{"id":"https://openalex.org/W7133193664","doi":"https://doi.org/10.48550/arxiv.2602.23636","title":"FlexGuard: Continuous Risk Scoring for Strictness-Adaptive LLM Content Moderation","display_name":"FlexGuard: Continuous Risk Scoring for Strictness-Adaptive LLM Content Moderation","publication_year":2026,"publication_date":"2026-02-27","ids":{"openalex":"https://openalex.org/W7133193664","doi":"https://doi.org/10.48550/arxiv.2602.23636"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.23636","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.23636","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.2602.23636","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114337254","display_name":"Zhihao Ding","orcid":"https://orcid.org/0000-0001-7778-6142"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Ding, Zhihao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121525783","display_name":"Jinming Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jinming","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127864828","display_name":"Ze Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Ze","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5127858487","display_name":"Jieming Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shi, Jieming","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":4,"corresponding_author_ids":["https://openalex.org/A5114337254"],"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.3982999920845032,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.3982999920845032,"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":0.1972000002861023,"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/T11644","display_name":"Spam and Phishing Detection","score":0.0812000036239624,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/moderation","display_name":"Moderation","score":0.8676999807357788},{"id":"https://openalex.org/keywords/consistency","display_name":"Consistency (knowledge bases)","score":0.6241999864578247},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.6100999712944031},{"id":"https://openalex.org/keywords/binary-number","display_name":"Binary number","score":0.5227000117301941},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.48570001125335693},{"id":"https://openalex.org/keywords/enforcement","display_name":"Enforcement","score":0.4415000081062317},{"id":"https://openalex.org/keywords/limiting","display_name":"Limiting","score":0.4366999864578247}],"concepts":[{"id":"https://openalex.org/C93225998","wikidata":"https://www.wikidata.org/wiki/Q1941972","display_name":"Moderation","level":2,"score":0.8676999807357788},{"id":"https://openalex.org/C2776436953","wikidata":"https://www.wikidata.org/wiki/Q5163215","display_name":"Consistency (knowledge bases)","level":2,"score":0.6241999864578247},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.6100999712944031},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6093000173568726},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.5227000117301941},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.48570001125335693},{"id":"https://openalex.org/C2779777834","wikidata":"https://www.wikidata.org/wiki/Q4202277","display_name":"Enforcement","level":2,"score":0.4415000081062317},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.4366999864578247},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.4309999942779541},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3828999996185303},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.36980000138282776},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.36559998989105225},{"id":"https://openalex.org/C182365436","wikidata":"https://www.wikidata.org/wiki/Q50701","display_name":"Variable (mathematics)","level":2,"score":0.33980000019073486},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.32760000228881836},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.28769999742507935},{"id":"https://openalex.org/C2983394010","wikidata":"https://www.wikidata.org/wiki/Q19885330","display_name":"Continuous variable","level":2,"score":0.28029999136924744},{"id":"https://openalex.org/C2780009758","wikidata":"https://www.wikidata.org/wiki/Q6804172","display_name":"Measure (data warehouse)","level":2,"score":0.25609999895095825}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.23636","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.23636","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.2602.23636","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.23636","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":[{"score":0.5769293904304504,"display_name":"Gender equality","id":"https://metadata.un.org/sdg/5"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Ensuring":[0],"the":[1,170],"safety":[2],"of":[3,28],"LLM-generated":[4],"content":[5],"is":[6,38],"essential":[7],"for":[8],"real-world":[9],"deployment.":[10,147],"Most":[11],"existing":[12,82],"guardrail":[13],"models":[14,84],"formulate":[15],"moderation":[16,64,159],"as":[17],"a":[18,25,61,111],"fixed":[19,26],"binary":[20,51],"classification":[21],"task,":[22],"implicitly":[23],"assuming":[24],"definition":[27],"harmfulness.":[29],"In":[30],"practice,":[31],"enforcement":[32],"strictness":[33,72,145],"-":[34,42],"how":[35],"conservatively":[36],"harmfulness":[37],"defined":[39],"and":[40,46,119,135,151,161,173],"enforced":[41],"varies":[43],"across":[44],"platforms":[45],"evolves":[47],"over":[48],"time,":[49],"making":[50],"moderators":[52],"brittle":[53],"under":[54,70,88,94,165],"shifting":[55],"requirements.":[56],"We":[57,125,168],"first":[58],"introduce":[59],"FlexBench,":[60],"strictness-adaptive":[62],"LLM":[63],"benchmark":[65],"that":[66,85,109,155],"enables":[67],"controlled":[68],"evaluation":[69],"multiple":[71],"regimes.":[73],"Experiments":[74,148],"on":[75,149],"FlexBench":[76,150],"reveal":[77],"substantial":[78],"cross-strictness":[79],"inconsistency":[80],"in":[81],"moderators:":[83],"perform":[86],"well":[87],"one":[89],"regime":[90],"can":[91],"degrade":[92],"substantially":[93,162],"others,":[95],"limiting":[96],"their":[97],"practical":[98,137],"usability.":[99],"To":[100],"address":[101],"this,":[102],"we":[103],"propose":[104],"FlexGuard,":[105],"an":[106],"LLM-based":[107],"moderator":[108],"outputs":[110],"calibrated":[112],"continuous":[113],"risk":[114,117],"score":[115],"reflecting":[116],"severity":[118],"supports":[120],"strictness-specific":[121],"decisions":[122],"via":[123,128],"thresholding.":[124],"train":[126],"FlexGuard":[127,156],"risk-alignment":[129],"optimization":[130],"to":[131,141,143,175],"improve":[132],"score-severity":[133],"consistency":[134],"provide":[136],"threshold":[138],"selection":[139],"strategies":[140],"adapt":[142],"target":[144],"at":[146],"public":[152],"benchmarks":[153],"demonstrate":[154],"achieves":[157],"higher":[158],"accuracy":[160],"improved":[163],"robustness":[164],"varying":[166],"strictness.":[167],"release":[169],"source":[171],"code":[172],"data":[174],"support":[176],"reproducibility.":[177]},"counts_by_year":[],"updated_date":"2026-03-03T06:18:10.843953","created_date":"2026-03-03T00:00:00"}