{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:E4S7OUFPNWR3J46L6TH6NOHKED","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"1a36b16ebadc40647dab2521c99b5cbc94ef2ac474328fbb465b856d42449363","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-13T08:22:19Z","title_canon_sha256":"6fbca8c139efd1bc41aee256e9a3d48c4d322e4aa42ec0a28b87e8b120d47d51"},"schema_version":"1.0","source":{"id":"2412.09952","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.09952","created_at":"2026-07-05T09:48:44Z"},{"alias_kind":"arxiv_version","alias_value":"2412.09952v1","created_at":"2026-07-05T09:48:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.09952","created_at":"2026-07-05T09:48:44Z"},{"alias_kind":"pith_short_12","alias_value":"E4S7OUFPNWR3","created_at":"2026-07-05T09:48:44Z"},{"alias_kind":"pith_short_16","alias_value":"E4S7OUFPNWR3J46L","created_at":"2026-07-05T09:48:44Z"},{"alias_kind":"pith_short_8","alias_value":"E4S7OUFP","created_at":"2026-07-05T09:48:44Z"}],"graph_snapshots":[{"event_id":"sha256:1eb3c7a22a934eeb5f64cf0f69de2f9a1ec07d7de3a3eb38f9decb95c7d17592","target":"graph","created_at":"2026-07-05T09:48:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2412.09952/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Scaling large language models (LLMs) significantly improves performance but comes with prohibitive computational costs. Mixture-of-Experts (MoE) models offer an efficient alternative, increasing capacity without a proportional rise in compute requirements. However, training MoE models from scratch poses challenges like overfitting and routing instability. We present an efficient training recipe leveraging pre-trained dense checkpoints, training an 8-Expert Top-2 MoE model from Llama 3-8B with less than $1\\%$ of typical pre-training compute. Our approach enhances downstream performance on acade","authors_text":"Aditya Vavre, Ashwath Aithal, Dennis Liu, Ethan He, June Yang, Nima Tajbakhsh, Zijie Yan","cross_cats":[],"headline":"","license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-13T08:22:19Z","title":"Llama 3 Meets MoE: Efficient Upcycling"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.09952","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:73229f6f6b61a2fe17e5e3f1ce9bf3588e0fca6a908c9144730843638db7da82","target":"record","created_at":"2026-07-05T09:48:44Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"1a36b16ebadc40647dab2521c99b5cbc94ef2ac474328fbb465b856d42449363","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-13T08:22:19Z","title_canon_sha256":"6fbca8c139efd1bc41aee256e9a3d48c4d322e4aa42ec0a28b87e8b120d47d51"},"schema_version":"1.0","source":{"id":"2412.09952","kind":"arxiv","version":1}},"canonical_sha256":"2725f750af6da3b4f3cbf4cfe6b8ea20d0e135ba139db73312e4b1fa6f55a920","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2725f750af6da3b4f3cbf4cfe6b8ea20d0e135ba139db73312e4b1fa6f55a920","first_computed_at":"2026-07-05T09:48:44.360797Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:48:44.360797Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"rlXpPSYL/KUz2AaMDooybOVH9hepN7goKUowwqJhjZcZJQZE1Z1/P5MfKZ5xdgF2WZ2qSURJSR4Jr33+aeQmCw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:48:44.361310Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.09952","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:73229f6f6b61a2fe17e5e3f1ce9bf3588e0fca6a908c9144730843638db7da82","sha256:1eb3c7a22a934eeb5f64cf0f69de2f9a1ec07d7de3a3eb38f9decb95c7d17592"],"state_sha256":"6610f868ea7dbcfd4be92f74bd7cec2e108390f1e5bba3e000871eaf212cc330"}