{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:S5NYABVGRGUZIBGMDTDB3KJ7LX","short_pith_number":"pith:S5NYABVG","schema_version":"1.0","canonical_sha256":"975b8006a689a99404cc1cc61da93f5dd55daaec9b04a4534240f7f7277306be","source":{"kind":"arxiv","id":"2607.17568","version":1},"attestation_state":"computed","paper":{"title":"CoCurve: Cross-Module Co-Pruning Curvature for Training-Free Structured LLM Pruning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chau Yuen, Tiantong Wang, Wei Yang Bryan Lim, Zhiren Gong, Zihao Zeng, Zijie Wang","submitted_at":"2026-07-20T05:27:57Z","abstract_excerpt":"Structured pruning compresses large language models (LLMs) by removing whole computational units, such as attention heads and feed-forward (FFN) channel groups. Most training-free methods, however, rank these units independently, implicitly treating the loss from pruning a set as the sum of its individual losses. This view fails for Transformers, whose sublayers are coupled through a shared residual stream. Two individually weak units can thus be jointly indispensable, yet independent scoring is blind to such dependence and removes them together. We introduce CoCurve (Cross-Module Co-Pruning C"},"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":"2607.17568","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-20T05:27:57Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c4c9d49633509baf35408bfb9a52859ef997f50477500a26c6fd610c161b1d35","abstract_canon_sha256":"e123e7e2d6b8d2843d3db0e3311d3c083a21b3b590e3e01d894a1aa9520dacbb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T01:21:52.651453Z","signature_b64":"N0ArWvEuD4sngg+iKM63PZTAMxOsQbvV/vH2fIglMs6emeolrzv0rAXwrCLqa3xwwUa/Kc0nn75rPGtPhYrwCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"975b8006a689a99404cc1cc61da93f5dd55daaec9b04a4534240f7f7277306be","last_reissued_at":"2026-07-21T01:21:52.650450Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T01:21:52.650450Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CoCurve: Cross-Module Co-Pruning Curvature for Training-Free Structured LLM Pruning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chau Yuen, Tiantong Wang, Wei Yang Bryan Lim, Zhiren Gong, Zihao Zeng, Zijie Wang","submitted_at":"2026-07-20T05:27:57Z","abstract_excerpt":"Structured pruning compresses large language models (LLMs) by removing whole computational units, such as attention heads and feed-forward (FFN) channel groups. Most training-free methods, however, rank these units independently, implicitly treating the loss from pruning a set as the sum of its individual losses. This view fails for Transformers, whose sublayers are coupled through a shared residual stream. Two individually weak units can thus be jointly indispensable, yet independent scoring is blind to such dependence and removes them together. We introduce CoCurve (Cross-Module Co-Pruning C"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.17568","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/2607.17568/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":"2607.17568","created_at":"2026-07-21T01:21:52.650903+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.17568v1","created_at":"2026-07-21T01:21:52.650903+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.17568","created_at":"2026-07-21T01:21:52.650903+00:00"},{"alias_kind":"pith_short_12","alias_value":"S5NYABVGRGUZ","created_at":"2026-07-21T01:21:52.650903+00:00"},{"alias_kind":"pith_short_16","alias_value":"S5NYABVGRGUZIBGM","created_at":"2026-07-21T01:21:52.650903+00:00"},{"alias_kind":"pith_short_8","alias_value":"S5NYABVG","created_at":"2026-07-21T01:21:52.650903+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/S5NYABVGRGUZIBGMDTDB3KJ7LX","json":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX.json","graph_json":"https://pith.science/api/pith-number/S5NYABVGRGUZIBGMDTDB3KJ7LX/graph.json","events_json":"https://pith.science/api/pith-number/S5NYABVGRGUZIBGMDTDB3KJ7LX/events.json","paper":"https://pith.science/paper/S5NYABVG"},"agent_actions":{"view_html":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX","download_json":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX.json","view_paper":"https://pith.science/paper/S5NYABVG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.17568&json=true","fetch_graph":"https://pith.science/api/pith-number/S5NYABVGRGUZIBGMDTDB3KJ7LX/graph.json","fetch_events":"https://pith.science/api/pith-number/S5NYABVGRGUZIBGMDTDB3KJ7LX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX/action/storage_attestation","attest_author":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX/action/author_attestation","sign_citation":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX/action/citation_signature","submit_replication":"https://pith.science/pith/S5NYABVGRGUZIBGMDTDB3KJ7LX/action/replication_record"}},"created_at":"2026-07-21T01:21:52.650903+00:00","updated_at":"2026-07-21T01:21:52.650903+00:00"}