{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:WX6EYHNBZC2TJLXAPEL6NWHODP","short_pith_number":"pith:WX6EYHNB","schema_version":"1.0","canonical_sha256":"b5fc4c1da1c8b534aee07917e6d8ee1bd4b14924a1f26b12bb9932f769db800b","source":{"kind":"arxiv","id":"2206.06888","version":1},"attestation_state":"computed","paper":{"title":"CERT: Continual Pre-Training on Sketches for Library-Oriented Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.PL"],"primary_cat":"cs.SE","authors_text":"Bei Chen, Bei Guan, Daoguang Zan, Dejian Yang, Jian-Guang Lou, Minsu Kim, Weizhu Chen, Yongji Wang, Zeqi Lin","submitted_at":"2022-06-14T14:44:34Z","abstract_excerpt":"Code generation is a longstanding challenge, aiming to generate a code snippet based on a natural language description. Usually, expensive text-code paired data is essential for training a code generation model. Recently, thanks to the success of pre-training techniques, large language models are trained on large-scale unlabelled code corpora and perform well in code generation. In this paper, we investigate how to leverage an unlabelled code corpus to train a model for library-oriented code generation. Since it is a common practice for programmers to reuse third-party libraries, in which case"},"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":"2206.06888","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2022-06-14T14:44:34Z","cross_cats_sorted":["cs.CL","cs.PL"],"title_canon_sha256":"82ac8980fbeaaa13c495db9075c179fd8ada61d679b956375b54c167c319feb2","abstract_canon_sha256":"2b49c2a29dc5b24f4a77f80ad9f4076accbdafac291b0f128590a39e7dd34ae0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:31:46.910739Z","signature_b64":"bFYxynVv0EJ9BoRW8zFCuN/w1IQJ6jsl153GtK2J+zyNJT2CwvVdqAStaw3FQKMwcCLweC1JoPfJvqAP46gxBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5fc4c1da1c8b534aee07917e6d8ee1bd4b14924a1f26b12bb9932f769db800b","last_reissued_at":"2026-07-05T04:31:46.910292Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:31:46.910292Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CERT: Continual Pre-Training on Sketches for Library-Oriented Code Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.PL"],"primary_cat":"cs.SE","authors_text":"Bei Chen, Bei Guan, Daoguang Zan, Dejian Yang, Jian-Guang Lou, Minsu Kim, Weizhu Chen, Yongji Wang, Zeqi Lin","submitted_at":"2022-06-14T14:44:34Z","abstract_excerpt":"Code generation is a longstanding challenge, aiming to generate a code snippet based on a natural language description. Usually, expensive text-code paired data is essential for training a code generation model. Recently, thanks to the success of pre-training techniques, large language models are trained on large-scale unlabelled code corpora and perform well in code generation. In this paper, we investigate how to leverage an unlabelled code corpus to train a model for library-oriented code generation. Since it is a common practice for programmers to reuse third-party libraries, in which case"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2206.06888","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/2206.06888/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":"2206.06888","created_at":"2026-07-05T04:31:46.910366+00:00"},{"alias_kind":"arxiv_version","alias_value":"2206.06888v1","created_at":"2026-07-05T04:31:46.910366+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2206.06888","created_at":"2026-07-05T04:31:46.910366+00:00"},{"alias_kind":"pith_short_12","alias_value":"WX6EYHNBZC2T","created_at":"2026-07-05T04:31:46.910366+00:00"},{"alias_kind":"pith_short_16","alias_value":"WX6EYHNBZC2TJLXA","created_at":"2026-07-05T04:31:46.910366+00:00"},{"alias_kind":"pith_short_8","alias_value":"WX6EYHNB","created_at":"2026-07-05T04:31:46.910366+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2402.13228","citing_title":"Smaug: Fixing Failure Modes of Preference Optimisation with DPO-Positive","ref_index":129,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22659","citing_title":"RealBench: A Repo-Level Code Generation Benchmark Aligned with Real-World Software Development Practices","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02431","citing_title":"ARIADNE: Agentic Reward-Informed Adaptive Decision Exploration via Blackboard-Driven MCTS for Competitive Program Generation","ref_index":28,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP","json":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP.json","graph_json":"https://pith.science/api/pith-number/WX6EYHNBZC2TJLXAPEL6NWHODP/graph.json","events_json":"https://pith.science/api/pith-number/WX6EYHNBZC2TJLXAPEL6NWHODP/events.json","paper":"https://pith.science/paper/WX6EYHNB"},"agent_actions":{"view_html":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP","download_json":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP.json","view_paper":"https://pith.science/paper/WX6EYHNB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2206.06888&json=true","fetch_graph":"https://pith.science/api/pith-number/WX6EYHNBZC2TJLXAPEL6NWHODP/graph.json","fetch_events":"https://pith.science/api/pith-number/WX6EYHNBZC2TJLXAPEL6NWHODP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP/action/storage_attestation","attest_author":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP/action/author_attestation","sign_citation":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP/action/citation_signature","submit_replication":"https://pith.science/pith/WX6EYHNBZC2TJLXAPEL6NWHODP/action/replication_record"}},"created_at":"2026-07-05T04:31:46.910366+00:00","updated_at":"2026-07-05T04:31:46.910366+00:00"}