{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:4VKVPDTRMAKTFKZUE6EQERXFQC","short_pith_number":"pith:4VKVPDTR","canonical_record":{"source":{"id":"2310.05079","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-08T09:05:14Z","cross_cats_sorted":[],"title_canon_sha256":"b64b137eeeca4d0188e5a8db7403a066aadffd9908b944cff7c24a9b7fa6a8d3","abstract_canon_sha256":"139be87526e2c3275aa220b0f3a3271226d6fd9b9f03984db47f11055b767f32"},"schema_version":"1.0"},"canonical_sha256":"e555578e71601532ab3427890246e580a42d24848be9b886d0e28005be328e5c","source":{"kind":"arxiv","id":"2310.05079","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.05079","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"arxiv_version","alias_value":"2310.05079v2","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.05079","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_12","alias_value":"4VKVPDTRMAKT","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_16","alias_value":"4VKVPDTRMAKTFKZU","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_8","alias_value":"4VKVPDTR","created_at":"2026-07-05T07:56:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:4VKVPDTRMAKTFKZUE6EQERXFQC","target":"record","payload":{"canonical_record":{"source":{"id":"2310.05079","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-08T09:05:14Z","cross_cats_sorted":[],"title_canon_sha256":"b64b137eeeca4d0188e5a8db7403a066aadffd9908b944cff7c24a9b7fa6a8d3","abstract_canon_sha256":"139be87526e2c3275aa220b0f3a3271226d6fd9b9f03984db47f11055b767f32"},"schema_version":"1.0"},"canonical_sha256":"e555578e71601532ab3427890246e580a42d24848be9b886d0e28005be328e5c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:56:00.108891Z","signature_b64":"nnYJrxpX6fsMQcPugMyoW9QBzxsGVXhzRjGjg1Bo+rlFqSiOx8/9IIM4iPai+HE60pDJkHUmQ0xQFNYUHsdwBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e555578e71601532ab3427890246e580a42d24848be9b886d0e28005be328e5c","last_reissued_at":"2026-07-05T07:56:00.108437Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:56:00.108437Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2310.05079","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:56:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CgFFkgX3v7tei0pEkKFYYvTkJChiNUhhyOi/FZAPAL0JDJPCUO4DZ76MnH81tdkHjHB7axFywfZuxXO3dPw5Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T05:00:34.801223Z"},"content_sha256":"f51e2d1846c95955f3154464adc9434ba81d8edd714f4194965195f26292cbd9","schema_version":"1.0","event_id":"sha256:f51e2d1846c95955f3154464adc9434ba81d8edd714f4194965195f26292cbd9"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:4VKVPDTRMAKTFKZUE6EQERXFQC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Cheng Zhang, George A. Constantinides, Ilia Shumailov, Jianyi Cheng, Yiren Zhao","submitted_at":"2023-10-08T09:05:14Z","abstract_excerpt":"The inference of Large language models (LLMs) requires immense computation and memory resources. To curtail these costs, quantisation has merged as a promising solution, but existing LLM quantisation mainly focuses on 8-bit. In this work, we explore the statistical and learning properties of the LLM layer and attribute the bottleneck of LLM quantisation to numerical scaling offsets. To address this, we adapt block quantisations for LLMs, a family of methods that share scaling factors across packed numbers. Block quantisations efficiently reduce the numerical scaling offsets solely from an arit"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.05079","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/2310.05079/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":"2026-07-05T07:56:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gHRAAFTRnLKS2OHKoijxO8WWM9U/NpPyTZra/x4x7L133kY5JV+adjq93074jhWdoS8QjQeZs91MUlp0HO8fBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T05:00:34.801967Z"},"content_sha256":"f9932680d279522a638ac6f49698306230b954b1172101a5574847deaa2fffa6","schema_version":"1.0","event_id":"sha256:f9932680d279522a638ac6f49698306230b954b1172101a5574847deaa2fffa6"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/bundle.json","state_url":"https://pith.science/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/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":"2026-08-21T05:00:34Z","links":{"resolver":"https://pith.science/pith/4VKVPDTRMAKTFKZUE6EQERXFQC","bundle":"https://pith.science/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/bundle.json","state":"https://pith.science/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4VKVPDTRMAKTFKZUE6EQERXFQC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:4VKVPDTRMAKTFKZUE6EQERXFQC","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":"139be87526e2c3275aa220b0f3a3271226d6fd9b9f03984db47f11055b767f32","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-08T09:05:14Z","title_canon_sha256":"b64b137eeeca4d0188e5a8db7403a066aadffd9908b944cff7c24a9b7fa6a8d3"},"schema_version":"1.0","source":{"id":"2310.05079","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.05079","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"arxiv_version","alias_value":"2310.05079v2","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.05079","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_12","alias_value":"4VKVPDTRMAKT","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_16","alias_value":"4VKVPDTRMAKTFKZU","created_at":"2026-07-05T07:56:00Z"},{"alias_kind":"pith_short_8","alias_value":"4VKVPDTR","created_at":"2026-07-05T07:56:00Z"}],"graph_snapshots":[{"event_id":"sha256:f9932680d279522a638ac6f49698306230b954b1172101a5574847deaa2fffa6","target":"graph","created_at":"2026-07-05T07:56:00Z","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/2310.05079/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The inference of Large language models (LLMs) requires immense computation and memory resources. To curtail these costs, quantisation has merged as a promising solution, but existing LLM quantisation mainly focuses on 8-bit. In this work, we explore the statistical and learning properties of the LLM layer and attribute the bottleneck of LLM quantisation to numerical scaling offsets. To address this, we adapt block quantisations for LLMs, a family of methods that share scaling factors across packed numbers. Block quantisations efficiently reduce the numerical scaling offsets solely from an arit","authors_text":"Cheng Zhang, George A. Constantinides, Ilia Shumailov, Jianyi Cheng, Yiren Zhao","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-08T09:05:14Z","title":"Revisiting Block-based Quantisation: What is Important for Sub-8-bit LLM Inference?"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.05079","kind":"arxiv","version":2},"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:f51e2d1846c95955f3154464adc9434ba81d8edd714f4194965195f26292cbd9","target":"record","created_at":"2026-07-05T07:56:00Z","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":"139be87526e2c3275aa220b0f3a3271226d6fd9b9f03984db47f11055b767f32","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-10-08T09:05:14Z","title_canon_sha256":"b64b137eeeca4d0188e5a8db7403a066aadffd9908b944cff7c24a9b7fa6a8d3"},"schema_version":"1.0","source":{"id":"2310.05079","kind":"arxiv","version":2}},"canonical_sha256":"e555578e71601532ab3427890246e580a42d24848be9b886d0e28005be328e5c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e555578e71601532ab3427890246e580a42d24848be9b886d0e28005be328e5c","first_computed_at":"2026-07-05T07:56:00.108437Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:56:00.108437Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nnYJrxpX6fsMQcPugMyoW9QBzxsGVXhzRjGjg1Bo+rlFqSiOx8/9IIM4iPai+HE60pDJkHUmQ0xQFNYUHsdwBw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:56:00.108891Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.05079","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f51e2d1846c95955f3154464adc9434ba81d8edd714f4194965195f26292cbd9","sha256:f9932680d279522a638ac6f49698306230b954b1172101a5574847deaa2fffa6"],"state_sha256":"02bea92d7c7d1590f92b7b5d6727588f5858b6b7c0ad1f446ebbf34d5f596bf8"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OxMXAHfdyb10d2RJ4c/Z6berrgqXTsS63JPDki9YL/2+cDT1X3kVjvux2t+LjvXh/gxf4QrJK2feS/N976C7Cw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T05:00:34.807844Z","bundle_sha256":"1f0585bf828c2c0acf647385b73f29a3811b33a1c6d1ed127c20d7811a474ec5"}}