{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YHCSDGTG5TCON25TWWFD6CGPEM","short_pith_number":"pith:YHCSDGTG","schema_version":"1.0","canonical_sha256":"c1c5219a66ecc4e6ebb3b58a3f08cf23061c6e82f3648168f1c03833546324c6","source":{"kind":"arxiv","id":"2506.08436","version":1},"attestation_state":"computed","paper":{"title":"Olica: Efficient Structured Pruning of Large Language Models without Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Huazhen Lin, Jiujun He","submitted_at":"2025-06-10T04:19:38Z","abstract_excerpt":"Most existing structured pruning methods for Large Language Models (LLMs) require substantial computational and data resources for retraining to reestablish the corrupted correlations, making them prohibitively expensive. To address this, we propose a pruning framework for LLMs called Orthogonal decomposition and Linear Calibration (Olica), which eliminates the need for retraining. A key observation is that the multi-head attention (MHA) layer depends on two types of matrix products. By treating these matrix products as unified entities and applying principal component analysis (PCA), we extra"},"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":"2506.08436","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-10T04:19:38Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"38a2da5d4ef8382ce2c3ec656fe076fa8e7d0abc122337fc52b39709f495c9ff","abstract_canon_sha256":"518a281ad3435b8801d6a966e6481647f0bfaef68e6d338e884f36ce643b37e9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:49.468404Z","signature_b64":"vsxgK3EyUgrJxtbfHgDuABl/6naB2n0yLLbcVkDKUf097L0BTgIOdtffbtqEf6gYtn4Jy229ZddLG9SZ/YJoCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c1c5219a66ecc4e6ebb3b58a3f08cf23061c6e82f3648168f1c03833546324c6","last_reissued_at":"2026-07-05T11:18:49.467931Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:49.467931Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Olica: Efficient Structured Pruning of Large Language Models without Retraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Huazhen Lin, Jiujun He","submitted_at":"2025-06-10T04:19:38Z","abstract_excerpt":"Most existing structured pruning methods for Large Language Models (LLMs) require substantial computational and data resources for retraining to reestablish the corrupted correlations, making them prohibitively expensive. To address this, we propose a pruning framework for LLMs called Orthogonal decomposition and Linear Calibration (Olica), which eliminates the need for retraining. A key observation is that the multi-head attention (MHA) layer depends on two types of matrix products. By treating these matrix products as unified entities and applying principal component analysis (PCA), we extra"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.08436","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/2506.08436/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":"2506.08436","created_at":"2026-07-05T11:18:49.467983+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.08436v1","created_at":"2026-07-05T11:18:49.467983+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.08436","created_at":"2026-07-05T11:18:49.467983+00:00"},{"alias_kind":"pith_short_12","alias_value":"YHCSDGTG5TCO","created_at":"2026-07-05T11:18:49.467983+00:00"},{"alias_kind":"pith_short_16","alias_value":"YHCSDGTG5TCON25T","created_at":"2026-07-05T11:18:49.467983+00:00"},{"alias_kind":"pith_short_8","alias_value":"YHCSDGTG","created_at":"2026-07-05T11:18:49.467983+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/YHCSDGTG5TCON25TWWFD6CGPEM","json":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM.json","graph_json":"https://pith.science/api/pith-number/YHCSDGTG5TCON25TWWFD6CGPEM/graph.json","events_json":"https://pith.science/api/pith-number/YHCSDGTG5TCON25TWWFD6CGPEM/events.json","paper":"https://pith.science/paper/YHCSDGTG"},"agent_actions":{"view_html":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM","download_json":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM.json","view_paper":"https://pith.science/paper/YHCSDGTG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.08436&json=true","fetch_graph":"https://pith.science/api/pith-number/YHCSDGTG5TCON25TWWFD6CGPEM/graph.json","fetch_events":"https://pith.science/api/pith-number/YHCSDGTG5TCON25TWWFD6CGPEM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM/action/storage_attestation","attest_author":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM/action/author_attestation","sign_citation":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM/action/citation_signature","submit_replication":"https://pith.science/pith/YHCSDGTG5TCON25TWWFD6CGPEM/action/replication_record"}},"created_at":"2026-07-05T11:18:49.467983+00:00","updated_at":"2026-07-05T11:18:49.467983+00:00"}