{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CFWSFMSEZMYZ7BL2XRIKEQKERI","short_pith_number":"pith:CFWSFMSE","schema_version":"1.0","canonical_sha256":"116d22b244cb319f857abc50a241448a225f67865cb09c48790b1d06a0a5502b","source":{"kind":"arxiv","id":"2608.07213","version":1},"attestation_state":"computed","paper":{"title":"From Test-Time Scaling to Reusable Memory: Measuring Crystallization in Text-to-SQL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"(2) Shanghai Jiao Tong University), Jiaqian Wang (1), Muning Wen (2) ((1) Xidian University, Wenjin Hou (1), Yuanxi Che (1), Yutao Qi (1)","submitted_at":"2026-08-07T13:28:27Z","abstract_excerpt":"Test-time scaling can correct difficult text-to-SQL queries, but the extra computation is normally discarded after each answer. Systems increasingly retain verified repair episodes, yet evaluations still report one end-to-end score. It cannot distinguish replay on recurring questions from help on unseen questions, or identify the responsible memory choice. We call measuring this future value the crystallization problem. Our controlled evaluation holds the single-shot solver fixed and varies one memory choice at a time. We separately measure replay, cross-question retention, and held-out same-d"},"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":"2608.07213","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-08-07T13:28:27Z","cross_cats_sorted":[],"title_canon_sha256":"7a849ae7f6cacdfe2b70207329d8874d9bcfc719a7e0bf40069b0bb51005edc2","abstract_canon_sha256":"eb705afb5bb28584dc420e3b59be76d40ab38a5a4daabcb5855ac8e6dc9312b5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-10T01:14:39.411150Z","signature_b64":"EUcKw4kS1JtdZsE91EuqLm80t+HkEQWls2WDflE1guP7piwtHjb0UQj1CDShehNuBxcHU6j2Qjf882hhlWeIDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"116d22b244cb319f857abc50a241448a225f67865cb09c48790b1d06a0a5502b","last_reissued_at":"2026-08-10T01:14:39.408008Z","signature_status":"signed_v1","first_computed_at":"2026-08-10T01:14:39.408008Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Test-Time Scaling to Reusable Memory: Measuring Crystallization in Text-to-SQL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"(2) Shanghai Jiao Tong University), Jiaqian Wang (1), Muning Wen (2) ((1) Xidian University, Wenjin Hou (1), Yuanxi Che (1), Yutao Qi (1)","submitted_at":"2026-08-07T13:28:27Z","abstract_excerpt":"Test-time scaling can correct difficult text-to-SQL queries, but the extra computation is normally discarded after each answer. Systems increasingly retain verified repair episodes, yet evaluations still report one end-to-end score. It cannot distinguish replay on recurring questions from help on unseen questions, or identify the responsible memory choice. We call measuring this future value the crystallization problem. Our controlled evaluation holds the single-shot solver fixed and varies one memory choice at a time. We separately measure replay, cross-question retention, and held-out same-d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.07213","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/2608.07213/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":"2608.07213","created_at":"2026-08-10T01:14:39.409449+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.07213v1","created_at":"2026-08-10T01:14:39.409449+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.07213","created_at":"2026-08-10T01:14:39.409449+00:00"},{"alias_kind":"pith_short_12","alias_value":"CFWSFMSEZMYZ","created_at":"2026-08-10T01:14:39.409449+00:00"},{"alias_kind":"pith_short_16","alias_value":"CFWSFMSEZMYZ7BL2","created_at":"2026-08-10T01:14:39.409449+00:00"},{"alias_kind":"pith_short_8","alias_value":"CFWSFMSE","created_at":"2026-08-10T01:14:39.409449+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/CFWSFMSEZMYZ7BL2XRIKEQKERI","json":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI.json","graph_json":"https://pith.science/api/pith-number/CFWSFMSEZMYZ7BL2XRIKEQKERI/graph.json","events_json":"https://pith.science/api/pith-number/CFWSFMSEZMYZ7BL2XRIKEQKERI/events.json","paper":"https://pith.science/paper/CFWSFMSE"},"agent_actions":{"view_html":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI","download_json":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI.json","view_paper":"https://pith.science/paper/CFWSFMSE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.07213&json=true","fetch_graph":"https://pith.science/api/pith-number/CFWSFMSEZMYZ7BL2XRIKEQKERI/graph.json","fetch_events":"https://pith.science/api/pith-number/CFWSFMSEZMYZ7BL2XRIKEQKERI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI/action/storage_attestation","attest_author":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI/action/author_attestation","sign_citation":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI/action/citation_signature","submit_replication":"https://pith.science/pith/CFWSFMSEZMYZ7BL2XRIKEQKERI/action/replication_record"}},"created_at":"2026-08-10T01:14:39.409449+00:00","updated_at":"2026-08-10T01:14:39.409449+00:00"}