{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5MSAVHF4IADIZHDWSPTKW5CDLO","short_pith_number":"pith:5MSAVHF4","schema_version":"1.0","canonical_sha256":"eb240a9cbc40068c9c7693e6ab74435ba8fb6e4f5c07a10fa649bc197b38050d","source":{"kind":"arxiv","id":"2605.29295","version":1},"attestation_state":"computed","paper":{"title":"EvoGM: Learning to Merge LLMs via Evolutionary Generative Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.NE","authors_text":"Chenhao Yi, Dongmei Jiang, Jianguo Zhang, Ran Cheng, Tao Jiang, Xinmeng Yu, Yan Li, Yiling Wu","submitted_at":"2026-05-28T03:22:54Z","abstract_excerpt":"Evolutionary model merging provides a powerful framework for the automated, training-free composition of LLMs through parameter-space search. However, existing methods predominantly rely on stochastic, hand-crafted operators that overlook the underlying performance landscape of the coefficient space. We propose Evolutionary Generative Merging (EvoGM), a framework that transcends manual heuristics by employing learnable generative modeling to optimize merging coefficients. Specifically, EvoGM features a dual-generator architecture with cycle-consistent learning to adaptively sample and refine p"},"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":"2605.29295","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.NE","submitted_at":"2026-05-28T03:22:54Z","cross_cats_sorted":[],"title_canon_sha256":"ce52c2ab433d9d9b65cd7f5634c80d56e837951d7a8682d1b1ed60ff899a6d14","abstract_canon_sha256":"d91fe56eaac130c981cccdef0091268739749e6c92d25591b64a9d06823f239f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:30.024951Z","signature_b64":"FPxg81E4DOGnIIp0X+yg/d//rKQwGrxrlVcqIbSMvUvZXbcvV/JU7zbH8sJ5DS0RuuaQRKM/q2BCnllUSW+NAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb240a9cbc40068c9c7693e6ab74435ba8fb6e4f5c07a10fa649bc197b38050d","last_reissued_at":"2026-05-29T01:05:30.023975Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:30.023975Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EvoGM: Learning to Merge LLMs via Evolutionary Generative Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.NE","authors_text":"Chenhao Yi, Dongmei Jiang, Jianguo Zhang, Ran Cheng, Tao Jiang, Xinmeng Yu, Yan Li, Yiling Wu","submitted_at":"2026-05-28T03:22:54Z","abstract_excerpt":"Evolutionary model merging provides a powerful framework for the automated, training-free composition of LLMs through parameter-space search. However, existing methods predominantly rely on stochastic, hand-crafted operators that overlook the underlying performance landscape of the coefficient space. We propose Evolutionary Generative Merging (EvoGM), a framework that transcends manual heuristics by employing learnable generative modeling to optimize merging coefficients. Specifically, EvoGM features a dual-generator architecture with cycle-consistent learning to adaptively sample and refine p"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29295","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/2605.29295/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":"2605.29295","created_at":"2026-05-29T01:05:30.024135+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.29295v1","created_at":"2026-05-29T01:05:30.024135+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29295","created_at":"2026-05-29T01:05:30.024135+00:00"},{"alias_kind":"pith_short_12","alias_value":"5MSAVHF4IADI","created_at":"2026-05-29T01:05:30.024135+00:00"},{"alias_kind":"pith_short_16","alias_value":"5MSAVHF4IADIZHDW","created_at":"2026-05-29T01:05:30.024135+00:00"},{"alias_kind":"pith_short_8","alias_value":"5MSAVHF4","created_at":"2026-05-29T01:05:30.024135+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO","json":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO.json","graph_json":"https://pith.science/api/pith-number/5MSAVHF4IADIZHDWSPTKW5CDLO/graph.json","events_json":"https://pith.science/api/pith-number/5MSAVHF4IADIZHDWSPTKW5CDLO/events.json","paper":"https://pith.science/paper/5MSAVHF4"},"agent_actions":{"view_html":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO","download_json":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO.json","view_paper":"https://pith.science/paper/5MSAVHF4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.29295&json=true","fetch_graph":"https://pith.science/api/pith-number/5MSAVHF4IADIZHDWSPTKW5CDLO/graph.json","fetch_events":"https://pith.science/api/pith-number/5MSAVHF4IADIZHDWSPTKW5CDLO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO/action/storage_attestation","attest_author":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO/action/author_attestation","sign_citation":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO/action/citation_signature","submit_replication":"https://pith.science/pith/5MSAVHF4IADIZHDWSPTKW5CDLO/action/replication_record"}},"created_at":"2026-05-29T01:05:30.024135+00:00","updated_at":"2026-05-29T01:05:30.024135+00:00"}