{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LXRJMMOFWSAFBR6A6GDSHIFU2N","short_pith_number":"pith:LXRJMMOF","schema_version":"1.0","canonical_sha256":"5de29631c5b48050c7c0f18723a0b4d346ea993a461007a272c1c72edf0224a0","source":{"kind":"arxiv","id":"2607.23545","version":1},"attestation_state":"computed","paper":{"title":"Language Shapes Instruction Hierarchy Compliance in Multilingual LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiwon Moon, Kyomin Jung, Yerin Hwang","submitted_at":"2026-07-26T08:50:28Z","abstract_excerpt":"Instruction hierarchy (IH) requires models to prioritize instructions by source, ensuring that higher-priority instructions override lower-priority ones. Despite its importance for safe and controllable deployment, existing evaluations have focused almost exclusively on English, leaving it unclear whether IH compliance remains stable in multilingual settings. We introduce XIH-Bench, a benchmark for multilingual IH evaluation with both same-language and cross-language conflicts across six languages, four domains, and three IH settings. Across models, we find two consistent patterns. First, IH c"},"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":"2607.23545","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-26T08:50:28Z","cross_cats_sorted":[],"title_canon_sha256":"bbdc88b107ba447386adf1a6745d53d39dfa1931da27a1fcc0b25bba085d42cf","abstract_canon_sha256":"78c4e2dbb8a6abaffd1ba02b8249efc50af03cb8ca552018886ea1cfb57d331a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-28T01:22:58.554651Z","signature_b64":"LGE83ZIvcgBsRsL6f5iHNF/aen9dwM6PFUB6TUoVy3s4WGRcSDc1uhjVYCwyH/oOCRGNcjZ1aEEEaeVIj+TwCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5de29631c5b48050c7c0f18723a0b4d346ea993a461007a272c1c72edf0224a0","last_reissued_at":"2026-07-28T01:22:58.553819Z","signature_status":"signed_v1","first_computed_at":"2026-07-28T01:22:58.553819Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Shapes Instruction Hierarchy Compliance in Multilingual LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiwon Moon, Kyomin Jung, Yerin Hwang","submitted_at":"2026-07-26T08:50:28Z","abstract_excerpt":"Instruction hierarchy (IH) requires models to prioritize instructions by source, ensuring that higher-priority instructions override lower-priority ones. Despite its importance for safe and controllable deployment, existing evaluations have focused almost exclusively on English, leaving it unclear whether IH compliance remains stable in multilingual settings. We introduce XIH-Bench, a benchmark for multilingual IH evaluation with both same-language and cross-language conflicts across six languages, four domains, and three IH settings. Across models, we find two consistent patterns. First, IH c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.23545","kind":"arxiv","version":1},"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/2607.23545/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":"2607.23545","created_at":"2026-07-28T01:22:58.554242+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.23545v1","created_at":"2026-07-28T01:22:58.554242+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.23545","created_at":"2026-07-28T01:22:58.554242+00:00"},{"alias_kind":"pith_short_12","alias_value":"LXRJMMOFWSAF","created_at":"2026-07-28T01:22:58.554242+00:00"},{"alias_kind":"pith_short_16","alias_value":"LXRJMMOFWSAFBR6A","created_at":"2026-07-28T01:22:58.554242+00:00"},{"alias_kind":"pith_short_8","alias_value":"LXRJMMOF","created_at":"2026-07-28T01:22:58.554242+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N","json":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N.json","graph_json":"https://pith.science/api/pith-number/LXRJMMOFWSAFBR6A6GDSHIFU2N/graph.json","events_json":"https://pith.science/api/pith-number/LXRJMMOFWSAFBR6A6GDSHIFU2N/events.json","paper":"https://pith.science/paper/LXRJMMOF"},"agent_actions":{"view_html":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N","download_json":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N.json","view_paper":"https://pith.science/paper/LXRJMMOF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.23545&json=true","fetch_graph":"https://pith.science/api/pith-number/LXRJMMOFWSAFBR6A6GDSHIFU2N/graph.json","fetch_events":"https://pith.science/api/pith-number/LXRJMMOFWSAFBR6A6GDSHIFU2N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N/action/storage_attestation","attest_author":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N/action/author_attestation","sign_citation":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N/action/citation_signature","submit_replication":"https://pith.science/pith/LXRJMMOFWSAFBR6A6GDSHIFU2N/action/replication_record"}},"created_at":"2026-07-28T01:22:58.554242+00:00","updated_at":"2026-07-28T01:22:58.554242+00:00"}