{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ACDEOYH53JMNDGA4OWHPLAYEP3","short_pith_number":"pith:ACDEOYH5","schema_version":"1.0","canonical_sha256":"00864760fdda58d1981c758ef583047ed1e4ad5707c976cae6a31ec4d9b4ccbd","source":{"kind":"arxiv","id":"2606.29639","version":1},"attestation_state":"computed","paper":{"title":"Two-Stage Prompt Optimization for Few-Shot Relation Extraction: From Reasoning-Guided Search to Gradient-Guided Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aunabil Chakma, Eduardo Blanco, Mihai Surdeanu","submitted_at":"2026-06-28T23:12:28Z","abstract_excerpt":"Automatic prompt optimization is still underexplored for episodic few-shot relation extraction with smaller language models. We propose a two-stage framework that combines reasoning-based prompt optimization with gradient-based prompt optimization. The first stage can use any reasoning-based optimizer to make broadprompt improvements in natural language. The second stage applies our GradPO, which uses loss and gradient signals to identify high-impact prompt spans and refine them with local edits. Experiments on FS-TACRED and FS-FewRel show that local refinement usually improves prompts found b"},"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":"2606.29639","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-28T23:12:28Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"68325b7b9b2a2656606ca3bf849d9534f6ff3cc02aa184b15670f9ed5d9a360a","abstract_canon_sha256":"cad9e2d61c7f86591852c81042f022decfb999423fd27477ec97521299f15ad5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:18:20.815281Z","signature_b64":"RxaC47MxPAeeVv0V+qqHfrJT4qOLMl9TY+O5Hj21L//z9NHYC944S0Yp5EOOHxJboh6NL32p3Az0bp66PwvbAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00864760fdda58d1981c758ef583047ed1e4ad5707c976cae6a31ec4d9b4ccbd","last_reissued_at":"2026-06-30T01:18:20.814670Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:18:20.814670Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Two-Stage Prompt Optimization for Few-Shot Relation Extraction: From Reasoning-Guided Search to Gradient-Guided Refinement","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aunabil Chakma, Eduardo Blanco, Mihai Surdeanu","submitted_at":"2026-06-28T23:12:28Z","abstract_excerpt":"Automatic prompt optimization is still underexplored for episodic few-shot relation extraction with smaller language models. We propose a two-stage framework that combines reasoning-based prompt optimization with gradient-based prompt optimization. The first stage can use any reasoning-based optimizer to make broadprompt improvements in natural language. The second stage applies our GradPO, which uses loss and gradient signals to identify high-impact prompt spans and refine them with local edits. Experiments on FS-TACRED and FS-FewRel show that local refinement usually improves prompts found b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.29639","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/2606.29639/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":"2606.29639","created_at":"2026-06-30T01:18:20.814766+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.29639v1","created_at":"2026-06-30T01:18:20.814766+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.29639","created_at":"2026-06-30T01:18:20.814766+00:00"},{"alias_kind":"pith_short_12","alias_value":"ACDEOYH53JMN","created_at":"2026-06-30T01:18:20.814766+00:00"},{"alias_kind":"pith_short_16","alias_value":"ACDEOYH53JMNDGA4","created_at":"2026-06-30T01:18:20.814766+00:00"},{"alias_kind":"pith_short_8","alias_value":"ACDEOYH5","created_at":"2026-06-30T01:18:20.814766+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/ACDEOYH53JMNDGA4OWHPLAYEP3","json":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3.json","graph_json":"https://pith.science/api/pith-number/ACDEOYH53JMNDGA4OWHPLAYEP3/graph.json","events_json":"https://pith.science/api/pith-number/ACDEOYH53JMNDGA4OWHPLAYEP3/events.json","paper":"https://pith.science/paper/ACDEOYH5"},"agent_actions":{"view_html":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3","download_json":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3.json","view_paper":"https://pith.science/paper/ACDEOYH5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.29639&json=true","fetch_graph":"https://pith.science/api/pith-number/ACDEOYH53JMNDGA4OWHPLAYEP3/graph.json","fetch_events":"https://pith.science/api/pith-number/ACDEOYH53JMNDGA4OWHPLAYEP3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3/action/storage_attestation","attest_author":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3/action/author_attestation","sign_citation":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3/action/citation_signature","submit_replication":"https://pith.science/pith/ACDEOYH53JMNDGA4OWHPLAYEP3/action/replication_record"}},"created_at":"2026-06-30T01:18:20.814766+00:00","updated_at":"2026-06-30T01:18:20.814766+00:00"}