{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FAPBVMYA27MXQS6R3PQDCYAXUW","short_pith_number":"pith:FAPBVMYA","schema_version":"1.0","canonical_sha256":"281e1ab300d7d9784bd1dbe0316017a582149e0d210ee5b3f417208429a91b0a","source":{"kind":"arxiv","id":"2501.16383","version":2},"attestation_state":"computed","paper":{"title":"RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hanyu Wei, Huangqi Yu, Kehong Yuan, Linge Li, Wang Shen, Zhe Chen, Zunhai Su","submitted_at":"2025-01-25T01:45:29Z","abstract_excerpt":"Key-Value (KV) cache facilitates efficient large language models (LLMs) inference by avoiding recomputation of past KVs. As the batch size and context length increase, the oversized KV caches become a significant memory bottleneck, highlighting the need for efficient compression. Existing KV quantization rely on fine-grained quantization or the retention of a significant portion of high bit-widths caches, both of which compromise compression ratio and often fail to maintain robustness at extremely low average bit-widths. In this work, we explore the potential of rotation technique for 2-bit KV"},"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":"2501.16383","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-25T01:45:29Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"64ee418bd6ed50bd4ecae54b8d06af4e8ee08a2040092e16c585d6e0134ea8f2","abstract_canon_sha256":"c0e9af561a1c87bb77c9044746bbe48c5e3eacd78cf00866544cb8c023cd91cc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:28.781573Z","signature_b64":"FsjGG3C78xeKdX6wUGl40zonWDnZraxWOuk17oVKgT5U46RlSoEGoYDhVNPGT2DVM16i6FlwendsL6wQ84OvCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"281e1ab300d7d9784bd1dbe0316017a582149e0d210ee5b3f417208429a91b0a","last_reissued_at":"2026-07-05T10:08:28.781066Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:28.781066Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Hanyu Wei, Huangqi Yu, Kehong Yuan, Linge Li, Wang Shen, Zhe Chen, Zunhai Su","submitted_at":"2025-01-25T01:45:29Z","abstract_excerpt":"Key-Value (KV) cache facilitates efficient large language models (LLMs) inference by avoiding recomputation of past KVs. As the batch size and context length increase, the oversized KV caches become a significant memory bottleneck, highlighting the need for efficient compression. Existing KV quantization rely on fine-grained quantization or the retention of a significant portion of high bit-widths caches, both of which compromise compression ratio and often fail to maintain robustness at extremely low average bit-widths. In this work, we explore the potential of rotation technique for 2-bit KV"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.16383","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/2501.16383/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":"2501.16383","created_at":"2026-07-05T10:08:28.781121+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.16383v2","created_at":"2026-07-05T10:08:28.781121+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.16383","created_at":"2026-07-05T10:08:28.781121+00:00"},{"alias_kind":"pith_short_12","alias_value":"FAPBVMYA27MX","created_at":"2026-07-05T10:08:28.781121+00:00"},{"alias_kind":"pith_short_16","alias_value":"FAPBVMYA27MXQS6R","created_at":"2026-07-05T10:08:28.781121+00:00"},{"alias_kind":"pith_short_8","alias_value":"FAPBVMYA","created_at":"2026-07-05T10:08:28.781121+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.17757","citing_title":"OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2504.19874","citing_title":"TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19660","citing_title":"OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17613","citing_title":"VeriCache: Turning Lossy KV Cache into Lossless LLM Inference","ref_index":61,"is_internal_anchor":false},{"citing_arxiv_id":"2601.14004","citing_title":"Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models","ref_index":287,"is_internal_anchor":false},{"citing_arxiv_id":"2602.02958","citing_title":"Quant VideoGen: Auto-Regressive Long Video Generation via 2-Bit KV-Cache Quantization","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2602.10718","citing_title":"SnapMLA: Efficient Long-Context MLA Decoding via Hardware-Aware FP8 Quantized Pipelining","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03562","citing_title":"HeadQ: Model-Visible Distortion and Score-Space Correction for KV-Cache Quantization","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24971","citing_title":"PolyKV: A Shared Asymmetrically-Compressed KV Cache Pool for Multi-Agent LLM Inference","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03562","citing_title":"HeadQ: Model-Visible Distortion and Score-Space Correction for KV-Cache Quantization","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW","json":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW.json","graph_json":"https://pith.science/api/pith-number/FAPBVMYA27MXQS6R3PQDCYAXUW/graph.json","events_json":"https://pith.science/api/pith-number/FAPBVMYA27MXQS6R3PQDCYAXUW/events.json","paper":"https://pith.science/paper/FAPBVMYA"},"agent_actions":{"view_html":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW","download_json":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW.json","view_paper":"https://pith.science/paper/FAPBVMYA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.16383&json=true","fetch_graph":"https://pith.science/api/pith-number/FAPBVMYA27MXQS6R3PQDCYAXUW/graph.json","fetch_events":"https://pith.science/api/pith-number/FAPBVMYA27MXQS6R3PQDCYAXUW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW/action/storage_attestation","attest_author":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW/action/author_attestation","sign_citation":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW/action/citation_signature","submit_replication":"https://pith.science/pith/FAPBVMYA27MXQS6R3PQDCYAXUW/action/replication_record"}},"created_at":"2026-07-05T10:08:28.781121+00:00","updated_at":"2026-07-05T10:08:28.781121+00:00"}