{"id":"067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c","title":"Bootstrapped On-Policy Distillation","argument":"The authors of “Bootstrapped On-Policy Distillation” argue that on-policy self-distillation (OPSD) can become less effective as training continues: a frozen privileged teacher may drift away from the evolving student and give misleading supervision. They propose Bootstrapped On-Policy Distillation (BOPD), which first uses OPSD to transfer task-relevant information into an intermediate model, then distills from that model into a fresh copy of the initial model. The idea is to use reference-conditioned guidance early, then pass on what the adapted model learned without reference conditioning. The authors report that, across models, BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers. Their results suggest this two-stage transfer can make privileged reference supervision more useful for those tasks. The abstract does not provide the experimental details needed to assess the size or durability of the reported gains. Will BOPD’s reported gains persist under continued training?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The available evidence is the arXiv page abstract; it does not give experimental details such as the models, evaluation setup, or size of the reported gains. The abstract reports improvements across models but does not establish whether those gains persist under continued training.","speaker":{"name":"Wu, Yue, Zhou, Xueyu, Hu, Yangrong, Huang, Jian","roleAtSource":"Article author"},"source":{"title":"Bootstrapped On-Policy Distillation","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.13680","publishedAt":"2026-11-10","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors say OPSD can lose effectiveness during continued training because the frozen privileged teacher becomes increasingly mismatched with the evolving student and may provide misleading supervision.","excerpt":"as training progresses, the growing distributional mismatch between the privileged teacher and reference-free student can cause the teacher to provide misleading supervision, ultimately degrading student performance.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"BOPD uses OPSD to transfer task-relevant information into an intermediate model, then distills that model into a fresh copy of the initial model.","excerpt":"first uses OPSD to internalize task-relevant information from reference-conditioned supervision into an intermediate model, and then distills this reference-free intermediate model into a fresh copy of the initial model.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"The authors report that BOPD improves mathematical reasoning and code generation over both initial models and constructed OPSD teachers across models.","excerpt":"Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"The authors’ proposed sequence is to use privileged reference supervision to acquire useful knowledge, then transfer it through a model without reference conditioning.","excerpt":"privileged reference supervision can be more effectively exploited by first transferring its task-relevant signal into a task-adapted model and then distilling from that model without reference conditioning.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13680","checkedAt":"2026-11-12T00:04:07.850Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Will BOPD’s reported gains persist under continued training?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Bootstrapped On-Policy Distillation","summary":"The authors of “Bootstrapped On-Policy Distillation” argue that on-policy self-distillation (OPSD) can become less effective as training continues: a frozen privileged teacher may drift away from the evolving student and give misleading supervision. They propose Bootstrapped On-Policy Distillation (BOPD), which first uses OPSD to transfer task-relevant information into an intermediate model, then distills from that model into a fresh copy of the initial model. The idea is to use reference-conditioned guidance early, then pass on what the adapted model learned without reference conditioning. The authors report that, across models, BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers. Their results suggest this two-stage transfer can make privileged reference supervision more useful for those tasks. The abstract does not provide the experimental details needed to assess the size or durability of the reported gains. Will BOPD’s reported gains persist under continued training?","question":"Will BOPD’s reported gains persist under continued training?","claims":[{"text":"The authors say OPSD can lose effectiveness during continued training because the frozen privileged teacher becomes increasingly mismatched with the evolving student and may provide misleading supervision.","excerpt":"as training progresses, the growing distributional mismatch between the privileged teacher and reference-free student can cause the teacher to provide misleading supervision, ultimately degrading student performance.","support":"source-stated"},{"text":"BOPD uses OPSD to transfer task-relevant information into an intermediate model, then distills that model into a fresh copy of the initial model.","excerpt":"first uses OPSD to internalize task-relevant information from reference-conditioned supervision into an intermediate model, and then distills this reference-free intermediate model into a fresh copy of the initial model.","support":"source-stated"},{"text":"The authors report that BOPD improves mathematical reasoning and code generation over both initial models and constructed OPSD teachers across models.","excerpt":"Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers.","support":"source-stated"},{"text":"The authors’ proposed sequence is to use privileged reference supervision to acquire useful knowledge, then transfer it through a model without reference conditioning.","excerpt":"privileged reference supervision can be more effectively exploited by first transferring its task-relevant signal into a task-adapted model and then distilling from that model without reference conditioning.","support":"source-stated"}],"limitations":["The available evidence is the arXiv page abstract; it does not give experimental details such as the models, evaluation setup, or size of the reported gains.","The abstract reports improvements across models but does not establish whether those gains persist under continued training."],"mediaId":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","correctionReason":null,"id":"067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c","publicationId":"1bcf077e-b2cc-451d-84b9-e6a09991ed73","version":1,"sourceUrl":"https://arxiv.org/abs/2611.13680","sourceTitle":"Bootstrapped On-Policy Distillation","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-10","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:09:26.440Z","createdAt":"2026-11-11T23:32:18.637Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","kind":"article","url":"https://arxiv.org/abs/2611.13680","sourceUrl":"https://arxiv.org/abs/2611.13680","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T23:31:20.181Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13680","checkedAt":"2026-11-12T00:04:07.850Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","confidence":0.8500000000000001,"provider":"typesafe","model":"jev-1.13.0","inputHash":"76f65fd2a97f33ffb8b3fff575349ce7cffd5ca53f1c8361da285cb3eecabe72","decidedAt":"2026-11-11T23:31:22.460Z"},"provenance":{"textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observationId":"ce90daaaa104e1bc0bb537476fc1204f1d5429f19f7efe8532aaf92243bf2b4b","retrievedAt":"2026-11-11T23:31:19.741Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"10c2ed27-4bc2-400c-8723-55f03f379723"},"reviewer":{"model":"gpt-6-luna","runId":"a2d18108-5b8a-4981-80b8-92b6ff1fdac2"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-11T23:32:18.637Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Wu, Yue","url":null,"kind":"unknown"},{"name":"Zhou, Xueyu","url":null,"kind":"unknown"},{"name":"Hu, Yangrong","url":null,"kind":"unknown"},{"name":"Huang, Jian","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Wu, Yue","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wu, Yue","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Zhou, Xueyu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhou, Xueyu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Hu, Yangrong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Hu, Yangrong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Huang, Jian","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Huang, Jian","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}}],"participationCoverage":{"total":4,"returned":4,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.13680","snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c/source"}},"canonicalUrl":"https://escapemeridian.com/readings/bootstrapped-on-policy-distillation-067c23e6","formats":{"html":"/readings/bootstrapped-on-policy-distillation-067c23e6","markdown":"/readings/bootstrapped-on-policy-distillation-067c23e6.md","json":"/readings/bootstrapped-on-policy-distillation-067c23e6.json"},"attribution":"Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.","maintainedGuideLinks":[],"readerContext":{"version":"reader-context-v1","attribution":"Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.","terms":[],"entities":[{"id":"arxiv","name":"arXiv","kind":"organization","context":"arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed.","referenceUrls":["https://arxiv.org/","https://info.arxiv.org/help/api/index.html"],"path":"/entities/arxiv","json":"/entities/arxiv.json","markdown":"/entities/arxiv.md"}]},"researchQuestions":[],"withdrawn":false,"references":{"schemaVersion":1,"target":{"type":"reading","id":"067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c","revision":"1bcf077e-b2cc-451d-84b9-e6a09991ed73","revisionKind":"publication-id","title":"Bootstrapped On-Policy Distillation","url":"https://escapemeridian.com/readings/bootstrapped-on-policy-distillation-067c23e6"},"source":{"type":"source","id":"https://arxiv.org/abs/2611.13680","revision":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","revisionKind":"sha256-source-text","title":"Bootstrapped On-Policy Distillation","url":"https://arxiv.org/abs/2611.13680"},"status":"available","items":[],"next":null,"scope":"Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter."},"revisions":[{"id":"067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c","title":"Bootstrapped On-Policy Distillation","argument":"The authors of “Bootstrapped On-Policy Distillation” argue that on-policy self-distillation (OPSD) can become less effective as training continues: a frozen privileged teacher may drift away from the evolving student and give misleading supervision. They propose Bootstrapped On-Policy Distillation (BOPD), which first uses OPSD to transfer task-relevant information into an intermediate model, then distills from that model into a fresh copy of the initial model. The idea is to use reference-conditioned guidance early, then pass on what the adapted model learned without reference conditioning. The authors report that, across models, BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers. Their results suggest this two-stage transfer can make privileged reference supervision more useful for those tasks. The abstract does not provide the experimental details needed to assess the size or durability of the reported gains. Will BOPD’s reported gains persist under continued training?","topicIds":["intelligence-and-discovery"],"domain":"science","claimType":"attributed reading","limits":"The available evidence is the arXiv page abstract; it does not give experimental details such as the models, evaluation setup, or size of the reported gains. The abstract reports improvements across models but does not establish whether those gains persist under continued training.","speaker":{"name":"Wu, Yue, Zhou, Xueyu, Hu, Yangrong, Huang, Jian","roleAtSource":"Article author"},"source":{"title":"Bootstrapped On-Policy Distillation","publisher":"arXiv cs.AI","format":"Article","url":"https://arxiv.org/abs/2611.13680","publishedAt":"2026-11-10","dateBasis":"Date basis not provided."},"evidence":[{"text":"The authors say OPSD can lose effectiveness during continued training because the frozen privileged teacher becomes increasingly mismatched with the evolving student and may provide misleading supervision.","excerpt":"as training progresses, the growing distributional mismatch between the privileged teacher and reference-free student can cause the teacher to provide misleading supervision, ultimately degrading student performance.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"BOPD uses OPSD to transfer task-relevant information into an intermediate model, then distills that model into a fresh copy of the initial model.","excerpt":"first uses OPSD to internalize task-relevant information from reference-conditioned supervision into an intermediate model, and then distills this reference-free intermediate model into a fresh copy of the initial model.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"The authors report that BOPD improves mathematical reasoning and code generation over both initial models and constructed OPSD teachers across models.","excerpt":"Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"},{"text":"The authors’ proposed sequence is to use privileged reference supervision to acquire useful knowledge, then transfer it through a model without reference conditioning.","excerpt":"privileged reference supervision can be more effectively exploited by first transferring its task-relevant signal into a task-adapted model and then distilling from that model without reference conditioning.","url":"https://arxiv.org/abs/2611.13680","relationship":"Source-stated"}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13680","checkedAt":"2026-11-12T00:04:07.850Z","items":[],"note":"No suitable media was found by the supported page extraction."},"assumptions":[],"openQuestion":{"attribution":"Open question","text":"Will BOPD’s reported gains persist under continued training?"},"whatWouldChangeView":{"attribution":"Change condition not recorded","text":""},"editorialComparisons":[],"live":{"title":"Bootstrapped On-Policy Distillation","summary":"The authors of “Bootstrapped On-Policy Distillation” argue that on-policy self-distillation (OPSD) can become less effective as training continues: a frozen privileged teacher may drift away from the evolving student and give misleading supervision. They propose Bootstrapped On-Policy Distillation (BOPD), which first uses OPSD to transfer task-relevant information into an intermediate model, then distills from that model into a fresh copy of the initial model. The idea is to use reference-conditioned guidance early, then pass on what the adapted model learned without reference conditioning. The authors report that, across models, BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers. Their results suggest this two-stage transfer can make privileged reference supervision more useful for those tasks. The abstract does not provide the experimental details needed to assess the size or durability of the reported gains. Will BOPD’s reported gains persist under continued training?","question":"Will BOPD’s reported gains persist under continued training?","claims":[{"text":"The authors say OPSD can lose effectiveness during continued training because the frozen privileged teacher becomes increasingly mismatched with the evolving student and may provide misleading supervision.","excerpt":"as training progresses, the growing distributional mismatch between the privileged teacher and reference-free student can cause the teacher to provide misleading supervision, ultimately degrading student performance.","support":"source-stated"},{"text":"BOPD uses OPSD to transfer task-relevant information into an intermediate model, then distills that model into a fresh copy of the initial model.","excerpt":"first uses OPSD to internalize task-relevant information from reference-conditioned supervision into an intermediate model, and then distills this reference-free intermediate model into a fresh copy of the initial model.","support":"source-stated"},{"text":"The authors report that BOPD improves mathematical reasoning and code generation over both initial models and constructed OPSD teachers across models.","excerpt":"Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers.","support":"source-stated"},{"text":"The authors’ proposed sequence is to use privileged reference supervision to acquire useful knowledge, then transfer it through a model without reference conditioning.","excerpt":"privileged reference supervision can be more effectively exploited by first transferring its task-relevant signal into a task-adapted model and then distilling from that model without reference conditioning.","support":"source-stated"}],"limitations":["The available evidence is the arXiv page abstract; it does not give experimental details such as the models, evaluation setup, or size of the reported gains.","The abstract reports improvements across models but does not establish whether those gains persist under continued training."],"mediaId":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","correctionReason":null,"id":"067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c","publicationId":"1bcf077e-b2cc-451d-84b9-e6a09991ed73","version":1,"sourceUrl":"https://arxiv.org/abs/2611.13680","sourceTitle":"Bootstrapped On-Policy Distillation","publisher":"arXiv cs.AI","publishedAt":null,"originalPublished":{"value":"2026-11-10","precision":"day","basis":"publisher_citation_date"},"firstObservedAt":"2026-11-11T05:09:26.440Z","createdAt":"2026-11-11T23:32:18.637Z","historical":false,"entityIds":["arxiv"],"topicIds":["intelligence-and-discovery"],"media":[{"id":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","kind":"article","url":"https://arxiv.org/abs/2611.13680","sourceUrl":"https://arxiv.org/abs/2611.13680","label":"Read original","provenance":"source_link","verification":{"status":"verified","checkedAt":"2026-11-11T23:31:20.181Z","method":"acquired_source","contentType":"text/html"},"recommendationReason":"Recommended by Jev for understanding this source; checked by Luna."}],"sourceMedia":{"status":"not_found","sourceUrl":"https://arxiv.org/abs/2611.13680","checkedAt":"2026-11-12T00:04:07.850Z","items":[],"note":"No suitable media was found by the supported page extraction."},"editorialDecision":{"outcome":"accept","reasonCodes":["material_development"],"rubricVersion":"stream-admission-v5","mediaId":"746d0ceb7f16b1b3055c001f34c14dc76b786e65da3494dfff29221838f2a31f","confidence":0.8500000000000001,"provider":"typesafe","model":"jev-1.13.0","inputHash":"76f65fd2a97f33ffb8b3fff575349ce7cffd5ca53f1c8361da285cb3eecabe72","decidedAt":"2026-11-11T23:31:22.460Z"},"provenance":{"textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observationId":"ce90daaaa104e1bc0bb537476fc1204f1d5429f19f7efe8532aaf92243bf2b4b","retrievedAt":"2026-11-11T23:31:19.741Z","kind":"publisher-page","preparer":{"model":"gpt-6-luna","runId":"10c2ed27-4bc2-400c-8723-55f03f379723"},"reviewer":{"model":"gpt-6-luna","runId":"a2d18108-5b8a-4981-80b8-92b6ff1fdac2"}},"audit":{"status":"uncertain","reviewedAt":"2026-11-11T23:32:18.637Z","model":"anthropic/claude-opus-5.5"},"sourceAttribution":{"authors":[{"name":"Wu, Yue","url":null,"kind":"unknown"},{"name":"Zhou, Xueyu","url":null,"kind":"unknown"},{"name":"Hu, Yangrong","url":null,"kind":"unknown"},{"name":"Huang, Jian","url":null,"kind":"unknown"}],"authorRole":"article-author","hosts":[],"status":"stated","authorRoster":{"truncated":false,"omitted":0,"assertionGroupsOmitted":0},"participation":[{"actor":{"name":"Wu, Yue","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Wu, Yue","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:0","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Zhou, Xueyu","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Zhou, Xueyu","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:1","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Hu, Yangrong","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Hu, Yangrong","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:2","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}},{"actor":{"name":"Huang, Jian","url":null,"kind":"unknown","actorId":null},"role":"author","scope":{"kind":"work-credit","locator":null},"literalCredit":"Huang, Jian","status":"stated","provenance":{"basis":"head-meta","sourceUrl":"https://arxiv.org/abs/2611.13680","locator":"head/meta[name=author|citation_author]"},"assertionId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30:participation:3","identity":{"version":"entity-identity-v1","status":"unresolved","actorId":null,"canonicalActorId":null,"candidateActorIds":[],"assertions":[],"sourceObservedAt":"2026-11-12T00:04:07.850Z","knowledgeAsOf":null},"sourceVersion":{"snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","sourceUrl":"https://arxiv.org/abs/2611.13680","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z"}}],"participationCoverage":{"total":4,"returned":4,"complete":true,"fullAssertions":"source-metadata-endpoint"},"distribution":null,"publisher":null,"sourceUrl":"https://arxiv.org/abs/2611.13680","snapshotId":"7ed426e2f52d232d294f6ed253100e9ff37aa30cd489232693c242c865dddb30","contentSha256":"63f6f63d3e6f81d573ff10f40767b7c5b91656e49ef16bf4cc633d1aadc49598","textHash":"afabcd7490c9873ecd850430645105085d27697b059fee61fd3348712b694eb2","observedAt":"2026-11-12T00:04:07.850Z","basis":"publisher-page-metadata","sourceLinks":44,"linksTruncated":false,"metadataUrl":"/api/pipeline/stream/067c23e659ee9fcaa6dad37fc52853bbd37aaca9db6de06e2bb778cf34f7ff3c/source"}}}],"audits":[{"publicationId":"1bcf077e-b2cc-451d-84b9-e6a09991ed73","verdict":"uncertain","findings":[{"dimension":"attribution","status":"pass","explanation":"The reading consistently attributes the diagnosis, the BOPD proposal and the results to the paper's authors (Yue Wu, Xueyu Zhou, Yangrong Hu, Jian Huang). It frames the results as reported, not established.","excerpt":"Authors:Yue Wu, Xueyu Zhou, Yangrong Hu, Jian Huang"},{"dimension":"summary","status":"uncertain","explanation":"The content is faithful, but central jargon is never explained in plain language. 'Distillation', 'on-policy' (training on the model's own generated outputs), 'privileged teacher' and 'reference-conditioned' (a copy of the model that sees reference answers) are left unexplained, so a non-specialist cannot follow why the teacher drifts or why a fresh student helps. Smallest repair: add one plain-language clause per term. The summary also repeats the question at its end, though that is stylistic. It omits the source's note that OPSD showed 'great early improvements', which slightly flattens the source's argument.","excerpt":"using a reference-conditioned copy of the model as the teacher"},{"dimension":"date","status":"pass","explanation":"originalPublished 2026-11-10 matches the arXiv v1 submission date. The feed date of 2026-11-11 is correctly not used.","excerpt":"[Submitted on 10 Nov 2026]"},{"dimension":"support","status":"pass","explanation":"Each claim matches its verbatim abstract excerpt. The limitations accurately note that the abstract lacks experimental details and does not address whether BOPD's own gains persist. The question follows from the source's own concern about degradation under continued training, without implying an answer.","excerpt":"Experiments across models show that BOPD improves mathematical reasoning and code generation beyond both the initial models and the constructed OPSD teachers."},{"dimension":"media","status":"pass","explanation":"The selected media is the verified direct arXiv abstract page for this paper, which is the most direct access to the source.","excerpt":null}],"createdAt":"2026-11-11T23:32:18.637Z","model":"anthropic/claude-opus-5.5"}],"correctionUrl":"https://escapemeridian.com/corrections/"}