{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:6M7IFWBLV6R76V6V4D6ZCYCYK2","short_pith_number":"pith:6M7IFWBL","canonical_record":{"source":{"id":"2404.08335","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-12T09:01:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"d65d4a32d3f8944856d6c41bc4a6d359008b73f255ac1009df5cace4c265b236","abstract_canon_sha256":"71bc41c69d66f64d772086e93001e64e1d42a17206f97a55bdb0bb69ade7035b"},"schema_version":"1.0"},"canonical_sha256":"f33e82d82bafa3ff57d5e0fd916058569a5f0f91b11af1014403d96c1a37e4ed","source":{"kind":"arxiv","id":"2404.08335","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08335","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08335v2","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08335","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_12","alias_value":"6M7IFWBLV6R7","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_16","alias_value":"6M7IFWBLV6R76V6V","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_8","alias_value":"6M7IFWBL","created_at":"2026-07-05T10:46:53Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:6M7IFWBLV6R76V6V4D6ZCYCYK2","target":"record","payload":{"canonical_record":{"source":{"id":"2404.08335","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-12T09:01:14Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"d65d4a32d3f8944856d6c41bc4a6d359008b73f255ac1009df5cace4c265b236","abstract_canon_sha256":"71bc41c69d66f64d772086e93001e64e1d42a17206f97a55bdb0bb69ade7035b"},"schema_version":"1.0"},"canonical_sha256":"f33e82d82bafa3ff57d5e0fd916058569a5f0f91b11af1014403d96c1a37e4ed","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:46:53.828979Z","signature_b64":"h2AiWRfnSTj1ZIRc6rDa0ddamgzBQmMdIahlYxdXsiEBEvUQOy53v53bwxlIo5ZRTGktsy//72+8lQra5Ts2BA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f33e82d82bafa3ff57d5e0fd916058569a5f0f91b11af1014403d96c1a37e4ed","last_reissued_at":"2026-07-05T10:46:53.828459Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:46:53.828459Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2404.08335","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":"2026-07-05T10:46:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7R9ftMHZBCpgnZwqnIwkYij2kQCuUgmUtrjtoCrPSZvVMRa6WBkJAGagUfY2JO+xdA0XZ9FRVMGhkxZ65YjyBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T00:17:28.544498Z"},"content_sha256":"39309c2d2a6c730781c342d0e2969e805c426e3e5c86e230ffd00ea3e4a251a4","schema_version":"1.0","event_id":"sha256:39309c2d2a6c730781c342d0e2969e805c426e3e5c86e230ffd00ea3e4a251a4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:6M7IFWBLV6R76V6V4D6ZCYCYK2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Toward a Theory of Tokenization in LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Jiantao Jiao, Kannan Ramchandran, Nived Rajaraman","submitted_at":"2024-04-12T09:01:14Z","abstract_excerpt":"While there has been a large body of research attempting to circumvent tokenization for language modeling (Clark et al., 2022; Xue et al., 2022), the current consensus is that it is a necessary initial step for designing state-of-the-art performant language models. In this paper, we investigate tokenization from a theoretical point of view by studying the behavior of transformers on simple data generating processes. When trained on data drawn from certain simple $k^{\\text{th}}$-order Markov processes for $k > 1$, transformers exhibit a surprising phenomenon - in the absence of tokenization, th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08335","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/2404.08335/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":"2026-07-05T10:46:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/2zTVdI3qR7o/7QKKGwgKLL4MQ1/66fm9DHmy1G4OaEoLH6EWwHdXKbOkHxqVCnycYTHeRHLmlzLTLkyKyeoAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T00:17:28.545406Z"},"content_sha256":"b8926017b3d9d3f7494c9a62da203b83be086e59e10cf76c28d5317084bab724","schema_version":"1.0","event_id":"sha256:b8926017b3d9d3f7494c9a62da203b83be086e59e10cf76c28d5317084bab724"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/bundle.json","state_url":"https://pith.science/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/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":"2026-08-09T00:17:28Z","links":{"resolver":"https://pith.science/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2","bundle":"https://pith.science/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/bundle.json","state":"https://pith.science/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6M7IFWBLV6R76V6V4D6ZCYCYK2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:6M7IFWBLV6R76V6V4D6ZCYCYK2","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":"71bc41c69d66f64d772086e93001e64e1d42a17206f97a55bdb0bb69ade7035b","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-12T09:01:14Z","title_canon_sha256":"d65d4a32d3f8944856d6c41bc4a6d359008b73f255ac1009df5cace4c265b236"},"schema_version":"1.0","source":{"id":"2404.08335","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.08335","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"arxiv_version","alias_value":"2404.08335v2","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.08335","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_12","alias_value":"6M7IFWBLV6R7","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_16","alias_value":"6M7IFWBLV6R76V6V","created_at":"2026-07-05T10:46:53Z"},{"alias_kind":"pith_short_8","alias_value":"6M7IFWBL","created_at":"2026-07-05T10:46:53Z"}],"graph_snapshots":[{"event_id":"sha256:b8926017b3d9d3f7494c9a62da203b83be086e59e10cf76c28d5317084bab724","target":"graph","created_at":"2026-07-05T10:46:53Z","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/2404.08335/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While there has been a large body of research attempting to circumvent tokenization for language modeling (Clark et al., 2022; Xue et al., 2022), the current consensus is that it is a necessary initial step for designing state-of-the-art performant language models. In this paper, we investigate tokenization from a theoretical point of view by studying the behavior of transformers on simple data generating processes. When trained on data drawn from certain simple $k^{\\text{th}}$-order Markov processes for $k > 1$, transformers exhibit a surprising phenomenon - in the absence of tokenization, th","authors_text":"Jiantao Jiao, Kannan Ramchandran, Nived Rajaraman","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-12T09:01:14Z","title":"Toward a Theory of Tokenization in LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.08335","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":"sha256:39309c2d2a6c730781c342d0e2969e805c426e3e5c86e230ffd00ea3e4a251a4","target":"record","created_at":"2026-07-05T10:46:53Z","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":"71bc41c69d66f64d772086e93001e64e1d42a17206f97a55bdb0bb69ade7035b","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-12T09:01:14Z","title_canon_sha256":"d65d4a32d3f8944856d6c41bc4a6d359008b73f255ac1009df5cace4c265b236"},"schema_version":"1.0","source":{"id":"2404.08335","kind":"arxiv","version":2}},"canonical_sha256":"f33e82d82bafa3ff57d5e0fd916058569a5f0f91b11af1014403d96c1a37e4ed","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f33e82d82bafa3ff57d5e0fd916058569a5f0f91b11af1014403d96c1a37e4ed","first_computed_at":"2026-07-05T10:46:53.828459Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:46:53.828459Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"h2AiWRfnSTj1ZIRc6rDa0ddamgzBQmMdIahlYxdXsiEBEvUQOy53v53bwxlIo5ZRTGktsy//72+8lQra5Ts2BA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:46:53.828979Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.08335","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:39309c2d2a6c730781c342d0e2969e805c426e3e5c86e230ffd00ea3e4a251a4","sha256:b8926017b3d9d3f7494c9a62da203b83be086e59e10cf76c28d5317084bab724"],"state_sha256":"966b49d5ac5dc83300c012b8077c38b5600130257b79d9e22a6668efe61de268"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0Yupxzi7KYlpMwrQhEdZPEqVtaIpHv7/Z/7IUPMySbvljHOmoCljS7EcOpZdIV3ULsjwz4TCxTXCfsGxp9W5CQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T00:17:28.552502Z","bundle_sha256":"66da36a08b506767f34e1929213b31d45bc950a46650cdbfc6ef706209254bb2"}}