{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:2WDWFUCL2OPME57LOUKZWPKLTU","short_pith_number":"pith:2WDWFUCL","schema_version":"1.0","canonical_sha256":"d58762d04bd39ec277eb75159b3d4b9d038fed3949f2de2efa882f62cc230bdb","source":{"kind":"arxiv","id":"2509.14257","version":3},"attestation_state":"computed","paper":{"title":"Student-Centered Distillation Narrows the Agentic Gap Between Small and Large LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chengyu Wang, Jun Huang, Tong Xu, Yuanjie Lyu","submitted_at":"2025-09-12T15:34:07Z","abstract_excerpt":"Large Language Model agents achieve strong performance on multi-step reasoning and tool-use tasks, but their impressive capabilities typically rely on extremely large backbones. Existing distillation approaches train smaller students to imitate full teacher trajectories, yet reasoning and knowledge gaps between the teacher and student can cause compounding errors. We propose SCoRe, a student-centered framework in which the student generates training trajectories and the teacher corrects only the earliest error, producing training data matched to the student's abilities and exposing specific we"},"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":"2509.14257","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-09-12T15:34:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1363655b4f390c46fc4512ae17478b12cac85f89a9499e9b5341db78e3b55ff8","abstract_canon_sha256":"20bba2206db5348d4fa69f13dda4319c0c0c015b67fa4e3e136bb372070930c3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T01:24:03.581800Z","signature_b64":"QLxZGtcPTGN76gd1ZeBLnnGKMxSkRyhyHtoPb228gsaaq4I1F417f+C7fdSHoW2YTYukZNeg2K9eIxbBFtRCAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d58762d04bd39ec277eb75159b3d4b9d038fed3949f2de2efa882f62cc230bdb","last_reissued_at":"2026-07-24T01:24:03.580853Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T01:24:03.580853Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Student-Centered Distillation Narrows the Agentic Gap Between Small and Large LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chengyu Wang, Jun Huang, Tong Xu, Yuanjie Lyu","submitted_at":"2025-09-12T15:34:07Z","abstract_excerpt":"Large Language Model agents achieve strong performance on multi-step reasoning and tool-use tasks, but their impressive capabilities typically rely on extremely large backbones. Existing distillation approaches train smaller students to imitate full teacher trajectories, yet reasoning and knowledge gaps between the teacher and student can cause compounding errors. We propose SCoRe, a student-centered framework in which the student generates training trajectories and the teacher corrects only the earliest error, producing training data matched to the student's abilities and exposing specific we"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.14257","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/2509.14257/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":"2509.14257","created_at":"2026-07-24T01:24:03.581282+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.14257v3","created_at":"2026-07-24T01:24:03.581282+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.14257","created_at":"2026-07-24T01:24:03.581282+00:00"},{"alias_kind":"pith_short_12","alias_value":"2WDWFUCL2OPM","created_at":"2026-07-24T01:24:03.581282+00:00"},{"alias_kind":"pith_short_16","alias_value":"2WDWFUCL2OPME57L","created_at":"2026-07-24T01:24:03.581282+00:00"},{"alias_kind":"pith_short_8","alias_value":"2WDWFUCL","created_at":"2026-07-24T01:24:03.581282+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":8,"sample":[{"citing_arxiv_id":"2606.23104","citing_title":"ReNIO: Reweighting Negative Trajectory Importance for LLM On-Policy Distillation","ref_index":77,"is_internal_anchor":true},{"citing_arxiv_id":"2605.28775","citing_title":"Learn from Weaknesses: Automated Domain Specialization for Small Computer-Use Agents","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2606.06840","citing_title":"Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces","ref_index":162,"is_internal_anchor":true},{"citing_arxiv_id":"2604.00626","citing_title":"A Survey of On-Policy Distillation for Large Language Models","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12652","citing_title":"Multi-Rollout On-Policy Distillation via Peer Successes and Failures","ref_index":65,"is_internal_anchor":true},{"citing_arxiv_id":"2604.00626","citing_title":"A Survey of On-Policy Distillation for Large Language Models","ref_index":5,"is_internal_anchor":true},{"citing_arxiv_id":"2604.21590","citing_title":"AgenticQwen: Training Small Agentic Language Models with Dual Data Flywheels for Industrial-Scale Tool Use","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07276","citing_title":"Signal Reshaping for GRPO in Weak-Feedback Agentic Code Repair","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU","json":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU.json","graph_json":"https://pith.science/api/pith-number/2WDWFUCL2OPME57LOUKZWPKLTU/graph.json","events_json":"https://pith.science/api/pith-number/2WDWFUCL2OPME57LOUKZWPKLTU/events.json","paper":"https://pith.science/paper/2WDWFUCL"},"agent_actions":{"view_html":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU","download_json":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU.json","view_paper":"https://pith.science/paper/2WDWFUCL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.14257&json=true","fetch_graph":"https://pith.science/api/pith-number/2WDWFUCL2OPME57LOUKZWPKLTU/graph.json","fetch_events":"https://pith.science/api/pith-number/2WDWFUCL2OPME57LOUKZWPKLTU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU/action/storage_attestation","attest_author":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU/action/author_attestation","sign_citation":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU/action/citation_signature","submit_replication":"https://pith.science/pith/2WDWFUCL2OPME57LOUKZWPKLTU/action/replication_record"}},"created_at":"2026-07-24T01:24:03.581282+00:00","updated_at":"2026-07-24T01:24:03.581282+00:00"}