{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:EC3EUFUAFZB6GDUSQL6HR4UQ3V","short_pith_number":"pith:EC3EUFUA","schema_version":"1.0","canonical_sha256":"20b64a16802e43e30e9282fc78f290dd65678a4808e84f6ddc2ed0b256422aa9","source":{"kind":"arxiv","id":"2410.07383","version":1},"attestation_state":"computed","paper":{"title":"SparseGrad: A Selective Method for Efficient Fine-tuning of MLP Layers","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alexander Mikhalev, Alexander Panchenko, Anna Rudenko, Gleb Mezentsev, Ivan Oseledets, Viktoriia Chekalina","submitted_at":"2024-10-09T19:03:52Z","abstract_excerpt":"The performance of Transformer models has been enhanced by increasing the number of parameters and the length of the processed text. Consequently, fine-tuning the entire model becomes a memory-intensive process. High-performance methods for parameter-efficient fine-tuning (PEFT) typically work with Attention blocks and often overlook MLP blocks, which contain about half of the model parameters. We propose a new selective PEFT method, namely SparseGrad, that performs well on MLP blocks. We transfer layer gradients to a space where only about 1\\% of the layer's elements remain significant. By co"},"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":"2410.07383","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2024-10-09T19:03:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"518ca6e3587975b4a8e3d374eb5960e927901f852cf2ac4be736cd145b2b1c6a","abstract_canon_sha256":"bc75e3a909c3bf4be6d30e53a934cf0ab6e22feb9c72d4671af052a11d0badc4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:18:35.444638Z","signature_b64":"h+YGqRzIe6q+GVuLyftYAQ0lKbzvkG4wFG74z7koBXIeBL0WgQUbulwf1nw0DpoeeGPT1slShbjekSc5RZZqAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"20b64a16802e43e30e9282fc78f290dd65678a4808e84f6ddc2ed0b256422aa9","last_reissued_at":"2026-07-05T09:18:35.444219Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:18:35.444219Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SparseGrad: A Selective Method for Efficient Fine-tuning of MLP Layers","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Alexander Mikhalev, Alexander Panchenko, Anna Rudenko, Gleb Mezentsev, Ivan Oseledets, Viktoriia Chekalina","submitted_at":"2024-10-09T19:03:52Z","abstract_excerpt":"The performance of Transformer models has been enhanced by increasing the number of parameters and the length of the processed text. Consequently, fine-tuning the entire model becomes a memory-intensive process. High-performance methods for parameter-efficient fine-tuning (PEFT) typically work with Attention blocks and often overlook MLP blocks, which contain about half of the model parameters. We propose a new selective PEFT method, namely SparseGrad, that performs well on MLP blocks. We transfer layer gradients to a space where only about 1\\% of the layer's elements remain significant. By co"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.07383","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/2410.07383/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":"2410.07383","created_at":"2026-07-05T09:18:35.444281+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.07383v1","created_at":"2026-07-05T09:18:35.444281+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.07383","created_at":"2026-07-05T09:18:35.444281+00:00"},{"alias_kind":"pith_short_12","alias_value":"EC3EUFUAFZB6","created_at":"2026-07-05T09:18:35.444281+00:00"},{"alias_kind":"pith_short_16","alias_value":"EC3EUFUAFZB6GDUS","created_at":"2026-07-05T09:18:35.444281+00:00"},{"alias_kind":"pith_short_8","alias_value":"EC3EUFUA","created_at":"2026-07-05T09:18:35.444281+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/EC3EUFUAFZB6GDUSQL6HR4UQ3V","json":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V.json","graph_json":"https://pith.science/api/pith-number/EC3EUFUAFZB6GDUSQL6HR4UQ3V/graph.json","events_json":"https://pith.science/api/pith-number/EC3EUFUAFZB6GDUSQL6HR4UQ3V/events.json","paper":"https://pith.science/paper/EC3EUFUA"},"agent_actions":{"view_html":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V","download_json":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V.json","view_paper":"https://pith.science/paper/EC3EUFUA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.07383&json=true","fetch_graph":"https://pith.science/api/pith-number/EC3EUFUAFZB6GDUSQL6HR4UQ3V/graph.json","fetch_events":"https://pith.science/api/pith-number/EC3EUFUAFZB6GDUSQL6HR4UQ3V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V/action/storage_attestation","attest_author":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V/action/author_attestation","sign_citation":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V/action/citation_signature","submit_replication":"https://pith.science/pith/EC3EUFUAFZB6GDUSQL6HR4UQ3V/action/replication_record"}},"created_at":"2026-07-05T09:18:35.444281+00:00","updated_at":"2026-07-05T09:18:35.444281+00:00"}