{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:BZ4J6HSSNIBDPVRZONA5LPNQ4C","short_pith_number":"pith:BZ4J6HSS","canonical_record":{"source":{"id":"2505.18231","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T12:40:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0554b80fdb34b745e5598e379515eac2be1af30a498c28ded69f9538623b1195","abstract_canon_sha256":"5e13d7b07df271ef56109ef14a70fb23181d2e1fb6b6e12c71ed546430846c5a"},"schema_version":"1.0"},"canonical_sha256":"0e789f1e526a0237d6397341d5bdb0e0802a0269a72394bb58dbfa937135243a","source":{"kind":"arxiv","id":"2505.18231","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18231","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18231v3","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18231","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_12","alias_value":"BZ4J6HSSNIBD","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_16","alias_value":"BZ4J6HSSNIBDPVRZ","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_8","alias_value":"BZ4J6HSS","created_at":"2026-07-16T01:22:25Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:BZ4J6HSSNIBDPVRZONA5LPNQ4C","target":"record","payload":{"canonical_record":{"source":{"id":"2505.18231","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T12:40:07Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0554b80fdb34b745e5598e379515eac2be1af30a498c28ded69f9538623b1195","abstract_canon_sha256":"5e13d7b07df271ef56109ef14a70fb23181d2e1fb6b6e12c71ed546430846c5a"},"schema_version":"1.0"},"canonical_sha256":"0e789f1e526a0237d6397341d5bdb0e0802a0269a72394bb58dbfa937135243a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-16T01:22:25.472722Z","signature_b64":"m5L1HWBoNnUkpkg2+tLnOf/rLQ1IsTCDeuWoZxEFZ82j/auxPlnqk19YNuPflss0g+u2mYnj/ZyN9p8RuU2tCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0e789f1e526a0237d6397341d5bdb0e0802a0269a72394bb58dbfa937135243a","last_reissued_at":"2026-07-16T01:22:25.471790Z","signature_status":"signed_v1","first_computed_at":"2026-07-16T01:22:25.471790Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.18231","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-16T01:22:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8G8iPBR1zyZawgn8b4kOiOm4HRV8gNgaZprUpWw7iXYz91CP81bx41YYsb9NwltRLBjSxRAXacEyUtm//uRPCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T08:48:11.623184Z"},"content_sha256":"ba07f434af882e1b8a35217891269adbec6090ad3850655e9f53f6631a5c2a4a","schema_version":"1.0","event_id":"sha256:ba07f434af882e1b8a35217891269adbec6090ad3850655e9f53f6631a5c2a4a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:BZ4J6HSSNIBDPVRZONA5LPNQ4C","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Donghyun Son, Euntae Choi, Sungjoo Yoo","submitted_at":"2025-05-23T12:40:07Z","abstract_excerpt":"Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache. Vector Quantization (VQ) is recently adopted to alleviate this issue, but we find that the existing approach is susceptible to distribution shift due to its reliance on calibration datasets. To address this limitation, we introduce NSNQuant, a calibration-free Vector Quantization (VQ) technique designed for low-bit compression of the KV cache. By applying a three-step transformation-1) a token-wise normalization (Nor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18231","kind":"arxiv","version":3},"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/2505.18231/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-16T01:22:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pi0aKlrgyUnkfcvVOTWN36IL6oE19LmLBwawIXy1C9IsYbeoAHXm88KdWAaM4PpTPNZh2uAVt4ILgJ6LIB7+AA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T08:48:11.624072Z"},"content_sha256":"7a9d52da8debdd7097f1563200005a25b0da7930a03efb3bfd380e9df85159a7","schema_version":"1.0","event_id":"sha256:7a9d52da8debdd7097f1563200005a25b0da7930a03efb3bfd380e9df85159a7"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/bundle.json","state_url":"https://pith.science/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T08:48:11Z","links":{"resolver":"https://pith.science/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C","bundle":"https://pith.science/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/bundle.json","state":"https://pith.science/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BZ4J6HSSNIBDPVRZONA5LPNQ4C/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:BZ4J6HSSNIBDPVRZONA5LPNQ4C","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":"5e13d7b07df271ef56109ef14a70fb23181d2e1fb6b6e12c71ed546430846c5a","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T12:40:07Z","title_canon_sha256":"0554b80fdb34b745e5598e379515eac2be1af30a498c28ded69f9538623b1195"},"schema_version":"1.0","source":{"id":"2505.18231","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18231","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18231v3","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18231","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_12","alias_value":"BZ4J6HSSNIBD","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_16","alias_value":"BZ4J6HSSNIBDPVRZ","created_at":"2026-07-16T01:22:25Z"},{"alias_kind":"pith_short_8","alias_value":"BZ4J6HSS","created_at":"2026-07-16T01:22:25Z"}],"graph_snapshots":[{"event_id":"sha256:7a9d52da8debdd7097f1563200005a25b0da7930a03efb3bfd380e9df85159a7","target":"graph","created_at":"2026-07-16T01:22:25Z","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/2505.18231/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Model (LLM) inference is typically memory-intensive, especially when processing large batch sizes and long sequences, due to the large size of key-value (KV) cache. Vector Quantization (VQ) is recently adopted to alleviate this issue, but we find that the existing approach is susceptible to distribution shift due to its reliance on calibration datasets. To address this limitation, we introduce NSNQuant, a calibration-free Vector Quantization (VQ) technique designed for low-bit compression of the KV cache. By applying a three-step transformation-1) a token-wise normalization (Nor","authors_text":"Donghyun Son, Euntae Choi, Sungjoo Yoo","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T12:40:07Z","title":"NSNQuant: A Double Normalization Approach for Calibration-Free Low-Bit Vector Quantization of KV Cache"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18231","kind":"arxiv","version":3},"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:ba07f434af882e1b8a35217891269adbec6090ad3850655e9f53f6631a5c2a4a","target":"record","created_at":"2026-07-16T01:22:25Z","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":"5e13d7b07df271ef56109ef14a70fb23181d2e1fb6b6e12c71ed546430846c5a","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-23T12:40:07Z","title_canon_sha256":"0554b80fdb34b745e5598e379515eac2be1af30a498c28ded69f9538623b1195"},"schema_version":"1.0","source":{"id":"2505.18231","kind":"arxiv","version":3}},"canonical_sha256":"0e789f1e526a0237d6397341d5bdb0e0802a0269a72394bb58dbfa937135243a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0e789f1e526a0237d6397341d5bdb0e0802a0269a72394bb58dbfa937135243a","first_computed_at":"2026-07-16T01:22:25.471790Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-16T01:22:25.471790Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"m5L1HWBoNnUkpkg2+tLnOf/rLQ1IsTCDeuWoZxEFZ82j/auxPlnqk19YNuPflss0g+u2mYnj/ZyN9p8RuU2tCQ==","signature_status":"signed_v1","signed_at":"2026-07-16T01:22:25.472722Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.18231","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:ba07f434af882e1b8a35217891269adbec6090ad3850655e9f53f6631a5c2a4a","sha256:7a9d52da8debdd7097f1563200005a25b0da7930a03efb3bfd380e9df85159a7"],"state_sha256":"a0aa0385f84aed13e5392fcfed1e5da363c4dc34a763183a1322697f01d9936d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"MCOpB2qHfjQH0m+7Uu1XWJA7NvQUy3IXm5EPAIv5Zve8dVJiBBGU1fPzMjIMP2m/8AnT97uB1hKTGj6aOCheDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T08:48:11.636137Z","bundle_sha256":"bdbbedd36935e834941345b40e2440146da9323eb4953807f269a904936246d2"}}