{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NHTXTVY2ICHUSQGJSXP6L75UHR","short_pith_number":"pith:NHTXTVY2","schema_version":"1.0","canonical_sha256":"69e779d71a408f4940c995dfe5ffb43c736410b1263df80146913ed130783c55","source":{"kind":"arxiv","id":"2407.02273","version":6},"attestation_state":"computed","paper":{"title":"Language Model Alignment in Multilingual Trolley Problems","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'as Strausz, Bernhard Sch\\\"olkopf, Fernando Gonzalez, Francesco Ortu, Giorgio Piatti, Jiarui Liu, Max Kleiman-Weiner, Mrinmaya Sachan, Rada Mihalcea, Sydney Levine, Yejin Choi, Zhijing Jin","submitted_at":"2024-07-02T14:02:53Z","abstract_excerpt":"We evaluate the moral alignment of LLMs with human preferences in multilingual trolley problems. Building on the Moral Machine experiment, which captures over 40 million human judgments across 200+ countries, we develop a cross-lingual corpus of moral dilemma vignettes in over 100 languages called MultiTP. This dataset enables the assessment of LLMs' decision-making processes in diverse linguistic contexts. Our analysis explores the alignment of 19 different LLMs with human judgments, capturing preferences across six moral dimensions: species, gender, fitness, status, age, and the number of li"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2407.02273","kind":"arxiv","version":6},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-02T14:02:53Z","cross_cats_sorted":[],"title_canon_sha256":"1de70bd6ddcac44f84f6f03f24983d570488ec79ae9f82af86adf5855f12ac27","abstract_canon_sha256":"a1ac09b975ee8f50d5dca9ee8f21cbdd3c22ffa498cb343471ba1e7f155c3941"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:46.408276Z","signature_b64":"sJ48a37OWXy7N0YcS0ofNtKqqrwqF4hd66jEYsQAb3/9yT7aKiCfp344pDSXLzK3cqPYNNtIpTMeJCbxOj8SDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69e779d71a408f4940c995dfe5ffb43c736410b1263df80146913ed130783c55","last_reissued_at":"2026-07-05T11:10:46.407774Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:46.407774Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Model Alignment in Multilingual Trolley Problems","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'as Strausz, Bernhard Sch\\\"olkopf, Fernando Gonzalez, Francesco Ortu, Giorgio Piatti, Jiarui Liu, Max Kleiman-Weiner, Mrinmaya Sachan, Rada Mihalcea, Sydney Levine, Yejin Choi, Zhijing Jin","submitted_at":"2024-07-02T14:02:53Z","abstract_excerpt":"We evaluate the moral alignment of LLMs with human preferences in multilingual trolley problems. Building on the Moral Machine experiment, which captures over 40 million human judgments across 200+ countries, we develop a cross-lingual corpus of moral dilemma vignettes in over 100 languages called MultiTP. This dataset enables the assessment of LLMs' decision-making processes in diverse linguistic contexts. Our analysis explores the alignment of 19 different LLMs with human judgments, capturing preferences across six moral dimensions: species, gender, fitness, status, age, and the number of li"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.02273","kind":"arxiv","version":6},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2407.02273/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2407.02273","created_at":"2026-07-05T11:10:46.407835+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.02273v6","created_at":"2026-07-05T11:10:46.407835+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.02273","created_at":"2026-07-05T11:10:46.407835+00:00"},{"alias_kind":"pith_short_12","alias_value":"NHTXTVY2ICHU","created_at":"2026-07-05T11:10:46.407835+00:00"},{"alias_kind":"pith_short_16","alias_value":"NHTXTVY2ICHUSQGJ","created_at":"2026-07-05T11:10:46.407835+00:00"},{"alias_kind":"pith_short_8","alias_value":"NHTXTVY2","created_at":"2026-07-05T11:10:46.407835+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.03217","citing_title":"Moral Sensitivity in LLMs: A Tiered Evaluation of Contextual Bias via Behavioral Profiling and Mechanistic Interpretability","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10843","citing_title":"Training-Free Cultural Alignment of Large Language Models via Persona Disagreement","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2601.17637","citing_title":"Scaling Laws for Moral Machine Judgment in Large Language Models","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03217","citing_title":"Moral Sensitivity in LLMs: A Tiered Evaluation of Contextual Bias via Behavioral Profiling and Mechanistic Interpretability","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR","json":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR.json","graph_json":"https://pith.science/api/pith-number/NHTXTVY2ICHUSQGJSXP6L75UHR/graph.json","events_json":"https://pith.science/api/pith-number/NHTXTVY2ICHUSQGJSXP6L75UHR/events.json","paper":"https://pith.science/paper/NHTXTVY2"},"agent_actions":{"view_html":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR","download_json":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR.json","view_paper":"https://pith.science/paper/NHTXTVY2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.02273&json=true","fetch_graph":"https://pith.science/api/pith-number/NHTXTVY2ICHUSQGJSXP6L75UHR/graph.json","fetch_events":"https://pith.science/api/pith-number/NHTXTVY2ICHUSQGJSXP6L75UHR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR/action/storage_attestation","attest_author":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR/action/author_attestation","sign_citation":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR/action/citation_signature","submit_replication":"https://pith.science/pith/NHTXTVY2ICHUSQGJSXP6L75UHR/action/replication_record"}},"created_at":"2026-07-05T11:10:46.407835+00:00","updated_at":"2026-07-05T11:10:46.407835+00:00"}