{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:OKRP4FJMWUSKQA2YOHDZOAXN7C","short_pith_number":"pith:OKRP4FJM","schema_version":"1.0","canonical_sha256":"72a2fe152cb524a8035871c79702edf888854dec53e701769a2dc07a4adb3080","source":{"kind":"arxiv","id":"2408.04303","version":1},"attestation_state":"computed","paper":{"title":"Trans-Tokenization and Cross-lingual Vocabulary Transfers: Language Adaptation of LLMs for Low-Resource NLP","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alfiya Khabibullina, Fran\\c{c}ois Remy, Hayastan Avetisyan, Miryam de Lhoneux, Pieter Delobelle, Thomas Demeester","submitted_at":"2024-08-08T08:37:28Z","abstract_excerpt":"The development of monolingual language models for low and mid-resource languages continues to be hindered by the difficulty in sourcing high-quality training data. In this study, we present a novel cross-lingual vocabulary transfer strategy, trans-tokenization, designed to tackle this challenge and enable more efficient language adaptation. Our approach focuses on adapting a high-resource monolingual LLM to an unseen target language by initializing the token embeddings of the target language using a weighted average of semantically similar token embeddings from the source language. For this, "},"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":"2408.04303","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-08-08T08:37:28Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"bb527d9da28eec04ef73f7cbb8304ac0f391500cfdd249d099b83941bc39466b","abstract_canon_sha256":"cd3f3bf72bc2cf1be13f3555f06e79cb131eaf850c784d13a649108d6092b7f5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:53:31.409993Z","signature_b64":"qNQOAyaabuay+dd+uwaqdcJ0KiwAs1mtSdN6QQS/s8AYtAfde0ocTzissPebZUwUKVdvkEPdJXKk69tVy1eyAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"72a2fe152cb524a8035871c79702edf888854dec53e701769a2dc07a4adb3080","last_reissued_at":"2026-07-05T08:53:31.409513Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:53:31.409513Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Trans-Tokenization and Cross-lingual Vocabulary Transfers: Language Adaptation of LLMs for Low-Resource NLP","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alfiya Khabibullina, Fran\\c{c}ois Remy, Hayastan Avetisyan, Miryam de Lhoneux, Pieter Delobelle, Thomas Demeester","submitted_at":"2024-08-08T08:37:28Z","abstract_excerpt":"The development of monolingual language models for low and mid-resource languages continues to be hindered by the difficulty in sourcing high-quality training data. In this study, we present a novel cross-lingual vocabulary transfer strategy, trans-tokenization, designed to tackle this challenge and enable more efficient language adaptation. Our approach focuses on adapting a high-resource monolingual LLM to an unseen target language by initializing the token embeddings of the target language using a weighted average of semantically similar token embeddings from the source language. For this, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.04303","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/2408.04303/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":"2408.04303","created_at":"2026-07-05T08:53:31.409580+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.04303v1","created_at":"2026-07-05T08:53:31.409580+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.04303","created_at":"2026-07-05T08:53:31.409580+00:00"},{"alias_kind":"pith_short_12","alias_value":"OKRP4FJMWUSK","created_at":"2026-07-05T08:53:31.409580+00:00"},{"alias_kind":"pith_short_16","alias_value":"OKRP4FJMWUSKQA2Y","created_at":"2026-07-05T08:53:31.409580+00:00"},{"alias_kind":"pith_short_8","alias_value":"OKRP4FJM","created_at":"2026-07-05T08:53:31.409580+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28057","citing_title":"MultiHashFormer: Hash-based Generative Language Models","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23566","citing_title":"LangMAP: A Language-Adaptive Approach to Tokenization","ref_index":54,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C","json":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C.json","graph_json":"https://pith.science/api/pith-number/OKRP4FJMWUSKQA2YOHDZOAXN7C/graph.json","events_json":"https://pith.science/api/pith-number/OKRP4FJMWUSKQA2YOHDZOAXN7C/events.json","paper":"https://pith.science/paper/OKRP4FJM"},"agent_actions":{"view_html":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C","download_json":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C.json","view_paper":"https://pith.science/paper/OKRP4FJM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.04303&json=true","fetch_graph":"https://pith.science/api/pith-number/OKRP4FJMWUSKQA2YOHDZOAXN7C/graph.json","fetch_events":"https://pith.science/api/pith-number/OKRP4FJMWUSKQA2YOHDZOAXN7C/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C/action/storage_attestation","attest_author":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C/action/author_attestation","sign_citation":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C/action/citation_signature","submit_replication":"https://pith.science/pith/OKRP4FJMWUSKQA2YOHDZOAXN7C/action/replication_record"}},"created_at":"2026-07-05T08:53:31.409580+00:00","updated_at":"2026-07-05T08:53:31.409580+00:00"}