{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5FEICMMBMVHHACF327CWGKQ73L","short_pith_number":"pith:5FEICMMB","schema_version":"1.0","canonical_sha256":"e948813181654e7008bbd7c5632a1fdac3e5ce4b93d4b34bca3e71961bf942bf","source":{"kind":"arxiv","id":"2508.16134","version":1},"attestation_state":"computed","paper":{"title":"CommonKV: Compressing KV Cache with Cross-layer Parameter Sharing","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoyu Qiao, Lujun Li, Qingfu Zhu, Wanxiang Che, Yixuan Wang","submitted_at":"2025-08-22T06:55:45Z","abstract_excerpt":"Large Language Models (LLMs) confront significant memory challenges due to the escalating KV cache with increasing sequence length. As a crucial technique, existing cross-layer KV cache sharing methods either necessitate modified model architectures with subsequent pre-training or incur significant performance degradation at high compression rates. To mitigate these challenges, we propose CommonKV, a training-free method for cross-layer KV cache compression through adjacent parameters sharing. Inspired by the high similarity observed in cross-layer hidden states, we utilize Singular Value Deco"},"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":"2508.16134","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-22T06:55:45Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a93bf48a501eba0972bed1ae1ab9e3eb070bb474fba38edb9c0fe89ce0c456fe","abstract_canon_sha256":"c8d4d9b62402ba944cb8c4ea3693ab9bf4006647934265b2847d090e4de44ba3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:57:42.536774Z","signature_b64":"62Jro/gNDTNZ4Qj6crkmH68zQFo0LawtsyYZI2tw2N+V9qYx/4CV2aplnlLcdW/5aNUP8rFu7wGZ04IhFnHQBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e948813181654e7008bbd7c5632a1fdac3e5ce4b93d4b34bca3e71961bf942bf","last_reissued_at":"2026-07-05T11:57:42.536397Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:57:42.536397Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CommonKV: Compressing KV Cache with Cross-layer Parameter Sharing","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoyu Qiao, Lujun Li, Qingfu Zhu, Wanxiang Che, Yixuan Wang","submitted_at":"2025-08-22T06:55:45Z","abstract_excerpt":"Large Language Models (LLMs) confront significant memory challenges due to the escalating KV cache with increasing sequence length. As a crucial technique, existing cross-layer KV cache sharing methods either necessitate modified model architectures with subsequent pre-training or incur significant performance degradation at high compression rates. To mitigate these challenges, we propose CommonKV, a training-free method for cross-layer KV cache compression through adjacent parameters sharing. Inspired by the high similarity observed in cross-layer hidden states, we utilize Singular Value Deco"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.16134","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/2508.16134/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":"2508.16134","created_at":"2026-07-05T11:57:42.536454+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.16134v1","created_at":"2026-07-05T11:57:42.536454+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.16134","created_at":"2026-07-05T11:57:42.536454+00:00"},{"alias_kind":"pith_short_12","alias_value":"5FEICMMBMVHH","created_at":"2026-07-05T11:57:42.536454+00:00"},{"alias_kind":"pith_short_16","alias_value":"5FEICMMBMVHHACF3","created_at":"2026-07-05T11:57:42.536454+00:00"},{"alias_kind":"pith_short_8","alias_value":"5FEICMMB","created_at":"2026-07-05T11:57:42.536454+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.31105","citing_title":"GRKV: Global Regression for Training-Free KV Cache Compression in Long-Context LLMs","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17757","citing_title":"OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22782","citing_title":"Stochastic KV Routing: Enabling Adaptive Depth-Wise Cache Sharing","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L","json":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L.json","graph_json":"https://pith.science/api/pith-number/5FEICMMBMVHHACF327CWGKQ73L/graph.json","events_json":"https://pith.science/api/pith-number/5FEICMMBMVHHACF327CWGKQ73L/events.json","paper":"https://pith.science/paper/5FEICMMB"},"agent_actions":{"view_html":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L","download_json":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L.json","view_paper":"https://pith.science/paper/5FEICMMB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.16134&json=true","fetch_graph":"https://pith.science/api/pith-number/5FEICMMBMVHHACF327CWGKQ73L/graph.json","fetch_events":"https://pith.science/api/pith-number/5FEICMMBMVHHACF327CWGKQ73L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L/action/storage_attestation","attest_author":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L/action/author_attestation","sign_citation":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L/action/citation_signature","submit_replication":"https://pith.science/pith/5FEICMMBMVHHACF327CWGKQ73L/action/replication_record"}},"created_at":"2026-07-05T11:57:42.536454+00:00","updated_at":"2026-07-05T11:57:42.536454+00:00"}