{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ROTOHUFPYB2APP6SROBSWJI35N","short_pith_number":"pith:ROTOHUFP","schema_version":"1.0","canonical_sha256":"8ba6e3d0afc07407bfd28b832b251beb68b52dba2f92175bc533e34445df9256","source":{"kind":"arxiv","id":"2406.10305","version":2},"attestation_state":"computed","paper":{"title":"Unlock the Correlation between Supervised Fine-Tuning and Reinforcement Learning in Training Code Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.SE","authors_text":"Jie Chen, Liang Xiang, Xintian Han, Xun Zhou, Yu Ma","submitted_at":"2024-06-14T03:39:01Z","abstract_excerpt":"Automatic code generation has been a longstanding research topic. With the advancement of general-purpose large language models (LLMs), the ability to code stands out as one important measure to the model's reasoning performance. Usually, a two-stage training paradigm is implemented to obtain a Code LLM, namely the pretraining and the fine-tuning. Within the fine-tuning, supervised fine-tuning (SFT), and reinforcement learning (RL) are often used to improve the model's zero-shot ability. A large number of work has been conducted to improve the model's performance on code-related benchmarks wit"},"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":"2406.10305","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.SE","submitted_at":"2024-06-14T03:39:01Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"11d3c8d24dd2b65c8256125c6cf014bf9cb185d0b3ec51887c6eeb63f44b45e5","abstract_canon_sha256":"e29635f5d315749aa898f50c233c5fff0e49d0e7f3a4c0f128613f9c766908be"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:50:02.984126Z","signature_b64":"BRoesjacI+A20nMpdVbqbJUaiRiH2XzzNFUreGdlbHIzAnPKuDMeH9Y/ORqqZ5xqsGFIv3tBka2ca8g0K9ZWCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8ba6e3d0afc07407bfd28b832b251beb68b52dba2f92175bc533e34445df9256","last_reissued_at":"2026-07-05T09:50:02.983610Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:50:02.983610Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unlock the Correlation between Supervised Fine-Tuning and Reinforcement Learning in Training Code Large Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.SE","authors_text":"Jie Chen, Liang Xiang, Xintian Han, Xun Zhou, Yu Ma","submitted_at":"2024-06-14T03:39:01Z","abstract_excerpt":"Automatic code generation has been a longstanding research topic. With the advancement of general-purpose large language models (LLMs), the ability to code stands out as one important measure to the model's reasoning performance. Usually, a two-stage training paradigm is implemented to obtain a Code LLM, namely the pretraining and the fine-tuning. Within the fine-tuning, supervised fine-tuning (SFT), and reinforcement learning (RL) are often used to improve the model's zero-shot ability. A large number of work has been conducted to improve the model's performance on code-related benchmarks wit"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.10305","kind":"arxiv","version":2},"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/2406.10305/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":"2406.10305","created_at":"2026-07-05T09:50:02.983663+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.10305v2","created_at":"2026-07-05T09:50:02.983663+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.10305","created_at":"2026-07-05T09:50:02.983663+00:00"},{"alias_kind":"pith_short_12","alias_value":"ROTOHUFPYB2A","created_at":"2026-07-05T09:50:02.983663+00:00"},{"alias_kind":"pith_short_16","alias_value":"ROTOHUFPYB2APP6S","created_at":"2026-07-05T09:50:02.983663+00:00"},{"alias_kind":"pith_short_8","alias_value":"ROTOHUFP","created_at":"2026-07-05T09:50:02.983663+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18089","citing_title":"From Reasoning Traces to Reusable Modules: Understanding Compositional Generalization in Language Model Reasoning","ref_index":142,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16804","citing_title":"AutoOR: Scalably Post-training LLMs to Autoformalize Operations Research Problems","ref_index":68,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N","json":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N.json","graph_json":"https://pith.science/api/pith-number/ROTOHUFPYB2APP6SROBSWJI35N/graph.json","events_json":"https://pith.science/api/pith-number/ROTOHUFPYB2APP6SROBSWJI35N/events.json","paper":"https://pith.science/paper/ROTOHUFP"},"agent_actions":{"view_html":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N","download_json":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N.json","view_paper":"https://pith.science/paper/ROTOHUFP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.10305&json=true","fetch_graph":"https://pith.science/api/pith-number/ROTOHUFPYB2APP6SROBSWJI35N/graph.json","fetch_events":"https://pith.science/api/pith-number/ROTOHUFPYB2APP6SROBSWJI35N/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N/action/storage_attestation","attest_author":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N/action/author_attestation","sign_citation":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N/action/citation_signature","submit_replication":"https://pith.science/pith/ROTOHUFPYB2APP6SROBSWJI35N/action/replication_record"}},"created_at":"2026-07-05T09:50:02.983663+00:00","updated_at":"2026-07-05T09:50:02.983663+00:00"}