{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:WWBSSZ5VMCYJ427E4SFUZVTO7L","short_pith_number":"pith:WWBSSZ5V","schema_version":"1.0","canonical_sha256":"b5832967b560b09e6be4e48b4cd66efae9003a5c65cbda9433ed979caaaf3bf3","source":{"kind":"arxiv","id":"2606.23104","version":1},"attestation_state":"computed","paper":{"title":"ReNIO: Reweighting Negative Trajectory Importance for LLM On-Policy Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chen Lin, Kedi Chen, Wei Zhang","submitted_at":"2026-06-22T09:46:35Z","abstract_excerpt":"On-policy distillation (OPD) improves LLM reasoning by training a student model on its own generated outputs, but standard OPD treats all student-generated outputs (SGOs) equally regardless of their informativeness. We observe a consistent asymmetry in controlled filtering experiments: in both OPD and on-policy self distillation (OPSD), training only on incorrect SGOs outperforms training only on correct ones. Our further analysis suggests that models trained on correct-only SGOs tend to generate shorter reasoning traces and show weaker reflection behavior, while incorrect SGOs better preserve"},"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":"2606.23104","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-22T09:46:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9062c62a5bfe889d85b58c7558e144de278ff835c4d902c8b52e0783dd303bf5","abstract_canon_sha256":"0d15857deb3b4f592c9c9ce4c186248e5f195361e2b59b91f94a99a882894531"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T03:14:10.013239Z","signature_b64":"mtL7txGu2EQvzspQUUMkKEYm7ASTZUcZdlbYMJLTq/jNIjNsCvZLMeSQduNAqZuOuPKl2MYaeBs7r15P+289Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5832967b560b09e6be4e48b4cd66efae9003a5c65cbda9433ed979caaaf3bf3","last_reissued_at":"2026-06-23T03:14:10.012840Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T03:14:10.012840Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ReNIO: Reweighting Negative Trajectory Importance for LLM On-Policy Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chen Lin, Kedi Chen, Wei Zhang","submitted_at":"2026-06-22T09:46:35Z","abstract_excerpt":"On-policy distillation (OPD) improves LLM reasoning by training a student model on its own generated outputs, but standard OPD treats all student-generated outputs (SGOs) equally regardless of their informativeness. We observe a consistent asymmetry in controlled filtering experiments: in both OPD and on-policy self distillation (OPSD), training only on incorrect SGOs outperforms training only on correct ones. Our further analysis suggests that models trained on correct-only SGOs tend to generate shorter reasoning traces and show weaker reflection behavior, while incorrect SGOs better preserve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.23104","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/2606.23104/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":"2606.23104","created_at":"2026-06-23T03:14:10.012901+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.23104v1","created_at":"2026-06-23T03:14:10.012901+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.23104","created_at":"2026-06-23T03:14:10.012901+00:00"},{"alias_kind":"pith_short_12","alias_value":"WWBSSZ5VMCYJ","created_at":"2026-06-23T03:14:10.012901+00:00"},{"alias_kind":"pith_short_16","alias_value":"WWBSSZ5VMCYJ427E","created_at":"2026-06-23T03:14:10.012901+00:00"},{"alias_kind":"pith_short_8","alias_value":"WWBSSZ5V","created_at":"2026-06-23T03:14:10.012901+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L","json":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L.json","graph_json":"https://pith.science/api/pith-number/WWBSSZ5VMCYJ427E4SFUZVTO7L/graph.json","events_json":"https://pith.science/api/pith-number/WWBSSZ5VMCYJ427E4SFUZVTO7L/events.json","paper":"https://pith.science/paper/WWBSSZ5V"},"agent_actions":{"view_html":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L","download_json":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L.json","view_paper":"https://pith.science/paper/WWBSSZ5V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.23104&json=true","fetch_graph":"https://pith.science/api/pith-number/WWBSSZ5VMCYJ427E4SFUZVTO7L/graph.json","fetch_events":"https://pith.science/api/pith-number/WWBSSZ5VMCYJ427E4SFUZVTO7L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L/action/storage_attestation","attest_author":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L/action/author_attestation","sign_citation":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L/action/citation_signature","submit_replication":"https://pith.science/pith/WWBSSZ5VMCYJ427E4SFUZVTO7L/action/replication_record"}},"created_at":"2026-06-23T03:14:10.012901+00:00","updated_at":"2026-06-23T03:14:10.012901+00:00"}