{"id":"https://openalex.org/W4400582517","doi":"https://doi.org/10.1145/3660799","title":"Glitch Tokens in Large Language Models: Categorization Taxonomy and Effective Detection","display_name":"Glitch Tokens in Large Language Models: Categorization Taxonomy and Effective Detection","publication_year":2024,"publication_date":"2024-07-12","ids":{"openalex":"https://openalex.org/W4400582517","doi":"https://doi.org/10.1145/3660799"},"language":"en","primary_location":{"id":"doi:10.1145/3660799","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3660799","pdf_url":null,"source":{"id":"https://openalex.org/S4404663975","display_name":"Proceedings of the ACM on software engineering.","issn_l":"2994-970X","issn":["2994-970X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM on Software Engineering","raw_type":"journal-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://vtechworks.lib.vt.edu/bitstreams/6fc8fe0b-ce4a-4504-b983-5ba3127f7a1d/download","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5020034653","display_name":"Yuxi Li","orcid":"https://orcid.org/0009-0008-8032-3841"},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["AU","CN"],"is_corresponding":true,"raw_author_name":"Yuxi Li","raw_affiliation_strings":["Huazhong University of Science and Technology, Wuhan, China","UNSW, Sydney, Australia"],"raw_orcid":"https://orcid.org/0009-0008-8032-3841","affiliations":[{"raw_affiliation_string":"Huazhong University of Science and Technology, Wuhan, China","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"UNSW, Sydney, Australia","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100330541","display_name":"Yi Liu","orcid":"https://orcid.org/0000-0002-4978-127X"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Yi Liu","raw_affiliation_strings":["Nanyang Technological University, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0002-4978-127X","affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore, Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029243457","display_name":"Gelei Deng","orcid":"https://orcid.org/0000-0002-0046-6674"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Gelei Deng","raw_affiliation_strings":["Nanyang Technological University, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0002-0046-6674","affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore, Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076091595","display_name":"Ying Zhang","orcid":"https://orcid.org/0000-0002-2770-9189"},"institutions":[{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Ying Zhang","raw_affiliation_strings":["Virginia Tech, Blacksberg, USA"],"raw_orcid":"https://orcid.org/0000-0002-2770-9189","affiliations":[{"raw_affiliation_string":"Virginia Tech, Blacksberg, USA","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101980974","display_name":"Wenjia Song","orcid":"https://orcid.org/0009-0002-9597-3587"},"institutions":[{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Wenjia Song","raw_affiliation_strings":["Virginia Tech, Blacksberg, USA"],"raw_orcid":"https://orcid.org/0009-0002-9597-3587","affiliations":[{"raw_affiliation_string":"Virginia Tech, Blacksberg, USA","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085767410","display_name":"Ling Shi","orcid":"https://orcid.org/0000-0002-2023-0247"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Ling Shi","raw_affiliation_strings":["Nanyang Technological University, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0002-2023-0247","affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore, Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5000432413","display_name":"Kailong Wang","orcid":"https://orcid.org/0000-0002-3977-6573"},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kailong Wang","raw_affiliation_strings":["Huazhong University of Science and Technology, Wuhan, China"],"raw_orcid":"https://orcid.org/0000-0002-3977-6573","affiliations":[{"raw_affiliation_string":"Huazhong University of Science and Technology, Wuhan, China","institution_ids":["https://openalex.org/I47720641"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5068561556","display_name":"Yuekang Li","orcid":"https://orcid.org/0000-0003-4382-0757"},"institutions":[{"id":"https://openalex.org/I31746571","display_name":"UNSW Sydney","ror":"https://ror.org/03r8z3t63","country_code":"AU","type":"education","lineage":["https://openalex.org/I31746571"]},{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["AU","CN"],"is_corresponding":false,"raw_author_name":"Yuekang Li","raw_affiliation_strings":["Huazhong University of Science and Technology, Wuhan, China","UNSW, Sydney, Australia"],"raw_orcid":"https://orcid.org/0000-0003-4382-0757","affiliations":[{"raw_affiliation_string":"Huazhong University of Science and Technology, Wuhan, China","institution_ids":["https://openalex.org/I47720641"]},{"raw_affiliation_string":"UNSW, Sydney, Australia","institution_ids":["https://openalex.org/I31746571"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100355692","display_name":"Yang Liu","orcid":"https://orcid.org/0000-0001-7300-9215"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Yang Liu","raw_affiliation_strings":["Nanyang Technological University, Singapore, Singapore"],"raw_orcid":"https://orcid.org/0000-0001-7300-9215","affiliations":[{"raw_affiliation_string":"Nanyang Technological University, Singapore, Singapore","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5115695530","display_name":"Haoyu Wang","orcid":"https://orcid.org/0000-0003-1100-8633"},"institutions":[{"id":"https://openalex.org/I47720641","display_name":"Huazhong University of Science and Technology","ror":"https://ror.org/00p991c53","country_code":"CN","type":"education","lineage":["https://openalex.org/I47720641"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Haoyu Wang","raw_affiliation_strings":["Huazhong University of Science and Technology, Wuhan, China"],"raw_orcid":"https://orcid.org/0000-0003-1100-8633","affiliations":[{"raw_affiliation_string":"Huazhong University of Science and Technology, Wuhan, China","institution_ids":["https://openalex.org/I47720641"]}]}],"institutions":[],"countries_distinct_count":4,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5020034653"],"corresponding_institution_ids":["https://openalex.org/I31746571","https://openalex.org/I47720641"],"apc_list":null,"apc_paid":null,"fwci":3.7763,"has_fulltext":true,"cited_by_count":12,"citation_normalized_percentile":{"value":0.93965525,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":"1","issue":"FSE","first_page":"2075","last_page":"2097"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9990000128746033,"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":0.9990000128746033,"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":0.9832000136375427,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.958299994468689,"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/computer-science","display_name":"Computer science","score":0.6856652498245239},{"id":"https://openalex.org/keywords/glitch","display_name":"Glitch","score":0.6208652257919312},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5077637434005737},{"id":"https://openalex.org/keywords/cluster-analysis","display_name":"Cluster analysis","score":0.49783968925476074},{"id":"https://openalex.org/keywords/categorization","display_name":"Categorization","score":0.4892854392528534},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.35723066329956055},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3498676121234894},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.22190740704536438},{"id":"https://openalex.org/keywords/telecommunications","display_name":"Telecommunications","score":0.07355207204818726}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6856652498245239},{"id":"https://openalex.org/C191287063","wikidata":"https://www.wikidata.org/wiki/Q543281","display_name":"Glitch","level":3,"score":0.6208652257919312},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5077637434005737},{"id":"https://openalex.org/C73555534","wikidata":"https://www.wikidata.org/wiki/Q622825","display_name":"Cluster analysis","level":2,"score":0.49783968925476074},{"id":"https://openalex.org/C94124525","wikidata":"https://www.wikidata.org/wiki/Q912550","display_name":"Categorization","level":2,"score":0.4892854392528534},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35723066329956055},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3498676121234894},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.22190740704536438},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.07355207204818726},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.0}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3660799","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3660799","pdf_url":null,"source":{"id":"https://openalex.org/S4404663975","display_name":"Proceedings of the ACM on software engineering.","issn_l":"2994-970X","issn":["2994-970X"],"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/P4310319798","host_organization_name":"Association for Computing Machinery","host_organization_lineage":["https://openalex.org/P4310319798"],"host_organization_lineage_names":["Association for Computing Machinery"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the ACM on Software Engineering","raw_type":"journal-article"},{"id":"pmh:oai:vtechworks.lib.vt.edu:10919/120871","is_oa":true,"landing_page_url":"https://hdl.handle.net/10919/120871","pdf_url":"https://vtechworks.lib.vt.edu/bitstreams/6fc8fe0b-ce4a-4504-b983-5ba3127f7a1d/download","source":{"id":"https://openalex.org/S4306400248","display_name":"VTechWorks (Virginia Tech)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I859038795","host_organization_name":"Virginia Tech","host_organization_lineage":["https://openalex.org/I859038795"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"Article - Refereed"}],"best_oa_location":{"id":"pmh:oai:vtechworks.lib.vt.edu:10919/120871","is_oa":true,"landing_page_url":"https://hdl.handle.net/10919/120871","pdf_url":"https://vtechworks.lib.vt.edu/bitstreams/6fc8fe0b-ce4a-4504-b983-5ba3127f7a1d/download","source":{"id":"https://openalex.org/S4306400248","display_name":"VTechWorks (Virginia Tech)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I859038795","host_organization_name":"Virginia Tech","host_organization_lineage":["https://openalex.org/I859038795"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"","raw_type":"Article - Refereed"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G3481324499","display_name":null,"funder_award_id":"62302176","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4400582517.pdf","grobid_xml":"https://content.openalex.org/works/W4400582517.grobid-xml"},"referenced_works_count":15,"referenced_works":["https://openalex.org/W1997874197","https://openalex.org/W3100869665","https://openalex.org/W4205166846","https://openalex.org/W4240433220","https://openalex.org/W4313442384","https://openalex.org/W4317651364","https://openalex.org/W4319460874","https://openalex.org/W4384155751","https://openalex.org/W4384347367","https://openalex.org/W4385572273","https://openalex.org/W4387872955","https://openalex.org/W4388505012","https://openalex.org/W4399567237","https://openalex.org/W4400484590","https://openalex.org/W4402670895"],"related_works":["https://openalex.org/W2140328632","https://openalex.org/W2541225405","https://openalex.org/W2149184458","https://openalex.org/W4316660008","https://openalex.org/W2373956461","https://openalex.org/W1971856568","https://openalex.org/W1482101251","https://openalex.org/W1505247109","https://openalex.org/W2898595008","https://openalex.org/W2807479533"],"abstract_inverted_index":{"With":[0],"the":[1,32,49,77,102,135,142],"expanding":[2],"application":[3],"of":[4,34,52,70,76,137],"Large":[5],"Language":[6],"Models":[7],"(LLMs)":[8],"in":[9,101,160],"various":[10],"domains,":[11],"it":[12],"becomes":[13],"imperative":[14],"to":[15,99],"comprehensively":[16],"investigate":[17],"their":[18],"unforeseen":[19],"behaviors":[20],"and":[21,29,45,66,81],"consequent":[22],"outcomes.":[23],"In":[24],"this":[25],"study,":[26],"we":[27,55,105,140],"introduce":[28],"systematically":[30],"explore":[31],"phenomenon":[33],"\u201cglitch":[35],"tokens\u201d,":[36],"which":[37],"are":[38],"anomalous":[39],"tokens":[40,80,97],"produced":[41],"by":[42,84],"established":[43],"tokenizers":[44,65],"could":[46],"potentially":[47],"compromise":[48],"models\u2019":[50],"quality":[51],"response.":[53],"Specifically,":[54],"experiment":[56],"on":[57,92,130,146],"seven":[58],"top":[59],"popular":[60],"LLMs":[61,85],"utilizing":[62],"three":[63,127],"distinct":[64],"involving":[67],"a":[68,109],"totally":[69],"182,517":[71],"tokens.":[72,90,148],"We":[73],"present":[74,141],"categorizations":[75],"identified":[78],"glitch":[79,89,96,116,147],"symptoms":[82],"exhibited":[83],"when":[86],"interacting":[87],"with":[88],"Based":[91],"our":[93,123,138],"observation":[94],"that":[95,122],"tend":[98],"cluster":[100],"embedding":[103],"space,":[104],"propose":[106],"GlitchHunter":[107],",":[108],"novel":[110],"iterative":[111],"clustering-based":[112],"technique,":[113],"for":[114],"efficient":[115],"token":[117],"detection.":[118],"The":[119],"evaluation":[120],"shows":[121],"approach":[124],"notably":[125],"outperforms":[126],"baseline":[128],"methods":[129],"eight":[131],"open-source":[132],"LLMs.":[133,161],"To":[134],"best":[136],"knowledge,":[139],"first":[143],"comprehensive":[144],"study":[145],"Our":[149],"new":[150],"detection":[151],"further":[152],"provides":[153],"valuable":[154],"insights":[155],"into":[156],"mitigating":[157],"tokenization-related":[158],"errors":[159]},"counts_by_year":[{"year":2026,"cited_by_count":3},{"year":2025,"cited_by_count":4},{"year":2024,"cited_by_count":5}],"updated_date":"2026-05-21T06:26:12.895304","created_date":"2025-10-10T00:00:00"}
