{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:H2A5K2HQYE6Y2SP2FMMCZNLOST","short_pith_number":"pith:H2A5K2HQ","schema_version":"1.0","canonical_sha256":"3e81d568f0c13d8d49fa2b182cb56e94dda8070f09b6b5dbad2281fdccfe76fb","source":{"kind":"arxiv","id":"2501.14577","version":3},"attestation_state":"computed","paper":{"title":"ZETA: Leveraging Z-order Curves for Efficient Top-k Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boxing Chen, Boyu Wang, Charles Ling, Gezheng Xu, Jerry Huang, Peng Lu, Qiuhao Zeng","submitted_at":"2025-01-24T15:33:05Z","abstract_excerpt":"Over recent years, the Transformer has become a fundamental building block for sequence modeling architectures. Yet at its core is the use of self-attention, whose memory and computational cost grow quadratically with the sequence length $N$, rendering it prohibitively expensive for long sequences. A promising approach is top-$k$ attention, which selects only the $k$ most relevant tokens and achieves performance comparable to vanilla self-attention while significantly reducing space and computational demands. However, causal masks require the current query token to only attend to past tokens, "},"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":"2501.14577","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-24T15:33:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"95735502b4f92e0a6c6d4272c658f8c31179d271afbcf5a790a385a8142a294b","abstract_canon_sha256":"4162e71be1639bcf97673d3d50737f22c34f56678dd0b6e03e03e2893ad6c3f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:42:14.713590Z","signature_b64":"AjrChLhbGSfUNHu390010eQw7W1oxuCFOuY6LccxyFSrrI8GD9HTPBpn86vxnbbgPNzchJmUyU/dqTW2eBmBAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3e81d568f0c13d8d49fa2b182cb56e94dda8070f09b6b5dbad2281fdccfe76fb","last_reissued_at":"2026-07-05T10:42:14.713103Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:42:14.713103Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ZETA: Leveraging Z-order Curves for Efficient Top-k Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Boxing Chen, Boyu Wang, Charles Ling, Gezheng Xu, Jerry Huang, Peng Lu, Qiuhao Zeng","submitted_at":"2025-01-24T15:33:05Z","abstract_excerpt":"Over recent years, the Transformer has become a fundamental building block for sequence modeling architectures. Yet at its core is the use of self-attention, whose memory and computational cost grow quadratically with the sequence length $N$, rendering it prohibitively expensive for long sequences. A promising approach is top-$k$ attention, which selects only the $k$ most relevant tokens and achieves performance comparable to vanilla self-attention while significantly reducing space and computational demands. However, causal masks require the current query token to only attend to past tokens, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.14577","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/2501.14577/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":"2501.14577","created_at":"2026-07-05T10:42:14.713159+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.14577v3","created_at":"2026-07-05T10:42:14.713159+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.14577","created_at":"2026-07-05T10:42:14.713159+00:00"},{"alias_kind":"pith_short_12","alias_value":"H2A5K2HQYE6Y","created_at":"2026-07-05T10:42:14.713159+00:00"},{"alias_kind":"pith_short_16","alias_value":"H2A5K2HQYE6Y2SP2","created_at":"2026-07-05T10:42:14.713159+00:00"},{"alias_kind":"pith_short_8","alias_value":"H2A5K2HQ","created_at":"2026-07-05T10:42:14.713159+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.19472","citing_title":"Balancing Computation Load and Representation Expressivity in Parallel Hybrid Neural Networks","ref_index":15,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST","json":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST.json","graph_json":"https://pith.science/api/pith-number/H2A5K2HQYE6Y2SP2FMMCZNLOST/graph.json","events_json":"https://pith.science/api/pith-number/H2A5K2HQYE6Y2SP2FMMCZNLOST/events.json","paper":"https://pith.science/paper/H2A5K2HQ"},"agent_actions":{"view_html":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST","download_json":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST.json","view_paper":"https://pith.science/paper/H2A5K2HQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.14577&json=true","fetch_graph":"https://pith.science/api/pith-number/H2A5K2HQYE6Y2SP2FMMCZNLOST/graph.json","fetch_events":"https://pith.science/api/pith-number/H2A5K2HQYE6Y2SP2FMMCZNLOST/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST/action/storage_attestation","attest_author":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST/action/author_attestation","sign_citation":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST/action/citation_signature","submit_replication":"https://pith.science/pith/H2A5K2HQYE6Y2SP2FMMCZNLOST/action/replication_record"}},"created_at":"2026-07-05T10:42:14.713159+00:00","updated_at":"2026-07-05T10:42:14.713159+00:00"}