{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:2546PWFKC22VG4C2UZSY6E62KK","short_pith_number":"pith:2546PWFK","canonical_record":{"source":{"id":"2503.04992","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-06T21:42:35Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9190fe89c43ea51e42d3b03408006f12b09aa5099373308a05242281911607a9","abstract_canon_sha256":"0875522191e33b8547c51dd2330bc933e7369ee3ebd1aefdcd2d071721dc9cee"},"schema_version":"1.0"},"canonical_sha256":"d779e7d8aa16b553705aa6658f13da5280f88fd4ffddd45ef7612680fbad1577","source":{"kind":"arxiv","id":"2503.04992","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.04992","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"arxiv_version","alias_value":"2503.04992v4","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.04992","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_12","alias_value":"2546PWFKC22V","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_16","alias_value":"2546PWFKC22VG4C2","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_8","alias_value":"2546PWFK","created_at":"2026-07-05T11:13:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:2546PWFKC22VG4C2UZSY6E62KK","target":"record","payload":{"canonical_record":{"source":{"id":"2503.04992","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-06T21:42:35Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"9190fe89c43ea51e42d3b03408006f12b09aa5099373308a05242281911607a9","abstract_canon_sha256":"0875522191e33b8547c51dd2330bc933e7369ee3ebd1aefdcd2d071721dc9cee"},"schema_version":"1.0"},"canonical_sha256":"d779e7d8aa16b553705aa6658f13da5280f88fd4ffddd45ef7612680fbad1577","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:13:22.195743Z","signature_b64":"u7xv2YAKb/Eq02HSH4YG/TmSRawBQ/mrDT/LzkzGmIGVZ7Hz56AHP0LnLfPdxsR3ASCCHwUMicDBPPmVpNj7DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d779e7d8aa16b553705aa6658f13da5280f88fd4ffddd45ef7612680fbad1577","last_reissued_at":"2026-07-05T11:13:22.195273Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:13:22.195273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2503.04992","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KP+Urb7UkDNx7ZZU3DK5Ctq6N4opVq+vPWv2TfMctPmETG5M4WlTqOBqY7noPC5rgZW1TxzBFm21uY0IJvJdCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T12:40:15.164036Z"},"content_sha256":"8c81774e1e077ffd8699389910bafa274c95cca0502f280733d49575f2915176","schema_version":"1.0","event_id":"sha256:8c81774e1e077ffd8699389910bafa274c95cca0502f280733d49575f2915176"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:2546PWFKC22VG4C2UZSY6E62KK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Wanda++: Pruning Large Language Models via Regional Gradients","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Abhishek Kumar, Aram Galstyan, Athanasios Mouchtaris, Bhavana Ganesh, Goeric Huybrechts, Jack FitzGerald, Jonas M. K\\\"ubler, Kai Zhen, Markus M\\\"uller, Nathan Susanj, Rupak Vignesh Swaminathan, Sravan Babu Bodapati, Yifan Yang, Zheng Zhang","submitted_at":"2025-03-06T21:42:35Z","abstract_excerpt":"Large Language Models (LLMs) pruning seeks to remove unimportant weights for inference speedup with minimal accuracy impact. However, existing methods often suffer from accuracy degradation without full-model sparsity-aware fine-tuning. This paper presents Wanda++, a novel pruning framework that outperforms the state-of-the-art methods by utilizing decoder-block-level \\textbf{regional} gradients. Specifically, Wanda++ improves the pruning score with regional gradients for the first time and proposes an efficient regional optimization method to minimize pruning-induced output discrepancies betw"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.04992","kind":"arxiv","version":4},"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/2503.04992/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:13:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"v8WHzI+OrwVcOu1dz2Y3nTcgbRMEeIBK33SsWZwevGtBQQdqT60E0FQDYgJTITa8TIgSpUYMekWK/cgvsZC4Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T12:40:15.164419Z"},"content_sha256":"e9c38a2703e5c07f0295c77137b55a18abec0744b47736721cfdd301dac51bb6","schema_version":"1.0","event_id":"sha256:e9c38a2703e5c07f0295c77137b55a18abec0744b47736721cfdd301dac51bb6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/2546PWFKC22VG4C2UZSY6E62KK/bundle.json","state_url":"https://pith.science/pith/2546PWFKC22VG4C2UZSY6E62KK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/2546PWFKC22VG4C2UZSY6E62KK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-07T12:40:15Z","links":{"resolver":"https://pith.science/pith/2546PWFKC22VG4C2UZSY6E62KK","bundle":"https://pith.science/pith/2546PWFKC22VG4C2UZSY6E62KK/bundle.json","state":"https://pith.science/pith/2546PWFKC22VG4C2UZSY6E62KK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/2546PWFKC22VG4C2UZSY6E62KK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:2546PWFKC22VG4C2UZSY6E62KK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"0875522191e33b8547c51dd2330bc933e7369ee3ebd1aefdcd2d071721dc9cee","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-06T21:42:35Z","title_canon_sha256":"9190fe89c43ea51e42d3b03408006f12b09aa5099373308a05242281911607a9"},"schema_version":"1.0","source":{"id":"2503.04992","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2503.04992","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"arxiv_version","alias_value":"2503.04992v4","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.04992","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_12","alias_value":"2546PWFKC22V","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_16","alias_value":"2546PWFKC22VG4C2","created_at":"2026-07-05T11:13:22Z"},{"alias_kind":"pith_short_8","alias_value":"2546PWFK","created_at":"2026-07-05T11:13:22Z"}],"graph_snapshots":[{"event_id":"sha256:e9c38a2703e5c07f0295c77137b55a18abec0744b47736721cfdd301dac51bb6","target":"graph","created_at":"2026-07-05T11:13:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2503.04992/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) pruning seeks to remove unimportant weights for inference speedup with minimal accuracy impact. However, existing methods often suffer from accuracy degradation without full-model sparsity-aware fine-tuning. This paper presents Wanda++, a novel pruning framework that outperforms the state-of-the-art methods by utilizing decoder-block-level \\textbf{regional} gradients. Specifically, Wanda++ improves the pruning score with regional gradients for the first time and proposes an efficient regional optimization method to minimize pruning-induced output discrepancies betw","authors_text":"Abhishek Kumar, Aram Galstyan, Athanasios Mouchtaris, Bhavana Ganesh, Goeric Huybrechts, Jack FitzGerald, Jonas M. K\\\"ubler, Kai Zhen, Markus M\\\"uller, Nathan Susanj, Rupak Vignesh Swaminathan, Sravan Babu Bodapati, Yifan Yang, Zheng Zhang","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-06T21:42:35Z","title":"Wanda++: Pruning Large Language Models via Regional Gradients"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.04992","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:8c81774e1e077ffd8699389910bafa274c95cca0502f280733d49575f2915176","target":"record","created_at":"2026-07-05T11:13:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"0875522191e33b8547c51dd2330bc933e7369ee3ebd1aefdcd2d071721dc9cee","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-03-06T21:42:35Z","title_canon_sha256":"9190fe89c43ea51e42d3b03408006f12b09aa5099373308a05242281911607a9"},"schema_version":"1.0","source":{"id":"2503.04992","kind":"arxiv","version":4}},"canonical_sha256":"d779e7d8aa16b553705aa6658f13da5280f88fd4ffddd45ef7612680fbad1577","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"d779e7d8aa16b553705aa6658f13da5280f88fd4ffddd45ef7612680fbad1577","first_computed_at":"2026-07-05T11:13:22.195273Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:13:22.195273Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"u7xv2YAKb/Eq02HSH4YG/TmSRawBQ/mrDT/LzkzGmIGVZ7Hz56AHP0LnLfPdxsR3ASCCHwUMicDBPPmVpNj7DA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:13:22.195743Z","signed_message":"canonical_sha256_bytes"},"source_id":"2503.04992","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8c81774e1e077ffd8699389910bafa274c95cca0502f280733d49575f2915176","sha256:e9c38a2703e5c07f0295c77137b55a18abec0744b47736721cfdd301dac51bb6"],"state_sha256":"0d10cd71353d79ea31c9b7f2c99c0cd39e350784fa427a1a437acc57907b6109"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3u2X8WD2TffR4BIS+QFIWCjinbEDoNVAs7vHmOlHqZjVFul021eFOU5ZdJuRnz+t/9+bN1lZrCEQz/HVUyKkAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T12:40:15.166446Z","bundle_sha256":"d031a6a40a40661c734ab9a9d538106585fb5d2f71c2ab57246ecb36fe8ac177"}}