{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:YQMS7XRYQVPPVG6EYGORCWT4S2","short_pith_number":"pith:YQMS7XRY","canonical_record":{"source":{"id":"2405.15523","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T13:05:05Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4c778aaab62ff0979971fe4ce7d5c60c42bc668bf883226169db8091ef986ca8","abstract_canon_sha256":"6ba88108747ed09e573150bbbda0206a76673f1780c01ff1771ea690cbce1270"},"schema_version":"1.0"},"canonical_sha256":"c4192fde38855efa9bc4c19d115a7c9689bfffe842fba2ad3841562bf8638edd","source":{"kind":"arxiv","id":"2405.15523","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.15523","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"arxiv_version","alias_value":"2405.15523v2","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.15523","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_12","alias_value":"YQMS7XRYQVPP","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_16","alias_value":"YQMS7XRYQVPPVG6E","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_8","alias_value":"YQMS7XRY","created_at":"2026-07-05T11:03:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:YQMS7XRYQVPPVG6EYGORCWT4S2","target":"record","payload":{"canonical_record":{"source":{"id":"2405.15523","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T13:05:05Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4c778aaab62ff0979971fe4ce7d5c60c42bc668bf883226169db8091ef986ca8","abstract_canon_sha256":"6ba88108747ed09e573150bbbda0206a76673f1780c01ff1771ea690cbce1270"},"schema_version":"1.0"},"canonical_sha256":"c4192fde38855efa9bc4c19d115a7c9689bfffe842fba2ad3841562bf8638edd","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:03:14.812250Z","signature_b64":"n4KwJEdj6HxzmYn4+gVKHctCIE8PAX7UpRQfXPraMDXF4HV9l6fzTmxqAlac8dXMZXfqWLN9EQDC0FbyV2qmCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c4192fde38855efa9bc4c19d115a7c9689bfffe842fba2ad3841562bf8638edd","last_reissued_at":"2026-07-05T11:03:14.811830Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:03:14.811830Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.15523","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":"2026-07-05T11:03:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"txPMx+qIdqPtJB7beasCFE59NVYJqN8H2MFBf6quwELmo2J2/+LVx20+FRPIS6ofQVesgXDKIX19h+Mwk3cXAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T00:49:06.499755Z"},"content_sha256":"1c1f4b9a4fc22def28ad50a97a7206b713dd8814a6fbabe51369c73401ba8506","schema_version":"1.0","event_id":"sha256:1c1f4b9a4fc22def28ad50a97a7206b713dd8814a6fbabe51369c73401ba8506"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:YQMS7XRYQVPPVG6EYGORCWT4S2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"The Mosaic Memory of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Igor Shilov, Matthieu Meeus, Yves-Alexandre de Montjoye","submitted_at":"2024-05-24T13:05:05Z","abstract_excerpt":"As Large Language Models (LLMs) become widely adopted, understanding how they learn from, and memorize, training data becomes crucial. Memorization in LLMs is widely assumed to only occur as a result of sequences being repeated in the training data. Instead, we show that LLMs memorize by assembling information from similar sequences, a phenomena we call mosaic memory. We show major LLMs to exhibit mosaic memory, with fuzzy duplicates contributing to memorization as much as 0.8 of an exact duplicate and even heavily modified sequences contributing substantially to memorization. Despite models d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.15523","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/2405.15523/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":"2026-07-05T11:03:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"FJ+3eYcvwj+WkPzeSqf8tVnoGXkKwR6xah8UopM+nK5mG/dILivL8ueJJNPSBiwMKFXgHz43xFksnMwg1kcADg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T00:49:06.500658Z"},"content_sha256":"747b1ea8045166cf6eeee8018b0652aa6906fc82626827a189e706b4c214d0b3","schema_version":"1.0","event_id":"sha256:747b1ea8045166cf6eeee8018b0652aa6906fc82626827a189e706b4c214d0b3"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/bundle.json","state_url":"https://pith.science/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/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":"2026-08-07T00:49:06Z","links":{"resolver":"https://pith.science/pith/YQMS7XRYQVPPVG6EYGORCWT4S2","bundle":"https://pith.science/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/bundle.json","state":"https://pith.science/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YQMS7XRYQVPPVG6EYGORCWT4S2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:YQMS7XRYQVPPVG6EYGORCWT4S2","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":"6ba88108747ed09e573150bbbda0206a76673f1780c01ff1771ea690cbce1270","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T13:05:05Z","title_canon_sha256":"4c778aaab62ff0979971fe4ce7d5c60c42bc668bf883226169db8091ef986ca8"},"schema_version":"1.0","source":{"id":"2405.15523","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.15523","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"arxiv_version","alias_value":"2405.15523v2","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.15523","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_12","alias_value":"YQMS7XRYQVPP","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_16","alias_value":"YQMS7XRYQVPPVG6E","created_at":"2026-07-05T11:03:14Z"},{"alias_kind":"pith_short_8","alias_value":"YQMS7XRY","created_at":"2026-07-05T11:03:14Z"}],"graph_snapshots":[{"event_id":"sha256:747b1ea8045166cf6eeee8018b0652aa6906fc82626827a189e706b4c214d0b3","target":"graph","created_at":"2026-07-05T11:03:14Z","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/2405.15523/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As Large Language Models (LLMs) become widely adopted, understanding how they learn from, and memorize, training data becomes crucial. Memorization in LLMs is widely assumed to only occur as a result of sequences being repeated in the training data. Instead, we show that LLMs memorize by assembling information from similar sequences, a phenomena we call mosaic memory. We show major LLMs to exhibit mosaic memory, with fuzzy duplicates contributing to memorization as much as 0.8 of an exact duplicate and even heavily modified sequences contributing substantially to memorization. Despite models d","authors_text":"Igor Shilov, Matthieu Meeus, Yves-Alexandre de Montjoye","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T13:05:05Z","title":"The Mosaic Memory of Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.15523","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":"sha256:1c1f4b9a4fc22def28ad50a97a7206b713dd8814a6fbabe51369c73401ba8506","target":"record","created_at":"2026-07-05T11:03:14Z","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":"6ba88108747ed09e573150bbbda0206a76673f1780c01ff1771ea690cbce1270","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T13:05:05Z","title_canon_sha256":"4c778aaab62ff0979971fe4ce7d5c60c42bc668bf883226169db8091ef986ca8"},"schema_version":"1.0","source":{"id":"2405.15523","kind":"arxiv","version":2}},"canonical_sha256":"c4192fde38855efa9bc4c19d115a7c9689bfffe842fba2ad3841562bf8638edd","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c4192fde38855efa9bc4c19d115a7c9689bfffe842fba2ad3841562bf8638edd","first_computed_at":"2026-07-05T11:03:14.811830Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:03:14.811830Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"n4KwJEdj6HxzmYn4+gVKHctCIE8PAX7UpRQfXPraMDXF4HV9l6fzTmxqAlac8dXMZXfqWLN9EQDC0FbyV2qmCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:03:14.812250Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.15523","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1c1f4b9a4fc22def28ad50a97a7206b713dd8814a6fbabe51369c73401ba8506","sha256:747b1ea8045166cf6eeee8018b0652aa6906fc82626827a189e706b4c214d0b3"],"state_sha256":"20de45c958cc96c453b401ceb79d1f7eafeee131c75d9220ff6541e939ab3467"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7LlOPePS9AAF0lL3UmD/vyGp23cz7rQJ48owPHATgvwmdpHC4oYI8hZR/yMLcO2zPhakANVRGSLdG3v4ywR2Ag==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T00:49:06.505782Z","bundle_sha256":"255bf41ef933947d26e3cbf8c7e83886b94489072c74bbeb91af505cdffdfbd0"}}