{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KDFAPECLEN32P4OCADCIAUNPD6","short_pith_number":"pith:KDFAPECL","schema_version":"1.0","canonical_sha256":"50ca07904b2377a7f1c200c48051af1f9f38fb37e04f48979a7265ca632af7a2","source":{"kind":"arxiv","id":"2412.13649","version":3},"attestation_state":"computed","paper":{"title":"SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Deyu Zhou, Jialong Wu, Linhai Zhang, Yilong Lai, Yulan He, Zhenglin Wang","submitted_at":"2024-12-18T09:27:33Z","abstract_excerpt":"Key-Value (KV) cache has become a bottleneck of LLMs for long-context generation. Despite the numerous efforts in this area, the optimization for the decoding phase is generally ignored. However, we believe such optimization is crucial, especially for long-output generation tasks based on the following two observations: (i) Excessive compression during the prefill phase, which requires specific full context impairs the comprehension of the reasoning task; (ii) Deviation of heavy hitters occurs in the reasoning tasks with long outputs. Therefore, SCOPE, a simple yet efficient framework that sep"},"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":"2412.13649","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-12-18T09:27:33Z","cross_cats_sorted":[],"title_canon_sha256":"ed220e4f7f0f50ce3464886241ea7c53c3450061bd2cbbf2ebd5b372c7d94434","abstract_canon_sha256":"a719f0704e63757dfe9d23be991ba7bfb166249765734d0f5be7ed2907cd1fb5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:26.021285Z","signature_b64":"ckUYawBYihWR+g+gvQoyx2iWO9gJ/b05xBuRcK+UVgji804XUOmxkX6YXpO+JtT/N51fUes2noJwFIy/8xjeBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"50ca07904b2377a7f1c200c48051af1f9f38fb37e04f48979a7265ca632af7a2","last_reissued_at":"2026-07-05T11:14:26.020813Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:26.020813Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SCOPE: Optimizing Key-Value Cache Compression in Long-context Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Deyu Zhou, Jialong Wu, Linhai Zhang, Yilong Lai, Yulan He, Zhenglin Wang","submitted_at":"2024-12-18T09:27:33Z","abstract_excerpt":"Key-Value (KV) cache has become a bottleneck of LLMs for long-context generation. Despite the numerous efforts in this area, the optimization for the decoding phase is generally ignored. However, we believe such optimization is crucial, especially for long-output generation tasks based on the following two observations: (i) Excessive compression during the prefill phase, which requires specific full context impairs the comprehension of the reasoning task; (ii) Deviation of heavy hitters occurs in the reasoning tasks with long outputs. Therefore, SCOPE, a simple yet efficient framework that sep"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.13649","kind":"arxiv","version":3},"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/2412.13649/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":"2412.13649","created_at":"2026-07-05T11:14:26.020870+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.13649v3","created_at":"2026-07-05T11:14:26.020870+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.13649","created_at":"2026-07-05T11:14:26.020870+00:00"},{"alias_kind":"pith_short_12","alias_value":"KDFAPECLEN32","created_at":"2026-07-05T11:14:26.020870+00:00"},{"alias_kind":"pith_short_16","alias_value":"KDFAPECLEN32P4OC","created_at":"2026-07-05T11:14:26.020870+00:00"},{"alias_kind":"pith_short_8","alias_value":"KDFAPECL","created_at":"2026-07-05T11:14:26.020870+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08032","citing_title":"What to Keep, What to Forget: A Rate--Distortion View of Memory Compaction in LLMs and Agents","ref_index":126,"is_internal_anchor":true},{"citing_arxiv_id":"2605.23200","citing_title":"Adaptive Mass-Segmented KV Compression for Long-Context Reasoning","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2502.01941","citing_title":"Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03679","citing_title":"LightThinker++: From Reasoning Compression to Memory Management","ref_index":66,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6","json":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6.json","graph_json":"https://pith.science/api/pith-number/KDFAPECLEN32P4OCADCIAUNPD6/graph.json","events_json":"https://pith.science/api/pith-number/KDFAPECLEN32P4OCADCIAUNPD6/events.json","paper":"https://pith.science/paper/KDFAPECL"},"agent_actions":{"view_html":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6","download_json":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6.json","view_paper":"https://pith.science/paper/KDFAPECL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.13649&json=true","fetch_graph":"https://pith.science/api/pith-number/KDFAPECLEN32P4OCADCIAUNPD6/graph.json","fetch_events":"https://pith.science/api/pith-number/KDFAPECLEN32P4OCADCIAUNPD6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6/action/storage_attestation","attest_author":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6/action/author_attestation","sign_citation":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6/action/citation_signature","submit_replication":"https://pith.science/pith/KDFAPECLEN32P4OCADCIAUNPD6/action/replication_record"}},"created_at":"2026-07-05T11:14:26.020870+00:00","updated_at":"2026-07-05T11:14:26.020870+00:00"}