{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:HO6FZ33GFFR4QOIWLPT77IT2HM","short_pith_number":"pith:HO6FZ33G","schema_version":"1.0","canonical_sha256":"3bbc5cef662963c839165be7ffa27a3b3ccab4464e842c82bbb3fc0cf4bc8838","source":{"kind":"arxiv","id":"2606.26587","version":1},"attestation_state":"computed","paper":{"title":"SharQ: Bridging Activation Sparsity and FP4 Quantization for LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoqian Meng, Huaqing Zheng, Peng Zhang, Wenyuan Liu, Xindian Ma, Yafei Zhao, Yilun Luo","submitted_at":"2026-06-25T04:19:04Z","abstract_excerpt":"Low-bit floating-point formats and semi-structured sparsity are increasingly supported by modern accelerators, yet combining them for LLM activation compression remains challenging: activations contain input-dependent outliers that dominate block scales in FP4 quantization, and directly applying N:M sparsity masks discards moderate values, coupling sparsification loss with quantization error. We introduce SharQ, a training-free inference method that bridges activation sparsity and FP4 quantization through an online sparse--dense decomposition. For each activation tensor, SharQ generates an inp"},"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":"2606.26587","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-25T04:19:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"08d92f7fad7c721ae64273fb8e646ed488bb8ef758b56ea8b3778ce55137c0e0","abstract_canon_sha256":"98ad849ecb6007ed985c975ba26c5549e44633e5798c7c9076f83571539389e4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:15:35.647306Z","signature_b64":"2rLA9+UH6oSPXfkMrK7hgttID3BylGUPK4XlgxqctN0MIYEEGmK2spGl0ty5H2x80mun8Bc+YxhEbMOCkRQvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3bbc5cef662963c839165be7ffa27a3b3ccab4464e842c82bbb3fc0cf4bc8838","last_reissued_at":"2026-06-26T01:15:35.646933Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:15:35.646933Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SharQ: Bridging Activation Sparsity and FP4 Quantization for LLM Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Haoqian Meng, Huaqing Zheng, Peng Zhang, Wenyuan Liu, Xindian Ma, Yafei Zhao, Yilun Luo","submitted_at":"2026-06-25T04:19:04Z","abstract_excerpt":"Low-bit floating-point formats and semi-structured sparsity are increasingly supported by modern accelerators, yet combining them for LLM activation compression remains challenging: activations contain input-dependent outliers that dominate block scales in FP4 quantization, and directly applying N:M sparsity masks discards moderate values, coupling sparsification loss with quantization error. We introduce SharQ, a training-free inference method that bridges activation sparsity and FP4 quantization through an online sparse--dense decomposition. For each activation tensor, SharQ generates an inp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.26587","kind":"arxiv","version":1},"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/2606.26587/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":"2606.26587","created_at":"2026-06-26T01:15:35.646992+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.26587v1","created_at":"2026-06-26T01:15:35.646992+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.26587","created_at":"2026-06-26T01:15:35.646992+00:00"},{"alias_kind":"pith_short_12","alias_value":"HO6FZ33GFFR4","created_at":"2026-06-26T01:15:35.646992+00:00"},{"alias_kind":"pith_short_16","alias_value":"HO6FZ33GFFR4QOIW","created_at":"2026-06-26T01:15:35.646992+00:00"},{"alias_kind":"pith_short_8","alias_value":"HO6FZ33G","created_at":"2026-06-26T01:15:35.646992+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/HO6FZ33GFFR4QOIWLPT77IT2HM","json":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM.json","graph_json":"https://pith.science/api/pith-number/HO6FZ33GFFR4QOIWLPT77IT2HM/graph.json","events_json":"https://pith.science/api/pith-number/HO6FZ33GFFR4QOIWLPT77IT2HM/events.json","paper":"https://pith.science/paper/HO6FZ33G"},"agent_actions":{"view_html":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM","download_json":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM.json","view_paper":"https://pith.science/paper/HO6FZ33G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.26587&json=true","fetch_graph":"https://pith.science/api/pith-number/HO6FZ33GFFR4QOIWLPT77IT2HM/graph.json","fetch_events":"https://pith.science/api/pith-number/HO6FZ33GFFR4QOIWLPT77IT2HM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM/action/storage_attestation","attest_author":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM/action/author_attestation","sign_citation":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM/action/citation_signature","submit_replication":"https://pith.science/pith/HO6FZ33GFFR4QOIWLPT77IT2HM/action/replication_record"}},"created_at":"2026-06-26T01:15:35.646992+00:00","updated_at":"2026-06-26T01:15:35.646992+00:00"}