{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:DIZJS7MCJNJA6HYLWNQM3DQZZ4","short_pith_number":"pith:DIZJS7MC","schema_version":"1.0","canonical_sha256":"1a32997d824b520f1f0bb360cd8e19cf27f8777c44f30ad723ee8da2f7f10d0d","source":{"kind":"arxiv","id":"2608.03632","version":1},"attestation_state":"computed","paper":{"title":"When Teachers Mislead: Spurious-Signal-Aware On-Policy Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Huajun Chen, Qiang Zhang, Tiankai Li, Xiang Zhuang, Yinuo Jiang, YongJie Ye, Zhou Tao","submitted_at":"2026-08-04T13:17:30Z","abstract_excerpt":"On-Policy distillation (OPD) transfers teacher capabilities by supervising student-sampled trajectories with dense token-level teacher signals. Recent selective OPD methods improve this process by prioritizing signals that are confident, informative, or learnable. However, the assumptions overlook a fundamental failure mode of language models: their token-level judgments can be driven by input-agnostic language priors, formatting conventions, or stereotyped reasoning templates rather than task-specific evidence. We refer to such optimization-relevant but weakly input-grounded supervision as sp"},"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":"2608.03632","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-04T13:17:30Z","cross_cats_sorted":[],"title_canon_sha256":"3b68384ae46d0848ac04bd6c35b0f38b48015662990a620c4fb47a2e2c5d90c5","abstract_canon_sha256":"56287ddf81849f8b7cd85d8d88260db65387699cfe612f7b86e42571d7b58bfe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T01:36:47.575225Z","signature_b64":"XPOOk+ENWl0yZA2L+PlNNJd7I5lDvD/C82l9jhVHHz2CFxdSigZ3QrCuS1kosLokjA3omb0STCtXXWfqeTDnDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1a32997d824b520f1f0bb360cd8e19cf27f8777c44f30ad723ee8da2f7f10d0d","last_reissued_at":"2026-08-05T01:36:47.573739Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T01:36:47.573739Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Teachers Mislead: Spurious-Signal-Aware On-Policy Distillation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Huajun Chen, Qiang Zhang, Tiankai Li, Xiang Zhuang, Yinuo Jiang, YongJie Ye, Zhou Tao","submitted_at":"2026-08-04T13:17:30Z","abstract_excerpt":"On-Policy distillation (OPD) transfers teacher capabilities by supervising student-sampled trajectories with dense token-level teacher signals. Recent selective OPD methods improve this process by prioritizing signals that are confident, informative, or learnable. However, the assumptions overlook a fundamental failure mode of language models: their token-level judgments can be driven by input-agnostic language priors, formatting conventions, or stereotyped reasoning templates rather than task-specific evidence. We refer to such optimization-relevant but weakly input-grounded supervision as sp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.03632","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/2608.03632/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":"2608.03632","created_at":"2026-08-05T01:36:47.574240+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.03632v1","created_at":"2026-08-05T01:36:47.574240+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.03632","created_at":"2026-08-05T01:36:47.574240+00:00"},{"alias_kind":"pith_short_12","alias_value":"DIZJS7MCJNJA","created_at":"2026-08-05T01:36:47.574240+00:00"},{"alias_kind":"pith_short_16","alias_value":"DIZJS7MCJNJA6HYL","created_at":"2026-08-05T01:36:47.574240+00:00"},{"alias_kind":"pith_short_8","alias_value":"DIZJS7MC","created_at":"2026-08-05T01:36:47.574240+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/DIZJS7MCJNJA6HYLWNQM3DQZZ4","json":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4.json","graph_json":"https://pith.science/api/pith-number/DIZJS7MCJNJA6HYLWNQM3DQZZ4/graph.json","events_json":"https://pith.science/api/pith-number/DIZJS7MCJNJA6HYLWNQM3DQZZ4/events.json","paper":"https://pith.science/paper/DIZJS7MC"},"agent_actions":{"view_html":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4","download_json":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4.json","view_paper":"https://pith.science/paper/DIZJS7MC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.03632&json=true","fetch_graph":"https://pith.science/api/pith-number/DIZJS7MCJNJA6HYLWNQM3DQZZ4/graph.json","fetch_events":"https://pith.science/api/pith-number/DIZJS7MCJNJA6HYLWNQM3DQZZ4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4/action/storage_attestation","attest_author":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4/action/author_attestation","sign_citation":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4/action/citation_signature","submit_replication":"https://pith.science/pith/DIZJS7MCJNJA6HYLWNQM3DQZZ4/action/replication_record"}},"created_at":"2026-08-05T01:36:47.574240+00:00","updated_at":"2026-08-05T01:36:47.574240+00:00"}