{"id":"https://openalex.org/W7220653121","doi":"https://doi.org/10.48550/arxiv.2610.06513","title":"SOL: Measuring Gaps between Text Distributions by Double Sliced Wasserstein Metrics","display_name":"SOL: Measuring Gaps between Text Distributions by Double Sliced Wasserstein Metrics","publication_year":2026,"publication_date":"2026-10-05","ids":{"openalex":"https://openalex.org/W7220653121","doi":"https://doi.org/10.48550/arxiv.2610.06513"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.06513","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.06513","pdf_url":"https://arxiv.org/pdf/2610.06513","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/2610.06513","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5016760339","display_name":"Gregor Kornhardt","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Kornhardt, Gregor","raw_affiliation_strings":["Institute of Mathematics , TU Berlin , Berlin , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Mathematics , TU Berlin , Berlin , Germany","institution_ids":["https://openalex.org/I4577782"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068150443","display_name":"Moritz Piening","orcid":"https://orcid.org/0009-0003-3877-4511","observed_orcids":["https://orcid.org/0009-0003-3877-4511"]},"institutions":[{"id":"https://openalex.org/I74656192","display_name":"University of G\u00f6ttingen","ror":"https://ror.org/01y9bpm73","country_code":"DE","type":"education","lineage":["https://openalex.org/I74656192"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Piening, Moritz","raw_affiliation_strings":["Department of Computer Science , University of G\u00f6ttingen , G\u00f6ttingen , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science , University of G\u00f6ttingen , G\u00f6ttingen , Germany","institution_ids":["https://openalex.org/I74656192"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093022018","display_name":"Jannis Chemseddine","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Chemseddine, Jannis","raw_affiliation_strings":["Institute of Mathematics , TU Berlin , Berlin , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Mathematics , TU Berlin , Berlin , Germany","institution_ids":["https://openalex.org/I4577782"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5102760434","display_name":"Gabriele Steidl","orcid":"https://orcid.org/0000-0001-7486-6315","observed_orcids":["https://orcid.org/0000-0001-7486-6315"]},"institutions":[{"id":"https://openalex.org/I4577782","display_name":"Technische Universit\u00e4t Berlin","ror":"https://ror.org/03v4gjf40","country_code":"DE","type":"education","lineage":["https://openalex.org/I4577782"]}],"countries":["DE"],"is_corresponding":false,"raw_author_name":"Steidl, Gabriele","raw_affiliation_strings":["Institute of Mathematics , TU Berlin , Berlin , Germany"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Institute of Mathematics , TU Berlin , Berlin , Germany","institution_ids":["https://openalex.org/I4577782"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T10028","display_name":"Topic Modeling","score":0.9672999978065491,"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":0.9672999978065491,"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":0.01810000091791153,"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":0.005400000140070915,"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/generative-model-evaluation","display_name":"generative model evaluation","score":0.5740000009536743},{"id":"https://openalex.org/keywords/diffusion-models","display_name":"diffusion models","score":0.5929999947547913}],"concepts":[{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.6049000024795532},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.5626000165939331},{"id":"https://openalex.org/C2777634741","wikidata":"https://www.wikidata.org/wiki/Q768993","display_name":"Wasserstein metric","level":2,"score":0.5568000078201294},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.5162000060081482},{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.45239999890327454},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4205000102519989},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.41990000009536743},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.41749998927116394},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.38339999318122864},{"id":"https://openalex.org/C176217482","wikidata":"https://www.wikidata.org/wiki/Q860554","display_name":"Metric (unit)","level":2,"score":0.3589000105857849},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.3479999899864197},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.3434000015258789},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.33230000734329224},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.326200008392334},{"id":"https://openalex.org/C9679016","wikidata":"https://www.wikidata.org/wiki/Q1417473","display_name":"Principle of maximum entropy","level":2,"score":0.32280001044273376},{"id":"https://openalex.org/C28826006","wikidata":"https://www.wikidata.org/wiki/Q33521","display_name":"Applied mathematics","level":1,"score":0.3070000112056732},{"id":"https://openalex.org/C98385598","wikidata":"https://www.wikidata.org/wiki/Q1339385","display_name":"Empirical distribution function","level":2,"score":0.29820001125335693},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.26969999074935913},{"id":"https://openalex.org/C206654554","wikidata":"https://www.wikidata.org/wiki/Q5374247","display_name":"Empirical measure","level":2,"score":0.2662000060081482},{"id":"https://openalex.org/C177769412","wikidata":"https://www.wikidata.org/wiki/Q278090","display_name":"Prior probability","level":3,"score":0.26499998569488525},{"id":"https://openalex.org/C21031990","wikidata":"https://www.wikidata.org/wiki/Q355020","display_name":"Probability measure","level":2,"score":0.25529998540878296},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.25459998846054077},{"id":"https://openalex.org/C39927690","wikidata":"https://www.wikidata.org/wiki/Q11197","display_name":"Logarithm","level":2,"score":0.25200000405311584},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2515000104904175},{"id":"https://openalex.org/C197096303","wikidata":"https://www.wikidata.org/wiki/Q869887","display_name":"Probability mass function","level":3,"score":0.2515000104904175}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.06513","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.06513","pdf_url":"https://arxiv.org/pdf/2610.06513","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.2610.06513","doi":"https://doi.org/10.48550/arxiv.2610.06513","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.06513","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:2610.06513","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.06513","pdf_url":"https://arxiv.org/pdf/2610.06513","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/G442688124","display_name":null,"funder_award_id":"SPP2298","funder_id":"https://openalex.org/F4320320879","funder_display_name":"Deutsche Forschungsgemeinschaft"}],"funders":[{"id":"https://openalex.org/F4320320879","display_name":"Deutsche Forschungsgemeinschaft","ror":"https://ror.org/018mejw64"},{"id":"https://openalex.org/F4320321114","display_name":"Bundesministerium f\u00fcr Bildung und Forschung","ror":"https://ror.org/04pz7b180"},{"id":"https://openalex.org/F4320322892","display_name":"Centre National de la Recherche Scientifique","ror":"https://ror.org/02feahw73"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7220653121.pdf","grobid_xml":"https://content.openalex.org/works/W7220653121.grobid-xml"},"referenced_works_count":23,"referenced_works":["https://openalex.org/W1539309091","https://openalex.org/W2168151633","https://openalex.org/W4310421098","https://openalex.org/W4310923309","https://openalex.org/W4385570969","https://openalex.org/W4387964173","https://openalex.org/W4406032137","https://openalex.org/W4414939355","https://openalex.org/W4415800813","https://openalex.org/W4416149152","https://openalex.org/W6910504028","https://openalex.org/W7125157509","https://openalex.org/W7133235760","https://openalex.org/W7133939672","https://openalex.org/W7154221019","https://openalex.org/W7160636438","https://openalex.org/W7160879804","https://openalex.org/W7160955450","https://openalex.org/W7162128388","https://openalex.org/W7164037820","https://openalex.org/W7166872990","https://openalex.org/W7167058132","https://openalex.org/W7214189200"],"related_works":["https://openalex.org/W3129146692","https://openalex.org/W7164037820","https://openalex.org/W7150793535","https://openalex.org/W4385570969","https://openalex.org/W2979952454","https://openalex.org/W4416035978","https://openalex.org/W3034557228","https://openalex.org/W3214667981","https://openalex.org/W2799184518","https://openalex.org/W7125652609"],"abstract_inverted_index":{"Evaluating":[0],"text":[1,60],"generation":[2],"requires":[3],"measuring":[4],"how":[5],"well":[6],"the":[7,11,21,51,67,79,87,100,126,129],"generated":[8],"distribution":[9,52],"matches":[10],"data":[12],"distribution.":[13],"For":[14],"autoregressive":[15],"models,":[16],"this":[17],"is":[18,64,96,102],"done":[19],"by":[20,66,86],"perplexity.":[22],"Diffusion":[23],"and":[24,78,115],"flow-based":[25],"language":[26],"models":[27,150],"can":[28],"only":[29],"provide":[30],"a":[31,57,75,97,139,147],"likelihood":[32],"bound,":[33],"whose":[34],"tightness":[35],"differs":[36],"between":[37,59],"model":[38,113],"families.":[39],"Sample-based":[40],"substitutes":[41],"such":[42],"as":[43],"generative":[44],"perplexity":[45],"with":[46],"entropy":[47],"do":[48],"not":[49],"consider":[50],"fit.":[53],"We":[54,92,120],"propose":[55],"SOL,":[56],"distance":[58],"distributions.":[61],"Each":[62],"sequence":[63],"represented":[65],"empirical":[68],"measure":[69],"of":[70,81,149],"its":[71],"hidden":[72],"states":[73],"under":[74],"fixed":[76],"transformer":[77,101],"distributions":[80],"these":[82],"measures":[83],"are":[84],"compared":[85],"double":[88],"sliced":[89],"Wasserstein":[90],"distance.":[91],"prove":[93],"that":[94,106],"SOL":[95,107,123,144],"metric":[98],"if":[99],"injective.":[103],"Experiments":[104],"show":[105],"detects":[108],"distributional":[109],"failures,":[110],"recovers":[111],"expected":[112],"trends,":[114],"provides":[116],"stable":[117],"sample-based":[118],"estimates.":[119],"put":[121],"forward":[122],"to":[124,145],"fill":[125],"gap":[127],"in":[128],"current":[130],"evaluation":[131],"protocol":[132],"used":[133],"for":[134],"non":[135],"auto-regressive":[136],"models.":[137],"As":[138],"first":[140],"step":[141],"we":[142],"use":[143],"re-evaluate":[146],"variety":[148],"trained":[151],"on":[152],"OpenWebText.":[153]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-07T00:00:00"}