{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:AO5Z6HIL37OW6AGO2OERRFEGVE","short_pith_number":"pith:AO5Z6HIL","schema_version":"1.0","canonical_sha256":"03bb9f1d0bdfdd6f00ced389189486a93ba5497c2293eb86ba0ac4ba7f8d4525","source":{"kind":"arxiv","id":"2410.16215","version":1},"attestation_state":"computed","paper":{"title":"Pre-training Distillation for Large Language Models: A Design Space Exploration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Peng, Jiajie Zhang, Juanzi Li, Lei Hou, Xin Lv, Yushi Bai, Zijun Yao","submitted_at":"2024-10-21T17:16:13Z","abstract_excerpt":"Knowledge distillation (KD) aims to transfer knowledge from a large teacher model to a smaller student model. Previous work applying KD in the field of large language models (LLMs) typically focused on the post-training phase, where the student LLM learns directly from instructions and corresponding responses generated by the teacher model. In this paper, we extend KD to the pre-training phase of LLMs, named pre-training distillation (PD). We first conduct a preliminary experiment using GLM-4-9B as the teacher LLM to distill a 1.9B parameter student LLM, validating the effectiveness of PD. Con"},"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":"2410.16215","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-21T17:16:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"82aff5c73ceb62334dea0e87da8c47c90ef378b19522bba80fc86075ee35f2c2","abstract_canon_sha256":"8bc584666e352aff26b802bff3ce21690bdf0e902b077015b08f27073a6c4e8a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:23:31.554010Z","signature_b64":"M+wlIVVSSUf4Fx/kAWPVxQvLLWYxWLOb3keXc7cTC6S6PcuD3ToX5gfJoRKzmRPEUuZQgewarHkyTxxkVXoxAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"03bb9f1d0bdfdd6f00ced389189486a93ba5497c2293eb86ba0ac4ba7f8d4525","last_reissued_at":"2026-07-05T09:23:31.553532Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:23:31.553532Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pre-training Distillation for Large Language Models: A Design Space Exploration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Peng, Jiajie Zhang, Juanzi Li, Lei Hou, Xin Lv, Yushi Bai, Zijun Yao","submitted_at":"2024-10-21T17:16:13Z","abstract_excerpt":"Knowledge distillation (KD) aims to transfer knowledge from a large teacher model to a smaller student model. Previous work applying KD in the field of large language models (LLMs) typically focused on the post-training phase, where the student LLM learns directly from instructions and corresponding responses generated by the teacher model. In this paper, we extend KD to the pre-training phase of LLMs, named pre-training distillation (PD). We first conduct a preliminary experiment using GLM-4-9B as the teacher LLM to distill a 1.9B parameter student LLM, validating the effectiveness of PD. Con"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16215","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/2410.16215/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":"2410.16215","created_at":"2026-07-05T09:23:31.553591+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.16215v1","created_at":"2026-07-05T09:23:31.553591+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16215","created_at":"2026-07-05T09:23:31.553591+00:00"},{"alias_kind":"pith_short_12","alias_value":"AO5Z6HIL37OW","created_at":"2026-07-05T09:23:31.553591+00:00"},{"alias_kind":"pith_short_16","alias_value":"AO5Z6HIL37OW6AGO","created_at":"2026-07-05T09:23:31.553591+00:00"},{"alias_kind":"pith_short_8","alias_value":"AO5Z6HIL","created_at":"2026-07-05T09:23:31.553591+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.29128","citing_title":"Apertus LLM Family Expansion via Distillation and Quantization","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08738","citing_title":"SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08738","citing_title":"SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09924","citing_title":"Evolving Knowledge Distillation for Lightweight Neural Machine Translation","ref_index":36,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE","json":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE.json","graph_json":"https://pith.science/api/pith-number/AO5Z6HIL37OW6AGO2OERRFEGVE/graph.json","events_json":"https://pith.science/api/pith-number/AO5Z6HIL37OW6AGO2OERRFEGVE/events.json","paper":"https://pith.science/paper/AO5Z6HIL"},"agent_actions":{"view_html":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE","download_json":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE.json","view_paper":"https://pith.science/paper/AO5Z6HIL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.16215&json=true","fetch_graph":"https://pith.science/api/pith-number/AO5Z6HIL37OW6AGO2OERRFEGVE/graph.json","fetch_events":"https://pith.science/api/pith-number/AO5Z6HIL37OW6AGO2OERRFEGVE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE/action/storage_attestation","attest_author":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE/action/author_attestation","sign_citation":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE/action/citation_signature","submit_replication":"https://pith.science/pith/AO5Z6HIL37OW6AGO2OERRFEGVE/action/replication_record"}},"created_at":"2026-07-05T09:23:31.553591+00:00","updated_at":"2026-07-05T09:23:31.553591+00:00"}