{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NILQ7NL2BJPERU2XWSN2FVWFU3","short_pith_number":"pith:NILQ7NL2","schema_version":"1.0","canonical_sha256":"6a170fb57a0a5e48d357b49ba2d6c5a6c4ab0413fcae9077ac5511cd5726d639","source":{"kind":"arxiv","id":"2605.28384","version":1},"attestation_state":"computed","paper":{"title":"Meta-Attention: Bayesian Per-Token Routing for Efficient Transformer Inference","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alan Ferrari","submitted_at":"2026-05-27T12:21:28Z","abstract_excerpt":"Standard transformer architectures apply a single attention mechanism uniformly across all tokens and sequence positions, irrespective of local context or computational budget. We propose Meta-Attention, a framework that dynamically routes each token to the most appropriate attention strategy -- full softmax attention, linear (kernel) attention, or sliding-window local attention -- via a Bayesian Meta-Controller. Unlike prior routing approaches that use deterministic or prior-free learned routing, the Meta-Controller treats per-token mechanism selection as posterior inference under a compute-a"},"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":"2605.28384","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T12:21:28Z","cross_cats_sorted":[],"title_canon_sha256":"2c194302fa2c6074f663a031418fba8bca20b3e22c2a9cd51abf07e23ce2c98f","abstract_canon_sha256":"6ca8e04244276a47a0f6c83becedda37d36f227b7d5047c1d7e39890be528142"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:16.679685Z","signature_b64":"RKnCrKwpUpCKN5qbx4tPeyaww0hk2zx7kwUCZM0KQG2+1zpXLYr6jKMFoqrbSEfeRoamnminAhi456ThK4omCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6a170fb57a0a5e48d357b49ba2d6c5a6c4ab0413fcae9077ac5511cd5726d639","last_reissued_at":"2026-05-28T01:05:16.679290Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:16.679290Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Meta-Attention: Bayesian Per-Token Routing for Efficient Transformer Inference","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alan Ferrari","submitted_at":"2026-05-27T12:21:28Z","abstract_excerpt":"Standard transformer architectures apply a single attention mechanism uniformly across all tokens and sequence positions, irrespective of local context or computational budget. We propose Meta-Attention, a framework that dynamically routes each token to the most appropriate attention strategy -- full softmax attention, linear (kernel) attention, or sliding-window local attention -- via a Bayesian Meta-Controller. Unlike prior routing approaches that use deterministic or prior-free learned routing, the Meta-Controller treats per-token mechanism selection as posterior inference under a compute-a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28384","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/2605.28384/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":"2605.28384","created_at":"2026-05-28T01:05:16.679355+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28384v1","created_at":"2026-05-28T01:05:16.679355+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28384","created_at":"2026-05-28T01:05:16.679355+00:00"},{"alias_kind":"pith_short_12","alias_value":"NILQ7NL2BJPE","created_at":"2026-05-28T01:05:16.679355+00:00"},{"alias_kind":"pith_short_16","alias_value":"NILQ7NL2BJPERU2X","created_at":"2026-05-28T01:05:16.679355+00:00"},{"alias_kind":"pith_short_8","alias_value":"NILQ7NL2","created_at":"2026-05-28T01:05:16.679355+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/NILQ7NL2BJPERU2XWSN2FVWFU3","json":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3.json","graph_json":"https://pith.science/api/pith-number/NILQ7NL2BJPERU2XWSN2FVWFU3/graph.json","events_json":"https://pith.science/api/pith-number/NILQ7NL2BJPERU2XWSN2FVWFU3/events.json","paper":"https://pith.science/paper/NILQ7NL2"},"agent_actions":{"view_html":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3","download_json":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3.json","view_paper":"https://pith.science/paper/NILQ7NL2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28384&json=true","fetch_graph":"https://pith.science/api/pith-number/NILQ7NL2BJPERU2XWSN2FVWFU3/graph.json","fetch_events":"https://pith.science/api/pith-number/NILQ7NL2BJPERU2XWSN2FVWFU3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3/action/storage_attestation","attest_author":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3/action/author_attestation","sign_citation":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3/action/citation_signature","submit_replication":"https://pith.science/pith/NILQ7NL2BJPERU2XWSN2FVWFU3/action/replication_record"}},"created_at":"2026-05-28T01:05:16.679355+00:00","updated_at":"2026-05-28T01:05:16.679355+00:00"}