{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ELYJIHBD2RXK7GCHLEOFHXX3OB","short_pith_number":"pith:ELYJIHBD","schema_version":"1.0","canonical_sha256":"22f0941c23d46eaf9847591c53defb7067fd2a1135b2db9b6e1c0f51a7aa8522","source":{"kind":"arxiv","id":"2406.14491","version":2},"attestation_state":"computed","paper":{"title":"Instruction Pre-Training: Language Models are Supervised Multitask Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daixuan Cheng, Furu Wei, Junyu Bi, Minlie Huang, Shaohan Huang, Yuxian Gu","submitted_at":"2024-06-20T16:55:33Z","abstract_excerpt":"Unsupervised multitask pre-training has been the critical method behind the recent success of language models (LMs). However, supervised multitask learning still holds significant promise, as scaling it in the post-training stage trends towards better generalization. In this paper, we explore supervised multitask pre-training by proposing Instruction Pre-Training, a framework that scalably augments massive raw corpora with instruction-response pairs to pre-train LMs. The instruction-response pairs are generated by an efficient instruction synthesizer built on open-source models. In our experim"},"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":"2406.14491","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-20T16:55:33Z","cross_cats_sorted":[],"title_canon_sha256":"4e1a4c1de3d856a2bae66895c6403c411729b8ed390c29c5c7ddd8137fd8b45c","abstract_canon_sha256":"2ee3434b69843835386285dc02eacb289ee1ce403d57c7a04ea9b1ea9d38fd02"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:41:38.252413Z","signature_b64":"hM9bDIbE2Z8vSg85yQRHseNVolgu8cyZY/iTf2/pn0LXXaMfqbwGUHIkX69ydvZr2hSvcp0XVLheTEfM+vjiAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"22f0941c23d46eaf9847591c53defb7067fd2a1135b2db9b6e1c0f51a7aa8522","last_reissued_at":"2026-07-05T09:41:38.251911Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:41:38.251911Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Instruction Pre-Training: Language Models are Supervised Multitask Learners","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daixuan Cheng, Furu Wei, Junyu Bi, Minlie Huang, Shaohan Huang, Yuxian Gu","submitted_at":"2024-06-20T16:55:33Z","abstract_excerpt":"Unsupervised multitask pre-training has been the critical method behind the recent success of language models (LMs). However, supervised multitask learning still holds significant promise, as scaling it in the post-training stage trends towards better generalization. In this paper, we explore supervised multitask pre-training by proposing Instruction Pre-Training, a framework that scalably augments massive raw corpora with instruction-response pairs to pre-train LMs. The instruction-response pairs are generated by an efficient instruction synthesizer built on open-source models. In our experim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.14491","kind":"arxiv","version":2},"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/2406.14491/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":"2406.14491","created_at":"2026-07-05T09:41:38.251982+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.14491v2","created_at":"2026-07-05T09:41:38.251982+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.14491","created_at":"2026-07-05T09:41:38.251982+00:00"},{"alias_kind":"pith_short_12","alias_value":"ELYJIHBD2RXK","created_at":"2026-07-05T09:41:38.251982+00:00"},{"alias_kind":"pith_short_16","alias_value":"ELYJIHBD2RXK7GCH","created_at":"2026-07-05T09:41:38.251982+00:00"},{"alias_kind":"pith_short_8","alias_value":"ELYJIHBD","created_at":"2026-07-05T09:41:38.251982+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2509.07177","citing_title":"Towards EnergyGPT: A Large Language Model Specialized for the Energy Sector","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB","json":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB.json","graph_json":"https://pith.science/api/pith-number/ELYJIHBD2RXK7GCHLEOFHXX3OB/graph.json","events_json":"https://pith.science/api/pith-number/ELYJIHBD2RXK7GCHLEOFHXX3OB/events.json","paper":"https://pith.science/paper/ELYJIHBD"},"agent_actions":{"view_html":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB","download_json":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB.json","view_paper":"https://pith.science/paper/ELYJIHBD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.14491&json=true","fetch_graph":"https://pith.science/api/pith-number/ELYJIHBD2RXK7GCHLEOFHXX3OB/graph.json","fetch_events":"https://pith.science/api/pith-number/ELYJIHBD2RXK7GCHLEOFHXX3OB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB/action/storage_attestation","attest_author":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB/action/author_attestation","sign_citation":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB/action/citation_signature","submit_replication":"https://pith.science/pith/ELYJIHBD2RXK7GCHLEOFHXX3OB/action/replication_record"}},"created_at":"2026-07-05T09:41:38.251982+00:00","updated_at":"2026-07-05T09:41:38.251982+00:00"}