{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:Q7VDP2MPJDMCQG73GWNCHFKQ3O","short_pith_number":"pith:Q7VDP2MP","schema_version":"1.0","canonical_sha256":"87ea37e98f48d8281bfb359a239550db89dd3bb26c83cb00915e705b225859f8","source":{"kind":"arxiv","id":"2608.10694","version":1},"attestation_state":"computed","paper":{"title":"Optimize Cheap, Deploy Strong: Cost-Aware Cross-Tier Transfer for Evolutionary Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.NE"],"primary_cat":"cs.LG","authors_text":"Oshri Naparstek, Roi pony, Tal Oved, Udi Barzelay","submitted_at":"2026-08-11T09:17:22Z","abstract_excerpt":"Evolutionary optimization of LLM prompts and agentic programs (e.g., GEPA) is dominated by fitness evaluation: scoring each candidate runs an answering LLM over a validation set, so the evaluator's price tier dictates total search cost. We restructure that search by decoupling the three roles an LLM plays, running the high-volume answering role on the cheapest tier, reserving a strong model for the rare reflection/variation operator, then exploiting upward cross-tier transfer to deploy the cheaply evolved prompt on a stronger target. We contribute a cost-controlled characterization of when che"},"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":"2608.10694","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-08-11T09:17:22Z","cross_cats_sorted":["cs.AI","cs.CL","cs.NE"],"title_canon_sha256":"41d3786a26258a746e17c82fbf7a222a180132ef1ae4fd8c13490314a8c0d89e","abstract_canon_sha256":"b651a8f8418bb99d48c13e7c7f2e9c1d0ef15642d435ef3c2cbe27248cf690f9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-12T01:23:04.531541Z","signature_b64":"Ky8x3mBb4DnfirNU0zFNHvEY7JXQP49G11H6h6MI9SU4DT1oRfGmGmFBszEizX73J2Krqw7Iu6cHnI4kf9D/Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"87ea37e98f48d8281bfb359a239550db89dd3bb26c83cb00915e705b225859f8","last_reissued_at":"2026-08-12T01:23:04.529743Z","signature_status":"signed_v1","first_computed_at":"2026-08-12T01:23:04.529743Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimize Cheap, Deploy Strong: Cost-Aware Cross-Tier Transfer for Evolutionary Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.NE"],"primary_cat":"cs.LG","authors_text":"Oshri Naparstek, Roi pony, Tal Oved, Udi Barzelay","submitted_at":"2026-08-11T09:17:22Z","abstract_excerpt":"Evolutionary optimization of LLM prompts and agentic programs (e.g., GEPA) is dominated by fitness evaluation: scoring each candidate runs an answering LLM over a validation set, so the evaluator's price tier dictates total search cost. We restructure that search by decoupling the three roles an LLM plays, running the high-volume answering role on the cheapest tier, reserving a strong model for the rare reflection/variation operator, then exploiting upward cross-tier transfer to deploy the cheaply evolved prompt on a stronger target. We contribute a cost-controlled characterization of when che"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10694","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/2608.10694/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":"2608.10694","created_at":"2026-08-12T01:23:04.533613+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.10694v1","created_at":"2026-08-12T01:23:04.533613+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10694","created_at":"2026-08-12T01:23:04.533613+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q7VDP2MPJDMC","created_at":"2026-08-12T01:23:04.533613+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q7VDP2MPJDMCQG73","created_at":"2026-08-12T01:23:04.533613+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q7VDP2MP","created_at":"2026-08-12T01:23:04.533613+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/Q7VDP2MPJDMCQG73GWNCHFKQ3O","json":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O.json","graph_json":"https://pith.science/api/pith-number/Q7VDP2MPJDMCQG73GWNCHFKQ3O/graph.json","events_json":"https://pith.science/api/pith-number/Q7VDP2MPJDMCQG73GWNCHFKQ3O/events.json","paper":"https://pith.science/paper/Q7VDP2MP"},"agent_actions":{"view_html":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O","download_json":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O.json","view_paper":"https://pith.science/paper/Q7VDP2MP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.10694&json=true","fetch_graph":"https://pith.science/api/pith-number/Q7VDP2MPJDMCQG73GWNCHFKQ3O/graph.json","fetch_events":"https://pith.science/api/pith-number/Q7VDP2MPJDMCQG73GWNCHFKQ3O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O/action/storage_attestation","attest_author":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O/action/author_attestation","sign_citation":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O/action/citation_signature","submit_replication":"https://pith.science/pith/Q7VDP2MPJDMCQG73GWNCHFKQ3O/action/replication_record"}},"created_at":"2026-08-12T01:23:04.533613+00:00","updated_at":"2026-08-12T01:23:04.533613+00:00"}