{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WI56GZTLLTRVEEHEL4ECOM43NK","short_pith_number":"pith:WI56GZTL","schema_version":"1.0","canonical_sha256":"b23be3666b5ce35210e45f0827339b6a88ed6894749a688332b40d1964d3a58a","source":{"kind":"arxiv","id":"2509.14230","version":2},"attestation_state":"computed","paper":{"title":"NIRVANA: Structured Pruning Reimagined for Large Language Model Compression","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jingrui He, Mengting Ai, Sirui Chen, Tianxin Wei","submitted_at":"2025-09-17T17:59:00Z","abstract_excerpt":"While structured pruning presents a highly effective pathway for accelerating Large Language Model (LLM) inference, existing methods frequently suffer from significant performance degradation and demand computationally retraining to recover capabilities. To overcome these barriers, we present NIRVANA, a novel, hardware-aware structured pruning framework designed to preserve both zero-shot performance and the optimization landscape for downstream fine-tuning. Departing from traditional loss-based heuristics, our approach evaluates structural importance through a first-order function-space salie"},"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":"2509.14230","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-17T17:59:00Z","cross_cats_sorted":[],"title_canon_sha256":"72c797d3cd16c348d5407e88aa035f33b4c6110766ba72b75593029fb8fcc39b","abstract_canon_sha256":"9ee92e8621f44c685751c871b23863e21a7d9c856494a5b9584d19f143f70eaf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T01:20:37.145274Z","signature_b64":"V/L1zfEwgZjAEjYoe5FwwF9SSPO4LYHeLvuue+jWp5Cbd6TQAdzv8wefUBjk7QF1wB5yti2ixtRpty8JE+doCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b23be3666b5ce35210e45f0827339b6a88ed6894749a688332b40d1964d3a58a","last_reissued_at":"2026-07-21T01:20:37.144251Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T01:20:37.144251Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"NIRVANA: Structured Pruning Reimagined for Large Language Model Compression","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jingrui He, Mengting Ai, Sirui Chen, Tianxin Wei","submitted_at":"2025-09-17T17:59:00Z","abstract_excerpt":"While structured pruning presents a highly effective pathway for accelerating Large Language Model (LLM) inference, existing methods frequently suffer from significant performance degradation and demand computationally retraining to recover capabilities. To overcome these barriers, we present NIRVANA, a novel, hardware-aware structured pruning framework designed to preserve both zero-shot performance and the optimization landscape for downstream fine-tuning. Departing from traditional loss-based heuristics, our approach evaluates structural importance through a first-order function-space salie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.14230","kind":"arxiv","version":2},"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/2509.14230/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":"2509.14230","created_at":"2026-07-21T01:20:37.144706+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.14230v2","created_at":"2026-07-21T01:20:37.144706+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.14230","created_at":"2026-07-21T01:20:37.144706+00:00"},{"alias_kind":"pith_short_12","alias_value":"WI56GZTLLTRV","created_at":"2026-07-21T01:20:37.144706+00:00"},{"alias_kind":"pith_short_16","alias_value":"WI56GZTLLTRVEEHE","created_at":"2026-07-21T01:20:37.144706+00:00"},{"alias_kind":"pith_short_8","alias_value":"WI56GZTL","created_at":"2026-07-21T01:20:37.144706+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2512.22671","citing_title":"Fragile Knowledge, Robust Instruction-Following: The Width Pruning Dichotomy in Llama-3.2","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK","json":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK.json","graph_json":"https://pith.science/api/pith-number/WI56GZTLLTRVEEHEL4ECOM43NK/graph.json","events_json":"https://pith.science/api/pith-number/WI56GZTLLTRVEEHEL4ECOM43NK/events.json","paper":"https://pith.science/paper/WI56GZTL"},"agent_actions":{"view_html":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK","download_json":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK.json","view_paper":"https://pith.science/paper/WI56GZTL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.14230&json=true","fetch_graph":"https://pith.science/api/pith-number/WI56GZTLLTRVEEHEL4ECOM43NK/graph.json","fetch_events":"https://pith.science/api/pith-number/WI56GZTLLTRVEEHEL4ECOM43NK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK/action/storage_attestation","attest_author":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK/action/author_attestation","sign_citation":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK/action/citation_signature","submit_replication":"https://pith.science/pith/WI56GZTLLTRVEEHEL4ECOM43NK/action/replication_record"}},"created_at":"2026-07-21T01:20:37.144706+00:00","updated_at":"2026-07-21T01:20:37.144706+00:00"}