{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IWVM27HEGAHLTLY3OIJKT6T4PM","short_pith_number":"pith:IWVM27HE","schema_version":"1.0","canonical_sha256":"45aacd7ce4300eb9af1b7212a9fa7c7b193a77a07f6161ddc0402dc1ba44be8a","source":{"kind":"arxiv","id":"2502.12982","version":1},"attestation_state":"computed","paper":{"title":"Sailor2: Sailing in South-East Asia with Inclusive Multilingual LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Anh Dao, Changyu Chen, Chao Du, Cunxiao Du, Fajri Koto, Fan Zhou, Haonan Wang, Hoang H. Tran, Hynek Kydl\\'i\\v{c}ek, Jiaheng Liu, Kridtaphad Sae-Khow, Kunat Pipatanakul, Longxu Dou, Man Tsung Yeung, Matichon Maneegard, Mike Zhang, Min Lin, Min Si Thu, Narong Borijindargoon, Nirattisai Thongchim, Penghui Yang, Phakphum Artkaew, Qian Liu, Quan Nguyen, Qunshu Lin, Shiqi Chen, Sittipong Sripaisarnmongkol, Taechawat Konkaew, Tianyu Pang, Tongyao Zhu, Wannaphong Phatthiyaphaibun, Wei Lu, Xiachong Feng, Xin Mao, Xinyi Wan, Yongchi Zhao, Zeyi Liu, Zheng-Xin Yong, Zichen Liu, Zili Wang, Ziqi Jin","submitted_at":"2025-02-18T16:04:57Z","abstract_excerpt":"Sailor2 is a family of cutting-edge multilingual language models for South-East Asian (SEA) languages, available in 1B, 8B, and 20B sizes to suit diverse applications. Building on Qwen2.5, Sailor2 undergoes continuous pre-training on 500B tokens (400B SEA-specific and 100B replay tokens) to support 13 SEA languages while retaining proficiency in Chinese and English. Sailor2-20B model achieves a 50-50 win rate against GPT-4o across SEA languages. We also deliver a comprehensive cookbook on how to develop the multilingual model in an efficient manner, including five key aspects: data curation, p"},"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":"2502.12982","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-18T16:04:57Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"4f68fa8dfcc9cbc4ae63f46fb219a1bca2477612af7ab1d001caff6c7360d495","abstract_canon_sha256":"cb4d28280c381b6a466cdafadc22ef6eee257429a20a9b67b6d733e089acdae0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:27.494957Z","signature_b64":"Wl/OtMgPQuWvY5+DcLfVZ2lgMOV/AIdUcc86ggVArGKhcgBQupKeYxHZGsFmFe1agUqCiAs0bhtUbyTwMORpAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"45aacd7ce4300eb9af1b7212a9fa7c7b193a77a07f6161ddc0402dc1ba44be8a","last_reissued_at":"2026-07-05T10:16:27.494438Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:27.494438Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sailor2: Sailing in South-East Asia with Inclusive Multilingual LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Anh Dao, Changyu Chen, Chao Du, Cunxiao Du, Fajri Koto, Fan Zhou, Haonan Wang, Hoang H. Tran, Hynek Kydl\\'i\\v{c}ek, Jiaheng Liu, Kridtaphad Sae-Khow, Kunat Pipatanakul, Longxu Dou, Man Tsung Yeung, Matichon Maneegard, Mike Zhang, Min Lin, Min Si Thu, Narong Borijindargoon, Nirattisai Thongchim, Penghui Yang, Phakphum Artkaew, Qian Liu, Quan Nguyen, Qunshu Lin, Shiqi Chen, Sittipong Sripaisarnmongkol, Taechawat Konkaew, Tianyu Pang, Tongyao Zhu, Wannaphong Phatthiyaphaibun, Wei Lu, Xiachong Feng, Xin Mao, Xinyi Wan, Yongchi Zhao, Zeyi Liu, Zheng-Xin Yong, Zichen Liu, Zili Wang, Ziqi Jin","submitted_at":"2025-02-18T16:04:57Z","abstract_excerpt":"Sailor2 is a family of cutting-edge multilingual language models for South-East Asian (SEA) languages, available in 1B, 8B, and 20B sizes to suit diverse applications. Building on Qwen2.5, Sailor2 undergoes continuous pre-training on 500B tokens (400B SEA-specific and 100B replay tokens) to support 13 SEA languages while retaining proficiency in Chinese and English. Sailor2-20B model achieves a 50-50 win rate against GPT-4o across SEA languages. We also deliver a comprehensive cookbook on how to develop the multilingual model in an efficient manner, including five key aspects: data curation, p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12982","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/2502.12982/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":"2502.12982","created_at":"2026-07-05T10:16:27.494493+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.12982v1","created_at":"2026-07-05T10:16:27.494493+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12982","created_at":"2026-07-05T10:16:27.494493+00:00"},{"alias_kind":"pith_short_12","alias_value":"IWVM27HEGAHL","created_at":"2026-07-05T10:16:27.494493+00:00"},{"alias_kind":"pith_short_16","alias_value":"IWVM27HEGAHLTLY3","created_at":"2026-07-05T10:16:27.494493+00:00"},{"alias_kind":"pith_short_8","alias_value":"IWVM27HE","created_at":"2026-07-05T10:16:27.494493+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.22099","citing_title":"A Comparative Study of Language Models for Khmer Retrieval-Augmented Question Answering","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18083","citing_title":"A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM$\\Delta$ Integration into Upcycled MoE","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10391","citing_title":"Phoenix-VL 1.5 Medium Technical Report","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05662","citing_title":"XL-SafetyBench: A Country-Grounded Cross-Cultural Benchmark for LLM Safety and Cultural Sensitivity","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM","json":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM.json","graph_json":"https://pith.science/api/pith-number/IWVM27HEGAHLTLY3OIJKT6T4PM/graph.json","events_json":"https://pith.science/api/pith-number/IWVM27HEGAHLTLY3OIJKT6T4PM/events.json","paper":"https://pith.science/paper/IWVM27HE"},"agent_actions":{"view_html":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM","download_json":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM.json","view_paper":"https://pith.science/paper/IWVM27HE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.12982&json=true","fetch_graph":"https://pith.science/api/pith-number/IWVM27HEGAHLTLY3OIJKT6T4PM/graph.json","fetch_events":"https://pith.science/api/pith-number/IWVM27HEGAHLTLY3OIJKT6T4PM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM/action/storage_attestation","attest_author":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM/action/author_attestation","sign_citation":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM/action/citation_signature","submit_replication":"https://pith.science/pith/IWVM27HEGAHLTLY3OIJKT6T4PM/action/replication_record"}},"created_at":"2026-07-05T10:16:27.494493+00:00","updated_at":"2026-07-05T10:16:27.494493+00:00"}