{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QPGU7RWXDREDKHJCSXLJRCMPXX","short_pith_number":"pith:QPGU7RWX","schema_version":"1.0","canonical_sha256":"83cd4fc6d71c48351d2295d698898fbde18d2c625bfb093d6320724248d1af32","source":{"kind":"arxiv","id":"2502.02481","version":4},"attestation_state":"computed","paper":{"title":"Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bin Wang, Jian Luan, Menglong Cui, Pengzhi Gao, Wei Liu","submitted_at":"2025-02-04T16:57:03Z","abstract_excerpt":"Large language models (LLMs) have shown continuously improving multilingual capabilities, and even small-scale open-source models have demonstrated rapid performance enhancement. In this paper, we systematically explore the abilities of open LLMs with less than ten billion parameters to handle multilingual machine translation (MT) tasks. We conduct comprehensive evaluations on six popular LLMs and find that models like Gemma2-9B exhibit impressive multilingual translation capabilities. We then introduce the Parallel-First Monolingual-Second (PFMS) data mixing strategy in the continual pretrain"},"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":"2502.02481","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-04T16:57:03Z","cross_cats_sorted":[],"title_canon_sha256":"ee9cdd8bb459eb3345a1297cea154229251d77e47136514c04198eb6d3165c24","abstract_canon_sha256":"8746337cf13dfdba81364fec1071eaa4e21d055ada5f796eb688f428fd256ed9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:18:49.220509Z","signature_b64":"98iG050lq532ZhztDQ4PXm/ru1BjbPnUEJUcLeZpzwHpZF4NhznjZd0BoS6sQtrJTls0OHUY31/nci5oUN0pBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"83cd4fc6d71c48351d2295d698898fbde18d2c625bfb093d6320724248d1af32","last_reissued_at":"2026-07-05T10:18:49.220001Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:18:49.220001Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bin Wang, Jian Luan, Menglong Cui, Pengzhi Gao, Wei Liu","submitted_at":"2025-02-04T16:57:03Z","abstract_excerpt":"Large language models (LLMs) have shown continuously improving multilingual capabilities, and even small-scale open-source models have demonstrated rapid performance enhancement. In this paper, we systematically explore the abilities of open LLMs with less than ten billion parameters to handle multilingual machine translation (MT) tasks. We conduct comprehensive evaluations on six popular LLMs and find that models like Gemma2-9B exhibit impressive multilingual translation capabilities. We then introduce the Parallel-First Monolingual-Second (PFMS) data mixing strategy in the continual pretrain"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.02481","kind":"arxiv","version":4},"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/2502.02481/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":"2502.02481","created_at":"2026-07-05T10:18:49.220060+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.02481v4","created_at":"2026-07-05T10:18:49.220060+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02481","created_at":"2026-07-05T10:18:49.220060+00:00"},{"alias_kind":"pith_short_12","alias_value":"QPGU7RWXDRED","created_at":"2026-07-05T10:18:49.220060+00:00"},{"alias_kind":"pith_short_16","alias_value":"QPGU7RWXDREDKHJC","created_at":"2026-07-05T10:18:49.220060+00:00"},{"alias_kind":"pith_short_8","alias_value":"QPGU7RWX","created_at":"2026-07-05T10:18:49.220060+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.05444","citing_title":"Multilingual Coreference Resolution via Cycle-Consistent Machine Translation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2512.01512","citing_title":"MCAT: Scaling Many-to-Many Speech-to-Text Translation with MLLMs to 70 Languages","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX","json":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX.json","graph_json":"https://pith.science/api/pith-number/QPGU7RWXDREDKHJCSXLJRCMPXX/graph.json","events_json":"https://pith.science/api/pith-number/QPGU7RWXDREDKHJCSXLJRCMPXX/events.json","paper":"https://pith.science/paper/QPGU7RWX"},"agent_actions":{"view_html":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX","download_json":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX.json","view_paper":"https://pith.science/paper/QPGU7RWX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.02481&json=true","fetch_graph":"https://pith.science/api/pith-number/QPGU7RWXDREDKHJCSXLJRCMPXX/graph.json","fetch_events":"https://pith.science/api/pith-number/QPGU7RWXDREDKHJCSXLJRCMPXX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX/action/storage_attestation","attest_author":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX/action/author_attestation","sign_citation":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX/action/citation_signature","submit_replication":"https://pith.science/pith/QPGU7RWXDREDKHJCSXLJRCMPXX/action/replication_record"}},"created_at":"2026-07-05T10:18:49.220060+00:00","updated_at":"2026-07-05T10:18:49.220060+00:00"}