{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2SBQPZTGLAX4R3GLXLDO7TR7LQ","short_pith_number":"pith:2SBQPZTG","schema_version":"1.0","canonical_sha256":"d48307e666582fc8eccbbac6efce3f5c226b732d60f1a6e676d6b005ce930b67","source":{"kind":"arxiv","id":"2503.11280","version":5},"attestation_state":"computed","paper":{"title":"High-Dimensional Interlingual Representations of Large Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bryan Wilie, Junxian He, Pascale Fung, Samuel Cahyawijaya","submitted_at":"2025-03-14T10:39:27Z","abstract_excerpt":"Large language models (LLMs) trained on massive multilingual datasets hint at the formation of interlingual constructs--a shared subspace in the representation space. However, evidence regarding this phenomenon is mixed, leaving it unclear whether these models truly develop unified interlingual representations, or present a partially aligned constructs. We explore 31 diverse languages varying on their resource-levels, typologies, and geographical regions; and find that multilingual LLMs exhibit inconsistent cross-lingual alignments. To address this, we propose an interlingual representation fr"},"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":"2503.11280","kind":"arxiv","version":5},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-14T10:39:27Z","cross_cats_sorted":[],"title_canon_sha256":"ab388436de1734ea1c11585f6732784f21aac28632dfdc850e7fd1cba102bf9c","abstract_canon_sha256":"d40bde5f4c734d434af99975ae6ab323d73c29c11cae7a7efecb7812f1230757"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:54:45.215253Z","signature_b64":"lWSque5OXomek5HyflhVvMPKECE5MylAspGJtx4U6Quc8dY3IujfC9GO7ZFzdUWJpqQS9oB9J3SJANUrFvlbDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d48307e666582fc8eccbbac6efce3f5c226b732d60f1a6e676d6b005ce930b67","last_reissued_at":"2026-07-05T11:54:45.214780Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:54:45.214780Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"High-Dimensional Interlingual Representations of Large Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bryan Wilie, Junxian He, Pascale Fung, Samuel Cahyawijaya","submitted_at":"2025-03-14T10:39:27Z","abstract_excerpt":"Large language models (LLMs) trained on massive multilingual datasets hint at the formation of interlingual constructs--a shared subspace in the representation space. However, evidence regarding this phenomenon is mixed, leaving it unclear whether these models truly develop unified interlingual representations, or present a partially aligned constructs. We explore 31 diverse languages varying on their resource-levels, typologies, and geographical regions; and find that multilingual LLMs exhibit inconsistent cross-lingual alignments. To address this, we propose an interlingual representation fr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.11280","kind":"arxiv","version":5},"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/2503.11280/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":"2503.11280","created_at":"2026-07-05T11:54:45.214835+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.11280v5","created_at":"2026-07-05T11:54:45.214835+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.11280","created_at":"2026-07-05T11:54:45.214835+00:00"},{"alias_kind":"pith_short_12","alias_value":"2SBQPZTGLAX4","created_at":"2026-07-05T11:54:45.214835+00:00"},{"alias_kind":"pith_short_16","alias_value":"2SBQPZTGLAX4R3GL","created_at":"2026-07-05T11:54:45.214835+00:00"},{"alias_kind":"pith_short_8","alias_value":"2SBQPZTG","created_at":"2026-07-05T11:54:45.214835+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20546","citing_title":"Paths Not Taken: Understanding and Mending the Multilingual Factual Recall Pipeline","ref_index":44,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ","json":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ.json","graph_json":"https://pith.science/api/pith-number/2SBQPZTGLAX4R3GLXLDO7TR7LQ/graph.json","events_json":"https://pith.science/api/pith-number/2SBQPZTGLAX4R3GLXLDO7TR7LQ/events.json","paper":"https://pith.science/paper/2SBQPZTG"},"agent_actions":{"view_html":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ","download_json":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ.json","view_paper":"https://pith.science/paper/2SBQPZTG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.11280&json=true","fetch_graph":"https://pith.science/api/pith-number/2SBQPZTGLAX4R3GLXLDO7TR7LQ/graph.json","fetch_events":"https://pith.science/api/pith-number/2SBQPZTGLAX4R3GLXLDO7TR7LQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ/action/storage_attestation","attest_author":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ/action/author_attestation","sign_citation":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ/action/citation_signature","submit_replication":"https://pith.science/pith/2SBQPZTGLAX4R3GLXLDO7TR7LQ/action/replication_record"}},"created_at":"2026-07-05T11:54:45.214835+00:00","updated_at":"2026-07-05T11:54:45.214835+00:00"}