{"id":"https://openalex.org/W7204221856","doi":"https://doi.org/10.18653/2026.emnlp-main.1352","title":"Linear Probing Provides Robust and Efficient Detection of Machine-Generated Text","display_name":"Linear Probing Provides Robust and Efficient Detection of Machine-Generated Text","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7204221856","doi":"https://doi.org/10.18653/2026.emnlp-main.1352"},"language":"en","primary_location":{"id":"doi:10.18653/2026.emnlp-main.1352","doi":"https://doi.org/10.18653/2026.emnlp-main.1352","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1352","pdf_url":"https://aclanthology.org/2026.emnlp-main.1352.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2026.emnlp-main.1352.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134569541","display_name":"Gerrit Quaremba","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183935753","display_name":"King's College London","ror":"https://ror.org/0220mzb33","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I183935753"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Gerrit Quaremba","raw_affiliation_strings":["King's College London ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"King's College London ,","institution_ids":["https://openalex.org/I183935753"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148265123","display_name":"Hanqi Yan","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183935753","display_name":"King's College London","ror":"https://ror.org/0220mzb33","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I183935753"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Hanqi Yan","raw_affiliation_strings":["King's College London ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"King's College London ,","institution_ids":["https://openalex.org/I183935753"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137586960","display_name":"Elizabeth Black","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183935753","display_name":"King's College London","ror":"https://ror.org/0220mzb33","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I183935753"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Elizabeth Black","raw_affiliation_strings":["King's College London ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"King's College London ,","institution_ids":["https://openalex.org/I183935753"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111465891","display_name":"Denny Vrande\u010di\u0107","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4210126245","display_name":"Wikimedia Foundation","ror":"https://ror.org/032q98j12","country_code":"US","type":"nonprofit","lineage":["https://openalex.org/I4210126245"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Denny Vrande\u010di\u0107","raw_affiliation_strings":["Wikimedia Foundation"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Wikimedia Foundation","institution_ids":["https://openalex.org/I4210126245"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5125242127","display_name":"Elena Simperl","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I183935753","display_name":"King's College London","ror":"https://ror.org/0220mzb33","country_code":"GB","type":"education","lineage":["https://openalex.org/I124357947","https://openalex.org/I183935753"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Elena Simperl","raw_affiliation_strings":["King's College London ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"King's College London ,","institution_ids":["https://openalex.org/I183935753"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.68263624,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"30805","last_page":"30827"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12380","display_name":"Authorship Attribution and Profiling","score":0.8148999810218811,"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/T12380","display_name":"Authorship Attribution and Profiling","score":0.8148999810218811,"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":0.05000000074505806,"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":0.037300001829862595,"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/ai-generated-text-detection","display_name":"AI-generated text detection","score":0.996999979019165},{"id":"https://openalex.org/keywords/linear-probing","display_name":"linear probing","score":0.6800000071525574},{"id":"https://openalex.org/keywords/latent-representations","display_name":"latent representations","score":0.6800000071525574},{"id":"https://openalex.org/keywords/out-of-domain-generalization","display_name":"out-of-domain generalization","score":0.7570000290870667},{"id":"https://openalex.org/keywords/sample-efficiency","display_name":"sample efficiency","score":0.8360000252723694},{"id":"https://openalex.org/keywords/ai-generated-text","display_name":"AI-generated text","score":0.5580000281333923},{"id":"https://openalex.org/keywords/human-written-text","display_name":"human-written text","score":0.6779999732971191}],"concepts":[{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6509000062942505},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6029999852180481},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.5950000286102295},{"id":"https://openalex.org/C77170095","wikidata":"https://www.wikidata.org/wiki/Q1753188","display_name":"Linearity","level":2,"score":0.5527999997138977},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.5479000210762024},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.5020999908447266},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4674000144004822},{"id":"https://openalex.org/C61272859","wikidata":"https://www.wikidata.org/wiki/Q7834031","display_name":"Transferability","level":3,"score":0.45719999074935913},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.44749999046325684},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.4390999972820282},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4318999946117401},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.4316999912261963},{"id":"https://openalex.org/C70710897","wikidata":"https://www.wikidata.org/wiki/Q680081","display_name":"Separable space","level":2,"score":0.41290000081062317},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3962000012397766},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3601999878883362},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3441999852657318},{"id":"https://openalex.org/C2779304628","wikidata":"https://www.wikidata.org/wiki/Q3503480","display_name":"Face (sociological concept)","level":2,"score":0.34130001068115234},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3059999942779541},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3057999908924103},{"id":"https://openalex.org/C163175372","wikidata":"https://www.wikidata.org/wiki/Q3339222","display_name":"Linear model","level":2,"score":0.2987000048160553},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26759999990463257},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C6802819","wikidata":"https://www.wikidata.org/wiki/Q1072174","display_name":"Linear system","level":2,"score":0.25529998540878296}],"mesh":[],"locations_count":4,"locations":[{"id":"doi:10.18653/2026.emnlp-main.1352","doi":"https://doi.org/10.18653/2026.emnlp-main.1352","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1352","pdf_url":"https://aclanthology.org/2026.emnlp-main.1352.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2608.24780","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24780","pdf_url":"https://arxiv.org/pdf/2608.24780","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":"pmh:oai:kclpure.kcl.ac.uk:openaire/05052a34-f805-4984-9839-c14cedcd6ffa","doi":null,"is_oa":true,"landing_page_url":"https://kclpure.kcl.ac.uk/portal/en/publications/05052a34-f805-4984-9839-c14cedcd6ffa","pdf_url":null,"source":{"id":"https://openalex.org/S4306400216","display_name":"Research Portal (King's College London)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I183935753","host_organization_name":"King's College London","host_organization_lineage":["https://openalex.org/I183935753"],"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":"Quaremba, G, Yan, H, Black, E, Vrandecic, D & Simperl, E 2026, Linear Probing Provides Robust and Efficient Detection of Machine-Generated Text. in The 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026) .","raw_type":"info:eu-repo/semantics/publishedVersion"},{"id":"doi:10.48550/arxiv.2608.24780","doi":"https://doi.org/10.48550/arxiv.2608.24780","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.24780","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":"doi:10.18653/2026.emnlp-main.1352","doi":"https://doi.org/10.18653/2026.emnlp-main.1352","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.emnlp-main.1352","pdf_url":"https://aclanthology.org/2026.emnlp-main.1352.pdf","source":{"id":"https://openalex.org/S4363608991","display_name":"Conference on Empirical Methods in Natural Language Processing (EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G4269887845","display_name":null,"funder_award_id":"EP/S023356/1","funder_id":"https://openalex.org/F4320314731","funder_display_name":"UK Research and Innovation"},{"id":"https://openalex.org/G8801017843","display_name":"UKRI Centre for Doctoral Training in Safe and Trusted Artificial Intelligence","funder_award_id":"EP/S023356/1","funder_id":"https://openalex.org/F4320334627","funder_display_name":"Engineering and Physical Sciences Research Council"}],"funders":[{"id":"https://openalex.org/F4320314731","display_name":"UK Research and Innovation","ror":"https://ror.org/001aqnf71"},{"id":"https://openalex.org/F4320334627","display_name":"Engineering and Physical Sciences Research Council","ror":"https://ror.org/0439y7842"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204221856.pdf","grobid_xml":"https://content.openalex.org/works/W7204221856.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7130693797","https://openalex.org/W4412393837","https://openalex.org/W4412164711","https://openalex.org/W6958861474","https://openalex.org/W4391013637","https://openalex.org/W7220915169","https://openalex.org/W4377865296","https://openalex.org/W4412888845","https://openalex.org/W7221109101","https://openalex.org/W7218109084"],"abstract_inverted_index":{"Distinguishing":[0],"machine-generated":[1],"text":[2,6],"(MGT)":[3],"from":[4],"human-written":[5],"(HWT)":[7],"becomes":[8],"increasingly":[9],"important":[10],"due":[11],"to":[12,115],"potential":[13,150,171],"misuse.However,":[14],"most":[15],"supervised":[16],"detectors":[17,50],"often":[18],"degrade":[19],"out-of-domain":[20],"(OOD)":[21],"and":[22,34,39,60,70,96,166,168,178],"require":[23],"large,":[24],"diverse":[25,135],"training":[26],"sets.In":[27],"this":[28,76,121],"work,":[29],"we":[30,88,137],"analyze":[31],"the":[32,170],"linearity":[33],"quality":[35],"of":[36,49,92,146,154,172],"MGT":[37,59,130,165,180],"representations":[38,63],"show":[40,57,119],"that":[41,58,120,132,139],"simple":[42,93],"linear":[43,94,173],"probes":[44,95,125,174],"outperform":[45],"a":[46,72,127,143],"wide":[47],"range":[48],"while":[51],"being":[52],"substantially":[53],"more":[54],"sample-efficient.We":[55],"first":[56],"HWT":[61,167],"latent":[62,129],"are":[64],"linearly":[65],"separable":[66],"in":[67,81],"low-dimensional":[68],"space,":[69],"provide":[71],"plausible":[73],"explanation":[74],"for":[75,151],"separability":[77],"through":[78],"systematic":[79],"differences":[80,163],"their":[82,149],"representation":[83],"quality.Motivated":[84],"by":[85],"these":[86],"insights,":[87],"train":[89],"two":[90],"variants":[91],"evaluate":[97],"them":[98],"across":[99,134],"4":[100],"benchmarks":[101],"against":[102],"16":[103],"baselines.Probes":[104],"consistently":[105],"improve":[106],"OOD":[107],"detection":[108],"(+11":[109],"AUC),":[110],"requiring":[111],"solely":[112],"<100":[113],"samples":[114],"reach":[116],"nearpeak":[117],"performance.We":[118],"transferability":[122],"arises":[123],"because":[124],"recover":[126],"shared":[128],"direction":[131],"generalizes":[133],"settings.Finally,":[136],"demonstrate":[138],"probing":[140],"vectors":[141],"capture":[142],"continuous":[144],"spectrum":[145],"\"machineness\",":[147],"highlighting":[148],"fine-grained":[152],"estimation":[153],"AI-edited":[155],"text.Overall,":[156],"our":[157,183],"work":[158],"provides":[159],"insights":[160],"into":[161],"latentspace":[162],"between":[164],"demonstrates":[169],"as":[175,176],"robust":[177],"sample-efficient":[179],"detectors.We":[181],"release":[182],"code":[184],"on":[185],"github.":[186]},"counts_by_year":[],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2026-08-27T00:00:00"}