{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2MKXQZQGCFQ3BJXB3LFS6EKOEO","short_pith_number":"pith:2MKXQZQG","schema_version":"1.0","canonical_sha256":"d3157866061161b0a6e1dacb2f114e238eed6e4ee6d42ba7ed9d9b8b506d83aa","source":{"kind":"arxiv","id":"2605.28354","version":1},"attestation_state":"computed","paper":{"title":"Plan Before Search: Search Agents Need Plan","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Jiayi Ji, Qibin Hou, Wenwu Ou, Xiaoshuai Sun, Yufei Ma, Zhipeng Qian, Zihan Liang","submitted_at":"2026-05-27T11:54:15Z","abstract_excerpt":"Training large language models as retrieval-augmented reasoning agents typically combines reinforcement learning with an SFT cold start distilled from a stronger model. However, this paradigm overlooks two fundamental factors: the dependency structure among sub-skills, and the possibility that distillation is not the only route to capability acquisition. We study this through Plan, a structured agentic behavior for multi-hop retrieval that decomposes a question into ordered sub-questions before any retrieval is performed, so that each search step can be anchored to a pre-designed sub-question "},"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":"2605.28354","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-05-27T11:54:15Z","cross_cats_sorted":[],"title_canon_sha256":"0f87876a3e3eac9864313a2daad8703dccd5117139bbef045f650707a7ff7ff4","abstract_canon_sha256":"f214ca5d92163bc6a560da4d688718143034b068ef76801f671bba03a720e2cb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:15.582765Z","signature_b64":"l3enjcCoZoTEM7ktRpZc63EzoaDvPiSL4hL60lexfCaUCvt8BgmY9WPUIQizaYP9XpJzKg80HiaLSv45KY3ECA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d3157866061161b0a6e1dacb2f114e238eed6e4ee6d42ba7ed9d9b8b506d83aa","last_reissued_at":"2026-05-28T01:05:15.582069Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:15.582069Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Plan Before Search: Search Agents Need Plan","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ben Chen, Chenyi Lei, Huangyu Dai, Jiayi Ji, Qibin Hou, Wenwu Ou, Xiaoshuai Sun, Yufei Ma, Zhipeng Qian, Zihan Liang","submitted_at":"2026-05-27T11:54:15Z","abstract_excerpt":"Training large language models as retrieval-augmented reasoning agents typically combines reinforcement learning with an SFT cold start distilled from a stronger model. However, this paradigm overlooks two fundamental factors: the dependency structure among sub-skills, and the possibility that distillation is not the only route to capability acquisition. We study this through Plan, a structured agentic behavior for multi-hop retrieval that decomposes a question into ordered sub-questions before any retrieval is performed, so that each search step can be anchored to a pre-designed sub-question "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28354","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/2605.28354/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":"2605.28354","created_at":"2026-05-28T01:05:15.582368+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28354v1","created_at":"2026-05-28T01:05:15.582368+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28354","created_at":"2026-05-28T01:05:15.582368+00:00"},{"alias_kind":"pith_short_12","alias_value":"2MKXQZQGCFQ3","created_at":"2026-05-28T01:05:15.582368+00:00"},{"alias_kind":"pith_short_16","alias_value":"2MKXQZQGCFQ3BJXB","created_at":"2026-05-28T01:05:15.582368+00:00"},{"alias_kind":"pith_short_8","alias_value":"2MKXQZQG","created_at":"2026-05-28T01:05:15.582368+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/2MKXQZQGCFQ3BJXB3LFS6EKOEO","json":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO.json","graph_json":"https://pith.science/api/pith-number/2MKXQZQGCFQ3BJXB3LFS6EKOEO/graph.json","events_json":"https://pith.science/api/pith-number/2MKXQZQGCFQ3BJXB3LFS6EKOEO/events.json","paper":"https://pith.science/paper/2MKXQZQG"},"agent_actions":{"view_html":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO","download_json":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO.json","view_paper":"https://pith.science/paper/2MKXQZQG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28354&json=true","fetch_graph":"https://pith.science/api/pith-number/2MKXQZQGCFQ3BJXB3LFS6EKOEO/graph.json","fetch_events":"https://pith.science/api/pith-number/2MKXQZQGCFQ3BJXB3LFS6EKOEO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO/action/storage_attestation","attest_author":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO/action/author_attestation","sign_citation":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO/action/citation_signature","submit_replication":"https://pith.science/pith/2MKXQZQGCFQ3BJXB3LFS6EKOEO/action/replication_record"}},"created_at":"2026-05-28T01:05:15.582368+00:00","updated_at":"2026-05-28T01:05:15.582368+00:00"}