{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RAF46FCVWXXJT44MRTS4LN5QXR","short_pith_number":"pith:RAF46FCV","schema_version":"1.0","canonical_sha256":"880bcf1455b5ee99f38c8ce5c5b7b0bc72249e8736b1320407a5f8fa0900e4c7","source":{"kind":"arxiv","id":"2411.07133","version":3},"attestation_state":"computed","paper":{"title":"Stronger Models are NOT Stronger Teachers for Instruction Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bill Yuchen Lin, Fengqing Jiang, Luyao Niu, Radha Poovendran, Zhangchen Xu","submitted_at":"2024-11-11T17:06:48Z","abstract_excerpt":"Instruction tuning has been widely adopted to ensure large language models (LLMs) follow user instructions effectively. The resulting instruction-following capabilities of LLMs heavily rely on the instruction datasets used for tuning. Recently, synthetic instruction datasets have emerged as an economically viable solution to provide LLMs diverse and high-quality instructions. However, existing approaches typically assume that larger or stronger models are stronger teachers for instruction tuning, and hence simply adopt these models as response generators to the synthetic instructions. In this "},"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":"2411.07133","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2024-11-11T17:06:48Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"cd6741c35ffb0d30f7fb0e302e2b4ca5e3ab6eeda2c697d7f31a91daf5a1cc60","abstract_canon_sha256":"757341a9d75e6cc8aad39db3e1faec09b916d0d76fe62e4079a10b608d2e83b3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:20:17.524065Z","signature_b64":"jC15MIFBvy+3HIFCfu3bVsNdMYwyuNCMw1pDIT1dQeoea/0xFNgaBBKR2xlz7ry4mVWocH3mxKjaBXq1Wh6BAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"880bcf1455b5ee99f38c8ce5c5b7b0bc72249e8736b1320407a5f8fa0900e4c7","last_reissued_at":"2026-07-05T10:20:17.523414Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:20:17.523414Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stronger Models are NOT Stronger Teachers for Instruction Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bill Yuchen Lin, Fengqing Jiang, Luyao Niu, Radha Poovendran, Zhangchen Xu","submitted_at":"2024-11-11T17:06:48Z","abstract_excerpt":"Instruction tuning has been widely adopted to ensure large language models (LLMs) follow user instructions effectively. The resulting instruction-following capabilities of LLMs heavily rely on the instruction datasets used for tuning. Recently, synthetic instruction datasets have emerged as an economically viable solution to provide LLMs diverse and high-quality instructions. However, existing approaches typically assume that larger or stronger models are stronger teachers for instruction tuning, and hence simply adopt these models as response generators to the synthetic instructions. In this "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.07133","kind":"arxiv","version":3},"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/2411.07133/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":"2411.07133","created_at":"2026-07-05T10:20:17.523484+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.07133v3","created_at":"2026-07-05T10:20:17.523484+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.07133","created_at":"2026-07-05T10:20:17.523484+00:00"},{"alias_kind":"pith_short_12","alias_value":"RAF46FCVWXXJ","created_at":"2026-07-05T10:20:17.523484+00:00"},{"alias_kind":"pith_short_16","alias_value":"RAF46FCVWXXJT44M","created_at":"2026-07-05T10:20:17.523484+00:00"},{"alias_kind":"pith_short_8","alias_value":"RAF46FCV","created_at":"2026-07-05T10:20:17.523484+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20700","citing_title":"Beyond Templates: Dynamic Adaptation of Reasoning Demonstrations via Feasibility-Aware Exploration","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR","json":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR.json","graph_json":"https://pith.science/api/pith-number/RAF46FCVWXXJT44MRTS4LN5QXR/graph.json","events_json":"https://pith.science/api/pith-number/RAF46FCVWXXJT44MRTS4LN5QXR/events.json","paper":"https://pith.science/paper/RAF46FCV"},"agent_actions":{"view_html":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR","download_json":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR.json","view_paper":"https://pith.science/paper/RAF46FCV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.07133&json=true","fetch_graph":"https://pith.science/api/pith-number/RAF46FCVWXXJT44MRTS4LN5QXR/graph.json","fetch_events":"https://pith.science/api/pith-number/RAF46FCVWXXJT44MRTS4LN5QXR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR/action/storage_attestation","attest_author":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR/action/author_attestation","sign_citation":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR/action/citation_signature","submit_replication":"https://pith.science/pith/RAF46FCVWXXJT44MRTS4LN5QXR/action/replication_record"}},"created_at":"2026-07-05T10:20:17.523484+00:00","updated_at":"2026-07-05T10:20:17.523484+00:00"}