{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MUEYMPCGGHTXUWUFF4AEJPTEMQ","short_pith_number":"pith:MUEYMPCG","schema_version":"1.0","canonical_sha256":"6509863c4631e77a5a852f0044be646422fac310e9cebf6cf5e5de195320d5a8","source":{"kind":"arxiv","id":"2606.00079","version":1},"attestation_state":"computed","paper":{"title":"BitsMoE: Efficient Spectral Energy-Guided Bit Allocation for MoE LLM Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiayu Zhao, Minhao Fan, Song Chen, Tianrui Ma, Weichen Liu, Wentao Ren, Zihan Teng","submitted_at":"2026-05-22T13:05:53Z","abstract_excerpt":"Mixture-of-Experts (MoE) large language models reduce per-token computation through sparse expert activation, but their deployment remains memory-intensive because all expert weights must be kept resident in memory. Existing MoE compression methods struggle in the ultra-low-bit regime: pruning irreversibly removes model capacity, while coarse-grained quantization fails to allocate bits according to heterogeneous expert and weight-direction importance. We propose BitsMoE, a spectral-energy-guided bit-allocation framework for MoE LLM quantization. BitsMoE decomposes each MoE layer by SVD into a "},"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":"2606.00079","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-22T13:05:53Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6222859f76f4ffa7bc41e77a91e964871400800715d2645963b4a583f65af4c1","abstract_canon_sha256":"69d62bab87751658fa8616afa166180b1cb16274633fc46cedf4630877a3446b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T00:03:14.364835Z","signature_b64":"0pyQRLLPJJzi6Bi/1uIFruYnZFzbkBs+CiSpENDCBr4PDGK3HzSdyhEYNMiXEVeXhIahtafDBLKzLFVoOMD8DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6509863c4631e77a5a852f0044be646422fac310e9cebf6cf5e5de195320d5a8","last_reissued_at":"2026-06-02T00:03:14.364411Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T00:03:14.364411Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"BitsMoE: Efficient Spectral Energy-Guided Bit Allocation for MoE LLM Quantization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jiayu Zhao, Minhao Fan, Song Chen, Tianrui Ma, Weichen Liu, Wentao Ren, Zihan Teng","submitted_at":"2026-05-22T13:05:53Z","abstract_excerpt":"Mixture-of-Experts (MoE) large language models reduce per-token computation through sparse expert activation, but their deployment remains memory-intensive because all expert weights must be kept resident in memory. Existing MoE compression methods struggle in the ultra-low-bit regime: pruning irreversibly removes model capacity, while coarse-grained quantization fails to allocate bits according to heterogeneous expert and weight-direction importance. We propose BitsMoE, a spectral-energy-guided bit-allocation framework for MoE LLM quantization. BitsMoE decomposes each MoE layer by SVD into a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.00079","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/2606.00079/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":"2606.00079","created_at":"2026-06-02T00:03:14.364469+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.00079v1","created_at":"2026-06-02T00:03:14.364469+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.00079","created_at":"2026-06-02T00:03:14.364469+00:00"},{"alias_kind":"pith_short_12","alias_value":"MUEYMPCGGHTX","created_at":"2026-06-02T00:03:14.364469+00:00"},{"alias_kind":"pith_short_16","alias_value":"MUEYMPCGGHTXUWUF","created_at":"2026-06-02T00:03:14.364469+00:00"},{"alias_kind":"pith_short_8","alias_value":"MUEYMPCG","created_at":"2026-06-02T00:03:14.364469+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/MUEYMPCGGHTXUWUFF4AEJPTEMQ","json":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ.json","graph_json":"https://pith.science/api/pith-number/MUEYMPCGGHTXUWUFF4AEJPTEMQ/graph.json","events_json":"https://pith.science/api/pith-number/MUEYMPCGGHTXUWUFF4AEJPTEMQ/events.json","paper":"https://pith.science/paper/MUEYMPCG"},"agent_actions":{"view_html":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ","download_json":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ.json","view_paper":"https://pith.science/paper/MUEYMPCG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.00079&json=true","fetch_graph":"https://pith.science/api/pith-number/MUEYMPCGGHTXUWUFF4AEJPTEMQ/graph.json","fetch_events":"https://pith.science/api/pith-number/MUEYMPCGGHTXUWUFF4AEJPTEMQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ/action/storage_attestation","attest_author":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ/action/author_attestation","sign_citation":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ/action/citation_signature","submit_replication":"https://pith.science/pith/MUEYMPCGGHTXUWUFF4AEJPTEMQ/action/replication_record"}},"created_at":"2026-06-02T00:03:14.364469+00:00","updated_at":"2026-06-02T00:03:14.364469+00:00"}