{"id":"https://openalex.org/W7220447351","doi":"https://doi.org/10.48550/arxiv.2610.02302","title":"Intent-Hiding Jailbreaks: An Information-Theoretic Framework for Compositional Attacks","display_name":"Intent-Hiding Jailbreaks: An Information-Theoretic Framework for Compositional Attacks","publication_year":2026,"publication_date":"2026-10-01","ids":{"openalex":"https://openalex.org/W7220447351","doi":"https://doi.org/10.48550/arxiv.2610.02302"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.02302","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.02302","pdf_url":"https://arxiv.org/pdf/2610.02302","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2610.02302","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5136487598","display_name":"Fengwei Tian","orcid":"https://orcid.org/0009-0007-9850-0028","observed_orcids":["https://orcid.org/0009-0007-9850-0028"]},"institutions":[{"id":"https://openalex.org/I138006243","display_name":"University of Arizona","ror":"https://ror.org/03m2x1q45","country_code":"US","type":"education","lineage":["https://openalex.org/I138006243"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tian, Fengwei","raw_affiliation_strings":["School of Electrical, Computing and Software Engineering , University of Arizona , Tucson , AZ 85721 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical, Computing and Software Engineering , University of Arizona , Tucson , AZ 85721 , USA","institution_ids":["https://openalex.org/I138006243"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5004316408","display_name":"Ravi Tandon","orcid":"https://orcid.org/0000-0002-6182-6098","observed_orcids":["https://orcid.org/0000-0002-6182-6098"]},"institutions":[{"id":"https://openalex.org/I138006243","display_name":"University of Arizona","ror":"https://ror.org/03m2x1q45","country_code":"US","type":"education","lineage":["https://openalex.org/I138006243"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Tandon, Ravi","raw_affiliation_strings":["School of Electrical, Computing and Software Engineering , University of Arizona , Tucson , AZ 85721 , USA"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"School of Electrical, Computing and Software Engineering , University of Arizona , Tucson , AZ 85721 , USA","institution_ids":["https://openalex.org/I138006243"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I138006243"],"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":0.9990000128746033,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","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/T11424","display_name":"Security and Verification in Computing","score":9.999999747378752e-05,"subfield":{"id":"https://openalex.org/subfields/1708","display_name":"Hardware and Architecture"},"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":9.999999747378752e-05,"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/jailbreak-attacks","display_name":"jailbreak attacks","score":0.8610000014305115},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.9549999833106995},{"id":"https://openalex.org/keywords/information-theoretic-security","display_name":"information-theoretic security","score":0.671999990940094},{"id":"https://openalex.org/keywords/query-construction","display_name":"query construction","score":0.8489999771118164},{"id":"https://openalex.org/keywords/computational-hardness","display_name":"computational hardness","score":0.8799999952316284},{"id":"https://openalex.org/keywords/safety-thresholds","display_name":"safety thresholds","score":0.49900001287460327}],"concepts":[{"id":"https://openalex.org/C2778134712","wikidata":"https://www.wikidata.org/wiki/Q1047307","display_name":"Bundle","level":2,"score":0.7211999893188477},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.7156000137329102},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.5928000211715698},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5917999744415283},{"id":"https://openalex.org/C2776036281","wikidata":"https://www.wikidata.org/wiki/Q48769818","display_name":"Constraint (computer-aided design)","level":2,"score":0.5706999897956848},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5473999977111816},{"id":"https://openalex.org/C2775941552","wikidata":"https://www.wikidata.org/wiki/Q25212305","display_name":"Isolation (microbiology)","level":2,"score":0.39309999346733093},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.35370001196861267},{"id":"https://openalex.org/C2779662365","wikidata":"https://www.wikidata.org/wiki/Q5416694","display_name":"Event (particle physics)","level":2,"score":0.32670000195503235},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.30090001225471497},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2870999872684479},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.27619999647140503},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.27140000462532043},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.2671000063419342},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.25999999046325684},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.2529999911785126}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.02302","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.02302","pdf_url":"https://arxiv.org/pdf/2610.02302","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.02302","doi":"https://doi.org/10.48550/arxiv.2610.02302","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.02302","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.02302","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.02302","pdf_url":"https://arxiv.org/pdf/2610.02302","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G7215000263","display_name":"EAGER: DCL: SaTC: Enabling Interdisciplinary Collab: Impact-aware Machine Learning for Fair and Private Decision Making: Algorithms and Applications in Juvenile Justice Systems","funder_award_id":"2209951","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G7789950527","display_name":"CIF: Small: Learning To Correct Errors","funder_award_id":"2100013","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G8063809372","display_name":"Collaborative Research: SaTC: CORE: Medium: Graph Mining and Network Science with Differential Privacy: Efficient Algorithms and Fundamental Limits","funder_award_id":"2317192","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320306084","display_name":"U.S. Department of Energy","ror":"https://ror.org/01bj3aw27"},{"id":"https://openalex.org/F4320332161","display_name":"National Institutes of Health","ror":"https://ror.org/01cwqze88"},{"id":"https://openalex.org/F4320332359","display_name":"Office of Science","ror":"https://ror.org/00mmn6b08"},{"id":"https://openalex.org/F4320337506","display_name":"Advanced Scientific Computing Research","ror":"https://ror.org/0012c7r22"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7220447351.pdf","grobid_xml":"https://content.openalex.org/works/W7220447351.grobid-xml"},"referenced_works_count":11,"referenced_works":["https://openalex.org/W4283170666","https://openalex.org/W4385374425","https://openalex.org/W4389713766","https://openalex.org/W4401306886","https://openalex.org/W4403883848","https://openalex.org/W4407506574","https://openalex.org/W4414939355","https://openalex.org/W4414946721","https://openalex.org/W4415797606","https://openalex.org/W6947991961","https://openalex.org/W7160317400"],"related_works":["https://openalex.org/W7076365198","https://openalex.org/W7161970319","https://openalex.org/W4405219474","https://openalex.org/W4414602076","https://openalex.org/W4392425370","https://openalex.org/W7166138963","https://openalex.org/W4412944475","https://openalex.org/W7119531022","https://openalex.org/W7131890740","https://openalex.org/W7220686358"],"abstract_inverted_index":{"Recent":[0],"work":[1],"has":[2],"shown":[3],"that":[4,99,126,158,233],"large":[5],"language":[6],"models":[7],"(LLMs)":[8],"can":[9,236],"be":[10,150],"vulnerable":[11],"to":[12,146,261],"jailbreak":[13,213],"attacks":[14],"in":[15,30,119,128,152],"which":[16,103],"harmful":[17,27,80,116],"intent":[18,111,200],"is":[19,132,166],"obscured":[20],"through":[21],"composition":[22],"with":[23,60],"benign":[24,43],"tasks.":[25],"A":[26],"request":[28],"refused":[29],"isolation":[31],"may":[32],"elicit":[33,237],"a":[34,40,86,163,183,251],"different":[35],"response":[36],"when":[37],"embedded":[38],"within":[39],"larger,":[41],"seemingly":[42],"query.":[44,155,210],"We":[45,122,156,211],"study":[46,123],"these":[47,100],"compositional":[48,234],"intent-hiding":[49],"jailbreaks":[50],"from":[51],"an":[52,61,170],"information-theoretic":[53],"perspective.":[54],"Our":[55],"formulation":[56],"associates":[57],"each":[58],"task":[59,73,191],"estimated":[62,110],"probability":[63,78],"of":[64,79,134,217],"being":[65],"judged":[66],"harmful:":[67],"the":[68,71,76,83,90,93,109,115,120,135,138,153,179,188,208,218,241,245],"average":[69,84],"over":[70,85],"full":[72],"collection":[74],"defines":[75,92],"prior":[77],"intent,":[81],"while":[82,249],"selected":[87,143],"bundle":[88,181,222,254],"containing":[89],"target":[91,117,203,219,238,258],"posterior.":[94],"Selecting":[95],"auxiliary":[96],"tasks":[97,141],"so":[98],"averages":[101],"agree,":[102],"we":[104],"call":[105],"prior-posterior":[106,160],"matching,":[107],"leaves":[108],"unchanged":[112],"even":[113],"though":[114],"remains":[118],"bundle.":[121],"two":[124],"settings":[125],"differ":[127],"whether":[129],"query":[130,194,224],"construction":[131,195],"part":[133],"optimization.":[136],"In":[137,187],"query-independent":[139],"setting,":[140,190],"are":[142,196,205],"without":[144],"regard":[145],"how":[147],"they":[148],"will":[149],"expressed":[151],"final":[154],"show":[157,232],"exact":[159],"matching":[161],"under":[162,244],"bundle-size":[164],"constraint":[165],"computationally":[167],"hard,":[168],"derive":[169],"optimal":[171],"water-filling":[172],"solution":[173],"for":[174,263],"fractional":[175],"weights,":[176],"and":[177,193,199,202,215,226],"characterize":[178],"smallest":[180],"satisfying":[182],"prescribed":[184],"safety":[185],"threshold.":[186],"query-dependent":[189],"selection":[192],"considered":[197],"jointly,":[198],"concealment":[201],"preservation":[204,216,259],"evaluated":[206,246],"on":[207],"resulting":[209],"evaluate":[212],"effectiveness":[214],"behavior":[220],"across":[221],"sizes,":[223],"generators,":[225],"several":[227,264],"open-source":[228],"models.":[229,265],"These":[230],"results":[231],"queries":[235],"behaviors":[239],"beyond":[240],"direct-request":[242],"baseline":[243],"search":[247],"budgets,":[248],"revealing":[250],"trade-off:":[252],"as":[253],"size":[255],"increases,":[256],"response-level":[257],"tends":[260],"decrease":[262]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-06T00:00:00"}