{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RCOTM2AMTYM3ROICMKY5X4HM37","short_pith_number":"pith:RCOTM2AM","schema_version":"1.0","canonical_sha256":"889d36680c9e19b8b90262b1dbf0ecdfcbc3c45362105a7dd870cdccbffa6ea8","source":{"kind":"arxiv","id":"2411.18065","version":2},"attestation_state":"computed","paper":{"title":"FlexiBit: Fully Flexible Precision Bit-parallel Accelerator Architecture for Arbitrary Mixed Precision AI","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Benji Y.H. Huang, Faraz Tahmasebi, Hyoukjun Kwon, Yian Wang","submitted_at":"2024-11-27T05:16:26Z","abstract_excerpt":"Recent research has shown that large language models (LLMs) can utilize low-precision floating point (FP) quantization to deliver high efficiency while maintaining original model accuracy. In particular, recent works have shown the effectiveness of non-power-of-two precisions, such as FP6 and FP5, and diverse sensitivity to low-precision arithmetic of LLM layers, which motivates mixed precision arithmetic including non-power-of-two precisions in LLMs. Although low-precision algorithmically leads to low computational overheads, such benefits cannot be fully exploited due to hardware constraints"},"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":"2411.18065","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AR","submitted_at":"2024-11-27T05:16:26Z","cross_cats_sorted":[],"title_canon_sha256":"dd9df5d4aae7dfdb09e4a151030f8dae9aab56343615b6da0f4b2703607c1777","abstract_canon_sha256":"fbcfd5acc3ee3d8b53bb668b97af080e28639a32a7ae625f7bc345a646045ee5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:15:18.483414Z","signature_b64":"qYYRqiGTkg7Y0YxnpOPF0AENekiiaWCJVPJ2aWYHo2klUYvp084pBhDfOxDtDAFWuB341zmMW40GWThkNDlUCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"889d36680c9e19b8b90262b1dbf0ecdfcbc3c45362105a7dd870cdccbffa6ea8","last_reissued_at":"2026-07-05T11:15:18.482990Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:15:18.482990Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FlexiBit: Fully Flexible Precision Bit-parallel Accelerator Architecture for Arbitrary Mixed Precision AI","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AR","authors_text":"Benji Y.H. Huang, Faraz Tahmasebi, Hyoukjun Kwon, Yian Wang","submitted_at":"2024-11-27T05:16:26Z","abstract_excerpt":"Recent research has shown that large language models (LLMs) can utilize low-precision floating point (FP) quantization to deliver high efficiency while maintaining original model accuracy. In particular, recent works have shown the effectiveness of non-power-of-two precisions, such as FP6 and FP5, and diverse sensitivity to low-precision arithmetic of LLM layers, which motivates mixed precision arithmetic including non-power-of-two precisions in LLMs. Although low-precision algorithmically leads to low computational overheads, such benefits cannot be fully exploited due to hardware constraints"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.18065","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/2411.18065/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":"2411.18065","created_at":"2026-07-05T11:15:18.483046+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.18065v2","created_at":"2026-07-05T11:15:18.483046+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.18065","created_at":"2026-07-05T11:15:18.483046+00:00"},{"alias_kind":"pith_short_12","alias_value":"RCOTM2AMTYM3","created_at":"2026-07-05T11:15:18.483046+00:00"},{"alias_kind":"pith_short_16","alias_value":"RCOTM2AMTYM3ROIC","created_at":"2026-07-05T11:15:18.483046+00:00"},{"alias_kind":"pith_short_8","alias_value":"RCOTM2AM","created_at":"2026-07-05T11:15:18.483046+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.06052","citing_title":"XtraMAC: An Efficient MAC Architecture for Mixed-Precision LLM Inference on FPGA","ref_index":34,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37","json":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37.json","graph_json":"https://pith.science/api/pith-number/RCOTM2AMTYM3ROICMKY5X4HM37/graph.json","events_json":"https://pith.science/api/pith-number/RCOTM2AMTYM3ROICMKY5X4HM37/events.json","paper":"https://pith.science/paper/RCOTM2AM"},"agent_actions":{"view_html":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37","download_json":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37.json","view_paper":"https://pith.science/paper/RCOTM2AM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.18065&json=true","fetch_graph":"https://pith.science/api/pith-number/RCOTM2AMTYM3ROICMKY5X4HM37/graph.json","fetch_events":"https://pith.science/api/pith-number/RCOTM2AMTYM3ROICMKY5X4HM37/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37/action/storage_attestation","attest_author":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37/action/author_attestation","sign_citation":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37/action/citation_signature","submit_replication":"https://pith.science/pith/RCOTM2AMTYM3ROICMKY5X4HM37/action/replication_record"}},"created_at":"2026-07-05T11:15:18.483046+00:00","updated_at":"2026-07-05T11:15:18.483046+00:00"}