{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:V4U3Z6QIHPVYV335CXUALYL6CH","short_pith_number":"pith:V4U3Z6QI","schema_version":"1.0","canonical_sha256":"af29bcfa083beb8aef7d15e805e17e11c50bb8c86de2a348f72ee641fd94d0c1","source":{"kind":"arxiv","id":"2502.11766","version":1},"attestation_state":"computed","paper":{"title":"Warmup-Distill: Bridge the Distribution Mismatch between Teacher and Student before Knowledge Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fandong Meng, Jie Zhou, Jinan Xu, Yijin Liu, Yufeng Chen, Zengkui Sun","submitted_at":"2025-02-17T12:58:12Z","abstract_excerpt":"The widespread deployment of Large Language Models (LLMs) is hindered by the high computational demands, making knowledge distillation (KD) crucial for developing compact smaller ones. However, the conventional KD methods endure the distribution mismatch issue between the teacher and student models, leading to the poor performance of distillation. For instance, the widely-used KL-based methods suffer the mode-averaging and mode-collapsing problems, since the mismatched probabitliy distribution between both models. Previous studies mainly optimize this issue via different distance calculations "},"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":"2502.11766","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-17T12:58:12Z","cross_cats_sorted":[],"title_canon_sha256":"e2c5e01f2aa1dd2191c5e314b56f919137c1ebec6cf894d717240fa4159e7455","abstract_canon_sha256":"ef78c952f164e27842d552ba111932b20267fc91acb02d7f86f9c1b67909520e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:33.827074Z","signature_b64":"Hll9sLYbMdQaWCNx8p+hQ1X7rsyH87+yn26u58fD25lZweaRAynWg/nG8fo6gxUIjLs2cMIdfaQDaOrN5VezDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"af29bcfa083beb8aef7d15e805e17e11c50bb8c86de2a348f72ee641fd94d0c1","last_reissued_at":"2026-07-05T10:15:33.826652Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:33.826652Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Warmup-Distill: Bridge the Distribution Mismatch between Teacher and Student before Knowledge Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fandong Meng, Jie Zhou, Jinan Xu, Yijin Liu, Yufeng Chen, Zengkui Sun","submitted_at":"2025-02-17T12:58:12Z","abstract_excerpt":"The widespread deployment of Large Language Models (LLMs) is hindered by the high computational demands, making knowledge distillation (KD) crucial for developing compact smaller ones. However, the conventional KD methods endure the distribution mismatch issue between the teacher and student models, leading to the poor performance of distillation. For instance, the widely-used KL-based methods suffer the mode-averaging and mode-collapsing problems, since the mismatched probabitliy distribution between both models. Previous studies mainly optimize this issue via different distance calculations "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.11766","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/2502.11766/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":"2502.11766","created_at":"2026-07-05T10:15:33.826710+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.11766v1","created_at":"2026-07-05T10:15:33.826710+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.11766","created_at":"2026-07-05T10:15:33.826710+00:00"},{"alias_kind":"pith_short_12","alias_value":"V4U3Z6QIHPVY","created_at":"2026-07-05T10:15:33.826710+00:00"},{"alias_kind":"pith_short_16","alias_value":"V4U3Z6QIHPVYV335","created_at":"2026-07-05T10:15:33.826710+00:00"},{"alias_kind":"pith_short_8","alias_value":"V4U3Z6QI","created_at":"2026-07-05T10:15:33.826710+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH","json":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH.json","graph_json":"https://pith.science/api/pith-number/V4U3Z6QIHPVYV335CXUALYL6CH/graph.json","events_json":"https://pith.science/api/pith-number/V4U3Z6QIHPVYV335CXUALYL6CH/events.json","paper":"https://pith.science/paper/V4U3Z6QI"},"agent_actions":{"view_html":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH","download_json":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH.json","view_paper":"https://pith.science/paper/V4U3Z6QI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.11766&json=true","fetch_graph":"https://pith.science/api/pith-number/V4U3Z6QIHPVYV335CXUALYL6CH/graph.json","fetch_events":"https://pith.science/api/pith-number/V4U3Z6QIHPVYV335CXUALYL6CH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH/action/storage_attestation","attest_author":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH/action/author_attestation","sign_citation":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH/action/citation_signature","submit_replication":"https://pith.science/pith/V4U3Z6QIHPVYV335CXUALYL6CH/action/replication_record"}},"created_at":"2026-07-05T10:15:33.826710+00:00","updated_at":"2026-07-05T10:15:33.826710+00:00"}