{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:RL65NZWJHR4ABNW6F46LT3SRQN","short_pith_number":"pith:RL65NZWJ","canonical_record":{"source":{"id":"2502.05967","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-09T17:31:09Z","cross_cats_sorted":[],"title_canon_sha256":"b9478ef27866e375475f698c343831c5f4dd3e41000183c57047d7345d48d5b5","abstract_canon_sha256":"2cdf0630959265057947340756351c3d11866b85ff124ecfa609ee92922363ad"},"schema_version":"1.0"},"canonical_sha256":"8afdd6e6c93c7800b6de2f3cb9ee518374d0ff2d5802b482fb97178cfb10888d","source":{"kind":"arxiv","id":"2502.05967","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.05967","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"arxiv_version","alias_value":"2502.05967v3","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.05967","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_12","alias_value":"RL65NZWJHR4A","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_16","alias_value":"RL65NZWJHR4ABNW6","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_8","alias_value":"RL65NZWJ","created_at":"2026-07-05T11:16:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:RL65NZWJHR4ABNW6F46LT3SRQN","target":"record","payload":{"canonical_record":{"source":{"id":"2502.05967","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-09T17:31:09Z","cross_cats_sorted":[],"title_canon_sha256":"b9478ef27866e375475f698c343831c5f4dd3e41000183c57047d7345d48d5b5","abstract_canon_sha256":"2cdf0630959265057947340756351c3d11866b85ff124ecfa609ee92922363ad"},"schema_version":"1.0"},"canonical_sha256":"8afdd6e6c93c7800b6de2f3cb9ee518374d0ff2d5802b482fb97178cfb10888d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:05.997923Z","signature_b64":"NnxXvxO+M61qv2fe94Z3fIszUBagQwuO/AodBx3YbumE3qYWrM1+T8UUI55E/CY3ToSJ8RjmvEK2aeyQgVVIAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8afdd6e6c93c7800b6de2f3cb9ee518374d0ff2d5802b482fb97178cfb10888d","last_reissued_at":"2026-07-05T11:16:05.997419Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:05.997419Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2502.05967","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:16:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1BmEacYv9Dm23BFd6AzRkCEHVyeLdZ2CIpYMXsUmPSarcP9J3kUDctorswemdh8En2QisyQfcKB0PW5rII41CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T17:27:32.230883Z"},"content_sha256":"672562d2af4e4530e6f76aab7857f0b514b0f1e24f4e77d06379060cbfe83172","schema_version":"1.0","event_id":"sha256:672562d2af4e4530e6f76aab7857f0b514b0f1e24f4e77d06379060cbfe83172"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:RL65NZWJHR4ABNW6F46LT3SRQN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"$\\mu$nit Scaling: Simple and Scalable FP8 LLM Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Abhay Gupta, Davis Blalock, Mansheej Paul, Saaketh Narayan","submitted_at":"2025-02-09T17:31:09Z","abstract_excerpt":"Large Language Model training with 8-bit floating point (FP8) formats promises significant efficiency improvements, but reduced numerical precision makes training challenging. It is currently possible to train in FP8 only if one is willing to tune various hyperparameters, reduce model scale, or accept the overhead of computing dynamic scale factors. We demonstrate simple, scalable FP8 training that requires no dynamic scaling factors or special hyperparameters, even at large model sizes. Our method, $\\mu$nit Scaling ($\\mu$S), also enables simple hyperparameter transfer across model widths, mat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.05967","kind":"arxiv","version":3},"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/2502.05967/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:16:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iVDezesefu/CX0YsKnFWnlrMu6qB8qC0EmWqhXAX4cYlOU1C8W412+omp9AD/UC1iAq2mv8z8m/J6Eo9TvUmDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T17:27:32.231400Z"},"content_sha256":"b5784ff4829e7cae1f4299e899d09e560114790b94e250768bfc52807fa194ca","schema_version":"1.0","event_id":"sha256:b5784ff4829e7cae1f4299e899d09e560114790b94e250768bfc52807fa194ca"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/RL65NZWJHR4ABNW6F46LT3SRQN/bundle.json","state_url":"https://pith.science/pith/RL65NZWJHR4ABNW6F46LT3SRQN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/RL65NZWJHR4ABNW6F46LT3SRQN/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-09T17:27:32Z","links":{"resolver":"https://pith.science/pith/RL65NZWJHR4ABNW6F46LT3SRQN","bundle":"https://pith.science/pith/RL65NZWJHR4ABNW6F46LT3SRQN/bundle.json","state":"https://pith.science/pith/RL65NZWJHR4ABNW6F46LT3SRQN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/RL65NZWJHR4ABNW6F46LT3SRQN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:RL65NZWJHR4ABNW6F46LT3SRQN","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":"2cdf0630959265057947340756351c3d11866b85ff124ecfa609ee92922363ad","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-09T17:31:09Z","title_canon_sha256":"b9478ef27866e375475f698c343831c5f4dd3e41000183c57047d7345d48d5b5"},"schema_version":"1.0","source":{"id":"2502.05967","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2502.05967","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"arxiv_version","alias_value":"2502.05967v3","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.05967","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_12","alias_value":"RL65NZWJHR4A","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_16","alias_value":"RL65NZWJHR4ABNW6","created_at":"2026-07-05T11:16:05Z"},{"alias_kind":"pith_short_8","alias_value":"RL65NZWJ","created_at":"2026-07-05T11:16:05Z"}],"graph_snapshots":[{"event_id":"sha256:b5784ff4829e7cae1f4299e899d09e560114790b94e250768bfc52807fa194ca","target":"graph","created_at":"2026-07-05T11:16:05Z","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/2502.05967/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Model training with 8-bit floating point (FP8) formats promises significant efficiency improvements, but reduced numerical precision makes training challenging. It is currently possible to train in FP8 only if one is willing to tune various hyperparameters, reduce model scale, or accept the overhead of computing dynamic scale factors. We demonstrate simple, scalable FP8 training that requires no dynamic scaling factors or special hyperparameters, even at large model sizes. Our method, $\\mu$nit Scaling ($\\mu$S), also enables simple hyperparameter transfer across model widths, mat","authors_text":"Abhay Gupta, Davis Blalock, Mansheej Paul, Saaketh Narayan","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-09T17:31:09Z","title":"$\\mu$nit Scaling: Simple and Scalable FP8 LLM Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.05967","kind":"arxiv","version":3},"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":"sha256:672562d2af4e4530e6f76aab7857f0b514b0f1e24f4e77d06379060cbfe83172","target":"record","created_at":"2026-07-05T11:16:05Z","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":"2cdf0630959265057947340756351c3d11866b85ff124ecfa609ee92922363ad","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-09T17:31:09Z","title_canon_sha256":"b9478ef27866e375475f698c343831c5f4dd3e41000183c57047d7345d48d5b5"},"schema_version":"1.0","source":{"id":"2502.05967","kind":"arxiv","version":3}},"canonical_sha256":"8afdd6e6c93c7800b6de2f3cb9ee518374d0ff2d5802b482fb97178cfb10888d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8afdd6e6c93c7800b6de2f3cb9ee518374d0ff2d5802b482fb97178cfb10888d","first_computed_at":"2026-07-05T11:16:05.997419Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:16:05.997419Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NnxXvxO+M61qv2fe94Z3fIszUBagQwuO/AodBx3YbumE3qYWrM1+T8UUI55E/CY3ToSJ8RjmvEK2aeyQgVVIAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:16:05.997923Z","signed_message":"canonical_sha256_bytes"},"source_id":"2502.05967","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:672562d2af4e4530e6f76aab7857f0b514b0f1e24f4e77d06379060cbfe83172","sha256:b5784ff4829e7cae1f4299e899d09e560114790b94e250768bfc52807fa194ca"],"state_sha256":"d55e96f363e5c95bac881c654d8dc7f2e2439383374c70feaa94e459194822c4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KytPqFJjlm25m4WIC/BJkdl1w3wNT8RFdRx1EGq0Kn7SnqT5mDYrlFcV70mzTofaJCQ+CeHWuZBNFczMPSc/BQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T17:27:32.236212Z","bundle_sha256":"1c01facc942609a1553c60e06cb2141d7e4a6bb7d59c0b5f3e9e22d31e0fd3c9"}}