{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YWNPY4TGJF4R5ZKNU6ONK6K4Y4","short_pith_number":"pith:YWNPY4TG","schema_version":"1.0","canonical_sha256":"c59afc726649791ee54da79cd5795cc711c53d95d32838b1ef1c256128c97480","source":{"kind":"arxiv","id":"2410.10912","version":1},"attestation_state":"computed","paper":{"title":"AlphaPruning: Using Heavy-Tailed Self Regularization Theory for Improved Layer-wise Pruning of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Haiquan Lu, Michael W. Mahoney, Shiwei Liu, Yaoqing Yang, Yefan Zhou, Zhangyang Wang","submitted_at":"2024-10-14T03:35:11Z","abstract_excerpt":"Recent work on pruning large language models (LLMs) has shown that one can eliminate a large number of parameters without compromising performance, making pruning a promising strategy to reduce LLM model size. Existing LLM pruning strategies typically assign uniform pruning ratios across layers, limiting overall pruning ability; and recent work on layerwise pruning of LLMs is often based on heuristics that can easily lead to suboptimal performance. In this paper, we leverage Heavy-Tailed Self-Regularization (HT-SR) Theory, in particular the shape of empirical spectral densities (ESDs) of weigh"},"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":"2410.10912","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-14T03:35:11Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"34bafd6419a02c6ab531e09a6698125b19860c409543537ecc40f080061d580e","abstract_canon_sha256":"e7c94e38a070b870307a87974921d291cf1e05a955ecfb1acdf40ad0329b6ad1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:20:26.871924Z","signature_b64":"QBo/H9Su8L75pg1B07El5dI7DqMC58el9IjaAcnv7J0zm3y99kh9oUiwM1VYlXhpcNyZvp9i2EN1GFTkPV9qAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c59afc726649791ee54da79cd5795cc711c53d95d32838b1ef1c256128c97480","last_reissued_at":"2026-07-05T09:20:26.871463Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:20:26.871463Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AlphaPruning: Using Heavy-Tailed Self Regularization Theory for Improved Layer-wise Pruning of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Haiquan Lu, Michael W. Mahoney, Shiwei Liu, Yaoqing Yang, Yefan Zhou, Zhangyang Wang","submitted_at":"2024-10-14T03:35:11Z","abstract_excerpt":"Recent work on pruning large language models (LLMs) has shown that one can eliminate a large number of parameters without compromising performance, making pruning a promising strategy to reduce LLM model size. Existing LLM pruning strategies typically assign uniform pruning ratios across layers, limiting overall pruning ability; and recent work on layerwise pruning of LLMs is often based on heuristics that can easily lead to suboptimal performance. In this paper, we leverage Heavy-Tailed Self-Regularization (HT-SR) Theory, in particular the shape of empirical spectral densities (ESDs) of weigh"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.10912","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/2410.10912/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":"2410.10912","created_at":"2026-07-05T09:20:26.871523+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.10912v1","created_at":"2026-07-05T09:20:26.871523+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.10912","created_at":"2026-07-05T09:20:26.871523+00:00"},{"alias_kind":"pith_short_12","alias_value":"YWNPY4TGJF4R","created_at":"2026-07-05T09:20:26.871523+00:00"},{"alias_kind":"pith_short_16","alias_value":"YWNPY4TGJF4R5ZKN","created_at":"2026-07-05T09:20:26.871523+00:00"},{"alias_kind":"pith_short_8","alias_value":"YWNPY4TG","created_at":"2026-07-05T09:20:26.871523+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.03887","citing_title":"Omega-S: A Functional Resilience Index for LLM Fine-Tuning","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4","json":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4.json","graph_json":"https://pith.science/api/pith-number/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/graph.json","events_json":"https://pith.science/api/pith-number/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/events.json","paper":"https://pith.science/paper/YWNPY4TG"},"agent_actions":{"view_html":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4","download_json":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4.json","view_paper":"https://pith.science/paper/YWNPY4TG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.10912&json=true","fetch_graph":"https://pith.science/api/pith-number/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/graph.json","fetch_events":"https://pith.science/api/pith-number/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/action/storage_attestation","attest_author":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/action/author_attestation","sign_citation":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/action/citation_signature","submit_replication":"https://pith.science/pith/YWNPY4TGJF4R5ZKNU6ONK6K4Y4/action/replication_record"}},"created_at":"2026-07-05T09:20:26.871523+00:00","updated_at":"2026-07-05T09:20:26.871523+00:00"}