{"id":"https://openalex.org/W4412888328","doi":"https://doi.org/10.18653/v1/2025.findings-acl.581","title":"Probabilistic Aggregation and Targeted Embedding Optimization for Collective Moral Reasoning in Large Language Models","display_name":"Probabilistic Aggregation and Targeted Embedding Optimization for Collective Moral Reasoning in Large Language Models","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412888328","doi":"https://doi.org/10.18653/v1/2025.findings-acl.581"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.findings-acl.581","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.581","pdf_url":"https://aclanthology.org/2025.findings-acl.581.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: ACL 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-acl.581.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5007850737","display_name":"Chenchen Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chenchen Yuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100677556","display_name":"Zheyu Zhang","orcid":"https://orcid.org/0000-0003-0768-2713"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheyu Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024712654","display_name":"Shuo Yang","orcid":"https://orcid.org/0000-0003-0190-3319"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shuo Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017702643","display_name":"Bardh Prenkaj","orcid":"https://orcid.org/0000-0002-2991-2279"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bardh Prenkaj","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5024434748","display_name":"Gjergji Kasneci","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gjergji Kasneci","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":"11151","last_page":"11168"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.9628999829292297,"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/T10028","display_name":"Topic Modeling","score":0.9628999829292297,"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/probabilistic-logic","display_name":"Probabilistic logic","score":0.7687530517578125},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.6890990734100342},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.5911816954612732},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.35270652174949646},{"id":"https://openalex.org/keywords/theoretical-computer-science","display_name":"Theoretical computer science","score":0.34695714712142944}],"concepts":[{"id":"https://openalex.org/C49937458","wikidata":"https://www.wikidata.org/wiki/Q2599292","display_name":"Probabilistic logic","level":2,"score":0.7687530517578125},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.6890990734100342},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5911816954612732},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.35270652174949646},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.34695714712142944}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.18653/v1/2025.findings-acl.581","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.581","pdf_url":"https://aclanthology.org/2025.findings-acl.581.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: ACL 2025","raw_type":"proceedings-article"},{"id":"pmh:doi:10.48550/arxiv.2506.14625","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"publisher-specific-oa","license_id":"https://openalex.org/licenses/publisher-specific-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.findings-acl.581","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.findings-acl.581","pdf_url":"https://aclanthology.org/2025.findings-acl.581.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: ACL 2025","raw_type":"proceedings-article"},"sustainable_development_goals":[{"score":0.5400000214576721,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412888328.pdf","grobid_xml":"https://content.openalex.org/works/W4412888328.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/W2081900870","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W1497573972","https://openalex.org/W2932872266","https://openalex.org/W2071659383","https://openalex.org/W1846253165"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models":[2],"(LLMs)":[3],"have":[4],"shown":[5],"impressive":[6],"moral":[7,18,31,37,51,76,93,113],"reasoning":[8],"abilities.Yet":[9],"they":[10],"often":[11],"diverge":[12],"when":[13],"confronted":[14],"with":[15],"complex,":[16],"multi-factor":[17],"dilemmas.To":[19],"address":[20],"these":[21],"discrepancies,":[22],"we":[23],"propose":[24],"a":[25,34,58,68,90],"framework":[26],"that":[27,41],"synthesizes":[28],"multiple":[29,116],"LLMs'":[30],"judgments":[32],"into":[33,57],"collectively":[35],"formulated":[36],"judgment,":[38],"realigning":[39],"models":[40,117],"deviate":[42],"significantly":[43],"from":[44],"this":[45],"consensus.Our":[46],"aggregation":[47],"mechanism":[48],"fuses":[49],"continuous":[50],"acceptability":[52],"scores":[53],"(beyond":[54],"binary":[55],"labels)":[56],"collective":[59],"probability,":[60],"weighting":[61],"contributions":[62],"by":[63],"model":[64,105],"reliability.For":[65],"misaligned":[66],"models,":[67],"targeted":[69],"embedding-optimization":[70],"procedure":[71],"fine-tunes":[72],"token":[73],"embeddings":[74],"for":[75,121],"philosophical":[77],"theories,":[78],"minimizing":[79],"JS":[80],"divergence":[81],"to":[82],"the":[83,109],"consensus":[84,101],"while":[85],"preserving":[86],"semantic":[87],"integrity.Experiments":[88],"on":[89],"large-scale":[91],"social":[92],"dilemma":[94],"dataset":[95],"show":[96],"our":[97],"approach":[98],"builds":[99],"robust":[100],"and":[102,118],"improves":[103],"individual":[104],"fidelity.These":[106],"findings":[107],"highlight":[108],"value":[110],"of":[111],"datadriven":[112],"alignment":[114],"across":[115],"its":[119],"potential":[120],"safer,":[122],"more":[123],"consistent":[124],"AI":[125],"systems.":[126],"1":[127]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}