{"id":"https://openalex.org/W7204198567","doi":"https://doi.org/10.48550/arxiv.2608.21985","title":"Redteaming Leading Arabic LLMs with ASAS","display_name":"Redteaming Leading Arabic LLMs with ASAS","publication_year":2026,"publication_date":"2026-08-22","ids":{"openalex":"https://openalex.org/W7204198567","doi":"https://doi.org/10.48550/arxiv.2608.21985"},"language":null,"primary_location":{"id":"pmh:oai:arXiv.org:2608.21985","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21985","pdf_url":"https://arxiv.org/pdf/2608.21985","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.21985","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5013671492","display_name":"Fidaa Abed","orcid":"https://orcid.org/0000-0002-9127-8715","observed_orcids":["https://orcid.org/0000-0002-9127-8715"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abed, Fidaa","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109744163","display_name":"Haidar Khan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khan, Haidar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006686579","display_name":"M Saiful Bari","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bari, M Saiful","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Khan, Babar","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Khan, Babar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5014648332","display_name":"Abdalghani Abujabal","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Abujabal, Abdalghani","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":true,"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":0.9940999746322632,"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":0.9940999746322632,"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":0.002400000113993883,"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":0.0012000000569969416,"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/llm-safety","display_name":"LLM safety","score":0.8840000033378601},{"id":"https://openalex.org/keywords/adversarial-evaluation","display_name":"adversarial evaluation","score":0.7409999966621399},{"id":"https://openalex.org/keywords/safety-benchmark","display_name":"safety benchmark","score":0.7020000219345093},{"id":"https://openalex.org/keywords/modern-standard-arabic","display_name":"Modern Standard Arabic","score":0.5759999752044678},{"id":"https://openalex.org/keywords/cultural-alignment","display_name":"cultural alignment","score":0.4950000047683716},{"id":"https://openalex.org/keywords/gpt-4o","display_name":"GPT-4o","score":0.7269999980926514},{"id":"https://openalex.org/keywords/claude-3-7-sonnet","display_name":"Claude 3.7 Sonnet","score":0.9139999747276306}],"concepts":[{"id":"https://openalex.org/C96455323","wikidata":"https://www.wikidata.org/wiki/Q13955","display_name":"Arabic","level":2,"score":0.8192999958992004},{"id":"https://openalex.org/C2778243841","wikidata":"https://www.wikidata.org/wiki/Q56467","display_name":"Modern Standard Arabic","level":3,"score":0.6022999882698059},{"id":"https://openalex.org/C37736160","wikidata":"https://www.wikidata.org/wiki/Q1801315","display_name":"Adversarial system","level":2,"score":0.535099983215332},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5189999938011169},{"id":"https://openalex.org/C2780385302","wikidata":"https://www.wikidata.org/wiki/Q367158","display_name":"Protocol (science)","level":3,"score":0.4941999912261963},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.44440001249313354},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.37950000166893005},{"id":"https://openalex.org/C2777382242","wikidata":"https://www.wikidata.org/wiki/Q6017816","display_name":"Index (typography)","level":2,"score":0.3422999978065491},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.3084000051021576},{"id":"https://openalex.org/C17744445","wikidata":"https://www.wikidata.org/wiki/Q36442","display_name":"Political science","level":0,"score":0.2946999967098236},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.2892000079154968},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.28110000491142273},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.2768000066280365},{"id":"https://openalex.org/C125496544","wikidata":"https://www.wikidata.org/wiki/Q1422423","display_name":"Arabic languages","level":3,"score":0.2700999975204468},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.26600000262260437}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.21985","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21985","pdf_url":"https://arxiv.org/pdf/2608.21985","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.21985","doi":"https://doi.org/10.48550/arxiv.2608.21985","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.21985","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,"listed_in":[],"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":"pmh:oai:arXiv.org:2608.21985","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.21985","pdf_url":"https://arxiv.org/pdf/2608.21985","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,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"sustainable_development_goals_aurora":[{"display_name":"Quality education","score":0.6593066453933716,"id":"https://metadata.un.org/sdg/4"}],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204198567.pdf","grobid_xml":"https://content.openalex.org/works/W7204198567.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"As":[0],"the":[1,34,39],"adoption":[2],"of":[3,112],"large":[4],"language":[5,143],"models":[6,76,87,106],"(LLMs)":[7],"grows":[8],"in":[9,28,63,121,176],"Arabic-speaking":[10],"regions,":[11],"ensuring":[12],"their":[13],"safety":[14,24,54,101,119,154],"and":[15,56,85,91,127,132,151,170],"cultural":[16],"alignment":[17,144],"is":[18],"increasingly":[19],"critical.":[20],"However,":[21],"Arabic":[22,35,43,66,78,177],"LLM":[23,178],"remains":[25],"underexplored,":[26],"especially":[27],"adversarial":[29],"evaluation":[30,72],"settings.":[31],"We":[32,68],"introduce":[33],"Safety":[36],"Index":[37],"(ASAS),":[38],"first":[40],"fully":[41],"human-curated":[42],"benchmark":[44,169],"for":[45],"redteaming":[46,71,171],"LLMs.":[47],"ASAS":[48,164],"contains":[49],"801":[50],"prompts":[51],"spanning":[52],"8":[53,57],"categories":[55,123],"attack":[58],"strategies,":[59],"with":[60,77,130],"ideal":[61],"responses":[62,96],"Modern":[64],"Standard":[65],"(MSA).":[67],"conduct":[69],"a":[70,98,166],"across":[73,149],"seven":[74],"leading":[75],"capabilities,":[79],"including":[80],"GPT-4o,":[81],"Claude":[82],"3.7":[83],"Sonnet,":[84],"regional":[86],"such":[88,124],"as":[89,125],"ALLaM":[90],"FANAR.":[92],"Human":[93],"annotators":[94],"rate":[95],"using":[97],"structured":[99],"4-point":[100],"scale,":[102],"revealing":[103],"that":[104,142,152],"most":[105,136],"fail":[107],"to":[108,161,173],"defend":[109],"against":[110],"50%":[111],"unsafe":[113],"prompts.":[114],"Our":[115],"findings":[116],"highlight":[117],"major":[118],"gaps":[120],"high-harm":[122],"weapons":[126],"illicit":[128],"substances,":[129],"direct":[131],"obfuscation-based":[133],"attacks":[134],"proving":[135],"effective.":[137],"The":[138],"results":[139],"also":[140],"show":[141],"does":[145],"not":[146],"readily":[147],"transfer":[148],"languages,":[150],"automated":[153],"judges":[155],"(e.g.,":[156],"GPT-4o)":[157],"perform":[158],"poorly":[159],"compared":[160],"human":[162],"annotators.":[163],"provides":[165],"culturally":[167],"grounded":[168],"protocol":[172],"drive":[174],"progress":[175],"safety.":[179]},"counts_by_year":[],"updated_date":"2026-09-29T14:09:16.559479","created_date":"2026-08-26T00:00:00"}