{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:RGWD5HDFQ4EDDQ3S2ZT74PWHUB","short_pith_number":"pith:RGWD5HDF","schema_version":"1.0","canonical_sha256":"89ac3e9c65870831c372d667fe3ec7a05bf11ebeff3b0b0a63530748e7272371","source":{"kind":"arxiv","id":"2312.17238","version":1},"attestation_state":"computed","paper":{"title":"Fast Inference of Mixture-of-Experts Language Models with Offloading","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Artyom Eliseev, Denis Mazur","submitted_at":"2023-12-28T18:58:13Z","abstract_excerpt":"With the widespread adoption of Large Language Models (LLMs), many deep learning practitioners are looking for strategies of running these models more efficiently. One such strategy is to use sparse Mixture-of-Experts (MoE) - a type of model architectures where only a fraction of model layers are active for any given input. This property allows MoE-based language models to generate tokens faster than their dense counterparts, but it also increases model size due to having multiple experts. Unfortunately, this makes state-of-the-art MoE language models difficult to run without high-end GPUs. In"},"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":"2312.17238","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-12-28T18:58:13Z","cross_cats_sorted":["cs.AI","cs.DC"],"title_canon_sha256":"226648469ab81c9fc578e389d28db919aea1c0cde320a3ae42bb8429648b81eb","abstract_canon_sha256":"97257f25cfbb8d3f86629495107dd73d7f77e67255ba1992ad9d37f5c36fa5fd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:28:41.232553Z","signature_b64":"2yb5ydJmB2XzHsq4YjNuJf6fRdftnNOQOuW6wEiqE69RQOmGDL3zgYqMkRhwvkyyyw5TbCLc2CVNkOd3CUgOBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"89ac3e9c65870831c372d667fe3ec7a05bf11ebeff3b0b0a63530748e7272371","last_reissued_at":"2026-07-05T07:28:41.232127Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:28:41.232127Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Fast Inference of Mixture-of-Experts Language Models with Offloading","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.DC"],"primary_cat":"cs.LG","authors_text":"Artyom Eliseev, Denis Mazur","submitted_at":"2023-12-28T18:58:13Z","abstract_excerpt":"With the widespread adoption of Large Language Models (LLMs), many deep learning practitioners are looking for strategies of running these models more efficiently. One such strategy is to use sparse Mixture-of-Experts (MoE) - a type of model architectures where only a fraction of model layers are active for any given input. This property allows MoE-based language models to generate tokens faster than their dense counterparts, but it also increases model size due to having multiple experts. Unfortunately, this makes state-of-the-art MoE language models difficult to run without high-end GPUs. In"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.17238","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/2312.17238/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":"2312.17238","created_at":"2026-07-05T07:28:41.232179+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.17238v1","created_at":"2026-07-05T07:28:41.232179+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.17238","created_at":"2026-07-05T07:28:41.232179+00:00"},{"alias_kind":"pith_short_12","alias_value":"RGWD5HDFQ4ED","created_at":"2026-07-05T07:28:41.232179+00:00"},{"alias_kind":"pith_short_16","alias_value":"RGWD5HDFQ4EDDQ3S","created_at":"2026-07-05T07:28:41.232179+00:00"},{"alias_kind":"pith_short_8","alias_value":"RGWD5HDF","created_at":"2026-07-05T07:28:41.232179+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26287","citing_title":"GeMoE: Gating Entropy is All You Need for Uncertainty-aware Adaptive Routing in MoE-based Large Vision-Language Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21428","citing_title":"Does Mixture-of-Experts Actually Help Inference on Consumer and Edge Hardware? An Empirical Study","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21868","citing_title":"WiSP: A Working-Set View of Mixture-of-Experts Serving on Extremely Low-Resource Hardware","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29982","citing_title":"Beyond Uniform Experts: Cost-Aware Expert Execution for Efficient Multi-Device MoE Inference","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2601.21198","citing_title":"ZipMoE: Efficient On-Device MoE Serving via Lossless Compression and Cache-Affinity Scheduling","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2401.04088","citing_title":"Mixtral of Experts","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2411.08982","citing_title":"Lynx: Enabling Efficient MoE Inference through Dynamic Batch-Aware Expert Selection","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21427","citing_title":"PALS: Power-Aware LLM Serving for Mixture-of-Experts Models","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20179","citing_title":"TIDE: Efficient and Lossless MoE Diffusion LLM Inference with I/O-aware Expert Offload","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2510.05497","citing_title":"Patterns behind Chaos: Forecasting Data Movement for Efficient Large-Scale MoE LLM Inference","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23150","citing_title":"Scaling Multi-Node Mixture-of-Experts Inference Using Expert Activation Patterns","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05899","citing_title":"VisMMOE: Exploiting Visual-Expert Affinity for Efficient Visual-Language MoE Offloading","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2604.20156","citing_title":"Temporally Extended Mixture-of-Experts Models","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB","json":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB.json","graph_json":"https://pith.science/api/pith-number/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/graph.json","events_json":"https://pith.science/api/pith-number/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/events.json","paper":"https://pith.science/paper/RGWD5HDF"},"agent_actions":{"view_html":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB","download_json":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB.json","view_paper":"https://pith.science/paper/RGWD5HDF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.17238&json=true","fetch_graph":"https://pith.science/api/pith-number/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/graph.json","fetch_events":"https://pith.science/api/pith-number/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/action/storage_attestation","attest_author":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/action/author_attestation","sign_citation":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/action/citation_signature","submit_replication":"https://pith.science/pith/RGWD5HDFQ4EDDQ3S2ZT74PWHUB/action/replication_record"}},"created_at":"2026-07-05T07:28:41.232179+00:00","updated_at":"2026-07-05T07:28:41.232179+00:00"}