{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:FNPD4PXGU4BFCGWCADRUSB7GTJ","short_pith_number":"pith:FNPD4PXG","schema_version":"1.0","canonical_sha256":"2b5e3e3ee6a702511ac200e34907e69a7fedfad8892125210de0673b00108196","source":{"kind":"arxiv","id":"2402.17764","version":1},"attestation_state":"computed","paper":{"title":"The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Ternary-weight LLMs achieve full-precision performance at far lower computational cost","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Furu Wei, Hongyu Wang, Jilong Xue, Lei Wang, Li Dong, Lingxiao Ma, Ruiping Wang, Shaohan Huang, Shuming Ma, Wenhui Wang","submitted_at":"2024-02-27T18:56:19Z","abstract_excerpt":"Recent research, such as BitNet, is paving the way for a new era of 1-bit Large Language Models (LLMs). In this work, we introduce a 1-bit LLM variant, namely BitNet b1.58, in which every single parameter (or weight) of the LLM is ternary {-1, 0, 1}. It matches the full-precision (i.e., FP16 or BF16) Transformer LLM with the same model size and training tokens in terms of both perplexity and end-task performance, while being significantly more cost-effective in terms of latency, memory, throughput, and energy consumption. More profoundly, the 1.58-bit LLM defines a new scaling law and recipe f"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2402.17764","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-02-27T18:56:19Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"07386fb3bc4c5e218ecb5dcf1df4250ca8801e5ca786069feec96a2427d82d23","abstract_canon_sha256":"426827275ed354a9f956802377eb8139613353790a25470518f41f878e208786"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:13.211486Z","signature_b64":"VY9IE30jHVjpx1rjd2udEpfXgJwjQgDRvRHozFAsvcHCnnzb/fwKJcAozR2hAjRCBhJmbw67xBFzh4QPyzw7Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2b5e3e3ee6a702511ac200e34907e69a7fedfad8892125210de0673b00108196","last_reissued_at":"2026-05-17T23:38:13.210836Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:13.210836Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Era of 1-bit LLMs: All Large Language Models are in 1.58 Bits","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Ternary-weight LLMs achieve full-precision performance at far lower computational cost","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Furu Wei, Hongyu Wang, Jilong Xue, Lei Wang, Li Dong, Lingxiao Ma, Ruiping Wang, Shaohan Huang, Shuming Ma, Wenhui Wang","submitted_at":"2024-02-27T18:56:19Z","abstract_excerpt":"Recent research, such as BitNet, is paving the way for a new era of 1-bit Large Language Models (LLMs). In this work, we introduce a 1-bit LLM variant, namely BitNet b1.58, in which every single parameter (or weight) of the LLM is ternary {-1, 0, 1}. It matches the full-precision (i.e., FP16 or BF16) Transformer LLM with the same model size and training tokens in terms of both perplexity and end-task performance, while being significantly more cost-effective in terms of latency, memory, throughput, and energy consumption. More profoundly, the 1.58-bit LLM defines a new scaling law and recipe f"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"It matches the full-precision (i.e., FP16 or BF16) Transformer LLM with the same model size and training tokens in terms of both perplexity and end-task performance, while being significantly more cost-effective in terms of latency, memory, throughput, and energy consumption.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the training procedure and scaling law developed for the 1.58-bit ternary setting will continue to produce competitive performance when model size or data volume increases beyond the scales tested.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"BitNet b1.58 shows that ternary 1.58-bit LLMs can match full-precision performance at substantially lower inference cost.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Ternary-weight LLMs achieve full-precision performance at far lower computational cost","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ddc322bdb29f2bfbbc4fa69825a4f1f1d3674a93e814ceed7cf341ae4361abfb"},"source":{"id":"2402.17764","kind":"arxiv","version":1},"verdict":{"id":"133e1b69-a422-44d1-aadb-0eb10183f4ad","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T20:06:27.706475Z","strongest_claim":"It matches the full-precision (i.e., FP16 or BF16) Transformer LLM with the same model size and training tokens in terms of both perplexity and end-task performance, while being significantly more cost-effective in terms of latency, memory, throughput, and energy consumption.","one_line_summary":"BitNet b1.58 shows that ternary 1.58-bit LLMs can match full-precision performance at substantially lower inference cost.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the training procedure and scaling law developed for the 1.58-bit ternary setting will continue to produce competitive performance when model size or data volume increases beyond the scales tested.","pith_extraction_headline":"Ternary-weight LLMs achieve full-precision performance at far lower computational cost"},"references":{"count":15,"sample":[{"doi":"","year":1911,"title":"PIQA: Reasoning about Physical Commonsense in Natural Language","work_id":"0d865a62-6376-4606-8d3a-eeb3b6e9ba6d","ref_index":1,"cited_arxiv_id":"1911.11641","is_internal_anchor":true},{"doi":"","year":null,"title":"arXiv preprint arXiv:2307.13304 , year=","work_id":"bd3fe3b4-ccc3-419b-969e-9a80ded56858","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1905,"title":"BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions","work_id":"511eeb84-4b95-46d5-b14f-50da43f4f19f","ref_index":3,"cited_arxiv_id":"1905.10044","is_internal_anchor":true},{"doi":"","year":2014,"title":"1.1 computing’s energy problem (and what we can do about it)","work_id":"85c69f95-61c6-46d7-a2cd-7076728504f6","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration","work_id":"ea9d1d72-db24-4cae-8c89-4ecd83dd87c1","ref_index":5,"cited_arxiv_id":"2306.00978","is_internal_anchor":true}],"resolved_work":15,"snapshot_sha256":"e30fcd93a9d7cb80a19c153ea14ecd4c6b57ef294d43ba70ac52e8d0802b8538","internal_anchors":9},"formal_canon":{"evidence_count":1,"snapshot_sha256":"c821ee11b34325432ec0696a2de201f332054e83b69091e7efddad8e94353a8e"},"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":"2402.17764","created_at":"2026-05-17T23:38:13.210947+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.17764v1","created_at":"2026-05-17T23:38:13.210947+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.17764","created_at":"2026-05-17T23:38:13.210947+00:00"},{"alias_kind":"pith_short_12","alias_value":"FNPD4PXGU4BF","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"FNPD4PXGU4BFCGWC","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"FNPD4PXG","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":18,"internal_anchor_count":18,"sample":[{"citing_arxiv_id":"2510.23912","citing_title":"Key and Value Weights Are Probably All You Need: On the Necessity of the Query, Key, Value weight Triplet in Self-Attention Transformers","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2512.06443","citing_title":"Vec-LUT: Vector Table Lookup for Parallel Ultra-Low-Bit LLM Inference on Edge Devices","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.13538","citing_title":"Locale-Conditioned Few-Shot Prompting Mitigates Demonstration Regurgitation in On-Device PII Substitution with Small Language Models","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03336","citing_title":"NativeTernary: A Self-Delimiting Binary Encoding with Unary Run-Length Hierarchy Markers for Ternary Neural Network Weights, Structured Data, and General Computing Infrastructure","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11558","citing_title":"A Composite Activation Function for Learning Stable Binary Representations","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2604.27396","citing_title":"VitaLLM: A Versatile, Ultra-Compact Ternary LLM Accelerator with Dependency-Aware Scheduling","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08894","citing_title":"Fitting Is Not Enough: Smoothness in Extremely Quantized LLMs","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06485","citing_title":"Litespark Inference on Consumer CPUs: Custom SIMD Kernels for Ternary Neural Networks","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00320","citing_title":"VitaLLM: A Versatile and Tiny Accelerator for Mixed-Precision LLM Inference on Edge Devices","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2604.20913","citing_title":"FairyFuse: Multiplication-Free LLM Inference on CPUs via Fused Ternary Kernels","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2604.10272","citing_title":"The Phase Is the Gradient: Equilibrium Propagation for Frequency Learning in Kuramoto Networks","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2604.20333","citing_title":"Quantization robustness from dense representations of sparse functions in high-capacity kernel associative memory","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2604.06836","citing_title":"STQuant: Spatio-Temporal Adaptive Framework for Optimizer Quantization in Large Multimodal Model Training","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2604.13806","citing_title":"Robust Ultra Low-Bit Post-Training Quantization via Stable Diagonal Curvature Estimate","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18976","citing_title":"STAR-Teaming: A Strategy-Response Multiplex Network Approach to Automated LLM Red Teaming","ref_index":44,"is_internal_anchor":true},{"citing_arxiv_id":"2604.21026","citing_title":"MCAP: Deployment-Time Layer Profiling for Memory-Constrained LLM Inference","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2605.02853","citing_title":"Trust, but Verify: Peeling Low-Bit Transformer Networks for Training Monitoring","ref_index":26,"is_internal_anchor":true},{"citing_arxiv_id":"2604.20333","citing_title":"Quantization robustness from dense representations of sparse functions in high-capacity kernel associative memory","ref_index":4,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ","json":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ.json","graph_json":"https://pith.science/api/pith-number/FNPD4PXGU4BFCGWCADRUSB7GTJ/graph.json","events_json":"https://pith.science/api/pith-number/FNPD4PXGU4BFCGWCADRUSB7GTJ/events.json","paper":"https://pith.science/paper/FNPD4PXG"},"agent_actions":{"view_html":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ","download_json":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ.json","view_paper":"https://pith.science/paper/FNPD4PXG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.17764&json=true","fetch_graph":"https://pith.science/api/pith-number/FNPD4PXGU4BFCGWCADRUSB7GTJ/graph.json","fetch_events":"https://pith.science/api/pith-number/FNPD4PXGU4BFCGWCADRUSB7GTJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ/action/storage_attestation","attest_author":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ/action/author_attestation","sign_citation":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ/action/citation_signature","submit_replication":"https://pith.science/pith/FNPD4PXGU4BFCGWCADRUSB7GTJ/action/replication_record"}},"created_at":"2026-05-17T23:38:13.210947+00:00","updated_at":"2026-05-17T23:38:13.210947+00:00"}