{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PUN4BJJKBJ7B5E3GMP357XHBWQ","short_pith_number":"pith:PUN4BJJK","schema_version":"1.0","canonical_sha256":"7d1bc0a52a0a7e1e936663f7dfdce1b439eada2010ae5b473626233cffb39528","source":{"kind":"arxiv","id":"2412.02252","version":2},"attestation_state":"computed","paper":{"title":"Compressing KV Cache for Long-Context LLM Inference with Inter-Layer Attention Similarity","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Da Ma, Hanqi Li, Hongshen Xu, Kai Yu, Lei Pan, Lu Chen, Shuai Fan, Situo Zhang, Su Zhu, Yuxun Miao, Zhi Chen","submitted_at":"2024-12-03T08:29:27Z","abstract_excerpt":"The rapid expansion of context window sizes in Large Language Models~(LLMs) has enabled them to tackle increasingly complex tasks involving lengthy documents. However, this progress comes at the cost of a substantial increase in memory usage during inference, primarily due to the linear growth of the key-value~(KV) cache. Existing KV cache compression methods often discard less relevant tokens, which can lead to significant performance degradation when critical information is lost. In this paper, we propose \\textsc{PoD}~(Proximal tokens over Distant tokens), a novel KV cache compression framew"},"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":"2412.02252","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-12-03T08:29:27Z","cross_cats_sorted":[],"title_canon_sha256":"7411289659b7c5f5a20db115f4fc3b6c6bc3578c8ab452c2057d96711840b270","abstract_canon_sha256":"fbc2a19a440ff8f9fa8c54eadb6e4e15c0d909474375d311a07cb71fd733b1c8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:47:16.202219Z","signature_b64":"BTwJwaDnI5fMYv12wU033QS35mmoK6RWWKELzkf7BfPR3me1tjsAGh6SkGQsx4aU0MF6MLULILggfND84KreBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7d1bc0a52a0a7e1e936663f7dfdce1b439eada2010ae5b473626233cffb39528","last_reissued_at":"2026-07-05T11:47:16.201712Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:47:16.201712Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Compressing KV Cache for Long-Context LLM Inference with Inter-Layer Attention Similarity","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Da Ma, Hanqi Li, Hongshen Xu, Kai Yu, Lei Pan, Lu Chen, Shuai Fan, Situo Zhang, Su Zhu, Yuxun Miao, Zhi Chen","submitted_at":"2024-12-03T08:29:27Z","abstract_excerpt":"The rapid expansion of context window sizes in Large Language Models~(LLMs) has enabled them to tackle increasingly complex tasks involving lengthy documents. However, this progress comes at the cost of a substantial increase in memory usage during inference, primarily due to the linear growth of the key-value~(KV) cache. Existing KV cache compression methods often discard less relevant tokens, which can lead to significant performance degradation when critical information is lost. In this paper, we propose \\textsc{PoD}~(Proximal tokens over Distant tokens), a novel KV cache compression framew"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.02252","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/2412.02252/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":"2412.02252","created_at":"2026-07-05T11:47:16.201775+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.02252v2","created_at":"2026-07-05T11:47:16.201775+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.02252","created_at":"2026-07-05T11:47:16.201775+00:00"},{"alias_kind":"pith_short_12","alias_value":"PUN4BJJKBJ7B","created_at":"2026-07-05T11:47:16.201775+00:00"},{"alias_kind":"pith_short_16","alias_value":"PUN4BJJKBJ7B5E3G","created_at":"2026-07-05T11:47:16.201775+00:00"},{"alias_kind":"pith_short_8","alias_value":"PUN4BJJK","created_at":"2026-07-05T11:47:16.201775+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/PUN4BJJKBJ7B5E3GMP357XHBWQ","json":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ.json","graph_json":"https://pith.science/api/pith-number/PUN4BJJKBJ7B5E3GMP357XHBWQ/graph.json","events_json":"https://pith.science/api/pith-number/PUN4BJJKBJ7B5E3GMP357XHBWQ/events.json","paper":"https://pith.science/paper/PUN4BJJK"},"agent_actions":{"view_html":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ","download_json":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ.json","view_paper":"https://pith.science/paper/PUN4BJJK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.02252&json=true","fetch_graph":"https://pith.science/api/pith-number/PUN4BJJKBJ7B5E3GMP357XHBWQ/graph.json","fetch_events":"https://pith.science/api/pith-number/PUN4BJJKBJ7B5E3GMP357XHBWQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ/action/storage_attestation","attest_author":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ/action/author_attestation","sign_citation":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ/action/citation_signature","submit_replication":"https://pith.science/pith/PUN4BJJKBJ7B5E3GMP357XHBWQ/action/replication_record"}},"created_at":"2026-07-05T11:47:16.201775+00:00","updated_at":"2026-07-05T11:47:16.201775+00:00"}