{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EM76QJS5PDVAOTVOV4JVDDNJWZ","short_pith_number":"pith:EM76QJS5","schema_version":"1.0","canonical_sha256":"233fe8265d78ea074eaeaf13518da9b65156b630c8d75965fc4d14a20d111379","source":{"kind":"arxiv","id":"2505.14840","version":1},"attestation_state":"computed","paper":{"title":"Subquadratic Algorithms and Hardness for Attention with Any Temperature","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CC"],"primary_cat":"cs.LG","authors_text":"Barna Saha, Boyang Huang, Christopher Ye, Shreya Gupta, Yinzhan Xu","submitted_at":"2025-05-20T19:12:43Z","abstract_excerpt":"Despite the popularity of the Transformer architecture, the standard algorithm for computing Attention suffers from quadratic time complexity in context length $n$. Alman and Song [NeurIPS 2023] showed that when the head dimension $d = \\Theta(\\log n)$, subquadratic Attention is possible if and only if the inputs have small entries bounded by $B = o(\\sqrt{\\log n})$ in absolute values, under the Strong Exponential Time Hypothesis ($\\mathsf{SETH}$). Equivalently, subquadratic Attention is possible if and only if the softmax is applied with high temperature for $d=\\Theta(\\log n)$. Running times of"},"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":"2505.14840","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-20T19:12:43Z","cross_cats_sorted":["cs.CC"],"title_canon_sha256":"de83df055703d3ea62832739a24cf388b578a6082587d160b4de55176fd11f56","abstract_canon_sha256":"1f59fef1d0fc3228578b9d665199bac96dec0cfa2a13b9e6265ae9852cbcd133"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:06:36.194480Z","signature_b64":"J19gG4P5D6pZzwmYDJisANWgWYoVqxg4F8VewpDv9kywIEPgEVdmegBBS86eHjJidEglNE7VXWzTzU5TKyvmDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"233fe8265d78ea074eaeaf13518da9b65156b630c8d75965fc4d14a20d111379","last_reissued_at":"2026-07-05T11:06:36.194026Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:06:36.194026Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Subquadratic Algorithms and Hardness for Attention with Any Temperature","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CC"],"primary_cat":"cs.LG","authors_text":"Barna Saha, Boyang Huang, Christopher Ye, Shreya Gupta, Yinzhan Xu","submitted_at":"2025-05-20T19:12:43Z","abstract_excerpt":"Despite the popularity of the Transformer architecture, the standard algorithm for computing Attention suffers from quadratic time complexity in context length $n$. Alman and Song [NeurIPS 2023] showed that when the head dimension $d = \\Theta(\\log n)$, subquadratic Attention is possible if and only if the inputs have small entries bounded by $B = o(\\sqrt{\\log n})$ in absolute values, under the Strong Exponential Time Hypothesis ($\\mathsf{SETH}$). Equivalently, subquadratic Attention is possible if and only if the softmax is applied with high temperature for $d=\\Theta(\\log n)$. Running times of"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.14840","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/2505.14840/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":"2505.14840","created_at":"2026-07-05T11:06:36.194082+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.14840v1","created_at":"2026-07-05T11:06:36.194082+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.14840","created_at":"2026-07-05T11:06:36.194082+00:00"},{"alias_kind":"pith_short_12","alias_value":"EM76QJS5PDVA","created_at":"2026-07-05T11:06:36.194082+00:00"},{"alias_kind":"pith_short_16","alias_value":"EM76QJS5PDVAOTVO","created_at":"2026-07-05T11:06:36.194082+00:00"},{"alias_kind":"pith_short_8","alias_value":"EM76QJS5","created_at":"2026-07-05T11:06:36.194082+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.19531","citing_title":"Minimalist Softmax Attention Provably Learns Constrained Boolean Functions","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ","json":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ.json","graph_json":"https://pith.science/api/pith-number/EM76QJS5PDVAOTVOV4JVDDNJWZ/graph.json","events_json":"https://pith.science/api/pith-number/EM76QJS5PDVAOTVOV4JVDDNJWZ/events.json","paper":"https://pith.science/paper/EM76QJS5"},"agent_actions":{"view_html":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ","download_json":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ.json","view_paper":"https://pith.science/paper/EM76QJS5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.14840&json=true","fetch_graph":"https://pith.science/api/pith-number/EM76QJS5PDVAOTVOV4JVDDNJWZ/graph.json","fetch_events":"https://pith.science/api/pith-number/EM76QJS5PDVAOTVOV4JVDDNJWZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ/action/storage_attestation","attest_author":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ/action/author_attestation","sign_citation":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ/action/citation_signature","submit_replication":"https://pith.science/pith/EM76QJS5PDVAOTVOV4JVDDNJWZ/action/replication_record"}},"created_at":"2026-07-05T11:06:36.194082+00:00","updated_at":"2026-07-05T11:06:36.194082+00:00"}