{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4Z2HN5REPEI5GYDHTKPTOCAWCX","short_pith_number":"pith:4Z2HN5RE","schema_version":"1.0","canonical_sha256":"e67476f6247911d360679a9f37081615dcc6515c3ce98b53ccd5c812de0a0a57","source":{"kind":"arxiv","id":"2502.13441","version":1},"attestation_state":"computed","paper":{"title":"The Self-Improvement Paradox: Can Language Models Bootstrap Reasoning Capabilities without External Scaffolding?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jianwei Yin, Mingshuai Chen, Ruochen Xu, Tiancheng Zhao, Yutao Sun, Zilun Zhang","submitted_at":"2025-02-19T05:37:08Z","abstract_excerpt":"Self-improving large language models (LLMs) -- i.e., to improve the performance of an LLM by fine-tuning it with synthetic data generated by itself -- is a promising way to advance the capabilities of LLMs while avoiding extensive supervision. Existing approaches to self-improvement often rely on external supervision signals in the form of seed data and/or assistance from third-party models. This paper presents Crescent -- a simple yet effective framework for generating high-quality synthetic question-answer data in a fully autonomous manner. Crescent first elicits the LLM to generate raw ques"},"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":"2502.13441","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T05:37:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e9d5cde7302601764dfca31f7dca06efaf623a7d252020a625fe39418fc31024","abstract_canon_sha256":"50b9351ab7a32b80deb5a980967fcc43a3774fcdfa82598d859f2cae1dc85db1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:47.316925Z","signature_b64":"Web3E69AdlNldFRIgu87MhvqbbcpjPUcZRcBTRSUUch5dRaLNLyM/4vSx3+z3ZvsWXnhbhVquDpnOi8UTX0hAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e67476f6247911d360679a9f37081615dcc6515c3ce98b53ccd5c812de0a0a57","last_reissued_at":"2026-07-05T10:16:47.316270Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:47.316270Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Self-Improvement Paradox: Can Language Models Bootstrap Reasoning Capabilities without External Scaffolding?","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jianwei Yin, Mingshuai Chen, Ruochen Xu, Tiancheng Zhao, Yutao Sun, Zilun Zhang","submitted_at":"2025-02-19T05:37:08Z","abstract_excerpt":"Self-improving large language models (LLMs) -- i.e., to improve the performance of an LLM by fine-tuning it with synthetic data generated by itself -- is a promising way to advance the capabilities of LLMs while avoiding extensive supervision. Existing approaches to self-improvement often rely on external supervision signals in the form of seed data and/or assistance from third-party models. This paper presents Crescent -- a simple yet effective framework for generating high-quality synthetic question-answer data in a fully autonomous manner. Crescent first elicits the LLM to generate raw ques"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13441","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/2502.13441/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":"2502.13441","created_at":"2026-07-05T10:16:47.316358+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13441v1","created_at":"2026-07-05T10:16:47.316358+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13441","created_at":"2026-07-05T10:16:47.316358+00:00"},{"alias_kind":"pith_short_12","alias_value":"4Z2HN5REPEI5","created_at":"2026-07-05T10:16:47.316358+00:00"},{"alias_kind":"pith_short_16","alias_value":"4Z2HN5REPEI5GYDH","created_at":"2026-07-05T10:16:47.316358+00:00"},{"alias_kind":"pith_short_8","alias_value":"4Z2HN5RE","created_at":"2026-07-05T10:16:47.316358+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.00323","citing_title":"Online Self-Calibration Against Hallucination in Vision-Language Models","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX","json":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX.json","graph_json":"https://pith.science/api/pith-number/4Z2HN5REPEI5GYDHTKPTOCAWCX/graph.json","events_json":"https://pith.science/api/pith-number/4Z2HN5REPEI5GYDHTKPTOCAWCX/events.json","paper":"https://pith.science/paper/4Z2HN5RE"},"agent_actions":{"view_html":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX","download_json":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX.json","view_paper":"https://pith.science/paper/4Z2HN5RE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13441&json=true","fetch_graph":"https://pith.science/api/pith-number/4Z2HN5REPEI5GYDHTKPTOCAWCX/graph.json","fetch_events":"https://pith.science/api/pith-number/4Z2HN5REPEI5GYDHTKPTOCAWCX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX/action/storage_attestation","attest_author":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX/action/author_attestation","sign_citation":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX/action/citation_signature","submit_replication":"https://pith.science/pith/4Z2HN5REPEI5GYDHTKPTOCAWCX/action/replication_record"}},"created_at":"2026-07-05T10:16:47.316358+00:00","updated_at":"2026-07-05T10:16:47.316358+00:00"}