{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:3P4DLB5HMPAAHOEIFAM3XVD6PP","short_pith_number":"pith:3P4DLB5H","canonical_record":{"source":{"id":"2205.10770","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-22T07:43:50Z","cross_cats_sorted":[],"title_canon_sha256":"b999b3640c58bcd6edceafc9b61810cd7795b73f75b3f9e1f5a564a0e9a8ffb7","abstract_canon_sha256":"78b9601a6e16c16d0b731e599c099c29721a2fc656c0c8f39124db3caddde996"},"schema_version":"1.0"},"canonical_sha256":"dbf83587a763c003b8882819bbd47e7bf62b2b7f48330c957cfd384704f372bd","source":{"kind":"arxiv","id":"2205.10770","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.10770","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"arxiv_version","alias_value":"2205.10770v2","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.10770","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_12","alias_value":"3P4DLB5HMPAA","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_16","alias_value":"3P4DLB5HMPAAHOEI","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_8","alias_value":"3P4DLB5H","created_at":"2026-07-05T05:12:51Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:3P4DLB5HMPAAHOEIFAM3XVD6PP","target":"record","payload":{"canonical_record":{"source":{"id":"2205.10770","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-22T07:43:50Z","cross_cats_sorted":[],"title_canon_sha256":"b999b3640c58bcd6edceafc9b61810cd7795b73f75b3f9e1f5a564a0e9a8ffb7","abstract_canon_sha256":"78b9601a6e16c16d0b731e599c099c29721a2fc656c0c8f39124db3caddde996"},"schema_version":"1.0"},"canonical_sha256":"dbf83587a763c003b8882819bbd47e7bf62b2b7f48330c957cfd384704f372bd","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:12:51.564604Z","signature_b64":"djFo/HAea+fN+An0sNsZXACskcnKiberwZ8uFgC8JGUvMo776zvbbl+cty0zl5ZpBpo+3lfuC/J+0G41J3ymBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbf83587a763c003b8882819bbd47e7bf62b2b7f48330c957cfd384704f372bd","last_reissued_at":"2026-07-05T05:12:51.564132Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:12:51.564132Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2205.10770","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T05:12:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"IczQu6xuABe9sYlCXuUL8k/0MzpboXm4gzTc+jaZIOhFJ/M4+HSmgU60J8dFtaJ39aXlmFPjWj6IpxY1UuEOCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T21:12:05.713445Z"},"content_sha256":"2ea8aba08dc0be5b5ed01130e5442ee64d60fd90f2ee36691b0fdf02ad24eb9b","schema_version":"1.0","event_id":"sha256:2ea8aba08dc0be5b5ed01130e5442ee64d60fd90f2ee36691b0fdf02ad24eb9b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:3P4DLB5HMPAAHOEIFAM3XVD6PP","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Memorization Without Overfitting: Analyzing the Training Dynamics of Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aram H. Markosyan, Armen Aghajanyan, Kushal Tirumala, Luke Zettlemoyer","submitted_at":"2022-05-22T07:43:50Z","abstract_excerpt":"Despite their wide adoption, the underlying training and memorization dynamics of very large language models is not well understood. We empirically study exact memorization in causal and masked language modeling, across model sizes and throughout the training process. We measure the effects of dataset size, learning rate, and model size on memorization, finding that larger language models memorize training data faster across all settings. Surprisingly, we show that larger models can memorize a larger portion of the data before over-fitting and tend to forget less throughout the training proces"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.10770","kind":"arxiv","version":2},"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/2205.10770/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":"2026-07-05T05:12:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uHhb9WtttuSscoXl6WPVFvDBFLfho8jWRuJI00TG5JG0eSBgwNJYJjVC9dLvJfJkLtaH9TtR//sGfA+EEn1tBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T21:12:05.713814Z"},"content_sha256":"efe1e85866eef406ffeb34f3f8f5ba5d6b1161ac630491300c6b127be7618fcb","schema_version":"1.0","event_id":"sha256:efe1e85866eef406ffeb34f3f8f5ba5d6b1161ac630491300c6b127be7618fcb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/bundle.json","state_url":"https://pith.science/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/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":"2026-07-06T21:12:05Z","links":{"resolver":"https://pith.science/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP","bundle":"https://pith.science/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/bundle.json","state":"https://pith.science/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3P4DLB5HMPAAHOEIFAM3XVD6PP/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:3P4DLB5HMPAAHOEIFAM3XVD6PP","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":"78b9601a6e16c16d0b731e599c099c29721a2fc656c0c8f39124db3caddde996","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-22T07:43:50Z","title_canon_sha256":"b999b3640c58bcd6edceafc9b61810cd7795b73f75b3f9e1f5a564a0e9a8ffb7"},"schema_version":"1.0","source":{"id":"2205.10770","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2205.10770","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"arxiv_version","alias_value":"2205.10770v2","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.10770","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_12","alias_value":"3P4DLB5HMPAA","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_16","alias_value":"3P4DLB5HMPAAHOEI","created_at":"2026-07-05T05:12:51Z"},{"alias_kind":"pith_short_8","alias_value":"3P4DLB5H","created_at":"2026-07-05T05:12:51Z"}],"graph_snapshots":[{"event_id":"sha256:efe1e85866eef406ffeb34f3f8f5ba5d6b1161ac630491300c6b127be7618fcb","target":"graph","created_at":"2026-07-05T05:12:51Z","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/2205.10770/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Despite their wide adoption, the underlying training and memorization dynamics of very large language models is not well understood. We empirically study exact memorization in causal and masked language modeling, across model sizes and throughout the training process. We measure the effects of dataset size, learning rate, and model size on memorization, finding that larger language models memorize training data faster across all settings. Surprisingly, we show that larger models can memorize a larger portion of the data before over-fitting and tend to forget less throughout the training proces","authors_text":"Aram H. Markosyan, Armen Aghajanyan, Kushal Tirumala, Luke Zettlemoyer","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-22T07:43:50Z","title":"Memorization Without Overfitting: Analyzing the Training Dynamics of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.10770","kind":"arxiv","version":2},"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":"sha256:2ea8aba08dc0be5b5ed01130e5442ee64d60fd90f2ee36691b0fdf02ad24eb9b","target":"record","created_at":"2026-07-05T05:12:51Z","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":"78b9601a6e16c16d0b731e599c099c29721a2fc656c0c8f39124db3caddde996","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2022-05-22T07:43:50Z","title_canon_sha256":"b999b3640c58bcd6edceafc9b61810cd7795b73f75b3f9e1f5a564a0e9a8ffb7"},"schema_version":"1.0","source":{"id":"2205.10770","kind":"arxiv","version":2}},"canonical_sha256":"dbf83587a763c003b8882819bbd47e7bf62b2b7f48330c957cfd384704f372bd","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dbf83587a763c003b8882819bbd47e7bf62b2b7f48330c957cfd384704f372bd","first_computed_at":"2026-07-05T05:12:51.564132Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T05:12:51.564132Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"djFo/HAea+fN+An0sNsZXACskcnKiberwZ8uFgC8JGUvMo776zvbbl+cty0zl5ZpBpo+3lfuC/J+0G41J3ymBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T05:12:51.564604Z","signed_message":"canonical_sha256_bytes"},"source_id":"2205.10770","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2ea8aba08dc0be5b5ed01130e5442ee64d60fd90f2ee36691b0fdf02ad24eb9b","sha256:efe1e85866eef406ffeb34f3f8f5ba5d6b1161ac630491300c6b127be7618fcb"],"state_sha256":"edd907eab8a69013e3bb32dd01665bbfe710770fbf786a908521f7e54faa098c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LfSijQ4aMPRCC+AOrdyePqKqllMfzreMBbmwhftzGKGAHWG/au6G0bhjd00RJ+38wX5JsVA7MHjlToYzGNXOBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T21:12:05.715683Z","bundle_sha256":"56672c387f115b7ae54d08b0bed25f575380a6c5ba0f303d7d7b494801ec86ec"}}