{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:2QAR4YWFPB5B3ZXPFYWED4II2A","short_pith_number":"pith:2QAR4YWF","schema_version":"1.0","canonical_sha256":"d4011e62c5787a1de6ef2e2c41f108d02203e343555d556f0e0fb53bdf60856d","source":{"kind":"arxiv","id":"2304.05216","version":1},"attestation_state":"computed","paper":{"title":"Towards Efficient Fine-tuning of Pre-trained Code Models: An Experimental Study and Beyond","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Dongmei Zhang, Ensheng Shi, Hongbin Sun, Hongyu Zhang, Lun Du, Shi Han, Yanlin Wang","submitted_at":"2023-04-11T13:34:13Z","abstract_excerpt":"Recently, fine-tuning pre-trained code models such as CodeBERT on downstream tasks has achieved great success in many software testing and analysis tasks. While effective and prevalent, fine-tuning the pre-trained parameters incurs a large computational cost. In this paper, we conduct an extensive experimental study to explore what happens to layer-wise pre-trained representations and their encoded code knowledge during fine-tuning. We then propose efficient alternatives to fine-tune the large pre-trained code model based on the above findings. Our experimental study shows that (1) lexical, sy"},"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":"2304.05216","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2023-04-11T13:34:13Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"6c10f322c692c2b6dbb2aba70a0f4e1ace410507109aad5f84894915f64dd43c","abstract_canon_sha256":"2b44a97b32209e906cc592cd85303e328f43c9c37b7c583b4ef87f36e8cc6256"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:59:57.302134Z","signature_b64":"F5WVLyg+1ZAoQMxuw58TiRpkgb3OaeCwo3hGbtKk0nB1vHyNtVb1RWOcTY8eYcisLL8FSerjRQxpLfUhfwMUCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d4011e62c5787a1de6ef2e2c41f108d02203e343555d556f0e0fb53bdf60856d","last_reissued_at":"2026-07-05T05:59:57.301738Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:59:57.301738Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Efficient Fine-tuning of Pre-trained Code Models: An Experimental Study and Beyond","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Dongmei Zhang, Ensheng Shi, Hongbin Sun, Hongyu Zhang, Lun Du, Shi Han, Yanlin Wang","submitted_at":"2023-04-11T13:34:13Z","abstract_excerpt":"Recently, fine-tuning pre-trained code models such as CodeBERT on downstream tasks has achieved great success in many software testing and analysis tasks. While effective and prevalent, fine-tuning the pre-trained parameters incurs a large computational cost. In this paper, we conduct an extensive experimental study to explore what happens to layer-wise pre-trained representations and their encoded code knowledge during fine-tuning. We then propose efficient alternatives to fine-tune the large pre-trained code model based on the above findings. Our experimental study shows that (1) lexical, sy"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.05216","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/2304.05216/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":"2304.05216","created_at":"2026-07-05T05:59:57.301795+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.05216v1","created_at":"2026-07-05T05:59:57.301795+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.05216","created_at":"2026-07-05T05:59:57.301795+00:00"},{"alias_kind":"pith_short_12","alias_value":"2QAR4YWFPB5B","created_at":"2026-07-05T05:59:57.301795+00:00"},{"alias_kind":"pith_short_16","alias_value":"2QAR4YWFPB5B3ZXP","created_at":"2026-07-05T05:59:57.301795+00:00"},{"alias_kind":"pith_short_8","alias_value":"2QAR4YWF","created_at":"2026-07-05T05:59:57.301795+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2412.00069","citing_title":"Condense, Don't Just Prune: Enhancing Efficiency and Performance in MoE Layer Pruning","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A","json":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A.json","graph_json":"https://pith.science/api/pith-number/2QAR4YWFPB5B3ZXPFYWED4II2A/graph.json","events_json":"https://pith.science/api/pith-number/2QAR4YWFPB5B3ZXPFYWED4II2A/events.json","paper":"https://pith.science/paper/2QAR4YWF"},"agent_actions":{"view_html":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A","download_json":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A.json","view_paper":"https://pith.science/paper/2QAR4YWF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.05216&json=true","fetch_graph":"https://pith.science/api/pith-number/2QAR4YWFPB5B3ZXPFYWED4II2A/graph.json","fetch_events":"https://pith.science/api/pith-number/2QAR4YWFPB5B3ZXPFYWED4II2A/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A/action/storage_attestation","attest_author":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A/action/author_attestation","sign_citation":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A/action/citation_signature","submit_replication":"https://pith.science/pith/2QAR4YWFPB5B3ZXPFYWED4II2A/action/replication_record"}},"created_at":"2026-07-05T05:59:57.301795+00:00","updated_at":"2026-07-05T05:59:57.301795+00:00"}