{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:WUAL7SQNPFEY5IZMFCR67UYHCN","short_pith_number":"pith:WUAL7SQN","schema_version":"1.0","canonical_sha256":"b500bfca0d79498ea32c28a3efd307136c4fd7740a6f2758354d87f6e8aec87d","source":{"kind":"arxiv","id":"2304.09145","version":3},"attestation_state":"computed","paper":{"title":"Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jinyang Guo, Ruihao Gong, Xiangguo Zhang, Xianglong Liu, Xiuying Wei, Yuhang Li, Yunchen Zhang","submitted_at":"2023-04-18T17:34:23Z","abstract_excerpt":"Post-training quantization~(PTQ) of transformer language models faces significant challenges due to the existence of detrimental outliers in activations. We observe that these outliers are concentrated in specific channels and are asymmetric across channels. To address this issue, we propose the Outlier Suppression+~(OS+) framework, which contains the channel-wise shifting for asymmetry and channel-wise scaling for concentration. We show that these operations can be seamlessly migrated into subsequent modules while maintaining equivalence. Second, we propose a fast and stable scheme to calcula"},"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":"2304.09145","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-04-18T17:34:23Z","cross_cats_sorted":[],"title_canon_sha256":"c14947fd0a285186ef6cea6a5a0300a19605c90b8c7ad2d53aa9164f36b3ecbf","abstract_canon_sha256":"36e35cd6ea25776d4592b78b3e6688393aac51eb5d385b308e1b31247c2264af"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:03:23.338449Z","signature_b64":"927OjodblZsU7ZdC5x83aj3J0asRFbxqb/PRN2BoQr332prsK0UwRSxa7pC5CMajr1WlSV151gvtPyHnyX2LBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b500bfca0d79498ea32c28a3efd307136c4fd7740a6f2758354d87f6e8aec87d","last_reissued_at":"2026-07-05T07:03:23.337978Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:03:23.337978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Outlier Suppression+: Accurate quantization of large language models by equivalent and optimal shifting and scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jinyang Guo, Ruihao Gong, Xiangguo Zhang, Xianglong Liu, Xiuying Wei, Yuhang Li, Yunchen Zhang","submitted_at":"2023-04-18T17:34:23Z","abstract_excerpt":"Post-training quantization~(PTQ) of transformer language models faces significant challenges due to the existence of detrimental outliers in activations. We observe that these outliers are concentrated in specific channels and are asymmetric across channels. To address this issue, we propose the Outlier Suppression+~(OS+) framework, which contains the channel-wise shifting for asymmetry and channel-wise scaling for concentration. We show that these operations can be seamlessly migrated into subsequent modules while maintaining equivalence. Second, we propose a fast and stable scheme to calcula"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.09145","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/2304.09145/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":"2304.09145","created_at":"2026-07-05T07:03:23.338032+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.09145v3","created_at":"2026-07-05T07:03:23.338032+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.09145","created_at":"2026-07-05T07:03:23.338032+00:00"},{"alias_kind":"pith_short_12","alias_value":"WUAL7SQNPFEY","created_at":"2026-07-05T07:03:23.338032+00:00"},{"alias_kind":"pith_short_16","alias_value":"WUAL7SQNPFEY5IZM","created_at":"2026-07-05T07:03:23.338032+00:00"},{"alias_kind":"pith_short_8","alias_value":"WUAL7SQN","created_at":"2026-07-05T07:03:23.338032+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.00539","citing_title":"GNMR: Runtime Stability Control for Low-Precision Large Language Model Training","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20295","citing_title":"Quant.npu: Enabling Efficient Mobile NPU Inference for on-device LLMs via Fully Static Quantization","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2307.06435","citing_title":"A Comprehensive Overview of Large Language Models","ref_index":257,"is_internal_anchor":false},{"citing_arxiv_id":"2405.16406","citing_title":"SpinQuant: LLM quantization with learned rotations","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2404.14294","citing_title":"A Survey on Efficient Inference for Large Language Models","ref_index":169,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22906","citing_title":"Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities","ref_index":163,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19167","citing_title":"LBLLM: Lightweight Binarization of Large Language Models via Three-Stage Distillation","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN","json":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN.json","graph_json":"https://pith.science/api/pith-number/WUAL7SQNPFEY5IZMFCR67UYHCN/graph.json","events_json":"https://pith.science/api/pith-number/WUAL7SQNPFEY5IZMFCR67UYHCN/events.json","paper":"https://pith.science/paper/WUAL7SQN"},"agent_actions":{"view_html":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN","download_json":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN.json","view_paper":"https://pith.science/paper/WUAL7SQN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.09145&json=true","fetch_graph":"https://pith.science/api/pith-number/WUAL7SQNPFEY5IZMFCR67UYHCN/graph.json","fetch_events":"https://pith.science/api/pith-number/WUAL7SQNPFEY5IZMFCR67UYHCN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN/action/storage_attestation","attest_author":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN/action/author_attestation","sign_citation":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN/action/citation_signature","submit_replication":"https://pith.science/pith/WUAL7SQNPFEY5IZMFCR67UYHCN/action/replication_record"}},"created_at":"2026-07-05T07:03:23.338032+00:00","updated_at":"2026-07-05T07:03:23.338032+00:00"}