{"id":"https://openalex.org/W7159061995","doi":"https://doi.org/10.48550/arxiv.2604.26506","title":"SafeReview: Defending LLM-based Review Systems Against Adversarial Hidden Prompts","display_name":"SafeReview: Defending LLM-based Review Systems Against Adversarial Hidden Prompts","publication_year":2026,"publication_date":"2026-04-29","ids":{"openalex":"https://openalex.org/W7159061995","doi":"https://doi.org/10.48550/arxiv.2604.26506"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.26506","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26506","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.26506","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5121770329","display_name":"Yuan Xin","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Xin, Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134907228","display_name":"Yixuan Weng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Weng, Yixuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134915252","display_name":"Minjun Zhu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhu, Minjun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134928756","display_name":"Ying Ling","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ling, Ying","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134917945","display_name":"Chengwei Qin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Qin, Chengwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134894231","display_name":"Michael Hahn","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Backes, Michael","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134883717","display_name":"Michael Backes","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yue","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134896632","display_name":"Yue Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Linyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":8,"corresponding_author_ids":["https://openalex.org/A5121770329"],"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":0.45249998569488525,"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":0.45249998569488525,"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":0.09969999641180038,"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":0.07199999690055847,"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/adversarial-system","display_name":"Adversarial system","score":0.9603000283241272},{"id":"https://openalex.org/keywords/vulnerability","display_name":"Vulnerability (computing)","score":0.6060000061988831},{"id":"https://openalex.org/keywords/forcing","display_name":"Forcing (mathematics)","score":0.5978999733924866},{"id":"https://openalex.org/keywords/resilience","display_name":"Resilience (materials science)","score":0.5584999918937683},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.49079999327659607},{"id":"https://openalex.org/keywords/generator","display_name":"Generator (circuit theory)","score":0.48170000314712524},{"id":"https://openalex.org/keywords/generative-grammar","display_name":"Generative grammar","score":0.47920000553131104}],"concepts":[{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.9603000283241272},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7128999829292297},{"id":"https://openalex.org/C95713431","wikidata":"https://www.wikidata.org/wiki/Q631425","display_name":"Vulnerability (computing)","level":2,"score":0.6060000061988831},{"id":"https://openalex.org/C197115733","wikidata":"https://www.wikidata.org/wiki/Q1003136","display_name":"Forcing (mathematics)","level":2,"score":0.5978999733924866},{"id":"https://openalex.org/C2779585090","wikidata":"https://www.wikidata.org/wiki/Q3457762","display_name":"Resilience (materials science)","level":2,"score":0.5584999918937683},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.5546000003814697},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.49079999327659607},{"id":"https://openalex.org/C2780992000","wikidata":"https://www.wikidata.org/wiki/Q17016113","display_name":"Generator (circuit theory)","level":3,"score":0.48170000314712524},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.47920000553131104},{"id":"https://openalex.org/C2780966255","wikidata":"https://www.wikidata.org/wiki/Q5474306","display_name":"Foundation (evidence)","level":2,"score":0.415800005197525},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.36800000071525574},{"id":"https://openalex.org/C137176749","wikidata":"https://www.wikidata.org/wiki/Q4105337","display_name":"Psychological resilience","level":2,"score":0.31130000948905945},{"id":"https://openalex.org/C2988773926","wikidata":"https://www.wikidata.org/wiki/Q25104379","display_name":"Generative adversarial network","level":3,"score":0.28200000524520874},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.27970001101493835},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C108827166","wikidata":"https://www.wikidata.org/wiki/Q175975","display_name":"Internet privacy","level":1,"score":0.27250000834465027},{"id":"https://openalex.org/C140547941","wikidata":"https://www.wikidata.org/wiki/Q7797194","display_name":"Threat model","level":2,"score":0.2685000002384186},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.26339998841285706},{"id":"https://openalex.org/C167063184","wikidata":"https://www.wikidata.org/wiki/Q1400839","display_name":"Vulnerability assessment","level":3,"score":0.2621000111103058}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.26506","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26506","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.26506","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.26506","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"As":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"are":[5],"increasingly":[6,79],"integrated":[7],"into":[8],"academic":[9],"peer":[10,45,141],"review,":[11],"their":[12],"vulnerability":[13],"to":[14,24,31,57,66,77,91],"adversarial":[15,19,39,71],"hidden":[16],"prompts,":[17],"i.e.,":[18],"instructions":[20],"embedded":[21],"in":[22],"submissions":[23],"manipulate":[25],"outcomes,":[26],"poses":[27],"a":[28,37,54,63,136],"critical":[29],"threat":[30],"scholarly":[32],"integrity.":[33],"We":[34],"propose":[35],"SafeReview,":[36],"co-evolutionary":[38,133],"training":[40,90,134],"framework":[41],"for":[42,138],"defending":[43],"LLM-based":[44],"review":[46,68],"systems":[47],"against":[48,107],"such":[49],"attacks.":[50],"SafeReview":[51,104],"jointly":[52],"trains":[53],"Generator":[55,74],"model":[56,65],"create":[58],"sophisticated":[59],"attack":[60],"prompts":[61],"and":[62,97,118],"Defender":[64,85],"preserve":[67],"integrity":[69],"under":[70,116],"manipulation.":[72],"The":[73],"is":[75,86],"optimized":[76],"produce":[78],"effective":[80],"prompt":[81,109],"injections,":[82],"while":[83],"the":[84,130],"strengthened":[87],"through":[88],"preference-based":[89],"maintain":[92],"consistent":[93],"reviews":[94],"between":[95],"clean":[96],"attacked":[98],"submissions.":[99],"Experimental":[100],"results":[101,128],"show":[102],"that":[103],"improves":[105],"robustness":[106],"adaptive":[108],"injection":[110],"attacks,":[111],"better":[112],"preserves":[113],"paper":[114],"ranking":[115],"attack,":[117],"generalizes":[119],"across":[120],"attacker":[121],"architectures":[122],"compared":[123],"with":[124],"static":[125],"defenses.":[126],"These":[127],"demonstrate":[129],"potential":[131],"of":[132],"as":[135],"foundation":[137],"securing":[139],"LLM-assisted":[140],"review.":[142]},"counts_by_year":[],"updated_date":"2026-05-30T06:14:24.967023","created_date":"2026-05-01T00:00:00"}