{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:L547C2RHY5P37AYXLN335X2CNS","short_pith_number":"pith:L547C2RH","schema_version":"1.0","canonical_sha256":"5f79f16a27c75fbf83175b77bedf426c88db1da7365f3afa8ea624cf78b9536b","source":{"kind":"arxiv","id":"2411.11266","version":5},"attestation_state":"computed","paper":{"title":"VersaTune: An Efficient Data Composition Framework for Training Multi-Capability LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bin Cui, Da Pan, Guosheng Dong, Keer Lu, Keshi Zhao, Shusen Zhang, Tengjiao Wang, Wentao Zhang, Xin Wu, Zheng Liang, Zhuoran Zhang","submitted_at":"2024-11-18T03:45:34Z","abstract_excerpt":"As demonstrated by the proprietary Large Language Models (LLMs) such as GPT and Claude series, LLMs have the potential to achieve remarkable proficiency across a wide range of domains, including law, medicine, finance, science, code, etc., all within a single model. These capabilities are further augmented during the Supervised Fine-Tuning (SFT) phase. Despite their potential, existing work mainly focuses on domain-specific enhancements during fine-tuning, the challenge of which lies in catastrophic forgetting of knowledge across other domains. In this study, we introduce **VersaTune**, a nove"},"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":"2411.11266","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-18T03:45:34Z","cross_cats_sorted":[],"title_canon_sha256":"fe4540faed8773c5b81bebd45ba5c4e8d333506f9f012624e9f496685e74c69a","abstract_canon_sha256":"48d8473c36946e512fad2e44625b217b18d7d368e7ed9d2a394453ce6918f228"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:05:22.278199Z","signature_b64":"FfO/KsKXPjWkakoPArgpXMLkwWFsBATlRvp6egy7r/T1cPmtD5x+Id8+CNbhXYJ25i+LmMqUgzCQ4GDtxjbKAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5f79f16a27c75fbf83175b77bedf426c88db1da7365f3afa8ea624cf78b9536b","last_reissued_at":"2026-07-05T11:05:22.277706Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:05:22.277706Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VersaTune: An Efficient Data Composition Framework for Training Multi-Capability LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bin Cui, Da Pan, Guosheng Dong, Keer Lu, Keshi Zhao, Shusen Zhang, Tengjiao Wang, Wentao Zhang, Xin Wu, Zheng Liang, Zhuoran Zhang","submitted_at":"2024-11-18T03:45:34Z","abstract_excerpt":"As demonstrated by the proprietary Large Language Models (LLMs) such as GPT and Claude series, LLMs have the potential to achieve remarkable proficiency across a wide range of domains, including law, medicine, finance, science, code, etc., all within a single model. These capabilities are further augmented during the Supervised Fine-Tuning (SFT) phase. Despite their potential, existing work mainly focuses on domain-specific enhancements during fine-tuning, the challenge of which lies in catastrophic forgetting of knowledge across other domains. In this study, we introduce **VersaTune**, a nove"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.11266","kind":"arxiv","version":5},"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/2411.11266/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":"2411.11266","created_at":"2026-07-05T11:05:22.277766+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.11266v5","created_at":"2026-07-05T11:05:22.277766+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.11266","created_at":"2026-07-05T11:05:22.277766+00:00"},{"alias_kind":"pith_short_12","alias_value":"L547C2RHY5P3","created_at":"2026-07-05T11:05:22.277766+00:00"},{"alias_kind":"pith_short_16","alias_value":"L547C2RHY5P37AYX","created_at":"2026-07-05T11:05:22.277766+00:00"},{"alias_kind":"pith_short_8","alias_value":"L547C2RH","created_at":"2026-07-05T11:05:22.277766+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20416","citing_title":"GraphGen: Enhancing Supervised Fine-Tuning for LLMs with Knowledge-Driven Synthetic Data Generation","ref_index":23,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS","json":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS.json","graph_json":"https://pith.science/api/pith-number/L547C2RHY5P37AYXLN335X2CNS/graph.json","events_json":"https://pith.science/api/pith-number/L547C2RHY5P37AYXLN335X2CNS/events.json","paper":"https://pith.science/paper/L547C2RH"},"agent_actions":{"view_html":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS","download_json":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS.json","view_paper":"https://pith.science/paper/L547C2RH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.11266&json=true","fetch_graph":"https://pith.science/api/pith-number/L547C2RHY5P37AYXLN335X2CNS/graph.json","fetch_events":"https://pith.science/api/pith-number/L547C2RHY5P37AYXLN335X2CNS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS/action/storage_attestation","attest_author":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS/action/author_attestation","sign_citation":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS/action/citation_signature","submit_replication":"https://pith.science/pith/L547C2RHY5P37AYXLN335X2CNS/action/replication_record"}},"created_at":"2026-07-05T11:05:22.277766+00:00","updated_at":"2026-07-05T11:05:22.277766+00:00"}