{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:EVVFELZD5IEJRVDIH7U2OJJM5S","short_pith_number":"pith:EVVFELZD","canonical_record":{"source":{"id":"2508.00180","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-31T21:49:20Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"a23b906780ba102c07e8570516498d9a9ca71fa88825d05aa56721c7e6204979","abstract_canon_sha256":"e94cfec2ff26d6eaa133affb58cef92e2c451f52eb5dc20b481b683fc4d3a073"},"schema_version":"1.0"},"canonical_sha256":"256a522f23ea0898d4683fe9a7252cec80efac3b197880e4e7b6c5f3142583b2","source":{"kind":"arxiv","id":"2508.00180","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.00180","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"arxiv_version","alias_value":"2508.00180v1","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.00180","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_12","alias_value":"EVVFELZD5IEJ","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_16","alias_value":"EVVFELZD5IEJRVDI","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_8","alias_value":"EVVFELZD","created_at":"2026-07-05T11:46:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:EVVFELZD5IEJRVDIH7U2OJJM5S","target":"record","payload":{"canonical_record":{"source":{"id":"2508.00180","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-31T21:49:20Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"a23b906780ba102c07e8570516498d9a9ca71fa88825d05aa56721c7e6204979","abstract_canon_sha256":"e94cfec2ff26d6eaa133affb58cef92e2c451f52eb5dc20b481b683fc4d3a073"},"schema_version":"1.0"},"canonical_sha256":"256a522f23ea0898d4683fe9a7252cec80efac3b197880e4e7b6c5f3142583b2","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:46:50.513469Z","signature_b64":"7pqXKwdQ25GadKlrjHmWX3E+kQ4ScLuGOlPngvH6RpkqJADw5rdbqsa5vTXQNznsjhiNO7Nf0L/qrNL+/r10DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"256a522f23ea0898d4683fe9a7252cec80efac3b197880e4e7b6c5f3142583b2","last_reissued_at":"2026-07-05T11:46:50.512695Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:46:50.512695Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2508.00180","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:46:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HVTSeTcy3s+yxADQm2DcgsRzW1Vvj0nLxTp2pweOhQ883D9QFhGeq5KQVQHcuBwE3SnIEDBU08dEHq9iCHaDAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T16:36:46.627491Z"},"content_sha256":"8488b1b3606f1f1bffd1b00f53e88cfa1d93e01eefe44b6c64fdda6aee900f84","schema_version":"1.0","event_id":"sha256:8488b1b3606f1f1bffd1b00f53e88cfa1d93e01eefe44b6c64fdda6aee900f84"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:EVVFELZD5IEJRVDIH7U2OJJM5S","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"EMA Without the Lag: Bias-Corrected Iterate Averaging Schemes","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Adam Block, Cyril Zhang","submitted_at":"2025-07-31T21:49:20Z","abstract_excerpt":"Stochasticity in language model fine-tuning, often caused by the small batch sizes typically used in this regime, can destabilize training by introducing large oscillations in generation quality. A popular approach to mitigating this instability is to take an Exponential moving average (EMA) of weights throughout training. While EMA reduces stochasticity, thereby smoothing training, the introduction of bias from old iterates often creates a lag in optimization relative to vanilla training. In this work, we propose the Bias-Corrected Exponential Moving Average (BEMA), a simple and practical aug"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.00180","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/2508.00180/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":"2026-07-05T11:46:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wpr8apYIce6Olf//JDQOPq0GVBjtRanVdIJywFWkMk7g97ItTN3wjlzxor5hO+OkrsI27oVzFT/BeEq39nC9Cw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T16:36:46.628050Z"},"content_sha256":"6e543dc24d84020053ab2ec9caccc3ca71ce509d789a2532949bb053017df58c","schema_version":"1.0","event_id":"sha256:6e543dc24d84020053ab2ec9caccc3ca71ce509d789a2532949bb053017df58c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/bundle.json","state_url":"https://pith.science/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/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":"2026-08-06T16:36:46Z","links":{"resolver":"https://pith.science/pith/EVVFELZD5IEJRVDIH7U2OJJM5S","bundle":"https://pith.science/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/bundle.json","state":"https://pith.science/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EVVFELZD5IEJRVDIH7U2OJJM5S/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:EVVFELZD5IEJRVDIH7U2OJJM5S","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":"e94cfec2ff26d6eaa133affb58cef92e2c451f52eb5dc20b481b683fc4d3a073","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-31T21:49:20Z","title_canon_sha256":"a23b906780ba102c07e8570516498d9a9ca71fa88825d05aa56721c7e6204979"},"schema_version":"1.0","source":{"id":"2508.00180","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.00180","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"arxiv_version","alias_value":"2508.00180v1","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.00180","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_12","alias_value":"EVVFELZD5IEJ","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_16","alias_value":"EVVFELZD5IEJRVDI","created_at":"2026-07-05T11:46:50Z"},{"alias_kind":"pith_short_8","alias_value":"EVVFELZD","created_at":"2026-07-05T11:46:50Z"}],"graph_snapshots":[{"event_id":"sha256:6e543dc24d84020053ab2ec9caccc3ca71ce509d789a2532949bb053017df58c","target":"graph","created_at":"2026-07-05T11:46:50Z","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/2508.00180/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Stochasticity in language model fine-tuning, often caused by the small batch sizes typically used in this regime, can destabilize training by introducing large oscillations in generation quality. A popular approach to mitigating this instability is to take an Exponential moving average (EMA) of weights throughout training. While EMA reduces stochasticity, thereby smoothing training, the introduction of bias from old iterates often creates a lag in optimization relative to vanilla training. In this work, we propose the Bias-Corrected Exponential Moving Average (BEMA), a simple and practical aug","authors_text":"Adam Block, Cyril Zhang","cross_cats":["cs.AI","stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-31T21:49:20Z","title":"EMA Without the Lag: Bias-Corrected Iterate Averaging Schemes"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.00180","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":"sha256:8488b1b3606f1f1bffd1b00f53e88cfa1d93e01eefe44b6c64fdda6aee900f84","target":"record","created_at":"2026-07-05T11:46:50Z","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":"e94cfec2ff26d6eaa133affb58cef92e2c451f52eb5dc20b481b683fc4d3a073","cross_cats_sorted":["cs.AI","stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-31T21:49:20Z","title_canon_sha256":"a23b906780ba102c07e8570516498d9a9ca71fa88825d05aa56721c7e6204979"},"schema_version":"1.0","source":{"id":"2508.00180","kind":"arxiv","version":1}},"canonical_sha256":"256a522f23ea0898d4683fe9a7252cec80efac3b197880e4e7b6c5f3142583b2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"256a522f23ea0898d4683fe9a7252cec80efac3b197880e4e7b6c5f3142583b2","first_computed_at":"2026-07-05T11:46:50.512695Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:46:50.512695Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7pqXKwdQ25GadKlrjHmWX3E+kQ4ScLuGOlPngvH6RpkqJADw5rdbqsa5vTXQNznsjhiNO7Nf0L/qrNL+/r10DQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:46:50.513469Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.00180","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:8488b1b3606f1f1bffd1b00f53e88cfa1d93e01eefe44b6c64fdda6aee900f84","sha256:6e543dc24d84020053ab2ec9caccc3ca71ce509d789a2532949bb053017df58c"],"state_sha256":"8ea6b0e9c4d35d38aa1cd326cff29a97bf4ac4df3243fb943118c8facfb9048f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UoivU47usGe9wYpqagE8C9H0ph6mkIP1NMl0YQN5a1m+sII4gI1YofF4n1s9FJq/5V/rZjXN9/F98MsY6ggQDw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T16:36:46.632856Z","bundle_sha256":"d6758a83140fc2e964d64aef80d50558c7e17861a0e66034781651b0cfcccf65"}}