{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JM7GCQRCVSUU6O5CRP3B435A54","short_pith_number":"pith:JM7GCQRC","schema_version":"1.0","canonical_sha256":"4b3e614222aca94f3ba28bf61e6fa0ef0ca6637eb1f57d1814213ff51f7fbce8","source":{"kind":"arxiv","id":"2504.12140","version":2},"attestation_state":"computed","paper":{"title":"Multilingual Contextualization of Large Language Models for Document-Level Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'e F. T. Martins, Miguel Moura Ramos, Patrick Fernandes, Sweta Agrawal","submitted_at":"2025-04-16T14:52:22Z","abstract_excerpt":"Large language models (LLMs) have demonstrated strong performance in sentence-level machine translation, but scaling to document-level translation remains challenging, particularly in modeling long-range dependencies and discourse phenomena across sentences and paragraphs. In this work, we propose a method to improve LLM-based long-document translation through targeted fine-tuning on high-quality document-level data, which we curate and introduce as DocBlocks. Our approach supports multiple translation paradigms, including direct document-to-document and chunk-level translation, by integrating"},"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":"2504.12140","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-16T14:52:22Z","cross_cats_sorted":[],"title_canon_sha256":"f08e5f150beeae707a9774391afc5a1c1e8b0d28da6aedc1a3d997d65a373d88","abstract_canon_sha256":"ecd12f5600996cc7e0464c232bf0a5083a87130cf2355c587bd30058fa546c20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:00:49.507881Z","signature_b64":"FWlWYYtTgYcfD+5vAUhjbfZlSwvOd0woGkShUo5PVmIxA2RYn7jKJcPVw68RjvzOoecXLl6uNNQQGCF3z0J7BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b3e614222aca94f3ba28bf61e6fa0ef0ca6637eb1f57d1814213ff51f7fbce8","last_reissued_at":"2026-07-05T12:00:49.507390Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:00:49.507390Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multilingual Contextualization of Large Language Models for Document-Level Machine Translation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andr\\'e F. T. Martins, Miguel Moura Ramos, Patrick Fernandes, Sweta Agrawal","submitted_at":"2025-04-16T14:52:22Z","abstract_excerpt":"Large language models (LLMs) have demonstrated strong performance in sentence-level machine translation, but scaling to document-level translation remains challenging, particularly in modeling long-range dependencies and discourse phenomena across sentences and paragraphs. In this work, we propose a method to improve LLM-based long-document translation through targeted fine-tuning on high-quality document-level data, which we curate and introduce as DocBlocks. Our approach supports multiple translation paradigms, including direct document-to-document and chunk-level translation, by integrating"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.12140","kind":"arxiv","version":2},"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/2504.12140/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":"2504.12140","created_at":"2026-07-05T12:00:49.507448+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.12140v2","created_at":"2026-07-05T12:00:49.507448+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.12140","created_at":"2026-07-05T12:00:49.507448+00:00"},{"alias_kind":"pith_short_12","alias_value":"JM7GCQRCVSUU","created_at":"2026-07-05T12:00:49.507448+00:00"},{"alias_kind":"pith_short_16","alias_value":"JM7GCQRCVSUU6O5C","created_at":"2026-07-05T12:00:49.507448+00:00"},{"alias_kind":"pith_short_8","alias_value":"JM7GCQRC","created_at":"2026-07-05T12:00:49.507448+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2511.01066","citing_title":"HPLT 3.0: Very Large-Scale Multilingual Resources for LLMs and MT. Mono- and Bi-lingual Data, Multilingual Evaluation, and Pre-Trained Models","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54","json":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54.json","graph_json":"https://pith.science/api/pith-number/JM7GCQRCVSUU6O5CRP3B435A54/graph.json","events_json":"https://pith.science/api/pith-number/JM7GCQRCVSUU6O5CRP3B435A54/events.json","paper":"https://pith.science/paper/JM7GCQRC"},"agent_actions":{"view_html":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54","download_json":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54.json","view_paper":"https://pith.science/paper/JM7GCQRC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.12140&json=true","fetch_graph":"https://pith.science/api/pith-number/JM7GCQRCVSUU6O5CRP3B435A54/graph.json","fetch_events":"https://pith.science/api/pith-number/JM7GCQRCVSUU6O5CRP3B435A54/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54/action/storage_attestation","attest_author":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54/action/author_attestation","sign_citation":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54/action/citation_signature","submit_replication":"https://pith.science/pith/JM7GCQRCVSUU6O5CRP3B435A54/action/replication_record"}},"created_at":"2026-07-05T12:00:49.507448+00:00","updated_at":"2026-07-05T12:00:49.507448+00:00"}