{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UMS4CLDAKUZKTBCNFOMIKUS6WG","short_pith_number":"pith:UMS4CLDA","schema_version":"1.0","canonical_sha256":"a325c12c605532a9844d2b9885525eb1ad6438faa67e30908a65c67eb138dfa0","source":{"kind":"arxiv","id":"2605.27088","version":1},"attestation_state":"computed","paper":{"title":"LLMs Are Already Good Tutors: Training-Free Prompt Optimization for Pedagogical Math Tutoring","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Eunjoo Lee, Hoilym Kwon, Jeongsu Moon, Kyungtae Joo, Minchul Shin, Sookbun Lee, Unggi Lee, Yeil Jeong","submitted_at":"2026-05-26T14:35:57Z","abstract_excerpt":"Aligning LLMs for math tutoring typically requires RL-based training with multi-GPU infrastructure. We investigate whether training-free prompt optimization-evolving only the system prompt via API calls-can serve as a practical alternative. We adapt 7 published methods and propose 5 education-specialized methods, evaluating these 12 methods under 5 conditions on 2 OOD benchmark suites. All 12 best-per-method configurations surpass the strongest RL-trained baseline (R_total = 0.633), and our ParetoGrad achieves the best Pareto balance across post-test solve rate, leak control, and helpfulness, "},"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":"2605.27088","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-26T14:35:57Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"cc93daaa75d2f8f6fa0944966f1401fe04e3b4d324e8fa4b3929cc0cd7b5c9d3","abstract_canon_sha256":"dcd7e12c7bff150cd2ed84f04176e99c12c23c53cd3d19e59615761fde0e3ee0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:06:27.725870Z","signature_b64":"xKjPXSZL2qESAgFdw+WflBlwfPRuMq0GDgB7tEbvLzKPn7W94zeAeljnA6U60Pv8hxZz3FtPL89E5GyLm7BVAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a325c12c605532a9844d2b9885525eb1ad6438faa67e30908a65c67eb138dfa0","last_reissued_at":"2026-05-27T01:06:27.725420Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:06:27.725420Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLMs Are Already Good Tutors: Training-Free Prompt Optimization for Pedagogical Math Tutoring","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Eunjoo Lee, Hoilym Kwon, Jeongsu Moon, Kyungtae Joo, Minchul Shin, Sookbun Lee, Unggi Lee, Yeil Jeong","submitted_at":"2026-05-26T14:35:57Z","abstract_excerpt":"Aligning LLMs for math tutoring typically requires RL-based training with multi-GPU infrastructure. We investigate whether training-free prompt optimization-evolving only the system prompt via API calls-can serve as a practical alternative. We adapt 7 published methods and propose 5 education-specialized methods, evaluating these 12 methods under 5 conditions on 2 OOD benchmark suites. All 12 best-per-method configurations surpass the strongest RL-trained baseline (R_total = 0.633), and our ParetoGrad achieves the best Pareto balance across post-test solve rate, leak control, and helpfulness, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.27088","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/2605.27088/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":"2605.27088","created_at":"2026-05-27T01:06:27.725488+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.27088v1","created_at":"2026-05-27T01:06:27.725488+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.27088","created_at":"2026-05-27T01:06:27.725488+00:00"},{"alias_kind":"pith_short_12","alias_value":"UMS4CLDAKUZK","created_at":"2026-05-27T01:06:27.725488+00:00"},{"alias_kind":"pith_short_16","alias_value":"UMS4CLDAKUZKTBCN","created_at":"2026-05-27T01:06:27.725488+00:00"},{"alias_kind":"pith_short_8","alias_value":"UMS4CLDA","created_at":"2026-05-27T01:06:27.725488+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/UMS4CLDAKUZKTBCNFOMIKUS6WG","json":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG.json","graph_json":"https://pith.science/api/pith-number/UMS4CLDAKUZKTBCNFOMIKUS6WG/graph.json","events_json":"https://pith.science/api/pith-number/UMS4CLDAKUZKTBCNFOMIKUS6WG/events.json","paper":"https://pith.science/paper/UMS4CLDA"},"agent_actions":{"view_html":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG","download_json":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG.json","view_paper":"https://pith.science/paper/UMS4CLDA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.27088&json=true","fetch_graph":"https://pith.science/api/pith-number/UMS4CLDAKUZKTBCNFOMIKUS6WG/graph.json","fetch_events":"https://pith.science/api/pith-number/UMS4CLDAKUZKTBCNFOMIKUS6WG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG/action/storage_attestation","attest_author":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG/action/author_attestation","sign_citation":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG/action/citation_signature","submit_replication":"https://pith.science/pith/UMS4CLDAKUZKTBCNFOMIKUS6WG/action/replication_record"}},"created_at":"2026-05-27T01:06:27.725488+00:00","updated_at":"2026-05-27T01:06:27.725488+00:00"}