{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ELFPNCT27ODWVCV473MG7EPT3Z","short_pith_number":"pith:ELFPNCT2","schema_version":"1.0","canonical_sha256":"22caf68a7afb876a8abcfed86f91f3de54a23bc463f7e858447938551a8112ab","source":{"kind":"arxiv","id":"2601.07475","version":2},"attestation_state":"computed","paper":{"title":"ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoqian Meng, Peng Zhang, Wenyuan Liu, Xindian Ma, Yafei Zhao, Yilun Luo","submitted_at":"2026-01-12T12:27:22Z","abstract_excerpt":"The emergence of fine-grained numerical formats like NVFP4 presents new opportunities for efficient Large Language Model (LLM) inference. However, it is difficult to adapt existing Post-Training Quantization (PTQ) strategies to these formats: rotation-based methods compromise fine-grained block isolation; smoothing techniques struggle with significant 4-bit quantization errors; and mixed-precision approaches often conflict with hardware constraints on unified-precision computation. To address these challenges, we propose ARCQuant, a framework that boosts NVFP4 performance via Augmented Residua"},"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":"2601.07475","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-01-12T12:27:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"15eac9e656bbe099838e79d0c7185c1ff86ca1ebd3e6188e2e12e093757a46ab","abstract_canon_sha256":"da300a283ad60d7384a38b61d8a50d5f20708814a1ee7eaa2c4fbea8e9bc5be8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:17:16.248450Z","signature_b64":"F8kWTQqVpkYyIeM+VKvmWcjWdhZofB/WDJMW9u4AKbJ/qObqNeEmcvN/nqPfD6imqxxGH7HIRX0nxboIymgrDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"22caf68a7afb876a8abcfed86f91f3de54a23bc463f7e858447938551a8112ab","last_reissued_at":"2026-07-07T02:17:16.247527Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:17:16.247527Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoqian Meng, Peng Zhang, Wenyuan Liu, Xindian Ma, Yafei Zhao, Yilun Luo","submitted_at":"2026-01-12T12:27:22Z","abstract_excerpt":"The emergence of fine-grained numerical formats like NVFP4 presents new opportunities for efficient Large Language Model (LLM) inference. However, it is difficult to adapt existing Post-Training Quantization (PTQ) strategies to these formats: rotation-based methods compromise fine-grained block isolation; smoothing techniques struggle with significant 4-bit quantization errors; and mixed-precision approaches often conflict with hardware constraints on unified-precision computation. To address these challenges, we propose ARCQuant, a framework that boosts NVFP4 performance via Augmented Residua"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.07475","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/2601.07475/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":"2601.07475","created_at":"2026-07-07T02:17:16.247669+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.07475v2","created_at":"2026-07-07T02:17:16.247669+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.07475","created_at":"2026-07-07T02:17:16.247669+00:00"},{"alias_kind":"pith_short_12","alias_value":"ELFPNCT27ODW","created_at":"2026-07-07T02:17:16.247669+00:00"},{"alias_kind":"pith_short_16","alias_value":"ELFPNCT27ODWVCV4","created_at":"2026-07-07T02:17:16.247669+00:00"},{"alias_kind":"pith_short_8","alias_value":"ELFPNCT2","created_at":"2026-07-07T02:17:16.247669+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2606.26587","citing_title":"SharQ: Bridging Activation Sparsity and FP4 Quantization for LLM Inference","ref_index":176,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12245","citing_title":"SOAR: Scale Optimization for Accurate Reconstruction in NVFP4 Quantization","ref_index":27,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z","json":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z.json","graph_json":"https://pith.science/api/pith-number/ELFPNCT27ODWVCV473MG7EPT3Z/graph.json","events_json":"https://pith.science/api/pith-number/ELFPNCT27ODWVCV473MG7EPT3Z/events.json","paper":"https://pith.science/paper/ELFPNCT2"},"agent_actions":{"view_html":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z","download_json":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z.json","view_paper":"https://pith.science/paper/ELFPNCT2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.07475&json=true","fetch_graph":"https://pith.science/api/pith-number/ELFPNCT27ODWVCV473MG7EPT3Z/graph.json","fetch_events":"https://pith.science/api/pith-number/ELFPNCT27ODWVCV473MG7EPT3Z/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z/action/storage_attestation","attest_author":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z/action/author_attestation","sign_citation":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z/action/citation_signature","submit_replication":"https://pith.science/pith/ELFPNCT27ODWVCV473MG7EPT3Z/action/replication_record"}},"created_at":"2026-07-07T02:17:16.247669+00:00","updated_at":"2026-07-07T02:17:16.247669+00:00"}