{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:M245ZHWACQGBIXAYNSZZ2VHP75","short_pith_number":"pith:M245ZHWA","schema_version":"1.0","canonical_sha256":"66b9dc9ec0140c145c186cb39d54efff5d923d3bad1e5f495067a9c0ee2c31c3","source":{"kind":"arxiv","id":"2411.02063","version":1},"attestation_state":"computed","paper":{"title":"Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bowen Zhou, Ermo Hua, Ganqu Cui, Kaiyan Zhang, Ning Ding, Xingtai Lv","submitted_at":"2024-11-04T13:06:17Z","abstract_excerpt":"Improving the effectiveness and efficiency of large language models (LLMs) simultaneously is a critical yet challenging research goal. In this paper, we find that low-rank pre-training, normally considered as efficient methods that will compromise performance, can be scalably effective when reduced parameters are precisely targeted. Specifically, applying the low-dimensional module only to the attention layer -- resolves this issue and enhances both effectiveness and efficiency. We refer to this structure as Low-dimensional Projected Attention (LPA) and provide an explanatory analysis. Through"},"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":"2411.02063","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-11-04T13:06:17Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"b6895c9d7038f6abcdbd0f185e91a2a6b944fe14de8c5750a1c603cd00132092","abstract_canon_sha256":"ed4e22549071b4c834fe8b3d2cdb1b6862f4f5f8276bdf475f493b90b3770407"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:30:40.770973Z","signature_b64":"wWFY6caHtiGLGfDwNju7lHN4LEnAX46DWxXyKgI9S3XE0T2MIRQHrLMDxCPeMiYEiRoRDwN1ZpzepqJ5wETwBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66b9dc9ec0140c145c186cb39d54efff5d923d3bad1e5f495067a9c0ee2c31c3","last_reissued_at":"2026-07-05T09:30:40.770432Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:30:40.770432Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scalable Efficient Training of Large Language Models with Low-dimensional Projected Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bowen Zhou, Ermo Hua, Ganqu Cui, Kaiyan Zhang, Ning Ding, Xingtai Lv","submitted_at":"2024-11-04T13:06:17Z","abstract_excerpt":"Improving the effectiveness and efficiency of large language models (LLMs) simultaneously is a critical yet challenging research goal. In this paper, we find that low-rank pre-training, normally considered as efficient methods that will compromise performance, can be scalably effective when reduced parameters are precisely targeted. Specifically, applying the low-dimensional module only to the attention layer -- resolves this issue and enhances both effectiveness and efficiency. We refer to this structure as Low-dimensional Projected Attention (LPA) and provide an explanatory analysis. Through"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.02063","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/2411.02063/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":"2411.02063","created_at":"2026-07-05T09:30:40.770491+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.02063v1","created_at":"2026-07-05T09:30:40.770491+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.02063","created_at":"2026-07-05T09:30:40.770491+00:00"},{"alias_kind":"pith_short_12","alias_value":"M245ZHWACQGB","created_at":"2026-07-05T09:30:40.770491+00:00"},{"alias_kind":"pith_short_16","alias_value":"M245ZHWACQGBIXAY","created_at":"2026-07-05T09:30:40.770491+00:00"},{"alias_kind":"pith_short_8","alias_value":"M245ZHWA","created_at":"2026-07-05T09:30:40.770491+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.19255","citing_title":"Multi-matrix Factorization Attention","ref_index":25,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75","json":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75.json","graph_json":"https://pith.science/api/pith-number/M245ZHWACQGBIXAYNSZZ2VHP75/graph.json","events_json":"https://pith.science/api/pith-number/M245ZHWACQGBIXAYNSZZ2VHP75/events.json","paper":"https://pith.science/paper/M245ZHWA"},"agent_actions":{"view_html":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75","download_json":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75.json","view_paper":"https://pith.science/paper/M245ZHWA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.02063&json=true","fetch_graph":"https://pith.science/api/pith-number/M245ZHWACQGBIXAYNSZZ2VHP75/graph.json","fetch_events":"https://pith.science/api/pith-number/M245ZHWACQGBIXAYNSZZ2VHP75/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75/action/timestamp_anchor","attest_storage":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75/action/storage_attestation","attest_author":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75/action/author_attestation","sign_citation":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75/action/citation_signature","submit_replication":"https://pith.science/pith/M245ZHWACQGBIXAYNSZZ2VHP75/action/replication_record"}},"created_at":"2026-07-05T09:30:40.770491+00:00","updated_at":"2026-07-05T09:30:40.770491+00:00"}