{"id":"https://openalex.org/W4416033900","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1285","title":"SteerVLM: Robust Model Control through Lightweight Activation Steering for Vision Language Models","display_name":"SteerVLM: Robust Model Control through Lightweight Activation Steering for Vision Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4416033900","doi":"https://doi.org/10.18653/v1/2025.findings-emnlp.1285"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1285","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1285","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1285.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.findings-emnlp.1285.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5114073724","display_name":"Anushka Sivakumar","orcid":null},"institutions":[{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Anushka Sivakumar","raw_affiliation_strings":["Department of Computer Science Virginia Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science Virginia Tech","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5019616787","display_name":"Andrew Zhang","orcid":"https://orcid.org/0000-0002-9432-2793"},"institutions":[{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Andrew Zhang","raw_affiliation_strings":["Department of Computer Science Virginia Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science Virginia Tech","institution_ids":["https://openalex.org/I859038795"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5081902603","display_name":"Zaber Ibn Abdul Hakim","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zaber Ibn Abdul Hakim","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5101461165","display_name":"Christopher Thomas","orcid":"https://orcid.org/0000-0002-3226-396X"},"institutions":[{"id":"https://openalex.org/I859038795","display_name":"Virginia Tech","ror":"https://ror.org/02smfhw86","country_code":"US","type":"education","lineage":["https://openalex.org/I859038795"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Chris Thomas","raw_affiliation_strings":["Department of Computer Science Virginia Tech"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Computer Science Virginia Tech","institution_ids":["https://openalex.org/I859038795"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.36987543,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"23640","last_page":"23665"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.5914999842643738,"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"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.5914999842643738,"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/T12031","display_name":"Speech and dialogue systems","score":0.09130000323057175,"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/T11450","display_name":"Model-Driven Software Engineering Techniques","score":0.035599999129772186,"subfield":{"id":"https://openalex.org/subfields/1712","display_name":"Software"},"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/control","display_name":"Control (management)","score":0.4803999960422516},{"id":"https://openalex.org/keywords/robust-control","display_name":"Robust control","score":0.3635999858379364},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.34529998898506165},{"id":"https://openalex.org/keywords/control-system","display_name":"Control system","score":0.3393999934196472},{"id":"https://openalex.org/keywords/work","display_name":"Work (physics)","score":0.33820000290870667},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.31949999928474426}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5758000016212463},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.4803999960422516},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4077000021934509},{"id":"https://openalex.org/C133731056","wikidata":"https://www.wikidata.org/wiki/Q4917288","display_name":"Control engineering","level":1,"score":0.365200012922287},{"id":"https://openalex.org/C31531917","wikidata":"https://www.wikidata.org/wiki/Q915157","display_name":"Robust control","level":3,"score":0.3635999858379364},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.34529998898506165},{"id":"https://openalex.org/C17500928","wikidata":"https://www.wikidata.org/wiki/Q959968","display_name":"Control system","level":2,"score":0.3393999934196472},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.33820000290870667},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.31949999928474426},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3190000057220459},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.3052000105381012},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.2912999987602234},{"id":"https://openalex.org/C47446073","wikidata":"https://www.wikidata.org/wiki/Q5165890","display_name":"Control theory (sociology)","level":3,"score":0.29010000824928284},{"id":"https://openalex.org/C2777212361","wikidata":"https://www.wikidata.org/wiki/Q5127848","display_name":"Class (philosophy)","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.26499998569488525},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.26489999890327454}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.findings-emnlp.1285","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1285","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1285.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-emnlp.1285","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-emnlp.1285","pdf_url":"https://aclanthology.org/2025.findings-emnlp.1285.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4416033900.pdf","grobid_xml":"https://content.openalex.org/works/W4416033900.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"This":[0],"work":[1],"introduces":[2],"SteerVLM,":[3],"a":[4,111,141],"lightweight":[5],"steering":[6,66,80,92,124,132],"module":[7,67,81],"designed":[8],"to":[9,20,37,72,116],"guide":[10],"Vision":[11],"Language":[12],"Models":[13],"(VLMs)":[14],"towards":[15],"outputs":[16],"that":[17],"better":[18],"adhere":[19],"desired":[21],"instructions.Our":[22],"approach":[23],"learns":[24],"from":[25],"the":[26,42,75,118],"latent":[27],"embeddings":[28],"of":[29,74,101,122],"paired":[30],"prompts":[31],"encoding":[32],"target":[33],"and":[34,89,120,133,139],"converse":[35],"behaviors":[36],"dynamically":[38],"adjust":[39],"activations":[40],"connecting":[41],"language":[43],"modality":[44],"with":[45],"image":[46],"context.This":[47],"provides":[48],"fine-grained,":[49],"inference-time":[50],"control":[51,84,147],"over":[52],"complex":[53],"output":[54],"semantics":[55],"without":[56,93],"modifying":[57],"model":[58,83,146],"weights":[59],"while":[60],"preserving":[61],"performance":[62],"on":[63,131],"off-target":[64],"tasks.Our":[65],"requires":[68],"learning":[69],"parameters":[70],"equal":[71],"0.14%":[73],"original":[76],"VLM's":[77],"size.Additionally,":[78],"our":[79],"gains":[82],"via":[85],"dimension-wise":[86],"activation":[87,149],"modulation":[88],"adaptive":[90],"layer-wise":[91],"requiring":[94],"pre-extracted":[95],"static":[96],"vectors":[97],"or":[98],"manual":[99],"tuning":[100],"intervention":[102,129],"points.Furthermore,":[103],"we":[104],"introduce":[105],"VNIA":[106],"(Visual":[107],"Narrative":[108],"Intent":[109],"Alignment),":[110],"multimodal":[112,145],"dataset":[113],"specifically":[114],"created":[115],"facilitate":[117],"development":[119],"evaluation":[121],"VLM":[123],"techniques.Our":[125],"method":[126],"outperforms":[127],"existing":[128],"techniques":[130],"hallucination":[134],"mitigation":[135],"benchmarks":[136],"for":[137,144],"VLMs":[138],"proposes":[140],"robust":[142],"solution":[143],"through":[148],"engineering.":[150]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-11-08T00:00:00"}
