{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:MFVIJU56MWZXKZZU7PSBABUZ6E","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9a8275002d4ae4875a038c95e1aff2efe8cc91eda3c923152003115de935b6ec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T00:25:20Z","title_canon_sha256":"04d0b35a6e463659d541d1fd3d84ca265b91927d9083b294cd758d3fb7828740"},"schema_version":"1.0","source":{"id":"2405.03917","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.03917","created_at":"2026-07-05T08:16:12Z"},{"alias_kind":"arxiv_version","alias_value":"2405.03917v1","created_at":"2026-07-05T08:16:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.03917","created_at":"2026-07-05T08:16:12Z"},{"alias_kind":"pith_short_12","alias_value":"MFVIJU56MWZX","created_at":"2026-07-05T08:16:12Z"},{"alias_kind":"pith_short_16","alias_value":"MFVIJU56MWZXKZZU","created_at":"2026-07-05T08:16:12Z"},{"alias_kind":"pith_short_8","alias_value":"MFVIJU56","created_at":"2026-07-05T08:16:12Z"}],"graph_snapshots":[{"event_id":"sha256:6b73dea0f822271b06da3474826b62221df113dd064d45c9dd861765a2ecead9","target":"graph","created_at":"2026-07-05T08:16:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2405.03917/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Efficient deployment of Large Language Models (LLMs) requires batching multiple requests together to improve throughput. As the batch size, context length, or model size increases, the size of the key and value (KV) cache can quickly become the main contributor to GPU memory usage and the bottleneck of inference latency. Quantization has emerged as an effective technique for KV cache compression, but existing methods still fail at very low bit widths. We observe that distinct channels of a key/value activation embedding are highly inter-dependent, and the joint entropy of multiple channels gro","authors_text":"Anshumali Shrivastava, Jonah Yi, Tianyi Zhang, Zhaozhuo Xu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T00:25:20Z","title":"KV Cache is 1 Bit Per Channel: Efficient Large Language Model Inference with Coupled Quantization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.03917","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f4109c7be0b4fa5645b477cf83a45fd6daf03530653c226b6b88c13458add08c","target":"record","created_at":"2026-07-05T08:16:12Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9a8275002d4ae4875a038c95e1aff2efe8cc91eda3c923152003115de935b6ec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-07T00:25:20Z","title_canon_sha256":"04d0b35a6e463659d541d1fd3d84ca265b91927d9083b294cd758d3fb7828740"},"schema_version":"1.0","source":{"id":"2405.03917","kind":"arxiv","version":1}},"canonical_sha256":"616a84d3be65b3756734fbe4100699f1260eb00928055aa33ab483a75c180c6d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"616a84d3be65b3756734fbe4100699f1260eb00928055aa33ab483a75c180c6d","first_computed_at":"2026-07-05T08:16:12.718469Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:16:12.718469Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JX53W5VZV1fG0cndjGJ3QF8Fl/sxZCAj+Sb+UEMQfwYWAk4rmfu2GhTz2FAl7oesGXqgzdpvBHQV57GQTrePAg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:16:12.718958Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.03917","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f4109c7be0b4fa5645b477cf83a45fd6daf03530653c226b6b88c13458add08c","sha256:6b73dea0f822271b06da3474826b62221df113dd064d45c9dd861765a2ecead9"],"state_sha256":"68b41fc35ef0e93933a02a8c4c3663778848bc6d995bf1aaea1a412863a458c8"}