{"id":"https://openalex.org/W7204990097","doi":"https://doi.org/10.3390/make8110373","title":"Boundary Screening Bounds the Evidence in Span Masking but Does Not Predict Linear-Probe Representation Quality at Small Scale","display_name":"Boundary Screening Bounds the Evidence in Span Masking but Does Not Predict Linear-Probe Representation Quality at Small Scale","publication_year":2026,"publication_date":"2026-11-09","ids":{"openalex":"https://openalex.org/W7204990097","doi":"https://doi.org/10.3390/make8110373"},"language":"en","primary_location":{"id":"doi:10.3390/make8110373","doi":"https://doi.org/10.3390/make8110373","is_oa":true,"landing_page_url":"https://doi.org/10.3390/make8110373","pdf_url":null,"source":{"id":"https://openalex.org/S4210213891","display_name":"Machine Learning and Knowledge Extraction","issn_l":"2504-4990","issn":["2504-4990"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["cwts-core","doaj","jufo-1","norway-1","poland-20","russia-white-list-1"],"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning and Knowledge Extraction","raw_type":"journal-article"},"volume":null,"type":"article","is_traditional":true,"indexed_in":["crossref","doaj"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.3390/make8110373","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5088438050","display_name":"Suwichai Phunsa","orcid":"https://orcid.org/0000-0003-1093-6999","observed_orcids":["https://orcid.org/0000-0003-1093-6999"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Suwichai Phunsa","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0003-1093-6999","affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5066024716","display_name":"Thawatchai Chomsiri","orcid":"https://orcid.org/0000-0001-5998-9376","observed_orcids":["https://orcid.org/0000-0001-5998-9376"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Thawatchai Chomsiri","raw_affiliation_strings":[],"raw_orcid":"https://orcid.org/0000-0001-5998-9376","affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5066024716"],"corresponding_institution_ids":[],"apc_list":{"value":1800,"currency":"CHF","value_usd":2172},"apc_paid":{"value":1800,"currency":"CHF","value_usd":2172},"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.0,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":"8","issue":"11","first_page":"373","last_page":"373"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9829000234603882,"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":0.9829000234603882,"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":0.007699999958276749,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.0012000000569969416,"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/masked-language-modeling","display_name":"masked language modeling","score":0.6470000147819519},{"id":"https://openalex.org/keywords/representation-quality","display_name":"representation quality","score":0.7070000171661377},{"id":"https://openalex.org/keywords/transformer-encoder","display_name":"Transformer encoder","score":0.5210000276565552},{"id":"https://openalex.org/keywords/gaussian-chain","display_name":"Gaussian chain","score":0.5019999742507935},{"id":"https://openalex.org/keywords/auxiliary-objectives","display_name":"auxiliary objectives","score":0.6710000038146973}],"concepts":[{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5673999786376953},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.557200014591217},{"id":"https://openalex.org/C2778753569","wikidata":"https://www.wikidata.org/wiki/Q1960395","display_name":"Span (engineering)","level":2,"score":0.5302000045776367},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.4758000075817108},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.4683000147342682},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.454800009727478},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.42899999022483826},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4106999933719635},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.38769999146461487},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3874000012874603},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.3804999887943268},{"id":"https://openalex.org/C149629883","wikidata":"https://www.wikidata.org/wiki/Q660926","display_name":"Fraction (chemistry)","level":2,"score":0.35589998960494995},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.3508000075817108},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.35010001063346863},{"id":"https://openalex.org/C42023084","wikidata":"https://www.wikidata.org/wiki/Q5249231","display_name":"Decision boundary","level":3,"score":0.33980000019073486},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.33799999952316284},{"id":"https://openalex.org/C163716315","wikidata":"https://www.wikidata.org/wiki/Q901177","display_name":"Gaussian","level":2,"score":0.33739998936653137},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.33570000529289246},{"id":"https://openalex.org/C62354387","wikidata":"https://www.wikidata.org/wiki/Q875399","display_name":"Boundary (topology)","level":2,"score":0.3314000070095062},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.3231000006198883},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3140999972820282},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.31380000710487366},{"id":"https://openalex.org/C94375191","wikidata":"https://www.wikidata.org/wiki/Q11205","display_name":"Arithmetic","level":1,"score":0.31150001287460327},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.27459999918937683},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.2702000141143799},{"id":"https://openalex.org/C133226019","wikidata":"https://www.wikidata.org/wiki/Q209812","display_name":"Linear span","level":2,"score":0.26420000195503235},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.26030001044273376}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.3390/make8110373","doi":"https://doi.org/10.3390/make8110373","is_oa":true,"landing_page_url":"https://doi.org/10.3390/make8110373","pdf_url":null,"source":{"id":"https://openalex.org/S4210213891","display_name":"Machine Learning and Knowledge Extraction","issn_l":"2504-4990","issn":["2504-4990"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["cwts-core","doaj","jufo-1","norway-1","poland-20","russia-white-list-1"],"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning and Knowledge Extraction","raw_type":"journal-article"},{"id":"doi:10.20944/preprints202609.0101.v1","doi":"https://doi.org/10.20944/preprints202609.0101.v1","is_oa":true,"landing_page_url":"https://doi.org/10.20944/preprints202609.0101.v1","pdf_url":null,"source":{"id":"https://openalex.org/S6309402219","display_name":"Preprints.org","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":null,"host_organization_name":null,"host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"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":"posted-content"}],"best_oa_location":{"id":"doi:10.3390/make8110373","doi":"https://doi.org/10.3390/make8110373","is_oa":true,"landing_page_url":"https://doi.org/10.3390/make8110373","pdf_url":null,"source":{"id":"https://openalex.org/S4210213891","display_name":"Machine Learning and Knowledge Extraction","issn_l":"2504-4990","issn":["2504-4990"],"is_oa":true,"is_in_doaj":true,"is_core":true,"listed_in":["cwts-core","doaj","jufo-1","norway-1","poland-20","russia-white-list-1"],"host_organization":"https://openalex.org/P4310310987","host_organization_name":"Multidisciplinary Digital Publishing Institute","host_organization_lineage":["https://openalex.org/P4310310987"],"host_organization_lineage_names":["Multidisciplinary Digital Publishing Institute"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Machine Learning and Knowledge Extraction","raw_type":"journal-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4387846608","https://openalex.org/W4386566778","https://openalex.org/W3090048242","https://openalex.org/W4386566638","https://openalex.org/W4361230791","https://openalex.org/W4287888493","https://openalex.org/W7219373864","https://openalex.org/W2962753370","https://openalex.org/W4385567255","https://openalex.org/W7140140792"],"abstract_inverted_index":{"Span":[0],"masking":[1,87],"is":[2,75,153,190],"usually":[3],"justified":[4],"informally:":[5],"contiguous":[6],"masks":[7],"are":[8,176],"said":[9],"to":[10,100,143,159],"force":[11],"phrase-level":[12],"rather":[13,55],"than":[14,56],"token-level":[15],"inference.":[16],"We":[17],"make":[18],"that":[19,30],"argument":[20],"quantitative":[21],"and":[22,111,163],"then":[23],"test":[24],"it.":[25],"A":[26],"boundary-screening":[27],"theorem":[28],"shows":[29],"the":[31,43,50,57,98,104,119,134,187],"evidence":[32,121],"a":[33,37,78,85,95,146,156,183],"context":[34],"carries":[35],"about":[36],"masked":[38,53,60],"set":[39],"passes":[40],"entirely":[41],"through":[42],"block":[44],"boundaries,":[45],"so":[46],"it":[47],"scales":[48],"with":[49,68],"number":[51,58],"of":[52,59,81,97],"blocks":[54],"tokens.":[61],"The":[62],"downstream":[63],"prediction":[64],"this":[65,128],"suggests,":[66],"registered":[67],"decision":[69],"rules":[70],"fixed":[71,86],"before":[72],"any":[73],"run,":[74],"refuted.":[76],"Across":[77],"sixteen-fold":[79],"range":[80],"span":[82],"length":[83],"at":[84,127,185],"ratio,":[88],"probe":[89],"accuracy":[90],"varies":[91],"by":[92],"0.80":[93],"points,":[94],"fraction":[96],"2.0":[99],"4.3":[101],"point":[102],"changes":[103],"same":[105],"design":[106],"resolves":[107],"from":[108,112],"auxiliary":[109,165],"objectives":[110,152],"training":[113],"length.":[114],"Mask":[115],"geometry":[116],"therefore":[117],"bounds":[118],"available":[120],"without":[122],"predicting":[123],"tested":[124],"representation":[125],"quality":[126],"scale.":[129],"Three":[130],"predictions":[131],"do":[132],"hold:":[133],"closed":[135],"form":[136],"for":[137,150],"Gaussian":[138],"chains":[139],"matches":[140],"direct":[141],"computation":[142],"10-13":[144],"nats,":[145],"radial":[147],"conservation":[148],"law":[149],"cosine":[151],"recovered":[154],"in":[155],"trained":[157,179],"transformer":[158],"within":[160],"3":[161],"percent,":[162],"an":[164],"masked-language":[166],"term":[167],"improves":[168],"lexical":[169],"probes":[170],"while":[171],"degrading":[172],"semantic":[173],"ones.":[174],"Encoders":[175],"4.5M":[177],"parameters":[178],"on":[180],"74M":[181],"tokens,":[182],"scale":[184],"which":[186],"random-initialised":[188],"baseline":[189],"already":[191],"strong.":[192]},"counts_by_year":[],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2026-09-02T00:00:00"}