{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZUMZPCVGS7JWTYE6OKMP46CGF7","short_pith_number":"pith:ZUMZPCVG","schema_version":"1.0","canonical_sha256":"cd19978aa697d369e09e7298fe78462ff2b669a00d80e87e652aab4557b20801","source":{"kind":"arxiv","id":"2411.12892","version":1},"attestation_state":"computed","paper":{"title":"Selective Attention: Enhancing Transformer through Principled Context Control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Amit Roy-Chowdhury, Jiasi Chen, Mingchen Li, Samet Oymak, Xiangyu Chang, Xuechen Zhang","submitted_at":"2024-11-19T22:17:18Z","abstract_excerpt":"The attention mechanism within the transformer architecture enables the model to weigh and combine tokens based on their relevance to the query. While self-attention has enjoyed major success, it notably treats all queries $q$ in the same way by applying the mapping $V^\\top\\text{softmax}(Kq)$, where $V,K$ are the value and key embeddings respectively. In this work, we argue that this uniform treatment hinders the ability to control contextual sparsity and relevance. As a solution, we introduce the $\\textit{Selective Self-Attention}$ (SSA) layer that augments the softmax nonlinearity with a pri"},"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":"2411.12892","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-11-19T22:17:18Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"387ce805b6757f181828abc2fadc880d739a8911ef3604ddd4f97d5031c13ebb","abstract_canon_sha256":"fb5cc12a6d6ba534d84e7cff5c6957b0b9482f6046f8cadd7e979e8bf52f816f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:38:01.905627Z","signature_b64":"T6emfMvhq9Y+PuOGkbERvpt282HvpK4ol+ZZhayrL/pkjTpxDfhvlD+fJLl8bcO/09Z5UCSkijqlXuaBKfbsCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd19978aa697d369e09e7298fe78462ff2b669a00d80e87e652aab4557b20801","last_reissued_at":"2026-07-05T09:38:01.905138Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:38:01.905138Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Selective Attention: Enhancing Transformer through Principled Context Control","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Amit Roy-Chowdhury, Jiasi Chen, Mingchen Li, Samet Oymak, Xiangyu Chang, Xuechen Zhang","submitted_at":"2024-11-19T22:17:18Z","abstract_excerpt":"The attention mechanism within the transformer architecture enables the model to weigh and combine tokens based on their relevance to the query. While self-attention has enjoyed major success, it notably treats all queries $q$ in the same way by applying the mapping $V^\\top\\text{softmax}(Kq)$, where $V,K$ are the value and key embeddings respectively. In this work, we argue that this uniform treatment hinders the ability to control contextual sparsity and relevance. As a solution, we introduce the $\\textit{Selective Self-Attention}$ (SSA) layer that augments the softmax nonlinearity with a pri"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.12892","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/2411.12892/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":"2411.12892","created_at":"2026-07-05T09:38:01.905190+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.12892v1","created_at":"2026-07-05T09:38:01.905190+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.12892","created_at":"2026-07-05T09:38:01.905190+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZUMZPCVGS7JW","created_at":"2026-07-05T09:38:01.905190+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZUMZPCVGS7JWTYE6","created_at":"2026-07-05T09:38:01.905190+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZUMZPCVG","created_at":"2026-07-05T09:38:01.905190+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/ZUMZPCVGS7JWTYE6OKMP46CGF7","json":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7.json","graph_json":"https://pith.science/api/pith-number/ZUMZPCVGS7JWTYE6OKMP46CGF7/graph.json","events_json":"https://pith.science/api/pith-number/ZUMZPCVGS7JWTYE6OKMP46CGF7/events.json","paper":"https://pith.science/paper/ZUMZPCVG"},"agent_actions":{"view_html":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7","download_json":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7.json","view_paper":"https://pith.science/paper/ZUMZPCVG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.12892&json=true","fetch_graph":"https://pith.science/api/pith-number/ZUMZPCVGS7JWTYE6OKMP46CGF7/graph.json","fetch_events":"https://pith.science/api/pith-number/ZUMZPCVGS7JWTYE6OKMP46CGF7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7/action/storage_attestation","attest_author":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7/action/author_attestation","sign_citation":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7/action/citation_signature","submit_replication":"https://pith.science/pith/ZUMZPCVGS7JWTYE6OKMP46CGF7/action/replication_record"}},"created_at":"2026-07-05T09:38:01.905190+00:00","updated_at":"2026-07-05T09:38:01.905190+00:00"}