{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:CKESZHD7AOIG3HWXUMHSPJC2R6","short_pith_number":"pith:CKESZHD7","schema_version":"1.0","canonical_sha256":"12892c9c7f03906d9ed7a30f27a45a8fbe966a9249639edca5283669ed28c790","source":{"kind":"arxiv","id":"2311.08640","version":4},"attestation_state":"computed","paper":{"title":"Multistage Collaborative Knowledge Distillation from a Large Language Model for Semi-Supervised Sequence Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Andrew Drozdov, Andrew McCallum, Benjamin Rozonoyer, Jay-Yoon Lee, Jiachen Zhao, Md Arafat Sultan, Mohit Iyyer, Wenlong Zhao","submitted_at":"2023-11-15T01:28:28Z","abstract_excerpt":"We study semi-supervised sequence generation tasks, where the few labeled examples are too scarce to finetune a model, and meanwhile, few-shot prompted large language models (LLMs) exhibit room for improvement. In this paper, we present the discovery that a student model distilled from a few-shot prompted LLM can commonly generalize better than its teacher to unseen examples on such tasks. We find that the student is able to learn a general pattern from the high-quality pseudolabels produced by the teacher during knowledge distillation (KD), and favorably not a general pattern from the low-qua"},"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":"2311.08640","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-15T01:28:28Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"30560fd66542df310d52a3725e30037eaff34ba4898de58c2c3a403f1194a64a","abstract_canon_sha256":"87c4f9fb8a8c7642ba284976864ac46148149c56c20df45d1df8dfa9b091fd87"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:51:44.711239Z","signature_b64":"PL9B9xxMSeBihX+jugSlK7a8+YiZjRpfL/HZzz5P7ZOKa9EXZvLmoxnA4q91xnsLfKzaOBQTFvOImjefNEnoBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"12892c9c7f03906d9ed7a30f27a45a8fbe966a9249639edca5283669ed28c790","last_reissued_at":"2026-07-05T08:51:44.710769Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:51:44.710769Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multistage Collaborative Knowledge Distillation from a Large Language Model for Semi-Supervised Sequence Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Andrew Drozdov, Andrew McCallum, Benjamin Rozonoyer, Jay-Yoon Lee, Jiachen Zhao, Md Arafat Sultan, Mohit Iyyer, Wenlong Zhao","submitted_at":"2023-11-15T01:28:28Z","abstract_excerpt":"We study semi-supervised sequence generation tasks, where the few labeled examples are too scarce to finetune a model, and meanwhile, few-shot prompted large language models (LLMs) exhibit room for improvement. In this paper, we present the discovery that a student model distilled from a few-shot prompted LLM can commonly generalize better than its teacher to unseen examples on such tasks. We find that the student is able to learn a general pattern from the high-quality pseudolabels produced by the teacher during knowledge distillation (KD), and favorably not a general pattern from the low-qua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.08640","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/2311.08640/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":"2311.08640","created_at":"2026-07-05T08:51:44.710826+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.08640v4","created_at":"2026-07-05T08:51:44.710826+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.08640","created_at":"2026-07-05T08:51:44.710826+00:00"},{"alias_kind":"pith_short_12","alias_value":"CKESZHD7AOIG","created_at":"2026-07-05T08:51:44.710826+00:00"},{"alias_kind":"pith_short_16","alias_value":"CKESZHD7AOIG3HWX","created_at":"2026-07-05T08:51:44.710826+00:00"},{"alias_kind":"pith_short_8","alias_value":"CKESZHD7","created_at":"2026-07-05T08:51:44.710826+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":119,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6","json":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6.json","graph_json":"https://pith.science/api/pith-number/CKESZHD7AOIG3HWXUMHSPJC2R6/graph.json","events_json":"https://pith.science/api/pith-number/CKESZHD7AOIG3HWXUMHSPJC2R6/events.json","paper":"https://pith.science/paper/CKESZHD7"},"agent_actions":{"view_html":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6","download_json":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6.json","view_paper":"https://pith.science/paper/CKESZHD7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.08640&json=true","fetch_graph":"https://pith.science/api/pith-number/CKESZHD7AOIG3HWXUMHSPJC2R6/graph.json","fetch_events":"https://pith.science/api/pith-number/CKESZHD7AOIG3HWXUMHSPJC2R6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6/action/storage_attestation","attest_author":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6/action/author_attestation","sign_citation":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6/action/citation_signature","submit_replication":"https://pith.science/pith/CKESZHD7AOIG3HWXUMHSPJC2R6/action/replication_record"}},"created_at":"2026-07-05T08:51:44.710826+00:00","updated_at":"2026-07-05T08:51:44.710826+00:00"}