{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AKBT3LNKPOX7Z3OVM34OS6MTFT","short_pith_number":"pith:AKBT3LNK","schema_version":"1.0","canonical_sha256":"02833dadaa7baffcedd566f8e979932cf6076fd6cc06ae829c9a4e043b3b8ca3","source":{"kind":"arxiv","id":"2507.18553","version":4},"attestation_state":"computed","paper":{"title":"The Geometry of LLM Quantization: GPTQ as Babai's Nearest Plane Algorithm","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DS","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Dan Alistarh, Elvir Crn\\v{c}evi\\'c, Jiale Chen, Torsten Hoefler, Yalda Shabanzadeh","submitted_at":"2025-07-24T16:22:18Z","abstract_excerpt":"Quantizing the weights of large language models (LLMs) from 16-bit to lower bitwidth is the de facto approach to deploy massive transformers onto more affordable accelerators. While GPTQ emerged as one of the standard methods for one-shot post-training quantization at LLM scale, its inner workings are described as a sequence of algebraic updates that obscure geometric meaning or worst-case guarantees. In this work, we show that, when executed back-to-front (from the last to first dimension) for a linear layer, GPTQ is mathematically identical to Babai's nearest plane algorithm for the classica"},"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":"2507.18553","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-24T16:22:18Z","cross_cats_sorted":["cs.DS","cs.IT","math.IT"],"title_canon_sha256":"3c2a6661aa336e8f0810828f71942ffce9f1f08ae6aafdb6ced9ee1a5168a10d","abstract_canon_sha256":"42dd6193a44b913ede3289344994f7bf42fdcf8254bf1bba72a8be78bfaff252"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:45:13.341076Z","signature_b64":"NRNiRIrPa6pTrq5E7ZuEPi+hAXbz2vzQHcybcNdDLIbCCcTlb+8WI8jDCR+oTdxCE5kue3VyG7aJLZZ9zgQRDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"02833dadaa7baffcedd566f8e979932cf6076fd6cc06ae829c9a4e043b3b8ca3","last_reissued_at":"2026-05-18T02:45:13.340581Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:45:13.340581Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Geometry of LLM Quantization: GPTQ as Babai's Nearest Plane Algorithm","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.DS","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Dan Alistarh, Elvir Crn\\v{c}evi\\'c, Jiale Chen, Torsten Hoefler, Yalda Shabanzadeh","submitted_at":"2025-07-24T16:22:18Z","abstract_excerpt":"Quantizing the weights of large language models (LLMs) from 16-bit to lower bitwidth is the de facto approach to deploy massive transformers onto more affordable accelerators. While GPTQ emerged as one of the standard methods for one-shot post-training quantization at LLM scale, its inner workings are described as a sequence of algebraic updates that obscure geometric meaning or worst-case guarantees. In this work, we show that, when executed back-to-front (from the last to first dimension) for a linear layer, GPTQ is mathematically identical to Babai's nearest plane algorithm for the classica"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.18553","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2507.18553","created_at":"2026-05-18T02:45:13.340665+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.18553v4","created_at":"2026-05-18T02:45:13.340665+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.18553","created_at":"2026-05-18T02:45:13.340665+00:00"},{"alias_kind":"pith_short_12","alias_value":"AKBT3LNKPOX7","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"AKBT3LNKPOX7Z3OV","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"AKBT3LNK","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2601.17187","citing_title":"High-Rate Quantized Matrix Multiplication I","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2602.05902","citing_title":"CoreQ: Learning-Free Mismatch Correction and Successive Rounding for Quantization","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.13768","citing_title":"High-Rate Quantized Matrix Multiplication II","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT","json":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT.json","graph_json":"https://pith.science/api/pith-number/AKBT3LNKPOX7Z3OVM34OS6MTFT/graph.json","events_json":"https://pith.science/api/pith-number/AKBT3LNKPOX7Z3OVM34OS6MTFT/events.json","paper":"https://pith.science/paper/AKBT3LNK"},"agent_actions":{"view_html":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT","download_json":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT.json","view_paper":"https://pith.science/paper/AKBT3LNK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.18553&json=true","fetch_graph":"https://pith.science/api/pith-number/AKBT3LNKPOX7Z3OVM34OS6MTFT/graph.json","fetch_events":"https://pith.science/api/pith-number/AKBT3LNKPOX7Z3OVM34OS6MTFT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT/action/storage_attestation","attest_author":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT/action/author_attestation","sign_citation":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT/action/citation_signature","submit_replication":"https://pith.science/pith/AKBT3LNKPOX7Z3OVM34OS6MTFT/action/replication_record"}},"created_at":"2026-05-18T02:45:13.340665+00:00","updated_at":"2026-05-18T02:45:13.340665+00:00"}