{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:UID6RT3D7SKV7Z7X6E4QOXJGJR","short_pith_number":"pith:UID6RT3D","schema_version":"1.0","canonical_sha256":"a207e8cf63fc955fe7f7f139075d264c69e1228736e4b7a233c7a89f7c764bc3","source":{"kind":"arxiv","id":"2305.12129","version":1},"attestation_state":"computed","paper":{"title":"Lifting the Curse of Capacity Gap in Distilling Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Benyou Wang, Chen Zhang, Dawei Song, Jiahao Liu, Jingang Wang, Yang Yang, Yunsen Xian","submitted_at":"2023-05-20T07:30:55Z","abstract_excerpt":"Pretrained language models (LMs) have shown compelling performance on various downstream tasks, but unfortunately they require a tremendous amount of inference compute. Knowledge distillation finds a path to compress LMs to small ones with a teacher-student paradigm. However, when the capacity gap between the teacher and the student is large, a curse of capacity gap appears, invoking a deficiency in distilling LMs. While a few studies have been carried out to fill the gap, the curse is not yet well tackled. In this paper, we aim at lifting the curse of capacity gap via enlarging the capacity o"},"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":"2305.12129","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-20T07:30:55Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"56e8f30c6654c6f23715f55bf53e92bc162a7b457762596a80aa8bfebb4764e2","abstract_canon_sha256":"ea872d563f85e8325845193e3db17796ded82115c5dd6eb1db85c1b60b7b7a28"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:12:06.819143Z","signature_b64":"5NsuHwnZHoWV2CsA710GnUk0ErQ+lj69c42WRnLZTIDPc2qUI1jyubJ7IHjkSQZRv2Jotp4Fr3TVdNhaG3jmCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a207e8cf63fc955fe7f7f139075d264c69e1228736e4b7a233c7a89f7c764bc3","last_reissued_at":"2026-07-05T06:12:06.818662Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:12:06.818662Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Lifting the Curse of Capacity Gap in Distilling Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Benyou Wang, Chen Zhang, Dawei Song, Jiahao Liu, Jingang Wang, Yang Yang, Yunsen Xian","submitted_at":"2023-05-20T07:30:55Z","abstract_excerpt":"Pretrained language models (LMs) have shown compelling performance on various downstream tasks, but unfortunately they require a tremendous amount of inference compute. Knowledge distillation finds a path to compress LMs to small ones with a teacher-student paradigm. However, when the capacity gap between the teacher and the student is large, a curse of capacity gap appears, invoking a deficiency in distilling LMs. While a few studies have been carried out to fill the gap, the curse is not yet well tackled. In this paper, we aim at lifting the curse of capacity gap via enlarging the capacity o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.12129","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/2305.12129/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":"2305.12129","created_at":"2026-07-05T06:12:06.818721+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.12129v1","created_at":"2026-07-05T06:12:06.818721+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.12129","created_at":"2026-07-05T06:12:06.818721+00:00"},{"alias_kind":"pith_short_12","alias_value":"UID6RT3D7SKV","created_at":"2026-07-05T06:12:06.818721+00:00"},{"alias_kind":"pith_short_16","alias_value":"UID6RT3D7SKV7Z7X","created_at":"2026-07-05T06:12:06.818721+00:00"},{"alias_kind":"pith_short_8","alias_value":"UID6RT3D","created_at":"2026-07-05T06:12:06.818721+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2404.14294","citing_title":"A Survey on Efficient Inference for Large Language Models","ref_index":135,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR","json":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR.json","graph_json":"https://pith.science/api/pith-number/UID6RT3D7SKV7Z7X6E4QOXJGJR/graph.json","events_json":"https://pith.science/api/pith-number/UID6RT3D7SKV7Z7X6E4QOXJGJR/events.json","paper":"https://pith.science/paper/UID6RT3D"},"agent_actions":{"view_html":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR","download_json":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR.json","view_paper":"https://pith.science/paper/UID6RT3D","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.12129&json=true","fetch_graph":"https://pith.science/api/pith-number/UID6RT3D7SKV7Z7X6E4QOXJGJR/graph.json","fetch_events":"https://pith.science/api/pith-number/UID6RT3D7SKV7Z7X6E4QOXJGJR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR/action/storage_attestation","attest_author":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR/action/author_attestation","sign_citation":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR/action/citation_signature","submit_replication":"https://pith.science/pith/UID6RT3D7SKV7Z7X6E4QOXJGJR/action/replication_record"}},"created_at":"2026-07-05T06:12:06.818721+00:00","updated_at":"2026-07-05T06:12:06.818721+00:00"}