{"id":"https://openalex.org/W4416035466","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.1356","title":"Pre-trained Language Models Learn Remarkably Accurate Representations of Numbers","display_name":"Pre-trained Language Models Learn Remarkably Accurate Representations of Numbers","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416035466","doi":"https://doi.org/10.18653/v1/2025.emnlp-main.1356"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.emnlp-main.1356","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.1356","pdf_url":"https://aclanthology.org/2025.emnlp-main.1356.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.emnlp-main.1356.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032449871","display_name":"Marek Kadl\u010d\u00edk","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Marek Kadl\u010d\u00edk","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102979706","display_name":"Michal \u0160tef\u00e1nik","orcid":"https://orcid.org/0000-0003-1766-5538"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Michal \u0160tef\u00e1nik","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5073653815","display_name":"Timothee Mickus","orcid":"https://orcid.org/0000-0001-9538-7209"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Timothee Mickus","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5022262572","display_name":"James K. Kuchar","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Josef Kucha\u0159","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5104202378","display_name":"Michal Spiegel","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Michal Spiegel","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5032449871"],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.18095134,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"26693","last_page":"26702"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.12880000472068787,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.12880000472068787,"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":0.11420000344514847,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.06499999761581421,"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/representation","display_name":"Representation (politics)","score":0.3443000018596649},{"id":"https://openalex.org/keywords/natural-language","display_name":"Natural language","score":0.3070000112056732},{"id":"https://openalex.org/keywords/identification","display_name":"Identification (biology)","score":0.28700000047683716},{"id":"https://openalex.org/keywords/abstraction","display_name":"Abstraction","score":0.2867000102996826},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.28450000286102295},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.2775000035762787}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5649999976158142},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4717999994754791},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.47110000252723694},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3467000126838684},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.3443000018596649},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3070000112056732},{"id":"https://openalex.org/C116834253","wikidata":"https://www.wikidata.org/wiki/Q2039217","display_name":"Identification (biology)","level":2,"score":0.28700000047683716},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.2867000102996826},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.28450000286102295},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2669999897480011},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26010000705718994},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.2563999891281128}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.emnlp-main.1356","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.1356","pdf_url":"https://aclanthology.org/2025.emnlp-main.1356.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.emnlp-main.1356","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.emnlp-main.1356","pdf_url":"https://aclanthology.org/2025.emnlp-main.1356.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G5112024542","display_name":"High Performance Language Technologies","funder_award_id":"101070350","funder_id":"https://openalex.org/F4320320300","funder_display_name":"European Commission"},{"id":"https://openalex.org/G8051717526","display_name":null,"funder_award_id":"Grant","funder_id":"https://openalex.org/F4320320300","funder_display_name":"European Commission"},{"id":"https://openalex.org/G8318064016","display_name":null,"funder_award_id":"Horizon","funder_id":"https://openalex.org/F4320320300","funder_display_name":"European Commission"}],"funders":[{"id":"https://openalex.org/F4320314707","display_name":"Government of the United Kingdom","ror":"https://ror.org/05wnh3t63"},{"id":"https://openalex.org/F4320314731","display_name":"UK Research and Innovation","ror":"https://ror.org/001aqnf71"},{"id":"https://openalex.org/F4320320300","display_name":"European Commission","ror":"https://ror.org/00k4n6c32"},{"id":"https://openalex.org/F4320334322","display_name":"HORIZON EUROPE Framework Programme","ror":null}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416035466.pdf","grobid_xml":"https://content.openalex.org/works/W4416035466.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Pretrained":[0],"language":[1],"models":[2],"(LMs)":[3],"are":[4,45],"prone":[5],"to":[6,27],"arithmetic":[7],"errors.Existing":[8],"work":[9],"showed":[10],"limited":[11],"success":[12],"in":[13,35,111],"probing":[14,43,63],"numeric":[15,67],"values":[16,68],"from":[17,69],"models'":[18],"representations,":[19],"indicating":[20],"that":[21,41,65,82,95,116],"these":[22,128],"errors":[23,110],"can":[24,126],"be":[25],"attributed":[26],"the":[28,48,84,96,118,121],"inherent":[29],"unreliability":[30],"of":[31,51,78,108],"distributionally":[32],"learned":[33,52],"embeddings":[34,54,71,119],"representing":[36],"exact":[37],"quantities.However,":[38],"we":[39,59,93],"observe":[40],"previous":[42],"methods":[44],"inadequate":[46],"for":[47],"emergent":[49],"structure":[50],"number":[53],"with":[55,72,90,120],"sinusoidal":[56],"patterns.In":[57],"response,":[58],"propose":[60],"a":[61,76,105],"novel":[62],"technique":[64],"decodes":[66],"input":[70],"near-perfect":[73],"accuracy":[74],"across":[75],"range":[77],"open-source":[79],"LMs.This":[80],"proves":[81],"after":[83],"sole":[85],"pre-training,":[86],"LMs":[87],"represent":[88],"numbers":[89],"remarkable":[91],"precision.Finally,":[92],"find":[94],"embeddings'":[97],"precision,":[98],"judged":[99],"by":[100],"our":[101,123],"probe's":[102],"accuracy,":[103],"explains":[104],"large":[106],"portion":[107],"LM's":[109],"elementary":[112],"arithmetic,":[113],"and":[114],"show":[115],"aligning":[117],"pattern":[122],"probes":[124],"discover":[125],"mitigate":[127],"errors.":[129]},"counts_by_year":[],"updated_date":"2026-03-21T08:13:44.787528","created_date":"2025-11-08T00:00:00"}