{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VC25QYOI33WIDEDUDZTSENNAUQ","short_pith_number":"pith:VC25QYOI","schema_version":"1.0","canonical_sha256":"a8b5d861c8deec8190741e672235a0a42c9d8fc5e32a9c3f5b5531c090f89e69","source":{"kind":"arxiv","id":"2407.03211","version":2},"attestation_state":"computed","paper":{"title":"How Does Quantization Affect Multilingual LLMs?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ahmet \\\"Ust\\\"un, Dennis Aumiller, Hongyu Chen, Kelly Marchisio, Sara Hooker, Saurabh Dash, Sebastian Ruder","submitted_at":"2024-07-03T15:39:40Z","abstract_excerpt":"Quantization techniques are widely used to improve inference speed and deployment of large language models. While a wide body of work examines the impact of quantization on LLMs in English, none have evaluated across languages. We conduct a thorough analysis of quantized multilingual LLMs, focusing on performance across languages and at varying scales. We use automatic benchmarks, LLM-as-a-Judge, and human evaluation, finding that (1) harmful effects of quantization are apparent in human evaluation, which automatic metrics severely underestimate: a 1.7% average drop in Japanese across automati"},"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":"2407.03211","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-07-03T15:39:40Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6998260ee6fb070c4a13dc8edf9d8bed1ba4ce272424bb4515ff7bde0cf285e3","abstract_canon_sha256":"0d78ef58ef0bc24844d725742a094206d1bb78b30ae01134b7e0ea71a57c694f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:29.763212Z","signature_b64":"igOOMPc2lS2AALhwUoppzQtYvnjzRRSXDPPEtgy3YANWgYT9ZM2Gs01wmxkB5oBzjMoVnxv+Fr6kjiLFthayCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a8b5d861c8deec8190741e672235a0a42c9d8fc5e32a9c3f5b5531c090f89e69","last_reissued_at":"2026-07-05T09:19:29.762793Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:29.762793Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"How Does Quantization Affect Multilingual LLMs?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Ahmet \\\"Ust\\\"un, Dennis Aumiller, Hongyu Chen, Kelly Marchisio, Sara Hooker, Saurabh Dash, Sebastian Ruder","submitted_at":"2024-07-03T15:39:40Z","abstract_excerpt":"Quantization techniques are widely used to improve inference speed and deployment of large language models. While a wide body of work examines the impact of quantization on LLMs in English, none have evaluated across languages. We conduct a thorough analysis of quantized multilingual LLMs, focusing on performance across languages and at varying scales. We use automatic benchmarks, LLM-as-a-Judge, and human evaluation, finding that (1) harmful effects of quantization are apparent in human evaluation, which automatic metrics severely underestimate: a 1.7% average drop in Japanese across automati"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.03211","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/2407.03211/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2407.03211","created_at":"2026-07-05T09:19:29.762847+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.03211v2","created_at":"2026-07-05T09:19:29.762847+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.03211","created_at":"2026-07-05T09:19:29.762847+00:00"},{"alias_kind":"pith_short_12","alias_value":"VC25QYOI33WI","created_at":"2026-07-05T09:19:29.762847+00:00"},{"alias_kind":"pith_short_16","alias_value":"VC25QYOI33WIDEDU","created_at":"2026-07-05T09:19:29.762847+00:00"},{"alias_kind":"pith_short_8","alias_value":"VC25QYOI","created_at":"2026-07-05T09:19:29.762847+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.03330","citing_title":"FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15208","citing_title":"Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ","json":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ.json","graph_json":"https://pith.science/api/pith-number/VC25QYOI33WIDEDUDZTSENNAUQ/graph.json","events_json":"https://pith.science/api/pith-number/VC25QYOI33WIDEDUDZTSENNAUQ/events.json","paper":"https://pith.science/paper/VC25QYOI"},"agent_actions":{"view_html":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ","download_json":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ.json","view_paper":"https://pith.science/paper/VC25QYOI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.03211&json=true","fetch_graph":"https://pith.science/api/pith-number/VC25QYOI33WIDEDUDZTSENNAUQ/graph.json","fetch_events":"https://pith.science/api/pith-number/VC25QYOI33WIDEDUDZTSENNAUQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ/action/storage_attestation","attest_author":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ/action/author_attestation","sign_citation":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ/action/citation_signature","submit_replication":"https://pith.science/pith/VC25QYOI33WIDEDUDZTSENNAUQ/action/replication_record"}},"created_at":"2026-07-05T09:19:29.762847+00:00","updated_at":"2026-07-05T09:19:29.762847+00:00"}