{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5FIO742N4YDBQOLSLECYHMPIID","short_pith_number":"pith:5FIO742N","schema_version":"1.0","canonical_sha256":"e950eff34de606183972590583b1e840c5a7fb023b27b13a56863ead93070904","source":{"kind":"arxiv","id":"2401.06477","version":4},"attestation_state":"computed","paper":{"title":"Kun: Answer Polishment for Chinese Self-Alignment with Instruction Back-Translation","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chenghua Lin, Ge Zhang, Jiawei Guo, Jie Fu, Qi Jia, Shuyue Guo, Tianyu Zheng, Wenhao Huang, Xingwei Qu, Xinrun Du","submitted_at":"2024-01-12T09:56:57Z","abstract_excerpt":"In this paper, we introduce Kun, a novel approach for creating high-quality instruction-tuning datasets for large language models (LLMs) without relying on manual annotations. Adapting a self-training algorithm based on instruction back-translation and answer polishment, Kun leverages unlabelled data from diverse sources such as Wudao, Wanjuan, and SkyPile to generate a substantial dataset of over a million Chinese instructional data points. This approach significantly deviates from traditional methods by using a self-curation process to refine and select the most effective instruction-output "},"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":"2401.06477","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T09:56:57Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cd093bedf1a8db9c8cd795ea558177824b88b1451ab2ba2c5fde5fcf4f29e74f","abstract_canon_sha256":"26b33bd2731d5567bed3bf6fdcdd54fca08100e0a795ece3110ab1fd2d8adbb3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:30:58.709293Z","signature_b64":"MtBGk701XtyKi9FJu+D6khS8OtXD7xoXN4fGaM36rDANGcMohnjSalS3R/CT4HuFoXNZ+TglZ5naqtm5U9vLAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e950eff34de606183972590583b1e840c5a7fb023b27b13a56863ead93070904","last_reissued_at":"2026-07-05T09:30:58.708782Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:30:58.708782Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Kun: Answer Polishment for Chinese Self-Alignment with Instruction Back-Translation","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chenghua Lin, Ge Zhang, Jiawei Guo, Jie Fu, Qi Jia, Shuyue Guo, Tianyu Zheng, Wenhao Huang, Xingwei Qu, Xinrun Du","submitted_at":"2024-01-12T09:56:57Z","abstract_excerpt":"In this paper, we introduce Kun, a novel approach for creating high-quality instruction-tuning datasets for large language models (LLMs) without relying on manual annotations. Adapting a self-training algorithm based on instruction back-translation and answer polishment, Kun leverages unlabelled data from diverse sources such as Wudao, Wanjuan, and SkyPile to generate a substantial dataset of over a million Chinese instructional data points. This approach significantly deviates from traditional methods by using a self-curation process to refine and select the most effective instruction-output "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.06477","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/2401.06477/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":"2401.06477","created_at":"2026-07-05T09:30:58.708844+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.06477v4","created_at":"2026-07-05T09:30:58.708844+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.06477","created_at":"2026-07-05T09:30:58.708844+00:00"},{"alias_kind":"pith_short_12","alias_value":"5FIO742N4YDB","created_at":"2026-07-05T09:30:58.708844+00:00"},{"alias_kind":"pith_short_16","alias_value":"5FIO742N4YDBQOLS","created_at":"2026-07-05T09:30:58.708844+00:00"},{"alias_kind":"pith_short_8","alias_value":"5FIO742N","created_at":"2026-07-05T09:30:58.708844+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.22462","citing_title":"IFEvalCode: Controlled Code Generation","ref_index":83,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID","json":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID.json","graph_json":"https://pith.science/api/pith-number/5FIO742N4YDBQOLSLECYHMPIID/graph.json","events_json":"https://pith.science/api/pith-number/5FIO742N4YDBQOLSLECYHMPIID/events.json","paper":"https://pith.science/paper/5FIO742N"},"agent_actions":{"view_html":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID","download_json":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID.json","view_paper":"https://pith.science/paper/5FIO742N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.06477&json=true","fetch_graph":"https://pith.science/api/pith-number/5FIO742N4YDBQOLSLECYHMPIID/graph.json","fetch_events":"https://pith.science/api/pith-number/5FIO742N4YDBQOLSLECYHMPIID/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID/action/storage_attestation","attest_author":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID/action/author_attestation","sign_citation":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID/action/citation_signature","submit_replication":"https://pith.science/pith/5FIO742N4YDBQOLSLECYHMPIID/action/replication_record"}},"created_at":"2026-07-05T09:30:58.708844+00:00","updated_at":"2026-07-05T09:30:58.708844+00:00"}