{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ASBVJQKM757SZR545FFU7EGKBM","short_pith_number":"pith:ASBVJQKM","schema_version":"1.0","canonical_sha256":"048354c14cff7f2cc7bce94b4f90ca0b118c26e8c4027733a584b6c4ecc2da03","source":{"kind":"arxiv","id":"2409.03733","version":2},"attestation_state":"computed","paper":{"title":"Planning In Natural Language Improves LLM Search For Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Catherine Wu, Evan Wang, Federico Cassano, Hugh Zhang, Sean Hendryx, Summer Yue, Vaskar Nath, Will Song, Yunfeng Bai, Ziwen Han","submitted_at":"2024-09-05T17:44:49Z","abstract_excerpt":"While scaling training compute has led to remarkable improvements in large language models (LLMs), scaling inference compute has not yet yielded analogous gains. We hypothesize that a core missing component is a lack of diverse LLM outputs, leading to inefficient search due to models repeatedly sampling highly similar, yet incorrect generations. We empirically demonstrate that this lack of diversity can be mitigated by searching over candidate plans for solving a problem in natural language. Based on this insight, we propose PlanSearch, a novel search algorithm which shows strong results acros"},"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":"2409.03733","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-09-05T17:44:49Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"91e7c178051a591eb8b9d13f069e5d41f4ff22e7965c9683912c5372d20fa858","abstract_canon_sha256":"0baacaa6aa7c09e41d5ab2c11eba98fa4e32787bdd5956c790bad27e4bd68bc2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:22:42.239287Z","signature_b64":"9vB57Bpcj01D8/0tpjj8vlWnueEYroZ6ILZcPh660QpOAYnNqBS6Uq+B1EvV1sXREzWMBFYagjbJAqxCFfnVDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"048354c14cff7f2cc7bce94b4f90ca0b118c26e8c4027733a584b6c4ecc2da03","last_reissued_at":"2026-07-05T09:22:42.238824Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:22:42.238824Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Planning In Natural Language Improves LLM Search For Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Catherine Wu, Evan Wang, Federico Cassano, Hugh Zhang, Sean Hendryx, Summer Yue, Vaskar Nath, Will Song, Yunfeng Bai, Ziwen Han","submitted_at":"2024-09-05T17:44:49Z","abstract_excerpt":"While scaling training compute has led to remarkable improvements in large language models (LLMs), scaling inference compute has not yet yielded analogous gains. We hypothesize that a core missing component is a lack of diverse LLM outputs, leading to inefficient search due to models repeatedly sampling highly similar, yet incorrect generations. We empirically demonstrate that this lack of diversity can be mitigated by searching over candidate plans for solving a problem in natural language. Based on this insight, we propose PlanSearch, a novel search algorithm which shows strong results acros"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.03733","kind":"arxiv","version":2},"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/2409.03733/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":"2409.03733","created_at":"2026-07-05T09:22:42.238881+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.03733v2","created_at":"2026-07-05T09:22:42.238881+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.03733","created_at":"2026-07-05T09:22:42.238881+00:00"},{"alias_kind":"pith_short_12","alias_value":"ASBVJQKM757S","created_at":"2026-07-05T09:22:42.238881+00:00"},{"alias_kind":"pith_short_16","alias_value":"ASBVJQKM757SZR54","created_at":"2026-07-05T09:22:42.238881+00:00"},{"alias_kind":"pith_short_8","alias_value":"ASBVJQKM","created_at":"2026-07-05T09:22:42.238881+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17628","citing_title":"OPD-Evolver: Cultivating Holistic Agent Evolver via On-Policy Distillation","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31511","citing_title":"Falsification, Not Exposure: An Internally Preregistered Placebo-Controlled Decomposition of Self-Repair Feedback in Frozen Small Code Models","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12429","citing_title":"Muse Spark Safety & Preparedness Report","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15607","citing_title":"Syntax Without Semantics: Teaching Large Language Models to Code in an Unseen Language","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18747","citing_title":"Code as Agent Harness","ref_index":159,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19341","citing_title":"Evaluation-driven Scaling for Scientific Discovery","ref_index":149,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10449","citing_title":"AdverMCTS: Combating Pseudo-Correctness in Code Generation via Adversarial Monte Carlo Tree Search","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM","json":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM.json","graph_json":"https://pith.science/api/pith-number/ASBVJQKM757SZR545FFU7EGKBM/graph.json","events_json":"https://pith.science/api/pith-number/ASBVJQKM757SZR545FFU7EGKBM/events.json","paper":"https://pith.science/paper/ASBVJQKM"},"agent_actions":{"view_html":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM","download_json":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM.json","view_paper":"https://pith.science/paper/ASBVJQKM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.03733&json=true","fetch_graph":"https://pith.science/api/pith-number/ASBVJQKM757SZR545FFU7EGKBM/graph.json","fetch_events":"https://pith.science/api/pith-number/ASBVJQKM757SZR545FFU7EGKBM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM/action/storage_attestation","attest_author":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM/action/author_attestation","sign_citation":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM/action/citation_signature","submit_replication":"https://pith.science/pith/ASBVJQKM757SZR545FFU7EGKBM/action/replication_record"}},"created_at":"2026-07-05T09:22:42.238881+00:00","updated_at":"2026-07-05T09:22:42.238881+00:00"}