{"id":"https://openalex.org/W7162759913","doi":"https://doi.org/10.48550/arxiv.2605.29816","title":"Harnessing non-adversarial robustness in large language models","display_name":"Harnessing non-adversarial robustness in large language models","publication_year":2026,"publication_date":"2026-05-28","ids":{"openalex":"https://openalex.org/W7162759913","doi":"https://doi.org/10.48550/arxiv.2605.29816"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.29816","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29816","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,"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2605.29816","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137313847","display_name":"Qinghua Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhou, Qinghua","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137344422","display_name":"Ellina Aleshina","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Aleshina, Ellina","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137346513","display_name":"Andrey Lovyagin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lovyagin, Andrey","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137359448","display_name":"Oleg Somov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Somov, Oleg","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137365361","display_name":"Mikhail Seleznyov","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Seleznyov, Mikhail","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137372869","display_name":"Alexander Panchenko","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Panchenko, Alexander","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137317926","display_name":"Ivan Oseledets","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Oseledets, Ivan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137375454","display_name":"Elena Tutubalina","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tutubalina, Elena","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135253578","display_name":"Ivan Y. Tyukin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tyukin, Ivan Y.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4900999963283539,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4900999963283539,"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":0.12240000069141388,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.07729999721050262,"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/debiasing","display_name":"Debiasing","score":0.8751000165939331},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.8550999760627747},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.34869998693466187},{"id":"https://openalex.org/keywords/retraining","display_name":"Retraining","score":0.33320000767707825}],"concepts":[{"id":"https://openalex.org/C2779458634","wikidata":"https://www.wikidata.org/wiki/Q24963715","display_name":"Debiasing","level":2,"score":0.8751000165939331},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.8550999760627747},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7175999879837036},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4074999988079071},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.34869998693466187},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3467999994754791},{"id":"https://openalex.org/C2778712577","wikidata":"https://www.wikidata.org/wiki/Q3505966","display_name":"Retraining","level":2,"score":0.33320000767707825},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.28790000081062317},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.25859999656677246},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.21729999780654907}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.29816","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29816","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,"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":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2605.29816","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.29816","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,"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0,48],"work":[1],"presents":[2],"an":[3],"approach":[4],"for":[5,117,140],"addressing":[6],"the":[7,42,65],"challenge":[8],"of":[9,36,44,64],"robustness":[10,54,86,107,141,152],"in":[11,95],"Large":[12],"Language":[13],"Models":[14],"(LLMs)":[15],"to":[16,55,150],"alterations":[17,58],"and":[18,74,125,130,135,147,153],"potential":[19],"errors":[20],"caused":[21],"by":[22,101],"semantically":[23],"similar":[24],"but":[25],"textually":[26],"different":[27],"prompts.":[28],"Recent":[29],"works":[30],"have":[31],"shown":[32],"that":[33,106,138],"these":[34],"kinds":[35],"prompt":[37,57,158],"variations":[38],"can":[39,52,108],"significantly":[40],"impact":[41],"performance":[43],"LLMs":[45],"on":[46],"tasks.":[47],"central":[49],"question":[50,71],"is:":[51],"LLMs'":[53],"semantically-neutral":[56],"be":[59,109,144],"acquired":[60],"without":[61],"expensive":[62],"retraining":[63],"entire":[66],"model?":[67],"We":[68,119],"address":[69],"this":[70,102],"both":[72,133],"theoretically":[73],"through":[75,132],"experiments.":[76],"Our":[77],"theoretical":[78],"analysis":[79],"reveals":[80],"a":[81,88,112,145],"crucial":[82],"factor":[83],"impacting":[84],"model":[85],"-":[87],"systematic":[89],"expected":[90],"shift":[91],"or":[92],"perturbation-induced":[93],"bias":[94],"neural":[96],"network":[97],"module":[98],"outputs.":[99],"Motivated":[100],"analysis,":[103],"we":[104],"show":[105],"achieved":[110],"via":[111],"simple":[113],"fine-tuning":[114],"process:":[115],"debiasing":[116,123,139],"robustness.":[118],"identify":[120],"conditions":[121],"when":[122,126],"helps":[124],"it":[127],"does":[128],"not,":[129],"demonstrate,":[131],"theory":[134],"extensive":[136],"experiments,":[137],"may":[142],"indeed":[143],"quick":[146],"efficient":[148],"tool":[149],"enhance":[151],"provide":[154],"certification":[155],"against":[156],"random":[157],"perturbations.":[159]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-05-30T00:00:00"}