{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7WQTLDTN6IDF5XMC7V5NKT3P3Z","short_pith_number":"pith:7WQTLDTN","schema_version":"1.0","canonical_sha256":"fda1358e6df2065edd82fd7ad54f6fde7e1448598486a6c4adfe65449ef7ceee","source":{"kind":"arxiv","id":"2507.10540","version":3},"attestation_state":"computed","paper":{"title":"FusionFactory: Fusing LLM Capabilities with Multi-LLM Log Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bryan Catanzaro, Haozhen Zhang, Jiaxuan You, Mohammad Shoeybi, Mostofa Patwary, Pengrui Han, Tao Feng, Zijie Lei","submitted_at":"2025-07-14T17:58:02Z","abstract_excerpt":"The rapid advancement of large language models (LLMs) has created a diverse landscape of models, each excelling at different tasks. This diversity drives researchers to employ multiple LLMs in practice, leaving behind valuable multi-LLM log data. This naturally leads to the question of whether such logs can be fully leveraged to fuse LLMs' complementary capabilities. Although prior work has explored various strategies for integrating multiple LLMs, we argue that practical fusion must meet two essential requirements: (1) compatibility with real-world serving scenarios (e.g., local and API-based"},"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":"2507.10540","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-14T17:58:02Z","cross_cats_sorted":[],"title_canon_sha256":"138aaf5349a7d7fdc6b48a22689f720cb1319daad536457c6b58fa2a8106584f","abstract_canon_sha256":"d5fb55fd6a2aa8daf2c5d7ae64d66195b1e2cc3940911956d0489dbac77e80ad"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-02T01:17:13.126140Z","signature_b64":"+hodMcV6e9k7FQ0THQkRNyItKrK9N9eQvqcEP7KNv0E3FphEHcukPvVv4lgGJBfigN98ztL9rIHFyEx0I2mEDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fda1358e6df2065edd82fd7ad54f6fde7e1448598486a6c4adfe65449ef7ceee","last_reissued_at":"2026-07-02T01:17:13.125630Z","signature_status":"signed_v1","first_computed_at":"2026-07-02T01:17:13.125630Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FusionFactory: Fusing LLM Capabilities with Multi-LLM Log Data","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bryan Catanzaro, Haozhen Zhang, Jiaxuan You, Mohammad Shoeybi, Mostofa Patwary, Pengrui Han, Tao Feng, Zijie Lei","submitted_at":"2025-07-14T17:58:02Z","abstract_excerpt":"The rapid advancement of large language models (LLMs) has created a diverse landscape of models, each excelling at different tasks. This diversity drives researchers to employ multiple LLMs in practice, leaving behind valuable multi-LLM log data. This naturally leads to the question of whether such logs can be fully leveraged to fuse LLMs' complementary capabilities. Although prior work has explored various strategies for integrating multiple LLMs, we argue that practical fusion must meet two essential requirements: (1) compatibility with real-world serving scenarios (e.g., local and API-based"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10540","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/2507.10540/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":"2507.10540","created_at":"2026-07-02T01:17:13.125700+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.10540v3","created_at":"2026-07-02T01:17:13.125700+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10540","created_at":"2026-07-02T01:17:13.125700+00:00"},{"alias_kind":"pith_short_12","alias_value":"7WQTLDTN6IDF","created_at":"2026-07-02T01:17:13.125700+00:00"},{"alias_kind":"pith_short_16","alias_value":"7WQTLDTN6IDF5XMC","created_at":"2026-07-02T01:17:13.125700+00:00"},{"alias_kind":"pith_short_8","alias_value":"7WQTLDTN","created_at":"2026-07-02T01:17:13.125700+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2607.00053","citing_title":"SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasks","ref_index":59,"is_internal_anchor":true},{"citing_arxiv_id":"2604.01674","citing_title":"Can Heterogeneous Language Models Be Fused?","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15022","citing_title":"Route to Rome Attack: Directing LLM Routers to Expensive Models via Adversarial Suffix Optimization","ref_index":13,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z","json":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z.json","graph_json":"https://pith.science/api/pith-number/7WQTLDTN6IDF5XMC7V5NKT3P3Z/graph.json","events_json":"https://pith.science/api/pith-number/7WQTLDTN6IDF5XMC7V5NKT3P3Z/events.json","paper":"https://pith.science/paper/7WQTLDTN"},"agent_actions":{"view_html":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z","download_json":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z.json","view_paper":"https://pith.science/paper/7WQTLDTN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.10540&json=true","fetch_graph":"https://pith.science/api/pith-number/7WQTLDTN6IDF5XMC7V5NKT3P3Z/graph.json","fetch_events":"https://pith.science/api/pith-number/7WQTLDTN6IDF5XMC7V5NKT3P3Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z/action/storage_attestation","attest_author":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z/action/author_attestation","sign_citation":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z/action/citation_signature","submit_replication":"https://pith.science/pith/7WQTLDTN6IDF5XMC7V5NKT3P3Z/action/replication_record"}},"created_at":"2026-07-02T01:17:13.125700+00:00","updated_at":"2026-07-02T01:17:13.125700+00:00"}