{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3W4TP4TDOK7R2KAVRJ6PBW7JV5","short_pith_number":"pith:3W4TP4TD","schema_version":"1.0","canonical_sha256":"ddb937f26372bf1d28158a7cf0dbe9af465b2aa99eb16f4e99523928076f5dd0","source":{"kind":"arxiv","id":"2510.21270","version":2},"attestation_state":"computed","paper":{"title":"Sparser Block-Sparse Attention via Token Permutation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.CL","authors_text":"Chenkun Tan, Dong Zhang, Fangxu Liu, Kai Song, Pengyu Wang, Shaojun Zhou, Shiguo Lian, Xinghao Wang, Xipeng Qiu, Zhaoxiang Liu","submitted_at":"2025-10-24T09:11:50Z","abstract_excerpt":"Scaling the context length of large language models (LLMs) offers significant benefits but is computationally expensive. This expense stems primarily from the self-attention mechanism, whose $O(N^2)$ complexity with respect to sequence length presents a major bottleneck for both memory and latency. Fortunately, the attention matrix is often sparse, particularly for long sequences, suggesting an opportunity for optimization. Block-sparse attention has emerged as a promising solution that partitions sequences into blocks and skips computation for a subset of these blocks. However, the effectiven"},"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":"2510.21270","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-10-24T09:11:50Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"c558737a8844134915601e51d1a568059e4aee5264aaf1425d8d0ce846bb02f4","abstract_canon_sha256":"f384821d7dfb4797dddd8226e955b5ffdd40de8cefdd69307c3351018a4b3a88"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:08.848155Z","signature_b64":"3WmDgEWOcXooz2oB1T3KgXONtC5BCG3VsTWazaX2iEddmB9AouCf56xNrDThQHJAuK7n0C8VBT55yX04Q75ZAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ddb937f26372bf1d28158a7cf0dbe9af465b2aa99eb16f4e99523928076f5dd0","last_reissued_at":"2026-05-25T02:01:08.847523Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:08.847523Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Sparser Block-Sparse Attention via Token Permutation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.CL","authors_text":"Chenkun Tan, Dong Zhang, Fangxu Liu, Kai Song, Pengyu Wang, Shaojun Zhou, Shiguo Lian, Xinghao Wang, Xipeng Qiu, Zhaoxiang Liu","submitted_at":"2025-10-24T09:11:50Z","abstract_excerpt":"Scaling the context length of large language models (LLMs) offers significant benefits but is computationally expensive. This expense stems primarily from the self-attention mechanism, whose $O(N^2)$ complexity with respect to sequence length presents a major bottleneck for both memory and latency. Fortunately, the attention matrix is often sparse, particularly for long sequences, suggesting an opportunity for optimization. Block-sparse attention has emerged as a promising solution that partitions sequences into blocks and skips computation for a subset of these blocks. However, the effectiven"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.21270","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/2510.21270/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":"2510.21270","created_at":"2026-05-25T02:01:08.847615+00:00"},{"alias_kind":"arxiv_version","alias_value":"2510.21270v2","created_at":"2026-05-25T02:01:08.847615+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.21270","created_at":"2026-05-25T02:01:08.847615+00:00"},{"alias_kind":"pith_short_12","alias_value":"3W4TP4TDOK7R","created_at":"2026-05-25T02:01:08.847615+00:00"},{"alias_kind":"pith_short_16","alias_value":"3W4TP4TDOK7R2KAV","created_at":"2026-05-25T02:01:08.847615+00:00"},{"alias_kind":"pith_short_8","alias_value":"3W4TP4TD","created_at":"2026-05-25T02:01:08.847615+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/3W4TP4TDOK7R2KAVRJ6PBW7JV5","json":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5.json","graph_json":"https://pith.science/api/pith-number/3W4TP4TDOK7R2KAVRJ6PBW7JV5/graph.json","events_json":"https://pith.science/api/pith-number/3W4TP4TDOK7R2KAVRJ6PBW7JV5/events.json","paper":"https://pith.science/paper/3W4TP4TD"},"agent_actions":{"view_html":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5","download_json":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5.json","view_paper":"https://pith.science/paper/3W4TP4TD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2510.21270&json=true","fetch_graph":"https://pith.science/api/pith-number/3W4TP4TDOK7R2KAVRJ6PBW7JV5/graph.json","fetch_events":"https://pith.science/api/pith-number/3W4TP4TDOK7R2KAVRJ6PBW7JV5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5/action/storage_attestation","attest_author":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5/action/author_attestation","sign_citation":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5/action/citation_signature","submit_replication":"https://pith.science/pith/3W4TP4TDOK7R2KAVRJ6PBW7JV5/action/replication_record"}},"created_at":"2026-05-25T02:01:08.847615+00:00","updated_at":"2026-05-25T02:01:08.847615+00:00"}