{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3PZX3FDIGWZUBEP764V7LAZOCI","short_pith_number":"pith:3PZX3FDI","schema_version":"1.0","canonical_sha256":"dbf37d946835b34091fff72bf5832e122b2c718e9d14161637aabb1982806372","source":{"kind":"arxiv","id":"2507.15844","version":3},"attestation_state":"computed","paper":{"title":"Hierarchical Budget Policy Optimization for Adaptive Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hao Li, Jun Xiao, Linjuan Wu, Peisheng Jiang, Shangke Lyu, Weiming Lu, Xingyu Wu, Yongliang Shen, Yuchen Yan, Yueting Zhuang","submitted_at":"2025-07-21T17:52:34Z","abstract_excerpt":"Large reasoning models achieve remarkable performance through extensive chain-of-thought generation, yet they suffer from a critical inefficiency: applying uniformly extensive reasoning regardless of problem complexity. We present Hierarchical Budget Policy Optimization (HBPO), a reinforcement learning framework that enables models to learn problem-specific reasoning depths without sacrificing capability. Unlike existing approaches that impose rigid constraints or rely on discrete mode selection, HBPO partitions the exploration space into budget-constrained hierarchies (512-2560 tokens), each "},"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":"2507.15844","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-07-21T17:52:34Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"b818768177e489d901938a2367c256d9d97aeab78a4720f917e8b0c7a9647f67","abstract_canon_sha256":"3ccd4d04f11785c23d19ce065c1ddbbcab3fa7e7575bd401e95df46f6143ecdf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:49:57.454079Z","signature_b64":"G7obdUpvYd3a7lknJuopJXOty7ndbQ2y2jPzMwB0LcFx1wKh/qGtbntKICD5mAD8On6eDVkKPOvvnv/qr6++Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbf37d946835b34091fff72bf5832e122b2c718e9d14161637aabb1982806372","last_reissued_at":"2026-07-05T11:49:57.453621Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:49:57.453621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hierarchical Budget Policy Optimization for Adaptive Reasoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Hao Li, Jun Xiao, Linjuan Wu, Peisheng Jiang, Shangke Lyu, Weiming Lu, Xingyu Wu, Yongliang Shen, Yuchen Yan, Yueting Zhuang","submitted_at":"2025-07-21T17:52:34Z","abstract_excerpt":"Large reasoning models achieve remarkable performance through extensive chain-of-thought generation, yet they suffer from a critical inefficiency: applying uniformly extensive reasoning regardless of problem complexity. We present Hierarchical Budget Policy Optimization (HBPO), a reinforcement learning framework that enables models to learn problem-specific reasoning depths without sacrificing capability. Unlike existing approaches that impose rigid constraints or rely on discrete mode selection, HBPO partitions the exploration space into budget-constrained hierarchies (512-2560 tokens), each "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.15844","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/2507.15844/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":"2507.15844","created_at":"2026-07-05T11:49:57.453677+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.15844v3","created_at":"2026-07-05T11:49:57.453677+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.15844","created_at":"2026-07-05T11:49:57.453677+00:00"},{"alias_kind":"pith_short_12","alias_value":"3PZX3FDIGWZU","created_at":"2026-07-05T11:49:57.453677+00:00"},{"alias_kind":"pith_short_16","alias_value":"3PZX3FDIGWZUBEP7","created_at":"2026-07-05T11:49:57.453677+00:00"},{"alias_kind":"pith_short_8","alias_value":"3PZX3FDI","created_at":"2026-07-05T11:49:57.453677+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/3PZX3FDIGWZUBEP764V7LAZOCI","json":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI.json","graph_json":"https://pith.science/api/pith-number/3PZX3FDIGWZUBEP764V7LAZOCI/graph.json","events_json":"https://pith.science/api/pith-number/3PZX3FDIGWZUBEP764V7LAZOCI/events.json","paper":"https://pith.science/paper/3PZX3FDI"},"agent_actions":{"view_html":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI","download_json":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI.json","view_paper":"https://pith.science/paper/3PZX3FDI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.15844&json=true","fetch_graph":"https://pith.science/api/pith-number/3PZX3FDIGWZUBEP764V7LAZOCI/graph.json","fetch_events":"https://pith.science/api/pith-number/3PZX3FDIGWZUBEP764V7LAZOCI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI/action/storage_attestation","attest_author":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI/action/author_attestation","sign_citation":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI/action/citation_signature","submit_replication":"https://pith.science/pith/3PZX3FDIGWZUBEP764V7LAZOCI/action/replication_record"}},"created_at":"2026-07-05T11:49:57.453677+00:00","updated_at":"2026-07-05T11:49:57.453677+00:00"}