{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:27I2ALV7AJURYSRRMIBH4MONJ5","short_pith_number":"pith:27I2ALV7","schema_version":"1.0","canonical_sha256":"d7d1a02ebf02691c4a3162027e31cd4f6bbfc04f64bff9eade1cd9aa1fb7121a","source":{"kind":"arxiv","id":"2507.19906","version":2},"attestation_state":"computed","paper":{"title":"CaliDrop: KV Cache Compression with Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Juntao Li, Min Zhang, Ping Li, Qingrong Xia, Quantong Qiu, Xinyu Duan, Yi Su, Yuechi Zhou, Zhefeng Wang","submitted_at":"2025-07-26T10:34:53Z","abstract_excerpt":"Large Language Models (LLMs) require substantial computational resources during generation. While the Key-Value (KV) cache significantly accelerates this process by storing attention intermediates, its memory footprint grows linearly with sequence length, batch size, and model size, creating a bottleneck in long-context scenarios. Various KV cache compression techniques, including token eviction, quantization, and low-rank projection, have been proposed to mitigate this bottleneck, often complementing each other. This paper focuses on enhancing token eviction strategies. Token eviction leverag"},"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":"2507.19906","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-07-26T10:34:53Z","cross_cats_sorted":[],"title_canon_sha256":"05d527725700940ac194e579cbe55de484d1cf8de1f1e19ec0fbfaef3f1082c1","abstract_canon_sha256":"76b3c2cb50e5e3a69bbb9959feb7a02265fbad70cf084eb551c840281568b7f8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:48:06.137499Z","signature_b64":"+MuptTHBqBlb/iK4lS7F9PEwpuwg1uE4+Y7vTjoSrIzXRIlhd+ynWJGOaxFlGEfhF9GIwoE2+kzsVG/F9VRBCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d7d1a02ebf02691c4a3162027e31cd4f6bbfc04f64bff9eade1cd9aa1fb7121a","last_reissued_at":"2026-07-05T11:48:06.136958Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:48:06.136958Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CaliDrop: KV Cache Compression with Calibration","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Juntao Li, Min Zhang, Ping Li, Qingrong Xia, Quantong Qiu, Xinyu Duan, Yi Su, Yuechi Zhou, Zhefeng Wang","submitted_at":"2025-07-26T10:34:53Z","abstract_excerpt":"Large Language Models (LLMs) require substantial computational resources during generation. While the Key-Value (KV) cache significantly accelerates this process by storing attention intermediates, its memory footprint grows linearly with sequence length, batch size, and model size, creating a bottleneck in long-context scenarios. Various KV cache compression techniques, including token eviction, quantization, and low-rank projection, have been proposed to mitigate this bottleneck, often complementing each other. This paper focuses on enhancing token eviction strategies. Token eviction leverag"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.19906","kind":"arxiv","version":2},"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/2507.19906/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":"2507.19906","created_at":"2026-07-05T11:48:06.137021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.19906v2","created_at":"2026-07-05T11:48:06.137021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.19906","created_at":"2026-07-05T11:48:06.137021+00:00"},{"alias_kind":"pith_short_12","alias_value":"27I2ALV7AJUR","created_at":"2026-07-05T11:48:06.137021+00:00"},{"alias_kind":"pith_short_16","alias_value":"27I2ALV7AJURYSRR","created_at":"2026-07-05T11:48:06.137021+00:00"},{"alias_kind":"pith_short_8","alias_value":"27I2ALV7","created_at":"2026-07-05T11:48:06.137021+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/27I2ALV7AJURYSRRMIBH4MONJ5","json":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5.json","graph_json":"https://pith.science/api/pith-number/27I2ALV7AJURYSRRMIBH4MONJ5/graph.json","events_json":"https://pith.science/api/pith-number/27I2ALV7AJURYSRRMIBH4MONJ5/events.json","paper":"https://pith.science/paper/27I2ALV7"},"agent_actions":{"view_html":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5","download_json":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5.json","view_paper":"https://pith.science/paper/27I2ALV7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.19906&json=true","fetch_graph":"https://pith.science/api/pith-number/27I2ALV7AJURYSRRMIBH4MONJ5/graph.json","fetch_events":"https://pith.science/api/pith-number/27I2ALV7AJURYSRRMIBH4MONJ5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5/action/storage_attestation","attest_author":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5/action/author_attestation","sign_citation":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5/action/citation_signature","submit_replication":"https://pith.science/pith/27I2ALV7AJURYSRRMIBH4MONJ5/action/replication_record"}},"created_at":"2026-07-05T11:48:06.137021+00:00","updated_at":"2026-07-05T11:48:06.137021+00:00"}