{"id":"https://openalex.org/W7214025871","doi":"https://doi.org/10.18653/2026.findings-emnlp.1451","title":"Seeing Through Conflicts: Improving Instruction Hierarchy Alignment in Vision-Language Models","display_name":"Seeing Through Conflicts: Improving Instruction Hierarchy Alignment in Vision-Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7214025871","doi":"https://doi.org/10.18653/2026.findings-emnlp.1451"},"language":"en","primary_location":{"id":"doi:10.18653/2026.findings-emnlp.1451","doi":"https://doi.org/10.18653/2026.findings-emnlp.1451","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1451","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1451.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"bronze","oa_url":"https://aclanthology.org/2026.findings-emnlp.1451.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5120371561","display_name":"Nicholas Sansoterra","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Nicholas Sansoterra","raw_affiliation_strings":["The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5152456922","display_name":"Zishuo Zheng","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Zishuo Zheng","raw_affiliation_strings":["The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5152473152","display_name":"Sachin Kumar","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I52357470","display_name":"The Ohio State University","ror":"https://ror.org/00rs6vg23","country_code":"US","type":"education","lineage":["https://openalex.org/I52357470"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sachin Kumar","raw_affiliation_strings":["The Ohio State University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Ohio State University","institution_ids":["https://openalex.org/I52357470"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I52357470"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.76511185,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"30912","last_page":"30927"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.7240999937057495,"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":0.7240999937057495,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.26489999890327454,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10181","display_name":"Natural Language Processing Techniques","score":0.0010999999940395355,"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/vision-language-models","display_name":"vision-language models","score":0.996999979019165},{"id":"https://openalex.org/keywords/adversarial-robustness","display_name":"adversarial robustness","score":0.6769999861717224}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8064000010490417},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.739799976348877},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5322999954223633},{"id":"https://openalex.org/C31170391","wikidata":"https://www.wikidata.org/wiki/Q188619","display_name":"Hierarchy","level":2,"score":0.5184999704360962},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.43619999289512634},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.4230000078678131},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.41290000081062317},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.4115000069141388},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.35920000076293945},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3125999867916107},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.30790001153945923},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.2824000120162964},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.2784999907016754},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.25440001487731934}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.18653/2026.findings-emnlp.1451","doi":"https://doi.org/10.18653/2026.findings-emnlp.1451","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1451","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1451.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2609.22234","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.22234","pdf_url":"https://arxiv.org/pdf/2609.22234","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.2609.22234","doi":"https://doi.org/10.48550/arxiv.2609.22234","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.22234","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":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.18653/2026.findings-emnlp.1451","doi":"https://doi.org/10.18653/2026.findings-emnlp.1451","is_oa":true,"landing_page_url":"https://doi.org/10.18653/2026.findings-emnlp.1451","pdf_url":"https://aclanthology.org/2026.findings-emnlp.1451.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7214025871.pdf","grobid_xml":"https://content.openalex.org/works/W7214025871.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4407683663","https://openalex.org/W7211900944","https://openalex.org/W7133561896","https://openalex.org/W4389260586","https://openalex.org/W4416047691","https://openalex.org/W7221043152","https://openalex.org/W4398795714","https://openalex.org/W4403443878","https://openalex.org/W4386240819","https://openalex.org/W4415947874"],"abstract_inverted_index":{"Instruction":[0],"hierarchy":[1],"(IH)":[2],"alignment":[3,45],"teaches":[4],"language":[5],"models":[6,20,77],"to":[7,72,108],"prioritize":[8],"higher-level":[9],"instructions":[10,27,84],"when":[11,76],"inputs":[12],"conflict.While":[13],"studied":[14],"primarily":[15],"in":[16,31],"text-only":[17,67],"settings,":[18],"vision-language":[19],"(VLMs)":[21],"introduce":[22],"new":[23],"challenges":[24],"for":[25],"IH:":[26],"may":[28],"be":[29],"embedded":[30],"images,":[32],"split":[33],"across":[34,85],"modalities,":[35],"visually":[36],"transformed,":[37],"or":[38,81],"encountered":[39],"during":[40],"agentic":[41],"tasks.Positing":[42],"multimodal":[43,73,118],"IH":[44,68],"as":[46],"a":[47],"reasoning":[48],"problem,":[49],"we":[50],"train":[51],"VLMs":[52],"using":[53],"reinforcement":[54],"learning":[55],"with":[56],"rule-based":[57],"rewards,":[58],"comparing":[59],"text-only,":[60],"image-only,":[61],"and":[62,110,133],"mixed-modality":[63,94],"supervision.We":[64],"find":[65],"that":[66,121],"training":[69,87,95,106],"partially":[70],"transfers":[71],"attacks,":[74],"failing":[75],"must":[78],"decode,":[79],"reconstruct,":[80],"reason":[82],"over":[83],"modalities.Imagebased":[86],"improves":[88],"robustness":[89,129],"beyond":[90,102],"textonly":[91],"supervision,":[92],"while":[93,114],"performs":[96],"best":[97],"overall.Importantly,":[98],"the":[99,103],"benefits":[100],"generalize":[101],"synthetic":[104],"typographic":[105],"setting":[107],"real-image":[109],"web-agent":[111],"safety":[112],"tasks,":[113],"largely":[115],"preserving":[116],"general":[117],"capability,":[119],"showing":[120],"lightweight,":[122],"verifiable":[123],"supervision":[124],"can":[125],"meaningfully":[126],"improve":[127],"VLM":[128],"under":[130],"adversarial,":[131],"cross-modal,":[132],"interactive":[134],"instruction":[135],"conflicts.1":[136]},"counts_by_year":[],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2026-09-23T00:00:00"}