{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NB4RRKL5TSVS3CHAEPAWK3BKMA","short_pith_number":"pith:NB4RRKL5","schema_version":"1.0","canonical_sha256":"687918a97d9cab2d88e023c1656c2a601377ba88df5b7436dc6711b221a122ff","source":{"kind":"arxiv","id":"2502.13542","version":1},"attestation_state":"computed","paper":{"title":"Activation-aware Probe-Query: Effective Key-Value Retrieval for Long-Context LLMs Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Cheng Deng, Haoyang Li, Jiachuan Wang, Jiaqi Tang, Jun Wang, Lei Chen, Qingfa Xiao, Shuangyin Li, Yongqi Zhang","submitted_at":"2025-02-19T08:50:44Z","abstract_excerpt":"Recent advances in large language models (LLMs) have showcased exceptional performance in long-context tasks, while facing significant inference efficiency challenges with limited GPU memory. Existing solutions first proposed the sliding-window approach to accumulate a set of historical \\textbf{key-value} (KV) pairs for reuse, then further improvements selectively retain its subsets at each step. However, due to the sparse attention distribution across a long context, it is hard to identify and recall relevant KV pairs, as the attention is distracted by massive candidate pairs. Additionally, w"},"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":"2502.13542","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T08:50:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"403baff129cbe590eb16a6762c643bbc4e161879d9beab4dd108dce6e0b960ad","abstract_canon_sha256":"114ad8fa2a3d615bd7a624e6f8bd577d7b29df2063551a21f6c92a55739ff11b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:48.933267Z","signature_b64":"joSzdUg7jU1yriMzQVFxrDzkKQYTzY5tyZGGUVO6WH7i36oT1lHC1SCom44+48gryOE+0kS0cWNx8U9tocNICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"687918a97d9cab2d88e023c1656c2a601377ba88df5b7436dc6711b221a122ff","last_reissued_at":"2026-07-05T10:16:48.932834Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:48.932834Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Activation-aware Probe-Query: Effective Key-Value Retrieval for Long-Context LLMs Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Cheng Deng, Haoyang Li, Jiachuan Wang, Jiaqi Tang, Jun Wang, Lei Chen, Qingfa Xiao, Shuangyin Li, Yongqi Zhang","submitted_at":"2025-02-19T08:50:44Z","abstract_excerpt":"Recent advances in large language models (LLMs) have showcased exceptional performance in long-context tasks, while facing significant inference efficiency challenges with limited GPU memory. Existing solutions first proposed the sliding-window approach to accumulate a set of historical \\textbf{key-value} (KV) pairs for reuse, then further improvements selectively retain its subsets at each step. However, due to the sparse attention distribution across a long context, it is hard to identify and recall relevant KV pairs, as the attention is distracted by massive candidate pairs. Additionally, w"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13542","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/2502.13542/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":"2502.13542","created_at":"2026-07-05T10:16:48.932885+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13542v1","created_at":"2026-07-05T10:16:48.932885+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13542","created_at":"2026-07-05T10:16:48.932885+00:00"},{"alias_kind":"pith_short_12","alias_value":"NB4RRKL5TSVS","created_at":"2026-07-05T10:16:48.932885+00:00"},{"alias_kind":"pith_short_16","alias_value":"NB4RRKL5TSVS3CHA","created_at":"2026-07-05T10:16:48.932885+00:00"},{"alias_kind":"pith_short_8","alias_value":"NB4RRKL5","created_at":"2026-07-05T10:16:48.932885+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/NB4RRKL5TSVS3CHAEPAWK3BKMA","json":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA.json","graph_json":"https://pith.science/api/pith-number/NB4RRKL5TSVS3CHAEPAWK3BKMA/graph.json","events_json":"https://pith.science/api/pith-number/NB4RRKL5TSVS3CHAEPAWK3BKMA/events.json","paper":"https://pith.science/paper/NB4RRKL5"},"agent_actions":{"view_html":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA","download_json":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA.json","view_paper":"https://pith.science/paper/NB4RRKL5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13542&json=true","fetch_graph":"https://pith.science/api/pith-number/NB4RRKL5TSVS3CHAEPAWK3BKMA/graph.json","fetch_events":"https://pith.science/api/pith-number/NB4RRKL5TSVS3CHAEPAWK3BKMA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA/action/storage_attestation","attest_author":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA/action/author_attestation","sign_citation":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA/action/citation_signature","submit_replication":"https://pith.science/pith/NB4RRKL5TSVS3CHAEPAWK3BKMA/action/replication_record"}},"created_at":"2026-07-05T10:16:48.932885+00:00","updated_at":"2026-07-05T10:16:48.932885+00:00"}