{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:L55F7NMRYN7VIPPN5R7ECR4N24","short_pith_number":"pith:L55F7NMR","schema_version":"1.0","canonical_sha256":"5f7a5fb591c37f543dedec7e41478dd71ebde67e78d6b9e6a16cf05d739f6460","source":{"kind":"arxiv","id":"2102.07492","version":3},"attestation_state":"computed","paper":{"title":"DOBF: A Deobfuscation Pre-Training Objective for Programming Languages","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Baptiste Roziere, Guillaume Lample, Marc Szafraniec, Marie-Anne Lachaux","submitted_at":"2021-02-15T11:50:47Z","abstract_excerpt":"Recent advances in self-supervised learning have dramatically improved the state of the art on a wide variety of tasks. However, research in language model pre-training has mostly focused on natural languages, and it is unclear whether models like BERT and its variants provide the best pre-training when applied to other modalities, such as source code. In this paper, we introduce a new pre-training objective, DOBF, that leverages the structural aspect of programming languages and pre-trains a model to recover the original version of obfuscated source code. We show that models pre-trained with "},"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":"2102.07492","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-15T11:50:47Z","cross_cats_sorted":[],"title_canon_sha256":"cd459ed0d8a447ca61a00ced9ffaab873440f21d8042417092c7191948107161","abstract_canon_sha256":"5db40195e49bd354901f5ce043114730eac375fbcd2c6f352304fe1e78319e39"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:26:38.832351Z","signature_b64":"6GrQJvacQcUr2RpZlmgJGz1EhA4PP6WrWfE871lvJ0+VdhcO2TVkCfE2ZYoyhat9KqOfdDBciEa0k5+cSJNOCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5f7a5fb591c37f543dedec7e41478dd71ebde67e78d6b9e6a16cf05d739f6460","last_reissued_at":"2026-07-05T03:26:38.831922Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:26:38.831922Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DOBF: A Deobfuscation Pre-Training Objective for Programming Languages","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Baptiste Roziere, Guillaume Lample, Marc Szafraniec, Marie-Anne Lachaux","submitted_at":"2021-02-15T11:50:47Z","abstract_excerpt":"Recent advances in self-supervised learning have dramatically improved the state of the art on a wide variety of tasks. However, research in language model pre-training has mostly focused on natural languages, and it is unclear whether models like BERT and its variants provide the best pre-training when applied to other modalities, such as source code. In this paper, we introduce a new pre-training objective, DOBF, that leverages the structural aspect of programming languages and pre-trains a model to recover the original version of obfuscated source code. We show that models pre-trained with "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.07492","kind":"arxiv","version":3},"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/2102.07492/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":"2102.07492","created_at":"2026-07-05T03:26:38.831986+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.07492v3","created_at":"2026-07-05T03:26:38.831986+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.07492","created_at":"2026-07-05T03:26:38.831986+00:00"},{"alias_kind":"pith_short_12","alias_value":"L55F7NMRYN7V","created_at":"2026-07-05T03:26:38.831986+00:00"},{"alias_kind":"pith_short_16","alias_value":"L55F7NMRYN7VIPPN","created_at":"2026-07-05T03:26:38.831986+00:00"},{"alias_kind":"pith_short_8","alias_value":"L55F7NMR","created_at":"2026-07-05T03:26:38.831986+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2109.00859","citing_title":"CodeT5: Identifier-aware Unified Pre-trained Encoder-Decoder Models for Code Understanding and Generation","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2305.06161","citing_title":"StarCoder: may the source be with you!","ref_index":143,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24","json":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24.json","graph_json":"https://pith.science/api/pith-number/L55F7NMRYN7VIPPN5R7ECR4N24/graph.json","events_json":"https://pith.science/api/pith-number/L55F7NMRYN7VIPPN5R7ECR4N24/events.json","paper":"https://pith.science/paper/L55F7NMR"},"agent_actions":{"view_html":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24","download_json":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24.json","view_paper":"https://pith.science/paper/L55F7NMR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.07492&json=true","fetch_graph":"https://pith.science/api/pith-number/L55F7NMRYN7VIPPN5R7ECR4N24/graph.json","fetch_events":"https://pith.science/api/pith-number/L55F7NMRYN7VIPPN5R7ECR4N24/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24/action/storage_attestation","attest_author":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24/action/author_attestation","sign_citation":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24/action/citation_signature","submit_replication":"https://pith.science/pith/L55F7NMRYN7VIPPN5R7ECR4N24/action/replication_record"}},"created_at":"2026-07-05T03:26:38.831986+00:00","updated_at":"2026-07-05T03:26:38.831986+00:00"}