{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:ZXVAOPAEQTKCC7ZF5NFE22D4O2","short_pith_number":"pith:ZXVAOPAE","schema_version":"1.0","canonical_sha256":"cdea073c0484d4217f25eb4a4d687c76bdf297fec41dff592d78a59a38c0c7c4","source":{"kind":"arxiv","id":"2305.11627","version":3},"attestation_state":"computed","paper":{"title":"LLM-Pruner: On the Structural Pruning of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Gongfan Fang, Xinchao Wang, Xinyin Ma","submitted_at":"2023-05-19T12:10:53Z","abstract_excerpt":"Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges in both the deployment, inference, and training stages. With LLM being a general-purpose task solver, we explore its compression in a task-agnostic manner, which aims to preserve the multi-task solving and language generation ability of the original LLM. One challenge to achieving this is the enormous size of the training corpus of LLM, which makes both data transfer and"},"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":"2305.11627","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-19T12:10:53Z","cross_cats_sorted":[],"title_canon_sha256":"9e368394d7d5a76b4c33e77157031aefcde228ea0a024329944b0d28dea66073","abstract_canon_sha256":"099b83db3fc694e4c7789dedbf7924189ac1ccd8ea1646c1a16bb3a8033087b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:55:03.632717Z","signature_b64":"daprK9OXBxHhL+oPW1B84sA1uP1Zrw88AoviLg8lj5iAF04Eg3kBBE8hJE7lDPDgIQPA/MIFASR2eFqf4flMCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cdea073c0484d4217f25eb4a4d687c76bdf297fec41dff592d78a59a38c0c7c4","last_reissued_at":"2026-07-05T06:55:03.632298Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:55:03.632298Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM-Pruner: On the Structural Pruning of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Gongfan Fang, Xinchao Wang, Xinyin Ma","submitted_at":"2023-05-19T12:10:53Z","abstract_excerpt":"Large language models (LLMs) have shown remarkable capabilities in language understanding and generation. However, such impressive capability typically comes with a substantial model size, which presents significant challenges in both the deployment, inference, and training stages. With LLM being a general-purpose task solver, we explore its compression in a task-agnostic manner, which aims to preserve the multi-task solving and language generation ability of the original LLM. One challenge to achieving this is the enormous size of the training corpus of LLM, which makes both data transfer and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.11627","kind":"arxiv","version":3},"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/2305.11627/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":"2305.11627","created_at":"2026-07-05T06:55:03.632360+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.11627v3","created_at":"2026-07-05T06:55:03.632360+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.11627","created_at":"2026-07-05T06:55:03.632360+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZXVAOPAEQTKC","created_at":"2026-07-05T06:55:03.632360+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZXVAOPAEQTKCC7ZF","created_at":"2026-07-05T06:55:03.632360+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZXVAOPAE","created_at":"2026-07-05T06:55:03.632360+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.06254","citing_title":"SecRL-Prune: Structured Reinforcement Learning-Based Pruning of CodeLLMs for Preserving Adversarial Code Mutation","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28438","citing_title":"When AI Reviews Its Own Code: Recursive Self-Training Collapse in Code LLMs","ref_index":72,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26538","citing_title":"CascadeFormer: Depth-Tapered Transformers Motivated by Gradient Fan-in Asymmetry","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2310.02277","citing_title":"Junk DNA Hypothesis: Pruning Small Pre-Trained Weights Irreversibly and Monotonically Impairs \"Difficult\" Downstream Tasks in LLMs","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2307.06435","citing_title":"A Comprehensive Overview of Large Language Models","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2401.05459","citing_title":"Personal LLM Agents: Insights and Survey about the Capability, Efficiency and Security","ref_index":253,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2","json":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2.json","graph_json":"https://pith.science/api/pith-number/ZXVAOPAEQTKCC7ZF5NFE22D4O2/graph.json","events_json":"https://pith.science/api/pith-number/ZXVAOPAEQTKCC7ZF5NFE22D4O2/events.json","paper":"https://pith.science/paper/ZXVAOPAE"},"agent_actions":{"view_html":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2","download_json":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2.json","view_paper":"https://pith.science/paper/ZXVAOPAE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.11627&json=true","fetch_graph":"https://pith.science/api/pith-number/ZXVAOPAEQTKCC7ZF5NFE22D4O2/graph.json","fetch_events":"https://pith.science/api/pith-number/ZXVAOPAEQTKCC7ZF5NFE22D4O2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2/action/storage_attestation","attest_author":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2/action/author_attestation","sign_citation":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2/action/citation_signature","submit_replication":"https://pith.science/pith/ZXVAOPAEQTKCC7ZF5NFE22D4O2/action/replication_record"}},"created_at":"2026-07-05T06:55:03.632360+00:00","updated_at":"2026-07-05T06:55:03.632360+00:00"}