{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4HLLBK5Z4FRDX4HK6KU6TXWXTO","short_pith_number":"pith:4HLLBK5Z","schema_version":"1.0","canonical_sha256":"e1d6b0abb9e1623bf0eaf2a9e9ded79bb6fc6bcaf694d88e7742a5044209f9eb","source":{"kind":"arxiv","id":"2605.18299","version":1},"attestation_state":"computed","paper":{"title":"SD-Search: On-Policy Hindsight Self-Distillation for Search-Augmented Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.IR"],"primary_cat":"cs.AI","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Lingtao Mao, Wenwu Ou, Xuxin Zhang, Yufei Ma, Zhipeng Qian, Zihan Liang","submitted_at":"2026-05-18T12:18:47Z","abstract_excerpt":"Search-augmented reasoning agents interleave internal reasoning with calls to an external retriever, and their performance relies on the quality of each issued query. However, under outcome-reward reinforcement learning, every search decision in a rollout shares the same trajectory-level reward, leaving individual queries without step-specific credit. Recent process-supervision approaches address this gap by drawing step-level signals from outside the policy, relying either on a much larger teacher model, or on sub-question annotations produced by a stronger external system. In contrast, we pr"},"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":"2605.18299","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-18T12:18:47Z","cross_cats_sorted":["cs.CL","cs.IR"],"title_canon_sha256":"48e1eef21f2ac5a6b7c3b97ef423db18906b9a91218bfd5e5a9fbdab9d529357","abstract_canon_sha256":"cba01f48da7c5546f4b6d1a76845004433400995c028357d97368dbcd90e1249"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:05:54.049422Z","signature_b64":"JI5sqLB3n3oSq6agVwH3Jms5e7EWcosV6gE8xtewseWjWyas4UPnDinPXWeXVPp+t4H7STc5thP8rMXg7sA4Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e1d6b0abb9e1623bf0eaf2a9e9ded79bb6fc6bcaf694d88e7742a5044209f9eb","last_reissued_at":"2026-05-20T00:05:54.048744Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:05:54.048744Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SD-Search: On-Policy Hindsight Self-Distillation for Search-Augmented Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.IR"],"primary_cat":"cs.AI","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Lingtao Mao, Wenwu Ou, Xuxin Zhang, Yufei Ma, Zhipeng Qian, Zihan Liang","submitted_at":"2026-05-18T12:18:47Z","abstract_excerpt":"Search-augmented reasoning agents interleave internal reasoning with calls to an external retriever, and their performance relies on the quality of each issued query. However, under outcome-reward reinforcement learning, every search decision in a rollout shares the same trajectory-level reward, leaving individual queries without step-specific credit. Recent process-supervision approaches address this gap by drawing step-level signals from outside the policy, relying either on a much larger teacher model, or on sub-question annotations produced by a stronger external system. In contrast, we pr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18299","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/2605.18299/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T23:33:35.213326Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T23:21:58.903210Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"3473efb3b6cedf2efe9069f140bcc8e725265afed62574014594851fe719a24f"},"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":"2605.18299","created_at":"2026-05-20T00:05:54.048872+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18299v1","created_at":"2026-05-20T00:05:54.048872+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18299","created_at":"2026-05-20T00:05:54.048872+00:00"},{"alias_kind":"pith_short_12","alias_value":"4HLLBK5Z4FRD","created_at":"2026-05-20T00:05:54.048872+00:00"},{"alias_kind":"pith_short_16","alias_value":"4HLLBK5Z4FRDX4HK","created_at":"2026-05-20T00:05:54.048872+00:00"},{"alias_kind":"pith_short_8","alias_value":"4HLLBK5Z","created_at":"2026-05-20T00:05:54.048872+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/4HLLBK5Z4FRDX4HK6KU6TXWXTO","json":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO.json","graph_json":"https://pith.science/api/pith-number/4HLLBK5Z4FRDX4HK6KU6TXWXTO/graph.json","events_json":"https://pith.science/api/pith-number/4HLLBK5Z4FRDX4HK6KU6TXWXTO/events.json","paper":"https://pith.science/paper/4HLLBK5Z"},"agent_actions":{"view_html":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO","download_json":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO.json","view_paper":"https://pith.science/paper/4HLLBK5Z","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18299&json=true","fetch_graph":"https://pith.science/api/pith-number/4HLLBK5Z4FRDX4HK6KU6TXWXTO/graph.json","fetch_events":"https://pith.science/api/pith-number/4HLLBK5Z4FRDX4HK6KU6TXWXTO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO/action/storage_attestation","attest_author":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO/action/author_attestation","sign_citation":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO/action/citation_signature","submit_replication":"https://pith.science/pith/4HLLBK5Z4FRDX4HK6KU6TXWXTO/action/replication_record"}},"created_at":"2026-05-20T00:05:54.048872+00:00","updated_at":"2026-05-20T00:05:54.048872+00:00"}