{"id":"https://openalex.org/W4412889812","doi":"https://doi.org/10.18653/v1/2025.acl-long.1326","title":"Internal Value Alignment in Large Language Models through Controlled Value Vector Activation","display_name":"Internal Value Alignment in Large Language Models through Controlled Value Vector Activation","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889812","doi":"https://doi.org/10.18653/v1/2025.acl-long.1326"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.1326","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1326","pdf_url":"https://aclanthology.org/2025.acl-long.1326.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.1326.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5032830409","display_name":"He Jin","orcid":"https://orcid.org/0000-0001-7574-3018"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haoran Jin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102012996","display_name":"Meng Li","orcid":"https://orcid.org/0000-0002-3937-1250"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Meng Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5018582673","display_name":"Xiting Wang","orcid":"https://orcid.org/0000-0002-1846-1118"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiting Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5114506825","display_name":"Zhihao Xu","orcid":"https://orcid.org/0000-0001-9558-1079"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhihao Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044042138","display_name":"Minlie Huang","orcid":"https://orcid.org/0000-0001-7111-1849"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Minlie Huang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":null,"display_name":"Yantao Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yantao Jia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5085254654","display_name":"Defu Lian","orcid":"https://orcid.org/0000-0002-3507-9607"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Defu Lian","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":"27347","last_page":"27371"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.7766000032424927,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.7766000032424927,"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":0.7727000117301941,"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/value","display_name":"Value (mathematics)","score":0.7427806258201599},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5297628045082092},{"id":"https://openalex.org/keywords/machine-learning","display_name":"Machine learning","score":0.13952293992042542}],"concepts":[{"id":"https://openalex.org/C2776291640","wikidata":"https://www.wikidata.org/wiki/Q2912517","display_name":"Value (mathematics)","level":2,"score":0.7427806258201599},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5297628045082092},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.13952293992042542}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.1326","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1326","pdf_url":"https://aclanthology.org/2025.acl-long.1326.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.1326","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1326","pdf_url":"https://aclanthology.org/2025.acl-long.1326.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":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889812.pdf","grobid_xml":"https://content.openalex.org/works/W4412889812.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052"],"abstract_inverted_index":{"Aligning":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"with":[5,127],"human":[6],"values":[7,41,62,81,115,125],"has":[8],"attracted":[9],"increasing":[10],"attention":[11],"since":[12],"it":[13],"provides":[14],"clarity,":[15],"transparency,":[16],"and":[17,55,68,96,120,122,129],"the":[18,39,107],"ability":[19],"to":[20,22,59],"adapt":[21],"evolving":[23],"scenarios.In":[24],"this":[25],"paper,":[26],"we":[27,71,86],"introduce":[28,87],"a":[29,47,73,88],"Controlled":[30],"Value":[31],"Vector":[32],"Activation":[33],"(ConVA)":[34],"method":[35,93,105],"that":[36,103],"directly":[37],"aligns":[38],"internal":[40],"of":[42,99],"LLMs":[43],"by":[44],"interpreting":[45],"how":[46],"value":[48,75,90,100],"is":[49],"encoded":[50],"in":[51,63],"their":[52],"latent":[53],"representations":[54],"modifies":[56],"relevant":[57],"activations":[58],"ensure":[60,65],"consistent":[61],"LLMs.To":[64],"an":[66],"accurate":[67],"unbiased":[69],"interpretation,":[70],"propose":[72],"context-controlled":[74],"vector":[76,91],"identification":[77],"method.To":[78],"consistently":[79],"control":[80,109],"without":[82,116],"sacrificing":[83],"model":[84],"performance,":[85],"gated":[89],"activation":[92],"for":[94],"effective":[95],"minimum":[97],"degree":[98],"control.Experiments":[101],"show":[102],"our":[104],"achieves":[106],"highest":[108],"success":[110],"rate":[111],"across":[112],"10":[113],"basic":[114],"hurting":[117],"LLM":[118],"performance":[119],"fluency,":[121],"ensures":[123],"target":[124],"even":[126],"opposite":[128],"potentially":[130],"malicious":[131],"input":[132],"prompts.":[133]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}