0.4424000084400177},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.4323999881744385},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4228000044822693},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.4049000144004822},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.38909998536109924},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37119999527931213},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.3698999881744385},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3684000074863434},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.35910001397132874},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.3540000021457672},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.34439998865127563},{"id":"https://openalex.org/C2778751112","wikidata":"https://www.wikidata.org/wiki/Q835016","display_name":"Window (computing)","level":2,"score":0.3321000039577484},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3140000104904175},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.30799999833106995},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.29269999265670776},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.2824999988079071},{"id":"https://openalex.org/C146072743","wikidata":"https://www.wikidata.org/wiki/Q192161","display_name":"Formal language","level":2,"score":0.2718000113964081},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.26989999413490295},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2694000005722046},{"id":"https://openalex.org/C311688","wikidata":"https://www.wikidata.org/wiki/Q2393193","display_name":"Time complexity","level":2,"score":0.26930001378059387},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.26669999957084656}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.22117","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22117","pdf_url":"https://arxiv.org/pdf/2608.22117","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.2608.22117","doi":"https://doi.org/10.48550/arxiv.2608.22117","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.22117","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:2608.22117","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.22117","pdf_url":"https://arxiv.org/pdf/2608.22117","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":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204190651.pdf","grobid_xml":"https://content.openalex.org/works/W7204190651.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4405783272","https://openalex.org/W7160635552","https://openalex.org/W4416970394","https://openalex.org/W7223574330","https://openalex.org/W7127285768","https://openalex.org/W7223564422","https://openalex.org/W7162140279","https://openalex.org/W7168234578","https://openalex.org/W7204148737","https://openalex.org/W4414191365"],"abstract_inverted_index":{"Transformers":[0],"separate":[1],"cross-token":[2],"mixing":[3],"in":[4,9,153,185,255,261,271],"self-attention":[5],"from":[6,173],"token-wise":[7],"transformation":[8],"feed-forward":[10],"networks.":[11],"We":[12,76],"ask":[13],"whether":[14],"combining":[15],"these":[16,52],"operations":[17],"can":[18],"lower":[19,252,266],"predictive":[20],"loss":[21,167,183],"under":[22],"fixed":[23],"data":[24,131],"and":[25,66,83,88,96,113,123,132,141,163,219,237,258],"parameter":[26,119],"budgets.":[27],"To":[28,201],"do":[29],"so,":[30],"we":[31,205],"introduce":[32],"the":[33,64,68,91,147,180,216,233,243,262],"Token-Aggregated":[34],"Nonlinear":[35],"Gating":[36],"Operator":[37],"(TANGO)":[38],"model.":[39],"TANGO":[40,85,145,188,236],"computes":[41],"a":[42,60,72,101,207],"nonlinear":[43],"feature-wise":[44],"gate":[45],"at":[46],"each":[47,55],"source":[48],"token.":[49],"Attention":[50,93,99],"averages":[51,194],"gates":[53,196],"for":[54],"destination.":[56],"The":[57,115],"average":[58],"modulates":[59],"linear":[61,74],"projection":[62],"of":[63,71,161,235,242],"destination":[65],"forms":[67],"diagonal":[69],"core":[70],"source-conditioned":[73],"operator.":[75],"test":[77,136,150],"this":[78,203,226],"proposal":[79],"by":[80],"comparing":[81],"full-prefix":[82,144],"windowed":[84],"with":[86,100,209],"looped":[87,181,244],"untied":[89],"Transformers,":[90],"Gated":[92],"Unit":[94],"(GAU),":[95],"Fast":[97],"Linear":[98],"Single":[102],"Head":[103],"(FLASH)":[104],"on":[105,129,134],"web":[106],"text,":[107],"Lean":[108],"formal":[109],"mathematics,":[110],"DeepMind":[111],"Mathematics,":[112],"code.":[114],"comparison":[116],"uses":[117],"two":[118,121],"scales,":[120],"depths,":[122],"three":[124,210,256],"seeds.":[125],"Checkpoints":[126],"are":[127],"selected":[128],"development":[130,166,248],"evaluated":[133],"held-out":[135],"data.":[137],"At":[138],"matched":[139],"parameters":[140],"training":[142],"data,":[143],"has":[146],"lowest":[148],"mean":[149,247],"negative":[151,249],"log-likelihood":[152,250],"all":[154,272],"16":[155],"settings.":[156],"In":[157],"eight":[158],"additional":[159],"combinations":[160],"size":[162],"dataset,":[164],"its":[165],"never":[168],"increases":[169,184],"as":[170],"depth":[171],"rises":[172],"4":[174],"to":[175,177,215,230,239],"8":[176],"16,":[178],"whereas":[179],"Transformer's":[182],"four.":[186,273],"Full-prefix":[187],"is":[189,251],"computationally":[190],"expensive":[191],"because":[192],"it":[193],"wide":[195],"over":[197],"every":[198],"visible":[199],"source.":[200],"reduce":[202],"cost,":[204],"evaluate":[206],"variant":[208,227],"narrower":[211],"gated-projection":[212],"sets":[213],"assigned":[214],"first,":[217],"middle,":[218],"last":[220],"applications.":[221],"Across":[222],"four":[223],"FineWeb-Edu":[224],"settings,":[225],"achieves":[228],"3.26":[229],"3.45":[231],"times":[232],"throughput":[234],"75%":[238],"96%":[240],"that":[241],"Transformer.":[245],"Its":[246],"than":[253,267],"TANGO's":[254],"settings":[257],"0.023":[259],"higher":[260],"fourth,":[263],"while":[264],"remaining":[265],"both":[268],"Transformer":[269],"baselines":[270]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-08-26T00:00:00"}