{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:62XTJQZ7MGNO7SIFZRRUMRWXRP","short_pith_number":"pith:62XTJQZ7","canonical_record":{"source":{"id":"2402.15319","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T13:39:16Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"bedb2ac721cc3f4b537071af89d8c26d4ce72df698bf8302c3e0486b7eafb6a9","abstract_canon_sha256":"0b9bdc31a5a2634a475fa018fee3d3093e3e516e524b03edd1651a3e918bce78"},"schema_version":"1.0"},"canonical_sha256":"f6af34c33f619aefc905cc634646d78bcfd5d824cefc0f0f66eda653fec2c569","source":{"kind":"arxiv","id":"2402.15319","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.15319","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"arxiv_version","alias_value":"2402.15319v2","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15319","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_12","alias_value":"62XTJQZ7MGNO","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_16","alias_value":"62XTJQZ7MGNO7SIF","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_8","alias_value":"62XTJQZ7","created_at":"2026-07-05T11:14:46Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:62XTJQZ7MGNO7SIFZRRUMRWXRP","target":"record","payload":{"canonical_record":{"source":{"id":"2402.15319","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T13:39:16Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"bedb2ac721cc3f4b537071af89d8c26d4ce72df698bf8302c3e0486b7eafb6a9","abstract_canon_sha256":"0b9bdc31a5a2634a475fa018fee3d3093e3e516e524b03edd1651a3e918bce78"},"schema_version":"1.0"},"canonical_sha256":"f6af34c33f619aefc905cc634646d78bcfd5d824cefc0f0f66eda653fec2c569","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:46.748820Z","signature_b64":"YaH2jKyXQPlZLZ2khZ4ebdWRiwujCP0+jWiIY//OpmHeqd+bPbh5K+60sh1uZxc/7eciM0Wek+SiKndLU61ZBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f6af34c33f619aefc905cc634646d78bcfd5d824cefc0f0f66eda653fec2c569","last_reissued_at":"2026-07-05T11:14:46.748358Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:46.748358Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.15319","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":"2026-07-05T11:14:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"c1QGfVJRQKuEQ5ipQa+AsTfOiYuAy7SqTEDPiVkS/N+0N/ow9JTxS1saQgWoO8+qdPt11RloV/UTecit2zfoBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T14:49:24.785936Z"},"content_sha256":"5860b49f1d421e713c3caedbcc0f4a393b1d60c044894c9fd9ba3cd99c595f4e","schema_version":"1.0","event_id":"sha256:5860b49f1d421e713c3caedbcc0f4a393b1d60c044894c9fd9ba3cd99c595f4e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:62XTJQZ7MGNO7SIFZRRUMRWXRP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"GPTVQ: The Blessing of Dimensionality for LLM Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Andrey Kuzmin, Cedric Bastoul, Eric Mahurin, Ivan Koryakovskiy, Markus Nagel, Mart van Baalen, Paul Whatmough, Peter Couperus, Tijmen Blankevoort","submitted_at":"2024-02-23T13:39:16Z","abstract_excerpt":"In this work we show that the size versus accuracy trade-off of neural network quantization can be significantly improved by increasing the quantization dimensionality. We propose the GPTVQ method, a new fast method for post-training vector quantization (VQ) that scales well to Large Language Models (LLMs). Our method interleaves quantization of one or more columns with updates to the remaining unquantized weights, using information from the Hessian of the per-layer output reconstruction MSE. Quantization codebooks are initialized using an efficient data-aware version of the EM algorithm. The "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15319","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/2402.15319/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":"2026-07-05T11:14:46Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eEs7KZCxLBAq6UZ/V4yGhCurjS8CZNZ+854OAwisrGl5ZqFhHiouboyn29PiBYjsJRvPdspXXiYpTSUAeO5PBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T14:49:24.786325Z"},"content_sha256":"893ce230b96bc7af16fa5faf4131c4ec578e1ca7c3c4698263937e7e65f60b2a","schema_version":"1.0","event_id":"sha256:893ce230b96bc7af16fa5faf4131c4ec578e1ca7c3c4698263937e7e65f60b2a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/bundle.json","state_url":"https://pith.science/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/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":"2026-07-07T14:49:24Z","links":{"resolver":"https://pith.science/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP","bundle":"https://pith.science/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/bundle.json","state":"https://pith.science/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/62XTJQZ7MGNO7SIFZRRUMRWXRP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:62XTJQZ7MGNO7SIFZRRUMRWXRP","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":"0b9bdc31a5a2634a475fa018fee3d3093e3e516e524b03edd1651a3e918bce78","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T13:39:16Z","title_canon_sha256":"bedb2ac721cc3f4b537071af89d8c26d4ce72df698bf8302c3e0486b7eafb6a9"},"schema_version":"1.0","source":{"id":"2402.15319","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.15319","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"arxiv_version","alias_value":"2402.15319v2","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15319","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_12","alias_value":"62XTJQZ7MGNO","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_16","alias_value":"62XTJQZ7MGNO7SIF","created_at":"2026-07-05T11:14:46Z"},{"alias_kind":"pith_short_8","alias_value":"62XTJQZ7","created_at":"2026-07-05T11:14:46Z"}],"graph_snapshots":[{"event_id":"sha256:893ce230b96bc7af16fa5faf4131c4ec578e1ca7c3c4698263937e7e65f60b2a","target":"graph","created_at":"2026-07-05T11:14:46Z","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/2402.15319/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In this work we show that the size versus accuracy trade-off of neural network quantization can be significantly improved by increasing the quantization dimensionality. We propose the GPTVQ method, a new fast method for post-training vector quantization (VQ) that scales well to Large Language Models (LLMs). Our method interleaves quantization of one or more columns with updates to the remaining unquantized weights, using information from the Hessian of the per-layer output reconstruction MSE. Quantization codebooks are initialized using an efficient data-aware version of the EM algorithm. The ","authors_text":"Andrey Kuzmin, Cedric Bastoul, Eric Mahurin, Ivan Koryakovskiy, Markus Nagel, Mart van Baalen, Paul Whatmough, Peter Couperus, Tijmen Blankevoort","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T13:39:16Z","title":"GPTVQ: The Blessing of Dimensionality for LLM Quantization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15319","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":"sha256:5860b49f1d421e713c3caedbcc0f4a393b1d60c044894c9fd9ba3cd99c595f4e","target":"record","created_at":"2026-07-05T11:14:46Z","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":"0b9bdc31a5a2634a475fa018fee3d3093e3e516e524b03edd1651a3e918bce78","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-23T13:39:16Z","title_canon_sha256":"bedb2ac721cc3f4b537071af89d8c26d4ce72df698bf8302c3e0486b7eafb6a9"},"schema_version":"1.0","source":{"id":"2402.15319","kind":"arxiv","version":2}},"canonical_sha256":"f6af34c33f619aefc905cc634646d78bcfd5d824cefc0f0f66eda653fec2c569","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f6af34c33f619aefc905cc634646d78bcfd5d824cefc0f0f66eda653fec2c569","first_computed_at":"2026-07-05T11:14:46.748358Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:14:46.748358Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YaH2jKyXQPlZLZ2khZ4ebdWRiwujCP0+jWiIY//OpmHeqd+bPbh5K+60sh1uZxc/7eciM0Wek+SiKndLU61ZBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:14:46.748820Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.15319","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5860b49f1d421e713c3caedbcc0f4a393b1d60c044894c9fd9ba3cd99c595f4e","sha256:893ce230b96bc7af16fa5faf4131c4ec578e1ca7c3c4698263937e7e65f60b2a"],"state_sha256":"9e1c524a8e9f587f1f9a50fea14c9c79770a07797bb2148eb1a9efdbc815270b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Vr4DQKxP8BEufmQhvPhtjZ+6et6kOJmicFreD9Po8c8Mu7L8ozaxxWxH280M+V+0r9eY1uhI7bY5W2mR5j3VAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T14:49:24.788214Z","bundle_sha256":"8e9caa5f848263c7bcaddb1da7a003916a68df86dcb9e9f3759e0ff389d33445"}}