{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HVH2OUCUO7I4JJMAPOQL2DAPYS","short_pith_number":"pith:HVH2OUCU","schema_version":"1.0","canonical_sha256":"3d4fa7505477d1c4a5807ba0bd0c0fc49da03f178a2ea3949600c662854a238e","source":{"kind":"arxiv","id":"2502.18394","version":7},"attestation_state":"computed","paper":{"title":"SPECTRE: An FFT-Based Efficient Drop-In Replacement to Self-Attention for Long Contexts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jacob Fein-Ashley, Neelesh Gupta, Rajgopal Kannan, Viktor Prasanna","submitted_at":"2025-02-25T17:43:43Z","abstract_excerpt":"Long-context transformers face significant efficiency challenges due to the quadratic cost of self-attention. However, many modern applications-from multi-turn dialogue to high-resolution vision-require contexts spanning tens of thousands of tokens. We introduce SPECTRE, a method that replaces each attention head with a fast real FFT, a content-adaptive spectral gate, and an inverse FFT, reducing per-layer complexity from $\\mathcal{O}(L^{2})$ to $O(L\\log L)$ while preserving the surrounding architecture. We extend this efficiency to autoregressive generation through our Prefix-FFT cache and en"},"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":"2502.18394","kind":"arxiv","version":7},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-25T17:43:43Z","cross_cats_sorted":[],"title_canon_sha256":"e45b7251bb78e54d07317759033231a2c683b74990be1099637b48850bc43506","abstract_canon_sha256":"eeb0fe9210714d2e22200a0f7b7e8e0996eb874c58a6ca0ab48eca76bd8bb7e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:04:46.355386Z","signature_b64":"KX+zEB48xuooT3tVDyfNPOeZKL6kdBwdq+BiX1djBSF6pa/lncKnsrZPbRuIIwHzrihQGVT3alkvdq2PR7pXDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3d4fa7505477d1c4a5807ba0bd0c0fc49da03f178a2ea3949600c662854a238e","last_reissued_at":"2026-07-05T11:04:46.354822Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:04:46.354822Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SPECTRE: An FFT-Based Efficient Drop-In Replacement to Self-Attention for Long Contexts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jacob Fein-Ashley, Neelesh Gupta, Rajgopal Kannan, Viktor Prasanna","submitted_at":"2025-02-25T17:43:43Z","abstract_excerpt":"Long-context transformers face significant efficiency challenges due to the quadratic cost of self-attention. However, many modern applications-from multi-turn dialogue to high-resolution vision-require contexts spanning tens of thousands of tokens. We introduce SPECTRE, a method that replaces each attention head with a fast real FFT, a content-adaptive spectral gate, and an inverse FFT, reducing per-layer complexity from $\\mathcal{O}(L^{2})$ to $O(L\\log L)$ while preserving the surrounding architecture. We extend this efficiency to autoregressive generation through our Prefix-FFT cache and en"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18394","kind":"arxiv","version":7},"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/2502.18394/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":"2502.18394","created_at":"2026-07-05T11:04:46.354882+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.18394v7","created_at":"2026-07-05T11:04:46.354882+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18394","created_at":"2026-07-05T11:04:46.354882+00:00"},{"alias_kind":"pith_short_12","alias_value":"HVH2OUCUO7I4","created_at":"2026-07-05T11:04:46.354882+00:00"},{"alias_kind":"pith_short_16","alias_value":"HVH2OUCUO7I4JJMA","created_at":"2026-07-05T11:04:46.354882+00:00"},{"alias_kind":"pith_short_8","alias_value":"HVH2OUCU","created_at":"2026-07-05T11:04:46.354882+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.14727","citing_title":"CHASM: Cross-frequency Harmonized Axis-Separable Mixing for Spectral Token Operators","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00292","citing_title":"Caracal: Causal Architecture via Spectral Mixing","ref_index":87,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS","json":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS.json","graph_json":"https://pith.science/api/pith-number/HVH2OUCUO7I4JJMAPOQL2DAPYS/graph.json","events_json":"https://pith.science/api/pith-number/HVH2OUCUO7I4JJMAPOQL2DAPYS/events.json","paper":"https://pith.science/paper/HVH2OUCU"},"agent_actions":{"view_html":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS","download_json":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS.json","view_paper":"https://pith.science/paper/HVH2OUCU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.18394&json=true","fetch_graph":"https://pith.science/api/pith-number/HVH2OUCUO7I4JJMAPOQL2DAPYS/graph.json","fetch_events":"https://pith.science/api/pith-number/HVH2OUCUO7I4JJMAPOQL2DAPYS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS/action/storage_attestation","attest_author":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS/action/author_attestation","sign_citation":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS/action/citation_signature","submit_replication":"https://pith.science/pith/HVH2OUCUO7I4JJMAPOQL2DAPYS/action/replication_record"}},"created_at":"2026-07-05T11:04:46.354882+00:00","updated_at":"2026-07-05T11:04:46.354882+00:00"}