{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2AVEJ224XGJVZ6B7SRXGKERUT2","short_pith_number":"pith:2AVEJ224","schema_version":"1.0","canonical_sha256":"d02a44eb5cb9935cf83f946e6512349e9446f936618d321b0e8449d212cb3b98","source":{"kind":"arxiv","id":"2605.20868","version":1},"attestation_state":"computed","paper":{"title":"Runtime-Certified Bounded-Error Quantized Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SY","eess.SY"],"primary_cat":"cs.LG","authors_text":"Dean Calver","submitted_at":"2026-05-20T08:04:40Z","abstract_excerpt":"KV cache quantization reduces the memory cost of long-context LLM inference, but introduces approximation error that is typically validated only empirically. Existing systems rely on average-case robustness, with no mechanism to detect or recover from failures at runtime. We present a tiered KV cache architecture that enables runtime-certified attention: INT8 keys and INT4 values are stored in GPU memory, while FP16 originals are retained in system RAM for deterministic fallback. A two-term error decomposition yields per-head, per-step bounds on (i) attention distribution distortion from key q"},"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":"2605.20868","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T08:04:40Z","cross_cats_sorted":["cs.AI","cs.SY","eess.SY"],"title_canon_sha256":"0dcf71d246e06d3bfb10f27bea61d882486410cdef4b167215729d76b82c5923","abstract_canon_sha256":"fbf043af13a8b67ce225f5911142002022a21a052e22da9a45ae1174f3ae3fee"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T01:05:25.363735Z","signature_b64":"lLuaWoFFhSg48rBIC67jfODee5I0NIu6a3miYqEBlRtgFLG7Es93Jue/w3Fni9IUU18cQwoM8WTcWmTPNCtNBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d02a44eb5cb9935cf83f946e6512349e9446f936618d321b0e8449d212cb3b98","last_reissued_at":"2026-05-21T01:05:25.363197Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T01:05:25.363197Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Runtime-Certified Bounded-Error Quantized Attention","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SY","eess.SY"],"primary_cat":"cs.LG","authors_text":"Dean Calver","submitted_at":"2026-05-20T08:04:40Z","abstract_excerpt":"KV cache quantization reduces the memory cost of long-context LLM inference, but introduces approximation error that is typically validated only empirically. Existing systems rely on average-case robustness, with no mechanism to detect or recover from failures at runtime. We present a tiered KV cache architecture that enables runtime-certified attention: INT8 keys and INT4 values are stored in GPU memory, while FP16 originals are retained in system RAM for deterministic fallback. A two-term error decomposition yields per-head, per-step bounds on (i) attention distribution distortion from key q"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.20868","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/2605.20868/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":"2605.20868","created_at":"2026-05-21T01:05:25.363273+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.20868v1","created_at":"2026-05-21T01:05:25.363273+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.20868","created_at":"2026-05-21T01:05:25.363273+00:00"},{"alias_kind":"pith_short_12","alias_value":"2AVEJ224XGJV","created_at":"2026-05-21T01:05:25.363273+00:00"},{"alias_kind":"pith_short_16","alias_value":"2AVEJ224XGJVZ6B7","created_at":"2026-05-21T01:05:25.363273+00:00"},{"alias_kind":"pith_short_8","alias_value":"2AVEJ224","created_at":"2026-05-21T01:05:25.363273+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/2AVEJ224XGJVZ6B7SRXGKERUT2","json":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2.json","graph_json":"https://pith.science/api/pith-number/2AVEJ224XGJVZ6B7SRXGKERUT2/graph.json","events_json":"https://pith.science/api/pith-number/2AVEJ224XGJVZ6B7SRXGKERUT2/events.json","paper":"https://pith.science/paper/2AVEJ224"},"agent_actions":{"view_html":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2","download_json":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2.json","view_paper":"https://pith.science/paper/2AVEJ224","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.20868&json=true","fetch_graph":"https://pith.science/api/pith-number/2AVEJ224XGJVZ6B7SRXGKERUT2/graph.json","fetch_events":"https://pith.science/api/pith-number/2AVEJ224XGJVZ6B7SRXGKERUT2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2/action/storage_attestation","attest_author":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2/action/author_attestation","sign_citation":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2/action/citation_signature","submit_replication":"https://pith.science/pith/2AVEJ224XGJVZ6B7SRXGKERUT2/action/replication_record"}},"created_at":"2026-05-21T01:05:25.363273+00:00","updated_at":"2026-05-21T01:05:25.363273+00:00"}