{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QIBI5TWXYZOOLLPBSOSMDIRXH2","short_pith_number":"pith:QIBI5TWX","schema_version":"1.0","canonical_sha256":"82028eced7c65ce5ade193a4c1a2373eacb244874bec2b02ef2d1c070bfd2a9a","source":{"kind":"arxiv","id":"2405.15025","version":2},"attestation_state":"computed","paper":{"title":"OAC: Output-adaptive Calibration for Accurate Post-training Quantization","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ali Edalati, Alireza Ghaffari, Boxing Chen, Lu Hou, Mahsa Ghazvini Nejad, Masoud Asgharian, Vahid Partovi Nia","submitted_at":"2024-05-23T20:01:17Z","abstract_excerpt":"Deployment of Large Language Models (LLMs) has major computational costs, due to their rapidly expanding size. Compression of LLMs reduces the memory footprint, latency, and energy required for their inference. Post-training Quantization (PTQ) techniques have been developed to compress LLMs while avoiding expensive re-training. Most PTQ approaches formulate the quantization error based on a layer-wise Euclidean loss, ignoring the model output. Then, each layer is calibrated using its layer-wise Hessian to update the weights towards minimizing the quantization error. The Hessian is also used fo"},"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":"2405.15025","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-23T20:01:17Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"43a2224253c45766f40454744e646231dc77dc80c6bca56051a02d079f3ed931","abstract_canon_sha256":"6e3b7e7427407c2f6cbcb5813b5a982a501ef04319734c3155a7a8fde9f8121d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:58:45.240588Z","signature_b64":"FqqadZzKKmNS5xrswAtCfeS/dDuygW6yWVJH4SYy1HmABpQG+cMLb8W4nnw3A+WyvP1EUKoOsq92bqb2/Gw7CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"82028eced7c65ce5ade193a4c1a2373eacb244874bec2b02ef2d1c070bfd2a9a","last_reissued_at":"2026-07-05T10:58:45.240167Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:58:45.240167Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OAC: Output-adaptive Calibration for Accurate Post-training Quantization","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ali Edalati, Alireza Ghaffari, Boxing Chen, Lu Hou, Mahsa Ghazvini Nejad, Masoud Asgharian, Vahid Partovi Nia","submitted_at":"2024-05-23T20:01:17Z","abstract_excerpt":"Deployment of Large Language Models (LLMs) has major computational costs, due to their rapidly expanding size. Compression of LLMs reduces the memory footprint, latency, and energy required for their inference. Post-training Quantization (PTQ) techniques have been developed to compress LLMs while avoiding expensive re-training. Most PTQ approaches formulate the quantization error based on a layer-wise Euclidean loss, ignoring the model output. Then, each layer is calibrated using its layer-wise Hessian to update the weights towards minimizing the quantization error. The Hessian is also used fo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.15025","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/2405.15025/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":"2405.15025","created_at":"2026-07-05T10:58:45.240224+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.15025v2","created_at":"2026-07-05T10:58:45.240224+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.15025","created_at":"2026-07-05T10:58:45.240224+00:00"},{"alias_kind":"pith_short_12","alias_value":"QIBI5TWXYZOO","created_at":"2026-07-05T10:58:45.240224+00:00"},{"alias_kind":"pith_short_16","alias_value":"QIBI5TWXYZOOLLPB","created_at":"2026-07-05T10:58:45.240224+00:00"},{"alias_kind":"pith_short_8","alias_value":"QIBI5TWX","created_at":"2026-07-05T10:58:45.240224+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2","json":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2.json","graph_json":"https://pith.science/api/pith-number/QIBI5TWXYZOOLLPBSOSMDIRXH2/graph.json","events_json":"https://pith.science/api/pith-number/QIBI5TWXYZOOLLPBSOSMDIRXH2/events.json","paper":"https://pith.science/paper/QIBI5TWX"},"agent_actions":{"view_html":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2","download_json":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2.json","view_paper":"https://pith.science/paper/QIBI5TWX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.15025&json=true","fetch_graph":"https://pith.science/api/pith-number/QIBI5TWXYZOOLLPBSOSMDIRXH2/graph.json","fetch_events":"https://pith.science/api/pith-number/QIBI5TWXYZOOLLPBSOSMDIRXH2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2/action/storage_attestation","attest_author":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2/action/author_attestation","sign_citation":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2/action/citation_signature","submit_replication":"https://pith.science/pith/QIBI5TWXYZOOLLPBSOSMDIRXH2/action/replication_record"}},"created_at":"2026-07-05T10:58:45.240224+00:00","updated_at":"2026-07-05T10:58:45.240224+00:00"}