{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:WRQFCMCRT5UV34EB7LQRQA65LE","short_pith_number":"pith:WRQFCMCR","schema_version":"1.0","canonical_sha256":"b4605130519f695df081fae11803dd593751aa8219eb936e2f48b2c0ef9944c6","source":{"kind":"arxiv","id":"2202.00828","version":1},"attestation_state":"computed","paper":{"title":"Co-training Improves Prompt-based Learning for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"David Sontag, Hunter Lang, Monica Agrawal, Yoon Kim","submitted_at":"2022-02-02T00:48:26Z","abstract_excerpt":"We demonstrate that co-training (Blum & Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data. While prompting has emerged as a promising paradigm for few-shot and zero-shot learning, it is often brittle and requires much larger models compared to the standard supervised setup. We find that co-training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model. In the case where we only have partial access to a prompt model (e.g., output probabilities from GPT-3 (Brown et al., 2020)) we lear"},"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":"2202.00828","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-02-02T00:48:26Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"9793896cf4db251885a64ad8ed2cd1bbdc48ebc5d390abeddde71265689c3302","abstract_canon_sha256":"ed11857b304e90e52a6454ff2b5b6c2b78196cffe646dcd5b8a272ae3a941d6e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:53:28.047236Z","signature_b64":"7evRr8Eodkos2K1fu6xgr5FeciK1Arpf0beohmJHuuv1uEpTgL7uEUYNDGBSGYEdShAkOvj1h4Sp2QWtyEKQAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b4605130519f695df081fae11803dd593751aa8219eb936e2f48b2c0ef9944c6","last_reissued_at":"2026-07-05T03:53:28.046779Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:53:28.046779Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Co-training Improves Prompt-based Learning for Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"David Sontag, Hunter Lang, Monica Agrawal, Yoon Kim","submitted_at":"2022-02-02T00:48:26Z","abstract_excerpt":"We demonstrate that co-training (Blum & Mitchell, 1998) can improve the performance of prompt-based learning by using unlabeled data. While prompting has emerged as a promising paradigm for few-shot and zero-shot learning, it is often brittle and requires much larger models compared to the standard supervised setup. We find that co-training makes it possible to improve the original prompt model and at the same time learn a smaller, downstream task-specific model. In the case where we only have partial access to a prompt model (e.g., output probabilities from GPT-3 (Brown et al., 2020)) we lear"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2202.00828","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/2202.00828/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":"2202.00828","created_at":"2026-07-05T03:53:28.046834+00:00"},{"alias_kind":"arxiv_version","alias_value":"2202.00828v1","created_at":"2026-07-05T03:53:28.046834+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2202.00828","created_at":"2026-07-05T03:53:28.046834+00:00"},{"alias_kind":"pith_short_12","alias_value":"WRQFCMCRT5UV","created_at":"2026-07-05T03:53:28.046834+00:00"},{"alias_kind":"pith_short_16","alias_value":"WRQFCMCRT5UV34EB","created_at":"2026-07-05T03:53:28.046834+00:00"},{"alias_kind":"pith_short_8","alias_value":"WRQFCMCR","created_at":"2026-07-05T03:53:28.046834+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.08243","citing_title":"Training Strategies for Efficient Embodied Reasoning","ref_index":69,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE","json":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE.json","graph_json":"https://pith.science/api/pith-number/WRQFCMCRT5UV34EB7LQRQA65LE/graph.json","events_json":"https://pith.science/api/pith-number/WRQFCMCRT5UV34EB7LQRQA65LE/events.json","paper":"https://pith.science/paper/WRQFCMCR"},"agent_actions":{"view_html":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE","download_json":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE.json","view_paper":"https://pith.science/paper/WRQFCMCR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2202.00828&json=true","fetch_graph":"https://pith.science/api/pith-number/WRQFCMCRT5UV34EB7LQRQA65LE/graph.json","fetch_events":"https://pith.science/api/pith-number/WRQFCMCRT5UV34EB7LQRQA65LE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE/action/storage_attestation","attest_author":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE/action/author_attestation","sign_citation":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE/action/citation_signature","submit_replication":"https://pith.science/pith/WRQFCMCRT5UV34EB7LQRQA65LE/action/replication_record"}},"created_at":"2026-07-05T03:53:28.046834+00:00","updated_at":"2026-07-05T03:53:28.046834+00:00"}