{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7KAAPLDQNOIGV26QVL7GIEMVGI","short_pith_number":"pith:7KAAPLDQ","schema_version":"1.0","canonical_sha256":"fa8007ac706b906aebd0aafe6411953202f1e8b7c3a90ea99efa032ca15844b4","source":{"kind":"arxiv","id":"2605.19660","version":1},"attestation_state":"computed","paper":{"title":"OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Chao Zhang, Dayou Du, Hongxia Yang, Jing Xiong, Ngai Wong, Rui Yang, Wei Wu, Xialie Zhuang, Yaxiu Liu, Yifan Zhang, Yik-Chung Wu, Yuchen Xie, Yulei Qian, Zunhai Su","submitted_at":"2026-05-19T10:53:03Z","abstract_excerpt":"The rapid advancement toward long-context reasoning and multi-modal intelligence has made the memory footprint of the Key-Value (KV) cache a dominant memory bottleneck for efficient deployment. While the established per-channel quantization effectively accommodates intrinsic channel-wise outliers in Key tensors, its efficacy diminishes under extreme compression. In this work, we revisit the inherent limitations of the per-channel quantization paradigm from both empirical and theoretical perspectives. Our analysis identifies Token Norm Imbalance (TNI) as the primary bottleneck to quantization f"},"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":"2605.19660","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-19T10:53:03Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"f4fc937ecf8f6376ab729a6ff59c00250462d1df59eb9fd77ecf1d6c2290389c","abstract_canon_sha256":"e6aecdc5a3fc467af05390086641e768cd3196dd96d8f2522ab13d0564e6c06a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:56.546324Z","signature_b64":"lbGiZHH3AlKLf296iHhzIAtpGv8Q19ReWQ0sx/RzFqc0+ac/4NuXiH33SvpSAeywzkUr6LrUxayNTziK/o1PAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fa8007ac706b906aebd0aafe6411953202f1e8b7c3a90ea99efa032ca15844b4","last_reissued_at":"2026-05-20T01:05:56.545654Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:56.545654Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Chao Zhang, Dayou Du, Hongxia Yang, Jing Xiong, Ngai Wong, Rui Yang, Wei Wu, Xialie Zhuang, Yaxiu Liu, Yifan Zhang, Yik-Chung Wu, Yuchen Xie, Yulei Qian, Zunhai Su","submitted_at":"2026-05-19T10:53:03Z","abstract_excerpt":"The rapid advancement toward long-context reasoning and multi-modal intelligence has made the memory footprint of the Key-Value (KV) cache a dominant memory bottleneck for efficient deployment. While the established per-channel quantization effectively accommodates intrinsic channel-wise outliers in Key tensors, its efficacy diminishes under extreme compression. In this work, we revisit the inherent limitations of the per-channel quantization paradigm from both empirical and theoretical perspectives. Our analysis identifies Token Norm Imbalance (TNI) as the primary bottleneck to quantization f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19660","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/2605.19660/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":"2605.19660","created_at":"2026-05-20T01:05:56.545783+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19660v1","created_at":"2026-05-20T01:05:56.545783+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19660","created_at":"2026-05-20T01:05:56.545783+00:00"},{"alias_kind":"pith_short_12","alias_value":"7KAAPLDQNOIG","created_at":"2026-05-20T01:05:56.545783+00:00"},{"alias_kind":"pith_short_16","alias_value":"7KAAPLDQNOIGV26Q","created_at":"2026-05-20T01:05:56.545783+00:00"},{"alias_kind":"pith_short_8","alias_value":"7KAAPLDQ","created_at":"2026-05-20T01:05:56.545783+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/7KAAPLDQNOIGV26QVL7GIEMVGI","json":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI.json","graph_json":"https://pith.science/api/pith-number/7KAAPLDQNOIGV26QVL7GIEMVGI/graph.json","events_json":"https://pith.science/api/pith-number/7KAAPLDQNOIGV26QVL7GIEMVGI/events.json","paper":"https://pith.science/paper/7KAAPLDQ"},"agent_actions":{"view_html":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI","download_json":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI.json","view_paper":"https://pith.science/paper/7KAAPLDQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19660&json=true","fetch_graph":"https://pith.science/api/pith-number/7KAAPLDQNOIGV26QVL7GIEMVGI/graph.json","fetch_events":"https://pith.science/api/pith-number/7KAAPLDQNOIGV26QVL7GIEMVGI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI/action/storage_attestation","attest_author":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI/action/author_attestation","sign_citation":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI/action/citation_signature","submit_replication":"https://pith.science/pith/7KAAPLDQNOIGV26QVL7GIEMVGI/action/replication_record"}},"created_at":"2026-05-20T01:05:56.545783+00:00","updated_at":"2026-05-20T01:05:56.545783+00:00"}