{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:24IQCLWZYWZWI7J7OK5WFJTXJY","short_pith_number":"pith:24IQCLWZ","schema_version":"1.0","canonical_sha256":"d711012ed9c5b3647d3f72bb62a6774e2ff8de8e89283a8b49fb65cf008f2e84","source":{"kind":"arxiv","id":"2105.13880","version":2},"attestation_state":"computed","paper":{"title":"Knowledge Inheritance for Pre-trained Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiajie Zhang, Jie Zhou, Jing Yi, Maosong Sun, Peng Li, Xu Han, Yankai Lin, Yujia Qin, Yusheng Su, Zhengyan Zhang, Zhiyuan Liu","submitted_at":"2021-05-28T14:43:26Z","abstract_excerpt":"Recent explorations of large-scale pre-trained language models (PLMs) have revealed the power of PLMs with huge amounts of parameters, setting off a wave of training ever-larger PLMs. However, it requires tremendous computational resources to train a large-scale PLM, which may be practically unaffordable. In addition, existing large-scale PLMs are mainly trained from scratch individually, ignoring that many well-trained PLMs are available. To this end, we explore the question how could existing PLMs benefit training large-scale PLMs in future. Specifically, we introduce a pre-training framewor"},"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":"2105.13880","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-05-28T14:43:26Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"1175b01442cf9a744f53d8bb528001ae881ef0b78f351157578ede79beac8b23","abstract_canon_sha256":"2e6f4b597544c2cf05ccf93a2d8fdf79fb7b44a15d55b6bb911a1412403be955"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:17:35.324062Z","signature_b64":"MGfB4HZmyH702V+X9dVl5Z42Xq8or5IblQZkLDhlClMPAIuUnCnjaA0C1f5VqrW4n8PTuTs3HJSC0hclSFEIDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d711012ed9c5b3647d3f72bb62a6774e2ff8de8e89283a8b49fb65cf008f2e84","last_reissued_at":"2026-07-05T04:17:35.323652Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:17:35.323652Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Knowledge Inheritance for Pre-trained Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiajie Zhang, Jie Zhou, Jing Yi, Maosong Sun, Peng Li, Xu Han, Yankai Lin, Yujia Qin, Yusheng Su, Zhengyan Zhang, Zhiyuan Liu","submitted_at":"2021-05-28T14:43:26Z","abstract_excerpt":"Recent explorations of large-scale pre-trained language models (PLMs) have revealed the power of PLMs with huge amounts of parameters, setting off a wave of training ever-larger PLMs. However, it requires tremendous computational resources to train a large-scale PLM, which may be practically unaffordable. In addition, existing large-scale PLMs are mainly trained from scratch individually, ignoring that many well-trained PLMs are available. To this end, we explore the question how could existing PLMs benefit training large-scale PLMs in future. Specifically, we introduce a pre-training framewor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.13880","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/2105.13880/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":"2105.13880","created_at":"2026-07-05T04:17:35.323709+00:00"},{"alias_kind":"arxiv_version","alias_value":"2105.13880v2","created_at":"2026-07-05T04:17:35.323709+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.13880","created_at":"2026-07-05T04:17:35.323709+00:00"},{"alias_kind":"pith_short_12","alias_value":"24IQCLWZYWZW","created_at":"2026-07-05T04:17:35.323709+00:00"},{"alias_kind":"pith_short_16","alias_value":"24IQCLWZYWZWI7J7","created_at":"2026-07-05T04:17:35.323709+00:00"},{"alias_kind":"pith_short_8","alias_value":"24IQCLWZ","created_at":"2026-07-05T04:17:35.323709+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2305.14233","citing_title":"Enhancing Chat Language Models by Scaling High-quality Instructional Conversations","ref_index":187,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13405","citing_title":"When is Warmstarting Effective for Scaling Language Models?","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.06377","citing_title":"The Master Key Hypothesis: Unlocking Cross-Model Capability Transfer via Linear Subspace Alignment","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY","json":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY.json","graph_json":"https://pith.science/api/pith-number/24IQCLWZYWZWI7J7OK5WFJTXJY/graph.json","events_json":"https://pith.science/api/pith-number/24IQCLWZYWZWI7J7OK5WFJTXJY/events.json","paper":"https://pith.science/paper/24IQCLWZ"},"agent_actions":{"view_html":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY","download_json":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY.json","view_paper":"https://pith.science/paper/24IQCLWZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2105.13880&json=true","fetch_graph":"https://pith.science/api/pith-number/24IQCLWZYWZWI7J7OK5WFJTXJY/graph.json","fetch_events":"https://pith.science/api/pith-number/24IQCLWZYWZWI7J7OK5WFJTXJY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY/action/storage_attestation","attest_author":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY/action/author_attestation","sign_citation":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY/action/citation_signature","submit_replication":"https://pith.science/pith/24IQCLWZYWZWI7J7OK5WFJTXJY/action/replication_record"}},"created_at":"2026-07-05T04:17:35.323709+00:00","updated_at":"2026-07-05T04:17:35.323709+00:00"}