{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QKWNJEWBRVI4OEJEWVNODNOPNZ","short_pith_number":"pith:QKWNJEWB","schema_version":"1.0","canonical_sha256":"82acd492c18d51c71124b55ae1b5cf6e445794b13796260a1c59baeaa93b1b5d","source":{"kind":"arxiv","id":"2506.18135","version":1},"attestation_state":"computed","paper":{"title":"SE-Merging: A Self-Enhanced Approach for Dynamic Model Merging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bo Zhang, Junchi Yan, Weinan Zhang, Xi Sun, Zhanpeng Zhou, Zijun Chen","submitted_at":"2025-06-22T18:38:41Z","abstract_excerpt":"Model merging has gained increasing attention due to its intriguing property: interpolating the parameters of different task-specific fine-tuned models leads to multi-task abilities. However, despite its empirical success, the underlying mechanisms of model merging remain poorly understood. In this work, we delve into the mechanism behind model merging from a representation perspective. Our analysis reveals that model merging achieves multi-task abilities through two key capabilities: i) distinguishing samples from different tasks, and ii) adapting to the corresponding expert model for each sa"},"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":"2506.18135","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-22T18:38:41Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"5724d6adbdcb07307e1f46c16d2dafdf7114d938922a55a7aad87c7f972ed0d2","abstract_canon_sha256":"66436758e04db8702c9043d027c50eeb1dfd6e4a91ead93e32c8d0570f3f456c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:25:29.490606Z","signature_b64":"KkgFTW+f5NkQPd8hI+D1BNCmZdQ1LzWP6sBq0hJhAqT8zkQe0ttfFz5x08l7lR1FP6En4gd611RL/HGQ9624Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"82acd492c18d51c71124b55ae1b5cf6e445794b13796260a1c59baeaa93b1b5d","last_reissued_at":"2026-07-05T11:25:29.490121Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:25:29.490121Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SE-Merging: A Self-Enhanced Approach for Dynamic Model Merging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Bo Zhang, Junchi Yan, Weinan Zhang, Xi Sun, Zhanpeng Zhou, Zijun Chen","submitted_at":"2025-06-22T18:38:41Z","abstract_excerpt":"Model merging has gained increasing attention due to its intriguing property: interpolating the parameters of different task-specific fine-tuned models leads to multi-task abilities. However, despite its empirical success, the underlying mechanisms of model merging remain poorly understood. In this work, we delve into the mechanism behind model merging from a representation perspective. Our analysis reveals that model merging achieves multi-task abilities through two key capabilities: i) distinguishing samples from different tasks, and ii) adapting to the corresponding expert model for each sa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18135","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/2506.18135/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":"2506.18135","created_at":"2026-07-05T11:25:29.490180+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18135v1","created_at":"2026-07-05T11:25:29.490180+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18135","created_at":"2026-07-05T11:25:29.490180+00:00"},{"alias_kind":"pith_short_12","alias_value":"QKWNJEWBRVI4","created_at":"2026-07-05T11:25:29.490180+00:00"},{"alias_kind":"pith_short_16","alias_value":"QKWNJEWBRVI4OEJE","created_at":"2026-07-05T11:25:29.490180+00:00"},{"alias_kind":"pith_short_8","alias_value":"QKWNJEWB","created_at":"2026-07-05T11:25:29.490180+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10617","citing_title":"SSR-Merge: Subspace Signal Routing for Training-Free LoRA Merging in Diffusion Models","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ","json":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ.json","graph_json":"https://pith.science/api/pith-number/QKWNJEWBRVI4OEJEWVNODNOPNZ/graph.json","events_json":"https://pith.science/api/pith-number/QKWNJEWBRVI4OEJEWVNODNOPNZ/events.json","paper":"https://pith.science/paper/QKWNJEWB"},"agent_actions":{"view_html":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ","download_json":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ.json","view_paper":"https://pith.science/paper/QKWNJEWB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18135&json=true","fetch_graph":"https://pith.science/api/pith-number/QKWNJEWBRVI4OEJEWVNODNOPNZ/graph.json","fetch_events":"https://pith.science/api/pith-number/QKWNJEWBRVI4OEJEWVNODNOPNZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ/action/storage_attestation","attest_author":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ/action/author_attestation","sign_citation":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ/action/citation_signature","submit_replication":"https://pith.science/pith/QKWNJEWBRVI4OEJEWVNODNOPNZ/action/replication_record"}},"created_at":"2026-07-05T11:25:29.490180+00:00","updated_at":"2026-07-05T11:25:29.490180+00:00"}